You are on page 1of 312

INFLATION-LINKED RESEARCH

March 2010

GLOBAL INFLATION-LINKED PRODUCTS
A USER’S GUIDE

PLEASE SEE ANALYST CERTIFICATIONS AND IMPORTANT DISCLOSURES ON THE LAST PAGE.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Introduction
Alan James +44 (0) 20 7773 2238 alan.james@barcap.com

Global inflation-linked markets have been through an extremely difficult period in the two years since the last edition of this guide was published in February 2008, but in most countries they have emerged from this turmoil stronger than ever. The economic shock accompanying the financial market meltdown from September 2008 prompted a rapid fall in inflation globally, with many countries experiencing deflation for the first time in 70 years, but this was mild compared to the expected fall implied by the cheapening of inflationlinked bonds. The extreme deleveraging phase that engulfed almost all financial markets included the majority of off-benchmark investors in inflation-linked bonds being stopped out of their positions. The impact was greater due to the acceleration in inflation in the preceding year, which had led to more investors holding inflation-linked bonds versus nominals as protection against inflation. After the extremes of Q4 08, the first half of 2009 was all about absorbing the float sold by these breakeven investors. During this period most governments were unable to increase inflation-linked issuance despite their much higher funding needs, but by the second half of 2009, as the markets normalised, supply began to increase and 2010 is set to see significantly more supply in government inflation-linked bonds than ever before. Inflation swap markets became dislocated from inflation-linked bond valuations during late 2008, generally cheapening less than bonds but with a commensurate drop in liquidity. Despite this illiquidity swap valuations were almost certainly less distorted during this phase than deleveraging bond markets and were also probably more representative of inflation expectations in the first half of 2009 as central banks began to engage in various forms of quantitative easing. As asset swap activity in government inflation-linked bonds resumed, this enabled relative valuations between inflation-linked bonds and swaps to move back towards their historical normal ranges. Inflation volatility markets suffered notably more than inflation swaps and liquidity still has not recovered in early 2010, despite markets once again being relatively well defined. While inflation uncertainty is notably higher in the aftermath of extreme fiscal injections and central bank quantitative easing programmes, implied volatility remains an order of magnitude above both realised market and actual inflation volatility. A greater share of government inflation-linked bonds is held by long-term investors than before the financial crisis. In 2009, significantly more of the allocations into the asset class were in index form than in preceding years and in this edition we significantly expand our discussion of both benchmark indices and index strategies. While several markets remain dominated by investors hedging their inflation-linked liabilities, either using bonds or derivative overlay strategies, we argue that unless such investors drive valuations to unrealistic levels then inflation-linked assets ought to have a place in both bond and diversified portfolios. The trend in recent years towards diversifying both broad and inflation focused portfolios into emerging market inflation-linked assets was only temporarily interrupted by the global financial dislocations. We expect this diversification trend to continue, which should help improve the depth of inflation-linked markets in emerging economies. Even so the developments in the last two years, particularly in countries such as Turkey where extreme dislocation was followed by a rapid recovery and growth of the market, highlight that the success of inflation-linked markets depends on domestic demand for the product. As with previous versions of this guide, this edition aims to provide a solid base for those looking to understand inflation-linked bonds and derivatives while also providing an invaluable reference source for market participants. The first section of the guide focuses on

15 March 2010

1

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

the products within the inflation-linked universe, the second covers the features of each individual market. This is followed by the section on inflation-linked indices and tradable inflation-linked index products, before addressing important themes for the asset class, including portfolio concepts, valuation themes and both the macro economic and component drivers of inflation. Government inflation-linked bonds tend to have a similar structure, where principal and income are adjusted for changes in the relevant consumer price index between issue and cash flow payment, subject to an indexation lag. In this guide, such bonds are referred to as “inflation-linked”, “inflation-indexed”, “index-linked”, “real return bonds”, “inflation bonds”, “inflation-protected securities”, or just “linkers”, with these terms used interchangeably. Other bonds with alternative links to inflation are referred to as structured notes or bonds, but the type of structure can vary markedly. The variety in inflation derivatives is even wider. While calculations vary between countries and products, this publication tries to outline the concepts behind them and their differences as well as providing the required rigorous detail. With global monetary and fiscal policy in recent years having been aggressively set out to avoid deflation, there are clear risks that this results in higher than intended inflation later in this economic cycle. Indeed as highlighted in the concluding article, ‘Run inflation run’, the economic benefits of a period of temporarily higher inflation may be considerable, which may mean that policymakers are likely to tolerate such a period. It may not be imminent however, with core inflation likely to stay subdued for some time as argued in ‘Why output gaps still matter for inflation’. Given this backdrop of heightened uncertainty, inflationlinked products are likely to continue to develop as a vital asset class for investors and policymakers alike.

15 March 2010

2

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

TABLE OF CONTENTS

INFLATION PRODUCTS

5

Inflation-linked bonds ............................................................................................................................... 6 Inflation derivatives..................................................................................................................................20 INFLATION MARKETS 39

US.................................................................................................................................................................40 Euro area ....................................................................................................................................................52 UK.................................................................................................................................................................73 Japan............................................................................................................................................................88 Canada........................................................................................................................................................97 Sweden .................................................................................................................................................... 101 Australia .................................................................................................................................................. 110 Brazil ......................................................................................................................................................... 118 Mexico ..................................................................................................................................................... 125 Argentina................................................................................................................................................. 128 Chile.......................................................................................................................................................... 133 Colombia ................................................................................................................................................. 137 Uruguay ................................................................................................................................................... 139 Israel ......................................................................................................................................................... 140 Turkey ...................................................................................................................................................... 143 South Africa............................................................................................................................................ 146 Poland ...................................................................................................................................................... 151 Iceland...................................................................................................................................................... 153 South Korea ............................................................................................................................................ 155 INFLATION INDICES 158

Barclays Capital inflation-linked indices........................................................................................... 159 Inflation-linked benchmark indices................................................................................................... 161 World Government Inflation-Linked Bond Index ........................................................................... 164 Euro Inflation-Linked Bond Index ...................................................................................................... 167 Emerging Markets Government Inflation-Linked Bond Index .................................................... 170 Sterling Inflation-Linked Bond Index ................................................................................................ 172 Global Inflation-Linked Bond Index (Series-L)................................................................................ 175 Comparator bond indices.................................................................................................................... 177 Benchmark Inflation-Linked bond index calculation and methodology .................................. 179 Barclays Capital Inflation–Linked Index Products .......................................................................... 190 Inflation Swap Index family................................................................................................................. 191 TWIST Index: Treasury-weighted inflation swap exposure across the curve ......................... 196 Barclays Capital US breakeven inflation benchmark indices ....................................................... 200 The INSTEP Index .................................................................................................................................. 204 CAD INSPIRE........................................................................................................................................... 206 AUD INSPIRE........................................................................................................................................... 209 Barclays Capital US TIPS Real Income Index family....................................................................... 212 AIMS Index.............................................................................................................................................. 214

15 March 2010

3

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

INFLATION THEMES

218

Breakeven inflation and the relationship between real and nominal yields............................. 219 Real return assets .................................................................................................................................. 226 Linkers in a portfolio context.............................................................................................................. 235 Modelling real yields and breakevens............................................................................................... 243 Seasonality and inflation markets ..................................................................................................... 249 Forwards and their usage.................................................................................................................... 261 Inflation volatility and deflation floors .............................................................................................. 268 Comparing baskets: The elements driving inflation...................................................................... 274 Why output gaps still matter for inflation ....................................................................................... 282 Run inflation run.................................................................................................................................... 292 APPENDICES 299

Key information sources...................................................................................................................... 300 Summary sovereign table.................................................................................................................... 302 Real Yield Histories................................................................................................................................ 304 Breakeven Inflation Histories .............................................................................................................. 307

15 March 2010

4

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

INFLATION PRODUCTS

15 March 2010

5

the cash flows on a linker will tend to be back-ended compared to a nominal bond. later coupons will be bigger than initial ones and the principal repayment at maturity will be above par. Government linkers typically pay coupons on a variable principal which increases or decreases with the price index over the life of the bond. but have similar. a linker is quoted in real yield. has become the standard quotation format for linkers. features. Assuming that there is inflation. with even the UK switching to this framework for all new bonds. The Canadian model has become the standard one for inflation-linked bonds.james@barcap. Thirteen of the 20 largest countries in the world weighted by GDP have inflation-linked markets.sooben @barcap. introduced by Canada in 1990. though there are countries such as Chile and Israel where there is more corporate debt that is inflation-linked than nominal. Corporate issuance of linkers where the inflation exposure is not swapped out has been rare outside UK among the developed markets. although non-standardised. and just like a nominal bond is quoted in nominal yield.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation-linked bonds Alan James +44 (0) 20 7773 2238 alan.com Inflation-linked bonds are now an established segment of their respective local bond markets for most issuers. with its innovative simplicity being that all calculations are done in “real” terms. this yield is therefore “real” in the sense of true purchasing power.com Khrishnamoorthy Sooben +44 (0) 20 777 37514 khrishnamoorthy. Linkers offer a “real” yield which is net of inflation. The Canadian model. without any assumption to be made about the future path of the price index as is necessary for instance in the traditional UK model. and the principal repayment at maturity is also adjusted for the change in the price index. Despite the financial dislocations in the second half of 2008 significantly affecting most government inflation-linked markets. Latin American markets have not adopted the Canadian format. although with varying degrees of activity. unadjusted for past inflation accrual AC is the real accrued coupon C is the real coupon RY is the real yield tm is the bond’s maturity 15 March 2010 6 . It is then straightforward to relate the real yield to the linker’s price using the generic present value formulation which relates the price of a nominal bond to its nominal yield: P DU = P CU + AC =∑ t =t 0 tm C 100 + t (1 + RY ) (1 + RY ) tm Where: PDU is the dirty price. One aspect to note is that for a particular issuer and maturity. the back-ended nature of linker cash flows effectively means that there is a bigger credit risk involved in holding an inflation-linked bond than a nominal one. The Canadian model The cash flows of an inflation-linked bond are dependent on the evolution of its linking price index. unadjusted for past inflation accrual PCU is the clean quoted price. The distinguishing feature between a nominal and an inflation-linked bond is the yield metric used. all G7 countries still run inflation-linked debt programmes alongside their nominal ones.

For coupons paid. and the Index Ratio for the trade settlement date. Canadian-style linkers are quoted in clean price terms (PCU). In many countries with the Canadian model. The real accrued coupon is calculated in the same way as for a nominal bond. while Sweden introduced floors for bonds issued from 15 March 2010 7 . The settlement price of a Canadian model bond is then determined as the product of the price PDU. the price index at maturity is lower than the Base Index. so in this case the Base Index is calculated from the same day. in the euro area or US. the Reference Index for 1 February 2010 is the CPI value of November 2009. the principal value at redemption is floored at par even if the index ratio is below 1. and likewise for the redemption amount. For example. This deflation floor only applies to the principal value. the (real) coupon rate is multiplied by the Index Ratio. bonds are issued with short first coupons with accrual calculated from what would have been the previous coupon date had the bond existed. including the final coupon. French supply the following year thus opted for the floor too. a final adjustment therefore needs to be applied to arrive at the true or monetary settlement price. and the sum (price PDU above) is multiplied by the Index ratio to arrive at the cash settlement amount.Barclays Capital | Global Inflation-Linked Products – A User’s Guide This price PDU does not include inflation which has already accrued on the bond’s principal value. which is also rounded to five decimal places. net of accrued interest. The inclusion of a deflation floor proved a popular addition to the Canadian model when the US Treasury first issued inflation-linked bonds (TIPS) with this feature. as have all subsequent euro issuers. For instance. but this does not necessarily coincide with the initial settlement date for the bond. the reference for the first day of a month is equal to the price index value three months back. but not Canada itself. which is also the date from which interest accrues. The inflation accrual adjustment reflects the cumulated change from the Base Index to the Reference Index value on the trade settlement date and is computed as an Index Ratio: Index Ratio = Reference Indext/Base Index In the Canadian model. coupons can be paid off a principal amount that is less than par. ie. Reference indices for intervening days are calculated by a linear interpolation on a standard actual/actual day count basis: Index = CPIm − 3 + (t − 1) × (CPIm − 2 − CPIm − 3) Dm Where: CPIm-2 is the price index for month m-2 CPIm-3 is the price index for month m-3 Dm is the number of days in month m m is the month in which settlement takes place t is the day of the month on which settlement takes place Both the Reference Index and the Base Index are truncated to six decimal places and then rounded to five decimal places before computing the Index Ratio. Typically. The quoted clean price and the real accrued coupon are added. in France. the daily Reference Index is based on a lag from the current settlement date. The “Base Index” is the price index reference value (Reference Index) on the date from which inflation accrues. ie. and most other issuers have followed suit. which is non-adjusted for accrued inflation.

In most Canadian-style markets. On 15 January 2010. In more precise terms. The Swedish market also trades almost entirely on a real yield basis. The absence of floors in JGBis is a significant reason why the market became extremely distressed in 2008/2009 as deflation fears increased. On the other hand. for a nominal bond. in linkers. However. The carry of a bond position is determined by the income provided by the bond and the cost of funding that position in the repo market. Japan chose not to include floors on its inflation-linked bonds. consistent with nominal bonds. when real yields were pushed well above nominals. recently issued TIPS only very briefly traded through the level at which their floor ensured a superior hold to maturity return to the nominal treasury curve. Therefore. in Japan. as the value of a floor would have been much higher than in other markets where deflation has been very rare: the interpretation of the value of bonds would have been much more difficult. extreme carry is seen only when the difference between the bond yield and the repo cost is significant. coupons are paid semi-annually. day count conventions for inflation accrual are based on a linear rate that assumes 30 days in each month. Sweden. the income is determined by two elements: the real yield but also the inflation which accrues over the holding period. an exact forward real yield or carry can therefore be calculated only up to the furthest settlement date for which the Reference index value is known. resulting in a less smooth accretion of inflation across a month. with quoted prices including inflation uplift. an exact carry was then calculable only up to the 1 February 2010 settlement date. While the shorter end of the US TIPS market had a similar supply/demand dislocation in Q4 08. the three-month lag is to the tenth of the month rather than the first. the latest euro HICPx value known was for November 2009. for example. which implies that the one-month real yield carry from one month to another can be very unstable. the inflation accrual and the change in the quoted clean price. carry is driven by the difference between the nominal yield and repo rate. given the seasonality in inflation indices. For nominal bonds. Japan announced that it was examining the possibility of issuing floored JGBIs. In 2009. the furthest settlement date for an exact carry computation then shifted to 1 15 March 2010 8 . The UK decided not to include a floor when it started issuing its new format bonds and nor did South Korea. extreme carry can also be induced by a very low or high inflation accrual over a month because of seasonality or an exceptional shock on prices. as Canadian conventions were only adopted after the market began. much more than in nominal space. where annual payments are the norm for both linker and nominal issues. with the release of the December 2009 inflation data. In Sweden. This inflation accrual element is inherently volatile from one month to another. on 10 January 2010. when the yield curve is extremely steep or very inverted. The return over the holding period has three components: any coupon payments and/or coupon accrual. and in South Africa the lag is four months. it is defined as the difference between the forward real yield and the current or spot real yield. For an inflation-linked bond. Japan also rounds to three decimal places rather than five when calculating the Index Ratio. carry on a linker is defined as the required change in the real yield over the holding period considered such that the return from holding the bond is equal to the repo cost. Germany and Greece. For example. ie. which means discontinuous accretion at month-end for months that are not this length. with the exception of France. Some countries have slightly different inflation lags. In common terminology. This means that for a euro HICPx inflation-linked bond. Given that the return depends on the inflation accrual. Carry in the Canadian model Carry is an important element for inflation-linked bonds. Also the magnitude of real yield carry can be very high in linkers.Barclays Capital | Global Inflation-Linked Products – A User’s Guide 1999.

but a fair assessment of its valuation should take into account the positive carry over the month. unadjusted for past inflation accrual. and can be a known or expected ratio Repo is the repo rate used to finance the position and is assumed to be constant In T0. but the basis point real yield carry in a given month can be approximated by: [M/M inflation accretion + (real yield– repo rate)/12]/forward modified duration of bond Shorter duration bonds thus have a notably higher sensitivity to inflation accretion than longer issues. at the start of a month during which the inflation accrual is known to be very high. We calculate the carry from date T0 to the forward date T2. consider a flat real yield curve of 2%. although it is only possible to lock in such forward levels if the future inflation can be traded. As discussed later in this guide. For a par bond with a duration of 10. The inflation lag in the Canadian model means that the valuation of the spot real yield may be “optically” distorted by future inflation accrual which is already known but which has not yet accrued on the bond. Looking at forward real yields offers a more realistic gauge of value ahead of extreme carry periods. Exhibit 1: Computation steps for carry calculations We assume that the bond is purchased for an inflation-unadjusted notional of 1 and financed in the repo market at a trade settlement date T0 and held until the settlement date T2. For example. The following notations are used: PtDU is the dirty price. an exact carry for a Canadian-style linker with a three-month lag can therefore generally be calculated over up to a two-week to a month and a half forward horizon. The formal computation steps for a carry calculation are laid out in Exhibit 1. one-month carry is worth 7. unadjusted for past inflation accrual. It is also possible to construct estimated forward real yields beyond the known inflation data using forecasts. a known monthly inflation accretion given the appropriate lag of 1% and repo funding at 5%. the real yield will tend to be biased down and therefore appear rich.5bp. unadjusted for past inflation accrual IRt is the Index ratio at settlement date t. using inflation futures or swaps. but for a 1y duration bond the carry is worth 75bp. at settlement date t ACt is the real accrued coupon at settlement date t C is the real coupon. Depending on the gap between the month of the last known CPI value and the current date. [1%+ (2%-5%)/12]/1. We assume that a coupon is paid on the settlement date T1. assuming an inflation index print date around mid-month.Barclays Capital | Global Inflation-Linked Products – A User’s Guide March 2010. for example. at settlement date t PtCU is the clean quoted price. in particular with seasonal swings likely ahead of periods of expected very positive and negative carry. the cash settlement price paid = = 15 March 2010 (P CU T0 + ACT0 * IRT0 ) PTDU * IRT0 0 9 . expected future carry can be an important feature of pricing both real yields and breakeven spreads versus nominal bonds. For example. [1% + (2%-5%)/12]/10.

the importance of inflation-linked bonds has varied across countries and over time. We also have: PTDU using the generic present 2 PTCU = PTDU − ACT2 2 2 PTCU 2 as an input. the total proceeds.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Using the non-arbitrage principle for forward calculations. issuance of inflation-linked paper declined in Mexico as the country gained investment grade status and demand for long peso paper increased. in the bond. ie. In practice. resale value plus any reinvested coupon. The additional inflation-unadjusted notional bought is therefore ie. but has already been discontinued and replaced by long nominal securities. The unadjusted forward price PTDU at settlement date T2 is then deduced from the equation: 2 ⎛ ⎞ ⎜1 + C ⎟ * PTDU * IRT = PTDU * IRT * ⎧1 + ⎛ T1 − T0 * repo ⎞⎫ ⎟⎬ ⎨ ⎜ 2 2 0 0 ⎜ PTDU ⎟ ⎠⎭ ⎩ ⎝ 360 1 ⎝ ⎠ ( ) In theory. Given the coupon to be paid in T1. and the need to countervail the related tendency towards financial dollarisation (the foreign currency denomination of financial assets) by introducing attractive dollar substitutes. Peru exhibited a similar pattern: issuance started in the early 2000s to help de-dollarise the economy. C DU PT1 . For instance. with both using Latin American inflation-linked bonds The development of inflation-linked government debt markets in many Latin American countries has been closely associated with their history of high and volatile inflation. By contrast. from the bond at date T2 should be equal to the cash settlement price in T0 uplifted with the repo cost. the forward bond price in T1 has to be computed first. This is assumed to be reinvested C * IRT1 PTDU * IRT1 1 . Total inflation-unadjusted notional held in T2 is 1+ C DU PT1 . The unadjusted forward price DU PT1 at settlement date T1 should satisfy the same non-arbitrage principle and is deduced from the equation: (C + P )* IR = (P DU T1 T1 DU T0 ⎧ ⎛T −T ⎞ ⎫ * IRT0 * ⎨1 + ⎜ 1 0 ⎟ * repo⎬ ⎩ ⎝ 360 ⎠ ⎭ ) The monetary coupon payment in T1 is equal to C * IRT1 . Accordingly. the forward real yield for T2 can then be obtained through the “Yield” function in Excel or the “YA” (Yield Analysis) page on Bloomberg. it is assumed that the coupon received will be reinvested in the bond at the calculated forward price for T1. Brazilian inflation-linked debt has increased significantly in recent years (by far the largest linker market in the emerging world) as the government made an effort to reduce exposure to currency and rollover risk by reducing the stocks of debt linked to the US dollar or the 15 March 2010 10 . the forward real yield can then be deduced from value formulation above.

unexpectedly low inflation. Why should governments issue linkers? Exploiting excessive market inflation expectations A government may have more faith than investors in the institutional arrangements in place to maintain an anti-inflationary bias. but investors remained unconvinced that there would be a significant long-term reduction in inflation. the gap between market and government expectations of inflation is likely to be small. the bonds are priced off a consensus expectation for the next inflation number. In most markets. By contrast with this tactical (and in some cases ephemeral) reliance on inflation linkers. While there may be times when divergences of expectations encourage issuance. If a government really has taken steps to bring down long-term inflation. due to a large extent to currency risk associated with large foreign currency-denominated public and private liabilities) when sharply raising inflation expectations left inflation indexation (legally banned in Argentina by the Convertibility Law in place during 1991-2001) as the only viable alternative to the dollar. During this period. While most Latin American linkers are quoted in real yield. ultimately bringing inflationary expectations down too.Barclays Capital | Global Inflation-Linked Products – A User’s Guide short-term interest rate. then it is in its interest to issue inflation-linked bonds while inflation expectations 15 March 2010 11 . This was a major factor influencing the UK’s decision to issue linkers in 1981: aggressive monetary and fiscal tightening had been implemented to bring inflation under control at the time. the expectations mismatch is unlikely to be the primary concern. Similarly. particularly when institutional changes have been made to fight inflation more directly. but otherwise. the price indexation methodologies can be very different. For more recently developing countries with less established monetary and fiscal institutions and capital markets there may still be occasions when governments perceive that the markets’ expectations of price increases are too high. in a format similar to Canadianstyle bonds. The experience of very high inflation in many Latin-American countries means that the lag used in the price indexation of linkers tends to be shorter than in the developed markets. Nominal bonds had enjoyed a windfall gain due to what was. In most developed economies this factor is notably less important than it has been in the past. the UK Treasury thus saved billions of pounds when inflation fell sharply and stayed low. as may have the resumption of issuance in Turkey in 2007. for the market. creating a very short de facto lag in the bonds. some were critical about the underperformance of linkers versus conventional bonds in this phase. the daily interpolating of the Reference Index means that the mechanics are very similar to the Canadian model. In the extreme case. the intermediate Index Reference values are not determined via linear interpolation. Positive credibility feedback A closely related benefit of inflation-linked bond issuance is that it can create a positive credibility feedback. Ex post. The substantial increase in issuance in Brazil in 2006 may have been partly motivated by such considerations. but assuming a geometric increase between the known published values. By issuing inflation-linked bonds. In Mexico the UDI inflation index is published twice a month. The details of conventions and calculations in each Latin American market are described in the relevant country sections. Uruguay and Argentina turned to inflation linking in the aftermath of the 2002 financial crisis (in turn. but such criticism was unjustified. adjusting to the actual value when it is released. the development of linker markets in Chile and Colombia owed itself to a long-standing policy of building a stable unit of account for long-term securities that dated back to the 1960s. IPCA consumer price indexed-linked bonds in Brazil have a lag that means their actual value is unknown in the early part of each month. With independent and transparent monetary policies.

the less incentive it has to reflate the economy and reduce the real value of the debt stock. The argument is weakened for countries that have significant inflation-linked liabilities. In practice. The more that revenues tend to grow faster than expenditures as prices rise. the government is better placed to cover this cost. Equally. liquidity. matching its future debt servicing costs with its revenues. then a government could save an inflation risk premium by issuing them in place of nominal debt. the more there is an incentive to issue inflation-linked bonds. However. except when there are substantial liabilities linked to inflation. Empirically in most markets. this consideration tends to gain increasing emphasis when monetary policy credibility comes under pressure. there is little pressure on public finances. Its major future “asset” is its entitlement to a future (real) stream of tax revenues. Saving a risk premium A popular early argument for inflation-linked issuance was that if government inflationlinked bonds really were risk-free financial assets. The more inflation-linked debt a country issues. While ex post the costs of inflation-linked bonds may be higher to issuers than nominal bonds would have been if there were higher-than-expected inflation. If investors are primarily interested in maintaining the future real value of their savings. The market may be more willing to believe in the institutional changes made to bring down inflation. or at least positive effects were more than offset by negatives. eg. prices are unlikely to rise 15 March 2010 12 . suggesting that risk premia are indeed more favourable for inflation-linked issuance at longer maturities. which will reflect both inflation and real economic activity. while long-term survey measures of inflation are normally relatively static with respect to maturity. Turkey is a clear example of where this may apply. Conceptually. Having at least a portion of liabilities linked to inflation should offer risk reduction benefits to the government borrower. while it may also have been a factor behind the significant increase in issuance in South Africa in 2009. as well as long-term liability benefits for a government. as for instance in the UK. It may be very significant for transitional economies that have undergone periods of high inflation though. for instance pension commitments linked to inflation that are liable to grow more quickly than inflation-linked tax revenues. again. Cyclical benefits Issuance of inflation-linked bonds can have significant cyclical. Early in the development of some of the major markets. there appeared to have been negative inflation risk premia. This argument is stronger the higher the percentage of taxation that comes from income and consumption taxes. it is debatable to what extent such a premium has been seen in the major markets. they should be prepared to pay an insurance premium for the privilege of owning a risk-free inflation hedge. The appropriate nature of liabilities The future expenditures and revenues of a government are almost all essentially real flows. This is another argument that is not particularly relevant for developed economies with totally independent monetary policies. the risk premium benefit of inflation-linked issuance is likely to be more pronounced at longer maturities. forward breakeven spreads between inflation-linked and nominal bonds usually increase at longer maturities. When growth is strong. The longer the expected lifespan of a particular government or policy regime the more the strategy may be beneficial. if the government is seen to be “putting its money where its mouth is”. as corporate taxes tend to be less dependent on domestic price levels. when growth is weak. as the inflation-protection element becomes more valuable. but inflation is likely to be higher.Barclays Capital | Global Inflation-Linked Products – A User’s Guide remain high.

Even when. including the US Treasury. being issued at a discount if inflation is expected over the life of the issue. while extending average duration via the nominal market is relatively straightforward. In several Latin American countries. A broader investor base not only cheapens funding on average. Maximising investor reach Clearly. issuing inflation-linked bonds can offer an attractive proposition. This diversification benefit can mean that it is in a government’s interest to issue inflation-linked bonds. such as in Mexico. have stressed this point. US state and local government pension funds hold less than 5% of their assets in government bonds. there is potential for a government that issues inflation bonds to reach investors who would not buy nominal government bonds. but it also reduces the reliance on particular sources of funds. The fiscal impact of a deflationary downturn on a country with a significant stock of inflation-linked bonds ought to be less severe than a country with only nominal debt. in the euro area. The largest issuers in recent years. for a country looking to extend the duration of its debt. unless they assign no probability to future inflation being lower than the market expects. a benefit also arises from reducing the exposure to short-term cash flow pressures. Risk diversification Even governments with no natural preference for either real or nominal liabilities should regard it as appropriate to have some inflation-linked liabilities within their debt. they are more natural buyers of inflation-linked bonds. however. Similarly. It is also a time when funding needs are high. and it is advantageous to extend the average life of the debt portfolio. due to fear of inflation eroding the value of nominal debt. This benefit may be a factor worth considering for transitional countries that have short-term cash constraints but ultimately sound finances. Conversely. In addition. The UK DMO puts particular emphasis on the fact that inflation and the government’s budgetary situation are likely to be highly correlated. even when implicit inflation is lower than the government’s inflation expectations. where governments are required to account for inflation as it accretes in linkers. Duration and cash flow benefits A standard inflation-linked bond has smaller nominal cash flows upfront than as the price level rises. as there are relatively few competing sources of inflation-linked supply. while their liabilities are linked to the cost of living. when after a strong growth period inflationary pressures may continue to grow even when output is already falling away. in nominal terms. again reducing risk. the main risk to this hypothesis is late in the economic cycle. Other than a “stagflation” scenario. the ability to reach an additional set of investors is a highly regarded prize. For example. A government is better off having a balanced liability portfolio in the face of economic uncertainty. but the counter-argument is that tax revenue also tends to lag output growth. It is a less important factor in developed countries. long-dated maturities have been issued in inflationlinked before the nominal market was sufficiently developed for long nominal issuance. Servicing linker costs should thus tend to be a fiscal stabiliser compared to servicing nominal debt. nominal curves have eventually extended to maturities as long as those of inflation-linked 15 March 2010 13 . and also to tap new money that would not have been allocated to nominal debt. where there is particularly high competition between government issuers.Barclays Capital | Global Inflation-Linked Products – A User’s Guide quickly. An inflation bond is. if so desired. As it is usually easier to sell longer-dated real return bonds than nominal issues. as it is likely that during such a phase inflation risk premia will be high until policy acts to contain inflationary pressures. issuing inflation-linked at the start of an economic upswing may be optimum timing.

After the US FOMC indicated that the main marketbased series of inflationary expectations that it focuses on is the breakeven implied by 5y5y forward TIPS. Partly. Putting a price on such benefits is difficult. The reason for the lower liquidity has a lot to do with the product matching long-term needs better than nominals. the lower liquidity is the price of success for meeting specific needs so well. The ability to easily discern markets’ inflation expectations may be of benefit to policy setters. For a sovereign with a very small debt. In addition the issuance of inflation-linked bonds should not be viewed in isolation. Having a market-based reference of inflation expectations from linkers may also be useful for economic agents in making decisions.Barclays Capital | Global Inflation-Linked Products – A User’s Guide bonds. either directly into inflation-linked bonds. this series stayed in a tight 40bp range until the strong inflation volatility experienced in 2008/2009. there may be considerable benefits if breakeven spreads between inflation-linked and nominal bonds help to avoid inflationary monetary and fiscal policy errors. Experience elsewhere also suggests that it can take several years before there is sufficient liquidity and acceptance of the asset class for the implied inflation to be a reliable enough guide to be a major benefit. Since this shock the liquidity gap has reduced though and in most countries it is notably lower than it was prior to the mid part of the decade. Drawbacks of inflation-linked issuance Criticism suggesting inflation products are less liquid than their nominal equivalents is fair. This could encourage more savings. While liquidity is lower. However. relatively stable spreads may also provide a self-reinforcing credibility tool for inflation targeting. which means that much less day-to-day trading is needed. where government supply ceased after 20 years in 2003. In particular. This was the situation in Australia. While it can be difficult for central bankers to ascertain how much of breakevens is true inflationary expectations and how much comes from risk premium. With central banks making no secret that they observe both spot and forward inflation-linked breakevens. it is unlikely to be very large. this differentiation becomes relatively less important. to the extent that the series itself becomes tied to policy credibility. which needs to keep a liquid nominal market to help the overall market function and in case circumstances become less favourable. Social benefits The existence of inflation-linked bonds may provide benefits to society beyond the funding considerations. Nonetheless. an inflation-linked market needs to be seen as relatively representative. or indirectly into assets for which there is a clearer real value if there are inflation-linked assets for comparison. The existence of inflation bonds could theoretically reduce inflation uncertainty. with liquidity differences exacerbated following market dislocations in the six months from September 2008. but as there seem to be few clear differences in behaviour between economic agents in similar countries with and without inflation bond markets. the duration of the long linkers remains notably longer due to the back-ended nature of linker cash flows. 15 March 2010 14 . South Africa has undergone a similar extension of its debt since the introduction of inflation-linked bonds. there is an argument for not issuing inflation-linked. In order to provide a significant benefit to policy makers. liquidity is a concern that a new government issuer has to overcome in order to sell inflation-linked bonds at attractive levels. with this argument also influencing the decision not to offer a deflation floor. a less frequent need to trade means that the relative cost of turnover is not that high. One of the major reasons put forward within Japan for an inflation-linked bond programme was that the resultant implied inflation rate would be a useful policy gauge. without a broad acceptance of the asset class by domestic investors this role failed to gain traction.

$100 now would have the purchasing power of $74 in 30 years’ time if inflation averaged 1%.5%. while the lower fiscal pressure resulting from a government funding with inflation bonds when there is a significant inflation risk premium ought to be deflationary. except in periods of extreme inflation. when 15 March 2010 15 . While there is some evidence to support the risks of creeping inflation from widespread indexation. In this case. and countries such as Israel and Iceland have tried to wean themselves off indexation as a consequence. The nominal volatility of returns on short-dated Canadian-style bonds is usually higher than those on nominal bonds – sometimes considerably higher at maturities of 1y and less. There is no reason why bonds cannot be linked to inflation without general indexation elsewhere. It should not be forgotten that even in a stable lowinflationary environment. appropriate monetary policy should be able to address this observable element.Barclays Capital | Global Inflation-Linked Products – A User’s Guide despite substantial domestic pension fund demand. If monetary policy is independent and can respond to the inflationary effects of any indexation. conceptually. where. the consumption of many items is relatively uneven and the individual may change purchasing patterns when inflation occurs. It should be relatively easy for a government to keep financial funding and other price setting at arms’ length. This line of reasoning was particularly prevalent in Germany. after multiple periods of hyper-inflation. but only $48 if inflation averaged 2. in the long term there is considerable uncertainty about the real value of nominal bond returns. While inflation-linked bonds represent an ideal long-term low-risk investment for an individual whose consumption pattern is represented relatively closely by a consumer price index basket. While there is always an element of basis risk for an individual. as their own consumption basket will not be the same as the relevant inflation index. While mathematically they are harder to quantify than nominal instruments. but why hold anything else. If bonds are linked to inflation. If saving is ultimately about deferred consumption. this is a long way from saying that it is the fault of inflation-linked bonds. On the other hand. with government issuance restarting in 2009 as funding needs rose. The danger is that if the cost of inflation is made less painful for individuals by widespread indexation. it is far from clear if this is the case in the shorter term. there will be increased pressure for other items to be linked to inflation too. or for other investors. no other financial asset can give close to the real value certainty of inflation-linked products. then for an investor the question should not be why hold inflation-linked bonds. there should be much less longer-term uncertainty for a product for which the value of the real cash flows is known in advance. more than 50% less. inflation may increase until it reaches levels that are once again painful. increasingly active quasiand non-government supply subsequently developed along with inflation swaps trading. In the short term. One persistent criticism of governments issuing inflation-linked bonds is that any form of inflation indexation is insidious and pernicious. the signalling benefit from inflation-linked bonds is likely to partially offset the inflationary bias. Even for a representative individual. there may be an inflationary bias created that is a negative social externality. consumption is based on nominal considerations. Government linkers: The risk-free asset class? Inflation-linked instruments are often criticised for being opaque and difficult to calculate. An extension to the dangers of indexation is that if there is a substantial risk premium in inflation-linked bonds and the implied inflation rate in the market is used as a basis for agents’ behaviour when setting prices and wages. it was made illegal during the period of the Deutsche Mark for any debt to be indexed.

or whose investments occur in currencies other than those in which consumption occurs. there is little incentive to consider government inflation-linked bonds as the risk-free asset class. For those working on behalf of individual investors. in the US. but there is much less argument for holding them as a core asset. In practice. 15 March 2010 16 . the indexation lag makes the real yield that is offered far from prefect at short maturities. the company involved retains no exposure to inflation. Bond issuance in the euro area remains relatively limited. In addition. as returns are often too low to justify the fees that most investment managers take. In some emerging markets – for instance. in the absence of a well-defined real rate swap curve that can be accessed on a funded basis. However. although issuance from 2008 has been much less than in 2005 to 2007. but it is also natural for many corporates to consider issuing them or pay inflation via swaps. There are only a very small number of unswapped US municipal issues despite the fact that funding levels for corporates to provide inflation to the market are notably more attractive than for the government. While there is an argument that a well-defined real rate swap curve is a better proxy for the risk free rate. non-government inflation-linked issuance remains a notable feature. the credit risk of government exposure adds an additional complication for individuals in lower-rated countries. it also makes sense for the inflation-linked bonds to be the risk-free asset for long-term investment. Non-government supply considerations Governments and agencies with revenues linked to inflation are among the most obvious issuers of inflation-linked bonds. who cannot avoid the underlying credit quality of the government. An international investor looking for a realistic comparison should scale up the credit default swap valuation for the increase in expected size of the inflation-linked bond’s final principal to measure the additional credit risk in the inflation bond relative to the nominal curve. for example. Factoring in this risk is complicated by credit default swaps only trading in nominal space. there are a variety of non-government issuers of inflation. In the UK. while there are clear arguments for endowments to use inflation-linked government bonds as the basis for investment decisions. a government inflation-linked bond may still provide the best proxy for a risk-free discount rate. the back-ended nature of real cash flows may make them less appealing than nominal bonds. This does not mean that they should not invest in them as part of a low-risk diversification strategy. and other riskier assets need to prove why they offer an attractive alternative. for an individual within a lower-rated country. For those investing for atypical individuals. While an inflation-inked bond offers a certainty of real cash flows in normal circumstances. However. A AAA-rated government inflation-linked bond ought to be considered the risk-free asset class for an average individual over the long term. Israel – inflation-linked issuance has been one of the few ways for corporates to access long-term capital. albeit in a narrower range than nominal bonds. with little supply from outside entities with tax-like revenue. such as in UK-defined benefit pension schemes. By contrast.Barclays Capital | Global Inflation-Linked Products – A User’s Guide temporary monthly inflation trends cause large deviations. whose consumption basket has little linkage to consumer price indices. they are only encouraged to do this actively if liabilities are tightly tied. ie. the ideal exposure for an individual should involve a real cash flow. Along similar lines. but there is a more significant paying flow in inflation swaps in both euro and individual country inflation. such as pension fund managers. While in most such countries nominal markets have now opened up to corporate issuers. there has been almost no corporate inflation-linked issuance in which the issuer does not asset swap in full. there is less of a case for corporates without clear inflation-linked costs to invest spare cash anywhere but in the nominal market.

particularly inflation paying. In July 2008. contractually-based inflation-linked revenues are the most natural source of supply. but no linkage between revenue and rising average prices. FAS133 considers the interest rate and the rate of inflation in the economic environment for the currency in which a debt instrument is denominated to be clearly and closely related. If there is a premium for the inflationlinked asset class. has been accounting treatment. making it extremely difficult for a US corporate to take exposure in inflation swaps without marking to market even if revenues are explicitly inflation-linked. encouraging direct payment of the inflation stream. Explicit. white goods retailers. so offsetting these with real liabilities ought to be appealing. every county with an inflationlinked market now either accepts international accounting standards. US Generally Accepted Accounting Principles (GAAP) only allow for hedging of interest rate. hedge accounting of such explicit cash flows is possible. a land-owning company might be a natural issuer of inflation even if it does not have direct revenue linked to inflation. eg. foreign exchange and credit. the UK had the most straightforward treatment. particularly for firms whose physical assets are a high proportion of their total value. even when there is no clear long-term revenue link to inflation. regulated utilities in some countries have the rate at which they can increase prices linked to inflation. the International Accounting Standards Board (IASB) amended the IAS 39 rule with clarification regarding the hedging of inflation. with the same reasoning as in the US. Such firms include those with large wage bills that track price inflation relatively closely. for instance supermarkets. as in Europe. encouraging the market to develop. but there is an argument for large firms with balanced inflation exposure to consider the inflation market as a diversified source of funding. Fundamental considerations Fundamentally. Other inflation linkages are subtler. In the US.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Accounting considerations One issue that has been a limiting factor for non-government inflation-linked issuance. allowing for issuance of non-leveraged bonds. The hedge accounting of inflation is allowed only if changes in inflation represent a contractually-specified element of cash flows of the instrument considered. With accreting inflation15 March 2010 17 . where there is a significant inflation risk premium. The main reason for the lack of a corporate inflation-linked bond market in the US is that such inflationlinked controls are relatively rare. In most countries other than the US. Bonds could be accounted for on an amortised cost basis with uplift accounted for as a finance charge. Corporates with significant revenues or assets linked to inflation have clear potential to benefit from issuing or paying inflation. This should be the case. but inflation swaps could also normally be held on an accruals basis. The most common reason for corporates to issue or pay inflation is if their revenues are inflation-linked. whose sales will be similar to the inflation basket and so their prices will rise in a similar vein. having it within the portfolio of a corporate can tend to act as a profit stabiliser. or has an accounting agency seeking convergence with them. the reasoning behind most corporate structural paying/issuing of inflationlinked instruments is not that different from a government’s. for instance. that there has been issuance from issuers with such indirect inflation linkages though. This is particularly the case for industries with strong cyclical cash flows. For example. Other than Argentina. Also. though there can also be shorter-term value and cash flow considerations. It is only in countries such as the UK and South Africa. just as having inflation exposure within a government debt portfolio acts as a fiscal stabiliser. then firms are likely to benefit from funding in inflation unless they have a cost base or liabilities more closely tied to inflation than their revenue or assets. Sometimes this cash flow linkage is obvious. However. Historically. Corporate balance sheets are full of real assets.

can be represented as in Figure 1. On the other hand. The benefits of extending investor reach by issuing inflation-linked bonds can be more important for corporate issuers than governments. inflation-linked bonds in most countries have with the uplift on the principal being taxable even though it is not paid out. ie. 15 March 2010 18 . may offer the best hope of a market for bonds outside of TIPS developing. as has been the case in the UK. where the problem of tax on unpaid uplift would not be an issue. very rarely involve any net inflation supply into the market. Companies can more easily take advantage of issuing into specific pockets of demand than governments. coupled sometimes with slightly more depth of demand for long-dated inflation linked than nominal exposure. There is certainly demand for such structures to develop as evidenced by the development of synthetic inflation-linked muni funds that gain exposure to CPI from TIPS. Typically. but these. The hedge which swaps out the inflation element depends on the format of the inflation-linked issue. Format of corporate inflation supply Corporates with inflation-linked revenues tend to be suited by issuing bonds with a similar accreting format to government inflation-linked issues. this can make inflation-linked issuance appealing to firms wanting to pay for their investment with longer-term funding. there are times when it is cheaper for firms to align their needs specifically using inflation swaps rather than issuing bonds.Barclays Capital | Global Inflation-Linked Products – A User’s Guide linked bond cash flows skewed later than those of nominal bonds. This helps to distinguish true inflation-linked supply from notes that are mainly issued into relatively niche retail markets with the inflation exposure swapped out. However. The lower liquidity of inflation-linked bonds than nominals can limit the depth of demand for non-government issuance. as generally their revenues accrete in a similar way. It is also a reason why in the US unswapped municipal issuance. but in hedging out the inflation component of all the cash flows. On the other hand. it is often cheaper for a corporate with explicit inflation-linked revenues to issue a nominal bond and then swap this exposure into inflation. but schematically. as with simple year-on-year coupon linked bonds. This has meant that in the UK in particular it is more common for unlisted companies. with the inflation risk swapped out. for publicly-listed companies unable to hedge account for inflation swaps. to pay inflation or real rates than public companies with similar inflation-related cash flows. for instance toll road prices. The following chapter discusses more complex structured notes in more detail. Outside the UK this kind of swapped note has been more prevalent in developed markets than true paying of inflation risk via unswapped corporate inflation-linked bonds. Thus. such as those controlled by private equity firms. Swapping out the inflation component of an inflation-linked bond An issuer may find an advantage in issuing an inflation-linked bond. such inflation notes offer coupons linked to year-on-year inflation rather than having a linkage via the principal value of the bond and are sold mainly to individuals via bank distribution networks. to inflation. particularly for smaller amounts. This avoids the problem that government-style. the mark-to-market valuations of long-term exposures may create unacceptable volatility in the results. for instance issuing more structured bonds and using inflation derivatives to align this with their liability needs. Added to often narrower assets swaps for inflation-linked government bonds than their nominal counterparts. for this to develop in a substantial way involves linking revenue streams of projects. it only tends to be markets in which inflation-linked pension demand is well developed in which true corporate inflation-linked bonds can flourish. whether the inflation paid out in an accreting or y/y format.

15 March 2010 19 . unless specific interest in a name offsets the relative richness of the hedging swap. pushed inflation swaps cheaper than bonds and had then led to some swapped issuance. In the UK. there is an advantage to issue debt in inflation-linked format and swapping out the inflation exposure only if the price of inflation is below the implicit breakeven rate between the cost of an equivalent nominal issuance and the real rate paid. such that there is no funding advantage in swapped issuance. In the euro area. heavy asset swapping of corporate linkers have. swaps have historically traded at a premium to bonds. but such bonds have usually been privately-placed. during some periods.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 1: Swapping out the inflation exposure in inflation-linked issuance Fixed or Libor payments Inflation -linked bond issuer Inflation-linked cash flows Hedging bank Bond price Inflation-linked cash flows Inflation -linked bond buyer Source: Barclays Capital Other than accounting considerations.

and this has led to a market for inflation caps and floors. a long/short position in an inflation swap implies receiving/paying inflation (ie.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation derivatives Khrishnamoorthy Sooben +44 (0) 20 777 37514 khrishnamoorthy. Alternatively. which is again the opposite of nominal swap parlance. As with inflation-linked bonds. The receiver/payer in an inflation swap will receive/pay accrued inflation and will pay/receive the fixed rate. Figure 2: Zero-coupon inflation swap structure notional * (1 + fixedrate) [ tenor −1 ] Counterparty B Counterparty A ⎞ ⎛ InflationI ndex t + tenor notional * ⎜ − 1⎟ ⎟ ⎜ InflationI ndex t ⎠ ⎝ Source: Barclays Capital One issue to be aware of is the potential confusion from the terminology used when inflation swaps are traded. In Europe and to a lesser extent the US. The fixed rate. ie. For standard US CPI and FRCPIx (French CPI ex-tobacco) swaps. Stating exactly which leg is being received and paid clears any confusion. This is opposite to the convention in the nominal swap market where the receiver/payer is understood to be receiving/paying the fixed rate. and is quoted as an annualised rate. Linear inflation derivatives Zero coupon inflation swaps A zero coupon (ZC) inflation swap is a pure inflation instrument. complex structured products issuance has emerged.sooben@ barcap. agreed at inception. Asset swapping has encouraged a parallel development with the cash market. reflects expected future inflation. with the nature of liabilities in the former having led to the development of a non-linear market. both parties will breakeven on the trade.com Alan James +44 (0) 20 7773 2238 alan.com Inflation derivatives are now established products. the net cash flow at maturity will be nil. only if annualised average inflation over the swap’s period is equal to the initial fixed rate. Paying or receiving in inflation swap parlance normally relates to the inflation leg.james@barcap. although the US remains a bond driven market with relatively limited swap activity. 15 March 2010 20 . The cash flow on the fixed leg is predetermined by the quoted swap rate and is effectively a breakeven inflation rate. it is therefore the “price” of expected inflation over the period of the swap. although the lagging principles may differ from the bond market. Liability-driven demand is an important feature of both the UK and euro markets. the floating element) versus paying/receiving the fixed leg. the inflation index reference value is calculated on the same three-month lag and interpolated principle as in their corresponding bond markets. we consider a 10-year ZC US CPI swap traded on 4 January 2010. Cash flows between the two parties in the transaction occur only at maturity and involve the exchange of a notional adjusted for inflation that has accrued over a specified time period against the notional capitalised with a fixed rate. For illustration. The leg depending on accrued inflation (referred to as the inflation or the floating leg) will vary solely on the basis of the final inflation index reference value at the end of the period of the swap. the inflation indexation mechanism in a swap is subject to a lag.

unlike in nominal swaps it is not sufficient to interpolate between annual points. This means that a euro HICPx swap traded on any given day of a particular month will have the same starting index reference value. The traded price indices in standard inflation swaps are naturally the same as for their corresponding bond markets. with some quoting on the basis of a lag to the settlement date and others versus the execution date. The settlement can be especially important when dealing with non-interpolated swaps. it is important to note that there is a discrete jump in quoted swap rates. with a two-month lag in the UK and three months in euro. in the exact same way as for US TIPS. 15 March 2010 21 . stating exactly which base month is being traded avoids confusion. In the interpolated case. For a comprehensive discussion of seasonality see the section ‘Seasonality and inflation markets’ later in this publication. The practical advantage of a non-interpolated convention is that a standard swap initiated and unwound in the same month will leave no residual inflation risk as the inflation accrual period traded will be exactly the same for both transactions. mainly because of seasonality. the interpolation imposes a drift on quoted rates. In particular. all standard euro HICPx swaps traded during January 2010 pay inflation accruing from October 2009 – referred to as the base month. However. In the UK. when the base month changes. The detailed computation steps to construct a projected CPI curve are provided in Exhibit 2. as the spot date is the trade date plus two business days. the lagging principles are notably different from the cash market as they trade on a noninterpolated basis. The quoted market for non-interpolated swaps remains consistent during a month. On each trading day. t+2 is standard. as they are generally in line with nominal swap markets instead. while in the US and euro area. as interpolation assumes that the price index will be growing at a constant rate over a year. which will be the published index value three months prior. care should be taken when establishing new positions in euro swaps at the end of each month. The inflation index reference value for 6 January 2010 is calculated as an interpolation between the October and November 2009 CPI values. as not all market quotes at this time will be on a consistent basis. While there are no discrete jumps in interpolated swaps. the fact that the index reference dates change every day implies that standard swaps on two different days will not be perfectly offsetting. Building a CPI curve from zero coupon inflation swaps Zero coupon inflation swaps provide the building blocks for the construction of a projected CPI curve from which other inflation-linked derivatives can be priced. For example. The second step therefore consists of an adjustment for the estimated seasonality during each month. which enables an easy reading of how the market is moving over that month. On the other hand. which makes the interpretation of data relatively difficult.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The start date from which inflation accrues on the swap is 6 January 2010. However. annual tenor swap quotes effectively provide the projected increase in the price index from a start index reference date over whole year periods. In February 2010. settlement dates may be different from bonds. a t+0 settlement date is typical. the base month for all swaps shifts one month and changes to December 2009. Here. For standard UK RPI and euro HICPx swaps. This is unlikely given the seasonal behaviour of price indices.

which is m months after the start date We have: PRCPI t = CPI 0 * (1 + ZCt ) t To interpolate between PRCPI t and PRCPI t +1 . of month of index reference date. whereby we calculate the annual trend rate of inflation between t and t+1 as: ln (PRCPI t +1 ) − ln (PRCPI t ) We then approximate PRCPI t . before applying the seasonality overlay. as: m PRCPI tm = CPI 0 * exp{ [ln(PRCPI t +1 ) − ln(PRCPI t )]* m / 12 } After applying the seasonality overlay. we can use a piece-wise flat interpolation method. in %.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Exhibit 2: Computation steps for constructing a projected CPI curve We use the following notations: CPI 0 : CPI index value for start index reference date PRCPI t : projected CPI for index reference date t years after the start reference date PRCPI tm : projected CPI for index reference date t years and m months after the start reference date ZC t : quoted annualised ZC swap rate with maturity t years MS m : month-on-month seasonality. we have: PRCPI tm = CPI 0 * exp{ [ln (PRCPI t +1 ) − ln (PRCPI t )]* m / 12 }* exp ∑ MS n n =1 m Figure 3: Projected FRCPI values after interpolation and seasonality overlay 140 Projected FRCPIx values (7 January 2010) 135 130 125 120 115 Jan 11 Jul 11 Jan 12 Jul 12 Jan 13 Jul 13 Jan 14 Jul 14 Jan 15 Jul 15 Jan 16 Source: Barclays Capital 15 March 2010 22 .

This expected cash flow is then discounted by a zero coupon nominal discount factor to give the current value of the swap position. 15 March 2010 23 . The magnitude of the oscillation due to seasonality on the projected ZC curve will naturally decrease in longer maturities as the ZC rate is expressed as an annualised rate. The difference between the notional inflated using the new projected index value for September 2014 and using the projection at inception date gives the expected future net cash flow at maturity of the swap. Therefore. 1 Only a five-year horizon is shown on the chart as the oscillatory pattern becomes less visible as the range of values on the Y-axis is increased. This swap will pay inflation accruing from September 2009 to September 2014 versus an annualised fixed rate (breakeven rate). but will have an oscillatory pattern. the new annualised rate corresponding to what the market is now pricing as inflation between September 2009 and September 2014 is needed. This is illustrated in Figure 3 1.5 1. Figure 4: ZC FRCPIx swaps curve after interpolation and seasonality overlay 3.5 ZC FRCPIX swaps curve (7 January 2010) 2. we assume a 5y ZC UK RPI swap traded in November 2009. a complete ZC swaps curve can be built by calculating the annualised rate of growth of the index from the starting reference date to each future reference date.0 2. as in Figure 4.5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Source: Barclays Capital Pricing ZC swap positions The unwind value of a ZC inflation swap trade varies solely on the basis of changes in the projected index value for the final inflation reference date. Exhibit 3 details the computation steps for a sample P&L calculation. the segments between two annual points on a projected CPI curve will not be straight lines. If the base month for standard ZC UK RPI swaps has changed. As an illustration. the projected index value for September 2014 may need to be computed using an interpolation plus seasonality overlay method analogous to the one detailed above. From the projected CPI curve.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Due to the seasonality overlay.0 1. then this rate will not be readily available from market quotations.0 Maturity 0. If the swap is unwound at a later point in time. This rate can be used to compute the projected index value for September 2014 at inception date. where projected annual FRCPIx values are computed from quoted ZC FRCPIx swaps on 7 January 2010 and intermediate values calculated via interpolation plus a seasonality overlay.

013% -0.3 Swap pays inflation accruing from September 2009 to September 2014 Projected index value for September 2014 at trade date = 215.0305^4) = 243.083% 15 March 2010 24 .068% -0.13% Projected index value for October 2013 = 216 * (1.392% (rounded) We assume the following m/m seasonality vector: January February March April May June -0.594% 0.392% 0.027% -0.3 * (1.279% 0.080% 0.58) = 3.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Exhibit 3: P&L calculation on seasoned ZC inflation swap position Trade date: 17 November 2009 Index traded: UK RPI Tenor: 5y 5y ZC swap rate: 3.41 (rounded 2dp) Trade unwind date: 11 December 2009 At the unwind date.99) – ln (243.0313^5) = 251.574% 0.05% 5y ZC swap rate: 3.58 (rounded 2dp) = 216 * (1.99 (rounded 2dp) Projected index value for October 2014 Trend rate of inflation between October 2013 and October 2014 = ln(251.173% -0.15% Notional: £100mn Base month on standard swaps on trade date: September 2009 Published September 2009 UK RPI value: 215.0315^5) = 251. we have: Base month on standard swaps: October 2009 Published October 2009 UK RPI value: 216 4y ZC swap rate: 3.116% July August September October November December -0.094% 0.

the effective inflation duration of an inflation swap will be 15 March 2010 25 . The difference in the cash flows is -£32922. if this decline is less than implied by the seasonal model then carry on the position that is receiving inflation can be positive.. the projected index value for September 2014 = = Projected index value for October 2013 * exp (Trend inflation * 11/12) * exp (November + December +….58 * exp (3. carry each month will be determined by the difference between the actual monthly inflation accretion and the traded breakeven level plus monthly seasonal factor.392%/12)} / exp(0. zero nominal duration.99/exp (3. the fixed rate of a ZC inflation swap is set at inception such that the value of the swap is nil. carry is a much more significant factor at shorter than long maturities.083% +….+ September m/m seasonal) = 243. -£28314).392% + 0. This means that if one goes long inflation on the 5y swap on 17 November 2009 and unwinds it with an equivalent swap. Hence.86. ie. even if inflation in a given month is negative. the P&L (excluding transaction costs) would be -£32922 times 0. i.-0. To look at the valuation of existing swaps from another perspective. Hence.86 (ie. then positive carry may be scant consolation and as with inflation-linked bonds. if the market price of the remaining position moves the other way.3) * £100mln = £116. Notional capitalised using projection on trade date = (Projected index value for September 2014 on trade date / September 2009 index value) * Notional = (251. the swap has no sensitivity to the nominal discount factor.3) * £100mln = £116.34/215. the position will be sensitive to changes in the ZC inflation swap rate. which is around 0. By definition.41/215.027%) = 251.774mln Notional capitalised using projection on unwind date = (Projected index value for September 2014 at unwind date / September 2009 index value) * Notional = (251. at inception. this sensitivity is known as the inflation duration. On the other hand.392% * 11/12) * exp (-0. Even though it is a zero coupon instrument.34 This is equivalent to working backwards from the projected index value for October 2014.741mln The capitalised notionals calculated correspond to the cash flows at maturity on the fixed leg of a swap paying inflation from September 2009 to September 2014.Barclays Capital | Global Inflation-Linked Products – A User’s Guide At the trade unwind date.e.34 To calculate the P&L on the position. This has to be discounted to the trade unwind date by the zero coupon discount factor.068%) = 251.: Projected index value for September 2014 = {Projected index value for October 2014/ exp(Trend inflation/12)} / exp (October m/m seasonal) = {251. we compare the notional capitalised on the basis of the two projections for the September 2014 index value. Obviously.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

less than its tenor due to the cash flow being paid at the end of the period. A seasoned swap trade is likely to have acquired value, for instance because breakevens have changed or because realised inflation levels have been different from the levels implied by the swap at inception. To determine the present value of a seasoned trade, its future expected final value has to be multiplied by a nominal discount factor to arrive at its present monetary value. Thus, the dv01 of an existing inflation swap with respect to movements in the inflation breakeven will be determined by the ratio of its remaining tenor to one plus the discount rate to the power of this tenor. As the inflation swap acquires value, it will start to gain a residual nominal duration, as a lower nominal yield takes the present value closer to the expected final cash value, although in the event of a negative expected final value, this residual duration is also negative. For a spot position, this nominal duration is usually a minor second order consideration, but the importance of effective convexity that stems from inflation and rates tending to move in the same direction can assume notably more importance for forward inflation swap positions.

Real rate swaps
While in a ZC inflation swap the cash flow only depends on the traded breakeven level and final inflation print, a real rate swap involves a real, inflation-linked, cash flow versus a floating leg, for example, a compounded Libor rate. In most developed markets, the majority of investors choose to separate decisions on nominal duration and inflation exposure even though most liabilities are in a real format, although real rate swaps do also trade. In most emerging markets, real rate swaps are much more commonly quoted than inflation swaps. Real rate swaps are usually quoted in a zero coupon format in markets in which ZC inflation swaps are established but can also trade in a coupon format in most of these markets. The effective duration of a ZC real rate swap is relatively long as the appropriate discount factor is real rather than nominal. Generally, investors matching liabilities who trade real rates will always trade ZC, but value investors, particularly those concerned about convexity of forward positions, often trade real rate swaps with fixed real rate coupons. In Latin America, almost all single-currency real rate swap activity is in coupon form. The parallel currency format of Latin American linker markets other than Brazil makes this kind of position, a quasi-cross currency swap, relatively straightforward.

Year-on-year inflation swaps
A format of inflation swap that used to be widely quoted in many markets but is now less common is the year-on-year (y/y) structure. This used to be the standard format for swapped corporate inflation-linked bonds sold to investors for whom the back-ended cash flows of government-style issues were not tax efficient, and hence was more prevalent than the zero structure in the early days of the euro market. The y/y structure involves one counterparty agreeing to receive an annual coupon determined by the y/y rate of inflation in return for paying a fixed rate. The fixed rate is the quoted rate, and is analogous to a breakeven inflation, although not in as pure a sense as in a ZC format. The yearly inflation period lags the payment dates in a similar way to the zero structure. In the US, y/y inflation swaps tend to be quoted in the monthly cash flow format, as this is the most tax-friendly structure, whereas in Europe the cash flow frequency is normally annual. Sometimes, the non-inflation leg of the swap will be floating Libor with a spread rather than a fixed rate. Although not quoted nearly as much on dealer screens as the zero-coupon structure, indicative y/y rates are easily derivable from the zeros in the same way as is true for coupon interest rates from zero-coupon interest rates. For longer maturities, convexity distortions can become relatively large, but most structured notes are quite short. The most common
15 March 2010 26

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

use of the y/y swap remains for hedging structured products, both new exposure and unwinds of old bonds that have been sold back to issuing banks. The floating leg of a y/y inflation swap is equivalent to a collection of consecutive forward inflation rates, and therefore the y/y structure can be useful for hedging inflation caps and floors, either on a stand-alone basis, or as features within structured products.

Figure 5: Year-on-year inflation swap periodic payments

Notional * dayfrac * fixedrate Counterparty A Counterparty B

⎧ ⎡⎛ CPI m−3 ⎞ ⎤ ⎫ ⎪ Notional* dayfrac ⎪ ⎢⎜ ⎟ − 1⎥ + X %⎬ ⎨ ⎜ ⎟ CPI m−15 ⎠ ⎦ ⎪ ⎣⎝ ⎪ ⎩ ⎭
Source: Barclays Capital

Real rate swaps versus $ Libor
Traditionally, the most common form of inflation-linked swaps in Latin America has been cross-currency swaps versus $ Libor. In Mexico, Argentina, Chile, Colombia and even Peru, local CPI-linked indices provide a reference value for real rates to be quoted. These positions are quoted in a coupon format and are effectively equivalent to buying or selling government inflation-linked bonds in these countries on an unhedged currency basis but without being exposed to the credit of the relevant government. In many of these markets, the increased importance of domestic demand has led to local real rates being quoted more frequently in recent years. It is straightforward to quote real rate swaps in other countries versus foreign currencies too, but it is rare to see, with such flows normally separated out.

Accreting inflation bond-style swaps
Transactions are sometimes structured to mimic existing inflation-linked bonds as a hedge for cash linker positions. Such accreting inflation bond-style swaps exactly replicate the cash flow structure of the bonds matching payment dates, and matching base indices with identical lags to the index, with the other side of the swap being a fixed rate or floating Libor plus spread. This kind of swap is rarely seen now. The linker-style structure is now traded far more commonly in another guise, as part of the inflation-linked asset swap, which is effectively a combination of the accreting bond style swap and an opposing position in the cash bond.

Inflation-linked asset swaps
Asset swapping of inflation-linked government bonds has been a major feature of the main developed inflation-linked markets for several years. An inflation-linked bond asset swap is essentially the same thing as a nominal bond asset swap, with the expected future nominal cash flows of the bond priced by the nominal swap curve of the currency concerned. It is then translated into a Libor flow, plus or minus a spread. The euro area was the first region where widespread asset swapping of government issues developed in 2003, following the first Italian issuance. This was the first occasion where there was a sufficiently liquid swaps market to allow viable pricing and with Libor spreads attractive enough to draw interest. Once the opportunity had been highlighted by activity on the part of speculative players, the natural buyers of asset swaps (ie, Libor-funding financial institutions) began to get involved in French CPIx-linked bonds as well as euro-linked issues. Asset swapping effectively tied
15 March 2010 27

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

together the growing bond and swap markets and has encouraged a complementary development since. With very few natural payers of euro area inflation other than governments, there is a natural tendency for asset swaps in inflation-linked bonds to be somewhat cheaper than in nominal government bonds. However, with a range of government issuers, the potential asset swap demand base is substantial, and has generally helped limit the degree to which breakevens in the bond and swap markets trade out of line. This was not the case however towards the end of 2008 and start of 2009, when European bond breakevens dropped significantly relative to inflation swaps; but the fact that many speculative players and natural buyers of asset swaps were no longer active meant that acute relative distortions persisted over several months. In the UK, there has been considerable asset swap activity in gilt linkers. However, up to 2007 this was dwarfed by asset swapping of non-government issuance, which drove the development of the swap market. When such corporate supply was heavy, gilt linkers sometimes traded significantly more expensive in asset swap than nominal government issues. In 2008 and 2009, UK corporate inflation-linked issuance other than from Network Rail was extremely limited, prompting gilt linkers to generally trade cheaper in asset swap despite unprecedented gilt linker asset swap buying. In the US there are very few sources of inflation other than TIPS, while Treasuries are not an asset class that is bought by asset swap investors as they are too expensive. This left US CPI swap breakevens much wider than TIPS breakevens, but prior to mid-2008 demand for TIPS asset swaps developed whenever issues traded at a discount to the general collateral (GC) repo rate spread to Libor offering a positive carry position. JGBi asset swap demand developed along similar lines, with demand cheaper than Tibor-GC. The widening of spreads between interbank and general collateral rates in the second half of 2007 helped shake up positions in both markets even as global nominal asset swap liquidity suffered. However, the deleveraging in 2008-09 distorted relative valuations to a larger extent than in the UK or euro market.

Asset swap pricing
In a developed asset swap market, pricing will be determined by supply and demand for bonds, factors driving swap valuations and also supply and demand specifically for asset swaps. In the UK and euro area, liquidity in the different market segments is realistically developed enough for this usually to be the case. Elsewhere, it is rare for more than two elements to be well defined at any one time, although they increasingly can be in South Africa. In markets such as the US and Sweden, liquidity and activity is concentrated in bonds and there can be extended periods where asset swap spreads will remain relatively static. Effectively, the zero coupon swap curve in such a case is defined by the bond curve unless there is a factor that changes this dynamic. By contrast, in Australia, there is usually more activity in swaps, with asset swap valuations coming into focus only when there is bond or corporate asset swap activity. For a market in which the inflation swap curve is relatively well defined, linker asset swap pricing is best thought of as akin to that in the nominal market with the additional step of transforming all real cash flows into nominal form, as ultimately the valuation is versus Libor or equivalent nominal floating rate and not a real floating leg. Hence, the inflationlinked bond cash flows should be projected into nominal space using the inflation swap discount factors and then discounted by the nominal swap curve. The difference between this valuation and that of the inflation swap curve will then provide the asset swap value.
15 March 2010 28

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

There are, however, several ways that this calculation can be done, three of which are particularly relevant to inflation-linked bonds. The first of these is the par-par asset swap calculation that is the conventional quoting method for euro inflation-linked asset swaps. Most other markets use a proceeds asset swap calculation. Both of these methods create inconsistencies comparing different bonds though and in particular comparing inflationlinked bonds with nominals. For comparative purposes, we use a zero spread (z-spread) calculation instead. While conceptually this method is straightforward, it is almost entirely an analytical tool, with trading in this format extremely difficult as pricing is on an iterative base that would make the paperwork intractable in anything but a contract for difference format. Set out below is an example of a par-par asset swap calculation for the OAT€i20. The inflation index used to transform the real cash flows is that implied by HICPx zero inflation swap breakevens, and the discount factors are of the Euribor curve. Being par/par, it is necessary to net out the deviation from par of the bond’s dirty cash price, from the PV of the par priced bond flows, and the PV of the Libor leg, to find the PV of the asset swap assuming no spread to Libor. Via the swap PV01, the necessary spread to give the overall structure a PV of zero can be found, which is the mid or fair-value asset swap spread, in this case 42bp assuming negligible value to the embedded inflation floor. The par par asset swap is not a particularly logical calculation to use for bonds, which will tend to deviate from nominal par over their lifetime as they gain inflation accretion, but it is relatively straightforward and became convention in the euro area when most asset swaps were on short maturity bonds with limited history. It should be highlighted that the frequency of the Libor payments will not necessarily be the same as the cash flows on the bond in an asset swap. For example, the underlying nominal swaps market in Europe most commonly trades on a semi-annual payment convention, whereas OAT€is carry annual coupons.

15 March 2010

29

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Figure 6: Example of par-par asset swap calculation (using the OAT€i20)
Bond price Bond Base Inflation Index 123.16 96.08560 Real Cash flow date 15 March 2010 26 July 2010 25 July 2011 25 July 2012 25 July 2013 25 July 2014 27 July 2015 25 July 2016 25 July 2017 25 July 2018 25 July 2019 25 July 2020 25 July 2020 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 100% -100% Coupon Notional Projected index values from swaps 108.22 109.25 111.00 113.17 115.36 117.88 120.57 123.61 126.91 130.26 133.73 137.08 137.08 2.56% 2.60% 2.65% 2.70% 2.76% 2.82% 2.89% 2.97% 3.05% 3.13% 3.21% 42.67% Nominal Nominal swap

Cash flow discount factor 1.000 0.996 0.983 0.961 0.935 0.904 0.871 0.836 0.801 0.766 0.732 0.699 0.699

A B C

PV bond cash flows PV Libor flat leg Bond price - 100

56.63% 30.07% 23.16% 3.41% 8.98 0.38%

A-B-C=D PV ASW at Libor flat E D/E
Source: Barclays Capital

Nominal swap PV01 Par/par ASW

Proceeds asset swap spreads are generated in a similar manner to par-par calculations with some minor differences. In a proceeds spread, there is no exchange of cash flows upfront as the swap notional is set to be equal to the bond’s dirty price. Therefore, the biggest difference is that the bond premium (or discount) in step C of Figure 6 must be discounted from the maturity date, reflecting the fact that the swap notional will be different from 100 at the final maturity of the asset swap. Proceeds asset swaps avoid any issues with historical accretion and hence are more appropriate for markets where bonds have been accreting inflation for years, as most TIPS had when the market took off and particularly in the UK. A par par asset swap tends to exaggerate the value of the asset swap for a bond trading above par, as bonds with inflation accretion will usually be. For comparison, we consider the proceeds asset swap on the same OAT€i20 below, although euro bonds are rarely quoted in this format. Neither par-par nor proceeds asset swaps are an ideal valuation measure to compare inflation swaps versus bonds. The back-ended cash flows of inflation-linked bonds mean that the value of a floating basis point now is less than that of an inflation-linked bond discounted on its own real yield curve. Hence, a 1bp move in real yield will be worth more than a 1bp in asset swap in either of these methodologies. Deviations of asset swap levels away from Libor flat will tend to become more distorted the longer the maturity of the linker. In an extreme case, a 50y bond may have an expected average principal in present value terms around twice its current level and hence a 1bp move in the real yield of the
15 March 2010 30

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

bond with no move in the inflation swap curve will move the asset swap by around 2bp. It is thus relatively difficult to compare asset swap levels across inflation-linked bonds and particularly difficult to compare them with asset swap levels of nominal bonds. For this reason, we prefer to use a z-spread asset swap methodology.

Figure 7: Example of proceeds asset swap calculation (using the OAT€i20)
Bond price Bond Base Inflation Index 123.16 96.08560 Real Cash flow date 15 March 2010 26 July 2010 25 July 2011 25 July 2012 25 July 2013 25 July 2014 27 July 2015 25 July 2016 25 July 2017 25 July 2018 25 July 2019 25 July 2020 25 July 2020 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 2.25% 100% -100% Coupon Notional Projected index values from swaps 108.22 109.25 111.00 113.17 115.36 117.88 120.57 123.61 126.91 130.26 133.73 137.08 137.08 2.56% 2.60% 2.65% 2.70% 2.76% 2.82% 2.89% 2.97% 3.05% 3.13% 3.21% 42.67% Nominal Nominal swap

Cash flow discount factor 1.000 0.996 0.983 0.961 0.935 0.904 0.871 0.836 0.801 0.766 0.732 0.699 0.699

A B A-B=C D E

PV bond cash flows PV Libor flat leg PV ASW at Libor flat Nominal swap PV01 Bond price

56.63% 53.25% 3.38% 9.06 123.16 0.30%

(C/D)/E Proceeds ASW

Note: Example only - the OAT€I20 asset swap would not normally be quoted in this format Source: Barclays Capital

Z-spread asset swaps are a widely used analytical tool within nominal bonds to smooth out micro distortions and compare relative value, but are more important for linkers as the distortions are notably larger. The calculation of the z-spread asset swap is done so that, by construction, a 1bp move in the real yield of the bond will move the asset swap by 1bp as well. A z-spread asset swap spread between a nominal and inflation-linked bond thus provides a consistent measure to the richness of bond versus swap breakevens at that maturity. The calculation involves taking the cash flows of a linker in a similar way to a proceeds asset swap and calculating their present value. The next step is to iteratively adjust the uplift factor affecting the bond cash flows in a parallel fashion until the present value matches that of the same cash flows priced from the swap leg. The amount that the swap curve has to be bumped to match the value of the bond cash flows is the z-spread asset swap. It will be narrower than asset swaps calculated by the other methods, substantially so in the case of long maturity bonds.

15 March 2010

31

57 3.08560 Projected index values from swaps 108.25% 2. Any such attempt to prevent distortions remains fraught with difficulties.979 0.41 0.56% 2. and as practitioners become comfortable with using z-spreads for analytics but asset swaps for trading purposes.25% 2. the instrument can be a single asset. A total return swap is an over-the-counter financial contract that provides a tailored economic exposure to the performance of an underlying instrument.21% 142.16 96.65% 2.70% 2. and it would still remove much of the distortion. Alternatively.97% 3. an index of bonds. The receiver of the total return of a linker (or basket of linkers or linker index) accepts its economic performance but without owning it on its 15 March 2010 32 .935 0.60% 2. There have been multiple attempts made to make a more representative tradable asset swap measure for inflation-linked bonds.05% 3. Total return swaps Total return swaps exist in most asset classes but have become increasingly widely used to mimic the returns of inflation-linked bonds.16 While the z-spread asset swap is an ideal analytical tool.88 120.25% 2.784 0.801 0.733 0.13% 3.73 137.995 0.22 2.766 0.25% 2.678 0. This approach may be feasible as long-term swap contracts usually have break clauses embedded within them anyway so they could instead be reworded as a reweighting clause.25% 100% 109.678 Cash flow date 15 March 2010 26 July 2010 25 July 2011 25 July 2012 25 July 2013 25 July 2014 27 July 2015 25 July 2016 25 July 2017 25 July 2018 25 July 2019 25 July 2020 25 July 2020 Real Coupon Nominal Cash flow PV Difference Z-spread Source: Barclays Capital 126.57 123.25% 2.904 0. the calls for an alternative measure may reduce. every five years. such a reset process could be less frequent.25% 2.712 0.000 0.25% 2. fixing the accretion rate at the actual rather than expected level.997 0. for instance by allowing the value of Libor to accrete at an agreed rate upfront (either the breakeven inflation rate or a conventional amount) that will compensate for the back-ended linker cash flows.25 111.699 1. eg. Conceivably. though. for example. linker asset swaps could be reset with each coupon to allow for the accretion of the bond.747 0.962 0.926 0.25% 2.871 0.983 0.25% 2.955 0.26 133.699 0.00 113.820 0.000 0. such as a bond. but this would involve a new transaction after every coupon point.61 126.91 130. or a basket of assets. it is a very difficult format to trade.17 115.76% 2.857 0.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 8: Example of zero-spread (z-spread) asset swap calculation (using the OAT€i20) Bond price Bond base inflation index 123.08 2.08 137.25% 2.67% Nominal swap Nominal swap discount discount factor factor (shifted up 29bp) 1.836 0.36 117.89% 2.82% 2.892 0.29% 123.

In the UK. The market for total return swaps in developed inflation-linked market indices has increased sharply. where the complexity and volume of structured products issued has led to the development of a market for inflation caps and floors. In the US or Europe. It is important to note that if the total return of the underlying instrument is negative. with those with a lower rating potentially requiring less spread to hold the assets on balance sheet. In markets such as US TIPS. Total return swaps can also be used as a way for investors to short the bonds without directly having exposure to the repo market. Figure 9: Total return swap structure notional × totalretur nt . This is particularly the case in the euro area. While there is a counterparty credit risk involved in a total return swap. whereas in less liquid or lower-rated markets. the development of the non-linear market has been driven by demand to hedge liabilities. For example. A market for US CPI caps and floors exists but is notably more illiquid than in Europe. active managers can use a total return swap on the Barclays US Inflation-Linked Index to gain their Beta exposure. TIPStions. while options on TIPS. it is likely to cost more than Libor to receive the total return. where general collateral repo yields are significantly lower than Libor. The credit rating of the institution offering the swap can also have a bearing on the pricing. The additional advantage for indexed investors is they do not have to rebalance their portfolio exposure as their beta index changes once it is held in total return swap format rather than in physical bonds. and the expertise developed to price them has encouraged the development of exotic inflation products elsewhere. receiving a total return swap is likely to be at Libor minus. 15 March 2010 33 . . are also quoted. simultaneously freeing up capital to pursue Alphagenerating trades. The total return swap effectively converts their benchmark from the index to Libor plus the spread charged for using the dealer’s balance sheet. LPI swaps are path-dependent capped and floored products. The spread applied to the floating side of the swap is required to compensate the dealer for the balance sheet cost of holding the asset or assets underlying the swap agreement.t −1 Counterparty A notional × (Libor ± spread ) Counterparty B Source: Barclays Capital Non-linear inflation derivatives The market for non-linear inflation derivatives has developed mainly on the back of hedging activity related to structured bond products issuance. The total return of the bond or bonds is received versus paying out a floating rate cash flow. partly due to the movement toward separating ‘Alpha’ and ‘Beta’ in portfolios. which are subject to a limited price indexation (LPI). One notable feature when distortions between bonds and swaps were very high in late 2008 was the use of total return swaps to express linker asset swap positions. although liquidity and activity remains poor and episodic. this can be mitigated by ISDA/CSA agreements or by frequent resets of the swap. it is likely to be a relatively cost-effective leveraging strategy for investors who are unable to repo bonds directly themselves.Barclays Capital | Global Inflation-Linked Products – A User’s Guide balance sheet. then the counterparty receiving the economic performance will have negative cash flows on both legs of the total return swap.

Range accrual notes helped push down cap/floor vol in the euro area significantly. with an abrupt break in September 2008. The pay-off of most bonds was relatively simple. although municipal issuance of inflation-structured notes has become an increasingly important element of the market and here it is demand for frequent upfront cash flows rather than taxation that means they have a similar structure. 15 March 2010 34 . with the total not significantly above that of 2003. Similar notes have also been sold with a UK RPI versus euro HICPx spread. Issuance of structured bonds with coupons floored above 0% became increasing popular. With realised inflation pushing towards 3% in 2007. with coupons set as a fixed rate X% plus year-on-year inflation rate. A combination of inflation and nominal CMS coupon-linked structured notes proved to be relatively popular with investors. These notes helped to correct a relatively extreme richening of French swaps versus euro inflation that had been driven by strong Livret-A related hedging demand for French CPIx. The market that has by far seen the highest level of issuance is Europe. the majority of inflation-structured notes were of this simple X% plus y/y inflation format and sold to individual investors. 2004 saw the emergence of leveraged notes. often forward starting. a large fixed nominal or floating rate would be paid if inflation remained within the range. floating rate notes have been offered at Euribor + Y%. in contrast to low realised inflation volatility. the main focus in euro inflation-linked structured note issuance was on inflationrange accruals. which typically paid a multiple of y/y inflation or a fixed rate plus a multiple. Buyers of such range accrual notes were effectively sellers of inflation volatility. 2005 and 2006 was €10bn. but investors in these bonds suffered when actual inflation started to move well above the upper end of the ranges offered towards the end of 2007. respectively. which typically aimed for the y/y inflation to remain in a tight range around the ECB’s target of inflation close to but below 2%. Similarly. Notes linked to the US CPI developed in a similar format. although with the y/y linked coupons paid on a monthly basis rather than annually as in Europe. particularly in 2006. for example. a note might pay the better of inflation or 10y CMS. In 2006-07. but the leverage factor in most notes issued during these three years implied a much greater impact on the swaps market. In addition to having an inflation floor. In such structures. compared to a low or no payout otherwise. sometimes higher than zero percent. with the euro area seeing over €18bn in issuance. such leveraged notes often have an inflation cap embedded within them as well. Coupons were usually floored at zero. The leverage factor amplified not only the impact in swaps space but also the implicit notional on any embedded floors. Activity in structured inflation products took off in 2003. but capped at a multiple of euro HICPx. Until 2004. For instance.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Structured inflation products The market for structured inflation bond products took off in 2003. although some issues involved lower fees and were targeted at institutional investors. Coupons were floored. One format that had a particularly important effect on the underlying markets was notes that offered a leveraged exposure to the spread of euro and French inflation. as the deflation scare and funding levels of issuers limited activity considerably. although floors set below zero were relatively common in 2004. but floors set at the fixed rate X% were not uncommon. Issuance in Europe during 2004. This was largely due to tax treatment for individuals in the US being more suited to the more frequent cash flows. These products therefore offered interesting value as implied cap and floors volatility was relatively high at that time. cap/floor implied volatility drifted higher as some assumptions in pricing models were changed but also because of some unwinding of the loss-making range accrual notes. €3bn and €7bn. or alternatively the worse of the two plus X%. especially given that the US experienced sub 1% y/y inflation in 2002. but activity has varied considerably in subsequent years.

One of the first large structured inflation notes in the euro area. was linked to the performance of the Italian MIB Stock Index but with the return floored at an inflation-protected principal. Inflation-linked credit products. plus euro-denominated issuance in Danish and Polish CPI. for instance inflation-linked CDOs. Products paying inflation with leverage once again became widespread. the fear of high inflation boosted demand for inflation-linked structured notes from retail investors. Hybrid inflation-linked notes have been a feature of the structured product market almost since its inception. The majority of notes sold there in 2003 and 2004 went to individuals via bank distribution networks. both in local and hard currencies. then became more popular. there is scope for hybrid notes to return as real yields rise.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Italy was the country in which the demand for structured inflation notes took off most notably. however. and in the euro area in particular. most commonly Spanish CPI. which significantly increased inflation fears despite negative realised inflation. Even though implied vols and swaps surged to unprecedented levels. and even before the worst of the market turbulence in 2008. All relied on a sufficiently high real yield to leverage up on the risky leg. Risk tolerance subsequently fell. demand for this kind of product fell sharply as real yields fell in 2005. Many of these notes were structured with a higher leverage than before and with floors above zero. but leveraged notes have been scarce. Similar best-of-hybrid products have also been seen. While more highly structured inflation-linked notes are unlikely to see any significant resurgence for some time. for instance notes linked to the S&P500 and the Euro Stoxx indices. Issuance started to recover again in Q2 09 after global central banks had engaged in quantitative easing strategies. as focus turned to deflation while an explosion in volatility made the cost of floors embedded within most structures extremely expensive. Issuance also thrived in the first half of the year as banks in particular took advantage of expensive credit/funding for financials. Other hybrid notes offered the best of inflation accretion and commodity basket performance. helping to lead to a rebound in total structured note issuance in 2006. While most issues sold in Italy were linked to euro inflation. there were inflation coupon notes sold linked to at least seven different euro area domestic inflation measures. there were as many notes linked to Italian inflation in these two years as total issuance linked to US CPI. sold by the Italian Post Office in 2001. there was a tendency towards relatively straightforward products. 15 March 2010 35 . the high funding levels meant that sufficiently attractive pay-offs could easily be structured to cater for the retail investor base. issuance dried out. In 2006 and 2007. Figure 10: Structured note inflation issuance in US CPI and euro area indices (bn) 25 US Euro (US$ equivalent) 20 15 10 5 0 2002 Source: Barclays Capital 2003 2004 2005 2006 2007 2008 2009 In 2008. A variety of other emerging market inflation indices were also offered in y/y inflation note formats. After summer in 2008.

while activity at the long end has also picked up notably on the back of pension fund demand. rather than a yearly basis. activity in Japanese core CPI floors became focused on final maturity. given the preponderance of pension liabilities that are LPI linked. The same pressures were seen even more extremely in the US despite the size of exposures being smaller. 1% and 2% floors and 2%. most commonly 10 years. having gravitated around structured note activity in Europe. which effectively killed off the market for competing products whose main additional feature was principal protection. When structured note demand in the US drove cap/floor vol there to several times that in euro HICPx.75% on the floors and 2. the cash flow on the first inflation caplet/floorlet is not known in advance at the trade inception date. 3% and 4% caps. This became very difficult in the US and euro area in late 2008. pricing of low strike floors comes most obviously from hedging with low strike nominal options. Nonetheless. As swap rates fell towards the end of 2008. High floors could previously be structured in notes as they were deep out of the money when inflation swap rates were stable around 2%. It was never a feasible proposition in Japan. the index reference date conventions are the same as their underlying swap markets: for example. However.5% to 3. having spiked higher earlier in the year. For standard caps and floors traded in the broker market. will be structure of 10 caplets with a predetermined strike rate. with the cash flow equal to the notional times the difference between the realised y/y and the strike rate. The deltahedging led to a self-reinforcing bearish momentum in euro HICPx swaps and cap/floor implied vol in the euro area shot higher. the caplet will pay a cash flow if year-on-year inflation is above the strike rate. 15 March 2010 36 . the development of the range accrual market in particular has helped better define the cap/floor vol smile. Inflation caps and floors are traded in most of the major inflation bases and not just because of embedded floors in structured notes. creating a rush to hedge. A 10-year cap. the sizeable street short in floors as a result of past structured issuance meant that losses had to be marked-to-market. Caps and floors are most commonly quoted at 5y and 10y maturities. as the strikes on these notes occurred at a varied range usually between 1 and 1. It did prompt some structured notes being sold elsewhere in the world taking advantage of this pricing though.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Caps and floors The inflation cap and floor market is most active in euro HICPx. A cap/floor is a series of caplets/floorlets on y/y inflation. Participants in other vol markets are increasingly getting involved in inflation caps and floors. even though breakevens were notably lower than in previous years. whether in floors or swaps. prices of 0% floors in Japan became less expensive in 2007. 0%.25% on the caps. trading on ZC euro HICPx swaps in July occur with an April base month. there was flow to hedge one with the other and notes also offered on this basis. One element to highlight is that unlike in nominal caps/floors. In a market where there is no underlying supply of inflation volatility. with the vol skew relatively extreme. This was only partially offset by some notes including caps too. even though there was very little activity in LPI swaps in 2008 and 2009 as realised inflation volatility exploded. The skew stemming from structured note floors in the euro market has been traditionally extreme. Coupled with the relaxation of investment restrictions on JGBis. While caps and floors related to structured notes have traditionally been the most important flows. This means that euro HICPx caps traded in July will have cash flows determined by y/yr for the months of April over the life of the cap. Each year. with this market also relatively distressed. which combined with initial restrictions on investment on bonds without a guaranteed principal led to floors pricing in Japan at extremely expensive levels. for instance. with a concentration of interest at integer values. The UK cap and floor market is relatively well defined as a result of LPI curves. limiting the degree to which it was the options market that drove dislocations in late 2008. by 2007 the market started to reach a point where flows unrelated to underlying product were becoming almost as important.

LPI is interesting from a modelling perspective because the annual accretion means that it is path-dependent. A much less frequent pension linkage. If products replicating the more exotic formula were structured. is considerable. However. LPI and other path-dependent options Conceptually.3). At the start of 2009. an element of non-linearity was added to the formula. the official rate-setting was further incremented with the condition that the magnitude of a rate change cannot exceed 1. Another significant inflation liability that involves an LPI formula. Models that were developed to price LPI swaps can generally be extended to capture most other non-linear pricing that has been seen in the inflation and real rate space. but negative inflation prompted some schemes to keep benefits static while others let them shrink. capped at 4% and floored at 0. there is an argument that de facto most RPI liabilities would behave as LPI(0.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The inflation volatility market in France grew as a result of liability-driven demand linked to the Livret A savings rate. there has been no demand to precisely hedge the new formula.5) exposures. Subsequently. which used to be the legal minimum. totalling over £500bn.5). but one which is still relatively important due to its popularity in management pension schemes even after 1997. but with the reduction in the prevalence of defined benefit pension schemes it will be many years before the size of liabilities on this lower base approaches the size of LPI(0. Modelling of inflation volatility product is discussed in more detail later in this publication in the section ‘Inflation Volatility Pricing Considerations’. but a zero floor is explicit in more recent Pensions Acts. meaning that simple cap/floor vol is only an approximation and that simple two currency inflation or real rate volatility models will not provide closed solutions. Another relatively common LPI linkage from pre-1997 that has prompted demand in recent years is LPI(0. The most common linkage. there is almost no incentive for new pension liabilities capped at 2. with it being the minimum of French CPIx y/y +0. the rate was computed as an average between 3-month Euribor and the y/y FRCPIx rate.5% to be hedged with RPI inflation.25% and 0. this would have significantly increased the implied vol at lower strikes in both inflation swaptions and caps. From 2008. Indeed. the underlying demand in the UK for LPI swaps. with supply of LPI from property rentals often at non-standard levels (albeit rarely outside the 0-5% range). albeit one 15 March 2010 37 . It was also decided that the Banque de France could recommend a change in the rate even between the twice-yearly revision dates. is LPI(3.5x the sum of CPIx and the average of 3mth Euribor and eonia. The y/y inflation linkage involved in this rate meant a significant degree of convexity pricing versus zero coupon swaps even though it had no theoretical floor attached to it. with the rate being set arbitrarily for most of 2008 and 2009.5). the most common of which is LPI(0. given that y/y RPI had not been negative from 1960 till then.infinity). LPI is not a unique concept to UK RPI. or LPI(0.infinity). due to the unacceptability to the relevant parties of pension payments being reduced in nominal terms.5) but only the cap. the legal minimum for 1997-2005. 1990s legislation arguably did not enshrine LPI(0. some pension liabilities linked to Irish CPI are also in an LPI format. This was a moot point until 2009. A result of this is that the development of the UK LPI swaps market in the last decade encouraged complexity to be added elsewhere. which accrete annually at the rate of RPI inflation capped and floored at a certain level. There are some instances of floored but uncapped exposure. It does help define a broad smile in the UK market though. All the above contributed to weaken the once strongly perceived formulaic link between French inflation and the Livret A rate. From 2004-08. With long RPI breakevens above 3% since 2005.2.5%. In particular. Most UK pension liabilities are LPI linked. the legal minimum was cut to LPI(0.5). was a cap at 5% and a floor at 0.

at 8. there are some challenges that makes pricing non trivial. Prior to 2009. What limited structured inflation note market there is also encourages activity in long tenors. although there is no interbank market. curve off which to price such exposures for less than one year. The market convention for quoting the options is typically in terms of at-the-money spot. the presence of inflation futures at least helped to offer a fully visible. while this is not trivial for bond options. Swaptions Swaptions in both UK RPI and euro HICPx area can be quoted. but the definition of the LPI swaps activity helps to provide a reference volatility curve and smile as a starting point for pricing. Swaption activity remains uncommon compared to LPI activity.Barclays Capital | Global Inflation-Linked Products – A User’s Guide where the cap and floor are so far out of the money that it is often ignored. Quoting in this manner reduces the complications involved with calculating forward prices in TIPS.75%. as well as the Netherlands becoming an increasingly important source of demand. is the so-called TFR rate in Italy. whereas a put TIPStion pays the maximum of the strike minus the price at expiry and zero. The more complete definition of the inflation swaps curve makes this pricing more straightforward. Both real rate and breakeven swaptions can be traded. A call TIPStion gives the buyer the right to purchase TIPS at a preset price on the expiration date of the option. A well defined swaps curve. TIPStions pay off just like standard options with the expected pay-off for a call being the maximum of the price at expiry minus the strike and zero.5% + 75% of Italian FOIx CPI inflation but with a cap of 7. Nominal swaption volatility remains the most important driver of real yield vol. The presence of conditional inflation indexation of Dutch pension funds has created the prospect for the supply of inflation volatility. while activity in the inflation cap and floor market helps to define breakeven swaption vol pricing. with increasing activity in long-term notes that are callable or puttable. Swaptions in the euro area have involved relatively short tenors. interest in the swaptions market is usually for very long-dated expiries and maturities. albeit illiquid. TIPStions Options on TIPS. The evolution of the TIPStions market was a natural extension of the TIPS market in that it gives portfolio managers and other real money investors an easy way of managing portfolio risk and returns without having to make outright purchases or sales of their TIPS holdings. While TIPStions are simple products at first glance. 15 March 2010 38 . The accretion of TFR funds. though activity in 2009 was almost non-existent. It also allows leveraged investors to implement strategies they have used in the cash market in a more efficient manner. with most options being for five years or more. We look at the pricing aspect in more detail in the ‘Inflation volatility and delation floors’ section. In the UK.33% inflation and at least implicitly with a floor at -2% inflation. is 1. known as TIPStions. The development of pension fund interest in swaptions helped drive demand for longer-term expiries as well as maturities. and options on TIPS breakevens have been trading in the OTC market since early 2005. traditionally a percentage of wages withheld and then paid when a worker leaves a company but now increasingly transferred into pension holdings. ie. while the most common maturity is 30 years. which is the case in both UK RPI and euro HICPx facilitates forward pricing in swaptions.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide INFLATION MARKETS 15 March 2010 39 .

as measured by the WTI. reaching a value of $600bn by February 2010.7% Medical Care 6. 10y and 30y maturities.3% Housing 39. although total return swaps on TIPS are regularly used for leveraged exposure. Rents data are gathered from 40. one of the most significant categories in terms of weights is housing. which are classified under eight major groups. Relative importance in essence means that if the price of a particular item rises more than the average price increase of items in the basket. Annual weights are typically released in mid-February for the previous year.1% Medical Care 5. TIPS had overtaken the UK inflation-linked market to become the largest market of its type.3% tion 5. The Inflation Index – CPI-U The Inflation Index used for TIPS is the not seasonally adjusted US City Average All Items Consumer Price Index for all Urban Consumers (CPI-U). The CPI-U measures price changes for urban consumers of a fixed basket of goods and services of constant quality and quantity. Barclays Capital Housing 43. To illustrate.6% Apparel 4. Since the CPI is a fixed-weight index.4% Food and Beverages 16.3% Transportat ion 17.4% tion 6. over half of which is an imputed measure called ‘Owners’ equivalent rent of primary residence’ (OER). the price of crude oil. then the relative importance of that item increases.5% Recreation 6. The Treasury has varied its issuance pattern over time and currently issues at 5y.6% Figure 12: December 2008 CPI-U weights Other Education Goods and and Communica Services Food and Beverages 3. which attempts to capture price changes if those consumers who 15 March 2010 40 . Figure 11: 1997 CPI-U weights Other Education Goods and and Communica Services 4. corporate and inflation swap activity is limited.9% Source: Haver Analytics. in 1997.2% to 9. A related concept is the relative importance of an item. most commonly known as TIPS. A result of that increase is that the relative importance of energy rose from 6. Despite the depth of the TIPS market and the commitment from the Treasury.com The US Treasury began issuing inflation-linked bonds.000 retail and service establishments. Prices are collected for over 200 categories. which include 21. The basket of goods and services and the item weights are determined from the Consumer Expenditure Survey (CEX).6% Apparel 3. the implicit weights remain the same from month to month.pond@barcap.3% 15. had risen from around $20 per barrel in January 2002 to near $90 per barrel in December 2007.7% Source: Haver Analytics.Barclays Capital | Global Inflation-Linked Products – A User’s Guide US Michael Pond +1 (212) 412 5051 michael.4% As can be seen from Figure 12.000 landlords and tenants. The charts highlight the change in the relative importance of the eight major categories from 1997 to weights set in December 2008. Barclays Capital Transportat ion 15.7% during the same time period. By 2000. Prices are collected from 85 urban areas.8% Recreation 5.

the weightings do change from month-to-month. This weight has grown since TIPS were first issued. While the BLS only publishes the relative importance for each December. As utility prices tend to fluctuate more than actual rents. home fuel oil and utility gas service (natural gas).gov/press/releases/rr1073. Gasoline (Motor Fuel) is the most important component of Energy CPI. Before 1983. so energy is even more important to inflation-linked investors now than it had been in prior years. number of rooms and type of air conditioning. BLS calculates the pure rent of the matched renters by removing the value of any landlord-provided utilities and furniture.7% of the total CPI basket.Barclays Capital | Global Inflation-Linked Products – A User’s Guide own their home were to rent instead. The characteristics included location. the BLS used an asset price approach in computing the shelter component of CPI. primarily natural gas prices. Energy currently makes up around 8. BLS measures the change in implicit rents over time by matching owner units to renter units with similar characteristics.htm) President Clinton and Treasury Secretary Rubin announce the terms and conditions of the first Treasury inflation-indexed security First 10y TIPS auction First 5y TIPS auction First 30y TIPS auction Final rules on fungible inflation-indexed STRIPS published Treasury begins selling series-I savings bonds Treasury announces regular quarterly schedule for TIPS and discontinues 5y TIPS Fed conducts first TIPS pass TIPS are stripped for the first time Treasury eliminates 30y TIPS because of lower borrowing needs First 5y TIPS matures Treasury conducts TIPS buyback Treasury expands 10y TIPS auctions to 4 per year with two new issues per year CPI futures begin trading at CME Treasury announces the introduction of 20y TIPS and reintroduction of 5y TIPS First 20y TIPS auction First reintroduced 5y TIPS auction First 10y TIPS matures TIPS Index market value hits $500bn First reintroduced 30y TIPS auction Source: US Treasury. both because its weight is higher and because it tends to be more volatile than the other components. Barclays Capital 15 March 2010 41 . the BLS relative importance data understate the average weight of energy. As seasonal factors affecting Energy CPI are at their worst in December. electricity.treas. Figure 13: TIPS program historical milestones Date 5/16/1996 9/25/1996 1/29/1997 4/9/1997 4/8/1998 6/30/1998 9/1/1998 9/29/1998 1999 11/30/2000 10/31/2001 7/15/2002 4/18/2002 4/30/2003 2/8/2004 5/5/2004 7/27/2004 10/26/2004 1/15/2007 1/22/2008 2/22/2010 Event Treasury Secretary Rubin announces the intention to issue Treasury inflationindexed securities (www. imputed pure rents tend to be negatively correlated with utility prices. structure type and other general traits such as age. because this method was driven by interest rates and house prices it was much more volatile than the current method and core CPI volatility has declined since. As owners pay for utilities separately. Energy prices can also have a significant impact on m/m CPI prints and in fact are historically responsible for over 50% of the volatility there.

the more common name used by market participants and the Treasury is Treasury Inflation Protected Securities (TIPS). average daily trading volume leveled off because of an increase in demand from structural investors such as pension funds and insurance companies where investments tend to be passive in nature. They also announced a major expansion of the program to include two 5y auctions and two 20y auctions per year. as the Treasury issued nearly as many TIPS that year as it did in the previous three. After the market cheapened significantly following this deleveraging process. with just one re-opening auction. The US Treasury first issued inflation-protected securities in 1997. CPI futures began trading at the CME in early 2004 and volume in the CPI swaps market increased significantly along with issuance in inflationlinked corporate notes. There was a significant shift in the investor base beginning in the second half of 2008. Despite this commitment. the iSTRIPS market was launched when the TII08s became the first TIPS to be stripped. existed in 2001 (Figure 14). While iSTRIPS have not been a huge success. the Treasury has come full circle and will again issue at those maturity points in 2010. iSTRIPS allow investors to trade the TIPS coupon and principal components separately. 10y and 30y securities and after many changes to the calendar. diversify its debt service costs and create liabilities that were more closely aligned with the government’s main asset – tax revenues. the Treasury reduced TIPS issuance commensurate with reductions in the nominal calendar until only an annual 10y note. although to date there has been only scarce interest in iSTRIPS. Hedge funds. There was also increased interest in real return mutual funds and other funds tied to the TIPS Index as investors began to make diversification allocations into TIPS as a new “asset class”. which is adjusted for inflation. in addition to the existing quarterly 10y note auction cycle. after making up around half of TIPS market flows in 2007. Increasing demand led to significant growth in 2004. we opt to use the latter given that it is market convention. Related to this demand. The inflation index used is the not seasonally adjusted headline CPI. there was significant growth in Inflation-Linked Total Return Swaps activity from investors looking to receive the return of the TIPS Index or a Global Inflation-Linked Index in a passive nature. but not in between. real money investors began to increase structural allocations significantly. in order to broaden its investor base. In this article. it was important that the Treasury encouraged stripping activity as a sign of it’s commitment to the TIPS program when many market observers were questioning the durability of the asset class. Alongside growth in the cash market was a developing inflation derivatives market. From 2005 to mid-2008.Barclays Capital | Global Inflation-Linked Products – A User’s Guide TIPS market history While officially called Treasury Inflation-Indexed Securities (TIIS). There is a par floor on the principal at maturity so the investor is protected from deflation from issue date to maturity. The small number of participants resulted in both low trading volume and a low beta relationship to nominal yields. particularly on the derivative and structured note side. Initial issuance was in 5y. Foreign central banks began to buy because of diversification benefits and as a de-facto currency hedge. where the principal component carries the floor. Consultants began recommending TIPS in earnest and due diligence and approval processes were introduced. The US inflation market continued to develop in 2005. Domestic 15 March 2010 42 . In November 2000. With five years of history and the 5y TIPS issued in 1997 having matured in 2002. largely exited the asset class during the financial crises. Breakevens were generally the main catalyst for investment decisions. A historical synopsis of the TIPS program is presented in Figure 13 There was limited initial support for TIPS as investors used them primarily as a tactical trading vehicle. TIPS are structured such that they pay a fixed coupon on a principal amount. the TIPS market finally started to gain broader acceptance.

TIPS issuance summary The following chart shows annual TIPS issuance since 1997. The Treasury held TIPS issuance steady at $5658bn gross and $35-38bn net during 2007-09.Barclays Capital | Global Inflation-Linked Products – A User’s Guide real money increased structural allocations in part because of medium-term inflation risks associated with stimulative fiscal and monetary policy and also because of a realization by many that the right allocation to TIPS within a well diversified portfolio is not zero. Figure 15 shows that TIPS issuance had also been generally increasing relative to other coupon-bearing Treasury issuance. but the other Treasury issuance was actually being reduced so the TIPS market grew on a relative basis. Barclays Capital Net Issuance Ex-SOMA 15 March 2010 43 . TIPS issuance grew only modestly. While issuance in the first three years of the program was just above $30bn per year. even as it increased nominal issuance significantly and the percentage of Treasury debt represented by TIPS fell from a peak of 10. issuance began to grow and the $63bn issued in 2004 was equal to all TIPS issuance in the prior three years. Excluding SOMA) Source: US Treasury. however. After picking up in 2006. by $6bn. TIPS issuance slowed in 2007 both on an outright basis and relative to nominal coupon issuance as the Treasury became proactive in slowing the growth rate of the program as it matured. Even with these increases. we expect TIPS as a percentage of outstanding Treasury debt to decline further to around 7% before starting to head higher in 2011 or 2012. was not a reflection on the TIPS program but rather on the impact of budget surpluses. from 2004-05. With the return to deficits in 2002. The reduction. particularly since 2003. Figure 14: Annual TIPS issuance ($bn) 90 80 70 60 50 40 30 20 10 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 (expected) Total Gross Issuance ($bn. where all Treasury issuance was being reduced. it declined to only $16bn in 2000 and 2001. After recommitting to the asset class and aiming to improve liquidity of the program.6% in late 2008 to less than 8% by the end of 2009. the Treasury has guided the market to expect $80-85bn gross TIPS issuance in 2010 and potentially going as high as $125bn in 2011.

trading volumes tended to spike only around auction weeks as they were seen as liquidity events in an otherwise low-volume product. supports this view. While auction periods are still seen as liquidity events. While trading volume increased only slightly over the 2000-02 period. Trading volume has since levelled off around $8-10bn/day. average daily trading volume increased significantly in late 2004 and was $3. this pattern has changed and trading volumes are more consistent. In the initial years of market development. Barclays Capital 15 March 2010 44 . The significant increase since 2005 in investors looking to receive the returns on the TIPS Index on a Total Return Swap.95bn in 2004 and took a sizable jump up to $8. $5. when it averaged $1.87bn per day. Figure 16: TIPS as a % of Treasury debt 16% 14% 12% 10% 8% 6% 4% 2% 0% Feb 97 TIPS as % of Treasury marketable debt TIPS as % of Treasury marketable debt (excluding bills) Feb 99 Feb 01 Feb 03 Feb 05 Feb 07 Feb 09 Feb 11 Source: US Treasury. who typically take more of a buyand-hold approach to investing and so tend to trade less. average daily trading volume has declined to a trend near $4. which by nature is a passive vehicle. Barclays Capital Increased issuance has driven an improvement in liquidity as well. This is because of the structural shift from hedge funds to real money investors. We believe this trend is being driven by increased activity by long-term investors whose structural investments tend to be passive and hence of a lower turnover than more tactical investors. Since the middle of 2008.73bn in 2003.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 15: TIPS issuance as a % of all Treasury note and bond issuance 12% 10% 8% 6% 4% 2% 0% 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Estimate TIPS as % of Total Gross Coupon Issuance Source: US Treasury.77bn in 2005.5-5bn.

Barclays Capital TIPS Index Market Value (RHS) Returns Through 2009. where the security pays a fixed coupon on the inflationadjusted principal. the biggest underperformance was -18.2% in 2009. Barclays Capital TIPS structure TIPS. Relative to the nominal comparator Index. but that was followed by the largest outperformance of 16.7% in 2008. This compares to the 6. The greatest annual return for TIPS was 17% in 2002 and the worst was -1.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 17: TIPS market value and average daily volume ($bn) 12 550 10 8 6 4 2 Jan-00 Mar-01 May-02 Jul-03 Sep-04 Nov-05 Jan-07 Mar-08 May-09 450 350 250 150 50 Rolling 3-month Daily Average Source: Bloomberg. along with most major inflation-linked bond markets except the majority of issues in the UK. follow the Canadian model. Figure 18: TIPS – historical performance and risk 20% US IL Returns US IL Ann Mthly Vol Figure 19: Return/risk versus nominals and equities 6 5 Equity IL Nominal Bonds 15% 4 3 10% 5% 2 1 0 0% -1 -2 -5% 1999 2001 2003 2005 2007 2009 -3 1998 2000 2002 2004 2006 2008 Source: Barclays Capital Source: Thomson Datastream. TIPS have had an annualized since inception return of 6.57% annualized return on a basket of comparable maturity nominal Treasuries. Federal Reserve. The principal is adjusted on a daily basis using an index ratio that quantifies the rate of growth in inflation or deflation between the issue date and settlement 15 March 2010 45 . the only year that the TIPS Index had a negative return.4% in 2008.65%.

the floor value of newer bonds tends to be more valuable because the index ratio is typically lower than seasoned TIPS. This means that there has been 2. is 1. the CPI-U for October 2009 applies. For example. and likewise for the par redemption amount (with the cash value subject to the par floor). so the inflation accrued since issuance would first need to be fully reversed out in a period of deflation. With about 3. Since the inflationadjusted principal is the par amount times the index ratio (which is the ratio of the reference CPI to the base CPI). The index is lagged three months from the settlement date. at maturity. this is another way of saying that. Reference Index = CPIm-3 + (t-1)/Dm x (CPIm-2 – CPIm-3) where: CPIm-2 = is the price index for month m-2 CPIm-3 = is the price index for month m-3 Dm = is the number of days in month m m = is the month in which settlement takes place t = is the day of the month on which settlement takes place. It is important to remember that the “strike” on the floor is at par. for 1 January 2010. the (real) semi-annual coupon rate is multiplied by the Index Ratio. For coupons paid.02332.99858.Barclays Capital | Global Inflation-Linked Products – A User’s Guide date. the investor gets the greater of par or the inflation-adjusted principal. Par × IndexRatio} The floor applies only to the principal amount at maturity – coupon payments can be and have been paid off an inflation-adjusted principal amount less than par. which is the trading price and does not include either the inflation or coupon accrual. We compute the index ratio as follows: Index Ratio = Reference Index/Base CPI Index. but the deflation effect on coupons becomes significant only in severe deflation environments. the TIIApr13 Index ratio. or an index ratio of 1. for example.71% annualized deflation to maturity for the floor to be at the money at maturity. TIPS have an embedded floor such that at maturity. The pay-off on the principal amount at maturity can be written as: Max{Par . and accrued are each multiplied by the Index Ratio to arrive at a cash settlement amount. 2010. the index ratio is floored at 1 as it applied to the principal. where the Base CPI Index is the Reference Index at issue date and. there needs to be 0. for a trade that settles on January 25. 15 March 2010 46 . Clean price. The floor would kick in if the index ratio fell below 1.2 years left. the first coupon payment on the TIIJul16s was paid off an index ratio of 0.33% cumulative inflation accrued since issuance in April 2008. For settlement amounts. The effect on the coupon can vary somewhat across issues with different coupons in significant deflation. As an example. The fact that there is not a floor on the coupon payments does complicate the calculation profile of TIPS somewhat. Floor In addition to the above structure. For this reason. not where the index ratio is at the time of purchase. real accrued interest is calculated as for ordinary Treasuries.

we have shown that they are not necessarily penalized on the expected after-tax total return versus nominal Treasuries. Major holders of US inflation-linked bonds A broad range of investors have been involved in the TIPS market. The final regulations generally require holders and issuers of inflation-indexed debt instruments to account for interest and original issue discount (OID) using constant yield principles. Other inflation-linked structures pay out the inflation on a monthly coupon rather than accreting on the principal.5%. while there are an increasing number of real return funds for whom TIPS are the core asset. In addition. as typically state taxes are paid on these bonds as well. albeit with an increasing amount directly mandated from pension funds and endowments. Tax On August 25.gov/gsrintax. To ameliorate this problem the Treasury in 1998 issued a Series I Savings Bond program targeted at individual investors. which for non tax-exempt investors such as insurance companies and individual investors may make ownership in TIPS unattractive. although this has been changing and we continue to see strong demand from long-term structural accounts such as pension funds and insurance companies. 1999. the inflation escalation of principal in the US is taxable as income annually. the Internal Revenue Service published ‘Final regulations’ covering the tax treatment of inflation-indexed instruments. so the investor is taxed on income actually received. many retail or other taxable investors view nominal Treasuries and corporate inflation-linked notes as more tax efficient. While it is true that TIPS are disadvantaged from a cash-flow perspective. However. the value increased significantly: for example. during the financial crises when breakevens out to the 9y turned negative and investors were increasingly risk averse.treas.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Before September 2008 the market put very little value on this embedded option. at one point the real yield spread between TIIApr13s and TIIJul13s was 200bp with most of this difference explained by the floor. The majority of TIPS holders continue to own them as diversification from core positions. The non-linear inflation market began quoting cumulative caps and floors around early 2009 and the value of the embedded option could then be priced separately from TIPS. Rules and regulations governing the tax treatment of TIPS can be found at the following link: ftp://ftp. particularly when real yields have moved above 2. However. These bonds are tax exempt for 30 years.pdf. Core and core plus-type total return funds and bond funds now commonly hold TIPS within their portfolios. Owing to this ‘phantom income’ tax issue associated with TIPS. the final regulations generally require holders and issuers of inflation-indexed debt instruments to account for inflation and deflation by making current adjustments to their OID accruals. but a key paragraph is detailed below: “The final regulations provide rules for the treatment of certain debt instruments that are indexed for inflation and deflation. Investors should consider the entire document.publicdebt. the advantage is not as clear cut as many perceive.” Thus. including Treasury Inflation-Indexed Securities. even though the Treasury will be making the inflation payment at maturity. This creates a phantom inflation tax. Mutual funds were the heaviest early buyers of TIPS and remain the largest managers of the asset class. Endowments and lottery funds have also 15 March 2010 47 .

which remain skewed heavily towards equities.gov/nara/cfr/waisidx_02/31cfr356_02. A TIPS security can be divided into its two components – coupon and principal.Barclays Capital | Global Inflation-Linked Products – A User’s Guide proved natural buyers of TIPS. The embedded floor in TIPS only 15 March 2010 48 .html. Foreign official institutions have become an increasingly important feature of the market. To date there has been only scarce interest in iSTRIPS. although inflation-linked structured issuance by insurance companies has become more common. Hence. All TIPS issues are now eligible for stripping and Barclays Capital has been an innovator in this area. state and local government pension scheme liabilities have relatively more explicit price indexation rather than wage indexation. first stripping TIPS in November 2000. The link for this register is as follows: www. Each coupon cash flow. The insurance sector is notably less important than in Europe. the development of a full yield curve has led to increased issuance. Importantly. On the other hand. increased investor demand and the development of a deep and liquid market. but until 2009 remained relatively small in TIPS compared to their nominal Treasury holdings.gpo. The US Federal Register sets forth basic conventions for the stripping and future settlement prices of zero-coupon inflation instruments. The absolute scale of private-defined benefit inflation-linked liabilities is considerably smaller than the more than $2trn state and local government sector. iSTRIPS The US Treasury over the years has developed the inflation-indexed security market in a similar fashion to the nominal Treasury market. and also much more frequently have indexation commitments beyond the period when a member of the pension scheme is an active contributor.36 Appendix B. STRIPS is an acronym for Separate Trading of Registered Interest and Principal Securities.access. Most trades have been in lieu of structured products or derivatives. federal pension reform is unlikely to affect the state sector significantly. One of the main reasons the Treasury reintroduced 5y TIPS issuance in 2004 was to encourage central bank buying. As in other countries. While iSTRIPS have not been a huge success. The complete formulas may be found at the following link for CFR 356.000 increments. it was important that the Treasury encouraged stripping activity as a sign of its commitment to the TIPS program when many market observers were questioning the durability of the asset class. but since most private funds do not have explicit cost-of-living adjustments (COLAs) allocations to TIPS would be mainly for diversification purposes. so the relative importance of the private sector defined-benefit sector may increase. Allowing TIPS notes and bonds to be stripped into zero coupon instruments is another step in this development process. Principal component There is only one principal component (corpus) per TIPS issue. State and local government schemes are already the largest pension fund buyers of TIPS and their liabilities mean the potential for increased buying is substantial. to date few TIPS have been bought for definedcontribution pension schemes. Pension reform may encourage more buying of TIPS by private defined-benefit pension schemes. The par amount is the original face value of the bond to be stripped in $1. or iSTRIP. along with the principal payment is made into a real zero-coupon instrument. mainly because inflation-linked life policies are much rarer. The principal component retains one of the key attractions to TIPS. At the start of 2009 we began to see structural investments in TIPS from foreign central banks and sovereign wealth funds because of diversification benefits and as a de-facto currency hedge.

Figure 20: Example of principal inflation strip (SIIP) TIPS 3. The embedded deflation floor in the TIPS security stays with the principal component.04516) * 1. however. So for example. the reference CPI at maturity of the bond were somehow less than the base CPI.04516 Source: Barclays Capital If.000 * 0. The US Treasury. coupons with the same maturity from different TIPS are now fungible and the coupon strip would be inflation adjusted at the same rate. the inflation-adjusted principal will be less than par and the investor will.06 Interest component The US Treasury faced a hurdle in the initial formation of the strips program. on January 23.261.99 = $1. if the January 2011 principal iSTRIP was priced at 99 for that settle date then market value would be calculated as: (216.000. 2011. removing the inflation indexation to allow for stripping and then re-adjusting the zero coupons for their inflation accrual. whichever is greater. a buyer of coupon iSTRIPS can effectively create “P” if necessary. regardless of the underlying security from which the interest payments were stripped. as each TIPS issue having its own base CPI would have a different inflation accrual index. resulting in an index ratio of less than 1.230. sets the stripped interest component and its adjusted payment valuation. Therefore.0. the reference CPI is equal to 219. receive the $1. To make issues fungible with each other.04516) * 1. The Treasury established that the adjusted valuation (AV) calculation would be as follows: 15 March 2010 49 . accordingly. making the coupon component a true real rate security.000 face value. All such components with the same maturity date have the same CUSIP number. an owner of the principal component will receive: (Reference CPI at maturity/base CPI) * par value (219.Barclays Capital | Global Inflation-Linked Products – A User’s Guide applies to the principal component so holders of the principal at maturity receive the inflation-adjusted principal value or the par amount.28558/174. The principal component trades at a discount to par and for trade settlement purposes will settle in the intervening period using the same methodology as above.000.5. the development of the inflation derivative market allows buyers of coupon iSTRIPS to purchase inflation floors.000.5/174. 5% 1/15/11 P = $1. which is established using the CPI reference value for its original issue (dated) date. The adjusted value represents the reset of the inflation accrual to 100. 2010 the reference CPI was 216.93 If.000 = $1. In this way.000. the Treasury had to create a two-step process: first.271. in the Federal Register. on January 15. with an inflation adjustment made to an investor at maturity.166.000 par amount Base CPI on issue date = 174.28558. substituting the current reference CPI into the equation. Hence. The interest component (coupon) from a particular TIPS issue is transferred at an adjusted value initially.

054. 000 * (219.86 Source: Barclays Capital In this example. The Bloomberg ticker for the index is CPURNSA <Index>.000. While zero-coupon-style swaps are the most active structure traded on the inter-broker market. 15 March 2010 50 . assume that in January 2011 the reference CPI (at maturity) is 219. in the same way as the French CPIx market. Using the example. A primary driver of US swap activity thus far has been hedging related to inflation-linked MTN deals. US inflation derivatives US zero-coupon CPI swaps have adopted an interpolated base index format. 10.04516) = $10.000. 2010 and a reference CPI of 216.000.000 bonds are created. with a $1. if we assume the price is 99 the coupon would settle at: Par x (Reference CPI-U/100) x market price or $1.000.000 notional stripped.5% 1/15/11 C = quoted coupon P = $1. a buyer/seller of a coupon would settle a trade as follows: Par x (Reference CPI U/100). y/y structures are most commonly demanded by the US retail sector.04516 base CPI on issue (dated) date AV = adjusted value AV = ((C/2) *P) *(100/CPI)) or ((0.000.195.5/100) = $2.1 of $1.28558.000 par amount CPI = 174.24 At maturity. Prior to maturity. This serves to smooth out the discontinuities in swap breakevens at month-end that occurred when the market first found its feet using the HICPx-style non-interpolated format. the amount payable on a coupon strip is made via the following formula: Figure 22: Amount payable on coupon inflation strip AP = amount payable at maturity RVCPI = reference value for CPI at maturity date AP = AV *(RVCPI/100) Source: Barclays Capital Following on our example for the principal strip. which also features an interpolated daily reference index.141.000 x (216. Bundled with other issues. This more closely aligns the swaps market methodology with the bond market. final payment would be $1.28558/100) x (0. there is sufficient liquidity created to generate round lots of bonds.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 21: Example of coupon inflation strip (SII) adjusted valuation TIPS 3. The index used is the CPI-U not seasonally adjusted index with a three-month lag. the same as that for TIPS.00.035/2) * 1000000) * (100/174. Barclays Capital’s indicative CPI-U zero-coupon swap Bloomberg page is BCAP3.227.99) = $2.5 and thus. assume that the January 2011 note is purchased with a settlement date of January 23.

However. Trading in asset swaps in the US has tended to occur on a tactical basis when valuations appear at the edge of a range. Typically. beginning in 2006. such as caps and floors. TRS provides an alternative to cash as a way to gain long or short exposure to cash instruments. but only in limited fashion thus far. Linking the cash TIPS and CPI swaps market is asset swaps. The total rate of return swaps market has taken off notably. Many investors use inflation-linked TRS to gain beta exposure to the asset class while generating alpha returns in some other product. CPI futures were introduced in the US in February 2004. along with other non-linear inflation products. The gap between cash breakevens and CPI can be explained by funding costs of a cash breakeven position including expected repo differentials and balance sheet costs.Barclays Capital | Global Inflation-Linked Products – A User’s Guide although issuance in this sector has been extremely limited since 2008. because this would be more useful for both risk management and relative value trading if it were introduced. we do see scope for reintroducing an inflation contract based on either monthly NSA CPI settings or monthly contracts on the y/y format. the initial contract specifications did not proved useful and it is no longer traded. TIPS and breakevens. While not likely soon. these corporate deals pay y/y inflation on a monthly basis plus a fixed spread (with a floor usually set at zero on the sum of inflation plus the fixed coupon). The general lack of payers in the inflation swap market thus far in the US means that CPI swaps breakevens and bond breakevens are more loosely connected than say in the euro area. 15 March 2010 51 . Options on CPI Swaps. Paying the inflation uplift out rather than accreting the principal as with TIPS is done primarily to avoid the phantom income tax problem associated with TIPS. similar to the HICPx future in Europe. have traded in the US. particularly by investors looking to match index returns.

Greece. MUICP inflation swaps traded before the launch of the first French euro HICPxlinked bond. electricity. the Eurostat statistical agency was charged with creating “common statistical standards for consumer price indices”. The weights are determined according to each country’s share of consumption expenditure within the euro area as measured by the “household final monetary consumption expenditure” in national accounts data. As most Italian domestic inflation liabilities also exclude tobacco. Activity in individual euro country inflation indices remains limited. The HICPx is computed as a weighted average of the individual euro area countries’ harmonised price indices. the country weights change over time and are actually reviewed each year and applied with the January data (ie. The depth and complexity of the inflation derivatives market has increased substantially. is the main inflation reference for monetary policy for the European Central Bank (ECB). while structured note activity has encouraged the development of a volatility market. with end users active in inflation swaps. also known as the MUICP or Monetary Union Index of Consumer Prices.com France kick-started the European linker market with an issue linked to French CPI extobacco in 1998 and launched its first bonds linked to the euro HICPx in 2001. is currently the main inflation-linking index for euro area government inflation-linked bonds. the January inflation figures published in February). gas and other fuels 16% Furnishings. Italy and Germany have since followed. all items excluding tobacco (referred to as HICPx). household equipment and routine maintenance of the house 7% In 1996.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Euro area Khrishnamoorthy Sooben +44 (0) 20 777 37514 khrishnamoorthy. but after France had issued its first bond. the Italian government used the same index for its first €i bond. Figure 23: Breakdown of euro HICPx by major category in 2009 Miscellaneous Food and nongoods and services alcoholic beverages Restaurants and Alcoholic 9% 17% hotels beverages 10% 1% Education 1% Recreation and culture 10% Communications 3% Transport 15% Source: Eurostat Health 4% Clothing and footwear 7% Housing. HICPx became the most widely used inflation swap base. confirming the benchmark status of euro HICPx for both bonds and swaps. The headline all-items HICP Index. The development of an inflation-linked debt programme in Germany has been slow. which has been defined by the ECB as a level of MUICP inflation close to but below 2%. 15 March 2010 52 . with 75% of such tradable issues by face value tied to it at the end of 2009. with Italy now the largest issuer in euro HICPx. The ECB has a mandate to maintain price stability. water. Countries joining the European Monetary Union are also added to the index. but is likely to gain momentum in the years to come. with exposures mostly hedged via the broader euro HICPx market Euro HICPx Index The Euro Area Harmonised Index of consumer prices. Therefore.sooben @barcap.

The most notable was the inclusion of discounted sales prices in Italy and Spain from 2001. Christmas-related price increases tend to push German prices much higher in December. There is still scope for further standardisation though. The use of hedonic pricing to adjust for changes in quality is likely to become more widespread as a result. Currently. with most such changes likely to further increase the seasonality of the series. This is a poor approximation for January m/m inflation as the seasonal factor for this month is the most extreme negative of the year. attempting to strip out the price of land from house prices) as this is an investment rather than consumption element. More important is the consideration of housing. As a result. which causes a complication for the bond market. actual plus imputed rentals). has conducted various pilot studies on this with mixed results. along with national statistics agencies. The ECB has highlighted the need to include owner-occupied housing.Barclays Capital | Global Inflation-Linked Products – A User’s Guide HICP indices. In the years since the formation of the monetary union. whereas national accounts data suggest that around 16% of all consumption spending is on housing (ie. Final January inflation data are released very late in February as a result of extra calculations needed for annual re-weighting. which over the long term may produce a marginal downwards bias in the index. The German resetting in February 2003 caused significant revisions to the inflation profile. There have been few major revisions to the composition of the euro HICPx in recent years. Following the German rebasing in February 2008. the index is published to two decimal places rather than one. for which the original reference HICP has been rescaled accordingly. leaving only limited uncertainty in the final release. but the impact for the euro area as a whole is muted. The rebasing of the HICPx Index at the end of February 2006 (base year 2005 = 100) did not materially affect valuations on bonds or swaps. have detailed re-weightings only every five years. which slightly reduced average inflation but greatly increased m/m volatility. This introduced a notable structural break in the package holiday subcomponent of the series in 2008. Housing rents make up 6% of the HICPx. are geometric chainweighted Laspeyres indices. in common with most other CPI indices nowadays. When an inflation reference value is unknown for settlement. the largest negative seasonals are in January and July due mainly to sales periods. there have been several changes to measurement by individual countries that have altered the seasonality of the aggregate index. but Eurostat has been focusing on gaining consistency. some euro countries. the official formula to calculate the index ratio is to extrapolate the last known y/y inflation rate to the latest index value. German HICP still does not include many temporarily discounted prices. Eurostat decided to allow changes to impact the HICP series only from January 2008. while a decision on the inclusion of owner-occupied housing was originally intended to be made by 15 March 2010 53 . This suggested that any inclusion would most likely be on a net acquisitions basis (ie. While there is annual chain indexation at the start of each year to reflect changes in consumption weights between and within countries. but seasonally-adjusted series are produced for the ECB monthly bulletin that provides information about the development of seasonal factors. There is relatively little standardisation of quality adjustment measurement at present. eg. Eurostat. but a flash estimate of MUICP inflation is released around the end of the month in which data are collected and individual countries release data in advance of the euro total. with owner-occupied housing currently excluded from HICP indices. Thus. the market rarely trades on the official convention at the end of February. with the most positive seasonality in March. Since February 2006. particularly Germany. preferring to short settle or not trade at all if the index ratio is unknown. for instance in its July 2005 monthly bulletin. All HICP data published by Eurostat is non-seasonally adjusted. Final euro area inflation data is usually released around the 17th of the following month. but all bonds and swap contracts are based on unrevised index values.

The unrevised index is used for bonds and swaps. While covering a notably smaller and more homogeneous area makes the French CPIx Index more volatile than euro HICPx. Although the level of French prices is only marginally above that of the euro area average. there has been a single CPI released mid-month.Barclays Capital | Global Inflation-Linked Products – A User’s Guide 2006. More substantive and difficult legislation would have had to be introduced to change this. this is mostly offset by having a lower weighting for fresh food and energy. France would become relatively more expensive as the euro area expands eastwards. while the index on Bloomberg is FRCPXTOB <Index>. This is particularly important for the healthcare component where rebates from the state are substantial in France. but it also meant that any inflation liabilities similarly had no link to tobacco. The bonds are linked to the nonseasonally adjusted CPIx series. Both use geometric aggregation at the lowest strata sub-indices. If agreement cannot be reached. and have the same methodology for quality adjustments. Since March 2005. French CPIx Index France first started issuing bonds linked to the French CPI excluding tobacco (CPIx). as indexation to tobacco prices had previously been banned in France. no decision has been taken by the start of 2010. There has been no clear. If owner-occupied housing were included in the HICPx. while CPI uses a gross basis. long-term bias between the two series. While in the long term it is likely that healthcare costs will increase notably higher than inflation. but also leads to inconsistencies between the two series when healthcare reform affects the degree of public subsidy. to publish earlier. This leads to a much higher weight for healthcare in the CPI (10%). the national statistics office in France. but the process of euro area standardisation has led to some changes in measurement of French components. inferences on the French data may be drawn if the euro flash estimate surprises in either direction. this may well be offset if the trend for reduced state subsidies continues. The index is usually released around the middle of the month. but Eurostat encouraged INSEE. it is possible that a separate house price measure is produced instead. Until 2005. which has increased seasonality somewhat in recent years. as estimates for most other large euro countries are published around this time. The difference is that HICP takes into account expenses net of rebates. for instance reducing the items on which they are available. The calculation methods for the French CPI and HICP are relatively similar. An unpublished early estimate of French HICP is provided to Eurostat for the euro HICP flash estimate and. which causes a jump in the HICP healthcare series without affecting the CPI. 15 March 2010 54 . just ahead of the euro area data. Daily interpolated index values for bonds are published on the Reuters page OATINFLATION01. If the series is rebased all reference calculations are adjusted accordingly. preliminary data were released early in the month with the final series published after the euro data. over the long run it would be likely to bias the price level upwards but it could also significantly increase medium-term inflation volatility.

Price convergence assumptions among countries in the euro area are an important element in projecting the evolution of a domestic index relative to the euro area. although ultimately. Eurostat Price convergence within the euro area Indices other than euro HICPx or FRCPIx are also traded in swaps although there are no active interbank markets for these. although empirically. where foreign trade is a significant share of GDP. Country-specific factors such as taxes and legislation will also be relevant. Ultimately. this implies that in the long run. At the other extreme. decisions that can be seen as neutral and appropriate for the area as a whole may not be suitable to some specific individual countries.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 24: French and European ex-tobacco inflation 5% 4% 3% 2% 1% 0% -1% -2% Jan 02 Euro HICPx y/y FRCPIx y/y Jan 03 Jan 04 Jan 05 Jan 06 Jan 07 Jan 08 Jan 09 Source: INSEE. However. For example. inflation rates will diverge during the convergence process among countries that initially have significantly different price levels. The evolution of domestic price levels with the euro area is therefore key in determining the fair value of spreads. It should be noted. The fact that the ECB’s policy is implemented at the euro area level means that at any time. as its top trading partners. then prices in Spain will tend to adjust upwards to catch up with the euro area average. however. The extent to which a particular country trades with other members of the euro area will also affect the convergence of prices: for example. but with two countries outside the euro area. by definition. This may accentuate price level differences. open and share the same currency. demand and supply factors may have a significantly bigger weight. price convergence towards the euro area average should be high. and with its economy highly dependent on international trade. France and Belgium. If euro area economies are closely related. then we can assume that domestic price levels will tend to converge. If there are strong convergence forces. price levels will be very similar across countries of the euro area. the UK and the US. there has been a tendency for a converging process. thereby mechanically pushing Spanish inflation relatively higher. Spain currently has a lower price level compared to the euro area average. Luxembourg trades mainly with Germany. The assumption of strong price convergence is however a debatable one. The inverse would apply for a country with a higher-than-average price level. for instance if the central bank rate is deemed too high or low for some countries. there may be little trend towards price convergence for Ireland. 15 March 2010 55 . The pricing rationale for such indices usually involves determining a spread versus the broad euro HICPx market. which acts as the benchmark reference.

Figure 25: Euro price levels converging (EA15=100) 130 120 110 100 90 80 70 60 1997 BE GR DE IE ES IT FR NL 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Source: Eurostat. The timing of the first issue just ahead of the start of the euro area was not a coincidence. Eurostat data up to 2008 show that the price level convergence between the major euro area countries has been relatively strong. determined by the share of its “household final monetary consumption expenditure”. the further its price level is from the current average. The legislation to enable the launch of the new asset class was passed on 3 July 1998. Barclays Capital Development of the government market France France first announced its intention to issue inflation-linked bonds on 3 December 1997. with the bond frequently re-opened by subsequent auctions. additions to the euro HICP have not introduced any strong bias in the index. this will introduce an upward bias on euro HICP inflation. with the euro area expanding towards eastern European countries. including the choice of inflation index to which it would be linked. however. the OATi 3% Jul 2009 was syndicated. convergence may not be as strong given that services are often not easily traded among countries. have an impact on euro HICP. In the past. the higher the weight of a new country in the euro area HICP basket. The inflation index was agreed as INSEE’s official measure of French national CPI. On 15 September 1998. First. However. The market was widely consulted on the main characteristics of the new bond. Second.Barclays Capital | Global Inflation-Linked Products – A User’s Guide that close trade relationships between a country and the rest of the euro area will tend to encourage price convergence in the products that are effectively traded. if the difference in the general price level is due primarily to the difference in price levels in the services sector. It was decided that the bonds would adopt the Canadian methodology that was fast becoming the preferred global structure. but including a principal floor as the US had done. where price levels are lower. excluding tobacco. The enlargement of the euro area will. This monetary union was expected to intensify competition for financing in nominal bonds. There was considerable debate ahead of the initial launch of French 15 March 2010 56 . the higher its potential effect on euro HICP. the higher will its inflation tend to diverge from other member countries at the beginning. Hence. and the hope was that France would gain a first-mover advantage by being the first euro area country to issue inflation-linked bonds.

international appeal would clearly be broader for a euro index. but then launched the OATi11 and OAT€i15 via auction later in the year. the OATi19 was issued via auction. the BTAN€i10. excluding August and December. again linked to the French national CPI ex-tobacco. In 2010. including some exchanges out of the OATi29. The strategy to reduce issuance from the end of 2008 and over 2009 was combined with a pragmatic approach. In October 2001. untested index with no track record. €17bn was issued in 2005. but with its size boosted by some direct exchanges out of the OATi09. the final decision almost certainly came down to practicalities. It is usual for one OATi and one OAT€i to be auctioned each month except when a new bond is launched. The same issuance route was followed. the disadvantages of Eurostat’s European Harmonised Index of Consumer Prices for the EMU area were material as it was a relatively new. The OATi17 was also launched via auction in September 2005. which left an index in flux and an associated fear of revision risk. Growth in the outstanding market value of these two bonds was slow but steady. France responded to an increase in interest and demand in the sector with a significant increase in the pace of supply. but with the possibility to issue significantly more if justified by demand. The OAT€i Jul 2032 was syndicated in 2002. in October 2008. France switched back to syndication. the first BTAN linked to euro HICPx. There were some fears ahead of this issue that the launch of a second inflation-linked product may harm the liquidity of existing OATi bonds. Auctions normally occur on the third Thursday of every month. In January 2010. After consultation with primary dealers. Not only did turnover in the new issue quickly grow. Full index coverage was not yet complete in some countries. In 2006. A second linker. as the AFT responded to the fall in demand for linkers. It has issued new bonds each year while auctioning existing issues nearly every month. compared with €24bn in 2004 when demand was particularly strong. In April 2006. this bond was launched via syndication. 14% of gross bond supply. but interest in the existing issues was heightened too. when fears about deflation were intense and risks of poor action were significant. France addressed this issue head on by launching the OAT€i 3% Jul 2012 linked to euro HICPx. the aim of the AFT is for about 10% of total issuance to be in inflationlinked bonds. For instance. the OATi 3. 15 March 2010 57 . Again. was launched a year later in September 1999. The arguments for the domestic index included the likelihood that national inflation would be a better liability match for the government. it was committed to a minimum of 10% of its total bond issuance each year to be in inflation-linked bonds. There was some disappointment that the instruments did not seem to be capturing the imagination of investors in euro area countries outside of France. which consisted of tapping specific issues that were in demand. but in fact the move gave a new lease of life to the sector as a whole. The AFT has steadily increased linker issuance since the product was launched. For the launch of the OAT€i40 in March 2007. The Agence France Trésor (AFT) decided to revert to a syndication method to launch its new OAT€i 2020 at the start of 2004.Barclays Capital | Global Inflation-Linked Products – A User’s Guide inflation-linked bonds as to whether to link the first issue to French inflation or to that of the then forthcoming euro area. the AFT recommended a T+3 settlement date for the BTAN€i (nominal BTANs settle T+1) to be consistent with other inflation-linked bonds. However. given that the only demand was for French inflation due to Livret A hedging. 2008 saw €15bn but issuance dropped to €12bn in 2009. At the time. Prior to 2009. In 1998. as in 2009. as for the OATi23 in February 2008.4% Jul 2029. The OATi Jul 2013 was the first issue to be launched via auction in 2003. €18bn was issued compared with €17bn in 2007. with an initial syndication and occasional re-openings. was launched via auction. only OATis were issued.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 26: French linkers: historical performance and risk 14% 12% 10% 8% 6% 4% 2% 0% -2% -4% 1999 2001 2003 2005 2007 2009 France IL Returns France IL Ann Mthly Vol Figure 27: Return/risk French IL versus nominals and equities 6 5 4 3 2 1 0 -1 -2 -3 1999 2001 2003 2005 2007 2009 Equity IL Nominal Bonds Source: Barclays Capital Source: Thomson Datastream.3% July 2030 via syndication in March 2007 for €3. Barclays Capital 15 March 2010 58 . A few weeks before launching the GGB€i30. It was reopened via syndication twice in 2004 and once in 2005 and 2006. again via syndication. Before it entered the euro area. but the last of these redeemed in 2007. probably due to their small weight in the market. with inflation accrual linked to HICPx. also in a Canadian format.9% Jul 2025 in March 2003 via syndication. Greece launched the GGB€i 2. The fact that they fell out of the main Barclays Capital inflation-linked indices at the end of December 2009 because of ratings downgrades may affect their structural valuations versus other issues and contribute to keeping poor liquidity conditions. It was tapped for €4bn in January 2008. Figure 28: Greek linkers: historical performance and risk 30% 25% 20% 15% 10% 5% 0% -5% -10% -15% -20% 2004 Source: Barclays Capital Figure 29: Return/risk Greek IL versus nominals and equities 5 4 3 2 1 0 -1 -2 -3 2004 Equity IL Nominal Bonds Greece IL Returns Greece IL Ann Mthly Vol 2005 2006 2007 2008 2009 2005 2006 2007 2008 2009 Source: Thomson Datastream. Greece made a private placement of a 50y bond linked to euro HICPx.15% 2032 on a real yield basis. Greece issued bonds linked to domestic CPI in small size in 1997. The bond was priced against an interpolated OAT€i real yield curve on 18 March. 25 July.5bn. with pricing against the OAT€i 3.25bn. for €1. The liquidity of the Greek issues has generally been lower than other European linkers. Barclays Capital Greece Greece launched the GGB€i 2. and even the same coupon payment date. GGB€i bonds have exactly the same calculation conventions as French OAT€i bonds.

but its 2010 Guidelines for Public Debt Management indicate that new securities will be placed via syndication. Effective from 15 March 2010 59 . The Tesoro had previously indicated it will launch 5y maturity bonds via auction. Both bonds have been re-opened since via auction. but the issue was quickly accepted.4bn. with relatively little going to other euro area countries. Further syndicated supply of the issue in 2004 brought its face value up to €13. The BTP€i 1. After almost one and a half years. Except for the 5y BTP€i12. The BTP€i 2012 was the first Italian linker to be launched via auction in March 2007. the launch of new bonds has been done via syndication. with the scheduled auction of October 2008 even cancelled.1% September 2017 was sold for €4bn via syndication. Italy overtook France as the country with the largest stock of bonds linked to euro HICPx. The majority of this bond quickly became held versus inflation swaps. with almost 40% being allocated to French investors. the day before nominal supply and settling on the last day of the month. its first 15y linker.65% Sep 2008 followed an almost identical model to French OAT€i bonds. the next new bond was a 10y. and tapped via auctions afterwards. The bond was initially priced using an interpolated spread to the nominal BTP curve. The choice of maturity for the first BTP€i was determined by heavy domestic retail demand for inflation-linked notes. with the majority placed in Italy. the 10y BTP€i 2. Italy has indicated its intention to continue with the monthly issuance of BTP€is. In May 2009. as a response to reduced demand.5bn. Italy also reduced its inflation-linked issuance significantly from the third quarter of 2008 and over 2009. except that it paid semiannual coupons like conventional BTP bonds. but a maturity matched conventional bond was auctioned the week after the launch.35% 2035 was syndicated in October 2004 for €4bn and has subsequently been re-opened by syndication and then auction. the BTP€i 2. much of which remains locked away to maturity. The BTP€i10 was syndicated in January 2005 as the second 5y issue and has been built up by auctions subsequently. These. Having started issuance with a relatively opportunistic 5y maturity. More than 220 investors bought the initial syndication. Italy hoped to capture both swaphedging demand and to appeal directly to individuals who had been buying the structured notes. In June 2006. In 2005.15% September 2014. In addition. The speed of the ground-breaking transaction took many in the market by surprise. a combination of derivative houses and long established. Auctions take place near the end of the month. together with Greece’s 50y linker private placement. Much of the remainder went to the UK and US. enabling a syndicated re-opening in October to bring the bond to over €10bn. The bond was initially syndicated in February 2004 for €5bn and subsequently built up via syndications and then auctions to a notional size of €14. but the Tesoro reverted to the syndication method a few months later in June to launch the BTP€i 2023. The first re-opening syndication in November 2003 redressed this imbalance. The BTP€i 2. sparked two-way interest in ultra-long euro HICPx swaps. in line with the target size indicated at launch.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Italy Italy announced its intention to issue its first inflation-linked BTP on 5 September 2003 and syndicated a €7bn 5y bond within five days. international inflation-linked investors. Italy came back to the new issue market in October 2009 with the syndicated launch of the BTP€i 2041. the 10y BTP€i 2019 was launched via syndication but thereafter re-opened via auction. Like France. The Public Debt Division of the Italian Department of the Treasury had noted that a considerable amount of swapped 5y MTN notes with inflation-linked coupons had been sold in Italy and without a 5y point on the OAT€i curve it was relatively difficult for issuers to hedge their inflation exposure. enabling straightforward trading of the breakeven inflation spread. Italy has issued private placements of two ultra long-dated euro HICPx linkers maturing in September 2057 and September 2062.

The 10y Bund€i 1. The bond was re-opened in September 2006 via syndication for €3. The third German linker came in June 2009.5bn. This reaffirmed the commitment of Germany to the inflation market. the OBL€i 2. although domestic Italian indexation has historically usually excluded tobacco. Although the bond was being priced against the 15 March 2010 60 . The German Finanzagetur indicated its intention to tap the bond up to three times to a volume of €10-15bn. The inaugural inflation-linked bond meant that all G7 countries were issuers of inflation-linked bonds. a maximum amount was announced for each individual BTP€i auctioned. As with a conventional BTP. Before the launch of the Bund €i16. given that the development of the German real €i curve has been slow compared to what was broadly expected when the programme was launched in 2006. deflation-led fall in breakevens during the second half of 2008. as it is the index most widely used in inflation swaps and MTN bonds as well as OAT€is. too. including €500mn retained. Barclays Capital Germany The intention to issue euro HICPx-linked bonds was announced by German finance ministry officials in November 2004. Figure 30: Italian linkers: historical performance and risk 14% 12% 10% 8% 6% 4% 2% 0% -2% 2005 Source: Barclays Capital Figure 31: Return/risk Italian IL versus nominals and equities 4 3 2 1 0 -1 -2 -3 2005 Equity IL Nominal Bonds Italy IL Returns Italy IL Ann Mthly Vol 2006 2007 2008 2009 2006 2007 2008 2009 Source: Thomson Datastream. The launch was conducted via auction.5bn retained to boost liquidity in secondary market trading. Previously. with inflation accrual calculated on a daily interpolated basis between the inflation data from three and two months previously. The fact that the Bund €i16 would redeem on 15 April 2016 meant that it would accrue less positive inflation seasonality compared to an equivalent OAT€i or BTP€i.5% April 2016 was issued via syndication with an initial size of €5.5bn. a BTP€i pays its coupon every six months. Calculations work in exactly the same way. Italy chose the same index as France mainly for market convenience. bonds linked to the euro HICPx had maturities of 25 July (French and Greek linkers) and 15 September (Italian linkers). BTP€i auction announcements carry a range for the total amount to be issued across the securites involved. Germany switched to an auction procedure for the second tap of the Bund€i16 in April 2007 for €2bn. including €0. and was priced against the nominal Bund January 2016. but its yield is quoted on an annual basis. was issued in October 2007 for €4bn. The second euro HICPx German linker. The Bund€i 2020 was the first euro linker launched after the episode of extreme. The focus during the launch of the first German linker was on seasonality pricing.25% April 2013. The Finanzagetur announced in its 2010 issuance outlook that it intends to issue €3-4bn inflation-linked bonds quarterly. but it was not until March 2006 that the initial bond was launched.Barclays Capital | Global Inflation-Linked Products – A User’s Guide the start of 2010.

initiated as from 2008. ie. the reference inflation index at maturity divided by the initial reference inflation index value for the underlying bond. The Italian Treasury followed suit in January 2008. they have helped to refine the seasonality pattern that is perceived by the market. 15 March 2010 61 . Coupon payments with the same maturity and linked to the same index are fungible and have a single ISIN. identified by its own ISIN. This means. While the value of the par floor option on its own is usually very small. the Agence France Trésor (AFT) made it possible to strip all French inflationlinked issues. Each linker under the stripping process can be decomposed into a series of coupon payments and a principal repayment. However.Barclays Capital | Global Inflation-Linked Products – A User’s Guide nominal Bund January 2016. making it possible to strip all BTP€is. that a coupon payment from the OAT€i15 to be paid on 25 July 2010 will be fungible with a coupon from the OAT€i20 to be paid on the same date. given that this three-component stripping model does not exist in any other inflation-linked bond market. given that coupons from two different bonds will have two different base reference inflation indices. Conventions for the principal component are exactly the same as for the underlying issue except that there is a floor on the principal. By trading on a different seasonality point. whereby each individual component is traded as a zero-coupon instrument. the German linkers are useful. This introduces a major innovation to the principle of iSTRIPS. for example. If inflation has occurred since the underlying bond was issued. the market was also focused on estimating fair value versus the OAT€i15. The coupon-stripping process for BTP€is. and estimating the difference in the seasonality component of the two bonds was key in determining the fair value of the Bund€i16. Figure 32: Euro linkers – historical performance and risk 14% 12% 10% 8% 6% 4% 2% 0% -2% -4% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: Barclays Capital Figure 33: Return/risk Euro IL versus nominals and equities 6 5 4 3 2 1 0 -1 -2 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: Thomson Datastream. there is an adjustment that transforms the base reference inflation index value for each coupon strip to 100. Barclays Capital Euro IL Returns Euro IL Ann Mthly Vol Equity IL Nominal Bonds Euro inflation-linked STRIPS In June 2007. Investors can actually request that the principal be split into a nominal component and a floored inflation uplift component. For French government issues. Hence an investor receives at least the face value of the position at maturity. the investor receives the face value multiplied by the index ratio. holds an additional complexity compared to French bonds. the guarantee of nominal principal repayment may be worth notably more for investors who are unable to buy bonds without floors. an adjustment is needed. From that point of view. In order to achieve fungibility for coupon iSTRIPS. each linker has one principal component.

The pressure for life insurers and pension funds to address liabilities with the introduction of IAS19 has been similar to that in the UK. regulatory and accounting factors have been less important in the development of the European linker market. de facto. implying less turnover in the liabilities and hence more incentive to hedge. for example. most are linked to indices very different from euro HICPx. the take-up of options that have an inflation linkage has been limited. was meant to depoliticise the ratesetting decision. The evolution of the framework allowing investors to hold inflation-linked bonds has been determinant.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation strips provide increased flexibility to hedge inflation-linked liabilities. insurance companies could not report principal accrual as earnings as it was unrealised. and was used in the twice-yearly revision of the rate. Demand may come from insurance and pension sectors in countries with inflation-linked liabilities but with difficultly accessing inflation swaps. So far. or new TFR money will be transferred by default to INPS where it will accrue at the normal TFR rate. Firms have been required to provide collective pensions options to employees. Pension reform in Europe may crystallise more demand to hedge long-dated. For accounting purposes. Since 1 July 2007. The revision of this rule in 2002 opened the linker market to a deep investor base. The amount retained is 6. The development of the FRCPIx inflation-linked bond and swap markets has been largely driven by the decision to partially link the remuneration rate on Livret A savings accounts to the FRCPIx inflation rate. inflation-linked strips can be more favourable than inflation swaps as they are considered unleveraged inflation-linked bonds on a held-to-maturity basis. although cash collected prior to 2006 remains with companies. The Livret A (called Livret Beu when distributed by the Credit Mutuel network) rate is used to determine the remuneration rate on various other savings accounts and in 2004. German investors were restricted from holding inflation-linked bonds prior to 2004. Funds collected on Livret A and Livret Bleu savings 15 March 2010 62 . taken in 2004. While previously. around €270bn of instant access account deposits became. accounting and taxation Compared to the UK. with a tendency for funds to be transferred to INPS. with no embedded derivatives under an IAS accounting regime. inflationlinked liabilities. For example. now they will be released only at retirement. euro countries where pension funds do not use many swaps in general and inflation swaps in particular also tend to have some of the longest inflation liabilities.91% of gross earnings and under the TFR the funds collected accrete in value annually at 1. Ultra-long iSTRIPS could therefore be of interest for pension funds in such countries. but far fewer liabilities are explicitly inflation-linked in the euro area and of these. Regulation. INPS. TFR funds were released to employees when they changed employers. The decision. linked to French inflation. which refers to money held back by firms from Italian employees’ salaries and paid back to them as severance pay when they retire. but activity in stripped euro linkers has never taken off. The original formula determined the rate as half the y/y FRCPIx rate plus half the 3mth Euribor rate plus 25bp. Spain. Co-incidentally. have long domestic inflation exposure within pension funds. all companies with 50 or more employees are required to transfer further funds collected into pension funds or to the control of the national government pensions system. with the French insurance sector now structurally a major player in the market. but the size of such exposures is likely to grow over time. prior to 2002 in France. for example. rounded to the nearest 25bp.5% plus 75% of the change in the level of the Italian FOI (workers) CPI inflation index ex tobacco. An example is the ‘TFR’ (trattamento di fine rapporto) in Italy. and one factor that delayed Germany from issuing in 2005 was ensuring that accounting and regulatory restrictions from investing in government issues were removed or addressed. Finland and Ireland.

Other than the Greek €i for domestic investors. However. although under some conditions. Euro non-government inflation-linked bonds There have been two distinct strands of non-government bonds that have been issued in the euro market with cash flows linked to inflation. the inflation accrual is taxable for domestics while there is no withholding tax payable for international investors. an agency administered by the state-owned financial institution. the change in the rate-setting formula and especially non-consistency in setting the rate according to the formula has weakened the formulaic link between French inflation and the Livret A. This is one of the main reasons for the existence of the structured inflation market. the official rate-setting was further incremented with the condition that the magnitude of the change in the Livret A rate cannot be more than 1. CDC. the distribution rights for the Livret A were extended to the whole banking network in France (as opposed to only La Banque Postale. and they send in the necessary initial documents that are on the Treasury website. commercial banks would be allowed to keep Livret A funds on their balance sheets. From 1 January 2009. The expected and realised increase in Livret A outstandings therefore implied a substantial increase in hedging demand and the impact on the French inflation market was obvious. Institutional investors pay tax both on interest received and annually on inflation as it accrues. From an ALM perspective. At the time of writing.25% and 0. which finances social housing schemes at lending rates that depend on the Livret A rate. as well as some offshore tax havens. the rate given by the formula has been overruled most of the time since the beginning of 2008 to the start of 2010. In our view. the basis of hedges and justification of those hedges from an accounting perspective. The reform of the Livret A undertaken in 2008 and effective from the start of 2009 had important implications with regards to the latter.5x the sum of CPIx and the average of 3mth Euribor and eonia. BTP€i bonds follow the same tax rules as conventional BTPs. There 15 March 2010 63 . which means that the once strong formulaic link. the Crédit Mutuel and the Caisse d’Epargne in the past).5%. an element of non-linearity has been added to the formula. Retail investors can pay all withholding tax at maturity or sale. with it now being the minimum of French CPIx y/y +0. This means that domestic entities are taxed on inflation uplift as well as real returns. ie. Livret A-related hedging demand has structurally decreased as a result of inconsistency in the rate-setting decisions. At the start of 2009. countries excluded from the “white list” of taxexempt countries include Switzerland. The aim was to reduce the volatility in the rate from Euribor fixings in the context of the financial turmoil of 2007. There has been issuance of bonds in a similar accreting Canadian-style format in line with that issued by governments. hedging of the exposure to French CPIx inflation is needed from commercial banks which distribute savings accounts linked to the Livret A rate. either through OATi bonds or FRCPIx swaps. Probably more relevant is the fact that the centralisation rules with the DFE changed and under the new regulation. OATis and OAT€i bonds are taxed in very much the same way as other French government bonds. International investors are exempted from paying withholding tax as long as they are within countries that Italy does not define as tax havens. a non-linear condition that makes hedging even less appealing than before. From 2008. Nevertheless.Barclays Capital | Global Inflation-Linked Products – A User’s Guide accounts have traditionally been centralised at the DFE (Direction des Fonds d’Epargne). This led to a surge in total Livret A and Bleu outstandings by more than €17bn in January 2009. there is therefore an automatic hedge for funds centralised at the DFE and used to finance social housing. However. the uplift of euro inflation-linked bonds is generally taxable. was broken. on which taxation is paid only on coupons as they are paid.

issued a €600mn 2016 French CPIxlinked bond in 2001. As of mid January 2010. France Telecom issued a 10y bond linked to euro HICPx for an initial €350mn amount. Given its revenues derived from regulated activities.5bn in CADESi issues in 2010. CADES stated that it aims to issue €0. which carries the same maturity date. issued a €750mn 2019 bond in February 2004.p. the owner of French railway infrastructure. while there was less than €3bn issued in 2005. and the inflation exposure of these notes has largely been hedged using inflation swaps. the popularity of this kind of product faded. Infrastrutture (ISPA). There was over €18bn of inflation structured note issuance in 2003. which grants loans to toll road companies. This sovereign agency was created in 1996 as a vehicle to consolidate and service the debts of the French social security funds. Veolia ought to be well suited to using the inflation-linked market as part of its funding strategy. Most structured note issuance has been to individual investors. equity-linked note issued by the Italian Post Office. particularly those with explicit inflation linkages for their tolls. In August 2008. having initially been syndicated for €800mn in February 2003. particularly in Italy. Inflation structured notes While there was issuance in inflation structured notes even before the first OAT€i was issued in 2001.A. Toll road operators may well become significant issuers/payers of inflation across Europe as the market develops. helping to define the curve when there were relatively few government issues. Other notable issuance has come from Réseau Ferré de France (RFF).Barclays Capital | Global Inflation-Linked Products – A User’s Guide has also been considerable issuance of structured notes whose cash flows are linked to inflation. it had €10bn face value of bonds linked to French CPIx. Its revenue comes from a ring-fenced. although there has been much more structured supply linked to this base. Terna is a natural payer of inflation. became the first Italian listed company to issue a bond linked to the Italian FOI Index ex tobacco in October 2007. making it a natural issuer/payer of French inflation. particularly as falling real yields made it increasingly difficult to structure sufficiently appealing cash flows. The €500mn issue matures in September 2023 and was priced against the BTP€i September 2023. with almost all of it sold into Italy. Total 2004 issuance was around €10bn. As with other structured products sold mainly to individuals. a euro HICPx-linked €600mn 2015 bond. Government-style bonds Agencies. quasi-agencies and regions have mostly issued inflation-linked bonds in a government-style format. it was in 2003 that the market really took off. with four CADESi benchmark issues with maturities ranging from July 2011 to 2019. The accreting style issuance has been directed at similar institutional investors to those buying government linkers and has generally not involved accompanying derivative transactions. Issuance picked up slightly in 2006 with around €7bn in issuance but 15 March 2010 64 . The Italian agency. which has inflation-linked revenues from some projects such as high-speed railways which it funds. These issues have been built up via multiple syndications and in practice have traded very much like OATis. This issue was the first benchmark inflation-accreting bond to be linked to Italian inflation. The largest non-government issuer has been the Caisse d’Amortissement de la Dette Sociale (CADES). fixed tax on income called CRDS. The first major true corporate inflation-accreting bond came from Veolia Environnement in June 2005. Caisse Nationale des Autoroutes (CNA). whose 2023 HICPx-linked bond reached €2bn face value. Terna S. notably an inflation-protected. specifically workers inflation (FOI) excluding tobacco. €12bn by market value. As an owner of a range of utilities in France and across Europe. albeit with lower liquidity.

but supply dried up in 2007. As distribution of this type of bond became more widespread. but the market is still far from having reached a self-sustained dynamism. there was increased interest from corporates as well as individuals. providing access to investors who are unable to take direct advantage of the development of the inflation swap market. eg. There are currently two contracts linked to the Eurozone Harmonised Index of Consumer Prices excluding tobacco (HICPx): The Chicago 15 March 2010 65 . Coupons were usually floored at the fixed rate. The development of these products marked a stepping stone as they helped increase liquidity in the market for inflation caps and floors. although in 2004 higher floors became more common. The widely dispersed nature of this issuance means that while there are several issues of over €500mn. Eurozone HICPx futures Eurozone HICPx futures are useful instruments to hedge short-term index fixings but activity dried almost completely over 2009. which is likely to be less sensitive to rate levels. holders of the notes suffered poor performance and demand for them waned. 2007 saw the appearance of inflation range accruals. with activity driven by demand from retail investors given the high inflation environment. before picking up strongly during the first half of 2008. Indeed. For example. There has been a clear trend towards more institutionally-focused structured issuance. notes paying a coupon linked to a multiple of the y/y inflation rate were widespread. on the back of evolutions in pension industries. With the slight revival of issuance in 2006. 2006 also saw several structures with coupons linked to the differential between euro and French ex-tobacco inflation. paying inflation from the previous year and high fixed coupons early in the life of the bond were commonly offered as enticements. products with coupon payments linked to inflation (mostly euro HICPx) with leverage became more popular. The area in which there is still ample room for further development is that of domestic inflation bases. The structuring of attractive payoffs was also relatively easy as result of expensive credit/funding for financials. Most issuance in 2003 was of bonds paying an annual coupon at the rate of inflation plus a fixed percentage but with a fixed principal. but with realised inflation breaking above the top of the ranges specified by the end of 2007. then this could force the development of two-way markets. If structural demand develops for individual domestic inflation. In 2003. liquidity is very limited. in 2006. but as yields fell and the type of demand changed. forward French inflation breakevens were higher than on the euro HICPx curve because of Livret A-related demand. Pay-offs here are dependent upon the length of time that inflation remains within a specified range. eg.5bn in 2007. Coupons were often backwardlooking. These notes proved popular as French inflation was expected to be lower than European inflation over the medium term. but fees were such that most institutional investors were deterred. most issuance was 5y. there was significant issuance in structures linked to Spanish inflation. A rise in real yields would increase issuance to individuals although the hefty levels of 2003-04 or 2008 will likely be seen only if there is once again an inflation scare. issuance has moved longer. In 2008. These notes often carried a fixed attractive coupon at the beginning and a cap and/or floor on the subsequent floating payments. when most of the structured notes issued in Spain were linked to the domestic index. The fees involved in this kind note are often lower due to smaller distribution fees. with 10-15y supply as common as 5y in recent years.Barclays Capital | Global Inflation-Linked Products – A User’s Guide volumes fell back to around €3. although opportunistic swapped issuers including the EIB and KfW have also been involved. Most structured inflation notes have been issued by financials. Issuance has been relatively significant from time to time on some bases. The issuance of such products thrived on what was perceived as an anomaly on the forward breakeven differentials in the swap market.

while the Jan ‘07 future is AAF7 <index>.45pm and 5pm central European time. where the code for the whole strip is HICA <index> CT.0 suggests deflation during the 12-month period. Prices are quoted as 100 minus the y/y change in euro HICPx during the previous month. which will almost always be released during the contract month.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Mercantile Exchange (CME) contract was first listed on 19 September 2005. but previously. Thus. rounded to four decimal places. Prices for the contract are available directly from the CME and can also be seen on Bloomberg. Source: CME The format of the Eurex HICPx future contract is almost identical to the CME.) 4. the front CME HICP future was a good predictor of inflation compared with most economists. Thus. thereby highlighting its use to ascertain fair economic value. The Reuters code is <0#EHI:>. there has been almost no activity since mid 2008. Prices can be monitored on Bloomberg. But the main innovation lies in the twice-daily auction mechanism. The price at which orders are matched is determined at the end of the twice-daily auction through a netting process. Every day between 9. Fixing risks are indeed very high on short-dated inflation swaps. By cash settlement on the day the HICP announcement is made. thereby hampering activity and resulting in wide bid/offer spreads. Available for trading on CME “Globex” weekdays from 8:00am to 4:00pm (London time).000. or 100 – [100*(HICPt/HICPt-12)-1]. 15 March 2010 66 . the designated market makers were expected to quote at least 12 of the 20 listed contract months. The contract trades until 4pm London time on the day before the inflation data is released. Each monthly contract refers to the inflation release from the previous month. The CME HICPx future contract is tradable on the CME electronic Globex platform between 8:00am and 4:00pm London time. the code is AAA <index> CT.000.00pm (London time) on the business day preceding the scheduled day the HICP announcement is made in the contract month.000 times Reference HICP (ex-Tobacco) Futures Index 100. The Reuters code is <0#HICP:> On both the CME and Eurex contracts. in an aim to to encourage liquidity. The Eurex future had the potential to be more successful: the fact that a strip of 20 monthly contracts is quoted should have encouraged its use by dealers as it enables the hedging of fixing risks even beyond the one-year point. to see the price of the whole futures strip. Figure 34: Contract specification for CME HICPx futures Contract size Reference HICP Futures Index Contract months Trading venue and hours Minimum price fluctuation Last trading day Final settlement price Contract valued at €10. while the Eurex HICPx future started trading on 21 January 2008. Note that a price of over 100. the value of the January 2008 contract expires based on the y/y inflation rate of December 2007 HICPx. (Leaving the notional value of the contract approximately €1. The price of each contract at expiry is 100 minus the y/y change in HICPx. 0.00 – annual inflation rate in the 12-month period preceding the contract month based on the Eurozone Harmonised Index of Consumer Prices excluding tobacco (“HICP”) published by Eurostat. where the contract code is AAxx <index>. The final settlement price shall be calculated as 100 less the annual % change in HICP over past 12 months.01 Index points or €100.45am and 10am and between 4. The first difference is that Eurex quotes a strip of 20 monthly contracts compared with only 12 for the CME alternative. The netting is designed to determine the price that results in the highest executable volume. Twelve consecutive calendar months. on Bloomberg.

The relatively high level of activity in asset swaps has generally helped align swap and bond breakevens more closely than has been seen in other markets. meaning that the base inflation index for the swap is the value of the HICPx three months before settlement. thereafter. but the investor base has probably shrunk relative to before the crisis The benchmark format for euro HICPx quotes is the zero-coupon structure. Prior to the financial turmoil starting in 2007. Activity in inflation caps and floors on euro HICPx has been boosted in 2007 following the issuance of inflation-range accrual notes. reflecting typical seasonality between months. but the inflation futures offer transparency in the 1y swap point. it accelerated further. The link between the vol and the swaps market became tight towards the end of 2008 when increasing deflation fears pushed short-dated swaps lower. There will be a discontinuity in the quoted “breakeven” at the time of the roll from one month to the next. The presence of embedded floors in previously issued notes and increased volatility in realised inflation led to a sharp repricing of the cap/floor vol surface in 2008-09 (we cover this in more detail in the ‘Inflation volatility and deflation floors’ section later in this publication). which coincided with a period of low liability-hedging demand as Dutch pension fund solvency ratios dropped. leading to a downward spiral of valuations at the short end of the curve as dealers sought to delta15 March 2010 67 . from proprietary desks and hedge funds for instance. direct interest in trading inflation swaps. However. Towards the end of the month. Typical monthly broker volumes moved from €500mn in mid-2002. Barclays Capital displays live prices for the currently trading base month on Bloomberg page BISW1. During the financial turmoil from 2007. Total volume traded surged in 2008 in the interbank market but mainly because of the Lehman collapse. which led to a significant rise in implied volatility. mainly from bank treasuries. most speculative investors significantly reduced their activity in the inflation swaps market. The most commonly-traded maturities are the 5y and 10y. Hefty issuance in leveraged notes over 2008 also encouraged hedging activity in the cap/floor market. but the depth of demand at such extreme liabilities remains relatively limited. Government linker asset swaps trading also represents a significant proportion of volumes in the euro area inflation swap markets. The increased trading activity on inflation options has thus provided a clearer picture of the implied inflation volatility.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation swaps and other derivatives Euro HICPx Euro HICPx swap broker volumes began to pick up substantially from late-2002. but annual maturities above 2y trade regularly out to 30y and sometimes longer. the cheapness of linker asset swaps attracted demand. meant volumes did not suffer even when issuance hedging flow slowed down. a level that was maintained in 2005 even though structured issuance fell notably. the most suitable hedge for most of these exposures. aided by the issuance of BTP€i08. In 2003. as well as defining a tradable short-end curve. Given that these notes rely on inflation remaining within a specified range. Activity at the front end is more limited. Liquidity in the market has been increasing since. while liability hedging rose steadily. driven by a rise in the need to hedge retail products and structured MTNs. The private placements of ultra long-dated linkers by Italy and Greece in 2007 has raised the prospect of the euro inflation curve stretching out to the 50y maturity as in the UK. swaps on the next base month also begin to trade. Ultimately. with typical monthly broker volumes rising to €15bn in 2007. the fact that many speculative players or natural investors in linker asset swaps were no longer active led to a wide distortion between swap and bond valuations in 2009. to €5bn by mid-2004. with a standard lag of three months. this means that dealers selling them were actually long inflation volatility and they have tried to offset their risks by selling inflation caps and floors in the market.

while liability hedging is significantly broader than just Livret A-related flows. However. it discourages short-term tactical trading within the same monthly inflation base that is seen in HICPx and makes it more difficult to compare market movements within a given month. it was not until trading of the basis against euro HICPx began. outright paying flows have often been more common than in euro HICPx. These notes were actually hedged in the corresponding swap markets. In aggregate. Some domestic real money investors started using swaps to match cash flows as the market developed. with relatively little apparent sensitivity to breakeven levels. including zero coupon across the curve. tailored maturity inflation swaps matching the inflation element of reset on the Livret A rate and Livret A swaps covering the nominal as well as the inflation element of the reset. the flow became more and more skewed towards using derivatives as a solution. This avoids the discontinuity of the HICPx method and aligns swap methodology more closely with bond methodology. rather than the monthly format used in other euro area markets. and there is much less structured inflation issuance to be hedged. There is a significant percentage of OATis that are held in asset swap form until maturity as a result of the Livret A hedging pressure. However. just ahead of the launch of the first OATi and before the HICP market that began almost as soon as the euro currency was created in 1999. With banks being restricted on the amount of bonds they could hold versus their liabilities. Short-dated issuance by French agency CADES and the earlier development of the OATi bond market allowed the French CPI swaps to initially have better liquidity than HICPx. On the other hand. at least from a pure formulaic perspective. as it can impose a drift on the level of breakevens. Bank hedging demand has at times led to more significant deviations between OATi asset swaps and those of nominal OATs than in the €i market. there has been a tendency for hedging activity to fall when real yields fall and also to move longer on the curve.Barclays Capital | Global Inflation-Linked Products – A User’s Guide hedge their short floor positions. The decision to link the Livret A French public sector savings rate to inflation from August 2004 greatly heightened activity levels in French CPIx. 15 March 2010 68 . with broker screens becoming readily visible. but there is also active two-way asset swap flow. that the market gained any depth. French CPIx The French CPIx market is slightly less transparent than that in euro HICPx. with no broker market. first trading in 1998. as it is based on the same interpolated daily reference value as used for OATis. This should help realign structural valuations between the FRCPIx and euro HICPx swap markets to more fair levels. the richness of French CPI swaps triggered substantial issuance in structured notes paying euro versus French inflation. However. correcting the squeeze created by Livret A-related hedging. given the extension of Livret A distribution rights to the whole banking network in France. Activity in breakeven and real yield swaptions is muted. The default format for French CPIx zero-coupon swaps is also different. In 2006. compared with the size of the underlying bond market. usually with a leverage factor. A variety of swap types have been used to hedge liabilities. but there was an increase in trades during 2006 and 2007 when the underlying swap markets became more liquid. The French CPIx market is the oldest major euro area inflation market. the fact that the formula was changed in 2008 and then over-ruled most of the time in 2008 and 2009 implies that French inflation is no longer a good hedge for Livret A-related liabilities. FRCPIx again richened significanty versus euro HICPx swaps toward the end of 2008/beginning of 2009.

Italian FOIx CPI The traditionally most active inflation swaps market that does not have an underlying government bond market to hedge against is the market for Italian FOI inflation (ex tobacco). even in the case where the relevant index is the German CPI. with live prices available from some dealers. activity on other euro area domestic inflation swap markets remains quite muted. Pension liabilities in Germany are traditionally backed by company assets. but the German government’s intention to issue inflation-linked bonds. but with floor quotes outside of a deal very rare. the extent to which this will effectively create substantial demand for German CPI is debateable.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Other euro area inflation markets While an underlying government bond market and developments on the regulatory front (Livret A) have helped create liquidity in the FRCPIx swap market. where the basis risk of the relevant index versus German CPI is high. Italian and Spanish inflation swap markets are the most active of these other countries. In the latter case. The market for Italianlinked retail notes started in 2002. Bloomberg code ITCPIUNR. but in 2009 activity in even those domestic indices fell notably. has become the predominantly used measure because it is the basis of most Italian inflation-linked liabilities. there were sufficient dealers pushing for hedging business that embedded floors were priced cheaply. the higher real yield that a German inflation basis offers compared with one linked to euro inflation may be of interest to retail and corporate investors. the tendency towards utility privatisation may well produce paying flows. including lagged basis month. Almost all retail issuance pays y/y inflation and much of it is floored. at which time there were small issues in both FOIx and the alternative NIC basis (Bloomberg code ITCPNIC). Other inflation bases are usually quoted as a spread against the more liquid euro HICPx basis. given that the basis risk versus European inflation is perceived to be low. even though this was euro HICPx. This was initially driven mainly by substantial retail note issuance in Italy on this basis. While there are very few explicit inflation liabilities in Germany. which technically is the price index for workers as opposed to the universal NIC measure. hedging needs may as well be met using the euro HICPx swaps market. liabilities are not solely to the national headline CPI but can be based on various other indices. However. we may expect activity on these swap markets to gain momentum if there is growing demand to hedge explicit liabilities linked to domestic inflation. However. supply in German inflation has come primarily from property securitisation and rental leases. The FOIx market is quoted with the same conventions as euro HICPx. First. while. Until now. as the market develops. usually at 0% inflation. Banks are willing to pay German inflation through swaps because of its weight in the euro HICPx. The FOIx measure. such that the basis risk involved is less of an issue than for other domestic inflation indices. there was some demand 15 March 2010 69 . Live prices are available from Barclays on BISW4. a trend towards funded schemes in Germany implies that embedded domestic inflation risks in liabilities are likely to be addressed on a larger scale. German CPI Any indexation to German CPI was illegal prior to 2003. As in euros. Furthermore. pension funds may prefer to hedge euro HICPx swaps. On the other side of the market. While Italian retail notes are predominantly sold to individuals. banks are likely to be happy to offer German inflation versus euro HICPx across the curve. which means that implicit or even explicit inflation risks tend to be ignored. and the market slowed as from 2005 with less structured issuance. spurred a market to develop. From a general perspective. with the pricing dependent on the expected evolution of each index relative to the euro area.

Demand for Spanish CPI is mainly from the insurance and retails sectors. but tightened significantly towards the end of 2009. was expected to prompt an increase in demand for Italian inflation.Barclays Capital | Global Inflation-Linked Products – A User’s Guide from corporates to hedge TFR exposure. rose strongly in 2006. effective since July 2007. probably due to some restructuring of some positions linked to Italian inflation. which is 1. and traded below in 2009 as inflation fell. The TFR reform. This led to a sharp relative richening in Spanish CPI swaps. most pension funds have accepted that they cannot match their liabilities. 15 March 2010 70 . inducing payers mainly from the infrastructure and property firms. However. and especially in long end. given the relative outlook in inflation. but the effect on FOIx swaps could be offset if TFR-related demand picks up. Inflation in Spain has been notably lower than the euro average over the 2008-09 crisis. and structured notes linked to Spanish inflation dried out in 2009. helped in part by speculative selling from some who viewed the spreads as unsustainable and a narrowing of realised inflation spreads. from the very rich levels observed in 2006. Insurance companies managing these pension liabilities should naturally be driven to receive Italian inflation. while natural supply comes from residential property leases and property securitisation. The volatility of Spanish CPI is notably greater than euro HICP given a heavy weight for the food and fuel components. The size of Dutch defined benefit pension schemes means that demand for domestic inflation ought to be considerable. in the absence of large natural inflation payers. This is a market of one-way flow. no marked pick-up in demand for the FOI ex-tobacco inflation has been observed. however. while the long end lagged versus euro probably due to infrastructure-related paying flows. Indexation is usually to either sector-specific or general wage inflation rather than Dutch CPI and. Economic slack in Italy and loss of competitiveness implies that the differential between the Italian and euro inflation is likely to shrink. with all inflation receiving by pension funds locked to maturity. including options offering a nominal guarantee but with a return target consistent with the TFR. thereby accentuating the richening versus euro HICPx swaps. The first half of 2008 saw heavy activity in Spanish CPI swaps on the back of structured note issuance. Structured note issuance. The interesting feature of structured note activity in Spain. Spreads have been volatile versus euro HICPx swaps. Dutch CPI The largest potential for domestic inflation protection demand. as opposed to other euro area countries. Spread levels to euro HICP swaps decreased in 2007. The rise in demand coincided with a period when actual inflation was high on the back of the real estate boom. Some funds are willing to pay a significant premium to source specific inflation though. These notes typically had 10y and 15y maturities. and the hedging of these triggered a strong demand and richening in Spanish CPI swaps as the street was left short. is undoubtedly in the Netherlands. There has been a very strong convergence in the Italian price level towards the euro average.5% plus 75% of FOI ex tobacco inflation. The pension liabilities linked to inflation should also grow more steadily than in the past given that money held back under the TFR will now be available to employees only at retirement. Spanish CPI The market for Spanish CPI developed in 2004 with some limited swapped structured note issuance. Companies with 50 employees or more are now required to provide collective pension options. but the spread collapsed in 2008. FOIx swaps have generally traded at a positive spread versus euro. The general price level in Spain is lower than the euro area and there had been convergence until 2008. with a tendency for funds to be transferred to the national government pensions system. during 2008 was that the underlying was commonly the domestic index.

but also Belgian CPI and HICP.Barclays Capital | Global Inflation-Linked Products – A User’s Guide At the time of writing. but with a very similar trend growth. so too is the basis risk between wages and Dutch CPI. in contrast to the IAS accounting framework. As with other smaller markets. but it may encourage innovation in other markets. Belgium has the potential to develop into a relatively significant inflation market. set up by the Dutch Ministry of Social Affairs and Employment to assess the pension fund industry. Pension fund inflation liabilities are more explicit than in most other euro countries. tobacco and fuel oils (Bloomberg code BECPHLTH). at the start of 2010. These characteristics may make it particularly ripe for development. Additionally. the Frijns Commission’s report actually highlighted the market for Dutch inflation as a missing one. Given the demographic structure in Ireland. Irish CPI Ireland has a longer history of inflation paying than anywhere else in the euro area. Supply from PPP was much less than expected in 2007. The Commission argued that pension funds should have a “real” commitment to their members. infrastructure and property developments may bring the most likely payers of Belgian inflation. floored at 0% each year and capped at 4%. Such issuance would significantly increase the scope for the development of a domestic inflation swaps market. implicitly implying that indexation should be treated more like a commitment rather than addressed on an adhoc basis when funding ratios are high enough. Also. However. the Frijns Commission. while the short-term basis risk between Dutch CPI and euro HICPx is significant. The FTK regulatory framework furthermore has made it easy to receive euro area inflation versus liabilities linked to domestic inflation. pension funds have been used to taking the basis risk created by long exposures to euro inflation. inflation indexation is conditional in the Netherlands. Further inflation paying is likely to come from infrastructure upgrading projects. particularly given the trend for pension funds of multi-national companies to base themselves in Ireland. pension fund demand for inflation tends to be in long maturities. This. We highlight though that a “real” framework would likely encourage the Netherlands to issue in the inflation-linked market. Until now. the absolute price level in the Netherlands has remained close to the euro area since the start of monetary union. which excludes alcohol. which makes pension funds in other countries reluctant to hedge domestic risk with euro HICPx. This is the closest to a core inflation measure traded in Europe and. coupled with the longer history of the market. suggesting that euro HICPx long term is a relatively effective hedge of Dutch inflationary risk. A move to a “real” framework would boost hedging demand but the potential for this to create a large domestic inflation market is uncertain. often results in relatively complex structures. notably. with a legal minimum of an LPI format. and this coincided with a wide divergence 15 March 2010 71 . The size of the market is such that it is unlikely to affect valuations elsewhere. This could be in Dutch inflation. Given the liquidity of the euro HICPx market and the absence of any other suitable alternative. Belgian Health Index CPI In our view. local government. conditional in the sense that pension funds usually grant indexation only when their nominal solvency ratio is high enough. recommended a move to a solvency ratio assessed in “real” terms. Assessing solvency in real terms should therefore automatically drive pension funds to take indexation more into account and increase inflation demand. which are usually swapped. Supply comes mainly by housing associations and Public Private Partnership (PPP) deals. supply has come mainly from rental leases paying Belgian Health inflation. The relevant index against which almost all inflation liabilities are based is the Health Index measure. has a seasonality that has actually offset euro HICPx in recent years.

Austria has seen little activity due to limited explicit liabilities linked to Austrian CPI. where price stability has been the norm. especially if the quantity of new supply continues to disappoint versus expectations. pension liabilities are mostly linked to the CPI and. this could spur demand for Greek inflation. 15 March 2010 72 . although price convergence is liable to lead to higher inflation than the euro average in all three countries over the coming decade given the much lower level of prices to start with. the lack of Finnish CPI payers restricts the development of a market. Recent entrants from east European have relatively undeveloped private pension systems. Consequently. The mortgage component is a function of both interest rates and long-term house price growth. Portugal has seen a higher degree of potential payers than elsewhere. Despite recent sharp price falls. Portugal and Greece are countries with fairly high interest in domestic inflation. so we would be surprised if domestic inflation markets emerged. even though there is interest to receive domestic inflation from pension funds. if proposals to unify Greece’s pension funds are adopted. which has been substantial since the mid-1990s. In Greece. in Finland. from this point of view. Irish inflation is often perceived as being expensive due to a lack of supply. and by the end of 2009 the y/y differential was 6. Other euro area countries As two regions that have experienced high inflation within recent memory. even though the government has previously issued swapped structured HICPx notes. the fact that Ireland’s major trade partners are countries outside the euro area means that price level convergence may not be very strong. producing a relatively balanced market. suggesting that Irish inflation should continue to fall relative to the euro HICPx in the coming years. By contrast.Barclays Capital | Global Inflation-Linked Products – A User’s Guide between Irish and euro inflation on the back of mortgage interest costs. The ECB rate-cutting cycle therefore pushed Irish CPI inflation much more negative than European inflation. Contracts more commonly contain inflation linkages than in northern European countries.1%. the price level in Ireland is still significantly above the euro area’s average. Irish CPI inflation may continue to trade at rich levels versus euro HICPx inflation. while pension funds are mostly choosing the more liquid euro HICPx options rather than addressing their liabilities directly. However.

or RPI. This was the basis on which the Bank of England’s inflation target was adjusted from 2.com The UK Treasury has been issuing gilts whose principal value is index-linked to the Retail Prices Index (RPI) since 1981. has been worth about 0. the inflation measure targeted by monetary policy was the RPI excluding mortgage interest payments. often referred to as the formula effect.6% Fuel and light 4.1% Personal goods and services 4.9% Household goods and services 13. the Bank of England’s Monetary Policy Committee has had an inflation target based on CPI. Previously.bettiss@barcap.8% 5% 4% 3% 2% 1% 0% Housing 23. but the name was changed from HICP to CPI in June 2003. RPIX.1% Source: National Statistics Figure 36: UK RPI.Barclays Capital | Global Inflation-Linked Products – A User’s Guide UK Chris Bettiss +44 (0) 20 7773 0836 chris. with re-weighting calculated for the January data. pension funds still own the largest proportion of the gilt linker market. The Retail Prices Index The Retail Prices Index. it produces an upward bias compared with a geometric aggregation. Pension fund liabilities are linked to RPI either directly or via Limited Price Indexation (LPI). when it was announced that the inflation target would be changed. Since December 2003. Weights are recalculated annually. has been used as a measure of UK inflation since 1947 and was the main measure of headline inflation for over 50 years.5% in recent years (ranging from 0. UK CPI is a harmonised index of consumer prices. In contrast to most consumer price indices collected internationally – including the UK CPI – the RPI is constructed with arithmetic rather than geometric aggregation. hence. though life insurers hold almost as many as they take on increasing pension exposures. Both use the RPI swap market to hedge their inflation exposures. As this aggregation is based on the average of relative prices rather than a ratio of averages. with the new index published on a Tuesday in the middle of the following month. Figure 35: Breakdown of RPI by major category Leisure goods and services 10. RPIX and CPI inflation 6% RPI y/y RPIX y/y CPI y/y Food. over £212bn. using Eurostat’s HICP principles. with tolerance bands remaining at 1% either side of the target.5% Transportation costs 14. This statistical bias. although it was never formally adopted as a targeted inflation series for monetary policy. Raw data for the RPI are collected in the middle of each month. By the end of 2009. beverages and tobacco 25. with the corporate linker market relatively moribund despite about £37bn in non government RPI-linked bonds outstanding at the end of 2009.5% RPIX to 2.0% CPI.9% -1% -2% 1997 1999 2001 2003 2005 2007 2009 Source: National Statistics 15 March 2010 73 . linkers made up 24% of UK government bond debt by market value.41-0.64% per annum throughout 2001-09).1% Clothing and footwear 3.

for £1bn. which move with the long-term average value of house prices. In the long run. particularly as components can be changed over time. the change in methodology should smooth but not bias the RPI-CPI spread. a higher weighting for airfares in CPI. 15 March 2010 74 . This means that the liabilities increase each year by the rate of RPI inflation capped at a certain level. usually 5% prior to April 2005. in particular the exclusion of owner-occupied housing costs. such as sharply falling used. Hence. mostly due to house price components. it was widely estimated (including by the Bank of England) that the long-term bias between RPI and CPI inflation would be about 75bp. based on a smoothed house price index and mortgage interest payments. with RPI based on only used car prices but the CPI based on new and used car prices.3%. This is mainly split between a depreciation component. in 2008 RPI inflation turned significantly negative even though CPI never fell below 1%. This was based mainly on the formula effect of approximately 50bp. weights can vary significantly. Currently 17% of the weight of the RPI is linked to house price elements that are not included in the CPI. RPI weights are calculated based on a variety of sources. The majority of pension fund liabilities accrue on a limited price indexation basis. This was despite pressures in other coverage and weights tending to offset the formula effect. but this could be reduced if owner-occupied housing costs were included in CPI.5%. with 17 bonds (colloquially referred to as “stocks”) still outstanding on a real yield curve that stretches out to 2055. Council tax makes up a further 4% of the RPI basket while not being included in CPI. While the data comprising most remaining series are the same for RPI and CPI.5% to 2. Apart from aggregation methodology. the 2% Sep 1996. In the following six years. albeit with a wide range of -3. This is because almost all inflation-linked liabilities are based on the RPI. the main difference between UK RPI and CPI is in the coverage. but because it is being introduced when rates are at a historic low it is a significant deflationary impetus through the period of rate normalisation. The mortgage rate series was changed from the standard variable rate to the average effective rate from the February 2010 RPI data. the difference between RPI and CPI has averaged 0. An ongoing commitment from both the Treasury and the investor base has seen the market grow to £341bn by market capitalisation at the end of 2009. RPI inflation can be significantly below that of CPI. the UK inflation-linked market will remain almost exclusively linked to the RPI. the most important of which is the expenditure and food survey. while the use of imputed values for CPI means a heavier weighting for financial services. The broader coverage of the latter measure means. but not new. The other significant coverage difference is car prices. 28 bonds in total have been issued. as well as standard variable mortgage rates. car prices. was auctioned by the UK Treasury on 27 March 1981. 33% of the total outstanding value of the gilt market. Ahead of the change in the monetary policy inflation target in 2003. which remains one of Eurostat’s long-term aims. Development of the UK inflation market The first index-linked gilt. Barclays Capital economists estimate a long-term RPI-CPI spread of about 0. Trying to tie down a long-term steady state for this basis is fraught with difficulty.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Even though monetary policy is now focused on CPI inflation.3%. The standard deviation of the basis since 1997 has been 1. for instance. and in periods of declining house prices and Bank of England rate cuts. plus the expectation that in the long run housing costs would grow in line with wage growth.9%. with an implicit floor of 0. whereas CPI weights are based on the household final consumption expenditure component of GDP. The switch to an average effective mortgage rate series from the Bank Rate sensitive standard variable rate series implies a less volatile mortgage interest payment series.

investors will be given the right to sell bonds back to the government at indexed par (par. annual RPI inflation moved between 6% and 26%. Formal remit revisions can happen at any time. Linker auctions are single-price – ie. The prospectuses of linkers issued before the 2035 bond contained “comfort language”. The remit contains an estimate of the total size of linker sales. just ahead of the beginning of the new fiscal year in April. Planned auction dates for the year are released at this time. which use a multiple price mechanism. although they are subject to remit revisions. While in the past only some nominal 15 March 2010 75 . Auction stocks are announced quarterly. The majority of issuance since November 1998 has been via auction. to be carried out in the new fiscal year. and with the exception of the boom and oil price shock at the end of the 1980s and the sharp rise in energy prices in 2008. By 2007. In the event of changes to the coverage. but two key times are early in the new fiscal year once the prior year’s finances are known and in the final quarter of the calendar year. all successful bidders pay the same price. starting with the first syndication by the UK government of the 2055 linker. a new curve has been built up of Canadian-style bonds. by cash value. adjusted for inflation). within two years of the inception of the market. indexation of debt was not a new idea in the UK – the UK Government’s National Savings department had been issuing inflation-linked savings certificates for retail investors since 1975. This rapid issuance of new bonds contrasts with the 2035 old style linker. just after the Pre-Budget Report is announced. In the ten years prior to the launch of the first linker. prompting a strong demand for inflation protection. although that is not of great comfort at present because all stocks under this protection are trading well above indexed par. and there is often guidance as to how plans might be altered in the event of changes to the health of public finances. which are auctions of about half the size of regular auctions that are announced at shorter notice and for which there is no post auction allotment. with the proviso that there is consultation with a body with “recognised expertise in the construction of price indices”.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The creation of a linker market was formally recommended by the “Committee to Review the Functioning of Financial Institutions (1977-80)” (known as the Wilson Committee. or calculation of the RPI. Early breakevens were about 9%. with six issues stemming maturities of 2017 and longer being launched by the end of that year and ten by 2010. in contrast to nominal gilt auctions. Other than a bond issued in 1983 that was convertible into nominal debt. For the 2035 and new Canadian-style bonds issued by the Debt Management Office (DMO). giving some protection against adverse RPI measurement changes. Since 2009 there is an additional 10% of stock for each auction that may be sold to successful bidders at the clearing price in a post-auction facility. although for the first year only pension funds were allowed to buy the new asset class. The DMO also sells linkers via mini-tenders. an expected inflation accrual cost considerably higher than has been realised. which was the only bond launched between 1992 and 2005. This generally coincides with the Budget. while Keynes recommended the move as early as 1924. all new issuance was in Canadian-style bonds. However. Government funding plans are laid out annually in a “Gilt Remit” within the Treasury’s “Debt and Reserves Management Report”. Issuance of indexed debt contributed to the credibility of the government’s anti-inflationary rhetoric by diminishing the incentive to debase the real value of the outstanding debt. This choice will be “conclusive and binding”. but the programme proved a windfall for government finances. all gilt linkers issued prior to 2005 had the same idiosyncratic format that has not been copied elsewhere. RPI dropped below 5%. has remained below 5% to date. the choice of linking index is at the Chancellor’s discretion. which the Bank of England (acting as “index trustee”) deems “materially detrimental”. Institutional buying at auctions must come via index-linked gilt-edged market makers (IL GEMMs). as inflation fell. after its Chair Sir Harold Wilson). Since September 2005. indeed almost all prior to 2009. with size details announced on the Tuesday in the week before an auction. the 2% 1999 dubbed “Maggie Mays”.

Syndicated gilt linker issuance was not used again until 2009. switch auctions. In the middle part of the decade inflation derivatives became the main source of inflation liability hedging. which were all issued with maturities of 22 November. the market remains far too small to address the liabilities completely. taps and other market activities. As equity markets declined in the following two years. it quickly grew to dominate end-user demand. the DMO conducted two syndications in Q3 09: the launch of a 2050 linker for £5bn notional in September 2009 and a new 2060 maturity nominal gilt in October 2009. It was not until 2004 that pension specific. Even with the government’s increased issuance. which saw £5bn notional of the new bond issued. in recent years all nominal GEMMs have also traded linkers. in contrast to older new-style linkers. there remain notably more RPI-linked liabilities in the UK than available assets. though the imbalance is steadily reducing given the sharp increase in government issuance. became the main driver of the UK swaps market. though with a few exceptions. the PS16/04 regulation hastened the use of inflation swaps by the insurance sector. given scheme closures and most remaining exposures being capped at 2. Although the DMO stated in its response to the consultation on the supplementary methods for distributing gilts that it expected syndicated issuance to occur no more than once a quarter. Syndicated issuance of gilt linkers was first used in September 2005 for the launch of the 2055 linker. However. rather than insurance business. when the DMO noted in its response to the consultation on the supplementary methods for distributing gilts. it is rare for the linker repo to stray far from general collateral rates for long. however. index-linked gilts remain the main stock of assets held against pension liabilities. when the Minimum Funding Requirement (MFR) encouraged pension funds to favour gilts. at 32% of the market compared with 34% in mid-2009 according 15 March 2010 76 . asset swaps were sufficiently attractive for significant supranational swapped supply. In particular. The DMO subsequently announced in its FY09/10 gilt remit that it would issue a combined £18bn of gilt linkers via syndication and mini-tender in FY09/10. that it would use syndicated issuance alongside the auction programme in FY09/10 to issue larger volumes of gilt linkers and long nominals than it judges would be possible via auction. due to the extreme extension of the curve and the innovation of the new bond format. after the syndicated launch of the 2042 linker in July 2009. they covered their life policy RPI exposure before focusing on immunising bought-out and in-house pension portfolios. Insurance companies now hold almost as many gilt linkers as pension funds. As recently as 2003 insurers and pension funds owned over 90% of all gilt linkers. such as the IL32. Fundamentally. While there has been some non-government issuance of sterling inflation-linked bonds since the mid-1980s and RPI swaps have traded since the early 1990s. IL GEMMs can also access the DMO’s standing repo facility (10% of the Bank of England’s Policy Rate at the end of 2009). The PPF liability measure is more conservative than most schemes report in their accounts with measurement based on nominal and inflation-linked gilt curves. the life insurance industry began to focus on the potentially more accurate liability-matching benefits of using inflation derivatives rather than bonds. but their share of the total market has fallen significantly since. This helped to kick start what was until then a niche RPI swaps market. largely due to inflation hedging increasingly occurring via derivatives. The Pension Protection Fund (PPF) survey of 7800 pension schemes showed estimated private sector pension liabilities to be £905bn at the end of December 2009. until the early 2000s gilt linkers were by far the dominant feature of the UK market.Barclays Capital | Global Inflation-Linked Products – A User’s Guide market makers were IL GEMMs. In 2000. These banks are also the route through which the DMO can conduct syndications. March 2009. Initially. but having been spurred on by heavy swap activity around the time of the IL55 launch in September 2005.5%. there are few new inflation liabilities accruing in the private sector. The 2050 linker was the first gilt linker to use the new March/September coupon series.

but the percentage of the market held this way has been declining steadily as investors have increasingly moved towards liability focused investment strategies even in their bond portfolios. The new-style issues have a virtually identical framework to US TIPS. or foreign. There is no longer a generally accepted discount curve dominating all others. This is likely to have encouraged the use of swaps. the purchaser immunises as much risk as possible. The owners of new facilities will be required to set aside funds for decommissioning costs. However. there can be a significant dislocation to the market. Buy-out demand was probably the largest single source of end-user demand in the linker market in 2008. When schemes close they often also choose to de-risk. Neither the Pensions Regulator nor the Pension Protection Fund has been actively pushing schemes backed by strong sponsors towards lower-risk or liability-driven solutions. The Pension Protection Fund encourages risk reduction by schemes that are in assessment for being taken over by the Fund. Pensions’ regulation encourages recognition of the nature of pension liabilities and has been a major factor behind increased inflation-linked demand since the 1990s. Over 5y indices are more widely used than all linker indices. the risk of pension funds has been transferred to the insurer). as the uncertainty of their future liabilities is considerably reduced without active members. particularly if closure also involved an injection of cash from its sponsor. typically with a much higher asset allocation into inflation than before the liabilities are transferred. Mechanics of UK linkers While new issuance is in Canadian-style linkers and most of the stock of UK linkers is now in new-style issues. which are likely to be closer than gilt linkers to the FRS17/IAS19 accounting definition. Details of how funds will be formally valued have yet to be announced but they may be invested more conservatively than those held versus ultra longdated pension liabilities. with the exception that they have no deflation floor (ie. but a scheme’s trustees decide whether to use this injected money to immunise risks. The Pensions’ Regulator has pushed for pension schemes to be safeguarded when firms are taken over by lower-rated. but it is not prescriptive beyond this universe even though its Section 179 liability estimate is referenced off the gilt linker curve.Barclays Capital | Global Inflation-Linked Products – A User’s Guide to National Statistics data. entities. there are still many old-style bonds. with insurers tending to have more regulatory flexibility if their base investment is in gilt linkers than swaps. Buy-outs are typically priced off a gilt linker curve. which according to the original White Paper will be valued against an ultra long real discount function. we expect nuclear liability hedging to become a new source of demand for ultra long linkers in the coming years following the announcement in 2008 that the government will allow new nuclear power plants. While there are few new RPI pension liabilities accruing in the UK. This means that when a bond drops below five years. but the bonds remaining in the index are likely to be supported due to an extension of the duration of the index. whereas in 2009 a more important driver appeared to be the significant number of very large pension schemes that were closing further accrual to existing members. This is mostly life insurers who are matching real annuity obligations and increasingly pension fund obligations that have been bought-out (ie. in recent years regulation has surprisingly not prescribed how pension funds should address their exposures. with the bond itself liable to underperform due to forced selling. and it is not a coincidence that the yield on the FTSE over 5y gilt linker index used to be the reference for MFR liability measurement. which has led to significant capital injections into pension funds. The Nuclear Liabilities Financing Assurance Board (NLFAB) will oversee investment of these funds. they 15 March 2010 77 . Typically when a scheme is bought out. unlike the MFR around the start of this decade. Traditionally most linkers were held versus an index benchmark rather than as directly matching liabilities.

the cash value of semi-annual coupons for old-style linkers are calculated as follows: Coupon paid = ⎛ C ⎞⎛ RPI m−8 ⎞ ⎟ ⎜ ⎟⎜ ⎝ 2 ⎠⎜ RPI i −8 ⎟ ⎝ ⎠ C RPIt m i is the quoted annual coupon is the RPI for month t is the payment month is issue month Where: Similarly. Old-style linkers also have no deflation floor.25% IL17 is shown below. the eight-month lag means that the principal value of the 2. Instead of trading in real space.15014 Dirty price 293. investors “gain” the inflation for the eight months prior to the bond’s issue. linkers first issued in the 1980s now trade with prices well above £200.11487 Accrued interest 2.5% 2016 (old-style) UK IL 1. paid on an actual/actual basis. Accrued interest is then calculated in the usual way for gilts. but calculations for old-style linkers are more complex and are set out below.Barclays Capital | Global Inflation-Linked Products – A User’s Guide can be redeemed below par if RPI falls over the lifetime of the bond). old-style linkers trade in clean price cash terms (not real). unlike for new-style linkers.78 Note: As of the close of 30 December 2009.5% IL16 and the new-style 1. Figure 37: Example of difference in pricing styles between old-style and new-style linkers Linker UK IL 2. and accrued interest is calculated on the cash value of the coupon. in a positive inflation environment. Source: Barclays Capital To trade in nominal space.493 103. with the traded price rising and falling to reflect inflation that has occurred.25% 2017 (new-style) Clean price 290. As this demonstrates. Since the price of an old-style linker already includes accrued inflation. it is necessary to know the inflated value of the next coupon so that accrued interest can be calculated. will be uplifted by the ratio of the RPI for December 2012 versus June 1984. but “lose” the inflation for eight months prior to the bond’s maturity.5% IL13. hence. The mechanics of new-style linkers are addressed earlier in this guide.79840 0. with a real price and with settlement amounts uplifted or reduced to reflect the inflation experienced over the life of the bond. but using the known inflated value.29 115. issued in February 1985 and redeeming in August 2013. but they have already accreted a considerable amount of inflation.72 Index ratio N/A 1. An example of the difference in price evolution between the old-style 2. So. and it takes some time to gather and publish the price information for the final month). no index ratio is used to create the settlement price. This term mismatch is not especially large in a benign inflation era. the effect has been large on the realised return. the cash value of the redemption amount is: Redemption value = ⎛ RPI r −8 ⎞ 100⎜ ⎜ RPI ⎟ ⎟ i −8 ⎠ ⎝ 78 15 March 2010 . at times. For example. the clean price of the old-style linker tends to drift higher. but history shows that. As a result. the inflation indexation for the coupon of an old-style linker is done with an eight-month lag (a coupon’s cash value will need to be known six months before it is due. Using this methodology.

10%). Once the “money yield” is found. 15 March 2010 79 . for old-style linkers. it is straightforward to calculate the settlement amount for an old-style linker. we are required to use an inflation assumption to determine the future value of the bond’s cash flows. A few extra stages are required to find the real yield to which the nominal price of the oldstyle linker equates. Therefore. Therefore. thereby allowing us to convert the nominal price of the old-style linker into a nominal yield (also referred to as the “gross redemption yield” or “money yield”). the coupon payments and redemption value of an old-style linker are then mapped out according to the assumed future path of RPI this creates. the internal rate of return. the inflation assumption is then removed to give the “real yield” by using the following calculation. originally. future RPI prints define the value of a UK linker’s future cash flows. Using this RPI assumption. which is translated into a real yield by using the same calculations as for a nominal bond and into a nominal settlement amount by using their index ratio.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Where: r is the redemption month. this assumption has been 3% per annum since the mid-1990s (prior to this. This is irrelevant for Canadian-style linkers because they are quoted at a real price. making the future cash flows of an old-style linker uncertain. thus preventing a nominal yield calculation. we simply add the accrued interest to the clean price (which already includes uplifted inflation). However. RPI t = RPI t −1(1+ f ) 1 12 Where f is the RPI assumption (in this case 3%). or “money yield”. From these cash flows. it is assumed that the RPI grows at an assumed rate beyond the last RPI print. which is the convention: ⎛ y⎞ ⎜1+ ⎟ 2 ⎛ g⎞ ⎝ 2⎠ ⎜1+ ⎟ = (1+ f ) ⎝ 2⎠ 2 Where g is the real yield. it was 5% and. To arrive at the nominal yield or “money yield” of an old-style linker. we must first know the nominal value of all of the bond’s cash flows to maturity. y is the money yield and f is the RPI assumption. an unknown RPI for month t is given by. However. To derive the yield metrics of a bond from its nominal price. By convention. This inflation assumption is then removed from the “money yield” to create the real yield. is then calculated for any given dirty price in the same way as the internal rate of return for a nominal bond.

16 March 2005 update. d2 = Cash flow due on the next but one quasi-coupon date per £100 face. hence. gives a closed solution real yield formula.03 ⎠ ⎝ 1+ f ⎠ 1 1 n = Number of coupon periods from next quasi-coupon date to redemption. is best valued using a money market yield). g = Semi-annual real yield. RPIL = Latest published RPI k = Number of months from the month whose RPI determines the next coupon to the month of the latest RPI a = RPIL 12 u RPIB 80 2k 15 March 2010 . Any errors of duplication are ours.gov.aspx?docname=/giltsmarket/formulae/yldeqns. The real yield formula covers traditional bonds with two or more cash flows left (when there is only one coupon remaining.uk/documentview. including accrued) per £100 face. a practical spreadsheet calculation is less so. weekends and holidays may mean true payment dates differ. It also highlights the detail of Canadianstyle linker calculations. The yield formula. r = Number of days from settlement date to next quasi-coupon date. Readers should refer to the above official publication to see complete details. u= ⎛ 1 ⎞2 ⎛ 1 ⎞2 ⎟ =⎜ ⎜ ⎟ ⎝ 1. and we have also trimmed and altered the wording of the explanatory notes.pdf&page =Formulae/Calc.dmo. c = (Real) coupon per £100 face. the bond has known nominal value and. w= 1 1+ g 2 f = Assumed inflation rate (3% is the current convention). www. expressed algebraically. d1 = Cash flow due on the next quasi-coupon date per £100 face. RPIB = The Base RPI for the bond – ie. The term “quasi-coupon date” is the theoretical cash flow date determined by the redemption date. is daunting. for the month eight months prior to issue date. s = Number of days in coupon run containing settlement date.Barclays Capital | Global Inflation-Linked Products – A User’s Guide DMO’s Real Yield Formula The DMO’s “Formulae for calculating gilt prices from yields”. for n ≥ 1 2(1− w ) ⎣ ⎦ Where: P = The “dirty” price (ie. r r r +n ⎡ ⎤ acw 2 P = ⎢d1 + d 2 (uw ) + (1− w n−1)⎥(uw ) s + 100au s w s .

4 1. At those shorter maturities.1 1. November RPI printed 216.6 -2.0 0.1bp. if the price of the bond was unchanged.4 0.1 2. the real yield is more a function of the interaction of the nominal yield and the market’s view of inflation expectations.7 -0.32% 0.09% 0. of course.3 216.1 -0.3 0.5 -0. old-style linkers effectively price the assumed nominal cash flows.8 -1. For example: Figure 38: Sample mechanical real yield adjustments on old-style linkers for an RPI print (bp) Projected RPI (November 2009) Index value 216.7 1.3 2.2 0.1 2.6 1. for instance.9 217.1 m/m IL11 Real yield adjustment (14 December 2009) IL13 IL16 IL20 IL24 IL30 IL35 Note: Example uses projection for November 2009 RPI print.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Old-style linkers in practical terms The linker market and the nominal gilt market are.4 -4.6 -0. a change in inflation expectations brought about by the RPI release. its real yield on the above calculation would alter so that the m/m inflation rate would deviate from 0.5 13. If an average of 3% y/y RPI inflation was expected to prevail over the life of the old-style linker.6 -0.1 -8.0 -1.7 5.28% 0.2 1.9 -0.5 216.1 -2.19% 0. a projected set of cash flows assuming 3% inflation would have to be priced such that the money yield of the linker would be equal to the yield of the nominal gilt.28% m/m. this equated to a mechanical cheapening of the IL13 real yield by 1. and the relative pricing-in theory should depend on the outlook for the RPI.4 -5.3 2.05% 0. and the price of the old-style linker will mechanically adjust to reflect realised inflation. Each time an RPI inflation print is released.247% (ie.4 5. This is especially true of the shorter maturities that are less dominated by supply and demand considerations. The real yield on the old-style linker may also adjust for changes to its market price.7 216. which printed 216. due to.1 -0.0 -0.4 216. If exactly zero inflation were expected for an old-style linker.1 -2.6 0.6 216.4 0.4 -0. 3% annualised). all else being equal.6 -0. a projected set of cash flows that is created using the 3% inflation assumption will have to be priced such that the yield they would generate is 3% above that of the conventional gilt for the yield on the inflation-linked bond to be at fair relative value.9 -0.0 6.3 -3. 15 March 2010 81 .3 0.3 0.8 -1. Source: Barclays Capital In this example.4 3.1 0.46% -11.7 -3. However.2 -5. 0. As Figure 38 shows.9 1.1 216.23% 0. real yields on oldstyle linkers mechanically adjust upwards immediately after the RPI release by the amount of the difference between the actual data and the assumption.4 1. Hence.2 216.4 0.0 0. Mechanically.247% m/m inflation assumption embedded in the bond’s cash flows. the real yield on an old-style linker will only adjust on an RPI release to reflect this mechanical effect.0 0.5 -1.8 -0.5 -1.6 0.5 -0.4 -1. so the money yield that they generate is: y = NY + (3% – BE) where NY is the comparable nominal gilt yield and BE is the market’s assessment of the appropriate breakeven inflation rate. Therefore.9 1.0 0.1 1.14% 0.9 -0.4 -1. competing assets.9 2.1 0. it replaces one month of the 3% inflation assumption with the actual value.6.3 0.8 1.9 3.1 7.8 216.37% 0.2 -0. slightly higher than the 0.8 10.42% 0.6.

12 Source: Barclays Capital Although some investors may feel more comfortable with the gradual adjustments of the mark-to-market value of the inflation-linked bond with respect to the price index (with the 3-month interpolated lag) that the Canadian model incorporates. if the month-on-month RPI print were zero and if the price of an old-style linker with exactly one year to maturity is unchanged then mechanically the real yield would fall by 0. As breakevens and actual inflation trend away from 3%. the price of the old-style linker would have to fall 0.247% for the yield to remain unchanged. The discrete shifts in yield purely as a result of a new RPI release may seem counterintuitive.247%. old-style UK linkers have an advantage in terms of the speed with which they compensate the investor for the value of an inflation surprise. this argument does not offset the downside of the eight-month lag for coupon indexation. whereas the yield of a bond with a duration of 25 years would be reduced only 1bp on an unchanged price.14 2.247% will a shift in yield be created that is worthy of a strong reaction from market prices. the UK market can react immediately and price in all available information as soon as it is published.16 2. there are clear differences between the UK and Canadian methodologies that need to be controlled in order to compare the relative merits of positions. In the final year of a bond’s life it can be 15 March 2010 82 .18 2. are predicted by the market. However. As noted. Therefore. they add to mechanical yield volatility but not to price volatility. but this is simply a function of the difference between the “nominal” quoted price and the “real” quoted yield.19 2. To the extent that strong seasonals. As a result. be they negative or positive.Barclays Capital | Global Inflation-Linked Products – A User’s Guide As Figure 38 shows.20 2. the yield effect on an unchanged price will be much greater on a short bond than on a long one. the assumed RPI schedule will only be lifted higher by a m/m RPI print above 0.15 2.247%. the use of this assumption can become very misleading for shorter maturity bonds. Overall. as the real yield will appear to rise as the price also rises. Instead of waiting up to six weeks for the effect of a large monthon-month number to gradually filter through to the reference index ratio.13 2. In terms of carry. an unchanged price for a bond with 1-year duration would mean a yield reduced 25bp. the price adjustment of old-style linkers in response to a relatively high or low m/m RPI is the UK linker market’s equivalent of inflation accretion. In the event of a flat m/m RPI print.17 2. only if a m/m RPI print is notably above 0. The most significant difference to bear in mind is the effect that the presence of an embedded 3% assumption has on carry and the market’s reaction to carry. Figure 39: The effect of RPI releases on the assumed RPI schedule 220 219 218 217 216 215 214 213 212 May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May 2nd Cashflow New assumed RPI after Sep release New assumed RPI after Aug release Current Assumed RPI schedule Known 3rd Cashflow 2. with the real yield adjusting for the deviation from the 3% inflation assumption and the price adjusting for the realised inflation. and will fall for any number below it. Alternatively.

This also means that losses. Investors in corporate inflation-linked bonds do not enjoy the inflation relief afforded to gilt linkers. the relief will be based on the RPI change from December to December. so they get all gains – inflation-linked. certain issuers might be able to obtain an exemption. December.Barclays Capital | Global Inflation-Linked Products – A User’s Guide difficult to meaningfully compare its screen real yield with the yield on its nominal comparators or the real yield of longer-dated gilt linkers. It is worth noting that most index-linked gilts are held by pension funds or within the pension business lines of life assurance companies. There are two reasons it is not the same: if an investor’s tax year-end is. We prefer a z-spread asset swap comparison for micro valuations of old. Tax usually has a material effect on the market only at sub-5y maturities where institutional investors are less dominant. While this is an appropriate simple method to compare moves in real yields and breakeven. whereas indexation occurs with an eight-month lag. the market value of non-government bonds in the Barclays Capital Sterling Index was over £23bn. The UK’s Inland Revenue decided that since corporate issuers are allowed to offset the inflation uplift against taxable income – in the year that it accretes – corporate linker investors should not receive inflation relief. The RPI level that offers a consistent Libor spread compared with the nominal gilt then offers a more representative breakeven value. not on inflation compensation.and new-style gilt linkers because this measures the inflation elements versus a consistent curve. A more straightforward. and the difference is taxed. are not allowable against tax. the distortions due to the embedded 3% assumption in traditional linkers do not make the spreads between bonds using this approach representative. no inflation credit is applied. or otherwise – tax free. without a lag. method is to compare the yield of traditional linkers with that of new-style linkers forward to the end of their known carry period because this leaves the same known inflation data in the pricing of both bonds. and the starting value in any tax year is unlikely to be exactly indexed at par. The inflation uplift is taxable – ie. It is fairly straightforward to evaluate the nominal return for a short linker for any given percentage change in the RPI over the remainder of the bond’s indexed life at the current price. This lack of structured product demand has resulted in a less active inflation swaps market at the short end of the curve. value for a traditional linker is better assessed in an alternate way. say. Taxation issues UK gilt linkers have a more favourable tax treatment than either UK corporate or international government bond issues. However. For most corporation tax purposes. In this situation. but less complete. Expressed as a spread to Libor. The relatively favourable tax treatment for gilt linkers coupled with a range of governmentbacked national savings products means that a sterling structured inflation note market has not developed. This relief is deducted from the total return (calculated on a mark-to-market basis or on an accrual basis. especially at shorter maturities. in the event of falling RPI. an inflation tax relief is granted based on the inflation experienced between tax year-ends. it can then be compared with nominal gilts. according to an election made by the investor). This is almost the same as saying that the inflation increase in principal is not taxable. non-government inflation-linked issuance from anything other than government guaranteed Network Rail was extremely limited in 2008 15 March 2010 83 . Corporate UK linkers At the end of 2009. in contrast to other countries. Private individuals who hold UK index-linked gilts only pay tax on income accrued during the financial year. which do not pay tax. However. This means that index-linked enjoy a material tax advantage over nominal gilts – the aim and effect is that investors are only taxed on their real return.

PFI deals involve a private company building infrastructure and being paid an income stream until the asset comes under the ownership of the relevant authority. almost half of the M80 road widening project was financed by EIB. there were about £20bn of corporate inflationlinked bonds that had been monoline-wrapped to transform them into AAA issues based on utility. the logjam of PFI projects was eased in March 2009 by the government offering backup funding via the creation of the Treasury Infrastructure Financing unit. hence. if the bond route is not available. PFIs almost always issue bonds rather than paying swaps. as the water regulator has indicated that the regulatory structure will remain RPI-linked over the long term. despite the improvements in the credit markets. At its five-year fixing of water prices in November 2009. although other large PFI projects have also involved partial financing via inflation-linked. Yet. while many utility linkers have durations above 30.Barclays Capital | Global Inflation-Linked Products – A User’s Guide and 2009. most new corporate linkers were not bought by bond investors and. In the 15 March 2010 84 . Hence. most ended up with asset-swap investors. This contrasts sharply with 2005-07. Recent PFI deals that have gone through have also had a larger percentage of funding from the European Investment Bank (EIB). which saw unprecedented corporate inflation-linked bond issuance. Utility companies have been a major source of corporate linker issuance. the pipeline of deals remaining in this format is limited. particularly hospital bonds. issued the majority of their debt in inflation following the five-year fixing of water prices at the end of 2004. PFI and infrastructure projects. However. but not LPI and other more structured formats. Water companies. as with utility supply. hence. Of those issued between 2005 and 2007. it is rare for that of PFI bonds to be much longer than 15 years. The largest single utility issuer in corporate linkers is National Grid Plc. but less than £200mn issued in 2008 and 2009. the water regulator announced a £22bn investment programme in the network. unless they can prove they have an expectation of closely matching cash. Private Finance Initiative (PFI)-related deals have been another source of non-gilt issuance. are ideally suited to funding via inflation-linked issuance. Utility companies taken over by private equity and infrastructure consortia may be able to use inflation derivatives with mark-to-market considerations less of a concern. For instance. However. this kind of issuer has been numerous. in particular. who provided for the inflation swap needs of pension funds by translating these into real rate swap paying flow. National Grid Plc had over £4bn in inflationlinked bonds outstanding as at end-2009. the vast majority had been absorbed by asset swap investors. this demand effectively ended. Most hospitalrelated projects involve RPI-linked cash flows that will be paid to the financier once the hospital is operational and. PFI financing has been a less important factor than utilities in recent years as most. Many regulated utilities have a degree of RPI-based pricing restrictions and. hence. By the end of 2007. particularly those with inflation-linked revenues. are amortising either immediately or after a number of years when the building project is expected to be complete and the offsetting of cash flows begins. a clear inflation-linked revenue link. Almost all of these issues were wrapped with credit guarantees to enable AAA ratings and. In duration terms. Thus. the overhang of old monoline wrapped supply may limit the scope for corporate linker supply from utilities. particularly covered bond investors. However. which has electricity and gas prices linked to RPI. as AAA insurers themselves lost their ratings. during this period. Due to specific accounting restrictions. this may well prompt further issuance of inflation-linked bonds. If there is demand. IAS39 accounting standards make it difficult for corporates to pay inflation without facing the mark-to-market volatility of these positions on their balance sheets. Despite difficulties in sourcing PFI funding from 2008 due to adverse credit conditions. with the largest bond £450mn notional. While PFI has been a source of significant supply over the past decade. are not eligible for inclusion in the Barclays Capital Sterling Index. Straight RPI bonds can be sold and accounted for on an accruals basis.

as the launch of the IL55 in September 2005 and resultant tactical asset-swap activity provided sufficient momentum for the market to become self sustaining. even with asset swapping still being a feature of demand for Network Rail issuance. in recent years. we would not expect it much beyond 20 years on the curve.4bn over this period including private placements. EIB has been able to access inflation swap markets directly rather than hedging inflation-linked cash flows by issuing inflation-linked bonds. long-end flows. Property transactions are the main source of LPI supply. the frequent asset swapping of new non-government inflation-linked paper had provided an environment in which gilt linkers did not always trade at a discount in asset swap to nominal gilts. only contractually RPI-linked revenue streams are likely to be hedged by projects directly. It was traditionally characterised by just a few banks looking to match pension demand and corporate supply. While in recent years most long supply was absorbed by asset swappers. As a natural AAA name backed by a letter of support from the government. with these banks concealing what were often reasonably large. in line with its aim to provide an alternative to gilt linkers. and securitisations in the sector are likely to increase if the commercial property market stabilises. Network Rail has functioned as a very important source of non-government inflation-linked supply in the absence of any other significant activity. and distortions between inflation swaps and gilt linkers became extreme as position unwinds and an absence of inflation swap supply drove swaps richer than bonds. but infrequent. Previously. We believe there is substantial potential for inflation flows stemming from property in the coming years. but this is exaggerated by some unswapped issuance in the past from EIB. having launched three benchmark issues at 20y. and property flows remain important today. which issued £6. However. even so. Network Rail is unusual in conducting much of its real rate funding on a programmatic basis. Once leveraged investors ceased to be a supplier of non-government inflation-linked flows via asset swapping of wrapped corporate inflation-linked issuance while pension demand for swaps continued. For a property revenue stream linked directly to inflation. 15 March 2010 85 . given that the overhang of old monoline wrapped issues is concentrated beyond the 30y sector. liquidity in UK inflation swaps fell dramatically in 2008. UK inflation swaps saw notably more inflation duration traded than any other market between late 2005 and early 2008. The early corporate linker issuance in the 1980s was mostly linked to property. The only major non gilt inflation-linked supply during 2008 and 2009 came from transport infrastructure. via swaps. Inter-dealer broker trading still makes up a lower percentage of activity than in other major markets. This is no longer seen as.5bn of natural AAA bonds have been issued by foreign entities since the start of 2000. although overall LPI supply has reduced significantly in recent years. shorter on the curve an active swaps market sometimes enabled natural AAA issuers to issue to realmoney investors and swap back themselves. However. as many contracts have explicit inflation floors and caps. when it was lending to PFI projects. the increase in gilt linker supply in 2008 and 2009 was barely enough to offset the overall absence of non-government inflation-linked supply. straight paying of inflation or real rate swaps is possible. RPI and LPI swaps and other inflation derivatives While RPI swaps have been traded since at least 1994. specifically Network Rail. Over £3. prior to 2005 the derivatives market was more opaque than that in the euro area. the market gradually became unbalanced. 30y and 40y in 2007 and continuing to re-open these in the following two years. if non-government bond supply does resume.Barclays Capital | Global Inflation-Linked Products – A User’s Guide absence of a wrapped bond market.

Uncapped RPI linkages were relatively uncommon outside of the ex-public sector. Prior to 1997. Forward trading in inflation and real-rate swaps is present.5% cap so far out of the money that it creates no new hedging demand. increased property-related paying encouraged an improved liquidity in LPI in general and LPI (0. albeit rare. In 2005 a large proportion of pension demand in inflation swaps was in LPI (0. liquidity to take positions in short-dated RPI swaps remains poor. both zero coupon and in a par-swap format. under the Pensions Act of 2004. This led to relatively limited trading in LPI (0. While domestic pension funds provide the main source of demand for RPI swaps. and to some extent this aids liquidity and helps to reduce relative value distortions across the curve. Hence.3) and LPI (3.2. The outperformance of bond over swap breakevens was also briefly helped by PFI paying flows in mid-2009 after the logjam of projects was ended by the government offering backstop funding via the creation of the Treasury Infrastructure Financing unit. This flow was accelerated following the gapping wider of relative z-spreads versus nominal gilt asset swaps in March 2009 on the announcement of BoE gilt purchases. From a liability perspective. even at long maturities. Many of these distortions have corrected significantly. The most frequently traded maturity is still 30y. particularly between 1997 and 2005 when. At the same time. causing problems to the pensions sector. real-rate RPI swaps can also trade directly. there was a minimum statutory entitlement for new defined benefit liabilities to accrete annually at the rate of RPI inflation capped at 5%. it is no longer consistently superior to inflation-linked bonds. in which there was regular issuance via non-government asset swaps. despite the quantitative easing deliberately not focusing on linkers to avoid falling real yields. An exception was in H2 07 when higher and more volatile RPI forwards drove a modest discount for LPI (0. The level of linker asset swaps fell in 2009 from the extremes reached in Q4 08 as pension funds and life insurers who had already immunised inflation and duration were significant buyers of gilt linkers on asset swap. with LPI (0. to see trading in forward swaps versus other markets.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The absence of asset swap paying flow led to significant distortions between inflation bonds and swaps. While the underlying liquidity of the inflation swap market recovered significantly in 2009. and the inflation swap and gilt linker markets still interact via asset swap activity.5). which often drives sub-5y interest in other countries. providing a consistent comparison with nominal swaps. While the most common format of inflation swap in the UK market remains zero-coupon RPI swaps. While 10y is relatively frequently quoted. LPI (0. in practice linker yields were affected very significantly. which prompted renewed pension interest.5) was the most common pension specification but exposures varied scheme by scheme.5).5).5) in particular. Short-dated linker asset swaps also benefitted from demand from bank treasuries. LPI (0. with the 2.5) other popular variants.5) no longer traded at a discount to RPI swaps at a 30y maturity due to the supply/demand imbalance. By Q2 06. under the Pensions Act of 1995. they are also traded by a variety of other leveraged and real-money investors. not helped by the lack of a structured note market. the legal minimum for new pension liabilities is LPI (0. now concentrated in gilt linkers. Many property rental streams 15 March 2010 86 . most pension exposures are linked to LPI rather than RPI. despite the cap level being much closer to spot RPI levels than the 0% floor. but in the absence of significant supply in this format this drove LPI (0. It is also possible. with an implicit floor at 0.5). From April 2005. The standard initial lag for almost all swap transactions other than asset swaps is a two-month calendar lag.5) swaps rich versus RPI. LPI (0. This leaves very few new inflation liabilities accruing. with gilt linker asset swap levels becoming much cheaper than nominals in Q4 08 and long forward real rates pushing into negative territory.5).

As LPI bond issuance does not offer the same accounting advantage as RPI supply. the trends since 2007 have been towards simplification of risk and product. albeit also heavily skew. Such supply. the smile in RPI volatility to a greater extent than seen in other inflation markets.5).Barclays Capital | Global Inflation-Linked Products – A User’s Guide have an LPI-like structure. could be sufficient to kick start the swaption market. Barclays Capital 15 March 2010 87 . In addition to pure LPI flows related to LPI (0. However. By contrast with LPI. On the other hand. the volatility market is set to remain concentrated in derivatives format. there is also outright interest in straight RPI caps and floors at 3%. The LPI market has helped define. This is partly due to the lack of natural inflation volatility issuers and partly to a self-reinforcing lack of perception of liquidity. given how popular nominal rate and equity swaption strategies have proved among pension funds. the inflation and real-rate swaption markets remain surprisingly subdued. there may be scope from regulation. which is only partially RPI-linked to prompt conditional supply from regulated industries. but often the annual growth is capped at levels other than 5% and/or floored above zero.3) and LPI (3. with the extreme and unanticipated realised volatility of RPI inflation as well as market conditions discouraging non-linear risk taking. or a significant further pick-up in activity in LPI swaps. Figure 40: UK linkers: Historical performance and risk 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% -2% 1999 2001 2003 2005 2007 2009 -2 1999 2001 2003 2005 2007 2009 0 -1 1 UK IL Returns UK IL Ann Mthly Vol Figure 41: Return/risk linkers versus nominals and equities 4 3 2 Equity IL Nominal Bonds Source: Barclays Capital Source: Datastream.

where weights are adjusted each year to reflect changes in spending patterns. from January 2007 the MIAC decided to begin releasing CPI numbers based on the chain-weighted calculation method as a supplement to the fixed-weight CPI. 40% of JGBis originally issued were retired by the end of 2009. tend to be overestimated in the calculation because there is an increasing impact due to the fixed weightings. which compares the prices of goods and services according to their weights at the time of the base year (currently 2005) with the y/y change in those figures. such as fresh fish. including a preliminary mid-month index for Tokyo and a nationwide index covering the previous month. the upward bias in the inflation rate based on the fixed-weight index has not been particularly high – on average 0. the few corporate linkers issued so far as well as yen inflation swaps. goods that have undergone steep price drops in recent years. unbalanced ownership. the retail prices for each item in the various municipalities nationwide. the weight (quantity) of various goods and services is fixed at the base year. Japan’s CPI broadly conforms to the international standards regarding consumer price indices set by the International Labour Organization (ILO). which increases the further one moves away from the base year.000 of the total consumption expenditure of Japanese households (including imputed rent in goods and services). Goods undergoing price increases.2% larger than the inflation rate as measured by the chain-weighted core CPI. which are disseminated at 8:30 AM on Friday of the week that includes the 26th of each month. 15 March 2010 88 . The index is commonly referred to as the “core CPI.liiceanu@barcap. The breakdown of the Japanese core CPI is shown in Figure 42. The prices of items used in calculating the consumer price index are. As a result. However. with the last auction in August 2008. ordinary retail prices of items that are actually sold). The Japanese CPI is calculated based on the Laspeyres method. The basic 10 categories have been globally standardized.” and it is released monthly by the Ministry of Internal Affairs and Communications (MIAC) Statistics Bureau. the MoF’s efforts have been concentrated on absorbing excess supply from the secondary market via buyback auctions. all reference the non-seasonally adjusted consumer price index excluding perishable food items. with the likelihood that new bonds will include a principal floor to stimulate demand from domestic market participants. The MoF has penciled in a JPY300bn JGBi auction for FY 10. heavily tilted toward overseas leveraged investors. Accordingly. on the other hand. the Laspeyres method tends to show an upward bias. The data. helping 10y breakevens to rebound to around -50bp from under -300bp. such as IT goods. Within five years there were 16 bonds with total capitalization of nearly JPY10trn (USD96bn). To address this issue. but there are differences in their respective weights by country. coupled with the 2008 global financial market turmoil. To date. forced a suspension of the program. Since then. vegetables and fruits (Bloomberg ticker: JCPNJGBI <Index>. in principle.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Japan Stefan Liiceanu +81 3 4530 1554 stefan. do not reflect any increases in real volume accompanying such price drops (for the calculation of the price index for such items the MIAC uses the Hedonic adjustment method based on the correlation between the price of products and their performance derived from available sales data). perishables are excluded due to their very high volatility). Under the Laspeyres Price Index.com The Japanese Ministry of Finance (MoF) has been issuing inflation-linked JGBs (JGBis) since March 2004. according to retail price data (ie. currently cover 584 items (the number of items can vary depending on the base year employed) that account for at least 1/10. The Japanese CPI: Technical overview Inflation-linked Japanese government bonds.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Figure 42: Major items in the Japanese core CPI (2005-base index)
Misc., 6.1% Recreation, 11.5% Education, 3.8% Transp. & Telecom, 14.5% Medical Care, 4.7% Clothes & Footwear, 4.8% Furniture & Household Utensils, 3.6% Food, 22.7%

Housing, 21.3% Fuel, Light & Water, 7.1%

Source: Ministry of Internal Affairs and Communications (MIAC), Barclays Capital

Meanwhile, base revisions to the index are conducted once every five years for years ending in “5” and “0”. The last such adjustment took place on 25 August 2006, when the CPI data were rebased to the year 2005 (ie, the July 2006 nationwide CPI had the year 2005 as base and 2005-base data were made available retroactively from January 2005). A total of 34 new items were added to the coverage, including digital electronics such as flat-panel TVs and DVD recorders. The next rebasing will occur in mid-2011, when the index’s base year will change to 2010. Further details on the Japanese CPI in the English language are available from the MIAC Statistics Bureau (www.stat.go.jp/english/data/cpi/index.htm). In the context of monetary policy, the Japanese core CPI naturally plays an important role, being the BoJ’s “target” index, along with the corporate goods price index (CGPI). However, unlike some other central banks, the BoJ’s definition of desirable long-run price stability based on the core CPI is rather loose. According to the Bank, Monetary Policy Board members understand medium- to long-term price stability as “an approximate range between zero and 2%,” with “most members’ median figures falling on both sides of 1% 2”. Moreover, the BoJ stated that this view of price stability is flexible, “reflecting changes in the economic structure such as further progress in globalization and innovations in information and communication technologies” and that “as a rule Policy Board members will review it annually”.

Inflation-linked government bonds
Market development and supply
Discussions about the possible issuance of inflation-linked government bonds in Japan kicked off in September 2002, when, during a regular meeting with investors, the Ministry of Finance submitted a draft proposal on the introduction of this asset class. Market participants’ opinion varied considerably, some arguing that linkers would be a useful diversification instrument also seen in other developed markets, while others commented that without a deflation floor (one of the bond’s features proposed in the issuance plan) it would be hard to allocate capital to linkers. Some market participants also voiced concern about issuing such instrument in deflationary times (spot inflation had ranged between -0.7% y/y and -0.9% y/y throughout 2002).

2

See for example the BoJ Quarterly Bulletin, May 2006.

15 March 2010

89

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

As deflationary pressures eased and borrowing needs remained high, a year and half later, in March 2004, Japan issued its first inflation-indexed government bond (JGBi1), with a 10y maturity and a small pilot size of JPY100bn. The bond came at a low 15.5bp BEI (spot inflation was -0.1% y/y at the time), but shortly thereafter, market perceptions vis-à-vis the new instrument turned favourable and breakevens subsequently rallied to a high of 94bp during the same year (for details please refer to Figure 44). Although the bond’s maturity was not particularly long compared with the structure of other developed linker markets, and although Japan had had a prolonged deflationary experience since the late 1990s, this first linker issue, as well as all subsequent JGBis, did not feature a deflation floor, as per the initial proposal brought forward by the issuing authorities. This initially left a notable gap between supply and demand for deflation protection. The main reason that a floor was not included was to make the real yield a relatively representative value of expected long-run inflation. Otherwise, the value of the embedded floor would have been a significant part of the value of the new bond, given that inflation had averaged -0.3% y/y throughout 2003 and market perceptions generally were not particularly inflationary. The key product specifications of JGBis are illustrated below. Figure 43: JGBi product specifications

Maturity: Type of Issue: Coupon Frequency: Minimum face value unit: Issuance Method: Auction Method: Recent Issuance Frequency: Reference Index: Reference Index frequency: Reference Index Seasonality Adjustment: Indexation Lag: Indexation Style: Floor: Strippable:
Source: MOF

10 years Coupon-bearing bonds Semiannual JPY100,000 Public offering Yield-competitive auction/ Dutch-style auction Bimonthly (even months) Japan nationwide CPI ex-fresh food (Japan Core CPI) Monthly No seasonal adjustment 3 months Canadian Model (linear interpolation to the 10th of the month) No floor Not strippable

Since 2004 and until the global financial shock in 2008, the market evolved rapidly, both in terms of size and depth. By August 2008, the MoF had issued a total of 16 bonds and the overall market size had reached almost JPY10trn (about USD96bn, or nearly 1.4% of total JGBs outstanding), placing Japan ahead of other inflation-issuing countries with longer experience with the product, such as Australia and Sweden. JGBi auctions typically were held on a bimonthly basis, with issuance amounts of JPY500bn per auction (for details please see Figure 46). Up until October 2008, when issuance was temporarily suspended, JGBi auctions were Dutch style, with bids being filled from the lowest yield (highest price) until the entire amount offered has been raised. Under this approach, the auction clears at a highest accepted yield and all bidders pay the same price for the bonds. Re-opening auctions in FY 06 also were Dutch style, but from FY 07 all re-openings (three held to date) became price-competitive auctions. The auction process itself has been identical with that of nominal bonds, with the issuing authorities first announcing the exact details of the offering (linker coupon and size) at 10:30 on the morning of the auction, at which point the bidding begins. The bidding ends at 12:00 and the auction results are announced at 12:45 3.

3

News providers such as Reuters and Bloomberg disseminate the information in real time (for Bloomberg for instance please refer to <NI JPB>).

15 March 2010

90

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Meanwhile, JGBis became eligible for MoF’s buyback operations in January 2007. Until April 2008, the MoF held five buyback operations, on each occasion retiring a modest JPY40bn to JPY50bn worth of linkers. From April 2008 the Ministry stepped up its JGBi buybacks (auction sizes of approximately JPY80bn) in response to the sudden decline in linker prices that occurred around March (the on-the-run 10y linker’s BEI briefly touched the -2bp level, as illustrated in Figure 44), a time when deleveraging by overseas investors distorted many asset prices. As the global financial shock reached its climax during September-October 2008 and JGBi BEIs sank to unseen levels, the MoF cancelled two JGBi auctions planned for October 2008 and February 2009, and increased the size and frequency of buybacks. From October 2008, monthly JGBi buybacks averaged about JPY215bn, with two auctions of JPY100bn to JPY120bn. The precise buyback amounts have usually been determined on a quarterly basis, with the choice between retiring linkers, floating-rate JGBs or fixed-coupon JGBs made following MoF hearings with the primary dealer community (since October 2008 the weight has been placed on linkers rather than floating-rate or regular JGBs). Overall, the MoF held a total of 39 JGBi buyback auctions during 2007-09, retiring JPY3.74trn worth of linkers from the secondary market. This represents nearly 40% of the total JGBi issuance since the program’s inception and leaves the market size a more modest JPY6trn as of the end of 2009. Some issues, such as JGBi2 and JGBi6, have been aggressively retired, with a current outstanding market size of less than c.40% of the original amount issued. At the time of writing, of the existing 16 JGBi issues, four are especially large – namely, JGBi8, 12, 14 and 16, each accounting for at least 10% of the broad JGBi market. It can be said that the original purpose of JGBi buybacks – to stabilize the market – has been largely achieved. Long-dated BEIs recovered to near -50bp by the end of 2009 versus a low of -323bp in mid-December 2008, and offer-to-cover ratios at buyback auctions have declined significantly, falling to 2x or less from around 4x in early 2009 (Figure 45). Simply put, buybacks helped address distressed selling and thus absorb excess supply in the secondary market. Since the middle of 2009, most buyback auctions have been orderly, with tight tails and average prices higher than market levels prior to the auction. Note that in addition to MoF buybacks, from the end of 2008 JGBis have also become eligible for BoJ outright purchase (Rinban) operations, under which the Bank buys JPY1.8trn/month of government securities directly from dealers. The BoJ’s outright JGBi Figure 44: History of 10y JGBi breakevens (Mar 2004-Jan 2010)
150 100 50 0 -50 -100 -150 -200 -250 -300 -350 2004 2005 2006 2007 2008 2009 2010

Figure 45: Bid-to-cover ratios and auction tails in JGBi buybacks
6 5 4 3 2 1 0
22 Jan 07 20 Aug 07 18 Feb 08 19 May 08 04 Aug 08 24 Oct 08 14 Nov 08 12 Dec 08 29 Jan 09 20 Feb 09 13 Mar 09 24 Apr 09 19 May 09 25 Jun 09 16 Jul 09 24 Aug 09 15 Sep 09 23 Oct 09 13 Nov 09 17 Dec 09

+250 +200 +150 +100 +50 +0

Buyback Tail (sen, RHS) Offer-to-Cover Ratio (x, LHS)
Source: MoF, Barclays Capital

Note: from Mar 2004 to mid-Sep 2005 on-the-run JGBi BEI, thereafter seasonally adjusted CMT 10y BEI derived from parametric curve fitting. Source: Barclays Capital

15 March 2010

91

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

purchases have been more modest than the MoF’s buybacks, being held on a bi-monthly basis (usually odd months) in amounts of JPY40bn. As of the end of 2009, the BoJ’s total JGBi holdings amounted to JPY454.4bn, or just 13% of total JGBi buybacks by the Ministry of Finance. Future supply of JGBis will depend on how fast market conditions normalize and allow the MoF to resume its regular auctions. The FY 10 JGB issuance plan announced at the end of 2009 does incorporate a JPY300bn JGBi auction (and total buybacks, including linkers, of JPY3trn), which is to be implemented if the secondary market environment is favorable, ie, breakevens return to positive territory. Interestingly, the official minutes of regular meetings between the issuing authorities and dealers as well as domestic investors have revealed strong support for this asset class, particularly if it is restructured to include a principal floor. Although there has been divergence in domestic market participants’ opinions with regard to the optimal timing to resume JGBi issuance (the minutes showed than some preferred issuance of floored bonds to restart only after all current JGBis would have been retired, while others argued for earlier supply of JGBis), overall domestic market participants have argued that unlike floating-rate JGBs, inflation-linked bonds are a global asset class and urged MoF not to terminate the program. From the issuing authorities’ perspective, it also makes sense to foster the linker program. Japan’s overall debt levels have been rising sharply, with its gross debt-to-GDP ratio now near 200%. Rollover bond issuance is likely to soar in the next decade (official estimates for yearly rollover JGB supply until FY 20 are between JPY104trn and JPY140trn), and budget deficits of at least 3% of GDP are also likely to persist. In the past few years, domestic appetite for floating-rate bonds has deteriorated sharply, as did interest from Japanese households for retail JGBs. This leaves Japan’s overall funding channels at a disadvantage, particularly as overseas investors currently own less than 6% of total JGBs outstanding. As rising government bond supply volumes are usually accompanied by the necessity to diversify funding channels, inflation-indexed JGBs should be an appropriate choice of issuance in the medium to longer run. Meanwhile, as is the case in other developed linker markets, trading volume in inflationindexed JGBs has naturally been lighter compared with nominal bonds, but liquidity did improve significantly up until the middle of 2008, with bid/offer spreads in the 10y sector tightening from 20-25 sen to around 10 sen between 2005 and 2008. Liquidity tended to be somewhat lower in the front end of the JGBi curve, as issues JGBi1 and JGBi2 were originally small (JPY100bn and JPY300bn, respectively), but longer-dated linkers featuring larger sizes (eg, JGBi8, JGBi10, JGBi12, JGBi14 and JGBi16 all were originally issued in JPY1trn size) exhibited better liquidity. Following the 2008 market turmoil and the MoF’s aggressive JGBi buybacks, liquidity deteriorated to 2005 levels, with bid/offer spreads between 20 sen and 30 sen, and daily turnover of about JPY5-10bn, compared with JPY50bn or more at the peak of the market.

15 March 2010

92

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Figure 46: JGBi issuance – auction history, 2004-09
Redemptio n date 10 Mar 14 10 Jun 14 10 Dec 14 10 Jun 15 10 Sep 15 10 Dec 15 10 Mar 16 10 Jun 16 10 Sep 16 10 Dec 16 10 Mar 17 10 Jun 17 10 Sep 17 10 Dec 17 10 Mar 18 10 Jun 18 Amount issued (JPY bn) 99.8 299.5 499.7 499.6 499.6 499.6 499.7 499.7 499.9 499.7 499.8 499.8 499.7 499.4 499.5 499.7 499.7 499.5 499.9 499.8 499.6 Secondary non-comp. auction (JPY bn) 0 0 0.7 28.6 43.9 8.8 0 0 0 0 31.5 2.3 4.7 8.6 17.9 0.8 7.8 6.3 39.6 48.9 1.0 MoF + BoJ repurchas es (JPY bn) 0.3 179.4 201.2 257.8 265.6 326.8 250.5 398.1 168.6 593.5 239.3 288.9 230.9 331.5 157.8 300.5 Amt outst. by bond (as of Dec 09) 99.5 120.1 299.2 270.4 277.9 181.6 249.2 601.5 331.1 439.9 265.1 736.5 269.6 681.8 381.7 748.8 5,953.9

Aucti on no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Auction date 4 Mar 04 3 Jun 04 7 Dec 04 7 Jun 05 6 Sep 05 6 Dec 05 7 Mar 06 6 Jun 06 3 Aug 06 5 Oct 06 7 Dec 06 6 Feb 07 5 Apr 07 07 Jun 07 07 Aug 07 04 Oct 07 06 Dec 07 07 Feb 08 03 Apr 08 05 Jun 08 07 Aug 08

Bond JGBi1 JGBi2 JGBi3 JGBi4 JGBi5 JGBi6 JGBi7 JGBi8 JGBi8R JGBi9 JGBi10 JGBi10R JGBi11 JGBi12 JGBi12R JGBi13 JGBi14 JGBi14R JGBi15 JGBi16 JGBi16R

Coupon 1.2 1.1 0.5 0.5 0.8 0.8 0.8 1.0 1.1 1.1 1.2 1.2 1.3 1.2 1.4 1.4 Total mkt size

Bid/Cover ratio 4.84 7.50 3.58 2.51 3.41 3.23 3.69 3.59 4.23 5.34 4.84 3.21 4.49 4.85 4.11 4.01 3.45 3.63 2.92 4.07 2.52

Note: “Secondary Non-Comp. Auction” refers to the additional JGBi amounts issued in secondary non-competitive auctions (held about two hours after the main auction), where primary dealers have the right to buy up to 10% of amount issued in the main auction at the clearing price. Source: MoF, BoJ, Barclays Capital

Indexation features and calculations
The JGBi inflation indexation mechanism is very similar to the Canadian style adopted by most developed inflation markets. That is, within a given month, the rate of inflation accrual is constant at the rate of month-on-month inflation between the inflation index three and two months previously. However, there is one notable difference – JGBi inflation accrual is based on the 10th of the month (rather than the 1st) due to the relatively late release of inflation data each month, which would otherwise cause uncertainty at the end of some months (for example, the CPI data for March are usually published in early May rather than at the end of April due to the traditional Golden Week holidays). Both JGBi principal and coupons accrue based on the ratio of the Daily Reference CPI (DRI) value to the Base Reference Index (base CPI) at issuance (this number obviously does not change throughout the life of the bond, although the rebasing of the CPI every five years requires additional adjustments to link the old and new indices). This ratio is referred to as the “CPI ratio,” or, simply, the “inflation ratio.” The official calculation formula for the daily reference CPI and the CPI ratio is illustrated below. First, the daily reference CPI (DRI) for day N in month M is: 1) If N = 10, the reference CPI is the index three months previously, ie, CPIM-3 2) If N > 10, the reference CPI is:

15 March 2010

93

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

(N-10) DRI = CPIM-3 + (CPIM-2 – CPIM-3) x No. of days from 11th of month M to the 10th of month M+1 3) If N<10, the reference CPI is: No. of days from 11th of month M-1 to N DRI = CPIM-4 + (CPIM-3 – CPIM-4) x No. of days from 11th of month M-1 to the 10th of month M

Next, the CPI ratio for any given day N is calculated as DRIN/Base CPI An important feature of JGBis is that due to rounding conventions, inflation accrual does not develop smoothly across the month. As mentioned above, the ratio of the reference CPI on settlement to the base CPI on which the settlement price is based is only rounded to three decimal places (and the CPI itself is published to one decimal place, unlike other countries such as the US), in contrast to other markets where it is usually rounded to five decimal places. Because of this, carry on a 10y JGBi jumps almost 1bp when the rounded index ratio changes by 0.001, a situation not seen in other inflation markets. The number of these “carry jumps” exhibited by linkers’ CPI ratio in any given month obviously depends on the magnitude of the month-on-month change in the CPI. JGBis are traded in real price terms – ie, without incorporating inflation adjustment. In the broker market, linker prices move in 5 sen increments, and daily closes are also rounded to the nearest 5 sen. While this is not a strict rule, it is an appropriate degree of accuracy given the approximately 10 sen jumps in nominal price terms that occur with inflation accretion. Settlement and day-count conventions applied to the nominal market are also used for JGBis (T+3, ACT/365). The simple bond calculation between price and yield is used, and therefore, real yields are determined by the rate that equates the traded price with the sum of the bond’s cash flows discounted to present value.

Accounting and tax treatment
The accounting treatment of JGBI up until 2006 was not favorable in the sense that the principal portion of a JGBi was classified as an “embedded derivative” security by the Accounting Standards Board of Japan (ASBJ), reflecting the possibility that holders may not receive their principal back at par if deflation is persistent enough. Mark-to-market gains and losses on such securities had to be recognized via the income statement, exposing earnings to increased volatility, instead of through the balance sheet, as is the case for most ordinary bonds. Because of this, many domestic investors shied away from the linker market. In 2006, however, the accounting treatment of JGBis was relaxed, and now gains or losses on JGBis do not have to be immediately reflected in the income statement. Furthermore, at the end of October 2008 the ASBJ allowed domestic investors to book their holdings of illiquid instruments, including floating-rate JGBs and JGBis, at theoretical value rather than market levels. To date many domestic investors, including large banks, have used this accounting rule for floating-rate notes, but the absence of a benchmark forward CPI curve made the theoretical valuation of inflation-linked JGBs problematic from an accounting perspective, although economists’ long-term estimates of inflation running at least 10 years forward are available (Bloomberg’s CPIJ screen).

15 March 2010

94

there was still uncertainty as to the treatment for investors usually exempt from withholding tax for nominal JGBs. The only issuer of note of straight inflation accreting bonds was the Japan Finance Corporation for Municipal Enterprises (JFM). if any. One of the supportive factors for 4-5y 15 March 2010 95 . including the classical “additive structures” and “multiplicative structures” (eg. which in 2005 was able to issue JPY40bn of 10y bonds at yields well through those of matched-maturity JGBis (the first bond’s coupon was 0. Moreover. In Japan. Furthermore.45% versus prevailing government bond real yields of between 65-75bp) because it featured a deflation floor. No tax is applied when the principal falls in value due to a decline in prices. In the short end. investor interest picked up both in shorter and longer maturities. The EIB also issued a JPY50bn accreting bond as opportunistic swapped funding in 2004. while transactions in inflation structured notes are lagging the European and US markets. While the inflation element of 10y notes has been straightforward to hedge. all tenors including 1y inflation swaps have traded. All of the derivatives mentioned above feature the same three-month indexation lag and interpolation to the 10th of the month. “designated financial institutions” are not subject to Japanese withholding taxes. Corporate and derivative markets As of the end of 2009. prompting significant interest in buying 5y5y forward breakevens. forward-starting swaps. taxable. This bond offered a par floor that made it attractive to real money investors. provided the bonds are held in book entry form”. liquidity was initially concentrated in the ~6-10y zone (bid/offers similar to JGBis). which greatly limited upside at a time when JGBi breakevens were 85bp. while in the longer end both 15y and 20y inflation swaps have traded. equivalent to 96bp per annum. despite having an aggressive principal cap at 110%. more innovative 5y inflation-linked term deposits have been offered by regional financial institutions. Forward and relative spread trades have also become increasingly important to the pricing of the JPY swap breakeven curve. the corporate market in Japanese inflation-linked bonds was limited. Pricing the belly of the Japanese CPI curve has been particularly complicated by expectations of consumption tax (VAT) hikes. The National Tax Agency formally clarified at the end of 2005 that JGBis’ interest and the gains or losses on principal would be exempt from withholding taxes if they are held by “entities entrusted to manage corporate pensions approved under the tax systems in the UK and the US. mainly because the tax status of the new bonds was not set. although in recent years other structures such as year-on-year inflation swaps. Regarding the zero-coupon swap market in Japan. The increase in the inflation-adjusted principal. the note’s payoff is determined by y/y core CPI growth plus a spread and a floor. The main demand for structured products involving Japanese inflation has been in the 5y and 10y sectors. many of the popular structures in these areas have been priced. but since early 2006. In the domestic market. Most overseas investors were excluded from directly owning JGBis before April 2005. total return swaps and asset swaps have become increasingly traded. Further out the curve. since 2008. 30y inflation swaps are regularly priced but activity has been limited. also maturity matched to a JGBi.47% and the second one’s 0.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Income tax is payable on JGBi coupon payments. or by a multiple of inflation with a floor of 0% or higher). Even when holding restrictions were relaxed. there have been times when the 5y sector of the swaps curve has been driven rich. the typical inflation derivative transaction has been the zero-coupon inflation swap. at the redemption date is considered to be an interest payment and so. as such.

This risk is likely to be again priced into the JPY inflation swap curve.0% (the CPI pass-through at the time of the 3% sales tax implementation in FY 89 and its subsequent hike by 2% in FY 97 has been at least 70% within the first two months). Barclays Capital 15 March 2010 96 . so the shorter end of the swap curve has generally been extremely sensitive to the expected timing of a VAT increase. Since 2007. given the apparent strength of the German economy through its 3% VAT hike (indeed as of March 2007. Prior to 2008. which would be worth 30bp on a 5y swap. a VAT hike appears to be only a matter of time. Expectations for a VAT hike naturally weakened as business conditions deteriorated following the 2007-08 global financial turmoil. major Japanese think tanks expected two hikes. a level about 12bp cheaper relative to nominal asset swaps.5-2. Figure 47: Japanese linkers – historical performance and risk 15% 10% 5% 0% -5% -10% -15% 2005 Source: Barclays Capital Figure 48: Return/risk linkers versus nominals and equities 4 3 2 1 0 -1 -2 2005 Equity IL Nominal Bonds Japan IL Returns Japan IL Ann Mthly Vol 2006 2007 2008 2009 2006 2007 2008 2009 Source: Thomson Datastream. asset swap margins normalized as well. JGBis started to trade on asset swap as spreads have been tighter versus Libor relative to nominal asset swap spreads of similar maturities. mid-market. but with the Japanese government moving back to an expansionary fiscal policy and tax revenues currently lower than yearly JGB issuance associated with the budget deficit. currently hovering around Libor+15bp to Libor+24bp across most issues.Barclays Capital | Global Inflation-Linked Products – A User’s Guide swaps into the start of 2007 was increased expectations of consumption tax hikes sooner rather than later. with no significant maturity differentials. the major asset price disruptions in 2008 left JGBis exceptionally cheap on an asset swap basis. one as early as 2009 and the second one around 2012-13). with the long end of the JGBi curve at times indicated at Libor+80bp. as the market’s cheapness gradually corrected throughout 2009. long-dated JGBi asset swaps were quoted at Libor-4bp to Libor-8bp. The direct impact on the Japanese CPI basket of a hike of a similar magnitude would likely be to boost consumer prices by at least ~1. As has been the case in other inflation-linked bond markets. However.

The index uses geometric means at the first-stage aggregation of collected price data. Information on consumer expenditures is gathered through the Survey of Household Spending and the Food Expenditure Survey. food and recreation. which uses random samples of Canadian households. RRBs made up 6. making quality adjustments where possible. Source: Statistics Canada 15 March 2010 97 . the index comprises eight major components.2bn.4% of total marketable Canadian government debt. with an adjusted principle amount of $34. The fixed basket price index is an arithmetic average of price relatives for all single commodities contained in the basket. The CPI Index Canadian RRBs are indexed against the Not Seasonally Adjusted All Items Consumer Price Index. furnishings and equipment 11% Note: Reflects 2005 basket at April 2007 prices. clothing. Issuance has always been concentrated in the 30y sector to facilitate pension fund demand.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Canada Michael Pond +1 (212) 412 5051 michael.pond@barcap. The weights are determined based on family expenditure surveys that are conducted periodically. The current weights are based on the 2005 survey. 9. and there are currently five issues outstanding. consumer savings and investments etc. which includes owner-occupied and rented accommodation.5% of Canadian government bonds and 31% of those with longer than 10 years to maturity. As of February 2010.com Real Return Bonds (RRBs) were first issued by the Canadian government in December 1991. The CPI includes only consumer items and excludes personal income taxes. It includes all Canadian families and individuals living in urban or rural private households. The basket consists of about 600 goods and services including transportation. As Figure 49 shows. The index attempts to capture innovations in final prices. education and reading 12% Health and personal care 5% Transportation 20% Clothing and footwear 5% Food 17% Shelter 27% Household operations. which include any changes in the Goods and Services Tax as well as provincial retail sales taxes. Figure 49: CPI weights for major components Alcoholic beverages and tobacco products 3% Recreation. The CPI index reflects pure price movements only as the basket includes goods and services of identical or equivalent quantity and quality over time. The component with the highest weight is the shelter component. The index measures price changes using the cost of a fixed basket of commodities through time. The index is weighted to reflect typical spending patterns. housing.

reducing the available float. from time to time. respectively. Thus far.25% 2026 being issued in 1995. international investors have taken advantage of real yield differentials versus other more heavily traded international markets such as the UK and US. the 3. The BOC currently operates under a quarterly funding schedule with one 30-year RRB auction every three months. in 2005 (Figure 50).0% 2041 in 2007. up from C$300-400 before 2007. A large portion of RRBs outstanding are held as an offset to future pension liabilities. the 4. the Canadian inflation-linked bond market has been largely the domain of pension funds. Therefore. Because of the tendency for pension fund investing to be of a buy-and-hold nature. and RRB issuance can have a valuation impact on the long-end of other markets. the 4.0% 2036 in 2003 and the 2. the Treasury has decided to issue new bonds at four-year intervals. RRB issuance has been relatively constant despite a recent increase in total borrowing needs (Figure 51). However. This pattern means that the initial maturity extends by one year with each new issue. The Bank of Canada (BOC) acts on behalf of the Department of Finance for the purpose of managing the financing program.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The Real Return Bond market The initial Real Return Bond issue. RRBs tend to trade at general collateral levels in the repo market due to the willingness of the funds that own the bonds to lend them. the 4% 2031 in 1999.4% and 66%. down from 8. secondary market liquidity can be difficult to obtain. We expect the quarterly issuance pattern to continue as the government appears committed to using the program as a cost-effective way to diversify its investor base. Figure 50: RRB issuance as a percentage of Canadian government issuance 70% 60% 50% 40% 30% 20% 10% 0% RRB Issuance as a % of Total Bond Issuance RRB Issuance as a % of 30yr Issuance 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: Bank of Canada Given its maturity profile and size. RRBs made up 2% of all Canadian government bond issuance and 28% of issuance in the long-end. At present. was a 30-year maturity and is now the shortest RRB bond on the curve. Similar to the US. In 2009. bid/offer spreads tend to be relatively wide versus much more liquid markets such as the US. the Bank of Canada issues about C$500-700mn in RRBs each quarter. 15 March 2010 98 .25% 2021. However.

The change in methodology allowed for simpler valuation and has assisted in the relative value analysis of the product versus conventional bonds as well as cross-currency real yields. RRBs do not have a par floor on the inflation adjusted principal. pioneering a simplified approach to the indexation of inflation for real return bonds.0 100. This change eliminated the effect of real yields changing when the inflation index is published. including all newly issued UK inflation-linked gilts. The reference CPI for any day during the month is calculated by linear interpolation. The change in the indexation process introduced with the first RRB was quite dramatic.0 80. Coupons are accrued on an actual/actual basis and paid semiannually.0 60. Calculation methodology A reference CPI value is calculated for every day based upon the CPI values for three months and two months prior to the month containing the settlement date. This enabled a more contemporaneous measure of inflation and allowed the market to trade in real space without an embedded inflation assumption.0 0. This new methodology became known as the “Canadian model”.0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: Bank of Canada 2s 5s 10s 30s RRBs The Canadian model The Canadian Treasury was an innovator.0 40. The reference CPI for the first of each month is the index value of three months previously. Old style UK real yields vary each time there is an inflation release different from this assumption. Unlike some other countries that issue based on the Canadian model. Reference CPI for day ‘d’: (d − 1) (CPI t −2 − CPI t −3 ) + CPI t −3 m d = day of the month (eg. The concepts of forward real yields and forward breakevens have become determining factors in the relative valuation of international markets that have adopted this calculation method.0 20. The gross settlement price is calculated as follows: 15 March 2010 99 . and has been generally followed by all subsequent major issuers. The crucial change in structure was the use of an Index Ratio to inflate principal and coupon for a given settlement date.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 51: Canadian government bond and RRB issuance (C$bn) 120. the 1st implies d=1) m = number of days in that month The indexation factor is the reference CPI for the settlement date divided by the reference CPI for the base date. with the inflation lag reduced to three months from the eight months previously used by the UK.

Barclays Capital 15 March 2010 100 . which already hold most of the outstanding Canadian linkers. we look for a reasonable twoway market to begin to develop. the Treasury’s website provides more detail on these conditions: www. the Canadian Treasury is not ordinarily required to withhold tax from interest or principal paid on RRBs.fin. Demand for a Canadian inflation derivatives market is growing from pension funds. Capital gains are not taxed until realised.Barclays Capital | Global Inflation-Linked Products – A User’s Guide ( p + c)( CPI t ) CPI base p = clean price of the bond c = real accrued CPIt = Reference CPI at time t CPIbase = Base CPI Real return bonds are taxable for residents but are not subject to withholding tax for nonresidents. and the amount of corporate issuance as of the end of 2009 was rather negligible. Many Canadian investors have considered going to the US market as a proxy for Canadian inflation due to the lack of an inflation swap market in Canada. There were several larger inflation-linked bonds issued by provinces. as the interest for Canadian inflation swaps grows. RRBs’ income received and accrued is taxed in a given year while the inflation accretion on the principal is also taxed. seeking exposure through receiving inflation in swaps. Non-government issuance and derivatives Non-government issuance has been slow to develop in Canada. but the last of those was in 2008.html. with the same structure as the Canadian linkers and mostly with overlapping maturities. However. there appear to be other parties interested in paying inflation to reduce the exposure they get through governmental contracts or explicit inflation links.ca/invest/taxtreat-e. However. Figure 52: Canada RRB – historical performance and risk 20% Canada IL Return Canada IL Ann Mthly Vol Figure 53: Return/risk versus nominals and equities 6 5 Equity IL Nominal Bonds 4 3 2 15% 10% 5% 1 0 -1 0% -5% 1999 2001 2003 2005 2007 2009 -2 1999 2001 2003 2005 2007 2009 Source: Barclays Capital Source: Thomson Datastream. Swapped bank issuance picked up in 2008 but nearly came to a halt in 2009. For non-residents. For residents. Despite the lack of government issuance.gc. particularly Quebec and Ontario.

liiceanu@barcap. in order to exclude mortgage interest payments and direct effects of changes in indirect taxes and subsidies.com Chris Bettiss +44 (0) 20 7773 0836 chris. but there is an explicit government commitment to continue fostering the linker market.bettiss@barcap. while subsidies are excluded. With the share of linkers at this target and Swedish government debt levels barely growing. there has been no bias between HICP and CPI over the long run. The sector owning the majority of the market. us beverages & 6% tobacco 17% Communica tion 3% Transportat ion 14% Figure 55: Swedish inflation – CPI. Owneroccupied housing costs carry a weight of around 8% in the CPI. 2009 Furnishing & household goods 6% Miscellaneo Food. a chain-weighted index (with annual links going from December one year to December in the following year. From 1993. though the baskets differ (Figure 55). The price collection is performed around the middle of the month. CPIx and HICP 5% 4% 3% 2% 1% 0% -1% -2% 1998 Housing & utilities 27% Clothing & Restaurants footwear Recreation & hotels 6% Healthcare 6% & culture 3% 12% Source: Statistics Sweden 2000 2002 2004 2006 2008 2010 HICP Headline CPI Source: Statistics Sweden Underlying Inflation (CPIx) 15 March 2010 101 . Value-added tax is included in the index. and the index is published in the middle of the following month. Sweden’s monetary policy is based on underlying inflation (CPIx) rather than the headline CPI measure. The weights and samples are revised at the beginning of each year. The Swedish Consumer Price Index (CPI) Swedish inflation-linked bonds are linked to the Swedish CPI (domestically referred to as the KPI or Konsumentprisindex). HICP includes childcare costs and care for the elderly. the Riksbank adopted a 2% annual inflation rate target. issuance is extremely limited. accounting for about 7% of total inflation-indexed bonds outstanding. The nongovernment linker market remains embryonic. with symmetrical tolerance of 1%. On average. but are not included in Figure 54: Swedish CPI breakdown by major category. with 1980 as base year) compiled monthly and a useful proxy for the consumption patterns of the entire country.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Sweden Stefan Liiceanu +81 3 4530 1554 stefan. The index is published by Statistics Sweden and from January 2006 it features two decimal places (versus only one previously) in order to better align the index to European standards. The weights for the major groups are based on the Swedish national account statistics and the index uses regular prices paid by the public. Official Swedish statistics also include the release of HICP and this provides a consistent comparison to other European countries. For instance. 54% as of Q3 09.com The Swedish government has been issuing inflation-linked bonds since 1994 and has a long-run target that this should make up 25% of central government debt including expected inflation accretion. The composition of the index based on year 2009 weights is shown in Figure 54. which own linkers versus liabilities. is insurance companies.

0013 0. Historically once can distinguish three separate regimes in the development of the Swedish linker market.0014 1. Development of the market Although at times Sweden’s inflation-linked government bond program has been fraught with challenges.0022 Dec 0.0005 Oct 1.9979 0.9923 0.0005 0. ie.9986 1. which is the most favourable month for Swedish CPI seasonality (the September seasonal factor has been 1.9967 0.0041 0.9990 1. Most Swedish linkers have very favourable seasonality into redemption. In practice. in a tightening cycle.0006 Note: based on data sample from November 1999 to November 2009.0041 0. as the SNDO put it. 15 March 2010 102 . Most Swedish electricity is generated by hydroelectric power.9976 0.0008 Sep 1. redeeming at the start of December (for details please refer to Figure 57) and so benefiting from September CPI. seasonality in Swedish inflation is more extreme than in larger countries. making the 10y breakeven a less accurate measure.9957 0.0029 1.0010 0.0008 Aug 0. however.9961 0.0030 1. lasting roughly from 1994 until 1996 was a period when the program was poorly understood by investors and auctions frequently went 4 “Ten years with inflation-linked bonds – a new asset class has been established. 2004.9988 0. Dev. 0. The first one.Barclays Capital | Global Inflation-Linked Products – A User’s Guide HICP.9947 0.0053 0. one “cannot try inflation-linked bonds for a short time or hesitate along the way…the strategy must be carefully thought-out and long-term…the SNDO has issued inflation-linked bonds every year since 1994. although the 10y point on the curve has traditionally been the more liquid sector of the linker curve. but CPI will tend to be lower than the other measures as mortgage rates fall.9947 0. which means their final accrual is based on the very poor end-of-year and January seasonal factors.0028 0.0016 1.9990 1.9999 0. rather than because Swedish inflation is traditionally more volatile than in the euro area. Therefore. Barclays Capital The 5y breakeven is commonly quoted in the Riksbank’s quarterly Monetary Policy Report as a measure of inflation expectations.34% higher in that month). Figure 56: Swedish CPI monthly seasonal factors (NSA CPI/SA CPI ratio) Jan Average Min Max St.9962 0. The linker 1% April 2012 and the 0% coupon April 2014 bond.9969 0. the SNDO reduced the volume…but by nevertheless continuing to issue these bonds.0011 1.0006 Jul 0. The Swedish National Debt Office (SNDO) made it clear before the last euro referendum in September 2003.0032 1. 4 Ongoing commitment and proactive assessment on the part of issuing authorities of whether the product suits the domestic investor base have thus been a key feature of the Swedish linker program.0009 May 1.0009 Feb 0.0026 0. this leaves a bias towards higher CPI than HICP and CPIx.0006 Mar 1. particularly very cold winters or extended dry spells. seasonality shows inflation higher by 0. In that case.0034 on average in the past decade.0036 0. we have clearly demonstrated that we believe in the growth of this market”. including the 5y5y forward breakeven. This is due to the 10y sector of the nominal curve being relatively slow to develop. causing headline CPI to have a higher sensitivity than normal to the weather.0019 1. although in some years the demand has been sluggish.9936 0. even if Sweden were to adopt the euro. the Riksbank examines a range of maturities when assessing medium to long-term inflation expectations. resulting into relatively volatile carry. Meanwhile. Source: Statistics Sweden.0014 Jun 1.0025 1.” SNDO.0034 1.0023 1.0015 Apr 1. are issues that suffer from maturing at the start of April.0042 0. that there would be no change to the measurement index for domestic inflation-linked bonds.0007 Nov 1.

SEK bn) Share of linkers in central govt.9 10.5bn.000 5.000 2.992 3 44. the program reached a mature phase.000 1.6 25.5 4.5 3. inflation-linked bonds outstanding (LHS. SEK mn) 26. lasting from around 1997 to 2001.189 4.0 29.0 16. SEK mn) Bid-to-cover ratio (RHS. The lack of investor enthusiasm at this first auction was a harbinger of limited acceptance of the new product in the following three years or so and indeed. auctioned Dutch-style (single price auction).2bn of the bond versus original plans for SEK3.0 31. debt (RHS) Note: Inflation uplift included.6 17.174 Loan number 3106 3001 3105 3102 3103 3104 Total Issue date 27-Sep-05 01-Apr-94 26-Apr-99 01-Dec-95 01-Dec-97 19-Apr-99 Redemption date 01-Apr-12 01-Apr-14 01-Dec-15 01-Dec-20 01-Dec-28 01-Dec-28 Initial maturity 6.692 4 52. while the balance of linkers rose from SEK3. Sweden’s inflation-linked government bond program underwent numerous reforms that helped develop the market further.2 4. upon consultations with the government and the central bank. the SGIL 0% 4 Jan 2014 bond.9 Floored ? Yes No Yes No No Yes Coupon rate (%) 1. Figure 57: Swedish govt.0 3. the SNDO decided to launch a program of government inflation-linked bonds and started with a zerocoupon 20y instrument linked to the Swedish consumer price index.6 Residual maturity 2.4bn in 1996. from around 2002. Lastly.2 Note: Outstanding amounts as of end November 2009.783 169. Source: SNDO Figure 59: Linker amounts sold at auctions and bid-to-cover ratios (quarterly) 7.6 17.Barclays Capital | Global Inflation-Linked Products – A User’s Guide undersubscribed.9 18.000 6.3 22.9 18. Figure 58: Amount outstanding of govt.6 22. at times the SNDO was forced to cancel auctions and breakeven inflation rates were on the decline (please refer to Figure 62).059 52.509 198.986 4.017 52.0 3.924 61.0 0.5 20. During the second phase. The SNDO judged that a large portion of the bids were at too low prices and consequently it issued only SEK1.000 3. x) Source: SNDO.1bn in 1994 to SEK73. Barclays Capital 15 March 2010 103 . inflation-linked bonds Amount outstanding (real terms.000 12 10 8 6 4 2 0 2001 2003 2005 2007 2009 25% 20% 15% 10% 5% 0% 4.060 Amount outstanding (nominal terms.2 5.000 0 1999 Amount auctioned (LHS.075 42. Source: SNDO In April 1994. SEK mn) 25.5 Inflation accrual 7. inflation-linked bonds and share in government debt portfolio 250 200 150 100 50 0 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 Govt.1 22. with linkers enjoying stable demand from investors while evolving into a full-fledged debt management instrument for the issuing authorities.3 17.

while exchanges between bonds of different maturities are conducted on a “price risk neutral” basis. However. In other words. This procedure continues today. it had to be withdrawn from auction on more than one occasion. cash for cash. and thus perceptions vis-à-vis the merits of linkers moved to the back burner. when dealers could switch from the 0% April 2014 bond to a new issue to the 3. the SNDO announces the official bid yield for the bonds to be retired 15 minutes before the cut off of the auction and the results are released 10 minutes after the auction. As illustrated in Figure 61. 2. with outright auctions on the first and third days and a switch auction in between.5% Dec 2028 bond.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The long duration of the first bond (20y versus 8y for the longest-maturity nominal bond) is one of the reasons why the reception for this new instrument was chilly and the decline in spot inflation. a bond that it continued to sell throughout the year in small quantities at weekly auctions but in this bond’s case too. in January 1995 the SNDO issued SEK500mn of a 9y zero-coupon bond (0% April 2004 issue). tackling the problem from various angles. In addition. from around 2. To deal with this issue. with the SNDO announcing the terms for bond exchanges about a month in advance. The SNDO allowed private investors to purchase inflation-indexed bonds through primary dealers after each auction (June 1995) and later in 1997 it introduced retail-oriented inflation-indexed bonds (available to individuals and small companies and organizations). when a new bond is introduced. central government finances were stronger. the SNDO took a proactive and flexible stance. switch auctions played an important role in the restructuring of the SNDO’s debt portfolio. The second stage of the Swedish inflation-linked government bond market proved to be even more difficult as the CPI dipped into negative territory from September 1996 to May 1997 (average -0. The linker market-related reforms implemented during this stage can chronologically be summarized as follows: 1. In the following years. Exchange transactions of bonds with similar maturity are executed in “liquidity neutral” terms. leading some to doubt the SNDO’s long-term commitment to foster the inflation market. 15 March 2010 104 . also likely contributed.7% y/y). Implementation of switch auctions (officially called “exchange transactions”) enabling dealers to move from less popular zero-coupon inflation-linked bonds to couponbearing linkers (June 1998). with amounts scaled for duration gaps. due to sharp falls in the owner-occupied housing sub-component of the CPI (mortgage interest payments were on the decline as the Riksbank cut rates) and fundamental core CPI disinflation. in general the bonds offered in switch auctions have had maturities at least twice longer than those of the linkers bought back. On the day of the switch auction.4% y/y) and from July 1998 to February 1999 (average -0.5% in mid-1994 to 0% by August 1996. ie. equity markets were overall bullish. the first such auction was held on 23 June 1998. ie. investors can switch from their holdings of more seasoned bonds or simply buy outright. Linker auctions were usually held over three consecutive days. The switch option is technically available for auctions of new bonds as well.

the adoption of floored bonds did not negatively impact the performance of non-floored bonds but rather helped the overall stability of the market. more recently. SEK bn) Bid-to-cover ratio (LHS) Source: SNDO. Again. enabling conversion from the seasoned 0% 2014 bond into a new issue.2bp versus an average realized inflation rate of 0. Regarding the addition of deflation floors. a feature officially motivated by the need for the “international harmonization 5” of the Swedish inflation-linked market. inflation averaged a low 0.5% Dec 2028 non-floored bond issued in the previous year. and from non-floored bonds to floored bonds. A second switch auction was held five days later. Figure 64 illustrates the monthly average BEI spread between the 3. we can make two noteworthy observations. Indeed.3bp.5% Dec 2015 bond (floored issue) and the 0% April 2014 bond (nonfloored bond) as well as the evolution of spot inflation. although the deflationary experience was likely another objective reason behind this move. breakevens turned negative (the 2y BEI reached nearly -50bp while the 10y BEI was just 18bp above zero) but following the issuance of floored bonds. around 1998-99. the price differential between floored and non-floored bonds of similar maturity suggests the market has been consistently discounting the existence of the floor although it did not attach too much value to it. the 3.1bn of outstanding inflation-linked bonds (without inflation uplift). Barclays Capital Amount sold in the market (RHS. within the space of one year. average inflation was 1.5% y/y and the BEI spread was 6. not surprising given the Riksbank’s inflation target and the bond’s generally long duration. The first bond with a deflation floor was sold in April 1999 (the 3.7% y/y. the SNDO designed switch auctions both from non-floored zero-coupon bonds to non-floored coupon-bearing bonds. when Sweden experienced deflation. from August 2008 to November 2009.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 60: Maturity comparison of bonds exchanged in linker switch auctions 35 30 25 20 15 10 5 0 O ct 05 O ct 06 D ec 06 O ct 08 05 Se p ar 07 07 Se p ay 09 Figure 61: Linker amounts sold in switch auctions and bidto-cover ratios 14 12 10 8 6 4 2 0 9 8 7 6 5 4 3 2 1 0 O ct 05 O ct 06 O ct 08 05 ar 07 07 06 Se p Se p ec ay 09 4 3 3 2 2 1 1 0 M M M D Maturity of bond bought back (LHS) Maturity of bond issued/tapped (LHS) Maturity ratio bonds offered/bonds bought back (RHS) Source: SNDO. 5 “Swedish Debt Policy. Second.7% y/y. of the SEK169. being the destination of the 3. BEIs were stable at around 2% in early 2004 despite the arrival of another deflationary episode. the average BEI spread between the two bonds was 6. both bonds had the same coupon and maturity and terms were set at even yields. February 2000. 15 March 2010 M 105 . First. Between May 1999 and November 2003.5% Dec 2015 bond. as illustrated in Figure 62. approximately 75% is floored and the rest non-floored (see Figure 63).” Erik Thedéen. The addition of deflation floors on the principal payment to all inflation-linked bonds that would be issued in the future (April 1999).1bp on average. Thus. close to the central bank’s target and the average spread between the two bonds’ BEIs was a low 1. From December 2003 to February 2006.5% Dec 2028 bond). In other words. At present. Barclays Capital 3.

Lastly. new dealers can be appointed if one or more of the previous year’s dealers fail to perform.1bn. However. This system was more demanding for the newly authorized inflation-linked bond dealers. the redemption of bond 3101 in December 2008 largely contributed to this decline). 15 March 2010 106 . to further promote the inflation-linked bond market and especially increase investor awareness. requiring that they submit detailed business plans about how they intend to broaden the investor base. these reflect dealer performance in the previous year. and 5) online linker sales to the retail sector permitting individuals to purchase directly in the auction at the average yield level set after bidding. The outstanding balance of linkers more than doubled from December 2001 to its recent peak in September 2008. in real terms the size of the market was SEK192. 4) linker repo for dealers up to SEK200mn at rates 25bp below the central bank’s overnight rate.5% 01 Dec 2015 3.0% 01 Dec 2020 Sweden IL 0.5% 01 Dec 2028 4. moving from SEK94. The SNDO has a long-run target that 25% of its debt should be inflation-linked. with a target interval rather than a fixed amount announced ahead of auctions.0% 01 October 2001 Spot Inflation Jan 99 Jan 01 Jan 03 Jan 05 Jan 07 Jan 09 Jan 11 Source: Statistics Sweden. the SNDO implemented other facilities to further foster the market. 2) flexibility regarding the amount of linkers sold at each auction. consisting of a fixed payment of SEK1mn per dealer and a payment based on dealers’ share of the SNDO’s activities in inflation-linked bond issuance. the outstanding amount of inflation indexed-bonds has been on the decline since 2007 (for example. With the debt stock in line with the long-run target at the start of 2010. With government debt levels relatively stable and linkers’ share in the debt portfolio having been above the target level.0% 01 December 2008 Sweden IL 0. buybacks and switch auctions. once adjusting for expected future inflation accretion (the 2009-2011 Central Government Debt Management Proposed Guidelines indicate that “inflation-linked SEK debt should be steered in the long-term towards a percentage of 25 per cent of the central government debt”). During this third stage. the authorized dealer status is valid for one year and the following year. Barclays Capital 4. the SNDO introduced a new dealer and commission system from 2000. Furthermore. including: 1) occasional buybacks to provide liquidity when needed. the other side of the coin was that levels of compensation were higher – in any given year.14bn in January 2007 while in November 2009 it was SEK169. The third stage of the Swedish inflation-linked market has been one where investor acceptance of the product matured and the asset class secured a stable place in the country’s debt portfolio. 3) secondary auctions enabling dealers to buy an additional 20% of the allotted volume in the primary auction.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 62: Swedish breakevens and realized inflation 6 5 4 3 2 1 0 -1 -2 Jan 97 Sweden IL Sweden IL Sweden IL Sweden IL 1% 01 Apr 2012 3.0% 01 Apr 2014 Sweden IL 4. and the share of linkers in total government debt rose steadily (Figure 59).2bn to SEK227bn.

5% of linkers outstanding. the Premium Pension Authority and the Deposit Guarantee Board) also brought inflation-linked bonds closer to the spotlight. inflation-indexed debt featured pricipal floor. As mentioned above. Source: SNDO. which clearly helped the SNDO restructure the debt portfolio and stimulate domestic demand. after reviewing the track record of switches and auctions held in the previous few years. Figure 63: Share of floored linkers in total Swedish govt. though this was revised down to SEK9bn in the funding update published in March 2010. %) 5 4 3 2 1 0 -1 -2 99 00 01 02 03 04 05 06 07 08 09 10 At the end of 2009. liquidity measures in recent years have been less instrumental that those implemented at the end of the 1990s. bp) Sweden CPI Y/Y (RHS. the local insurance industry has been a key player. In June 2002. AP2 and AP3) reveal that between 2004 and 2007 they owned. 15 March 2010 107 . 5 0 -5 -10 -15 Note: BEI spread data are monthly averages of non-seasonally adjusted BEI spread. inflation-linked bond market 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Note: No inflation uplift included. 6 “Central Government Borrowing: Forecast and Analysis. The part-privatization of the national pension system and the conversion of cash balances to government bonds at three state entities in 2002 (the Swedish Nuclear Waste Fund.0% Apr '14 & Sweden IL 3. as illustrated in Figure 66 since 2004 secondary market turnover of linkers has been above 4%. nearly 75% of Swedish govt. boosting its share of total government linkers outstanding from around 35-40% in the early 2000s to above 50% in 2009 (Figure 66). Furthermore. . For example. the explicit indexation of the national pension scheme to wage inflation led to linker purchases by several of Sweden’s national pension funds. at times reaching nearly 10% of total government bond turnover and in addition the repo market has been functioning well. 18 June 2002.” SNDO.5% Dec '15 (LHS. Barclays Capital Figure 64: BEI spread of floored bond vs non-floored bond of similar maturity and spot inflation 35 30 25 20 15 10 BEI Spread Between Sweden IL 0. the financial statements of the first three national funds (so-called AP1. some 14. the SNDO has facilitated the purchase of this product for retail investors as well (ownership around 5%). with the rest of 25% being non-floored . on average. On the demand side. Source: Statistics Sweden. Barclays Capital Although important in themselves. the SNDO concluded that “there is no strong need for further restructuring 6” given that most inflation-indexed issues had reached satisfactory volumes and liquidity.Barclays Capital | Global Inflation-Linked Products – A User’s Guide issuance was scheduled to be SEK10bn a year in 2010 and 2011. Indeed. .

including inflation uplift. 29] (CPI t −2 − CPI t −3 ) + CPI t −3 30 This convention affects daily valuations. Barclays Capital 2005 2006 2007 2008 2009 Linker turnover as % of total govt. but in nominal terms. using a threemonth lag between the inflation release and the first of the month. Inflation accrual is calculated in a very similar manner to the three-month lag model. 15 March 2010 108 . Non-government linkers and inflation swaps At the end of 2009. There has been very little new issuance in recent years. Hence the reference day. but all coupons are paid on the first of the month. as there is linear interpolation but assuming 30-day months. However.4bn of issuance by the European Investment Bank since 2006 notably larger than all other supply combined. a very small figure compared to government bonds issuance (7% of total government inflation-indexed bonds outstanding in notional terms). d of the month is: Minimum of [d − 1. The clean nominal price (ie. With the SNDO decreasing its issuance. bond turnover (LHS) Linker repo transactions as % of outstanding (RHS) Note: Turnover represents linker spot transactions as % of aggregate spot government bond turnover.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 65: Secondary market turnover of Swedish govt. with the SEK5. The settlement price is then rounded to the nearest krona. Unlike the Canadian model. the prices for Swedish linkers are not expressed in real terms. with the 2020 bond the only benchmark that does not. Interest accrues on a European 30/360 basis. Bonds issued since 1998 have deflation floors. it remains to be seen whether non-government supply will emerge to take its place as in markets such as Australia. bond holdings of Swedish insurance comps as percentage of market size 60% 55% 50% 60% 40% 20% 0% 2010 45% 40% 35% 30% 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: Statistics Sweden. inflation-linked bonds and repo transactions 12% 10% 8% 6% 4% 2% 0% 2004 120% 100% 80% Figure 66: Inflation-linked govt. SNDO. Barclays Capital Features of Swedish linkers Calculations for Swedish linkers are slightly more complex than for those markets which employ the standard Canadian model. Virtually all corporate issuance is maturity matched to government issues and holds the same bond conventions. Source: Riksbank. The quoted real yield on Swedish linkers is consistent with the Canadian model though.8bn) by market value of nongovernment bonds linked to Swedish CPI. day count conventions are different. after inflation accrual) of coupon bonds is rounded to three decimal places before adding on accrued interest. The settlement convention for Swedish linkers is T+3. there was SEK13bn (about USD1.

Figure 67: Swedish linkers – historical performance and risk 16% 14% 12% 10% 8% 6% 4% 2% 0% 1999 2001 2003 2005 2007 2009 Sweden IL Returns Sweden IL Ann Mthly Vol Figure 68: Return/risk versus nominals and equities 6 5 4 3 2 1 0 -1 -2 1999 Equity IL Nominal Bonds 2001 2003 2005 2007 2009 Source: Barclays Capital Source: Thomson Datastream. there is scope for the inflation swaps market to develop. particularly given the presence of supra-national supply.Barclays Capital | Global Inflation-Linked Products – A User’s Guide While conceptually the well-defined bond market means that there is plenty of scope for a liquid inflation swaps market to develop. However. with the 10y the only maturity in which they are not uncommon. the Swedish CPI swaps market still remains surprisingly underdeveloped. Barclays Capital 15 March 2010 109 . as the liquidity of the nominal swaps curve has improved in recent years. particularly at the long end where it was previously practically non-existent. Quotes remain infrequent across the curve.

when it suspended its TIB programme due to ongoing budget surpluses. helping to fill the gap left by the government.nsf/DetailsPage/6440. the composition of the basket and other features of the CPI are reviewed “from time to time to ensure that it continues to meet community needs”. 8 Australian Bureau of Statistics. but at less frequent intervals (currently the base is year 1989-90=100). Sources and Methods”. Meanwhile.abs. The Australian CPI is published and maintained by the Australian Bureau of Statistics (ABS) on a quarterly basis (three months ending March.liiceanu@barcap. 8 The ABS undertakes these reviews at approximately five-year intervals with the timing generally dependent on the availability of results from the Household Expenditure Survey. Weights were revised to reflect new expenditure patterns and the expanded population coverage. “Australian Consumer Price Index: Concepts. The eight cities are the six state capital cities (Sydney. 7 15 March 2010 110 . however. In the comprehensive review of CPI calculation methodology undertaken in 1998. CPI figures are compiled separately for each capital city and the overall CPI is derived by weighting price movements (or price relatives) between the base and current period by their shares of total household expenditure in the base period. the fifteenth series CPI was introduced in September 2005. 2005. Adelaide. The Australian inflation-linked government bond market is thus poised to revive as the Commonwealth government explicitly expressed its commitment for continued supply. June. The ABS estimates that the individual consumer population in these cities represents about 64% of Australian private households. The reference base period for the CPI is also updated.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Australia Stefan Liiceanu +81 3 4530 1554 stefan. These groups are further divided into 33 subgroups. As a result. a significant local inflation swap market has been gathering pace. the CPI figures are typically released within one month of the end of the quarter). Besides the government IL bond market. The composition of the CPI basket is based on the pattern of household expenditure observed in the weighting base period (the latest being 2003-04).gov. the CPI basket is divided into 11 major groups. Hobart) plus Darwin and Canberra. more commonly known as the Australian CPI 7. Brisbane. also conducted by the ABS. The Australian CPI Australian inflation-linked government bonds. are indexed to the Weighted Average of Eight Capital Cities: All-Groups Index. September and December. in recent years several state governments and infrastructure companies have issued inflation-indexed securities. Moreover. each representing a set of goods and services (Figure 69).au/AUSSTATS/abs@. issuance resumed. 9 For further details on the Australian CPI please refer to the Guide to the Consumer Price Index published by the ABS (www. and the size of the market expanded by a sharp 66% to AUD10bn. Non-government inflation-indexed bonds stood at AUD12. As is the case in other developed countries. and the majority of AUD-denominated inflation structures including swaps. the population covered was expanded from wage and salary earning households to include all metropolitan households. and the subgroups into a total of 90 expenditure classes. More recently. supported by large institutional participants such as asset managers and pension funds. In 2009. the ABS decided that the index would be modified from a measure of the change in living costs of employee households to a general measure of price inflation for the household sector.02005?OpenDocument). possibly across several maturity sectors. slightly larger that the size of the government IL market. Melbourne. with information about consumption trends of Australian households coming from the Household Expenditure Survey (HES).com The Australian government actively issued inflation-linked bonds (commonly called Treasury Indexed Bonds) from 1985 until 2003. with item weights being revised in line with 2003-04 HES expenditure patterns and “Financial & insurance Services” included in the basket 9.4bn at the end of 2009. Perth.

0 96 97 98 99 00 01 02 03 04 05 06 07 08 09 Australia IL 4.0% 20 Aug 2020 Australia IL 3% 20 Sep 2025 Source: Barclays Capital Being released on a quarterly basis rather than on a monthly basis as is the case in the US.61 Alcohol and Tobacco. 11.44 Housing. Australian breakeven inflation rates have been trading around a long-run average of 250bp (in Figure 70.8% y/y since 1990 (Figure 15 March 2010 111 . 1990 through 2009.53 Source: Australian Bureau of Statistics The Australian CPI is an important economic indicator not only for the bond market but also for the central bank. The graph shows the average q/q percentage change in the index over two periods – ie. 3.5 1.0 3. which introduced inflation targeting in 1993.5 4. over the medium term.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 69: Breakdown of the Australian CPI Financial & Insurance Services.0 2.73 Recreation. Figure 70: Long-run trends in Australian breakeven rates (%) 5.0 0. 9. UK.11 Health.5 0.70 Household Contents & Services. 9.0% 20 Aug 2010 Australia IL 4. 6.5 3. 19. seasonality of the Australian CPI is less relevant. 4.91 Food. The Reserve Bank of Australia’s monetary policy aims to achieve.0 4. as illustrated in Figure 72. a target for consumer price inflation of 2-3% as the bank judges that preserving the value of money is the “principal way in which monetary policy can help to form a sound basis for long-term growth in the economy. with the exception of a brief period around the 2008 global financial turmoil. 2000 through 2009 and. and the euro area.31 Transportation.31 Education.79 Clothing & Footwear.” The central bank’s commitment to containing inflation helps explain the fact that.55 Communication. 2. 3. Inflation has averaged 2. 13.0% 20 Aug 2015 Australia IL 4.0 1. That said the index exhibits seasonal patterns to a certain extent. 15. note that the rebound from the 1% to the middle of the range in effect over the past decade or so has also been swift). from a longer-run perspective.5 2.

8% 0. against a backdrop of limited understanding of inflation-linked securities among investors. The number of TIBs issued was not large (the inflation-indexed programme never 10 11 Commonwealth Debt Management Report. In other words. was not adjusted for inflation at redemption). increasing on average by about AUD640mn a year in 1993-2000. The size of linker supply picked up considerably. although the size was a modest AUD100mn. the domestic inflation market had become more sophisticated. helped by issuance by a number of state governments and growing demand for long-term linkers from the emerging superannuation or pension fund industry. four Treasury Indexed Bonds (TIBs) were issued by the Commonwealth government. Australian Financial Markets Association.3%). however. Annual turnover averaged a modest AUD12bn during 2001-06 versus AUD410.9%) and the UK (3. “2006 Australian Financial Markets Report.2% 0. Two of these issues were capital-indexed bonds (CIB). Prices tend to increase most in the second and third quarters of the year. 10 Between 1988 and 1993. the Treasury ceased issuing linkers. while the other two were rarer interest-indexed bonds (IIB). paying a fixed coupon plus an inflation accrual on the principal every period (the principal. The maturities of these first bonds were 10y and 20y. Barclays Capital 1990-2009 2000-2009 Q2 Q3 Q4 The government bond market Australia’s first index-linked bond was issued by the State Electricity Commission of Victoria in August 1983. Two years later in July 1985.” 15 March 2010 112 . TIB liquidity was generally low as the bonds were issued largely to buy-and-hold investors.6% 0.3bn for nominal Commonwealth Government Securities (CGS). As the Australian government’s fiscal situation improved sharply in 1988. Figure 71: The Australian CPI (y/y) 14% 12% 10% 8% 6% 4% 2% 0% -2% 81 83 85 87 89 91 93 95 97 99 01 03 05 07 09 Source: ABS Figure 72: Seasonality patterns – average q/q changes 1. only capital-indexed bonds were brought to the markets.9bn (Figure 73). with supply being “tailored to identify market demand”.0% Q1 Source: ABS. However. It even bought back some of the existing bonds as part of its debt reduction policy. Q4 exhibits the slowest increases in consumer prices.Barclays Capital | Global Inflation-Linked Products – A User’s Guide 71).2% 1. the government resumed its index-linked issuance programme in 1993. 1996. with Q3 recording the fastest pace of consumer price appreciation. from AUD1.6bn to AUD5. all with long maturities. However. levels comparable with those in the US (2. Under the new programme.0% 0.4% 0. TIB trading volume was less than 3% of that of regular government bonds despite the size of the total market being around 10% of the nominal market 11. In contrast.

however.096 2. Source: Australian Office of Financial Management (AOFM) TIB outstanding (LHS.0% Face value (AUD bn) 1. AUD bn) Ratio of linkers to nominal treasury bonds (RHS) 16% 14% 12% 10% 8% 6% 4% 2% 0% Figure 74: Australian government inflation-linked bonds Bond Australia IL 4.010 Note: Numbers in real terms. In an announcement regarding the issuance of Commonwealth government 15 March 2010 113 .5% of GDP in 1995-96 to 1. The government’s review concluded that an interest rate market completely dominated by banks and corporate paper would be vulnerable to economic shocks. which sought to determine whether the CGS market was a viable going concern given the sharp fall in the Commonwealth government’s financing requirement over the previous few years and the abundance of cash available from the sale of government assets (the Australian government’s net debt had fallen from 18. the very favourable fiscal conditions implied no need for active issuance of government securities. Figure 73: Australian Treasury indexed bonds outstanding 12 10 8 6 4 2 0 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 Note: Mid-year levels except 2009. Rather disappointingly. thus posing significant threats to financial stability and the accessibility of refinancing capital for corporates. where Dec levels are shown. In 2009.0% 4. the inflation-indexed government bond programme was revived after a six-year hiatus.773 4. On the other hand.0% 2015 Australia IL 4.0% 2020 Australia IL 3. there was no room for a continuation of the Treasury Indexed Bond Programme.00% 2025 Issue date 9 Feb 93 17 May 94 14 Oct 96 8 Oct 2009 Redemption date 20 Aug 10 20 Aug 15 20 Aug 20 20 Sep 25 Coupon 4. Because of this.Barclays Capital | Global Inflation-Linked Products – A User’s Guide had more than five bonds). and was structured in such a way that it supported the 3y and 10y Treasury bond futures contracts. there was concern about banks’ disproportionately large role within the financial markets and hence it was argued that the CGS market should be maintained. On purely economic grounds.452 2. the decision was taken to support government debt liquidity.0% 2010 Australia IL 4. however.0% 3.3% of GDP by 2004-05). the Australian Treasury announced the suspension of the Treasury Indexed Bond programme with the publication of its budget in 2003. This announcement followed a one-year period of analysis and consultation with more than 90 domestic and foreign market participants. Source: Australian Office of Financial Management (AOFM) Meanwhile.0% 4. and additional supply came in the form of re-opening seasoned bonds on average 5-6 times a year.

First. linker repo transactions as a percentage of outstanding government ILBs jumped from around 20% to an average of 80%. both income and capital generated by TIBs are indexed to inflation. in mid-November 2009 the AOFM also issued via an AUD300mn tap of the 4% Aug 2020 bond). The actual amount sold was four times larger. Following consultation with various market participants. an important addition to Australia’s new linker programme is the expansion of the AOFM’s securities repo facility to include TIBs (both seasoned bonds and the new 2025 bond). since Treasury Indexed Bonds (TIB) would serve as both a pricing benchmark and a risk management tool… indexed financing can be attractive for those infrastructure projects whose revenues are linked to inflation… in addition. coming at AUD4bn. Meanwhile.5-2bn until the middle of 2010). to extend the Commonwealth inflation indexed yield curve beyond 2025 at an appropriate time”. had been steering debate on asset management in the direction of safer investment guidelines. Such a move.75% 15 May 2021 issue). the Australian Office of Financial Management (AOFM) stated that resumption of linker supply “could assist in the debt financing of long-term infrastructure. or nearly 5% of GDP in 2009). leading to Australia’s first budget deficit in seven years and the largest one on record (AUD57. For example. however. As history has shown.Barclays Capital | Global Inflation-Linked Products – A User’s Guide bonds made public in May 2009. while tax revenue plunged. both principal and coupon payments have a deflation guarantee. The AOFM expressed its commitment for additional issuance that would contribute to the benchmark status of the new linker (TIB issuance is officially expected to be AUD1. Australian linkers are similar to old-style UK index-linked bonds in that the next coupon amount is always known on or before the current coupon payment date. “the amount of inflation indexation in any given coupon period is equal to the average percentage 15 March 2010 114 . As a result. by 66% (besides the auction of the new 2025 issue. the government’s expenditure base had increased sharply. One can think of at least two more reasons for the resumption of inflation-indexed bonds not explicitly stated in the AOFM note. Second. As shown in Figure 74. redeeming on 20 September 2025. Technical features of Australian inflation-linked bonds As is the case with other international linker markets. the longest-dated government inflation-linked bond in the Australian market. the maturity of the new bond is 16y. As a result the Australian government inflation-linked bond market expanded sharply. the AOFM decided to issue the new linker at the beginning of October 2009 via syndication. to boost the efficiency of its domestic government IL market the Swedish debt management office decided in the early 2000s to allow dealers to borrow bonds up to a SEK200mn ceiling at rates 25bp below the central bank’s overnight interest rate. with an announced supply size of “at least” AUD1bn. this step should clearly contribute to the improvement in secondary-market liquidity. double-digit negative returns on Australian households’ mandatory superannuation (pension) funds owing to financial market volatility. would also require the launch of nominal bonds of similar maturity to enable the observation of BEIs (currently the longest-dated nominal bond is the 5. indexed instruments have advantages for investors with inflation-linked liabilities”. and coupons are payable on a quarterly basis. subject to investor demand. According to the prospectus for Treasury Index Bonds published in 1995. The AOFM’s notice added that “it is also planned. combined with proposals to raise the age at which individuals can access these funds. being identical to the three existing capital-indexed bonds: indexation is relative to the Weighted Average of Eight Capital Cities – All Groups Index (the Australian CPI). similar to other countries affected by the 2007-09 global financial crisis. and linkers obviously fit the description.6bn. The structure of the new IL security is unchanged.

for example if the next interest payment is in November. while conventional bonds pay semi-annual coupons. these securities contain an embedded put at maturity that protects against deflation over the life of the bond. inflation-indexed pay coupons on a quarterly basis.0116. Note that Kt and Kt-1 are rounded to two decimal places.Barclays Capital | Global Inflation-Linked Products – A User’s Guide change in the Consumer Price Index over two quarters ending in the quarter. “f” is the number of days from the date of settlement to the next interest payment date. − (1+p/100)-f/d 100 V = 1/(1+i). Canadian.65%. Kt can also be expressed as Kt = Kt-1*(1+p/100). This means that the bonds have a six-month indexation lag compared to eight months in the UK. with the highest power being “n” (the number of full quarters between the next interest payment date and the date of maturity).n] j − “Kt” is the nominal value of the principal at the next interest payment date (whether or not there is an interest payment due). “x” is a valued of either 0 or 1 depending on whether there is an interest payment at the next interest payment date. however. Kt-1 is equal to AUD100. “g” is the fixed quarterly interest rate payable (equal to the annual fixed rate divided by 4). or 0. capital-indexed bonds protect both coupon and principal from deflation over the life of the bond. then “p” is based on the average movement in the CPI over the two quarters ended in the June quarter preceding. the bond is trading ex-dividend). P = V f/d [g(x + an) + 100Vn] Kt Where. For example. where CPIt is the CPI for the second quarter of the relevant two-quarter period. The interest on Australian linkers is accrued on an actual/actual basis. then “i” is equal to 4. an = V + V2 + …+Vn = − − − − − ΣV . The settlement price for AUD100 face value of Australian government inflation-linked bonds is provided by the following formula. “x” is 1 if there is an interest payment and 0 if there is no interest payment (ie. Furthermore. − “p” is the average percentage change in the CPI over two quarters ending in the quarter which is two quarters prior to that in which the next interest payment falls. which is two quarters prior to that in which the next interest payment falls”. as is the case with other developed linker markets. Mathematically expressed. Interestingly. “d” is the number of days in the quarter ending on the next interest payment date. where Kt-1 is the cash value at the previous payment date. Mathematically. Unlike other markets that offer an inflation floor. and CPIt-2 is the CPI for the quarter immediately 115 15 March 2010 .65/400. If there has been no previous payment date. The calculation of interest and principal payments for Australian index-linked bonds is significantly different than US. “p” is (100/2)*[(CPIt/CPIt-2)-1]. where j∈[1. with “i” being the annual percentage real yield (quoted real yield) divided by 400. Euro and Swedish bonds. Australian linkers trade ex-dividend for seven days prior to the payment date. “an” is the sum of the power series or V. if the annual yield is 4. and the bonds are quoted on a yield basis.

income from Treasury Indexed Bonds derived through interest or discount or through capital accruals throughout the life of the bonds was taxed according to the laws of the Commonwealth and states. Taxation Up until 2009. 0. where “g” and “Kt” are the variables defined in the calculation of settlement prices above. The Australian market has a large number of smaller inflation-indexed issuers including state governments. Most of the initial offerings were capital-indexed bond type. Non-government linkers and Australian inflation derivatives Although the Commonwealth Government has been the dominant issuer of Australian inflation-linked bonds. being perceived as more appealing by overseas investors. also 15 March 2010 116 . Interest payments for Australian linkers are calculated as g*Kt/100. as well as private issuers. one third of which was issued by State governments. quasi-government authorities and government-regulated utilities companies. Hints of such pick-up in inflation supply could be seen in the entity’s 2008 Annual Report.50 being rounded up). More recently. The Queensland Treasury Corporation. its securities represent only about half of the market. Illustrating robust investor demand for inflationindexed bonds in August 2009. Moreover. local Australian governments also have a fairly long experience with inflation-linked securities as many of them began tapping the markets in the late 1980s.75% November 2020 capital indexed bond at an outright real yield of 3. or IAB) were brought to the markets. The initial issue size was AUD603mn. Subsequent interest and/or principal payments will in such cases be reduced by the difference between the fixed interest payment that was paid in the period and the payment that would have been made under the above formula except for this provision. the government announced it will legislate to make Commonwealth Government Securities eligible for exemption from non-resident interest withholding tax. However. no interest payment is based on a nominal value of less than AUD100.Barclays Capital | Global Inflation-Linked Products – A User’s Guide prior to the relevant two quarter period. This follows the removal of the interest withholding tax on publicly issued corporate bonds some 10 years earlier and on state government securities in 2008. One such example is an IAB issued by the State Government of Victoria in 1993. This was followed by a total of AUD460mn worth of linker supply (maturities of 18y and 28y) at the end of 2007 by the Treasury of New South Wales. non-government Australian CPI-linked bonds outstanding amounted to about AUD12. a large amount at that time. on 21 August 2009. The Treasurer of the Commonwealth of Australia argued that such a move would closer align Australia’s tax treatment of securities to those systems in other developed markets and that this would also lead to Australian bonds. commenting that “the potential to expand the use of CPI linked debt by a broader range of clients. Just like the Commonwealth State. “p” is also known as the “Australia CPI factor average change (ACIF)” and is regularly calculated by the Reserve Bank of Australia (RBA). following the exact format of inflation-linked CGS. Interest payments are rounded to the nearest cent (ie. As of the end of 2009. If the nominal value of the principal falls to below AUD100. then the interest payment would be based on a nominal value of AUD100. These figures are also available on Bloomberg as ACIF Index. the Treasury Corporation of Queensland issued AUD268mn of 15y capitalindexed bonds in mid-2006. the third largest issuer of inflation-linked bonds in Australia.4bn. the Treasury of New South Wales issued AUD800mn of a new 3. particularly for PTEs (public trading enterprises) with revenue streams linked to the CPI. TIBs included.75% (a spread of nearly 60bp to the 2020 Commonwealth government TIB) on a book-build basis. is currently being explored… this may lead to further client demand for CPI linked funding”. but subsequently annuity-style indexed securities (called Index Annuity Bonds.

This type of issuance has been sporadic but has featured hefty deal sizes. As the projects feature amortising cash flows. However. Therefore. The pick-up in nongovernment inflation-linked bond issuance has thus clearly fostered the development of an AUD inflation swap market. Thus.” a public-private partnership). When demand from end-investors for such bonds is not strong enough.Barclays Capital | Global Inflation-Linked Products – A User’s Guide announced in its borrowing programme for 2009-10 that it would fund half of the estimated deficit of AUD22. the nominal bonds were issued along with paying CPI swaps to the tune of AUD1bn. while the Commonwealth Government ceased to provide liquidity to the Australian inflation markets during 2003-2009. paying cash flows consisting of both principal and interest.75y. Barclays Capital 15 March 2010 117 . with payments adjusted to the rate of inflation.75y and 14. the consortium has incentives to issue inflation-linked bonds with maturities similar to that of the lease contract in order to achieve a better asset-liability profile. in many cases the structure of these bonds is Credit Foncier type. Figure 75: Australian linkers – historical performance and risk 14% 12% 10% 8% Australia IL Returns Australia IL Ann Mthly Vol Figure 76: Return/risk linkers versus nominals and equities 8 6 4 2 Equity IL Nominal Bonds 6% 0 4% 2% 0% 1999 2001 2003 2005 2007 2009 Source: Barclays Capital -2 -4 1999 2001 2003 2005 2007 2009 Source: Thomson Datastream. which at the end of 2006 financed the renewal of the Sydney rail network (a fleet of 78 trains) partly via AUD300mn worth of 29y CPI-linked annuity bonds and partly via AUD1. the consortium typically issues only one part of the debt linked to inflation and the rest in nominal form with an inflation outlay. A good example is Reliance Rail. with activity in the swap market spiking around the time of these large infrastructure projects. the government enters into a long-term lease contract with the consortium. Australian governments responsible for building infrastructure projects have been turning to consortiums typically made of construction companies and financial institutions (a recent example is the consortium charged with the construction of nine public schools under the so-called “New Schools Privately Financed Projects.6bn floating-rate bullet bonds with maturities between 9. trading has been more or less sporadic. Once the project is finalised. ie.5bn via domestic bonds including the 2030 capital indexed bond. Inflationindexed security supply related to infrastructure projects has also surged in recent years. local government were quick to pick up the slack. paying inflation via the swap market.

January 2010 Housing Comuni Goods cation 4% 6% Clothing 7% Education 7% Personal Services 10% Personal Care 11% Source: IBGE. FGV.5% +/.2%). Figure 77 plots the IPCA rate (official inflation target rate. Barclays Capital Food and Beverages 23% Transport 19% Dwellings 13% Jan 01 Jan 03 Jan 05 IPCA (% Y/Y) Lower bound Jan 07 Jan 09 Jan 11 Inflation target Upper bound Source: IBGE. It enjoys a significant degree of “defacto” (not “de-jure”) independence. Figure 77: The Brazilian inflation targeting regime 20 18 16 14 12 10 8 6 4 2 0 Jan 99 Forecast Figure 78: IPCA weights (%). with its governor being nominated by the Brazilian President and the board of directors sanctioned by the Senate.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Brazil Marcelo Salomon +55 (0) 11 5509 3295 marcelo. it was during its tenth anniversary. There are no fixed mandates for the board or the governors. For the first time the Brazilian Central Bank (BCB) was able to implement countercyclical monetary policy. Barclays Capital 15 March 2010 118 . The BCB’s only mandate is to implement monetary policy in line with these targets. UK and France. Brazil was the fourth-largest issuer after the US. that we believe the regime finally reaped reward from its investment. However. floating not only in the upper but also in the lower range of the bands.com Guilherme Loureiro +55 11 5509 3295 guilherme. reflecting both domestic and external shocks. helping to minimise the downturn of economic activity at a time when capital flows were drying up dramatically and the BRL depreciated substantially.loureiro@barcap. Each July the National Monetary Council (chosen by the Finance and Planning Ministers along with the BCB President) sets the two-year ahead mid-point of the target along with the tolerable range of fluctuation (currently at 4.salomon@barcap. Having been the dominant form of local debt through hyperinflation periods. observed and forecast) along with the bands and mid-point targets of the Brazilian IT regime since it was implemented. linkers fell to a small share of national debt until 2003. After a volatile initial few years. there has been a major move towards increased linker issuance to replace floating and foreign currency debt.5% mid-point of the target. We believe this action signals the way for smoother business cycles and that it is a new step in the real interest rate convergence story in Brazil.com Brazil has been issuing inflation-linked bonds since 1964. with the uplifted notional of IPCAlinked bonds equivalent to US$200bn. Inflation targeting The Brazilian inflation targeting (IT) regime has been consistently put under pressure since it was implemented in 1999. At end-2009. inflation converged towards the current 4. when the Brazilian economy weathered the global financial meltdown surprisingly well. Subsequently.

Figure 79: IPCA and IGP-M consumer price component (% m/m) 3. with food/beverages. reflecting the changes in relative prices. we believe it will continue to lead the pack by introducing new mechanisms to contain exaggerated credit expansion. having the largest share. and the m/m change for February represents the full-month February average compared with the January full-month average. FGV. Barclays Capital IGP-M Wholesale Component Source: IBGE Source: FGV. Geographical coverage comprises 11 metropolitan areas and these indices are aggregated to form the national one.5% 2. As standard in CPI index calculations. In addition. it also introduced another layer of uncertainty: how will monetary authorities deal with asset/credit bubbles? In our view the Brazilian IT will be enhanced by the use of non-interest rate tightening mechanisms to rein in such bubbles. reflects average prices during the month of February. In other words.0% 2.0% -0. IBGE-reported weights move slightly each month. transportation and housing.0% 0.0% 1. with São Paulo having by far the biggest weight of 33%. the February index. calculated by the national statistics agency (IBGE).Barclays Capital | Global Inflation-Linked Products – A User’s Guide While the crisis crowned the Brazilian IT regime as a very successful framework of monetary policymaking. further prudential and regulatory measures should start to emerge as part of the BCB’s toolkit. Indexation Brazil developed several inflation indices during its high inflation period. Figure 78 shows the weights in January 2010. The most important are IPCA (BZPIIPCA <Index> on Bloomberg) and IGPM (IBREIGPM <Index> on Bloomberg). released in March. is the official national consumer price index (and the measure of inflation targeted by the central bank).5% -1.5% 1.0% Jan 02 Jan 04 Jan 06 Jan 08 IPCA Figure 80: IGP-M wholesale and consumer components (% y/y) 50% 40% 30% 20% 10% 0% -10% Jan 02 Jan 04 Jan 06 Jan 08 IGP-M Consumer Component IGP-M Consumer Component Source: IBGE.5% 0. Barclays Capital 15 March 2010 119 . The importance of the food and transportation components is exacerbated by their high level of volatility. The IPCA (December 1993 = 100). while the BCB is already a benchmark for regulatory and prudential regulation. IBGE publishes the IPCA around the 10th day of each calendar month covering the period of the previous calendar month. Current weights are based on the consumption survey taken in 2002 and 2003 and reflect the consumption patterns in households with incomes of 1-40x the minimum wage.5% 3. weights implicit in the index level are fixed. Specifically. However. as they are affected by the swings in international food and energy prices.

0 8. and i A is the inflation projected by ANBIMA. Government bonds: NTN-Bs and NTN-Cs The Treasury used to issue two inflation-linked securities: NTN-Bs (BNTNB <Govt> on Bloomberg) and NTN-Cs (BNTNC <Govt> on Bloomberg).so that we have: t I t' = I t (1+ i A ) n/ N where n is the number of Brazilian business days between the evaluation date and the last day of the previous IPCA period coverage.0 5.0 6. N is the number of Brazilian business days between the last day of the previous IPCA period coverage and the last day of the next IPCA period coverage.RHS Source: National Treasury.LHS Average Duration (yrs) . affecting the value of the linkers. ANBIMA’s inflation projection might change according to other data releases. as measured by the IGP-M. The consumer price component of the IGP-M tends to behave similarly to the IPCA but the wholesale price component is considerably more volatile. The IGP-M is also published monthly. fluctuates more widely than IPCA inflation. Both inflation indices It are updated only once a month and evolve as a step function. inflation. Barclays Capital 6.0 2.0 2006 2007 2008 Average Yi eld (NTN-B May/15) . published by the private Getulio Vargas Foundation. It consists of three components: a measure of wholesale prices (60% of the total). Special attention should be paid to the dates when there are releases of other inflation indices. measures a broader set of prices. The market convention is to use the official ANBIMA’s (Brazilian Association of Financial and Capital Market Companies) inflation forecast and pro-rata it to define an index value. I . a measure of consumer prices (30% of the total).0 2009 Daily Average Trade (USDmn) Source: Barclays Capital 15 March 2010 120 .0 0. since they are generally correlated with IPCA or IGPM or both.0 4.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The IGP-M Index (August 1994 = 100). both materials and labour (10% of the total). To adjust the current price of the bonds correctly (see next section) one needs to account for the accrued inflation from the last date the index was updated to the settlement date of the bond (generally T+1). which can affect tradable goods prices significantly (Figure 80). but around the 30th day of each month. The former have their principal Figure 81: NTN-B market information by maturity 450 400 350 300 250 200 150 100 50 0 ago/10 nov/11 ago/12 ago/14 ago/20 ago/24 ago/30 ago/40 mai/11 mai/13 mai/15 mai/17 mai/35 mai/45 ago/50 Figure 82: IPCA-linked debt duration and NTN-B (May15) yield 10. and a measure of construction costs. Instead of calendar months.5 4.5 4.0 5. BRL changes have a faster and higher pass-though into IGPM than into IPCA. As a result. partly because of the prevalence of raw foods in the index and partly because of exchange rate movements. it covers a period from the 21st day of a given month through to the 20th day of the following month.

NTNs are quoted on a yield basis using the Brazilian Business/252 day count convention and annual compounding. led the Treasury to shift towards IPCA-linkers. c' = 1 + c 2 − 1 is the effective coupon. I t' is the current index level. In the past few years. due to the local convention. its main aim was to minimise the debt’s vulnerability. to establish new benchmarks on the longer ends of the IPCA yield curve and to broaden the alternatives of inflation liked bonds. Y is the quoted yield and Pt is the current price of the bond. The strategy was mainly to increase the maturity of the public debt. when the National Treasury started publishing the guidelines for annual debt through the Annual Borrowing Plan. ti refers to the period in years from the settlement date to each coupon payment date and tn is the maturity of the bond in years. with one auction a month featuring all on-the-run issues and the second comprising issues up to the 10y benchmark. the SELIC and FX-linked share of debt dropped significantly to 40% from nearly 80%. the effective coupon c’ (paid every 6m). with the curve extending out to 2050. NTN-Bs. ( ) Government debt structure The Federal Government has been trying to improve the quality of public debt since 2001.Barclays Capital | Global Inflation-Linked Products – A User’s Guide indexed to IPCA. while the 15 March 2010 121 . the National Treasury issued three new NTN-B bonds maturing in August 2030. Broadly speaking. reduce the long-term financing costs and to ensure the maintenance of prudent risk levels. They usually have semi-annual payments of fixed-rate coupons on the indexed principal. demand for IPCA-linked securities has been increasing. The results of this initiative became more apparent by 2003. the market for NTN-Cs was substantially larger than that of the NTN-Bs. Issuance in recent years has been twice monthly. trade currently (1st bimester of 2010) on average US$1. we stress the aims of the most important measures:1) to gradually increase the proportion of fixed and inflation-linked share of the debt while reducing that of the FX and SELIC shares. I is 0 the inflation index at base date. Until 2005. but with principal indexed to IGP-M. Furthermore. In February 2010. All IPCA linkers have 6% real coupons. Liquidity is often poor for NTN-Cs. The yield to price formula is given by: Pt = 1000 1 ⎤ I 't ⎡ n c' ⎢∑ (1 + Y )ti + (1 + Y )tn ⎥ I 0 ⎣ i =1 ⎦ Where t corresponds to the settlement date. which also have long-term IGPM liabilities acquired in the past. Therefore. as a sizeable portion of the outstanding bonds are held by buy-and-hold pension funds. newly-issued bonds start up with a large inflation adjustment and a nominal invoice payment necessary to acquire the bond that is materially above par. from a supply perspective. NTN-Bs account for about 90% of the total outstanding amount of inflation-linked securities (BRL410bn). From that year to December 2009. where c is the annual coupon. NTN-Cs are similar to NTN-Bs. given IPCA’s central role within the inflation-targeting regime. though there have been some zero coupon bonds issued with indexed principal. ( ) Linkers have their principal indexed from a base date that does not coincide with the issuance date. However. 2040 and 2050. 1 with the Brazilian Bus/252 convention used. From a more detailed perspective. 1 is given by: c ' = 1 + c 2 − 1. 2) to lengthen the debt’s maturity and 3) to implement both strategies reducing the average cost of debt. the combination of the 1999 devaluation and the high FX pass-through of the IGPM. However. on the other hand. as the end of February 2010.1bn on a daily basis (Figure 82). The base date is set at 15 July 2000 to all NTN-Bs. NTN-Cs issuance has been extinguished and.

and intermediate rates for holding periods in between). while the average (12-month) implicit cost of public debt declined from 14. Figure 83: Inflation-linked bonds (% federal government debt) and target (annual borrowing plan) 35% 30% 25% 20% 15% 10% 5% 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2003 2004 2005 2006 2007 2008 2009 2010F Inflation-linked bonds (%Total Debt) Upper Bound Lower Bound Source: National Treasury. the stop-and-go cycle of foreign capital forced the depreciation of the BRL. the Government’s goal is to consolidate the share of fixed-income bonds and inflation-linked bonds at 55-65% of debt. the average duration of the public debt increased to 3. Barclays Capital 15 March 2010 . Focusing on the evolution of the inflation-linked component of the debt.3 years from 5. in absolute terms we anticipate further issuances of inflation-linked bonds as the gross public indebtedness should continue to expand in the next few years. From 200509.5%. The tax rate ranges from 15-22. we note that its share grew to nearly 27% of debt in 2009 from 10% in 2003 (Figure 83 and Figure 84).7% from 8. The forecast of the ABP 2010 is that the internal public debt expands to BRL1600-1730bn in 2010 (from BRL1497bn in 2009) which is compatible with new net issuances of NTN-B reaching about BRL100bn in 2010. From 200509.5% if held for less than 180 days.8 years). The debt management gains affected more than just public finances.4%. inflation-linked debt should remain broadly unchanged.9% (Figure 82). leading the BCB to start a tightening cycle to rein in inflation (both observed and expected) and stem the process of currency depreciation. Barclays Capital Figure 84: Debt composition 100% 80% 60% 40% 20% 0% 2009 122 0% Fixed-Rate FX-linked Inflation-Linked Other Selic-Linked Source: National Treasury.2% to 9. the IPCA-linked bond (NTN-B maturing in May 2015) yield declined to 6. Along with large international reserves this helped to pave the way for counter-cyclical monetary policymaking during the recent financial global meltdown.Barclays Capital | Global Inflation-Linked Products – A User’s Guide fixed-income and inflation-linked component moved to 60% from around 20%. In the same period. A larger share of fixed and inflation-linked debt (smaller FX and Selic) broke the cycle where a weaker BRL and higher Selic rate would lift the debt-to-GDP ratio.8 (with the total debt maturing in more than 5 years rising to 41% from 34%). Taxation Local residents in Brazil pay a withholding tax of 15% or more on the income from bonds. Reducing the FX/Selic component of the debt enhanced the perception of fiscal solvency in periods of stress. 15% for periods above 720 days. its average duration increased to 6. Although. In previous crises. with the precise bracket depending on the holding period (22. in relative terms. According to the 2010 Annual Borrowing Plan (ABP 2010).5 years (from 2. worsening fundamentals and further weakening BRL.

gov.receita.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Since early 2006. Non-resident investors domiciled in tax havens are taxed at the same rates as local investors.0% 3. new instruments continue to emerge in the private fixed income market. increase if USD/BRL eventually starts to trade below 1. Thus. with limited interest from foreign investors. However. It represented 3. Risks. qualifying foreign investors are exempt from withholding taxes. The tax is now levied on all inflows (equity and fixed income) except foreign direct investment and it is collected only on the way in. One of them is the bank financial note (Letra Financeira). with all the political cost associated with that. This largely reflects both limited liquidity and the asymmetric tax treatment they receive. Foreign investors do not enjoy the tax withholding exemption when purchasing private debt instruments (non-government inflation linkers are included here). It is important to point out that a zero tax rate is different from a non-existent tax. There is no discrimination between long-term and short-term flows. Therefore. This measure was aimed at containing BRL’s appreciation as the government seemed concerned about an eventual de-industrialisation process. The government is by far the largest player. Despite being relatively small in size and suffering from unequal taxation. the use of inflation-linked debentures is stagnant over time (Figure 86). private issuance is mostly concentrated in debentures (Figure 85).2% Debentures 74% Source: ANBIMA. however. Barclays Capital 15 March 2010 123 .br for a list of these). When we look at the other 6%. which was just approved by the government and which should be Figure 85: Non-government linkers breakdown Others 11% CD 15% Figure 86: Inflation-linked debentures (% total debentures) 4.2% 4.6% 3.8% 3.0% Dec 05 Dec 06 Dec 07 Dec 08 Nov 09 Source: ANBIMA.7% of the total stock of debentures outstanding by the end of 2009. Any new tax in Brazil has to be approved by Congress and the Senate.4% 3. Other inflation-linked assets and derivatives The market for corporate linkers in Brazil is still underdeveloped. Barclays Capital 3. the withholding income tax rate applicable to sovereign local bonds has been set to zero for foreign investors who are not in tax havens (see www. the government raised it from 0% to 2%. which they benefit from when buying government debt. In October 2009. According to ANBIMA (Brazilian Association of Financial and Capital Market Companies) the inflation linked market (private and government) represent 17. accounting for 94% of market share. At current levels of BRL we do not expect new FX measures.fazenda.70. it is much simpler for the Executive to raise a tax rate when the tax already exists. The IOF tax on foreign capital flows is a good example of this.5% of the total fixed income market.

The cost of the pass-through service charge by the dealers has declined to a few basis points. On the derivatives side. Offshore total return swaps are common investment vehicles for foreign investors looking to circumvent the burden of opening and managing local investment accounts in Brazil. These swaps can be registered in CETIP (OTC) or BM&F (the main local exchange) and are usually traded as zero coupon. Dealers with an onshore presence buy the bonds on their books and pass the total return to investors offshore though an ISDA swap. Longer tenors are available. while the reimplementation of the IOF tax has increased interest from international investors in total return swaps. leaving their books with basis risk and cash flow mismatches. but with very limited liquidity and concentrated on short tenors (up to 2y). The funding leg of the swap may be the local overnight rate CDI or Libor.Barclays Capital | Global Inflation-Linked Products – A User’s Guide formally regulated by the monetary council in early 2010. in case the client wants to keep the FX exposure along with the local interest rate exposure. potentially representing a new investment alternative for global and domestic investors. Regular offshore swaps against CDI or Libor may also be found on a limited basis and with fairly wide bid/ask spreads. Once its details are announced and the first issuances approved. onshore inflation-linked swaps over IPCA and IGPM are available. as is also the case for the IPCA futures contract (WLA <Index> on Bloomberg). It is a new longer-term debenturelike instrument that can be issued by banks. the few dealers quoting the offshore swap need duration hedges using the local linkers. The leg of the swap linked to inflation pays the changes in the inflation index plus a real rate coupon that is quoted at the onset of the trade. but can also be coupon bearing. albeit with very large bid/ask spreads or coinciding with the maturity dates of bonds. Since the local inflation swap market is not well developed. The other leg of the swap is usually the accumulated overnight rate (CDI). Barclays Capital 15 March 2010 124 . we will be able to gauge the significance of this new instrument. but it can also be a fixed Libor plus spread or other formats. Figure 87: Brazilian linkers – historical performance and risk 25% Brazil IL Returns Brazil IL Ann Mthly Vol 20% Figure 88: Return/risk – linkers versus nominals and equities 8 7 6 5 Equity IL Nominal Bonds 15% 4 3 2 1 10% 5% 0 -1 0% 2004 Source: Barclays Capital 2005 2006 2007 2008 2009 -2 2004 2005 2006 2007 2008 2009 Source: Thomson Datastream.

c. Price information is collected in 46 cities and metropolitan areas. as higher inflation led to accelerated amortisation of loans in real terms. the UDI Index changes by the Figure 89: CPI weights Other 6. Banxico also reports a breakdown of the headline index into core (69. it is a function of bi-weekly inflation and has been published since 4 April.com Mexico has been issuing inflation-linked bonds for over 30 years. By day 10 of each month. The UDI is unusual in that it fixes off a CPI Index that is published twice monthly rather than monthly. despite no restrictions or withholding tax for international investors.9% Food 22. with foreigners holding less than 3% as at the end of 2009. with around 15% of outstanding government debt denominated in the inflation-linked UDI Index at the end of 2009 (UDI98bn.6% Furniture 4.zuniga@barcap.6% Housing 26. The CPI considers fixed weights and is based on the consumption basket of 2000.1%. Barclays Capital 15 March 2010 125 . which is due on the 10th and 25th day of each month (or previous day if not a business day) for the previous half-month period. The release occurs one day before publication in the official gazette.4% 2001 2003 2005 2007 CPI % y/y 2009 Inflation target Source: INEGI. The UDI Index (MUDI <Index> on Bloomberg) is released twice a month by Banxico. it publishes values for the period between day 26 of the previous month and day 10 of next month.9% Clothes 5.com Jimena Zuniga +1 (212) 412 5361 jimena.org. Pension funds are the largest holders of government UDIBonos.7% Figure 90: Inflation dynamics 20 18 16 14 12 10 8 6 4 2 0 1999 Educ. The central bank (Banxico) adopted the Inflation Targeting (IT) regime in 1999 with an initial 13% aim for that year. The CPI (‘Indice Nacional de Precios al Consumidor’ – INPC) is calculated bi-weekly by Banxico. In each period.4% of the total) items.mx) in accordance with an annual calendar. particularly those of perishable food items. 1995. Barclays Capital Source: Central bank.banxico.melzi@barcap. Banxico publishes index values for the period between days 11 and 25 of the previous calendar month. followed by the food component. but with the objective of bringing inflation down to 3% by 2003 and beyond. Indexation Mexico’s current inflation-linked bond market started in May 1996 as a result of the ‘Tequila crisis’ in late-1994. This created an incentive to issue credit in UDI (‘Unidades de Inversion’) to preserve their real value. 11. non-core prices inject considerable volatility into the index. which releases the data on its web page (www.6% of the total) and non-core (30. On day 25.USD16bn equivalent).5% Transport 13. The housing component has the heaviest weight in the index.4% Health 8. While core inflation is fairly stable in time.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Mexico Roberto Melzi +52 55 5241 3260 roberto. The latter has remained the target with a tolerance range of +/.

who are not subject to withholding taxes or any other local Mexican taxes on purchasing government securities. UDIBonos are quoted in real yields. Cetes.50 Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual act/360 act/360 act/360 act/360 act/360 act/360 act/360 act/360 act/360 act/360 act/360 TOTAL 7. other than the halfmonthly inflation periods and short lag.5bn notional (approximately MXN425bn or USD33bn) of these bonds were outstanding. Liquidity as of January 2010 is concentrated in the on-the-run Jun12 and Nov35 issues.4 7. The total return will depend on the realisation of the UDI Index. couponbearing bonds auctioned by Banxico as agent of the Treasury (MUDI <Govt> on Bloomberg).00 4. The UDIBonos market The UDIBonos are UDI-denominated.Barclays Capital | Global Inflation-Linked Products – A User’s Guide daily geometric equivalent of the corresponding bi-weekly inflation rate according to: UDI t = UDI t −1 × (1 + π )1 / n where π is the most recent reported bi-weekly inflation rate and n is the number of days during the period to which the inflation rate corresponds. As of end-2009. while the 10y onthe-run Jun19 is less actively traded. hold just 3% of the outstanding amount of these bonds.1 4. Barclays Capital Domestic investors hold 94% of the total outstanding amount of UDIBonos.50 3.00 3.6 10.2 3. MBonos. 15 March 2010 126 . UDI98. semi-annual (182-day). the real yield calculation is simply the “yield to maturity” of the bond quoted in UDI. Out of the MXN2.50 4.50 5. Figure 91: UDI Bonos outstanding Maturity date Issue date Coupon Coupon freq.50 3. which will affect the interest accrued and principal of the bond. UDIBonos represent 15% – a proportion that has steadily declined from the 30% peak in the mid1990s. Given the ‘bullet’ structures of UDIBonos. conceptually the calculations are similar to those of the Canadian-model. with a typical bid-ask spread of 5bp. the government remains interested in developing the UDI market in light of pension funds’ need to hedge future inflation-linked liabilities.4 8% 5% 7% 3% 5% 11% 11% 14% 3% 9% 25% Source: Bloomberg. UDIBonos.50 4. The average trade ticket is MXN100mn. Euro-clearable.3 98. Hence. Foreign investors.5 5.6 24. Day count Outstanding UDIbn % total Dec-10 Jan-11 Jun-12 Dec-12 Dec-13 Dec-14 Jun-16 Dec-17 Jun-19 Dec-25 Nov-35 Oct-07 Jan-01 Jan-09 Jan-03 Jan-04 Jan-05 Jul-06 Jan-08 Jul-09 Jan-06 Jan-06 3.50 4. Bondes and Bondes D) outstanding.1 8.25 6.25 5.6 10.3 3. That said. with local pension funds being the single key holders (52% of the total).8 13.77trn (USD216bn) of the Mexican government ‘bonded-debt’ (ie.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation derivatives and non-government UDI debt OTC inflation-linked swaps are traded in both UDI/Libor and UDI/TIIE formats. UDI/TIIE swaps are fixed-for-floating real rate swaps. although throughout 2009 activity was sporadic with only three months registering activity. Average monthly issuance of these securities has been UDI0.5bn. Cross-currency basis risk is present in this type of swap. In general. Private sector issuance of medium-term and long-term UDI-denominated securities has been on the rise since 2003. Figure 92: Mexican linkers: historical performance and risk 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% 2004 Source: Barclays Capital Figure 93: Return/risk of linkers versus nominals and equities 5 4 3 2 1 0 -1 -2 2004 Equity IL Nominal Bonds Mexico IL Returns Mexico IL Ann Mthly Vol 2005 2006 2007 2008 2009 2005 2006 2007 2008 2009 Source: Thomson Datastream. much of this acceleration came from issuance securitised against inflation-linked mortgages and this flow understandably fell away as global credit conditions deteriorated in the second half of 2008. Act/360) and the other receiving/paying the six-month USD Libor floating rate.27bn since 2006. close to the peak level (UDI13.9bn) registered at the end of 2008. the stock of this type of debt was about UDI13. liquidity is lower than in the UDIBonos market. However. The former are offshore fixed real-for-floating cross-currency swaps. As of November 2009. the notional amount is exchanged at the start of the contract. with typical bid/ask spreads of 10bp/15bp and average ticket size of MXN100mn. Barclays Capital 15 March 2010 127 . with renewed force between 2006 and 2008. while 25k/50k DV01 are traded daily. where one counterparty pays/receives a fixed UDI (real) rate and the other receives/pays the 28-day TIIE floating nominal rate. In both formats. 5y and 10y maturities are most liquid. with one counterparty paying/receiving fixed UDI rate (semi-annual.

while the CER is calculated by the central bank. housing and basic services (12. equivalent to approximately 27% of Argentina’s total outstanding debt. the 15 March 2010 128 .6%). Despite the government’s efforts.com Sebastian Vargas +54 (0) 114850 1230 sebastian.vargas@barcap. has detracted from the appeal of this market among institutional and other long-term investors at whom these assets were originally targeted. In addition. The bid-offer spread is typically 2-5bp in real yield terms. is published daily by the central bank and is calculated using the geometric mean of the changes in the consumer price index (CPI) with a one-month lag. the new methodology was unable to restore confidence in the new CPI.3%). a series of swaps aimed to extend maturity and switch to nominal bonds also reduced the amount outstanding. In addition. apparel (7. medical attention and healthcare expenses (5. transportation and communications (16. household equipment and maintenance (4. In the first 10 working days of each month of the year.3%). The new methodology was introduced in the midst of controversies over inflation under-reporting by the INDEC. This has. other goods and services (6.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Argentina Guillermo Mondino +1 212 412 7961 guillermo. among other negative consequences. coupled with the ongoing controversy about the measurement of the Consumer Price Index.6%). which.9%). The CPI methodology was changed in May 2008 to introduce a new CPI (base 2008) to replace the old CPI (base 1999). there has been significant controversy over the CPI Index and. in 2009. Since 2007. The wedge between genuine inflation and the CPI inflation is still significant. launched in February 2002. which was discontinued. Inflation under-reporting has significantly reduced liquidity in this market and the government no longers issues. the size of government CER-linked debt is USD38. The CPI is calculated by the Instituto Nacional de Estadistica y Censos (INDEC). Recent promises to restore INDEC credibility made by Minister Boudou have not yet materialised and the development of this asset class in Argentina seems highly dependent on future political developments.com Argentinian inflation linkers are an endangered species. It is important to note that the Argentine government stopped issuing inflation-linked debt in 2005.4% of the Barclays EM Government Inflation-linked Bond Index. Argentinean bonds represent 10. The index is broken down into the following nine groups (weights in parentheses): food and beverages (37. and education (4. leisure (5. hence the CER. significantly affected both domestic and global demand for inflation-linked bonds. The chosen basket is intended to represent the expenditures structure of the population considered in the National Survey of Household Expenditure (ENGH) of 2004-05. There is a widely-held belief by investors that inflation is being understated. This new CPI replaced the old index.3bn (including loans and other non-marketable debt).9%). the INDEC makes public the index corresponding to the previous month.mondino@barcap.1%).3%). CER Inflation-linked Index Inflation-linked bonds issued were issued in 2002 for the first time in the recent past and are indexed to consumer prices through the CER Index (Coeficiente de Estabilización de Referencia). The CER Index. These bonds are quoted on broker screens in both ARS and USD (quotes in USD use an implicit exchange rate). The CER-linked bond market As of September 2009.1%).

Banks and depositors were compensated with BODEN 2007 and BODEN 2012. Bonos de Consolidación (BOCON) 4. The structure of the marketable inflation-linked debt can be summarised as follows: 1. respectively. The central bank received BODEN 2011 and BODEN 2013. All of these bonds incorporated a GDP-linked unit (or GDP warrant.5%. which began to trade separately in November 2005.19%. in exchange. DISCOUNT.and ARSdenominated) bonds. USD. BODEN were largely issued to compensate for losses incurred by individuals and financial institutions as a result of the compulsory peso conversion adopted by the government during the 2002 economic crisis. Of the $82bn of eligible debt. 30/360 Amortisation Am: 84 equal monthly installments (first 83 of 1. starting 5/30/04 Am: 8 S/A equal installments of 12. The government also issued PAR/DISC and QUASI PAR regulated by the Argentine local law and linked to inflation. As a result. who. Pension funds were compensated with BODEN 14. trading in those bonds is less liquid than in the DISC (particularly in the QUASI PAR. two swaps performed during 2008 aimed to extend maturity and switch currency denomination (to nominal bonds) have reduced the amount outstanding of CER denominated debt even further. last 1.5178) S/A. Bonos Garantizados (BOGAR) Bonos del estado nacional (BODEN) Bonos del estado nacional are the BODEN (CER-linked) and BONAR (USD. The following table summarises the structure of the BODEN bonds: Figure 94: BODEN bond structure Bond BODEN 2011 CCY CER Issue date 30 Apr 03 Maturity 30 Apr 11 Type Sinking fund Coupon 2%* Monthly. DISCOUNT and QUASI-PAR bonds carried a haircut of 66.23%). about 76% was tendered by holders. In addition. Bonos de reestructuración (PAR/DISCOUNT/QUASI PAR) 3.3% and 30.Barclays Capital | Global Inflation-Linked Products – A User’s Guide nationalisation of pension funds in October 2008 will potentially reduce the traditional demand for this kind of instruments. which was customised for buy-and-hold private pension funds). Argentina extended a global exchange offer to the holders of its defaulted debt. while public sector workers and pensioners received BODEN 2008. EUR and JPY. received PAR. as is usually referred to in the market). 30/360 Note: *Principal is adjusted for inflation using CER Index (T-10 business days)/Initial CER (note this varies for the different BODENs). Source: Mecon Bonos de reestructuración (PAR/DISC/QUASI PAR) In February 2005.1%. starting 3/31/11 BODEN 2014 CER 30 Sep 04 30 Sep 14 Sinking fund 2%*(initial CER = 1. The following table summarises the structure of these bonds: 15 March 2010 129 . Both the PAR and the QUASI PAR were targeted at long-term local investors. Bonos del estado nacional (BODEN) 2. or QUASI-PAR bonds denominated in ARS.

The following table summarises the structure of some of the most liquid BOCON issues: Figure 96: BOCONs structure Bond PRE 08 CCY CER Issue date 3 Feb 02 Maturity 3 Jan 10 Type Sinking fund/ capitalised Coupon 2%* monthly. 4. is additionally guaranteed by the central 15 March 2010 130 . 30/360 Note: *Principal is adjusted for inflation using CER index (T-10 business days).Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 95: PAR/DISC and QUASI PAR bond structure.Cap: interest fully capitalised in the first 10yrs QUASI PAR CER 31 Dec 03 31 Dec 45 Step-up cpn/sinking fund/capitalised Note: *Principal is adjusted for inflation using CER Index (T-10 business days)/Initial CER (1. 30/360 Amortisation Am: 20 equal S/A instalments starting 6/30/29 DISC CER 31 Dec 03 31 Dec 33 Step-up cpn/sinking fund/capitalised Am: 20 equal S/A installments. An earlier series of BOCON was issued as compensation to families of victims who were jailed or disappeared during the military dictatorship.23%.4549). 3. starting: 2/3/06 – Cap: thru 1/3/06 Am: 72 equal monthly instalments of 1. The first 119 of 0. 30/360 PRO 12 CER 3 Feb 02 3 Jan 16 Sinking fund/ capitalised Sinking fund 2%* monthly. 30/360 PRO 13 CER 15 Mar 04 15 Mar 24 2%* monthly. S/A. S/A. They are divided into Bocones de Deudas Previsionales (Pre 8 and Pre 9) and Proveedores (Pro 11.77% 1625yrs.25%. Source: Mecon Bonos Garantizados (BOGAR) Issued by the Fondo Fiduciario de Desarrollo Provincial (a trust). is secured by a pledge of up to 15% of the province’s share in shared tax revenues and. 30/360 PRO 11 CER 3 Feb 02 3 Dec 10 Sinking fund 2%* monthly. which.75%. 1. Payment is secured by a government guarantee. or for additional amounts authorised through new decrees.04% Cap: thru 1/3/06 Am: 120 equal monthly instalments of 0.77% 6-10yrs.95% the last one of 0.35% except the last two of 2.06% 6-10yrs. Pro 13 and Pro 12).24%. 30/360* 0% in the first 10yrs.Cap: 3. 30/360 Amortisation Am: 48 equal monthly instalments. 2.08% the last ones of 2.04% first 5yrs. starting: 4/15/08 Am: 106 equal monthly instalments the first 105 of 0. S/A.83% thereafter*. starting 6/30/34. BOGAR bonds were used to restructure the debt of a number of provinces. Source: Mecon Bonos de Consolidación (BOCON) The Bonos de Consolidación (BOCON) were issued by the National Government to restructure its obligations to pensioners and suppliers. The first 47 of 2. .63% first 5yrs. 1. 1. Individual reopenings are not reported.18% 6-15yrs. in turn.84% the last one of 0. beyond that ceiling. 5.79% first 5yrs. The Pre 08 and Pro 12 are the most liquid BOCON bonds. starting: 4/15/14 PRE 09 CER 15 Mar 04 15 Mar 14 Sinking fund 2%* monthly. starting 9/30/24 . Bond PAR CCY CER Issue date 31 Dec 03 Maturity 31 Dec 38 Type Step-up cpn/sinking fund Coupon 0.31% thereafter*. Re-openings of BOCON subsequently took place opportunistically for amounts within those originally authorised in 2002. 30/360 2.83% except the last one of 1. starting: 3/3/02 Am: 120 equal instalments. 0% thereafter Am: 20 equal S/A installments.48% thereafter*.

Actual/365 2%* monthly. Real yield-to-maturity calculation To calculate the real yield to maturity for CER-linked bonds. The only complication in this formula is to adjust the quoted all-in peso market price to take into account the change in the CER Index: Pr ice = 15 March 2010 Mkt Pr ice CERT − d / CER0 131 . The CER index was fixed as 1 ARS on 2 February 2002 and tracks the Argentine CPI with a one-month lag: CERt = Ft * CERt −1 where Ft is Daily CER The Daily CER factor factor Ft is calculated as follows: a) Ft = CPI j − 2 CPI j − 3 ( ) ) 1/ k for days 1-6 of each month.Cap: thru 9/4/02 Partially capitalised: first 3yrs of the 2% cpn: 60% will be paid in cash. starting 3/4/05 . it is sufficient to solve for R in the equation: Pr ice = ∑ Where c(t ) qt t =1 (1 + R / q ) (in years) and n c(t ) is the cash flow at time t R is the real yield to maturity q - compounded. Actual/365 Amortisation Am: 156 monthly rising payments. Source: Mecon CER and real yield calculations CER calculation CER are units of account. the first step is to set up the entire real cash flows structure (including amortisation and capitalisation). between the second and third month previous to the current month. b) Ft = CPI j −1 CPI j − 2 ( 1/ k for days 6 to the last day of each month.9999). Since these real cash flows are essentially deterministic to calculate the real yield to maturity. From 2/4/05-8/4/05 the 2% coupon will be capitalised. Paid in cash thereafter Note: *Principal is adjusted for inflation using CER index (T-5 business days)/Initial CER (0. Figure 97: BOGAR structure Bond BOGAR18 BOGAR20 CCY ARS ARS Issue date 4 Feb 02 4 Feb 02 Maturity 4 Feb 18 4 Oct 20 Type Sinking fund Partially capitalised Coupon 2%* monthly. the financial intermediation tax and remaining fiscal resources (net of what corresponds to the state-managed social security system). the CER is based on the geometric mean of CPI variation. 10% will be capitalised and 30% represents the haircut. whose value in pesos is indexed to Argentine CPI.Barclays Capital | Global Inflation-Linked Products – A User’s Guide government. The CER time series can be viewed on Bloomberg using the code ACERCER <Index>. the CER is based on the geometric mean of CPI variation during the previous month: where k is the number of days in the current month and j is the current month.

0000 0. but activity has dried up since controversy over CPI measurements developed. at a 0% rate. Barclays Capital 1. Source: Mecon.5178 1. PRO 12 BOGAR 18 Note: T= payment date. The purchase of securities at the secondary market requires investors to deposit 30% of the total amount brought into the country. government and central bank bonds are not subject to withholding tax or any income tax.9999 T-10 T-10 T-10 T-10 T-5 Taxation and capital controls In Argentina. CERT − d is the index level for T-d (where T is the payment and d specific number of days) and is the base index level (fixed at inception of the specific bond). Bloomberg. d number of days. Non-government CER debt and derivatives There has been almost no significant non-government issuance linked to the CER Index.Barclays Capital | Global Inflation-Linked Products – A User’s Guide where Mkt Pr ice is the peso price quoted. Figure 99: Argentinian linkers – historical performance and risk Figure 100: Return/risk of linkers versus nominals and equities 4 3 2 1 0 -1 -2 2004 Equity IL Nominal Bonds 200% 150% 100% 50% 0% -50% -100% 2004 Argentina IL Returns Argentina IL Ann Mthly Vol 2005 2006 2007 2008 2009 2005 2006 2007 2008 2009 Source: Barclays Capital Source: Thomson Datastream. The following table summarises the change in CER calculation for the most liquid CER-linked bonds: Figure 98: Change in CER Index adjustment BODEN 08 BODEN 14 PAR. Argentina has capital controls for short-term capital flows.4053 1.4549 1. However. DISC PRE 08. CER swaps have traded versus USD Libor. Barclays Capital 15 March 2010 132 .

18. Being a highly open economy with few regulated prices.0 Entertain.melzi@barcap. The largest issuer is the central bank. international developments quickly transmit into domestic prices. Health.5 Inflation target mid-point Source: Barclays Capital Source: Central bank. 6.0 Clothing. The CPI is calculated by the National Institute of Statistics (INE). Barclays Capital 15 March 2010 133 . (including the vast majority of mortgages). but also in labour and house rental contracts. 13.4 7..Barclays Capital | Global Inflation-Linked Products – A User’s Guide Chile Roberto Melzi +52 55 5241 3260 roberto.9 Drinks. with a combined uplifted notional size of CLP8. using the CPI-linked UF unit.zuniga@barcap. the index is based on a nationwide geographical coverage including all the region’s capital cities (as opposed to the previous’ index focus on Gran Santiago only). This is particularly noticeable with oil prices (despite the operation of a stabilisation fund) and with foodstuff. 4. 19. but the government also issues debt using the central bank as agent. Indexation Chile has a long tradition with price indexation. CPI releases are not subject to revisions nor are they seasonally adjusted. indexation has been almost exclusively based on the ‘Unidad de Fomento’ (UF). with survey inflation expectations well-anchored at 3% over the 2y policy horizon..7 Food. whose base was one month (December 2008). which releases CPI for the previous month during the first eight days of the month within a pre-announced annual calendar.3 0 -2 -4 2000 2002 2004 2006 2008 CPI % y/y Transport. Most domestic debt is inflation indexed. the central bank enjoys ample credibility. 4. in contrast with the previous index. Despite CPI inflation volatility (mainly due to food and fuel prices).3 Furniture. while UFlinked derivatives are also traded. This fact partly explains why Chile escaped from the widespread dollarisation observed in other Latin American economies that.4 Diverse. It is present not only in financial markets. The central bank runs an inflation-targeting regime.com Few countries have indexation embedded within their financial markets as much as Chile.9bn) at the end of 2009.57tn ($16. Figure 101: CPI weights H &R. 2. with a 3% operational objective (in a +/-1% tolerance range) within a two-year policy horizon. 7.8 Education. The new CPI also adopted the entire average of 2009 prices as a base.com Jimena Zuniga +1 (212) 412 5361 jimena.5 Com munic. like Chile in the 1970s. which are related to soft commodity prices such as grains and milk. Since 1967. with daily readjustments based on the previous month’s CPI inflation prevailing since 1977. CPI-indexed bonds were first issued in 1966 by the central bank. 5. Since February 2010.2 Figure 102: Inflation dynamics 12 10 8 6 4 2 Housing. although indexed deposits to both wages and CPI occured as far back as 1959. 5. suffered from very high inflation. 5.

down from 93% in 1992.1) is the monthly CPI inflation of month m-1 rounded to the first decimal point. BCUCL <Govt> on Bloomberg) are standard bullet bonds with semi-annual interest payments that are issued with 5y. When the economy is growing below/above potential the government runs a cyclical/temporary deficit/surplus. Yet. 10y and 20y tenors. CHILBT <Govt> on Bloomberg). with CPI released early in each month to enable continuous calculation of the UF as each release fixes the index to the 9th of the following month. we do not foresee its debt growing significantly given the structural balance rule. the central bank is likely to continue to have more outstanding debt paper than the Treasury. the central bank has since August 2002 been increasing the supply of BCUs. with the market for both being dominated mainly by local pension funds and banks. PTFs (floating rate ‘pagares’) and PRCs (coupon-bearing 4y/20y ‘pagares’). for day t between the 10th of month m and 9th of month m+1: UFt = UFt −1 × (1 + round (π m−1 . zero-coupon ‘pagares’). In addition. BCUs (‘Bonos del Banco Central de Chile’. Foreign investors do not participate much due to taxation issues and the complexity of trading them (a local custodian account is needed). The average ticket size is UF100k. away from the inflation-linked PRBCs (5y. UFdenominated standard bullet bonds. 15 March 2010 134 . In other words. among others. with a typical bid/ask spread of 2bp and an average traded volume (which is quite volatile. BCUs are more liquid than BTUs.1bn). some additional supply from the Treasury could materialise. The same readjustment factor is used between the 10th day of the current month and the 9th day of the following in the daily equivalent of inflation calculated geometrically. the government (with the central bank acting as its agent) issues BTUs (‘Bonos de la Tesoreria General de la Republica en UF’.1))1 / d where d is the number of calendar days between those two dates and round (π m −1 . Thus. The UF-linked bond market The largest issuer of inflation-linked bonds is the central bank.3trn/USD20. Government supply of debt in general depends on its fiscal situation and is contingent on the ‘structural balance’ rule.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The UF Index (CHUF <Index > on Bloomberg) is calculated and published by the central bank. To promote the development of this market and foreign investors’ participation. depending on pension funds participation) of around UF2m per day. Given that the financial crisis increased the need to finance higher deficits in coming years. UF-linked debt represents 63%. As of June 2009. out of the central bank’s total issued securities (CLP11. implying higher/lower net issuance of debt (including BTUs).

00 3.00 3.00 3.00 5.8 19. the tax is zero.39 2.00 5.8 8. not to earnings due to indexation (in general the Chilean tax system considers tax bases that are CPI deflated).0 8.2 5% 8% 5% 4% 5% 5% 4% 7% 11% 4% 4% 1% 4% 5% 8% 9% 4% 0% 3% 4% BTUs Jul-14 Sep-15 Jul-19 Oct-23 Aug-24 Sep-25 Mar-27 Mar-28 Mar-29 Mar-38 Mar-39 Jul-09 Sep-05 Jul-09 Oct-03 Sep-04 Sep-05 Mar-07 Mar-08 Mar-09 Mar-08 Mar-09 3.8 10.10 3.0 2. However.4 23.5 11.2 10. have a more cumbersome treatment.4 10.60 3.00 4.00 3.2 8.0 25.00 3.00 Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 11.00 5.00 5.00 3. This rate applies to the BCU rate.0 11.5 11.0 20.0 19. Capital gains.00 2.00 3. The general rule for bonds is that they pay the general income tax.6 9.00 Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Semiannual Total Source: Bloomberg.2 19.2 11.00 3.00 3. 15 March 2010 135 .3 26.0 12.00 3.8 0.00 3.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 103: BCUs and BTUs outstanding Maturity date Issue date Coupon Day count Day count Outstanding UFm % total BCUs May-10 Aug-10 Sep-10 Sep-11 Apr-12 Sep-12 Oct-12 Apr-13 Jul-13 Oct-13 Nov-13 Apr-14 Jan-16 May-17 Jan-18 Jul-18 Oct-18 May-19 Sep-22 May-28 May-08 Aug-08 Sep-05 Sep-06 Apr-07 Sep-02 Oct-07 Apr-08 Jul-08 Oct-08 Nov-03 Apr-09 Jan-06 May-07 Jan-08 Jul-08 Oct-08 May-09 Sep-02 May-08 3. however.0 10.00 5.00 3.6 8% 7% 15% 12% 11% 7% 6% 12% 5% 11% 6% Taxation Chile has a general flat 4% tax rate for interest earned by foreigners in the domestic market.4 18. if the investor is domiciled in a country with which Chile has a double-taxation treaty.1 9.00 3.00 3.4 11.0 12.9 18. Barclays Capital act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 act/365 13.0 162.00 5.00 3.0 17. which is 35% for foreigners.0 10.50 4.00 3.00 3.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Non-government UF debt and derivatives
Out of the 677 corporate bonds outstanding at the end of October 2009, more than 92% were UF-denominated, with local pension funds and insurance companies being the most important holders. A similar story occurs in the securitisation market, where 77% of the 267 bonds outstanding are linked to UF. Regarding derivatives, two main types of swaps are traded OTC. The UF/Camara swap is a fixed-for-floating contract with a counterparty paying/receiving a fixed UF rate and the other receiving/paying a floating nominal rate that depends on the ‘Indice de Camara Promedio’ (ICP). The ICP Index represents the funding cost of local financial institutions and is published daily by Chile’s banking association (www.abif.cl). The index depends on the ‘Tasa Camara Interbancaria Promedio’, which is the average O/N interbank interest rate set by the central bank. The other key swap traded OTC is the UF/Libor, an offshore, fixed-for-floating crosscurrency swap, where a counterparty pays/receives a fixed UF rate and the other receives/pays the 6mth USD Libor floating rate. In both swaps, coupons are exchanged semi-annually in an ACT/360 day-count convention, with notional amounts exchanged at the start of the swaps. Finally, it is also possible to trade floating real rate versus fixed real rate swaps, with the floating real rate determined by deflating nominal ICP with the UF Index. Activity is mainly concentrated in the interbank market, with foreign hedge funds participating comparatively more in the nominal swap market. Liquidity is concentrated in the 1y, 5y and 10y segments of the curve (typical bid/ask of 5bp/10bp), although quotes can be obtained up to 20y. Local corporations use these markets for liability management. Finally, there is also a liquid market of inflation rate forwards (up to 1y), where investors can trade their views on inflation rates in the future directly. In combination with positions in nominal swaps, these ones provide a means to synthetically trade UF-denominated swap rates.

Figure 104: Chilean linkers – historical performance and risk
12% 10% 8% 6% 4% 2% 0% 2003 2004 2005 2006 2007 2008 2009 Chile IL Returns Chile IL Ann Mthly Vol

Figure 105: Return/risk of linkers versus equities
4 3 2 1 0 -1 -2 2003 Equity IL 2004 2005 2006 2007 2008 2009

Source: Barclays Capital

Source: Thomson Datastream, Barclays Capital

15 March 2010

136

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Colombia
Roberto Melzi +52 55 5241 3260 roberto.melzi@barcap.com Jimena Zuniga +1 (212) 412 5361 jimena.zuniga@barcap.com

Inflation-linked bonds were first issued in 1967 and the market has grown steadily since then. Around a quarter of government debt is linked to CPI via the UVR indexing unit, but taxation limits interest from international investors.

Indexation
The central bank (Banrep) conducts monetary policy within an inflation-targeting framework. In October 2009, Banrep set the 2010 inflation target to be the long term-goal of 3% (in a +/1% tolerance range), culminating a successful process of gradual target reductions and inflation convergence since the inception of the inflation-targeting regime. Although inflation averaged 6.2% in 2001-08, it plummeted to 2.0% by December 2009 and is widely expected to end 2010 within the central bank's new target range. As in other emerging countries, food prices are critical drivers of inflation volatility. Energy prices are relatively heavily regulated and typically reflect international prices more smoothly and with a lag. CPI inflation is calculated monthly by the statistics office (‘Departamento Administrativo Nacional de Estadistica’ – DANE). The CPI has December 2008 as its base, with the consumption basket weights fixed and representing the consumption survey of 2006-07. Geographical coverage includes the urban population of 24 cities. The CPI represents average monthly prices and is released by the fifth day of the following month. Since August 2000, the inflation indexing unit has been the ‘Unidad de Valor Real’ (UVR <Index> on Bloomberg), which is calculated/published by Banrep based on DANE’s CPI inflation rate for the previous month. The UVR Index is adjusted daily between the 16th day of the current month and the 15th day of the following month. For any day t within this period, the UVR is given by:

UVRt = UVR15 × (1 + π m −1 ,1) t / d
15 is the index value in day 15 of the previous month, where month’s m/m inflation rate and d is the number of days in the period.

UVR

π m −1

is the previous

Figure 106: CPI weights
Others Communic. 6% 4% Transport. 15% Entertain. 3% Education 6% Health 2% Apparel 5% Housing 31%

Figure 107: Inflation dynamics
11 Food 28% 10 9 8 7 6 5 4 3 2 2000 2002 2004 2006 2008 CPI % y/y

Inflation target mid-point
Source: DANE, Barclays Capital Source: Central bank, Barclays Capital

15 March 2010

137

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

The UVR-linked market
While there are two types of linkers, the UVR-denominated are the most liquid ones (page 2 of SENC1 on Bloomberg). These are fixed-rate, coupon-bearing bonds. As of end-October 2009, total gross domestic debt was COP132trn, out of which coupon-bearing TES bonds (both nominal and UVR-linked) represented almost 90%. Fixed-rate UVR-denominated debt represented 25% of the aforementioned total. Liquidity varies significantly, depending on market sentiment. In a rallying market, a typical bid/ask spread is 5bp for the most liquid paper (Sep10 and Mar13 and Feb23) and 10bp for their illiquid counterparts. When the market is under pressure the liquid bonds’ bid/ask spread widens to 10bp, while quotes for other bonds are practically non-existent. The average ticket size is UVR20m and many days can pass without activity in this market. The UVR/Libor OTC is a swap that used to trade but is practically non-existent now. This swap was an offshore fix-for-floating cross-currency swap, where a counterparty pays/receives a fixed UVR rate (semi-annual, Act/360) and the other receives/pays 6mth USD Libor floating. When and if prices are provided, bid/ask spreads are very wide. When inflation-linked corporate deals occur, activity in this market rises, although this is sporadic.

Taxation
Taxation issues limit foreign investors’ participation in the UVR-linked bond market as they are subject to withholding tax on both income and capital gains. The rate ranges from 7% if the bond matures within five years to 4% if it has a longer maturity. In addition, UVR bonds are subject to a 0.4% financial transaction tax.

Figure 108: Linkers – historical performance and risk
30% 25% 20% 15% 10% 5% 0% 2003 2004 2005 2006 2007 2008 2009 Colombia IL Returns Colombia IL Ann Mthly Vol

Figure 109: Return/risk of linkers versus nominals and equities
5 4 3 2 1 0 -1 2003 Equity IL Nominal Bonds

2004

2005

2006

2007

2008

2009

Source: Barclays Capital

Source: Thomson Datastream, Barclays Capital

15 March 2010

138

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Uruguay
Sebastian Vargas +54 (0) 114850 1230 sebastian.vargas@barcap.com

Uruguay resorted to a CPI-linked market in the aftermath of the 2002 financial crisis, as a means of creating a local currency market in the context of widespread financial dollarisation and with both high and uncertain inflation expectations. Owing to the government’s effort, the market expansion has been sustained. The creation of an inflation linked bond market is part of a consistent effort to provide a dollar substitute and is continuing at a growing speed. To foster this market, in 2002 the Uruguayan government replaced the old Unidad Reajustable (UR), which was adjusted with a wage index, with a newly-created Unidad Indexada (UI), which was adjusted with the CPI.

CPI and the Unidad Indexada (UI)
The CPI is calculated by the national statistics agency, INE, on monthly basis and is released over the first five days of each calendar month, covering the period of the previous month. The index is based on March 1999 and prices are collected in Montevideo. The UI was introduced in June 2002, with its level set at 1 Uruguayan peso per UI, and is calculated and published by the INE. The level of UI is calculated on a daily basis from the sixth day of the month up to the fifth day of the following month, using CPI levels from the previous month. Specifically, the daily factor UId,,M is computed according to the following formulae:

a) UI d ,M

⎛ CPI M − 2 = UI 5, M −1 ⎜ ⎜ CPI M −3 ⎝ ⎛ CPI M −1 = UI 5, M ⎜ ⎜ CPI M −2 ⎝

⎞ ⎟ ⎟ ⎠

d + DM −1 − 5 D M −1

from the first day of month M up to the fifth day;

b) UI d ,M

⎞ DM from the sixth day of month M to the end of the month, ⎟ ⎟ ⎠

d −5

where DM represents the amount of calendar days in month M and CPIM corresponds to the CPI level in month M.

Inflation-linked bonds
Since the creation of the UI, the government has issued several types of bonds linked to the index, in particular Letras, Notas BCU and Notas del Tesoro, with initial maturities ranging between three and 10 years. Recently, there has been notably larger supply in UIdenominated global bonds whose cash flows are UI-linked but paid in USD. Three benchmarks have been issued since 2006 with 2018, 2027 and 2037 maturities and now make up almost half of the government’s inflation-linked debt.

15 March 2010

139

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Israel
Arko Sen +44 (0) 20 3134 2839 arko.sen@barcap.com Daniel Hewitt +44 (0) 20 3134 3522 daniel.hewitt@barcap.com

Israel has the oldest continuous sovereign linker market, with CPI-linked government bonds first issued in 1955 and a wide spectrum of corporate inflation-linked bonds. This was in reaction to years of very high inflation rates. Although inflation has been below 10% since 1998, inflation-linked bonds are still an important instrument in Israel’s financial markets.

CPI
The inflation reference for linked bonds in Israel is the consumer price index (CPI). The index is calculated and published on the 15th of each month by Israel’s Central Bureau of Statistics. This is calculated from a monthly survey of 1,320 goods and services, by numerators or by phone, of c.2,700 stores, business and households in c.100 localities. The weightings in the basket are based on the household expenditure survey and are changed every few years. No schedule has been given for the next change, but weighting changes tend to be modest. The Israeli central bank targets 1-3% CPI growth per annum, and was successful in achieving this target range from 2004 until 2008, when inflation exceeded 5% at one point. Since then, the financial crisis has helped pull inflation lower, but new indirect taxes and house price increases have kept inflation near or above the upper end of the target range. Food accounts for only a small share of the CPI basket, and at this stage one of the key risks to inflation seems to be from housing, which accounts for 21% of the CPI basket and has remained buoyant throughout the global crisis. FX indexation in the Israeli economy has fallen substantially. The BOI estimates that the proportion of rental contracts quoted in USD has fallen from 90% to 15% over the past two years. As a result, the pass-through from the exchange rate to the housing component has fallen, from 90% to 48% in the short term (one month), and has disappeared over a longer term (year and above) where previously it was fully transmitted.

Figure 110: Israel’s CPI basket
Miscalleneo us , 4.46 Food and Beverages, 18.4

Figure 111: Inflation in Israel
8.0 6.0 4.0 Apparel, 3.2 2.0 0.0 Housing , 20.69 -2.0 -4.0 Dec-99

Transport and Communica tion , 21.14

Health , 5.22 Education , 12.53 Furnishing , 3.75 Housing maintenanc e , 10.62

Dec-01

Dec-03

Dec-05 Israel CPI

Dec-07

Dec-09

Source: Haver Analytics, Barclays Capital

Source: Bloomberg, Barclays Capital

15 March 2010

140

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Market development
There is about ILS124bn of tradable government CPI-linked bonds outstanding, equivalent to 48% of traded local government bonds. The traditional CPI-linked bonds in Israel are ‘Galil’ bonds. However, since 2006, the Israeli government has been issuing a new type of CPI-linked bond – the main difference versus the Galil bonds is the absence of a deflation floor for the value of the bond’s principal and coupon. The new inflation-linked bonds as of the end of 2009 had 2010, 2014, 2018, 2019 and 2036 maturities, with issues having been tapped monthly since early-2007. On some of them the amount outstanding has now reached similar levels to benchmark Galil bonds. The real curve in Israel has a greater maturity than the nominal curve (2036 against 2026 for the longest maturity nominal local government bond). However, the government’s emphasis on developing the unlinked curve has been slowly paying dividends. The share of the tradable local government bond market, which is unlinked (either to CPI or FX), has risen from over 8% in 1995 to about 58% by end-2008. In 2009 this trend paused as the share of CPI-linked issuance increased to 27% compared to 21-24% over 2006-07. The reform of Israeli pension funds, which started in January 2004, marked an important change in the local debt market: the issue of non-tradable government debt to these funds virtually ceased and these funds became more active in the tradable local debt market – both in linked and unlinked bonds. Initially, their buying interest was concentrated in CPIlinked bonds, but in recent years their buying of linked and unlinked bonds has stabilised. Domestic pension funds’ share of total holdings in tradable government bonds rose from 2% in 2002 to 13% by end-2008. Provident funds have generally declined in importance and hence so have their share of the holdings, with individuals overtaking these funds as the group holding the largest share of inflation-linked bonds (Figure 113).

Calculations
The annual coupon on Galil bonds are fixed real rates, while the value of principal is enlarged by the rate at which CPI has increased between the base date and the latest inflation observation. Indexation has a one- to a one-and-a-half-month lag: for example, the base CPI for a Galil bond issued on 20 August 2001 is the CPI for July 2001. Both the coupon and redemption value of Galil bonds in Israel have an implicit floor against deflation. By Figure 112: Distribution of tradable debt by type (%)
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1995 1997 1999 2001 2003 2005 2007 CPI-linked USD-linked Unlinked

Figure 113: Distribution of total bond holdings (%)
100 80 60 40 20 0 2005 Public Pension fund Foreigners 2006 2007 Mutual fund Banks BOI 2008 2009Sep*

Provident fund Insurance firms

Source: Ministry of Finance, Barclays Capital

Source: Ministry of Finance, Barclays Capital

15 March 2010

141

This market is likely to grow over the coming months/years as local corporates and banks adopt a more sophisticated approach to liability management and foreign interest in creating global inflation exposure increases. r = the fixed interest rate and T= number of days in the interest period. The capitalisation of the corporate linkers. and much of the current activity in this sphere is undertaken by Israeli corporates hedging their balance sheet exposures.Barclays Capital | Global Inflation-Linked Products – A User’s Guide contrast. hence the corporate linker market’s aggregate turnover is limited. was ILS181bn at end2009. As mentioned above. according to TASE. however. The latest CPI report is used to compute the index ratio from time of issue – this index is used to deflate the uplifted price and hence derive the implied real yield. or CPI known when the bond was first issued. While the corporate linker market is larger in absolute size. Non-government bonds and derivatives The corporate inflation bond market is larger than the sovereign market. For Galil bonds. The quoting convention is to use uplifted dirty prices. One of the more common products is real rate swaps where the floating leg is 3mth Telbor and the fixed leg is the real rate uplifted again by the ratio of the most recent CPI index over the CPI index at start. relatively modest-sized issues. it tends to be tightly held. The market for inflation derivatives is still in the early stages of development with the interbank market not yet very active. For the new CPI linked bonds. and with the large number of small. this induces volatility in the yield levels around CPI release dates. It is. a multiplicative formula is used: R = { (1+ r/100) ^ (T/365) -1 } * 100 Where R = interest rate for the interest period. The format of the linkage is as follows: Principal Payment = (M1 / M0) * 100 Coupon Payment = (M1 / M0) * 100 * Real Coupon Where M1 = latest CPI at the time of payment and M0 = base CPI. quite possible to trade inflation swaps on request. the latest CPI when the bond was first issued). so the principal and coupon can have step moves according to the release of the latest CPI report. There are about 444 corporate linkers listed on the TASE (Tel Aviv stock exchange) against 18 sovereign linkers. the interest payment is computed more simply as: R = r*T/365 The key difference between Israeli inflation-linked bonds and those of most other markets is that there is no official daily referenced index. 15 March 2010 142 . Another difference between Galil bonds and the new CPI linked bonds is the method of computing interest payments. the value of the principal of the new inflation-linked bonds can rise and fall with the change in CPI relative to the base CPI (ie.

2) the high exposure to external food and energy prices. headline inflation in Turkey can change significantly as a result of global commodity price developments combined with changes in the exchange rate – ie. in housing rent inflation. but with informal annual targets. Previously monetary policy targeted inflation reductions.5% for 2009.5% for 2011. y/y headline inflation was driven higher by food price Figure 114: Composition of Turkey’s CPI basket Other 11% Food and nonalcoholic drink 28% Figure 115: Inflation targeting is still evolving 14 12 10 8 CPI inflation Central bank target lower tolerance band Barcap forecast Upper tolerance band Communica tions 5% Hotels and restaurants 5% Transport 13% Alcohol and tobacco 5% Clothing and footwear 7% 6 4 2 0 Jan 04 Jan 05 Jan 06 Jan 07 Jan 08 Jan 09 Jan 10 Jan 11 Source: Central Bank of Turkey.com Christian Keller +44 (0) 20 7773 2031 christian.5%. with the uncertainty band of +/.2 unchanged. prices are collected twice each month (four times each month for vegetables.20% over a six-month horizon. 6. rising global food and energy prices combined with a weaker TRY can drive up headline CPI even as demand pressures remain relatively muted.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Turkey Arko Sen +44 (0) 20 3134 2839 arko. The index is published on the third day after the reference month and has a large coverage span of 26 regions. 81 provinces. according to our estimates. covering 447 items. using continuous household budget surveys. a 2012 issue. While demand conditions remained weak.5% target but well above the Central Bank of Turkey’s projection of 5. fruits.5% in the quarterly inflation report. Even with issuance likely to increase further and a sharp fall in real yields. Barclays Capital Household goods and services 8% Source: Turkish Statistical Institute Housing 19% 15 March 2010 143 .com Turkey issued its first Canadian-style linker in February 2007.sen@barcap. inflation risks still make inflation-linked bonds attractive. The prices include any relevant taxes. and a total of 30% over the longer term). which is calculated and published by the Turkish Statistical Institute. CPI inflation ended 2009 at 6. well below the official 7. From a structural perspective.keller@barcap. and related to this 3) the pass-through from exchange rates (c. After the initial inflation target of 4% had been missed in 2006 and 2007 – and seemed unattainable in the foreseeable future – the official targets were changed in mid-2008 to 7. Thus. Turkey’s inflation challenges remain: 1) the inertia in service price inflation. and petroleum products). Inflation targeting by the central bank formally began at the beginning of 2006. although there have been some breakthroughs in recent years – eg. Since then it has continued to build up the market for inflation linked bonds. especially in 2009 when linked issuance leapt to 16% of total supply.5% for 2010 and 5. Consumer price inflation Inflation-linked bonds are referenced off CPI. Bloomberg. Weightings and the item basket are updated annually. Generally. The data are not seasonally adjusted.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide developments. and coupled with base effects following the sharp disinflation from December 2008. but structurally too it is perhaps not hard to argue that Turkey will continue to offer one of the higher real yield levels in the EM space – as a function of its current account and fiscal deficit and also due to the still comparatively low monetary policy credibility under the relatively new inflation targeting framework. but were redeemed in 2000 and never enjoyed great popularity. Taken together. To date. however. At the end of January 2010. A small amount of inflation-linked bonds (albeit of a different format) had been issued in the 1990s.5% inflation target for 2011. which was reflected in the strength of foreign demand for Turkish linkers. Partly as a function of this and partly due to some fears of an inflation spike following the central bank’s sharp policy easing. the rise in headline inflation rate in H1 10 is likely to also raise expectations for 2011. This is. but it would not be a surprise if the Treasury chose to extend the curve longer. The initial launch of the February 2012 10% CPI-linked bond (TRY4.4bn by end-2007. They had been reluctant to buy these bonds when nominal yields were high but.0bn. In 2010. In 2008. CPI linked bonds found renewed favour among local investors. as a consequence of the rebound in oil and gas prices. One of the key reasons behind the surge in inflation-linked issuance over 2009-10 is the newfound interest for such bonds among local banks. on 15 March 2010 144 . we expect inflation to moderate toward the end of the year. In addition. the Treasury held only three auctions for a total sum of TRY2. breakeven inflation had reached startlingly low levels.8bn. for a total amount of TRY23bn.4bn in 2012 and 2013 bonds. all new inflation-linked bonds in Turkey have been issued with a 5y maturity. With base effects turning more favourable in the final months of the year and the output gap likely remaining large. The challenge building for monetary policy makers will also be to contain the deterioration in inflation expectations vis a vis the 5. The liquidity in the market has improved over the years but is still poor relative to the nominal yield curve. ending 2010 near the upper end of the tolerance band. which as of March had an amount oustanding of TRY7. highway fees. Market development The Turkish Treasury re-entered the linker market in February 2007 with a new instrument that followed the international standard for CPI-linkers (see below). a new 5y bond was issued in February. both real yields and nominal yields are poised to rise.3bn over the year. tobacco and alcohol. Anecdotal evidence also suggests that much of the inflation linked paper was originally bought for banks’ hold to maturity portfolios. ~US$3. the total amount of Lira-denominated debt (free float) amounted to TRY263bn of which CPI indexed debt amounted to TRY27bn. The government has announced a series of fiscal measures coming into effect in early 2010: increases in excise taxes on gasoline. Given that inflation expectations tend to be highly adaptive in Turkey. and various other taxes. Cyclically. It was in 2009 that the pace of issuance really accelerated as the Treasury held six auctions at a rate of up to one in a month.0bn sold) was met with strong demand and quarterly taps of the 2012 issue raised the nominal amount outstanding to USD5. this implies headline inflation rising toward 10% in the first half of 2010. no longer the case with real yields hovering around the 3-4% mark at the start of 2010. Turkey has historically offered one of the highest real yields in EM markets. with the collapse in policy rates and nominal yields. It seems clear that the Turkish Treasury remains committed to maintaining and building a real curve. the introduction of formula-based quarterly energy price adjustments since 2008 also implies further energy price increases for endconsumers in 2010. Two 2014 bonds were issued in 2009 with the amounts outstanding being increased to TRY8bn and TRY6. or about 16% of the year’s gross issuance.

Calculations Inflation-linked bonds in Turkey follow a Canadian format that investors are familiar with: a fixed annual coupon bond. Taking pure inflation views is still difficult as the repo market is not developed enough to take short positions. The Treasury updates the daily reference series for linkers after a new CPI report is published. with the principal value changing by the reference index. Non-government bonds Both the nominal and real corporate bond markets in Turkey have shown slow progress. 15 March 2010 145 . the general outperformance of the lira against the FX forwards has encouraged external borrowing by Turkish companies. although this likely has only a modest effect on pricing considerations. Turkish linkers are protected against deflation by a par floor. The problem has arisen because of the high nominal yields compared with developed markets and also as. As yet there has been no significant activity in inflation swaps. the market convention in Turkey has changed and CPI bonds are now quoted on a clean price basis similar to other markets. Artificially structured positions are possible by using cross currency swaps but these too are illiquid and investors would also need to assume the asset swap spread risk. This is due. especially as the government’s full balance sheet is likely leading to significant crowding out of the domestic bond market.7% average since 2004) though it has reached new lows following the financial crisis (5. linked to CPI with a two-to-three month lag. in part.Barclays Capital | Global Inflation-Linked Products – A User’s Guide which a 10y bond was issued for the first time in 2010.5 in November 2009). The format of the bond is as follows: Principal value = (Reference index maturity/Reference index issue data) * 100 Existing bonds have a coupon of 10% in real terms paid semi-annually. since 2002. given that inflation is still relatively high (8. to the dominance of borrowing via an extensive local banking system and recently because of the appeal of external borrowing. They are indexed against inflation as follows: Coupon payment = (Reference index coupon date/Reference index issue date) * 100 * Real coupon rate Daily reference index = CPI a-3 + (G-1 / AG) * (CPI a-2 – CPI a-3) Where CPI a-2 = CPI of (a-2) month and G = number of past days in month “a” Where CPI a-3 = CPI of (a-3) month and AG = number of total days in month “a” Notably. There is little indication that this will change.

helped by asset swapping of government bonds. which historically has been subject to periods of significant movement. The inflation-targeting regime was announced in 2000. South Africa remains vulnerable to episodes of high inflation volatility.6% y/y in August 2008. the pace of linker issuance by the National Treasury has increased significantly. as published by Statistics SA. the target of monetary policy was CPIX (a measure of CPI that excluded mortgage rate costs) covering metropolitan and urban areas.Barclays Capital | Global Inflation-Linked Products – A User’s Guide South Africa Jeff Gable +27 (0)11 895 5368 jeff. Inflation had been in the 3-6% band from September 2003 to March 2007. some are collected quarterly. This experience highlights that. helping to entrench the disinflationary trend and reinforce the credibility of monetary policy. inflation-linked products in South Africa reference the headline non-seasonally adjusted Consumer Price Index (CPI) for all urban areas.badsha@absacapital. The Monetary Policy Committee now uses the same inflation measure as its policy target. the first target year was 2002. with swaps trading developing subsequently. reaching a peak of 13. semiFigure 116: Weight of CPI in basket . Surveys of business/labour inflation expectations show above-target results. Comms and Insurance 13% Other 9% Food. medium-term inflation breakevens remain within the central bank’s 3-6% inflation target. however. Prices of goods and services included in the CPI are collected in the first seven days of the month. along with the continued development of an active repo market.com Bulent Badsha +27 (0)11 895 5323 bulent. The new CPI basket Starting in January 2009. In 1989. looking to keep inflation at 3-6%. Since early 2009. followed by a period of sharply rising inflation throughout much of 2007 and 2008. Culture. Alcohol and Tobacco 20% Clothing and footwear 4% Other housing and utilities 10% OER Households 12% contents 6% Figure 117: CPI inflation history 16 14 12 10 8 6 4 2 0 2004 CPIX (previous MPC target) CPI (new basket) SARB target upper bound SARB target lower bound Transport 19% 2005 2006 2007 2008 2009 Source: Barclays Capital Source: Barclays Capital 15 March 2010 146 . Following a surge in inflation during 2008-09.com Inflation-linked bonds have been available in South Africa since 2000. despite a solid policy commitment to low inflation. Hotels & Restaurants 7% Education.4% in 2004. Part of the difficulty is due to the currency. The 3-6% target has been unchanged since the introduction of the regime.6% in 1986. the South African Reserve Bank (SARB) adopted a strong anti-inflation stance as annual average CPI declined from a high of 18. Inflation and real rate swaps are also increasingly active.gable@absacapital. whereas inflation-linked products referenced CPI for metropolitan areas. While most prices are collected monthly. in line with the need for an extended period of deficit financing. Prior to January 2009. This additional issuance. to a low of just 1. has helped lead to a noticeable improvement in liquidity.

and though the average of analyst expectations for 2010-11 inflation is within the inflation band. The three further important changes all relate to weights. the most critical being the replacement of mortgage rates as a proxy for housing costs (as per the previous CPI basket) with an estimate of Owner Equivalent Rent (OER). but breakeven inflation measures suggest that the market expects inflation to remain firmly within the 36% inflation target (albeit towards the upper end). This is likely to mean a smoother trajectory for inflation through the interest rate cycle as changes in the policy rate under the previous CPI basket automatically induced a jump in the inflation series via mortgage interest costs.4 percentage points on measured CPI inflation in the January 2009 figure. The collection of prices depends on the frequency with which they tend to change. Every five years Stats SA conducts a Survey of Income and Expenditure of Households. the expectations of business and labour representatives are well above there. or household.95% earlier. In addition to the usual re-basing and re-weighting. The seasonality of this new series is expected to differ from that of the previous series. The survey is used to identify the goods and services purchased by a typical consumer. based on the 2005 expenditure survey.25% from 5.97% to 1. known as the R189.Barclays Capital | Global Inflation-Linked Products – A User’s Guide annually. however. The lower weight of electricity in the current CPI basket will be important in preventing the very large planned electricity price increases of the next few years from passing directly into inflation. the weight of vehicle purchases (new and used) rising to 11. such as airline tickets. The shift to the new basket affected a decline of about 1.68%. and the weight of electricity prices reducing from 2. The next survey revision is due only in 2014. Government bonds Market development The government first issued inflation-linked bonds in March 2000. The recent episode of high inflation has had a damaging effect on surveys of inflation expectations. that might have a high expenditure value in the basket but are only used by a small minority of South Africans. with the weight on foodstuffs reducing to 14. and which should be included in the basket of goods and services to monitor price changes.27% from 20.25% March 2013.99% previously. weights are determined for specific products in the basket. the revised methodology also involved a major reorganisation of the basket to “democratise” it by removing luxury items. or annually. The weighting of each product stays the same for the five-year period – ie. holders of inflation-linked bonds saw a direct index adjustment on an index rebase and so did not get affected by the revised y/y rate. with auctions now being held every Friday in linkers across the curve (there are 15 March 2010 147 . There are four other changes of note. From this survey. This resulted in the number of items in the basket falling from 1.21% in the current CPI basket. The government has underscored its commitment to the inflation-linked market by issuing inflation-linked bonds on a regular basis. but as with other inflation markets. Inflation returned to within the upper inflation band in October 2009. OER has a weight of 12. The most recent CPI basket reweighting took place in January 2009. This information is weighted according to the population census figures in order to represent all households spending patterns in South Africa accurately.124 to 386. citing confidence that inflation would decline over the medium term and underlining a policy commitment to low and stable inflation. but over a longer period the new series is not expected to have a significant effect on measured CPI. The frequency of amount of supply has increased significantly. until the results of the next Survey of Income and Expenditure of Households become available. launching the CPI-linked 6.

The first day of each month has a reference index equal to that of the CPI of four calendar months earlier – eg.25% Mar ‘13).2 for government nominal bonds. These two indices provide an index ratio for the value date: Index Ratio Rate = Reference CPIt/Reference CPIBase For settlement amounts. This understates the importance of linkers. Basically. South African CPI-linked bonds fall under section 24J of the Income Tax Act of 1962. with the R197 principal having increased to 1. which allows the SARB to lend up to ZAR1bn of inflationlinked bonds to the market for fixed periods. to improve liquidity. the R197. In addition. Reference indices for intervening days are calculated by straight-line interpolation. for 1 December 2009. versus ZAR72. and the largest linker issue. albeit from very low levels. However. the government announced in the 2010 Budget that two new ILBs maturing in 2017 and 2022 will soon be introduced. the US or the euro area. this is due to the small size of the market: linkers make up only 23% of the notional amount of government bonds issued.Barclays Capital | Global Inflation-Linked Products – A User’s Guide currently four bonds: R189 (6. In part. a reference CPI is then calculated. The nature of the investor base is also an important limiting factor for liquidity as linkers tend to be bought by buy-and-hold domestic institutional investors. In 2009. with inflation indexation calculated using a slightly augmented Canadian methodology.45% Dec ‘33)). is only ZAR28. The 15 March 2010 148 . and likewise for the par redemption amount (with the cash value subject to the par floor). real accrued interest is calculated as for ordinary South African bonds.67 times its original size by the start of 2010.8bn. This means that South African linkers have a lag that is a month longer than those in Canada. given that linkers have accrued significant inflation. R197 (5. however. For settlement on the first day of any calendar month. Calculations South African government inflation-linked bonds carry a principal deflation floor and are quoted on a real yield basis. Dirty price and accrued are each multiplied by the index ratio to arrive at a cash settlement amount. This formula is used to calculate a reference CPI for the official original issue date. and R202 (3. rather than by trading accounts. the CPI from four months previous is the reference CPI for that date.6% Mar ‘28). versus 27. Liquidity in the market is improving. with a 10-day “books closed” ex-dividend period. Each day has its own distinct reference index. the (real) semi-annual coupon rate is multiplied by the index ratio. For settlement date or cash flow payment date “t”. the government introduced a reverse repo facility in June 2006. R210 (2. Taxation South African inflation-linked bonds pay interest on a semi-annual basis.2. Both the reference index and the base index are rounded to 15 decimal places.5% Dec ‘23). as issuance has increased and real yields rise. there may be an increase in interest from foreign participants. Interest on bonds is taxed on a yield-to-maturity basis. the R186. including hedge funds. For coupons paid. the CPI is for August 2009 and for 1 January 2010 the CPI is for September 2009. inflation-linked bonds are treated like floating-rate instruments. making investment relatively straightforward. However. international investors are not subject to withholding tax. Since demand for inflation linked bonds has been strongest in the short end of the curve.0bn for the largest nominal issue. or “Base Reference index”. the turnover ratio (total turnover divided by the nominal outstanding issue of bonds at year-end) for government linker bonds was 3. Coupon payments and the difference between the acquisition cost and the nominal value of the bond are defined as interest and are liable for income tax. For domestic investors.

much of the issuance has occurred most recently. Much of the bank issuance remains driven by tailored. as banks became willing to facilitate corporate hedging in the form of real rates and pension fund investment via annuities. they are not a regular feature of the market. Parastatal (state-owned companies) inflation-linked issuance remains disappointing. liquidity in the swap market has dried up. Non-government bonds As of December 2009.5bn in new paper came to market from the likes of ACSA (ZAR500mn) and Eskom (ZAR1bn). with an outstanding value of ZAR36. with the aid from the SARB’s reverse repo facility enabling investors to take on short positions in the bonds to some degree. Even though the size of underlying flows has grown significantly as the curve has become better defined. In 2009. ranging from 10-20bp. There is less of a structural constraint on taking long/paid real yield positions in the swap market – but the repo market is developing. The sector’s reluctance to issue inflation seems likely to continue into 2010 despite the significant infrastructure requirements that need to be funded over the medium term. bringing their total outstanding figure to ZAR21bn. 15 March 2010 149 . with cash bonds quoted relatively cheaply. with the former being more frequently traded. inflation pricing in the derivative and cash markets moved significantly out of line in late 2008. While the swaps market does not have the consistent weekly liquidity injection of auctions. Though the first issue in the non-government market dates back to 2001. Real rate and inflation derivatives The inflation-linked derivatives market has only taken off since 2003. Currently. the Johannesburg Stock Exchange listed 54 non-government inflation-linked bonds. with more frequent trading in the bonds. While a few option trades have occurred. allowing swaps to benefit from the improved liquidity of the expanding bond market. bid-offer spreads remain quite wide. it quickly developed into a more liquid market than bonds. ZAR1. As in other markets. responsible for ZAR14bn in inflation-linked debt. as the swap market was mainly an inter-bank market rather than a buy-and-hold one. This prompted the first significant volume of asset swap buying of government linkers. Among concerns such as large refinancing risks and cost effectiveness. the fact that the market remains highly illiquid seems a key deterrent (about 3% of issue size was traded over 2009). it consists mainly of zero-coupon real-rate swaps and zero-coupon breakeven inflation swaps. Banks were the largest issuers in this market during 2009. When swap trading began. taking into account any adjustments in the principal amount and the coupon payments as a result of changes in the CPI.Barclays Capital | Global Inflation-Linked Products – A User’s Guide tax liability is determined annually. Liquidity is notably worse than the nominal swaps market. these do create swap market activity if there is asset swap flow. however. Recently. Asset swap trading has continued subsequently.7bn. it is still mostly a market between five market-making banks. For this reason. structured notes. with 2009 seeing total outstanding issuance double.

7.5 2. 442.0 2. Although this has the potential to push real swap yields higher. The local pension fund industry’s liability driven investor flows will typically drive real yield swaps lower relative to government real yields.9. increased government ILB supply could have the same effect on bond yields as well. 77% R197 (2023) R210 (2028) R202 (2033) Source: Barclays Capital 1.5 R189 (2013) Source: Barclays Capital Real bond-swap spreads are currently trading at historically wide levels.5 ZAR IL bonds % ZAR real yield swaps % Figure 119: Composition of the SA government bond market FY 2009/10 Rbn ILBs. Figure 120: ZAR linkers – historical performance and risk 35% 30% 25% 20% 15% 10% 5% 0% 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: Barclays Capital Figure 121: Return/risk versus nominals and equities 7 6 5 4 3 2 1 0 -1 -2 2001 2002 2003 2004 Source: Datastream. leading to increased interest from institutional investors to pay inflation linked swaps and thereby take advantage of the wide spreads on offer. while the lack of significant development in the corporate side of the market makes the market rather one-sided. Barclays Capital South Africa IL Returns South Africa IL Ann Mthly Vol Equity IL Nominal Bonds 2005 2006 2007 2008 2009 15 March 2010 150 . 23% 3. 130.0 Nominal bonds.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 118: ZAR real bond and swap curves 3.

a doubling of housing prices (20% of index). a town.7% of government debt.0 Dec-99 Housing. During 2007-09.9% per annum. first auctioned on 18 August 2004.1% since 2004. The main factors accounting for this increase were a tripling of food and beverage prices (30% of the index.0 6. and built up by 2008 to about PLN9. Market development The Polish linker market has grown slowly.0 0. but it has since gained momentum and averaged 4% during 2007-09. domestic investors were illplaced to deal with inflation-linked bonds and the bulk of the issue was place outside Poland.000 individual prices are collected each month.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Poland Daniel Hewitt +44 (0) 20 3134 3522 daniel. and a doubling in energy prices (11% of the index).0 4. global trends in food and commodity prices – have a large impact on the path of inflation. Clearly.1bn outstanding. which is published by Poland’s central statistical office around the middle of the following month. Fruit and vegetables are priced three times per month.0 Apparel. 19% Dec-01 Dec-03 Dec-05 Poland CPI Dec-07 Dec-09 Source: Haver Analytics.com Poland inflation-linked bonds comprise a small market.com Arko Sen +44 (0) 20 3134 2839 arko. The weightings are revised annually but importantly. a 3. 30% Figure 123: Inflation in Poland 12.0% 2016 bond. 1% Entertainment 14% Transport and Communication 14% Health. CPI is calculated using the data gathered from 209 research regions (ie. It is unlikely to be tapped again. inflation averaged only 1. factors outside the control of the MPC – eg.5% +/. This is understandable as inflation was very low during 2002-06. albeit from 33% three years previously). Barclays Capital 15 March 2010 151 . Since 1990. inflation doubled to 3. this will depend on global trends in prices. Prices of the other goods and services are collected once a month.9% per annum. 5% Others. or a part of the large city) across the country. About 292.800 consumer goods and services are surveyed between the first and the 25th day of each month. Barclays Capital Source: Bloomberg. Poland adopted inflation targeting in 1999. with only one significant sovereign issue. The index is dominated by the 30% weight of food and beverages.hewitt@barcap. During 2002-06. the statistical office only publishes m/m and y/y changes in CPI – not the underlying index.sen@barcap. The MPC pursues its inflation targets seriously and expects to keep inflation within the inflation target band. the weighting system has been based on a household budget survey. 5% 2. The government has only issued two bonds – the first in 2004 and the second in 2008 – and they only account for 1. 5% House maintenance and equipment. At time of issue. The prices of 1. CPI Inflation-linked bonds are referenced off the non-seasonally adjusted m/m CPI inflation. and unusually. The NBP (National Bank of Poland) inflation target has been 2.0 10. Figure 122: Poland’s CPI basket Education. averaging just 1.9% per annum.0 8. 5% Food and Beverages. However.

but there has been little progress here.0 -0. the majority of pension funds will only start to make payments in 2010.5 -1. The issuer’s original intention was to provide an inflation-linked asset to the growing pension industry. so the month-on-month CPI changes are combined in order to calculate an index that is based to July 2003.0 -1. real accrued interest is calculated with reference to the index ratio. There are several possible factors behind this.5 Equity IL Nominal Bonds Poland IL Returns Poland IL Ann Mthly Vol 2006 2007 2008 2009 -2. which a change in regulation could alter. thus the outlook for trading liquidity looks poor in the near term. with linear interpolation during the month. there were euro-structured notes with the government linker as the underlying asset.5 0. there have been no issues of corporate inflation-linked bonds. which might change. although in the past when the bond traded cheaply. although the Polish Ministry of Finance has yet to announce any formal changes. The only difference from standard Canadian methodology is that there is no CPI Index.0 2006 2007 2008 2009 Source: Thomson Datastream.5 1. Barclays Capital 15 March 2010 152 .75% of 2023. First. In 2010. Trading volumes in the sovereign linker market are low (even compared with other EM linker markets). initial regulation of the new Polish pension funds encouraged total returns-focused investment strategies. having first emerged from the government reforms of the previous “pay-as-you go” pension system in 1999. with inflation indexation calculated according to the Canadian methodology. and the amount outstanding of the 2016 bond has been eclipsed by other new entrants to the EM linkers market. As a result of the modest growth in this market. Local pension funds have about USD60bn of assets under management and have grown quickly. These pension funds hold very few inflation-linked bonds.0 1.0 2. the implied CPI Index from three months previously is the reference CPI for that date. For settlement amounts. with coupons paid annually. there are plans for only very modest linker issuance. although further developments in the pension fund industry could bring back government interest in this market over the medium term. The reference coefficient for month “T” is published to five decimal places and in the case of the August 2016 linker bond is 100 times the product of the monthly changes. Figure 124: Polish linkers: historical performance and risk 14% 12% 10% 8% 6% 4% 2% 0% 2005 Source: Barclays Capital Figure 125: Return/risk versus nominals and equities 3. For settlement on the first day of any calendar month. Second.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Since then only one new series has been started – the 2. which was first auctioned in July 2008 but has not been tapped since and only has PLN 450mn outstanding.5 2. suggesting that hedging inflation liabilities could gradually become a more pressing concern. Calculations The Polish government’s inflation-linked bonds carry a principal deflation floor and are quoted on a price basis.0 0.

the upper tolerance limit was 6%. The government’s rationale for re-defining the index as the CPI was for financial indexation that was consistent with the practice of other countries. Barclays Capital 15 March 2010 153 .james@barcap.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Iceland Alan James +44 (0) 20 7773 2238 alan. In its November 2009 inflation projections.5% before 2012. Inflation peaked above 18% at the start of 2009 following the sharp depreciation of the krona through 2008.5% at the end of 2002). Mortgage debt was partially indexed as long ago as 1955.9% two years earlier. rebased as of March each year and published monthly. though this fell from 17. but the market is dominated by the Housing Finance Fund mortgage agency. and it has a +/. at 14. allowing publication late in the month. The Central Bank of Iceland adopted a monetary policy inflation target of 2. Barclays Capital Figure 127: Icelandic inflation targeting has been unsuccessful 20 18 16 Clothing and footwear 6% 14 12 10 8 CPI Central Bank target Upper tolerance bound Lower tolerance bound Food. Icelandic Consumer Price Index (CPI) Iceland has a long history of indexation. drink and tobacco 18% Housing and utilities 24% 6 4 2 0 Mar 01 Mar 03 Mar 05 Mar 07 Mar 09 Source: Statistics Iceland.com Iceland issued its first government inflation-linked bond in 1964. and that would appear more permanent and credible. The consumer price index is a chain-weighted index. Figure 126: Components of Icelandic CPI Hotels and Others 8% restaurants 5% Health and recreaction 4% Communica tions 4% Transport 13% Furnishings 8% Source: Statistics Iceland. with the largest single component for imputed rents. Icelandic inflation is particularly exposed to currency movements. the Central Bank did not envisage CPI inflation dipping below the 4% upper tolerance bound before Q4 10 or below 2. Price collection since 2008 has taken place in the middle week of the month. particularly after the launch of the Swedish linker market in 1994. having previously been linked to a broader cost of living and prior to 1980 to building costs. The Central Bank’s inflation target takes priority over all other objectives. the challenges of inflation targeting for such an open economy were highlighted by inflation moving above 8% in mid-2006 following currency weakness.1% as of March 2009.5% tolerance band (at initial adoption. housing is the largest sector. consistent with Eurostat conventions. with 37% of the index made up of imported consumer goods at the 2009 reweighting. but it is only since 1995 that Icelandic linkers have been tied to CPI.1. with the market consolidated in 2004 into benchmark bonds that are similar to other linker markets. falling to 4. However.5% on the CPI in March 2001. As in many CPI indices. Central Bank of Iceland.

Given ongoing pension demand and relatively limited HFF issuance. Since 2004. Registered foreign investors are exempt from withholding tax on the issues. HFF has issued bonds dated 2014. with semi-annual coupons and the same 30/360 day count convention as for Swedish linkers. holding 65% of HFF benchmarks in Q4 09. indeed. The HFF was established in 1999. As of the end of November 2009. International investors became significant participants in HFF benchmarks in the middle part of the decade. when it assumed all the assets and obligations of its predecessor. though this has been offset by the increase in the uplifted size of the market due to inflation. 2034 and 2044. with a market cap of almost ISK800bn (~US$6. they briefly spiked richer in October 2008. 15 March 2010 154 . though the shortest issue is no longer re-opened and is now notably smaller than the other bonds.5bn).3%) when foreign selling was predominant. with the reference CPI at the start of 2010 more than 50% higher than when the benchmarks were issued during 2004. Pension funds are easily the largest investor base. the State Housing Board. and almost all of its debt is now CPI-linked. In July 2004. 2024.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Icelandic CPI-linked bonds Indexation remains widespread throughout the Icelandic domestic bond market. hence. shares the same ratings as the government: BBB+ from S&P and Baa3 from Moody’s as of the start of 2010. the Housing Finance Fund (HFF) is a notably larger issuer of debt than the government. with over half of all bonds linked to the consumer price index as of Q4 09. it consolidated most of its debt into new bonds that have a very similar format to those in the major inflation-linked markets. there were ISK526bn notional of benchmark HFF bonds. Issuance has slowed in recent years due to the weakness of the domestic housing market. While there is only one notable government inflation linked note. from as low as 63% two years previously. with domestic institutions holding over 80% of benchmark HFF bonds by Q4 09. Central Bank data show their importance has fallen notably. despite not having explicit inflation-linkages within their liabilities. it would not be surprising if the government were to issue inflation linked bonds directly in the coming years. This ongoing domestic demand meant the real yield of the 2044 stayed below 5% through the worst of the financial crisis. with the highs in real yield well before this in December 2007 (at 5. HFF is a sovereign-backed agency and. However. a zero coupon issue redeeming in 2015.

These changes largely reflect the national shift from agriculture to heavy industry in the 1970s and 1980s.0 -1. CPI The reference index for computing the inflation adjustment factor for the first South Korean linker is the South Korean Headline Consumer Price Index (CPI).0%. in terms of the 12month rate of change in the consumer price index. Although the CPI has been used as a measure of South Korean inflation since 1936. there is only one issue. using a standard Canadian format. which takes the form of a questionnaire. before the growth of the service sector in the 1990s. which has not been re-opened since July 2008. The issue is linked to headline Korean CPI. albeit without a par floor on the principal.2%. The central bank targets inflation on both the headline and core CPI measures.0 3.000 price quotes are obtained each month from 13. the KTBi Mar 2017. and the quotes are obtained by personal interview.0-4. Since 1998.0 5.huang@barcap. released on 1 February 2010. with a current target range of 2. For instance.9% in the last 10 years – with a standard deviation of 1. inflation receded from 6.2% in the last 30 years to just 2.com Matthew Huang +65 6308 3093 matthew.0 May 07 Nov 07 CPI % y/y May 08 Nov 08 May 09 Nov 09 BoK policy bad Breakevens to Bonds Breakevens to Swaps Source: Korea National Statistic Office Source: Bloomberg. Approximately 80.0 Cooked food 13% 4. with a tolerance range of plus/minus 1 percentage point around this target. Bank of Korea. and set the target in consultation with the government.0 0. eg.0 6. Since inflation targeting was institutionalised in 1998. The inflation target for 2010 onward is 3 percent. The major categories of goods and services and their weights are shown in Figure 128. As of January 2010. inflation in Korea has since been better contained and more stable.0 2. There are plans for resumed issuance starting June 2010.Barclays Capital | Global Inflation-Linked Products – A User’s Guide South Korea Wai Ho Leong +65 6308 3292 waiho.000 outlets. Barclays Capital 15 March 2010 155 . the headline CPI for January 2010.The Korean National Statistics Office publishes CPI data each month. there have been considerable changes in South Korea since its inception. the Bank of Korea has applied inflation targeting.0 1.com South Korea issued its first inflation-linked bond in March 2007. Previously the tolerance band was plus or minus 0. which is not seasonally adjusted. The Bank reviews the performance of inflation targeting annually and makes a public announcement of the results. The target horizon is three years (2010~12). The index Is revised every five years to reflect the changes in consumption structure of Figure 128: Breakdown of CPI by major category Recreation 6% Health 5% Education 11% Clothing 6% Housing 23% Transport 17% Misc 5% Food 14% Figure 129: CPI inflation and the breakeven of KTBi 7. The CPI is measured monthly by the Consumer Price Survey.5%. usually the first working day of the following month.leong@barcap.

For reference. we expect the priority for developing the linkers market will again come to the forefront of government priorities. The indices have been revised on a five-year basis to reflect the changes in consumption structure of urban households.77trn outstanding of the bond. market liquidity for the sole outstanding issue remains poor. less than 1% of the total outstanding government bond market. with minimal trading in secondary markets. Investment regulations will also help demand for Korean bonds. Government bonds Market development The government first issued inflation-linked bonds in March 2007. In April 2003. limiting the amount of secondary trading. but was postponed due to the financial crisis. the Nationwide Chain Index as of 2002 as a supplementary index was compiled and released. The government continues to express a commitment to expand its bond markets including the KTBi. Furthermore. when it launched the CPI-linked 2. For settlement on the first day of any calendar month. the financial crisis has led to a pause in new issuance of linkers. however. Currently there is KRW1. The last monthly auction was held in July 2008 for KRW61bn (c. with KRW900bn of the existing issue bought back by the Treasury in Q4 08. known as the KTBi. Investing in Korean linkers has been helped by the removal of withholding tax for foreign investors and the addition of Korean bonds into Euroclear’s settlement systems in 2009. In December 2006. The addition into Euroclear is still at an early stage with limited trading at present. bid-ask spreads for linkers are 5-10bp compared to 1-3bp for 5y and 10y KTBs.75% March 2017. the CPI from three months previously is the reference CPI for that date. At present. but with time this is likely to improve accessibility tremendously as any counterparty with a Euroclear account would have access to trade the bonds. The reference CPI for any day during the month is calculated by linear interpolation. A reference CPI value is calculated for every day based on the CPI values for three months and two months prior to the month containing the settlement date. We anticipate that liquidity will improve upon the government’s opening of a new issue in June 2010. the results of revised indexes were released with 2005 prices based. with no floor on the principal value. The investor base has tended to be foreign buy-and-hold institutional investors rather than trading accounts. As economic conditions in Korea and globally continue to stabilise. The finance ministry stated in its 2010 issuance strategy that it intended to resume inflation-linked issuance in June 2010. reducing the need for an onshore custodian account and reducing associated paperwork for foreign investors. In January 2002. the results of revised index as of 2000 were released. leading to continued support for the nascent Korean linkers market. heightened concern over rising inflation following historical amounts of monetary easing through 2009 could also provide demand for inflation-linked products auctioned in June.Barclays Capital | Global Inflation-Linked Products – A User’s Guide urban households. with a view to expanding the Treasury bond market base to resident investors and foreigners.USD61mn as of July 2008). A new issue was scheduled for September 2008. 15 March 2010 156 . Calculations South Korean government inflation-linked bonds are quoted using the standard Canadian model.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide Reference CPI for day d of month t: (d − 1) (CPI t −2 − CPI t −3 ) + CPI t −3 m d = day of the month eg.8%. Taxes were not removed for Korean resident investors and a 15.5%. much of the bond was absorbed by international investors via total return swaps (TRS). Total return swaps and structured notes An inflation swaps market has yet to develop in Korea. January 2010 break-even levels are 2.4% withholding tax on interest income and 24. Coupons are accrued on an actual/actual basis and paid semi-annually. Break-even levels are now well within the revised BoK 2010 inflation band of 2-4%. 15 March 2010 157 . however. pushing break-even inflation below 1% at the end of October for about a week. the breakeven was at the lower end of the 2. At present. Prior to the tolerance band revision. When the KTBi was first issued. As of 21 May 2009. The change was made amid concerns of insufficient demand for KTBs to fund Korea’s supplemental budget. there is now a differentiation between taxes passed for resident and non-resident investors. but most international transactions have been derivative related while cash buyers of the KTBi often do so in breakeven versus nominal swaps.4% withholding tax and a 24. we believe most investors of the inflation-linked bonds remain unchanged from the initial TRS investors as secondary trading is limited. which meant settlement and registration issues. withholding and capital gains taxes were assessed subject to levels determined by tax treaties. This too is a change as financial institutional investors were exempt from withholding tax prior to 1 January 2010. Standard rates of 15. The removal of taxes was made in a bid to increase foreign participation in Korean Treasury markets. Through Q4 08. 1st implies d=1 m = number of days in that month The indexation factor is the reference CPI for the settlement date divided by the reference CPI for the base date. Prior to the change. compared to headline inflation of 2. Taxation The tax treatment for the KTBi is the same as that for KTB. This period marks the low for break-even levels as they have slowly increased back into the BoK’s policy bands. demand and liquidity for linkers fell.5-3.2% on capital gains were assessed if not specified in a treaty otherwise. non-resident investors were made exempt from taxes on interest and capital gains taxes through a tax law change by the Korean parliament.5% BoK band through the end of 2009.2% capital gains tax applies.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide INFLATION INDICES 15 March 2010 158 .

) Euro Inflation-Linked (Government + Non-Government) Sterling Inflation-Linked (Gilt + Non-Gilt) Emerging Markets Government Inflation-Linked Universal Government Inflation-Linked (WGILB + EMGILB) US TIPS (Series-L) Comparator bond indices (a family of indices consisting of the nominal comparator government bonds for each index-eligible inflation-linked bond) Tradable inflation-linked index products Tradable inflation-linked index products are available for investors seeking efficient access to beta. Australia. Barclays Capital Indices are used on a daily basis by portfolio managers. enhanced beta or alpha in index form. Japan. product development.com The Barclays Capital Inflation-Linked Index offering includes market leading benchmark bond indices and tradable inflation-linked index products that are designed to offer efficient market access and/or rules-based alpha generation in index form.myers@barcap. but in many cases they are designed to offer exposure to a particular market segment or investment strategy in objective rules-based index form.harman@barcap. both as performance targets for global investors and informational measures of security. hedge funds and retail investors. In some cases these products can also be used as benchmarks. sovereign wealth funds. etc. sector. consultants. Euro. UK. pension funds.com Brian Upbin +1 212 526 6981 brian. researchers. and marketing groups. commercial banks. France) Real rate and zero coupon TWIST – Treasury-weighted inflation swap exposure across the curve 15 March 2010 159 . Within such firms. Our family of tradable index products include: Breakeven Inflation Index products US breakeven inflation indices INSTEP Inflation swaps (UK. operations.com Jenna Myers +1 212 526 0500 Jenna. traders. insurance companies. performance analysts. US.com Anand Venkataraman +44 (0) 20 7773 0852 anand. US.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Barclays Capital inflation-linked indices Scott Harman +44 (0) 20 7773 1775 scott. Barclays Capital is the leading index provider of both inflation-linked and broad-based fixed income benchmark indices. risk managers. Users of our inflation-linked indices span a wide variety of investor types including asset managers. Benchmark indices Benchmark indices are a critical component to the portfolio management process.venkataraman@ barcap. Japan.upbin@barcap. Flagship indices in the benchmark family include: World Government Inflation-Linked Bond Index (including benchmark sub-indices for each market: Euro. offering useful and well designed benchmarks for both dedicated inflation portfolios and other multi-asset managers. country and market performance.

Bond and index analytics including forward yields. Forward index information with projected index constituents at the next rebalancing date available for the majority of indices to help investors manage against their benchmark. Real time indices and analytics for the US. Portfolio & Risk Solutions team for more details or for requests for customised indices. carry and an interactive carry calculator. Accessing inflation-linked index data All of the indices profiled in this handbook are available on Barclays Capital Live (keyword: indices). breakeven. we offer overviews of our flagship inflation-linked indices (benchmark and other index products). the Barclays Capital portfolio and risk analytics platform Barclays Capital inflation-linked indices are also available via Bloomberg. euro area and UK markets. with details on specific index rules and return calculations. POINT®.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Other tradable Inflation Index products AIMS – Algorithmic Inflation Momentum Switching (AIMS) strategy INSPIRE Real income In this section of the publication. and a growing number of other third-party data providers. which provides access to a range of inflation-linked bond resources including: A complete database of bond and index returns in local currency plus returns hedged and unhedged in most major currencies. Please contact the Index. 15 March 2010 160 .

EGILB and EILB. and has covered the expansion of the international inflation-linked bond markets with successive benchmark index launches as either stand-alone indices or additions to existing flagship indices. France. The most notable expansions were the launch of the Emerging Markets Government InflationLinked Bond Index (EMGILB) and the Universal Government Inflation-Linked Bond (UGILB) Indices. US. broader benchmarks for the euro and sterling markets that include both government and non-government linkers. Euro and Euro Govt Inflation-Linked Indices following ratings downgrade to BBB+ Barclays Capital BESA South Africa Government Inflation-Linked Bond Index is rebranded as Barclay Capital / ABSA South Africa Government Inflation-Linked Bond Index and adopts ABSA capital market makers pricing for index calculation Universal and EM government inflation-linked bond indices are launched Inclusion of Greece into GILB. EGILB and EILB Japan joins GILB Addition of investability rule to bond criteria. Currently. World Government Inflation-Linked Index launched with a base date of 31 December 1996. Amendment of the market aggregate face value and bond minimum issue size requirements in relation to the application of FX rates Aug 2003 May 2003 Jan 2001 Oct 1997 Italy joins GILB. The founding markets included the UK. Gilt and Non-Gilt indices launched. In addition to these traditional bond indices. the Barclays Capital Inflation-Linked Index family tracks 19 countries and over 200 inflation-linked securities.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation-linked benchmark indices Introduction The Barclays Capital Inflation-Linked benchmark family includes flagship government benchmarks for developed and emerging markets. Figure 130: Benchmark Inflation Index family timeline Effective date Dec 2009 Feb 2009 Greece excluded from World Government. Australia and Canada Oct 2007 Jun 2006 Mar 2006 Apr 2005 Jan 2004 Source: Barclays Capital 15 March 2010 161 . other inflation-linked index products that can be used as benchmarks include breakeven inflation indices that measure the total return of a breakeven inflation position and inflation-swap indices that can be used for Liability Driven Investing (LDI) indices. Index methodology changed such that the returns are calculated month-to-date and intra-month coupons are held in cash and accrue at the local market monthly deposit rate Germany joins GILB. and comparator indices that consist of the nominal comparator bonds for government linkers. EGILB and EILB Euro and euro government inflation-linked indices launched to accommodate the growing Euro Inflation-Linked market Sterling All Bond. Sweden. Barclays Capital launched its first global benchmark index for this asset class with the World Government Inflation-linked Bond Index (WGILB) in October 1997.

Sweden.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Overview of flagship benchmark indices Figure 131: The Barclays Capital Inflation-Linked Government Bond Index Family Euro Sterling Universal Govt World Govt EM Govt Greece Govt Euro NonGovt UK NonGovt UK Govt Euro Govt AU Govt CA Govt JP Govt SE Govt US Govt Latam Govt EMEA Govt Asia Govt Greece Govt Source: Barclays Capital World Government Inflation-Linked Bond Index (WGILB) The Barclays Capital World Government Inflation-Linked Bond Index (WGILB) measures the performance of the major government inflation-linked bond markets. Sub-indices are also available that cover securities linked to EMU HICP and FRCPI inflation indices. France. whereas quasi-government and corporate debt is ineligible for the EM Govt Inflation-Linked Bond Index. 15 March 2010 162 .45trn and includes over 90 issues from nine markets. The Barclays Capital Euro Government Inflation-linked Bond Index covers euro-denominated debt from euro area member governments linked to domestic or EMU HICP inflation indices. Italy. hence there are indices for each individual country and maturity bucket within each country. inflation-linked issues from the EMU form a major part of the World Government Inflation-linked index. Currently.. Canada. UK. Markets currently included in the index. Euro Inflation-Linked Bond Index The Euro Inflation-Linked Bond Index combines euro government and non-government inflation-linked bonds and captures the evolution of the linker market in the euro area. Sweden. The index includes only government debt. The index is designed to include only those markets in which a global government inflation-linked fund is likely and able to invest. Japan and Germany. EM Government Inflation-linked Bond Index The Barclays Capital Emerging Markets Government Inflation-Linked Bond Index (EMGILB). i. Italy.e. Germany and Australia. France. Canada. have a face value of €500mn or more and are issued by an EMU member government or a quasi-government issuer. Bonds are eligible for this index if they are linked to the inflation of any EMU member country or the harmonised EMU HICP are denominated in euro. the index includes government bonds from France (linked to FRCPI and EMU HICP Ex-tobacco). measures the total return performance of inflation-linked bonds from the major emerging markets. The Euro Government Index market capitalisation is now over €270bn ($395bn equivalent) comprising 20 bonds with maturities out to 2041. Taken together. are the UK. the US. In size order these are: the US. Japan. direct obligations of the state issuer. launched in October 2007. The index currently has a market capitalisation of more than US$1. Australia. in order of inclusion. The structure of the index is a granular one. Italy and Germany.

Germany. The Barclays Sterling Inflation-Linked Bond Index was launched in January 2001 with a base date of 31 December 1999. in order of size – the US.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Sterling Inflation-Linked Bond Index Demand for long-dated inflation-linked assets from the UK pension industry led to quite spectacular growth in the sterling corporate inflation-linked bond market in the late 1990s and during much of the last decade. Universal Government Inflation-linked Bond Index The Barclays Capital Universal Government Inflation-Linked Bond Index combines markets from the World and EM Government indices in a single index designed to measure the performance of the major developed and emerging government inflation-linked bond markets.7trn and includes inflationlinked government bonds from 19 countries. The index combines bonds from the Gilt and non-Gilt Inflation-Linked bond markets with a cut-off of £100mn face value for non-gilt issues. South Africa. Poland and South Korea. Chile. The Non-Gilt portion of the index is further divided by rating and by sector. Argentina. Sweden. Australia. The index presently has a market capitalisation of over US$1. 15 March 2010 163 . Italy. Turkey. Canada. Japan. France. Greece. UK. Mexico. Colombia. Brazil.

Sweden (1994). Italy (2003). This index measures the performance of the major government inflation-linked bond markets and is designed to include only those markets in which a global government linker fund is likely and able to invest. Australia (1985). from 2% in September 1998 when France first issued. This is partly due to inflation accretion but by far the most important factor is the emergence of new inflation-linked bond markets. Japan (2005) and most recently Germany (2006). but due to a ratings downgrade in December 2009 became ineligible (Greece remains as part of the Universal Government Inflation-Linked Bond Index – see Figure 131). Investability is therefore a key criterion for inclusion of markets in this index. The share of the euro area has grown rapidly. Canada (1991). the index was dominated by issuance from the UK. Greek Government Inflation-Linked bonds became eligible for both the Euro and World Government Inflation-Linked Indices on 30 June 2006. the Barclays Capital World Government Inflation-Linked Bond Index (WGILB). The total market capitalisation of the index has grown more than ten-fold since the launch of the Index in December 1996. Markets currently included in the index in the order of their inception are the UK (1981). France (1998). with the euro government share of the World Index reaching around 27% by December 2009. Heavy issuance elsewhere has now reduced the UK share to around 23% and we now see the US as the dominant market with around 39% of the index. which at that time accounted for approximately 80% of the index by market value. Figure 132: Index market value by country 1600 1400 Market Value ($bn) 1200 1000 800 600 400 200 0 Dec-96 Australia France Italy Sweden US Canada Germany Japan UK Greece Figure 133: Index weights by country 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Dec-96 Dec-99 Dec-02 Dec-05 Dec-08 Germany UK Dec-99 Dec-02 Dec-05 Dec-08 Australia Italy US Source: Barclays Capital Canada Japan Greece France Sweden Source: Barclays Capital 15 March 2010 164 . At inception. in October 1997.Barclays Capital | Global Inflation-Linked Products – A User’s Guide World Government Inflation-Linked Bond Index Overview Barclays Capital launched its first inflation-linked bond index. the US (1997).

Bonds are allocated to maturity bands based on their remaining time to maturity. New bonds and taps/increases entering the index must have settled on or before the rebalancing date. AUD700mn. The most up to date set of rules and related notifications can be accessed at www. Min issue size Minimum issue sizes are set in local currency terms and may be reviewed annually by Barclays Capital taking into account local market conditions. Bonds must be capitalindexed and linked to a commonly used domestic inflation index. The face values are used rather than an inflation-adjusted value. Maturity bands are inclusive at the lower bound and exclude the upper bound. Minimum rating for inclusion is A3/A. EUR500mn. All spot and forward foreign exchange rates used are official WM Company mid rates from the London market at 4pm. The index uses standard settlement conventions for all calculations. The index is calculated daily and has a value for each calendar day. Underlying Market/bond type Maturity and issue Index calculation and monthly review Pricing methodology Settlement conventions Index frequency and reinvestment The index uses mid-market prices from local market close. domestic inflation indices and the EMU HICP are eligible. CAD600mn. Reference deposit rates are set on the last business day of the month for the next month. Income received during the month is put on deposit until the month-end rebalancing. Market calendars most appropriate for international investors are used. and USD500mn.for G7 and Euro-zone markets and Aa3/AA. In the euro area.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 134: World Government Inflation-Linked Bond Index inclusion criteria and review policy 12 Criteria for inclusion in the index Inclusion of markets in the World Govt Index The WGILB Index is built bottom up from a selection of country/currency indices based on rating and size. 12 15 March 2010 165 . The following minimum issue sizes presently apply.live. JPY50bn. and will be reviewed annually. Market size: Minimum aggregate issuance (non inflation-adjusted) for inclusion of bonds from a currency zone is $4bn.com (keyword: index).barcap. The index holding of each bond for the next month is set to the amount outstanding on the review date. These are in force for a full year. SEK4bn. Government domestic inflation-linked debt issued in domestic currency of that country only. Review procedure Maturity bands Index holdings The index is reviewed and rebalanced monthly on the last calendar day of each month. Limits are set in local currency terms using WM closing spot rates from the last business day of the year. Rating: Long-term local currency ratings from S&P and Moody’s are considered. and the lower of the two is used for the index. Reference rates are 1M LIBOR – 15bp (if LIBOR is not available in the market currency a local equivalent may be used – see Figure 24). Index rules are periodically reviewed by Barclays Capital Index research through formal index governance procedures. The indices are weighted using the market capitalisation as standard. issuance trends and movements in exchange rates. at which point it is reinvested back into the index. GBP300mn. Bonds must settle on or before the rebalancing date and have a minimum remaining life of at least one year. Issues that are not available in whole or part to international investors are not eligible for the index. The notional coupon of a bond must be fixed or zero.otherwise. effective at the end of the first quarter of the next year.

3-5y. Local and foreign currency indices The World Govt index and other composite indices are available as an aggregate local currency index.82% 39. 10-15y. World Ex-US.23 126. 5-15y.70% 2.24 565. CHF. 5-7y.40% 22. except for bonds with less than a year to maturity. Maturity buckets The standard maturity breakdowns consists of All Maturities. CAD. 7-10y.92 45.43 108.21 244. 1-10y. >10y and >15y.49 32. USD and ZAR . >5yr.91% 15.72 67.72 215.37% No of issues 96 3 5 21 11 6 5 3 7 16 5 17 29 AAA AAA AAA AAA AA AA AAA AAA AAA AAA AAA Index rating Base date World Govt Australia Govt Canada Govt Euro Govt France Govt FRCPI-Linked EMU HICP-Linked Germany Govt Italy Govt Japan Govt Sweden Govt UK Govt US Govt Source: Barclays Capital Price source 31-Dec-96 31-Dec-96 31-Dec-96 31-Dec-99 30-Sep-98 30-Sep-98 31-Oct-01 31-Mar-06 30-Sep-03 31-Mar-04 31-Dec-96 31-Dec-96 28-Feb-97 15 March 2010 166 . As of 31 December 2009.31 107. SEK. 1-3y. Foreign and hedged versions are also calculated in the following base currencies.86 Barclays Capital Barclays Capital Barclays Capital Barclays Capital Barclays Capital PMI Exchange Barclays Capital Barclays Capital 79.00% 26. GBP.81% 3.52 34.08% 8. AUD. SGD. 510y.12 44. Figure 136: World Government Inflation-Linked Index characteristics (as of 28 February 2010) Market cap (local bn) 1436.37 78.68 565.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Index structure Figure 135: World Government Inflation-Linked Bond Index structure World Govt Australia Govt Canada Govt France Govt Italy Govt Germany Govt Sweden Govt Japan Govt UK Govt US Govt FRCPI Linked EMU HICP Linked EMU HICP Linked EMU HICP Linked Euro Govt EMU HICP Linked Source: Barclays Capital Composite indices of countries such as World Ex-UK. the index covers 100% of each of the eligible markets. 1-5y.07 386.54% 7.38 215.39 Market cap ($ bn) 1436. UK+US+France are also available as standard or are customisable on request.00% 7.02 RBS Australia RBC Dominion 12.02 11.52 283.00% 0.85 6000.46% 3.41 327.41 92. JPY.12 157.39 Weight % of overall 100.57 43. EUR.82% 4.

Government debt is further divided by country of issue. Greek Government Inflation-Linked bonds became eligible for both the Euro and Euro Government Inflation-Linked Bond Indices on 30 June 2006. CNA.1bn and included just French domestic issuance. Figure 137: Euro Government Index breakdown by issue type 300 250 Market Value (€bn) 200 150 20% 100 50 0 Sep-98 Sep-00 Source: Barclays Capital Figure 138: Euro Government Index breakdown by issue type 100% 80% 60% 40% Greece Govt France EMU HICP-Linked Italy Govt Germany Govt France Govt FRCPI-Linked 0% Sep-98 Sep-00 Sep-02 Sep-04 Sep-06 Sep-08 France Govt FRCPI-Linked Italy Govt Sep-02 Sep-04 Sep-06 Sep-08 Greece Govt Source: Barclays Capital Germany Govt France EMU HICP-Linked 15 March 2010 167 . respectively. which presently includes issuers such as CADES. the index has grown in terms of market value and the number of markets covered by the index. reaching €290. but due to a ratings downgrade in December 2009 became ineligible (Greece remains as part of the Universal Government Inflation-Linked Bond Index). ISPA and RESFER and EMU member states. the Euro Inflation-Linked Index had a market value of €11.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Euro Inflation-Linked Bond Index Introduction The Euro-Inflation-Linked Index covers euro-denominated inflation-linked debt issued by both government and quasi-government agencies. who joined the Euro Inflation-Linked Index in September 2003 and March 2006. At inception (31 December 1999).6bn at year-end 2009. with market coverage extending to include issuance from Italy and Germany. Since its launch. The Euro Inflation-Linked Index family includes a broad Index and sub-indices by government and non-government sectors and by linking inflation index.

The face values are used rather than an inflation-adjusted value.or better. All Government and non-government issues must have a rating of A3/A. The government sector is further divided by country of issue or by the Linking Inflation Index. The index uses the lower of the Standard & Poor’s and Moody’s rating (long-term local currency rating for Government issues) to determine the index rating for each bond. The reference rate for the Euro and Euro Government Inflation-Linked Index is EUR 1M LIBOR – 15bp Review procedure Maturity bands Index holdings The index is reviewed and rebalanced monthly on the last calendar day of each month. Income received during the month is put on deposit until the month-end rebalancing. The index holding of each bond for the next month is set to the amount outstanding on the review date. New bonds and taps/increases entering the index must have settled on or before the rebalancing date. The non-government sector has only quasi-government issues.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 139: Euro Inflation-Linked Index inclusion criteria and review policy Criteria for inclusion in the index Issuer type Bond type Linking Index Rating Coupon type Maturity Issue date Min issue size Issuer must be an EMU member state or a quasi-government body of an EMU member state Bonds must be capital-indexed and linked to an eligible inflation index. at which point it is reinvested back into the index. 15 March 2010 168 . The indices are weighted using the market capitalisation as standard Index structure The Euro Inflation-linked Index comprises an overall index. Reference deposit rates are set on the last business day of the month for the next month. Bonds must have a minimum remaining life of at least one year on the rebalancing date. Index calculation and monthly review Pricing methodology Settlement conventions Index frequency and reinvestment The index uses mid-market closes from Barclays Capital Market Makers. All spot and forward foreign exchange rates used are official WM Company mid rates from the London market at 4pm. They should be denominated in Euros and pay coupon and principal in Euros. Bonds must settle on or before the rebalancing date to be eligible for the index The issue size must be equal to or in excess of €500mn. Market calendars most appropriate for international investors are used The index is calculated daily and has a value for each calendar day. and is split into government and non-government sectors and also by Inflation Index. Maturity bands are inclusive at the lower bound and exclude the upper bound. Bonds are allocated to maturity bands based on their remaining time to maturity. The linking Index can either be a domestic inflation index or the harmonised EMU HICP. The notional coupon of a bond must be fixed or zero. which are broken down by the Linking Inflation Index. The index uses standard settlement conventions for all calculations.

with individual country and sector indices available since the inception of each market as in the table below.e. 5-10y.49 92.52% 5.83% 5. >10y and >15y. JPY.20% 21 15 5 7 3 6 7 5 1 1 78. GBP. 5-7y.85 32.75 Weight % 100.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 140: The Barclays Capital Euro Inflation-Linked Bond Index structure Euro Inflation-Linked Bond Index Government Non-Government France Govt Italy Govt Germany Govt FRCPI Linked EMU HICP Linked Other FRCPI Linked EMU HICP Linked EMU HICP Linked EMU HICP Linked Euro Govt EMU HICP Linked Source: Barclays Capital The Euro Inflation-Linked Bond Index has a base date of 31 December 1999.03% 100.00% 79.00% 94. 10-15y.41% 4.31 283.32% 0.97% 27.80% 0. 1-3y. SEK.02% 10. USD and ZAR. Foreign and hedged indices are available in AUD.19 12.59% 68. Figure 141: Euro inflation-linked Index characteristics (as of 28 February 2010) Base date Euro Euro Govt EMU HICP Linked France EMU HICP Italy Germany FRCPI Linked Euro Non-Govt FRCPI Linked EMU HICP Linked Other Source: Barclays Capital Pricing source Market cap (€ bn) 299..37 16.97% 14.22% 31.94 2.00% 71.08% 26.12 203.84 15 March 2010 169 .28% Weight % sector No of issues 28 31-Dec-99 31-Dec-99 31-Oct-01 31-Oct-01 30-Sep-03 31-Mar-06 31-Dec-99 31-Dec-99 31-Dec-99 28-Feb-03 Barclays Capital Barclays Capital Barclays Capital Barclays Capital Barclays Capital Barclays Capital 100. 1-5y. CHF. 1-10y.83% 26.72% 32.41 79.40 0.45% 28. i. 5-15y. As well as an overall index. >5y. 7-10y.80% 11. 3-5y. CAD. sub-indices are available for the standard maturity breakdown.

with Brazil being the largest market having a weight of c.$100mn) Nominal ZAR400mn (c. If individual bonds within a market default then these bonds will be removed from the index at the earliest opportunity. Quasi-government and corporate debt is not included in the index. denominated in UVR. Bocones and peso bonds from debt restructuring. For details of bond inclusion rules for each market please refer to Figure 143 Bond inclusion rules Index calculation Index pricing methodology Treatment of income Source: Barclays Capital The index uses daily mid market closes from Barclays Capital market makers taken at local market close for all markets with the exception of South Africa.$90mn) Nominal UDI300mn (c. EEMEA: South Africa.55%. Income from coupon payments is held in cash until month-end when it is re-invested into the index. Countries currently included in the index are: LatAm: Argentina.$60mn) Nominal PLN500mn (c.$190mn) Nominal TRY500mn (c. size. The index was launched in October 2007.2 bn) Bocones CER750mn (c. with daily historical index levels and statistics available back to January 2004 and longer histories for individual countries/regions where available. i. general investability and availability of pricing information.$380mn) Nominal BRL400mn (c. direct obligations of the state issuer.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Emerging Markets Government Inflation-Linked Bond Index The Barclays Capital Emerging Markets Government Inflation-linked Bond Index (EMGILB) measures the total return performance of inflation-linked bonds from major emerging markets countries. denominated in UF. The index includes government debt.$270mn) Nominal UF1mn (c. Mexico.. Min issue Bodens CER5bn (c. denominated in UDI. Figure 143: Bond inclusion rules by EM market Issuer Argentina Republic of Argentina Bond type CER-linked Bodens.$380mn) Restructured debt CER750mn (c.$410mn) Nominal KRW500bn (c. where ABSA Capital market maker mid closes are used. In other markets the cash is held in the index without accruing any interest intra-month.$2.$540mn) Brazil Chile Colombia Mexico South Africa Poland Turkey South Korea Brazil Treasury Central Bank of Chile Republic of Colombia Mexico Ministry of Finance Republic of South Africa Poland Turkey South Korea NTN-B linked to IPCA consumer price index.$40mn) Nominal UVR1bn (c.e. the cash is re-invested intra-month into 1mth JIBAR – 15bp. payable in MXN Bonds linked to South Africa CPI and denominated in ZAR Bonds linked to Polish CPI and denominated in PLN Bonds linked to Turkish CPI and denominated in TRY Bonds linked to Korean CPI and denominated in KRW Source: Barclays Capital 15 March 2010 170 . Poland. payable in COP UDIBONOS linked to Unidades de Inversion (UDI). payable in CLP TES linked to Colombian CPI (UVR). The EMGILB Index currently consists of 57 bonds from nine countries and has an overall market capitalisation of around $246bn. Turkey Asia: South Korea Country rating If an index-eligible country defaults on all of its inflation-linked debt then the market will be removed from the index at the earliest opportunity. For South Africa. Brazil. and the eligibility of the overall market will be reviewed. Hedged and un-hedged currency indices are calculated using 4pm London rates from WM Company. denominated in BRL BCU linked to Chilean CPI (UF). Chile. Colombia. and depth of market. The five LatAm markets account for around 80% of the overall index capitalisation. Figure 142: EM Govt Inflation-Linked Index inclusion criteria and monthly review Criteria for inclusion in the index Markets included Markets are included based on quantitative and qualitative criteria including.

Figure 145: EM Govt Inflation-Linked Index characteristics (as of 28 February 2010) Base date EM By country/region LATAM Argentina Brazil Chile Colombia Mexico EEMEA Poland South Africa Turkey ASIA South Korea Source: Barclays Capital Price source Market cap (local $bn) Market cap $bn 250. 1-3y.69% 48 5 12 16 6 9 10 1 4 5 1 1 A A A BB B BB A BBB BBB 15 March 2010 171 . 7-10y. Please refer to our website www. 3-5y.72 50.96% 55.36 146. GBP.60 3.93 379. sub-indices are available for the standard maturity breakdown. The overall index is also broken down by regional indices for LatAm. USD.84 8. 5-15y. Figure 144: EM Govt Inflation Linked Bond Index structure Barclays Capital Emerging Markets Government Inflation-Linked Bond Index Latin America EEMEA Asia Argentina Brazil Chile Colombia Mexico Poland South Africa Turkey South Korea Source: Barclays Capital Indices are available in local currency and hedged and un-hedged into the following base currencies. i. 10-15y.91 Barclays Capital Barclays Capital Barclays Capital Barclays Capital Barclays Capital 47.43 139.26% 3.12% 4.97 252.24% 11. >5y.65% 11.43% 7.59 19. which are broken down further by maturity. 1-10y.73 79. A constrained index with a cap on maximum country weight of 25% is also available.27 12.00% No of issues 59 Index rating 31-Dec-03 31-Dec-03 31-Dec-03 30-Sep-03 30-Sep-02 31-Dec-02 31-Jan-03 31-Mar-00 31-Aug-04 31-Mar-00 28-Feb-07 31-Mar-07 31-Mar-07 Barclays Capital 2005.80% 3.94 42. EEMEA and Asia.com (keyword: index) for more details on this index. 5-7y.73 1.86% 20.69% 0.live. EUR. 1-5y.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Index structure As well as an overall index. 5-10y.56 15666.60 Weight 100.96 4312.19% 1.28 Barclays Capital ABSA Capital Barclays Capital 10.66 198. JPY and SGD.e. >10y and >15y.11 29.11% 0.barcap.18 8.84 1.17 27.

index activity has been relatively vibrant with a number of issuers becoming ineligible due to both rating downgrades and rating withdrawals – this trend particularly affected the monoline insurance wrapped issuers. For the avoidance of doubt. a trend brought about by the limited number of issuers in the Non-Gilt Inflation-Linked market and relatively poor liquidity.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Sterling Inflation-Linked Bond Index Launched in June 2001. For amortising structures. with the former dominating the All Bond Index making up 91% of the Overall Sterling Inflation-linked Bond Index by December 2009. where capital amount = face value x amortisation factor. much of the new NonGilt issuance has been taken up by asset swap investors and thus has not been eligible for index inclusion. this procedure will apply to new issues only – thus at the time that the market pricing rule was introduced (1 July 2006). then prevailing index constituents were not subject to the pricing check. The linking index can be UK RPI or LPI. The Sterling Inflation-Linked All Bond Index covers both Gilt and Non-Gilt inflation-linked debt issuance. Since 2005. All issues must have a rating of Baa3/BBB. Bonds must have a minimum remaining life of at least one year on the rebalancing date. typically from utility issuers. Supply has continued to be steady in the second half of 2009. available in full via market data services such as Bloomberg). Bonds must have an issue size of £100mn or above. then Fitch ratings are used. Any new issue that has not sufficient price coverage to enter the index at the first opportunity will be checked again the following month.4bn ($21.or above The notional coupon of a bond must be fixed or zero. All (real) cash flows must be fixed at issue. Amortising inflation-linked bonds must have a remaining capital amount greater than or equal to £100mn. the Barclays Sterling Inflation-Linked All Bond Index captures the performance and characteristics of the sterling inflation-linked bond market. At inception.4bn equivalent). Figure 146: Sterling Non-Gilt Index market value by sector 25 Insurance Wrapped 20 Market Value (£bn) 15 10 5 0 Dec-99 Corporate Agency / Supra Utilities Figure 147: Sterling Non-Gilt weights by sector 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Dec-99 Dec-01 Dec-03 Dec-05 Dec-07 Dec-09 Utilities Dec-01 Dec-03 Dec-05 Dec-07 Dec-09 Corporate Source: Barclays Capital Agency / Supra Insurance Wrapped Source: Barclays Capital Figure 148: Sterling Inflation-Linked Index Inclusion criteria and monthly review Criteria for inclusion in the index Bond type Linking index Rating Coupon type Maturity Issue date Min issue size All issues must be denominated in sterling and pay interest and principal in sterling. but after that becomes ineligible even if it is subsequently priced by more than one dealer. the Non-Gilt portion of the All-Bond Index has grown substantially and at 31 December 2009 consisted of 50 bonds with an Aggregate face value of £17. The test is implemented by looking for pricing coverage in the last calendar week of each month for all potential index entrants. Towards the end of 2009. Bonds must settle on or before the rebalancing date to be eligible for the index. the Sterling Non-Gilt Inflation-Linked Bond Index consisted of 12 issues with an aggregate face value of £740m. 172 Market pricing 15 March 2010 . To be eligible for inclusion in the index a new issue has to be priced by at least two dealers. Since inception. If neither of these rates the issue. with UKRAIL regularly tapping and the occasional new issues. the amortisation schedule must be fixed at issue and must be in the public domain (ie. The index uses the lower of the Standard & Poor’s and Moody’s rating to determine the index rating for each bond.

history is available for the index-linked Gilt market back to 1981. Figure 149: Sterling Inflation Linked Bond Index structure Sterling Inflation-Linked Bond Index UK Govt Non-Gilts Agency/ Supranational Corporate Utility Insurance Wrapped The Non-Gilt. 5-15y. All spot and forward foreign exchange rates used are official WM Company mid rates from the London market at 4pm are used for the Gilt Indices. The sector indices have been designed to reflect the types of non-government issuance seen to date. New bonds and taps/increases entering the index must have settled on or before the rebalancing date. Income received during the month is put on deposit until the month-end rebalancing. 15 March 2010 173 .15pm London time. >5y. 3-5y. 5-10y. 1-10y. 7-10y.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Index calculation and monthly review Pricing methodology Settlement conventions Index frequency and reinvestment The index uses mid-market real prices from the Barclays Capital market makers at 4. The face values are used rather than an inflation-adjusted value. Bonds are allocated to maturity bands based on their remaining time to maturity. overall and sector indices have a base date of 31 December 1999. Market calendars most appropriate for international investors are used. 5-7y. >10y and >15y. 1-5y. sub-indices are available for the standard maturity breakdown. Maturity bands are inclusive at the lower bound and exclude the upper bound. 1-3y. 10-15y. The indices are weighted using the market capitalisation as standard. ie. The index is calculated daily and has a value for each calendar day. Note that there are no foreign or hedged versions for the Non-Gilt Indices. at which point it is reinvested back into the index. The index holding of each bond for the next month is set to the amount outstanding on the review date. As well as an overall index. Review procedure Maturity Bands Index Holdings Source: Barclays Capital Index structure The sterling inflation-linked index family comprises an overall index as well as two index sub-groups: Gilts and Non-Gilts. The Non-Gilt index is further divided into rating and sector indices. The Gilt index has a base date of 31 December 1996. The reference rate for the Sterling Inflation-Linked Indices is GBP 1mth LIBOR – 15bp The index is reviewed and rebalanced monthly on the last calendar day of each month. The index uses standard settlement conventions for all calculations. Reference deposit rates are set on the last business day of the month for the next month.

77 3.06 1.05 6.13 No of issues 67 100.91 14.66% 14.17% 1.93% 0.09% 0.22 1.79% 19 8 19 4 15 March 2010 174 .84% 12.00% 64.95 1.32% Weight % sector 31-Dec-99 31-Dec-96 31-Dec-99 31-Aug-00 30-Apr-00 31-Dec-99 31-Dec-99 31-Dec-99 30-Apr-00 31-Dec-99 31-May-01 14.24 21.60% 4.44% 64.00% 0.76% 9.82% 1.69 4.71% 2.00% 90.15 Barclays Capital Barclays Capital Barclays Capital Barclays Capital Barclays Capital Barclays Capital 215.31% 50 19 8 11 12 Base date £ Overall Gilts Non-Gilts Agency/Supra Corporate Utility Insurance Wrapped By Rating Non-Gilt AAA Non-Gilt AA Non-Gilt A Non-Gilt BBB Source: Barclays Capital Pricing source Weight % 100.24% 5.91% 7.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 150: Sterling Inflation-Linked Index characteristics (as of 28 February 2010) Mcap local (£ bn) 237.00% 100.94 2.70% 22.20% 8.

The two main benchmarks in the series-L family are the Global Inflation-Linked (Series-L) and US TIPS (Series-L) indices.or higher using the middle rating of Moody’s. S&P and Fitch ratings. This difference has a small impact on the overall indices. Greece was eligible for the WGILB. consistent with float adjustments made for nominal US Treasuries in the US Aggregate Index. UK. The Global Inflation-Linked Index (Series-L) also includes inflation-linked bonds from Greece because it has a minimum rating requirement of Baa3/BBB. Differences between these two benchmarks and their WGILB counterparts are as follows: Country Inclusion List − Both the WGILB and Global Inflation-Linked (Series-L) indices include the following countries: US. Prior to 31 December 2009. which has a minimum rating of A3/A-. because some clients are still benchmarked to these legacy indices. as Greece only has two index eligible bonds (1. The WGILB Index includes bonds from Australia. the inflation-linked index family will be consolidated into a single family and common set of rules. The US TIPS component of the WGILB Index uses the full issue size to determine index weights and does not adjust for Federal Reserve or government holdings. these indices are available only for government linkers in developed markets. − − − Full issue size versus public float for US TIPS: − US TIPS in the Global Inflation Linked (Series-L) and the stand-alone US TIPS (Series-L) Indices both exclude government/Federal Reserve purchases of TIPS from the face amount outstanding of each bond in the index.7% of the WGILB Index by market value). They remain in the index until month-end and are reinvested back into the index at that time. Therefore. Germany and Japan. the amount outstanding of a TIP bond in the Global IL Index (Series-L) may be less than the amount outstanding of the same bond in the WGILB Index if any part of the issue was purchased by the US Government or Federal Reserve.3% of the Global IL Index by market value) and Australia has only three index eligible bonds (0. 15 March 2010 175 . These bonds do not qualify for the WGILB. Though covering most of the same markets as the World Government Inflation-Linked bond Index (WGILB). there are some differences in index rules and conventions. whereas the WGILB Index uses the local settlement convention for each market. which is not an eligible country for the Global Inflation-Linked (Series-L) Index. − Settlement assumption: The Global Inflation-Linked (Series-L) Index uses a T+1 settlement assumption for all bonds in the index (and first calendar day of the next month at each month-end). France. Reinvestment return: Series-L indices do not earn a reinvestment return on intra-month coupon payments. Canada. Designated as ‘Series-L’. but over time. they have been maintained. Italy.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Global Inflation-Linked Bond Index (Series-L) The Barclays Capital Inflation-Linked Index family also includes several legacy benchmarks that were published by Lehman Brothers and are now part of the Barclays Capital Index platform. Sweden.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide The index has a market capitalisation of US$1. prices from the previous business day are used for that particular market. Prior to May 2009. Sweden. New bonds and taps/increases entering the index must have been auctioned on or before the rebalancing date. Settlement conventions Index frequency Review procedure Maturity bands Index holdings Source: Barclays Capital The index uses a T+1 settlement convention for all calculations. The index holding of each bond for the next month is set to the amount outstanding on the review date. Germany and Japan (as of 1 January 2009). Canadian linker prices are taken from RBC Dominion and Swedish linker prices are taken from PMI Exchange.or above) using the middle rating of Moody’s. Government domestic inflation-linked debt issued in the domestic currency of that country only. settlement is the first calendar day of the next month. The index has a market capitalisation of US$1. Bonds must have a minimum remaining life of atleast one year. EUR100mn. All spot and forward foreign exchange rates used are official WM Company mid rates from the London market at 4pm. GBP100mn and USD250mn.4trn and includes 92 issues from 10 governments (as of 31 December 2009). The indices are weighted using the market capitalisation as standard. US TIPS were priced with bid prices. A full set of inflation-linked analytics is required for a bond/country to be added to the index. Euro government. SEK100mn. S&P and Fitch. US TIPS. on the last calendar day of the month. Maturity bands are inclusive at the lower bound and exclude the upper bound. Bonds are allocated to maturity bands based on their remaining time to maturity. Rating: Countries eligible for inclusion must have an investment grade sovereign rating (Baa3/BBB-/BBB. UK linkers and Japanese linker prices come from Barclays Capital market makers. Bonds must be capital-indexed and linked to a commonly used domestic inflation index. 15 March 2010 176 . The face values are used rather than an inflation-adjusted value. Index calculation and monthly review Pricing methodology The index uses mid-market prices from local market close and is priced daily. Greece. France. respectively. UK. If the last business day of the month is a public holiday in one of the major regional markets. The index is reviewed and rebalanced once a month. The index is calculated daily and has a value for each business day. At month-end. Italy. Canada. The notional coupon of a bond must be fixed. Figure 151: Inflation-Linked (Series-L) Index inclusion criteria and monthly review Criteria for inclusion in the index Inclusion of markets in the Global Inflation Linked Index Underlying market/bond type Min issue size Maturity and issue Countries currently covered include the US. The following minimum issue sizes are as follows: CAD 200mn.4trn and includes 92 issues from 10 governments (as of 31 December 2009). JPY10bn.

Thus by selecting the comparator nominal bonds and custom-weighting each nominal bond equal to that of the equivalent inflationlinked bond it is possible to make a more effective comparison by avoiding the complexities of broad market. How can we compare linker portfolio duration with that of a nominal bond portfolio? We could use historical yield beta analysis of each linker verses its benchmark as an approach to solving this problem but this has some drawbacks: 15 March 2010 177 . The Comparator Bond Indices allow for a simple and useful comparison of the relative performance of Inflation-Linked and Nominal Government bond indices. The basis for this technique is to construct an index of nominal comparator bonds that are both maturity matched with the equivalent inflation-linked index and custom weighted such that the weight of the maturity-matched. issued with the same (or closest) original term as the inflation-linked bond and liquidity. are modified to account for the distinct nature of inflation-linked and nominal securities.com Introduction Comparator bond Indices provide a simple and transparent framework for comparing returns on an inflation-linked Government bond index with that of the equivalent nominal government bond index. To make for a more effective comparison. Considerations In a pure credit context one would also try to match the duration profile of the target portfolio.venkataraman@ barcap. Therefore the Comparator Bond Indices. each of the nominal comparator bonds is weighted such that the market value of each nominal comparator bond matches the market value (inflationadjusted) of the associated inflation-linked bond. because we are dealing with two separate asset classes. on the last calendar day of each month. Generally comparing broad market. In addition. market-value weighted indices. The nominal comparator bonds generally switch more frequently than the corresponding inflation-linked bonds given the depth of the nominal market relative to the inflation-linked. market-value weighted nominal bond indices in a breakeven inflation (BEI) context with the equivalent inflation-linked bond index would be sub-optimal since nominal indices tend to be much broader and have different maturity profiles than their inflation-linked counterparts. The comparator nominal bonds are determined by market convention. while borrowing the principle of excess return analysis. this is problematic. but not limited to. this is not so straightforward when comparing two distinct asset classes such as inflation-linked and nominal government bonds. Some of the key considerations while establishing the applicable nominal comparator for each of the inflation-linked bonds include. nominal comparator bond is equal to that of the equivalent inflation-linked security in the corresponding indices.com Anand Venkataraman +44 (0) 20 7773 0852 anand. However.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Comparator bond indices Scott Harman +44 (0) 20 7773 1775 scott. the Comparator Bond indices select the comparator nominal bond utilised for the purposes of BEI. closest time-to-maturity as the inflationlinked bond. Generally these are of equal credit quality and similar in maturity to the equivalent inflation-linked bond. Methodology The Comparator Bond Index methodology is borrowed from the excess returns analysis of credit markets where excess returns are calculated by making a straight comparison between returns of a credit bond versus its benchmark.harman@barcap. however.

SGD and USD. eg. South Korea. the following foreign currencies are available: EUR. which cannot be standardised. SGD. eg. we would be unable to perform a historical analysis. the results would very much depend on the period chosen for analysis. World Government ex Japan Comparator Bond Index. This is not a wholly transparent process. JPY. The Composite Comparator Bond indices also include Greece and associated Composites. JPY. This is bound to be subjective. South Korea. For these reasons a simple maturity-based approach has been adopted. The remaining EM inflation-linked markets have less developed nominal bond markets than inflation. GBP. Betas would have to be re-estimated every month based on available information. For this reason we have no aggregate or composite Comparator Bond indices in the EM space. instead we would have to make a forwardlooking prediction perhaps based on other markets. GBP. USD and ZAR. For South Africa. such as Germany or Japan. EUR.Barclays Capital | Global Inflation-Linked Products – A User’s Guide At the inception of a new inflation-linked market. World Government + Greece Comparator Bond Index. Figure 152: Comparator Bond Index structure World Govt Comparator Bond Index Australia Govt Canada Govt France Govt Italy Govt Germany Govt Sweden Govt Japan Govt UK Govt US Govt Source: Barclays Capital Comparator bond indices are available for all of the inflation-linked markets covered within the World Government. CHF. The following foreign currencies are available for World and sub-indices thereof: AUD. Poland and Turkey. Euro Government and Euro Inflation-Linked Bond Index families. CAD. Poland and Turkey. Only those Emerging Market (EM) countries where the nominal government market has sufficient coverage have comparator bond indices. Popular Composite Comparator Bond Indices are also produced replicating the associated Inflation-Linked Composite Index. 15 March 2010 178 . SEK. Comparator bond indices are available in local and foreign currency (hedged and un-hedged). hence South Africa.

income from coupon (and principal payments in the case of amortising bonds) is held as cash and earns the monthly re-investment rate until the next rebalancing date. Market calendars most appropriate for international investors are used. Figure 139. Settlement conventions The index uses standard settlement and ex-dividend conventions for all calculations. Figure 142 and Figure 148 for the price sources for each inflation market. accrued and analytical values are carried over unchanged from the previous day. for all of the EM countries with the exception of South Africa. From 1 July 2006 onwards.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Benchmark Inflation-Linked Bond Index calculation and methodology Calculation and methodology for Flagship indices 13 Pricing methodology The Barclays Capital Inflation-Linked family of indices use mid quotes updated daily at local market close. For all markets the monthly reinvestment is applied but only markets where there is a suitable deposit rate available does the income that is received from the coupon accrue. In the case of South Africa coupon payments are reinvested using the method described above. Series-L Inflation-Linked Indices follow separate conventions. The table below contains the source for monthly money market rates for each currency bloc. This allows for simple calculation of returns without having to adjust the start and end dates according to the business calendar. Thus. EGILB. On non-business days the security price. This ensures that the index has no local currency performance on days when the local market is closed. 13 The calculation and methodologies outlined in this section apply to the WGILB.15%) on the last business day of the previous month. Figure 153 lists those countries where monthly reinvestment of coupons accrue and the associated deposit rates. less an accrual rate. 15 March 2010 179 . income from coupon was reinvested in the index as soon as it was received. Please refer to Figure 134. All spot and forward foreign exchange rates used are official WM Company mid rates from the London market close at 4pm London time. EMGILB and Sterling InflationLinked Indices. coupon payments do not accrue at local money market rates. Index frequency and reinvestment The index is calculated daily and has a value for each calendar day. On the next rebalancing date. the coupon payments and interest earned on it are re-invested back into the index. Money market rate for coupon re-investment For each currency bloc the monthly money market rate for re-investment of coupon payments is equal to the 1M Libor rate less 15bp (0. Until 30 June 2006.

The face value is used rather than an inflation-adjusted value. the weights are adjusted daily to account for price changes. 15 March 2010 180 . Bond selection rules are applied and bonds are allocated to the appropriate sector. Bonds are allocated to maturity bands based on their remaining time to maturity (for bullet bonds) or average life (for amortising bonds).15bp GBP 1M Libor . Details of the available maturity bands for each index are given in the respective index structure and rules section.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 153: Monthly rates for re-investment of intra-month coupons Market US UK France Italy Greece Germany Japan Sweden Canada Australia South Africa Currency USD GBP EUR EUR EUR EUR JPY SEK CAD AUD ZAR Reinvestment rate USD 1M Libor . on the last calendar day. the indices are reviewed and rebalanced based on inclusion criteria described earlier in the respective index sections of this guide.15bp EUR 1M Libor . Index holdings The index holding of each bond for the next month is set to the amount outstanding on the review date. Increases or buybacks to existing bonds are applied according to the same rule. Some markets have a full set of maturities while others are sparsely distributed across different maturities.15bp CAD 1M Libor . accrued interest and indexation. Bonds entering the index for the first time must have settled on or before the review date.15bp EUR 1M Libor . Maturity bands A variety of different maturity bands are utilised.15bp AUD 1M Libor .15bp EUR 1M Libor . Maturity bands are inclusive at the lower bound and exclude the upper bound.15bp Johannesburg Interbank Agreed Rate 1M – 15bp Bloomberg ticker US0001M BP0001M EE0001M EE0001M EE0001M EE0001M JP0001M SK0001M CD0001M AU0001M JIBA1M Source: Barclays Capital Monthly rebalancing Review procedure Once a month. rating and/or maturity indices.15bp EUR 1M Libor . As detailed in the index formulae section. The indices are weighted using market capitalisation as standard.15bp JPY 1M Libor – 15bp SEK 1M Libor . The holdings remain fixed for the whole of each calendar month.

– Cash held on day t.b – Clean settlement price of bond i at close of day t. starting with a base value of 100 on the base date. The following notations are used in the index formulae: Pi .b ) i Cash Held is reset to zero on date b.t )* N i .t – Coupon paid on bond i on day t – Any ex-dividend coupon that will be received on bond i – Gross redemption yield on bond i at the close of day t – The duration of bond i at the close of day t.t + Ai .b Ai . t – Business day on which the index is being calculated t-1 – The previous business day n – Number of calendar days since the previous business day t-1 b – Last business day of the previous month Ri .b – Face value amount outstanding of bond i on day t.t Pi . as given in Figure 154. ∑x i i .t is intended to show a summation occurring over bonds i which are eligible to be included in the index on day t. – Clean settlement price of bond i at close of day b. Note this is not inflation- adjusted but the face value.b ) n + ∑ (Ci . – Accrued interest for settlement of bond i for trading on day t.b CH t – The monthly re-investment rate for bond i on business day b. – Market value of bond i on day t in cash terms. mi .b would apply to the eligible list on 15 March 2010 181 .t * N i . t-1 – Face value amount outstanding of bond i on day t-1.b The notation ∑x i i . Ni.t Ai .t X i .Barclays Capital | Global Inflation-Linked Products – A User’s Guide Index formulae – local market indices Notation The Barclays return indices are calculated daily and are chain weighted for each day t.t = (Pi . N i . – Accrued interest for settlement of bond i for trading on day b. Ci . post any capitalisation change. where CH t = CH t −1 * (1 + Ri .t Yi . Hence day b.t Di .

Clean price index The Clean Price Index ( CPI t ) is defined as: CPI t = CPI b ∑ (P * ∑ (P i i i .b ) * N i .b i .t i .t + Ai .b Total Return Index The Total Return Index ( TRI t ) is calculated as: TRI t = TRI b ∑ ((P + A + X )* N ) + CH * ∑ (( P + A + X ) * N ) i .t i . Gross Price Index The Gross Price Index ( GPI t ) is analogous to the clean price index and is defined as: GPI t = GPIb ∑ (( P * ∑ (( P i i i .b The CPI is calculated daily as shown above for all eligible bonds.b i Defining the Income Index in this way ensures that the Total Return Index can also be calculated by the equation: TRI t = TRI b * (GPI t + ΔI t ) GPI b where ΔIt is the increase in income (if any) on day t.t ) * N i .b + X i .Barclays Capital | Global Inflation-Linked Products – A User’s Guide Market capitalisation The Total Market Capitalisation M t of all the constituent bonds on day t is given by: M t = ∑ mi . I t = I b + GPI b * CH t ∑ ( Pi .b ) i .b i .b + Ai .b + X i . 15 March 2010 182 .t + X i .b i i .b ) * N i .b i t Income Index The Income Index ( I t ) is a cumulative figure of coupon income received in the year to date.t * N i . Coupons are added to the income index for each stock paying a coupon today that was in the index yesterday.t i .b i .t i .b ) + Ai . It is reset to zero at the beginning of each calendar year.b ) i .b ) * N i .

15 March 2010 183 .t Mt where W i. the duration of each bond is weighted by its market capitalisation.t is the real gross redemption yield (either on an annual or semi-annual basis) of bond i at the close of day t. Wt = ∑W i i . Average (Macaulay) Duration To calculate the Average (Macaulay) Duration (D t). ΔI t = I t It can be shown that this calculation of the TRI is perfectly consistent with the previous equation stated above. Dt = ∑D i i . and D i.t i i .Barclays Capital | Global Inflation-Linked Products – A User’s Guide Normally ΔI t = I t − I b but.t i .t is the modified duration of bond i at the close of day t. Average Real Yield (annual and semi-annual) To calculate the Average Real Yield (Y t).t i . the modified duration of each bond is weighted by its market capitalisation. if t and b span a year-end.t i i . Modified duration quoted is the sensitivity of the quoted price with respect to changes in real yields.t is the duration of bond i at the close of day t.t ∗ mi . than weighting individual redemption yields by market capitalisation alone.t Mt where D i. Yt ∑ (Y * m * D = ∑ (m * D ) i . the real gross redemption yield of each bond is weighted by its market capitalisation and its duration (Macaulay).t ) where: Y i.t is the duration (Macaulay). Average Modified Duration (annual and semi-annual) To calculate the Average Modified Duration (W t). This measure (which includes weighting by duration) is a better approximation of the true gross redemption yield of all the cash flows in the index.t ∗ mi .

t is the Local Currency Return Index of a Multi-Currency Index on day t. Total Return Index The Total Return Index is calculated as: TRI M . Ft = ∑ Fi .t i i .t is the life in years of bond i at the close of day t. but with weights in US dollar terms.t = TRI M .t where L i.b L ⎛ TRI L.t L . and F i. the life in years of each bond is weighted by its face value.b 15 March 2010 is the Local Total Return Index on day t.b ⎟ ⎟ ⎠ LM .t is the face value of the bond i at the close of day t.b L .b ) Where: TRI M . CVt = ∑ Fi . TRI L .t ∗ Pi . Face Value To calculate the Face Value (F t) of the index. The formula for the Total Return Index is given below (other indices are calculated on a similar basis). 184 .t i Clean Market Value To calculate the Clean Market Value (CV t) of the index. Lt ∑L *F = ∑F i . the face value of each bond is multiplied by its clean price and this value is summed across all bonds in the index.b ⎞ * S LM .Barclays Capital | Global Inflation-Linked Products – A User’s Guide Average Life To calculate the Average Life (L t).t i i .b ∑ ⎜ TRI × M ⎜ ∗ ⎝ ∑ (M × S L L . is the Local Total Return Index on day b.t TRI L . the face value of each bond is summed across all bonds in the index.t i Aggregate local currency indices The aggregate local indices for multi-currency portfolios (World and World Ex-UK) are calculated from the individual country indices with local performance.

S LF . TRI L. S LF . coupled with a rationale for the choice of methodology employed. is the local total market value at time b. The method assumes that the exchange rate used to convert the value of non-US dollar holdings into US dollars at the end of the month is simply the one-month forward rate from the previous month. While the differences between the three are minor. and is. The following sections provide a detailed explanation of how these versions of index are calculated. Methodology rationale – hedged indices Three different procedures were considered for calculating the hedged performance of the index. The perfect foresight method The perfect foresight method is the easiest method for calculating hedged returns.t = Local Total Return Index at time t. TRI F . The same technique is used to calculate the associated clean and gross price indices.t S LF . ease of understanding and replication.Barclays Capital | Global Inflation-Linked Products – A User’s Guide S LM .c = Spot exchange rate between local and foreign currency at commencement date of the index.t * Where: S LF . This has the advantage of being the simplest method of the three to apply to the index.t = Spot exchange rate between local and foreign currency on day t.t = Foreign Total Return Index at time t. Foreign and hedged return indices The Barclays Capital Inflation-Linked indices are also available in foreign currency and hedged versions. and summation is over all countries in the Multi-Currency Index. 15 March 2010 185 . it has the major disadvantage that it would be impossible to replicate in practice. however.b is the spot exchange rate between the local currency and the currency of the Multi- Currency Index on day b. The formula below is used to calculate the Foreign Total Return Index.b M L . The decision was based on a number of factors including transparency. Foreign Currency-Total Return Index The foreign currency versions of the local currency index are calculated using the local index and the spot foreign exchange rate between the local and “foreign” currencies. referred to as the ‘perfect foresight method’. This method inherently assumes that we hedge exactly the correct amount of each currency. it was ultimately decided that the third option would be used for calculating the hedged returns.c TRI F.t = TRI L . therefore. The three options that were considered are presented below.

and this is the method used for the Index. but only provides a marginal reduction in the currency mismatch. The current value method is relatively straightforward to apply and is easily replicable. Any subsequent appreciation or depreciation in the value of the index is un-hedged until the next hedge is taken out. The expected value method The expected value method requires the calculation of the expected amount of the holdings at the end of the month. the interest that will accrue and coupons that will be paid are calculated and added to the beginning-of-month holdings. income is reinvested across all the bonds in proportion to their weight. This takes place as soon as income is received and hence the weight will depend on the local market price and. At the beginning of each month. Given 15 March 2010 186 . Hedged return calculations The index uses the current value method to execute a one-month hedge at the beginning of each calendar month. One-month forward rate agreements are entered into for the full market value of the index at the beginning of each month. This residual currency exposure will then have to be converted into US dollars using the spot exchange rate at that time. Exchange rate source WM Company closing mid values are used for both spot and forward rates at 4pm London time. This sum is then hedged into US dollars.Barclays Capital | Global Inflation-Linked Products – A User’s Guide As this method would violate the general principle that the index be replicable. This method is better in theory than the current value method. on the currency cross rates. The current value method The final method considered was to hedge only the amount outstanding in each of the currencies at the end of each month after any reweighting of the index constituents. Single Currency Index hedge The hedged return on a single currency index or portfolio can be viewed as consisting of three parts: Local return Currency return on the un-hedged portion of the fund Profit or loss on the hedge itself Multi-Currency Index hedge In an index. This hedge covers the current value of the Index plus the predictable components of the change in value. in the case of a multi-currency basket. This increase in accuracy involves significantly more complex calculations to estimate the end-of-month currency exposure. This is referred to as the ‘current-value method’. as it provides a better match to the true currency exposure. This does not assume perfect foresight so there will be an element of currency mismatch at the end of the month if the value of the portfolio holdings in the currency changes. the perfect foresight method is not used for calculating the hedged returns. If the last calendar day of a month is a non business day the rates correspond to the previous business day.

as stated above. but in practice only certain periods (tenors) are quoted. Suppose we are 10 days into the month and that the last business day of this month is the 28th. This is to provide continuity between the monthly and daily hedged total return series.1M S LF . s . Pricing the offsetting forward To mark the initial one-month forward position we use an offsetting forward to the end of the month. s −1 Forward Return = FR = FLF . and we need to use interpolation to arrive at a rate for the desired period. Here we need to offset the starting onemonth forward with an 18-day forward (ie. In our example we would calculate the 18-day forward rate as the current spot rate plus the premium or discount between spot and 1-mth forward pro-rated for 18 days. The easiest way to explain this is to look at an example. s −1 Currency Return on Un-hedged Portion = (1 + LR) x CR Hedge Return = FR – CR Hedged Index Return = LR + Currency Return on Un-hedged Portion + Hedge Return Hedged Index Value = Start Hedged Index Value * (1+Hedged Index Return) Where: 15 March 2010 187 . 28 –10 = 18 days). s − 1. this is a daily estimation of the Monthly-Hedged Index and not a true Daily-Hedged Index. The hedge return is then combined with the month-to-date local currency return and the un-hedged currency return to give an overall month-to-date hedged index return and hence the hedged index value. Instead we use a simpler breakdown: Index return (in the desired currency) Profit or less on the hedge itself (or in this case on a series of currency hedges). This is more precise than using the spot rate as it takes into account expected interest rate differentials for the remaining part of the month. Currency Return = CR = S LF . This is done by unwinding the forward position and adjusting the return on the hedge. Calculating a Daily Hedged Index In order to provide a daily estimate of the performance of the Monthly-Hedged Index we do a daily mark-to-market of the currency hedge. Formulae for monthly calculation – Single Currency Index Local Return = LR = TRI L . For the sake of simplicity we use a linear interpolation based on the current one-month forward rate and spot rate. In theory we could obtain an 18day rate directly for the forward market. we cannot observe the local currency return in the same way as described above. Note that.e S LF .e TRI L .Barclays Capital | Global Inflation-Linked Products – A User’s Guide this reinvestment strategy.

e − 1 + ∑ ⎢WL . R – Remaining days in hedge. s . i.b × LM.b Hedged Index Return = Un-hedged Index Return + Hedge Return = ⎤ ⎡ F -S TRI M .i FLF . R – Forward FX rate local currency into hedge currency on day i for forward period R. Formulae for daily calculation – Single Currency Index MTD Local Return = MTD LR = TRI L . s −1 Forward Return = FR = FLF . s −1 MTD Currency Return = MTD CR = S LF .b.b × FLM.e −1 TRI M .b. e ⎤ ⎥ S LF .Barclays Capital | Global Inflation-Linked Products – A User’s Guide s – Start date e – End date TRIL – Local Currency Total Return Index S LF – Spot foreign exchange rate between local currency and the hedge currency F LF. s −1 MTD Currency Return on un-hedged portion = (1 + MTD LR) x MTD CR Hedge Reversal Return = S LF . s.b S LF . F LF. e ⎥ TRI M .e TRI L . 1M – One-month forward foreign exchange rate between local currency and the hedge currency. s and the onemonth forward rate FLF. Formulae for monthly calculation – Multi-Currency Index Un-hedged Index Return = TRI M . This is calculated by linear interpolation between the spot rate S LF.e S LF . 1M where 1 < R < 1M.b ⎦ Hedge Return = ∑ ⎢W L ⎡ ⎣ L .i .SLF.1M S LF . LF. 1M . 1M. R −1 MTD Hedge Return = FR + Hedge Reversal Return – MTD CR MTD Hedged Index Return = MTD LR + (1+MTD LR) * MTD CR + MTD Hedge Return Hedged Index Value = Start of Month Hedged Index Value * (1 + MTD Hedged Index Return) Where: i – Intra-month date.b L ⎣ ⎦ 188 15 March 2010 .

b × S LF .b ∑ (M L.b ⎠⎥ ⎦ MTD Currency Return = ∑ ⎢W L ⎡ ⎢ ⎣ L .b = M L .b ⎛S ⎞⎤ × ⎜ LF. 1M – One-month forward FX rate local currency into hedge currency on day b WL .i − 1⎟⎥ ⎜F ⎟ ⎝ LF. i. R.i −1 TRI M . b. – R period Forward FX rate local currency into hedge currency on day i WL . 1M ⎞⎤ − 1⎟⎥ ⎢WL .b × ⎜ ∑ ⎜ S ⎟ L ⎢ ⎝ LF .b = M L .b × S LF .Barclays Capital | Global Inflation-Linked Products – A User’s Guide Hedged Index Value = Start Hedged Index Value * (1 + Hedged Index Return) Where: b – start date e – end date TRI M –un-hedged Total Return Index in hedge currency S LF – Spot FX rate local currency into hedge currency F LF.b ⎛ S ⎞⎤ × ⎜ LF .b ⎠⎥ ⎣ ⎦ Hedge Reversal Return = ∑ ⎢W L ⎡ ⎢ ⎣ L .b ∑ (M L.b – Weight of each local currency index on day b in the Multi-Currency Index: WL .b ) L 15 March 2010 189 .b. ⎠⎥ ⎦ MTD Hedge Return = Forward Return + Hedge Reversal Return – MTD Currency Return MTD Hedged Index Return = MTD Un-hedged Index Return + MTD Hedge Return Hedged Index Value = Start of Month Hedged Index Value x (1 + MTD Hedged Index Return) Where: i – intra month date R – Remaining days in hedge F LF. R.b Forward Return = ⎡ ⎛ FLF.b × S LF .b ) L Formulae for daily calculation – Multi-Currency Index MTD Un-hedged Index Return = TRI M . i.b × S LF . i − 1⎟⎥ ⎜S ⎟ ⎝ LF .b – Weight of each local currency index on day b in the Multi-Currency Index: WL .

Treasury weighted inflation swap exposure across the curve Breakeven Inflation Indices US Breakeven Inflation Index INSTEP Other Inflation Index products INSPIRE Real Income AIMS 15 March 2010 190 . and other access products designed to offer inflation exposure to less liquid markets. Inflation Swaps Index family Inflation Swaps – Constant Maturity swaps at different tenors for six markets. breakeven inflation index products. Barclays Capital calculates and publishes a number of tradable inflation-linked index products designed to offer efficient market access and/or rules-based alpha generation in index form. − − − − − US UK Euro France Japan TWIST. Variants include zero coupon and real return indices and are available as both unfunded (excess) and funded (total)returns. This section will profile some of our more notable inflation-linked index products. These indices include a family of inflation swaps indices.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Barclays Capital Inflation–Linked Index Products Overview In addition to the flagship benchmark inflation-linked indices profiled earlier.

out to 30 years in increments EUR HICPx. The extensive coverage features various markets and tenors supporting breakeven inflation swaps and real rate swaps. only breakeven swaps French CPIx. Each set references the return of zero-coupon swaps rebalancing every month end. 14 Index family overview Barclays Capital produce two sets of indices – breakeven inflation swap indices and real rate swap indices. out to 30 years in increments Japan CPI. out to 30 years in increments Excess return indices are generated from pure swap performance. The position is marked to market every day as it decays over the next month. At the next rebalance. 14 Calculations and Methodology apply only to Series-B Inflation Swap Indices 15 March 2010 191 . Total return indices are calculated as well and assume an additional cash funding component. out to 50 years in increments US CPI. selected tenors based on liquidity AUCPI.com The Barclays Capital Inflation Swap Indices are tradable products designed to replicate the performance of investing in inflation swaps.sharma@barcap. selected tenors. Indices are calculated for various tenors across the term structure and for the following inflation markets: UK RPI. The indices maintain a constant maturity from month to month. These swap indices can be used as customized liability management tools against which to benchmark inflation sensitive portfolios. On the last business day of each month a swap is initiated at the prevailing market rate for a specific tenor.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation Swap Index family Rahul Sharma +65 6308 3982 rahul. They can also be used as components in strategy index products. the 1m decayed position is unwound and a new swap is initiated at that day’s prevailing market rate for the same original tenor as before.

The inflation payer agrees to pay the cumulative percentage change in a referenced and lagged price index over the life of the swap. There are no interim payments. The real rate is calculated on the trade date by setting the initial net present value of the swap to zero. DCF is the day-count fraction. the fixed breakeven rate (BE). These swaps synthetically replicate TIPS style exposure since the NPV is driven by real yields and inflation. pension fund liabilities are becoming linked to real yields and not just inflation. The inflation receiver in turn pays a fixed compounded rate. ⎡ C P I ma t ⎤ N × D Fm at × ⎢ − 1 ⎥ − N × D F m a t × ( 1 + B E )T e no r − 1 ⎣ CP I ba se ⎦ ⎡ C P I m a t ⎤ T e no r − 1 = BE ⎢ ⎥ ⎣ CP I ba se ⎦ 1 [ ] Zero-coupon real rate swaps A real rate inflation swap involves an exchange of Libor-based payments versus a fixed real payment plus realised inflation. This is the breakeven inflation rate (BE) calculated by setting the initial net present value of the swap to zero. the inflation payer agrees to pay the cumulative percentage change in a referenced and lagged price index multiplied by a fixed real rate (RR). Increasingly. 15 March 2010 192 . The instrument is equivalent to holding a zero-coupon breakeven inflation swap and a nominal zero coupon swap. These swaps are exposed to realized and expected inflation. The inflation receiver in turn pays a compounded Libor-based rate (Lib). The swap NPV depends on the CPI ratio. Over the life of the swap. the notional and a nominal discount factor (DF). which is sensitive to inflation.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 155: Inflation swap indices 115 110 105 100 95 90 85 80 Oct 06 Apr 07 Oct 07 Apr 08 Oct 08 Apr 09 Oct 09 Figure 156: Real rate swap indices 125 120 115 110 105 100 95 90 85 Oct 06 Apr 07 Oct 07 Apr 08 Oct 08 Apr 09 Oct 09 US 10yr inflation swap index ER UK 10yr inflation swap index ER rebased EUR 10yr inflation swap index ER rebased Source: Barclays Capital US 10yr real rate swap index ER UK 10yr real rate swap index ER rebased EUR 10yr real rate swap index ER rebased Source: Barclays Capital Swap structures Zero-coupon breakeven inflation swaps A zero-coupon inflation swap is an exchange of an inflation-linked cash flow and a fixed cash flow at maturity. The fixed rate is the market breakeven of future inflation over the horizon of the trade. These swaps can be used to match a future cash flow.

Total return indices The total return series additionally assume a return coming from a cash investment. The index level on any index business day is derived from the index level at last rebalance and the NPV of the underlying swap. The cash deposit is marked to market each day using interpolated fixes. It rebalances in the same way and at the same time as the underlying swaps.Barclays Capital | Global Inflation-Linked Products – A User’s Guide ⎡Tenor ⎤ ⎡ CPImat ⎤ N × DFmat × ⎢ ×( 1 + RR )Tenor − 1⎥ − N × DFmat × ⎢ ( 1 + Lib× DCF) − 1⎥ ⎢ s ⎥ ⎣ CPIbase ⎦ ⎣ ⎦ ∏ Figure 157: An inflation swap transaction Fixed Breakeven Rate Figure 158: A real rate inflation swap transaction Libor compounded Client Barclays Fixed Breakeven Rate Client Barclays Actual Inflation Fixed Nominal Swap rate Actual Inflation Barclays Source: Barclays Capital Source: Barclays Capital Excess return indices Index performance is driven by the net present value of the underlying swap. 15 March 2010 193 . The official closing inflation curves utilise proprietary seasonality models. The excess return indices are also known as the unfunded series. The cash return applied for each market comes from a money market deposit with a tenor commensurate with the floating frequency of a standard interest rate swap in that market. The NPV is calculated using Barclays Capital analytics and closing curves. All of the indices assume one is receiving inflation and paying a fixed breakeven rate on the underlying swap. For example. the EUR swaps assume a cash return based on a 6m EURIBOR deposit held for one month. The total return indices are also known as the funded indices.

EUR. and FSWI is the sub page for the total return indices.0 -0.barcap. Portfolio & Risk Solutions team for more information.5 0. Please contact the Index.0 -2. are also available on Bloomberg page BSWM. including nominal swap indices.com – keyword index. 15 March 2010 194 . The indices are priced using Barclays Capital official closing curves. UK and US swap indices. The indices are published to four decimal places.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 159: Index family conventions assumptions summary table EUR Inflation lag Inflation lag interpolation Spot Floating leg Fixed leg Floating basis Fixed basis Roll convention Index calendar Index inception date Source: Barclays Capital US 3 months Linear between 2m and 3m T+2 Quarterly Semi-Annual ACT/360 30/360 MF London & New York 31/10/2006 UK 2 months No T Semi-Annual Semi-Annual ACT/365 ACT/365 MF London 31/05/2006 AUD 3 months Last available CPI is the base T+1 Semi-Annual Semi-Annual ACT/365 ACT/365 MF Sydney.0 -3. London & Tokyo 31/10/2007 Japan 3 months Linear T+2 Semi-Annual Semi-Annual ACT/360 ACT/365 MF London & Tokyo 27/04/2007 3 months No T+2 Semi-Annual Annual ACT/360 30/360 MF London & TARGET 29/09/2006 Figure 160: Total return and excess return inflation swap indices 115 110 105 100 95 90 Sep 06 Mar 07 Sep 07 Mar 08 Sep 08 Mar 09 Sep 09 EUR 10yr inflation swap index TR EUR 10yr inflation swap index ER Source: Barclays Capital Figure 161: Japan 10y inflation swap index versus breakevens 110 100 90 80 70 60 50 Apr 07 1.0 -1. https://live.5 Oct 07 Apr 08 Oct 08 Apr 09 Oct 09 Japan 10yr inflation swap index ER Japan 10yr swap breakeven yields (RHS) Source: Barclays Capital Additional information Swap present value are marked to market using standard market conventions. BSWI is the sub page for excess return inflation swaps.5 -2. Levels and supporting analytics are available on the index website.5 -3.0 0.5 -1.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide The indices aforementioned refer to the Barclays Capital Series-B inflation swap indices. Please also see the funding methodology amendment document published by the Index Products Group Research on 6 March 2009. 15 March 2010 195 . This document is an update to “The Barclays Capital Inflation swap Indices Family” index document published on 9 November 2007.

6. An Figure 162: The TWIST Index and real swap rates 105 103 101 99 97 95 93 91 89 87 1. TWIST Index constituents Barclays Capital Inflation Swap Indices (Series B) The TWIST Index is constructed using Barclays Capital US CPI Inflation Swap indices.com The Barclays Capital TWIST Index provides investors with cost-efficient tradable access to inflation using zero-coupon breakeven inflation swap indices which are termstructure weighted in line with the distribution of TIPS in the US Government InflationLinked Bond Index.93 0.16 Ann. Barclays Capital Zero Coupon US Inflation Swap indices. The TWIST Index is an excess return index and is tradable. The weighting scheme reflects broad exposure across the inflation curve and also means the index is sensitive to changes in linker issuance patterns. Return -1. The weighting is derived from the amount of outstanding bonds. Vol.35 85 0.37 Worst return -6.0 2.Barclays Capital | Global Inflation-Linked Products – A User’s Guide TWIST Index: Treasury-weighted inflation swap exposure across the curve Rahul Sharma +65 6308 3982 rahul.5 3.72 0.0 Oct-06 Apr-07 Oct-07 Apr-08 Oct-08 Apr-09 Oct-09 TWIST Index Source: Barclays Capital Source: Barclays Capital US 10yr Real Swap Rates (RHS) 15 March 2010 196 .55 Source: Barclays Capital 2. The TWIST Index can be used as a swap breakeven inflation benchmark or as an overlay tool for investors wishing to access inflation as a hedge for their portfolios.46 US CPI 0. The index is not correlated to real yields in contrast to inflation linked bond indices. The index delivers liquid breakeven inflation exposure in a cost-efficient manner as compared to other approaches. The index exhibits low correlation to real yields. split by certain maturity bucket.0 Figure 164: Summary monthly returns statistics Since inception (%) TWIST index Ann.5 3.5 Figure 163: Monthly returns correlations against TWIST index Since inception (%) Correlation 10y real 10y swap 10y bond swap rates BEI rates BEI rates TIPS index -0.14 Best return 3. referencing US CPI.57 Worst drawdown -13.15 0.5 1.0 0. are weighted each month to give a composite return that drives the performance of TWIST.sharma@barcap. in the Barclays Capital US Government Inflation-Linked Bond Index (the US TIPS index). TWIST Index overview The TWIST Index reflects the return of inflation swaps across the entire term structure of the US inflation market. Since the index is constructed using inflation swaps the index performance is exposed to both realized and expected inflation.

Each inflation swap index replicates the performance of continuously investing in zerocoupon breakeven inflation swaps. so the indices maintain a constant maturity exposure on a month-to-month basis. Barclays Capital US Government Inflation-Linked Index The Barclays Capital US Government Inflation-Linked Bond Index measures the performance of the US Treasury Inflation Protected Securities (TIPS) market. The swap is rebalanced at month-end.Barclays Capital | Global Inflation-Linked Products – A User’s Guide inflation swap is an instrument that exchanges a fixed breakeven rate for actual future inflation. The holder of the swap receives inflation and pays the fixed breakeven rate. TIPS in the index have a total outstanding issue size of $500mn or more. The index includes TIPS with one or more years remaining until maturity. Figure 165: Barclays Capital US inflation swap indices 110 105 100 95 90 85 80 75 Oct-06 Apr-07 Oct-07 Apr-08 Oct-08 Apr-09 Oct-09 Figure 166: US Government Inflation-Linked index 245 225 205 185 165 145 125 105 85 Feb 97 Feb 99 Feb 01 Feb 03 Feb 05 Feb 07 Feb 09 US TIPS all maturities TR index Source: Barclays Capital USD 2yr inflation swap index USD 5yr inflation swap index USD 10yr inflation swap index USD 20yr inflation swap index Source: Barclays Capital 15 March 2010 197 . Sub-indices available by maturity. Largest component of the World Government Inflation-Linked Index.

Figure 167: The mapping of the weights Inflation-Linked maturity bucket US Government Inflation-Linked 1-3y bucket US Government Inflation-Linked 3-5y and 5-7y bucket US Government Inflation-Linked 7-10y bucket US Government Inflation-Linked >10y bucket Source: Barclays Capital Swap index used in the mapping scheme (series B) USD Zero Coupon Inflation Swap 2y ER index USD Zero Coupon Inflation Swap 5y ER index USD Zero Coupon Inflation Swap 10y ER index USD Zero Coupon Inflation Swap 20y ER index Figure 168: Historical swap index weighting for TWIST Index 0.Barclays Capital | Global Inflation-Linked Products – A User’s Guide TWIST Index weighting scheme Inflation swap indices weighted by TIPS face value In the TWIST Index a selection of inflation swap indices are weighted by the face value. This scheme is fixed.40 0. of bonds in the Barclays Capital Government Inflation-Linked Bond Index (the TIPS Index). (split by maturity baskets). Index return is a composite return measure from the last rebalance date.25 0. The face value weighting is re-evaluated on the last business day of each month when the index rebalances.30 0.00 Oct-06 Apr-07 Oct-07 Apr-08 Oct-08 Apr-09 Oct-09 USD 2yr USD 10yr Source: Barclays Capital Figure 169: Weighting as of end-February 2010 USD 2yr 17% USD 20yr 29% USD 5yr 34% USD 10yr 20% Source: Barclays Capital USD 5yr USD 20yr 15 March 2010 198 .35 0.20 0. Note that the 3-5y and 5-7y buckets of the TIPS index are both matched off against the 5y Inflation Swap index. as per the weighting scheme.10 0.15 0.05 0. The weighting scheme This is described in the table in Figure 167.

86 1.54 -2.09 2.19 2.92 0.01 0.5 2.86 2.33 -0. Portfolio & Risk Solutions for more information.0 1.11 -2.55 -1.57 0. Index levels and analytics are available on the Barclays Capital Index Website.37 -2.87 -0. keyword:index The index is also available on Bloomberg page BXIITWSU Index <GO>.5 1.64 -0.56 0. The index rebalances at month-end.37 -0.5 Oct-06 Apr-07 Oct-07 Apr-08 Oct-08 Apr-09 Oct-09 TWIST index monthly US 10yr Breakeven Swap Rates (RHS) Source: Barclays Capital 85 0.72 -0.0 2.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Performance and additional information Figure 170: TWIST versus swap breakeven rates 103 101 99 97 95 93 91 89 87 3.52 0.5 2.41 0.04 0.02 Source: Barclays Capital Additional information The index inception date is 31 October 2006.barcap.5 3.39 1.92 0.02 Feb Mar Apr May Jun Jul Aug Sep Oct Nov -0.32 -0.38 3.52 -1. Please contact the Index. 15 March 2010 199 .59 -0.34 -6.5 85 0.5 1. https://live.0 Oct-06 Apr-07 Oct-07 Apr-08 Oct-08 Apr-09 Oct-09 TWIST index monthly 10yr Generic Bond Breaeven rates (RHS) Source: Barclays Capital Figure 172: TWIST Index monthly returns since inception (%) Jan 2006 2007 2008 2009 0.0 2.92 Dec -0.0 1.91 -0.79 -1.0 Figure 171: TWIST versus bond breakeven rates 103 101 99 97 95 93 91 89 87 3.17 0.com.0 0. The index is published on New York and London business days.19 -0.68 0.7 2.19 -0.

which in principle is the rate of inflation that will equate the returns on an inflation-linked bond and a “comparator” nominal bond issue of the same term. Global Rates Weekly.. the short positions (and.venkataraman @barcap. therefore. 15 March 2010 200 .com Investors seeking a hedge for inflation have several options at their disposal. long positions in inflation-protected securities. thereby letting the cash “lent” during the period earn a return determined by an appropriate general collateral repo rate.harman@barcap.let us count the ways”. 23 July 2009.. Another investment that has shown an even higher correlation with inflation is an investment in breakeven inflation. Index description The Barclays Capital US breakeven inflation benchmark indices are designed to provide access to breakeven inflation by capturing the returns of a simultaneous long position in inflation-linked securities (represented by OTR US TIPS) and a short position (facilitated by a repo agreement) in suitable nominal comparator US Treasury bonds. 15 How to Hedge Inflation. while minimising exposure to real yields Index rationale Scott Harman +44 (0) 20 7773 1775 scott. 16 On-the-run represents the most recently issued TIPS of a given tenor. This methodology forms the basis for construction of the Barclays Capital US Breakeven Inflation Benchmark Indices. However. the reverse repo agreements) are scaled appropriately as described earlier. The short position in nominal comparator Treasury bonds is facilitated by entering into reverse repo agreements (of buying the Treasury bonds for a certain period and selling them back at the end of the period). to minimise exposure to real yields. The Barclays Capital US Breakeven Inflation Aggregate Index represents a composite breakeven inflation position by taking long positions in all the prevailing bonds spanning the entire term structure of the OTR TIPS market and short positions in their respective nominal comparator Treasury bonds. or the use of inflation swaps. including investments in commodities.com The Barclays Capital US breakeven inflation benchmark indices are designed to represent a breakeven inflation position by taking positions across the term structure of the on-the-run (OTR) TIPS market. December 2009) Anand Venkataraman +44 (0) 20 7773 0852 anand.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Barclays Capital US breakeven inflation benchmark indices (This is an edited extract from Barclays Capital US Breakeven Inflation Benchmark Indices: Capturing US Breakeven Inflation. 15 Such an investment in breakeven inflation can be obtained by taking a long position in inflation-linked securities and a short position in a suitable nominal comparator bond. Barclays Capital. 16 The indices also aim to minimise exposure to real yields by scaling the total returns of the nominal comparator Treasury bonds by a ratio equal to the duration of the OTR TIPS to its corresponding nominal comparator Treasury bond.

The breakeven inflation index at a specific tenor (for instance. The single bond indices are total return bond indices calculated in the same manner as other inflation-linked bond indices and nominal Treasury bond indices. The Breakeven Inflation Index also includes a return on the cash lent as part of the repo agreement (as described earlier).Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 173: Anatomy of a breakeven inflation position Treasury Bond bond Cash Cash Breakeven Inflation Position TIPS Treasury Bond bond Cash Repo Seller Source: Barclays Capital Barclays Capital US Breakeven Inflation Benchmark Indices designed to reflect breakeven inflation position Indices represented by long positions in OTR TIPS and short positions in nominal comparator Treasury bonds Indices designed to minimise exposure to real yields Figure 174: Barclays Capital US Breakeven Inflation Aggregate Index Barclays Capital US Breakeven Inflation Aggregate Index US 5yr Breakeven Index US 10yr Breakeven Index US 20yr Breakeven Index New Breakeven Tenor 5yr TIPS OTR Index Source: Barclays Capital 5yr Nom Index 10yr TIPS OTR Index 10yr Nom Index 20yr TIPS OTR Index 20yr Nom Index New Tenor (TIPS) New Tenor(Nom) Index framework and calculation methodology The Barclays Capital US Breakeven Inflation Aggregate Index is an equally weighted composite index comprising individual breakeven inflation indices at all the prevailing tenor points on the OTR TIPS market term structure. 15 March 2010 201 . respectively. the Barclays Capital US five-year Breakeven Inflation Index) is constructed using two single bond total return indices. one consisting of the OTR TIPS bond and the other consisting of the nominal comparator Treasury bond.

b = ⎛ 1+ ⎜ rt ⎝ Where Rt . one-week.b index consisting of the Rt . The interpolated repo rate on any day is calculated based on days remaining until end of prevailing repo period by linear interpolation of prevailing day’s overnight.b − SF b × (Nom t .b SF b ⎛ LD SF b = ⎜ ⎜ ND ⎝ Where LDb NDb ⎞ ⎟ ⎟ ⎠ is the real semi-annual modified duration for OTR TIPS at the tenor is the semi-annual modified duration of the corresponding nominal comparator bond The repo return component of the index is designed to represent the return on cash “lent” as part of the reverse repo agreement. is calculated using the following equation: ILt = ILb × [(TIPS t . The repo return component is given by: repo _ period ⎞ ⎛ 1+ ⎜ rb . at any given tenor point (for Instance. two-week. A new reverse repo agreement is entered into at each rebalance date maturing at the next index rebalance date (usually one month) since the indices rebalance monthly. as described below. three-week and one-month general collateral repo rates as applicable.b is the repo return component between t and b is the one-month general collateral repo rate is the interpolated repo rate on day t.b − R t .1m rt repo _ period (ie.Barclays Capital | Global Inflation-Linked Products – A User’s Guide This index. days _ till _ repo _ period _ end the prevailing repo period 15 March 2010 202 . 5y).b ) + 1] Where ILt ILb is the Index level on day t is the Index level at previous rebalance day b is the total return between t and b of a single bond index consisting of the OTR TIPS is the total return between t and b of a single bond nominal comparator Treasury bond is the repo return component between t and b is the scaling factor and is calculated as: b b TIPS i .t Nomt . is the actual length in days of the repo period at the previous rebalancing date days between previous rebalancing date and the following rebalancing date) is the number of days remaining as on t till the end of rb.1m × ⎟ 360 ⎝ ⎠ −1 days _ till _ repo _ period _ end ⎞ × ⎟ 360 ⎠ R t .

and liquidity among other determinants.b )]× w i . At each monthly rebalancing. 15 March 2010 203 . Notwithstanding.t − SFi × (Nomi . and all remaining tenors will be equally weighted. the applicable nominal comparator Treasury bond is determined at each monthly rebalancing based on market conventions.t w i . In the event that the Treasury establishes a new tenor point at which TIPS will be issued (for instance.t Nomi . Index rebalancing Presently. a tenor point would be removed from the Aggregate Index if there had been no new TIPS issuance for a 2y period at the tenor point. b SFi n At inception.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The Barclays Capital US Breakeven Inflation Aggregate Index is calculated as: ⎫ ⎧n ILC t = ILC b × ⎨∑ [(TIPSi . the Aggregate Index would weight all the tenor points equally at all times. To reflect the mechanics of a breakeven inflation position consistently. Note that each tenor point is equally weighted is the repo return component as on day t is the scaling factor for a given tenor calculated as described earlier is the number of available tenor points on the OTR TIPS market term structure ILC b TIPS i .b Rt . the new tenor would be included in the Aggregate Index at the subsequent monthly rebalancing. the Breakeven Inflation Index at each individual tenor point would switch the underlying inflation-linked security only in case a new TIPS is issued at the given tenor point. 10y and 20y points. the Barclays Capital US Breakeven Inflation Aggregate Index had equalweighted exposure to the three prevailing OTR tenor points: the 5y. Upon the two-year anniversary since the last occasion a TIPS was issued at a given tenor point. Conversely. bond(s) issued with the same (or closest) original term as the OTR TIPS.t − Rt . The nominal comparator bonds generally switch more frequently than the corresponding OTR TIPS. Some key considerations while establishing the applicable nominal comparator Treasury bond for an OTR TIPS are bond(s) having the closest time-to-maturity as the OTR TIPS.b + 1 ⎬ ⎭ ⎩ i =1 Where: ILCt is the Aggregate Index level on day t is the Aggregate Index level at previous rebalance day b is the total return of the single bond TIPS total return index at a given tenor on day t since last rebalance day b is the total return of the single nominal comparator bond index for the corresponding tenor on day t since last rebalance day b is the weight of an individual tenor within the Aggregate Index. all of which rebalance on the last calendar day of each month in line with conventional bond Indices. thereby becoming the prevailing OTR TIPS.17 17 Please refer to the Inflation-Linked Daily for contemporary comparator bond listings. due to the depth of the Treasury market relative to the TIPS market. that tenor point would be removed from the Barclays Capital US Breakeven Inflation Aggregate Index at monthly rebalancing. the new 30y in February 2010). the Barclays Capital US Breakeven Inflation Benchmark Index family comprises 10 separate indices (Figure 174).

Figure 175: NSTEP Index historical performance 140 135 130 125 120 115 110 105 100 95 90 Jan 00 Jul 01 Jan 03 Jul 04 Jan 06 Jul 07 Jan 09 Figure 176: Chart Title 114 112 110 108 106 104 102 100 98 Sep 06 M ar 07 Sep 07 M ar 08 Sep 08 M ar 09 Sep 09 EUR 10yr ZC Inflati on Sw ap index (TR) INSTEP index Source: Barclays Capital INSTEP Index rebased Source: Barclays Capital 15 March 2010 204 . The short position is based on borrowing nominal bonds in the market and having the cash lent earn a repo rate. The index provides beta-style access to breakeven inflation in the euro government bond market. The INSTEP Index constituents are: − − − The Barclays Capital Euro Government Inflation-Linked All Maturities Index The Barclays Capital Euro Government Comparator All Maturities Index EUR General Collateral 1m Repo (EUREPO) Rate The index is a bond space alternative to breakevens and is similar to the Barclays Capital Inflation Swap Indices (series-B).sharma@barcap.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The INSTEP Index Rahul Sharma +65 6308 3982 rahul. The index exhibits low correlation with real yields. An introduction to the INSTEP Index The INSTEP index is long an all-maturities inflation-linked bond index and short a nominal comparator bond Index.com The Barclays Capital INSTEP Index offers investors duration-adjusted breakeven inflation exposure in the euro government bond market. which replicate the performance of investing in zerocoupon inflation swaps in which one party pays a fixed rate and receives the cumulative percentage change in a price index over the life of the swap. The future path of the INSTEP Index is sensitive to changes in future inflation expectations and to the actual path of inflation indexation. The linker and nominal bonds exposure is rates-duration-hedged to minimise real yield risk.

funded at 1m GC repo. for the back-test prior to March 2002. and the net position is insensitive to changes in real yields. The index rebalances on the last business day of each month. by a durationadjusted amount. However. BBA fixes were used. Portfolio & Risk Solutions team for more information. Instep = Linker . The index is published to four decimal places.Repo) Figure 177: INSTEP Index monthly returns correlations with yields Index/Rate Euro 10y bond breakevens Euro gov inflation-linked index average real yield EUR 10y breakeven inflation swap index EUR 10y real rate swap index Source: Barclays Capital Monthly correlation 0. The original version of the INSTEP Index.barcap.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The INSTEP strategy The INSTEP Index strategy is long linkers.Nov 09 Oct 06 . the nominal breakeven comparator index.Nov 09 An inflation-linked bond is sensitive to real yields and inflation indexation. Please contact the Index.Nov 09 Feb 00 . On each rebalance date. A nominal bond. The index is published with a one day lag on the Barclays Capital Index Website at https://. was launched in August 2008. the INSTEP index goes long the linkers index and short. due to the repo component. based on a prior methodology. given the above. Additional information The INSTEP Index inception date is 31 January 2000. The index is also available on Bloomberg page BXIIEXTE Index.Dur ratio x (Nominal . Also.932 0.Nov 09 Oct 06 . The repo component The repo position is rebalanced each month.live.921 -0. This applies on an average annual-modified-duration basis since the index durations we are using to hedge are average durations for all maturities indices. The repo deposit is marked-to-market each day. which is backward looking. European Banking Federation repo fixings (Bloomberg page EBF) are used for the index calculation. is sensitive to real yields and forward-looking inflation expectations. 15 March 2010 205 .236 0. short nominal comparators (duration ratio adjusted). the net index position has a small sensitivity to 1m rates. The real yield components offset each other.com (keyword: index).381 Period Dec 03 . The maturity of each deposit is assumed to coincide with the next index rebalance date. The index is calculated on TARGET business days.

In particular. Known as the breakeven inflation rate.15 0. with currently five issues outstanding and a face value of CAD28bn.00 -0.com Yuan Tian +44 (0)20 7773 1426 yuan. such as inflation-linked bonds or inflation swaps. justifying the use of a rolling optimisation model. this measure contains information on the average expected inflation over a certain tenor.77% Govt Semi Govt Non Govt 77. The Canadian inflation market is relatively small and illiquid. Demand for a Canadian inflation derivatives market is so strong at the moment.tian@barcap. UK and EUR inflation. UK and euro area. a market-based measure of the expected path of future inflation is contained within inflationlinked bonds or swaps.christodoulou @barcap. The Canadian government started issuing inflation-linked bonds in 1991. Figure 178: Canada inflation is highly correlated with developed inflation markets 0. despite significant demand for inflation protection from both domestic pension funds and international investors. that some Canadian investors have even considered going to the US inflation swap market as a proxy for Canadian inflation. 15 March 2010 206 .Barclays Capital | Global Inflation-Linked Products – A User’s Guide CAD INSPIRE Michalis Christodoulou +44 (0)20 3134 2751 michalis.com Marcela Barreto +44 (0)20 3134 0750 marcela. Canadian Inflation exhibits significant time-varying correlation features with US. Non-government inflation-linked bond issuance has been slow to develop in Canada. Inflation-linked instruments. Barclays Capital offers an innovative solution to tackle such a challenge: The CAD INSPIRE Index Family takes advantage of Canadian inflation being correlated with developed inflation markets. allow investors to express views or hedge against the future path of inflation.66% Jan-81 Jan-86 Jan-91 Jan-96 Jan-01 Jan-06 Source: Barclays Capital Source: Barclays Capital The lack of an inflation swap market in Canada makes it challenging for investors to create economically efficient portfolios and hedge against inflationary trends.25 0.barreto@barcap.20 0.05 Jan-76 CAD yoy inflation US yoy inflation UK yoy inflation EUR yoy inflation Figure 179: Canada inflation-linked bond market breakdown 3.10 0. with the amount of corporate issuance being rather negligible as of the end of 2009.com The Barclays Capital CAD INSPIRE Index family is designed to provide synthetic inflation protection for Canada using an optimised weighted combination of actively traded and liquid inflation swap indices from the US.05 0.56% 18.

g. 15 March 2010 207 . Therefore. Minimising transaction costs embedded in the index while achieving the best possible correlation features. an inflation measure has been constructed that reflects similar statistical properties to the breakevens. The illiquid nature of the Canadian inflation market calls for an alternative inflation measure to breakevens. Inflation measure and replica The breakeven rate from an inflation swap is equal to the average expected inflation over the tenor of the swap.. using liquid markets’ inflation instruments. The spirit of the INSPIRE framework is to replicate inflationary trends in illiquid countries. Replicate the Canadian inflation measure as a dynamically weighted combination of the US. maintaining cost efficiency. UK. the 20yr inflation measure) 2. making it a good inflation measure for any given tenor. UK and EUR zero-coupon inflation swap breakeven rates.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation being a global phenomenon implies that inflationary trends are shared across different regions which will in turn feed into market expectations on future inflation and thus breakevens. Taking Canada as our illiquid inflation market. EUR model BE w 104 102 100 98 96 94 92 90 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Source: Barclays Capital Inflation measures and breakevens CAD AUD IM CAD INSPIRE indices CAD INSPIRE 20yr Index INSPIRE model’s main objectives Replicating inflationary trends of CAD by dynamically combining US. UK and EUR inflation swap breakevens. Compute a specific tenor inflation measure for Canada (e. the INSPIRE framework follows a two-step process: 1. Overview of the INSPIRE framework Find optimal linear combination of core countries' breakevens that replicate inflation measure in Canada 106 Optimised weights INSPIRE optimisation Core Countries US.

5% 3. The 20y Canadian inflation measure shown is in line with inflationary developments worldwide.0% 2. The CAD INSPIRE Index is offered in 5y.0% 3. indicating a very good fit throughout the sample for various tenors. UK and EUR to produce the CAD INSPIRE Index return.0% 0. subject to cost constraints.5% 1. An analogous empirical measure to the breakevens is therefore constructed by averaging the interest rate forwards over a certain tenor and subtracting a constant real rate. UK and EUR breakeven replica is around 80-95%.0% 2.0% 3. Figure 180: CAD 20y inflation measure 4. Clients.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Forward interest rates contain information about average expected inflation and expected real rate for a future period.0% 0. Type BXIICI05 Index <GO> for the CAD INSPIRE 5y Index Type BXIICI10 Index <GO> for the CAD INSPIRE 10y Index Type BXIICI20 Index <GO> for the CAD INSPIRE 20y Index 15 March 2010 208 .0% Apr-01 Apr-03 Apr-05 Apr-07 Apr-09 CAD INSPIRE replica CAD 20yr inflation measure Source: Barclays Capital Source: Barclays Capital Inspire optimisation model’s fit The dynamic INSPIRE optimisation model estimates monthly weights such that it minimises the difference between our Canadian empirical inflation measure and a weighted linear combination of US. 10y and 20y maturities and returns are expressed in AUD. UK and EUR breakevens.5% 2.5% 2. The correlation achieved between our Canadian inflation measure and its corresponding US.5% 0. It can be offered at any additional tenor on request and can be used flexibly according to investors’ needs. once given permission.0% 1.5% 4. can access the index data on Bloomberg.0% 1.5% 4.5% 1. Additional information Barclays Capital CAD INSPIRE indices start on 30 January 2004. The resulting monthly weights are then applied to the returns of the zero-coupon inflation swap indices of US.0% May-00 May-02 May-04 May-06 May-08 CAD 20yr inflation measure Figure 181: CAD INSPIRE fit versus inflation measure 4.5% 0.5% 3.

which accounts for 42% of the total inflation issuance. Australian Inflation exhibits significant time-varying correlation features with US. Over the past seven years. Known as the breakeven inflation rate.3bn. The current outstanding notional of AUD inflation-linked bonds issued is approximately AUD24. UK and EUR inflation.6bn.com The Barclays Capital AUD INSPIRE Index family is designed to provide synthetic inflation protection for Australia using an optimised weighted combination of actively traded and liquid inflation swap indices from the US. The AUD INSPIRE Index takes advantage of Australian inflation being correlated with developed inflation markets. allow investors to express views or hedge against the future path of inflation.Barclays Capital | Global Inflation-Linked Products – A User’s Guide AUD INSPIRE Michalis Christodoulou +44 (0)20 3134 2751 michalis. The AUD inflation swap market started in 2006. Inflation-linked instruments.christodoulou @barcap.barreto@barcap. In particular. with limited liquidity and no standard benchmark. 15 March 2010 209 . There are four such bonds currently outstanding with a total face value of AUD10. Barclays Capital offers an innovative solution to tackle such a challenge: The AUD INSPIRE Index Family.com Yuan Tian +44 (0)20 7773 1426 yuan. and trading activities have been relatively sporadic.com Marcela Barreto +44 (0)20 3134 0750 marcela. a market-based measure of the expected path of future inflation is contained within inflationlinked bonds or swaps. Figure 183: Australian inflation-linked bond current market breakdown GOVT 29% SEMIGOVT NONGOVT 42% Figure 182: Australian inflation is highly correlated with developed inflation markets 25% 20% 15% 10% 5% 0% -5% Mar-76 Mar-81 Mar-86 Mar-91 Mar-96 Mar-01 Mar-06 Source: Barclays Capital AUD yoy inflation US yoy inflation UK yoy inflation EUR yoy inflation 29% Source: Barclays Capital Given the illiquidity of the Australian inflation swap market. The Australian market has a large number of small issuers that issue inflation-linked bonds. the Australian government has issued only one inflationlinked bond.tian@barcap. UK and euro area. such as inflation-linked bonds or inflation swaps. The Australian inflation market is relatively under-developed and illiquid. this measure contains information on the average expected inflation over a certain tenor. it is challenging for investors to create economically efficient portfolios and hedge against Inflationary trends. justifying the use of a rolling optimisation model.

EUR BE model ? w 100 95 90 85 Jan 04 Nov 04 Sep 05 Jul 06 May 07 Mar 08 Jan 09 Nov 09 105 replicate inflation measure in Australia 110 AUD INSPIRE AUD INSPIRE 10yr Index AUD INSPIRE 5yr Index Source: Barclays Capital INSPIRE model’s main objectives Replicating inflationary trends of AUD by dynamically combining US. UK. Compute a specific tenor inflation measure for Australia (e. maintaining cost efficiency Overview of the INSPIRE framework Find optimal linear combination of core countries' breakevens that Inflation measures and breakevens AUD Optimised weights IM INSPIRE optimisation Core Countries US. making it a good inflation measure for any given tenor. The illiquid nature of the Australian inflation market calls for an alternative inflation measure to breakevens. using liquid markets’ inflation instruments. UK and EUR inflation swap breakevens. Minimising transaction costs embedded in the index while achieving the best possible correlation features. UK and EUR zero-coupon inflation swap breakeven rates. The spirit of the INSPIRE framework is to replicate inflationary trends in illiquid countries. Taking Australia as our illiquid inflation market. the 5y or 10y inflation measure) 2..g. 15 March 2010 210 . which will in turn feed into market expectations on future inflation and thus breakevens.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation being a global phenomenon implies that inflationary trends are shared across different regions. Replicate the Australian inflation measure as a dynamically weighted combination of the US. Inflation measure and replica The breakeven rate from an inflation swap is equal to the average expected inflation over the tenor of the swap. the INSPIRE framework follows a two-step process: 1.

0% 2. an inflation measure has been constructed that reflects similar statistical properties to the breakevens. can access the index data on Bloomberg.0% AUD 10yr inflation measure 4.0% 0.0% 2.0% 3. indicating a very good fit throughout the sample for various tenors. subject to cost constraints. The 10y Australian inflation measure shown is in line with inflationary developments worldwide. Clients. It can be offered at any additional tenor upon request and can be used flexibly according to investors’ needs. The resulting monthly weights are then applied to the returns of the zero-coupon inflation swap indices of US. An analogous empirical measure to the breakevens is therefore constructed by averaging the interest rate forwards over a certain tenor and subtracting a constant real rate. Forward interest rates contain information about average expected inflation and the expected real rate for a future period. UK and EUR to produce the AUD INSPIRE Index return. once given permission. Additional information Barclays Capital AUD INSPIRE indices start on 30 January 2004.0% 1.0% 3.0% 1. The correlation achieved between our Australian inflation measure and its corresponding US. The AUD INSPIRE Index is offered in 5y.0% Apr-01 Jun-02 Aug-03 Oct-04 Dec-05 Feb-07 Apr-08 Jun-09 Source: Barclays Capital Figure 185: AUD INSPIRE fit versus inflation measure 5. UK and EUR breakeven replica is around 80-90%. 10y and 20y maturities and returns are expressed in AUD. UK and EUR breakevens. Type BXIIAII05 Index <GO> for the AUD INSPIRE 5y Index Type BXIIAII10 Index <GO> for the AUD INSPIRE 10y Index Type BXIIAII20 Index <GO> for the AUD INSPIRE 20y Index 15 March 2010 211 . Figure 184: AUD 10y inflation measure 5.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Therefore.0% 4.0% 0.0% Apr-01 Jun-02 Aug-03Oct-04 Dec-05Feb-07 Apr-08 Jun-09 Source: Barclays Capital AUD 10yr inflation measure AUD INSPIRE replica Inspire optimisation model’s fit The dynamic INSPIRE optimisation model estimates monthly weights such that it minimises the difference between our Australian empirical inflation measure and a weighted linear combination of US.

Key characteristics of the index Rules-based index aimed at replicating the performance of monthly constant real cash flows by using TIPS. The rate of the constant real cash flows is set at a level that makes the present value of the cash flows equal to par. assuming the current index holdings (as of October 2009 ) do not change over time. The Barclays Capital US TIPS Real Income Index family consists of the Barclays Capital US TIPS Real Income 2019 index and the Barclays Capital US TIPS Real Income 2029 Index. Figure 186 compares the annual real cash flows (coupon + principal) of the Barclays Capital 1-10 year US TIPS Index versus the US TIPS Real Income 2019 Index.com (Previously published in Barclays Capital US TIPS Real Income 2019/2029 Index. market value weights are calculated using the outstanding amount and the market price of bonds.mazoy@barcap. Once the size of the cash flows is determined. Bi-weekly rebalancing. Eligible bonds are US TIPS issued on or before the Index Rebalance Day and maturing at least 20 calendar days after the Index Rebalance Day. TIPS coupons are immediately reinvested into the index upon payment.jiang@barcap. and it is adjusted bi-weekly to reflect changes in real rates. The US TIPS Real Income Index bond weights are determined with the aim of generating a flat real cash flow profile over the index tenor. The stream of cash flows has a specific maturity date and is constructed in a way that the invested principal is depleted over the index tenor. The portfolio of replicating US TIPS is determined on each index rebalance date using an algorithm based on cash flow matching and risk minimization. As compared with market value weighted TIPS indices. the Barclays Capital US TIPS Real Income indices have the advantage of not being influenced by the government bond issuance pattern. a portfolio of US TIPS bonds is constructed aiming to have the cash flows of the portfolio as close as possible to the targeted monthly cash flows. 15 March 2010 212 . November 2009) The Barclays Capital US TIPS Real Income 2019/2029 indices track the performance of a stream of constant real cash flows by adaptively taking positions on a basket of US TIPS bonds Index description The Barclays Capital US TIPS Real Income indices are designed to track the performance of a stream of constant real cash flows by adaptively taking positions on a basket of US TIPS. On the other hand. therefore creating a cash flow pattern that may deviate significantly from the targeted constant cash flows.com Jenny Jiang +44 (0) 20 7773 0822 Jenny.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Barclays Capital US TIPS Real Income Index family Jose Mazoy +44 (0) 20 3134 0998 jose. The Barclays Capital US TIPS Real Income indices are designed for investors seeking an investment solution that provides steady inflation-protected cash flows over a specific period of time.

15 March 2010 213 . Step 3: Determination of weights by an optimization algorithm incorporating cash-flow matching and risk minimization. The termination dates of the Barclays Capital US TIPS Real Income 2019/2029 indices are 31 October 2019/31 October 2029. The indices utilize a patent-pending proprietary algorithm designed by Pacific Investment Management Company. LLC. Step 2: Selection of eligible TIPS from the universe of TIPS bonds to replicate the target real constant cash flows. Bloomberg Tickers: Barclays Capital US TIPS Real Income 2019 Index (BXIIRI19 Index). Real Constant Cash Flows Selection of Replicating US TIPS Optimization Algorithm Cash flow Matching Risk Minimization Replicating Bond Weights Additional Information The base date of the Barclays Capital US TIPS Real Income 2019/2029 indices is 14 October 2009. respectively.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 186: Real cash flows comparison 18% 16% Barclays Capital US TIPS Real Income 2019 index Barclays Capital US TIPS (Tenor 1-10 years) index 14% 12% 10% 8% 2010 Source: Barclays Capital 2011 2012 2013 2014 2015 2016 2017 Index framework Step 1: Determination of the target real constant cash flows using information from the real curve. Barclays Capital US TIPS Real Income 2029 Index (BXIIRI29 Index). The indices follow the New York business day calendar.

but switch to short TIPS positions if inflation expectations fall. This price pressure caused by size imbalances presents an opportunity to extract excess return provided we identify the changes in current and future inflation expectations in a timely manner. Index rationale Investors who are concerned that realized inflation risk would erode the value of their nominal fixed income investment tend to focus on inflation-linked assets.500 1.com The Barclays Capital AIMS (Algorithmic Inflation Momentum Switching) Index seeks to enhance returns generated by a long-only portfolio of inflation-linked securities. The AIMS algorithm uses publicly available data and exploits the link between US inflation expectations and movements in TIPS prices. The Barclays Capital AIMS Index seeks to enhance returns generated by a long-only portfolio of inflation-linked securities by adopting a fully non-discretionary algorithm.500 4. The nominal investment grade bond market and the TIPS market have long exhibited imbalances in supply and demand.com Jose Mazoy +44 (0) 20 3134 0998 jose.000 4.500 2. The index takes a long position in a TIPS basket as default. issued by the US Treasury. The rationale of the underlying AIMS strategy is that investors switch from nominal Treasuries (and other bonds) to TIPS when concerned about inflation.Barclays Capital | Global Inflation-Linked Products – A User’s Guide AIMS Index Yuan Tian +44 (0) 20 7773 1436 yuan. which pay a fixed coupon on a notional that accretes according to the US CPI-U Inflation Index. but underperformed nominal Treasuries in 2008 due to a flight to quality and the relative illiquidity of TIPS compared with nominal Treasuries. stronger price pressure can be observed on TIPS. During periods of heightened inflation expectations. TIPS outperformed nominal Treasury bonds during periods of rising breakeven inflation rates. The nominal Treasuries market is a substantial multiple of the size of the TIPS market. Figure 187:US Treasury versus TIPS Index market capitalization 5.000 500 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Figure 188: US Treasury Index TRI versus TIPS Index TRI 240 220 200 180 160 140 120 100 80 Dec-98 Aug-00 Apr-02 Dec-03 Aug-05 Apr-07 Dec-08 US Treasury All Maturities US Govt Inflation-Linked All Maturities Source: Barclays Capital US Treasury Index Market Cap US Govt Inflation-Linked Index Market Cap Source: Barclays Capital 15 March 2010 214 .mazoy@barcap.000 1. One way of tracking inflation is through Treasury Inflation Protected Securities (TIPS). which allow them to retain the real value of their investment. but switches to a short TIPS position if the signal indicates lowered inflation expectations.tian@barcap.000 3.500 3. Historically.000 2.

any increase or decrease in the current CPI will immediately be priced into the TIPS market value. Therefore. We use a simpler approach in the AIMS index and take the slope of the nominal yield curve as a measure of future inflation expectations. Inflation surveys or models could be used to try to infer its value.75 unit of TIPS YES: Long TIPS position No: Short TIPS position The Index is rolled every month following the release of US CPI AIMS signal indicates inflation concerns? Source: Barclays Capital 15 March 2010 215 . reflecting a current inflationary trend. The allocation to the TIPS position is determined in the following way at the index roll date: Long 1 unit TIPS if Δ C PI t * ≥ 0 or Δ Slope t * ≥ 0 Otherwise short 0. The current inflationary trend can be monitored by the evolution of the CPI y/y print. the principal in the TIPS is adjusted upward and the interest on the bond is then accrued on a higher principal. reflecting market expectations of higher or lower inflation in the future. since a steep yield curve reflects the expectation of higher nominal rates in the future and therefore higher inflation. The TIPS principal and coupon are also linked to changes in the CPI: as CPI rises. Figure 189: US TIPS y/y return and CPI y/y Index 25% 20% 15% 10% 5% 0% -5% -10% Feb-98 Feb-00 Feb-02 Feb-04 Feb-06 Feb-08 US TIPS index y/y return Source: Barclays Capital Figure 190: Swap slope as an approximation for future inflation 6% 5% 4% 3% 2% 1% 0% -1% -2% -3% Change in the 2s10s slope = change in future inflation r Expected increase in future inflation.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Index framework Inflation concerns are both determined by the current realization of inflation as well as future inflation expectations. among other factors 2y 10y Expiry CPI y/y Index Source: Barclays Capital The AIMS mechanics On a monthly basis. the AIMS algorithm switches between long and short TIPS exposure based on two signals: 1) the current CPI y/y print relative to its previous value. Future inflation expectations are not directly observable. and 2) the current value of the US swap 2s10s slope relative to its previous value.

2 0.8 -1. The index takes a long or a short position in a one-month total return swap on the Barclays Capital US Government Inflation-Linked Bond Index Figure 192: AIMS Excess Return Index and Long TIPS Excess Return Index 200 180 160 140 120 100 80 Mar-97 Mar-99 Mar-01 Mar-03 Mar-05 Mar-07 Mar-09 AIMS ER Index Source: Barclays Capital Figure 193: AIMS Excess Return Index return distribution 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% AIMS return distribution Long only TIPS ER Index -6.4% -4.8% 0. a steepening yield curve may be signalling rising long-term inflation expectations.7% 5. the AIMS allocation was mostly long.0 -0. Figure 191: AIMS Index monthly allocation and signal components 6% 5% 4% 3% 2% 1% 0% -1% -2% -3% Mar-97 Sep-98 Mar-00 Sep-01 Mar-03 CPI YOY Sep-04 Slope Mar-06 Sep-07 Mar-09 1. this was due to the fact that the signal determination takes into account the CPI y/y print as well as the slope of the swap curve.2% Source: Barclays Capital 15 March 2010 216 .8 0.4% 2. It has exhibited a bias towards being long TIPS.4 -0.0% 7.2 1.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The AIMS Index has taken a short TIPS position 20% of the time since its inception in 1997 and historically. The index has mostly outperformed the benchmark long TIPS portfolio during periods in which it has taken a short position.4 0.2 -0. Despite falling CPI y/y print.1% -1.0 Monthly Allocation Source: Barclays Capital The AIMS Index performance and characteristics The AIMS Index seeks to enhance returns generated by a long-only portfolio of inflationlinked securities. It is worth noting that during the sharp global recession in 2008-09.0 0.6 -0.6 0.

15 March 2010 217 . The US swap 2s10s slope is calculated using the US ISDA fixings published at 11:30 EST. Bloomberg Tickers: AIMS (BXIIAIMS). AIMS TR series 1 (BXIIAI1). The CPI y/y change is calculated by taking the y/y percentage change on the US CPI for all urban consumers. AIMS Index Inception date is 19 March 1997. This number is published at 8:30 EST at about the middle of each month.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Index characteristics The Barclays Capital AIMS Index is available in USD. The AIMS Index is available in total return and excess return versions. Target AIMS ( BXIITAIM).AIMS TR series 2 (BXIIAI2).

Barclays Capital | Global Inflation-Linked Products – A User’s Guide INFLATION THEMES 15 March 2010 218 .

From an investor’s perspective differentiating between the elements is relatively unimportant. which helps to define inflation-linked bonds as a separate asset class from nominals. a required real yield that investors demand over and above those inflationary expectations. in practice separating out the inflation expectations and risk premium component is extremely complex.james@barcap. to give: (1+n) = (1+r) (1+f) (1+p) Where: n = yield on nominal bond r = real yield on index-linked bond f = inflationary expectations p = risk premium The risk premium reflects the assumption that investors want additional compensation for accepting undesirable inflation risk when holding nominal bonds. the yield premium investors demand in order to hold an inflation-linked bond rather than a nominal because the latter is usually more liquid. The formula states that a nominal bond yield is made up of three components: inflationary expectations. and a risk premium factor. but in practice this is usually offset by a liquidity discount. Calculating it by simply subtracting a real yield from a nominal yield is a crude form of a properly compounded calculation. plus the challenges in assigning effective nominal duration to inflation-linked bonds. which has led to the market “shortcut”: n = r + bei Where: bei = “breakeven inflation” Breakeven inflation In principle. If inflation and interest rates are relatively low.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Breakeven inflation and the relationship between real and nominal yields Alan James +44 (0) 20 7773 2238 alan. While more accurate “Fisher breakevens” have been quoted in the past. devised by Irving Fisher in the inter-war years long before the start of government inflation-linked bond markets. Historically. While real and nominal yields are clearly related and usually move closely together. the formula can be approximated with an additive form: n=r+f+p However. the correlation between the two is often unstable. The Fisher equation. breakeven inflation is the rate of inflation that will equate the returns on an inflation-linked bond and a “comparator” nominal bond of the same term. makes a theoretical connection between real rates and the nominal world. 15 March 2010 219 . market convention has now moved decisively towards simple spreads other than in high-yielding markets. particularly in the UK.com This section discusses breakeven inflation and the difficulty in its interpretation. ie. The presence of inflation-linked bonds allows the substitution of actual real yields for “required real yields” in the formula. breakeven inflation rates in the UK were sometimes calculated in a more complicated way for old style linkers whose real yields require an inflation assumption (3% is the market convention). In this case a semi-annual Fisher breakeven could be used as the start of an iterative process in which the inflation assumption was then adjusted towards a more accurate final ‘true’ breakeven inflation rate.

The Federal Reserve thus refers to inflation compensation. For instance. ie. there is a limited correlation between survey measures of inflation expectations and inflation breakevens even at relatively short maturities. in this example only 1bp. as it is likely to be driven around by relative supply and demand factors. the breakeven rate. in countries where indexation is more embedded. A truer measure of breakeven inflation would be achieved if we were lucky enough to have zero-coupon linkers with no lag and a zero-coupon nominal of identical term. In Europe and the US. By contrast. with a real yield at 2% and a nominal yield at 4%. These distortions are relatively small compared to other difficulties involved that have to be accepted – invariably there is a term mismatch between linker and comparator. with almost no statistical significance between 5y and over expectations and breakevens. Practically it may be feasible to observe when there are sharp changes in this risk premium but not to put any precise value on it.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The equation to calculate Fisher breakeven for a market with an annual yield convention is: (1 + bei ) = (1 + r ) (+ ) 1 n While. the distortion for an annual bond is only 4bp. such as Israel and some Latin American markets. Central banks have frequently attempted to ascertain as pure a measure as possible of inflation expectations. Breakeven spreads are sometimes quoted versus an interpolated nominal curve. Logically the inflation risk premium should tend to increase as the tenor extends. and there is the fact that. The ECB and Bank of England more frequently refer to inflation swaps than the Fed but similarly try to observe and understand changes in the market rather than attempting to mechanically break apart the breakeven component. 1y professional forecasts and market implied inflation tend to move relatively closely together. it is quite possible for inflation risk premia to be negative even before 15 March 2010 220 . the calculation is: ⎛1 n ⎞ ⎜ + ⎟ 2⎠ 1 + bei ) = ⎝ ( 2 ⎛1 r ⎞ ⎜ + ⎟ ⎝ 2⎠ The approximation of using a simple spread is not extreme if yields are relatively low. While conceptually the idea of an inflation risk premium. The degree of liquidity discount for inflation-linked bonds relative to nominal government issues can also vary significantly. demand from investors to hedge their inflation exposures is unlikely to smoothly increase at longer maturities. because of the indexation lag. but with no more success than market participants or academics. for a semi-annual market. particularly in France. In many markets there is no unanimity over which nominal bond to use as a comparator. 2 Separating inflation expectations and risk premia – an elusive goal Deconstructing the breakeven into inflationary expectations and the various elements of risk premia is notoriously difficult. and has various methods of estimating the liquidity discount applicable to TIPS but has not put forward any compelling way of splitting inflation expectations from risk premium. In practice. Given that most investors have a nominal benchmark. although the indexation imperfection would remain. there is reinvestment risk with nominal coupons larger than those in linkers. there is no reason that this factor should be at all stable. an amount of yield that investors will forgo in order to obtain inflation protection is inherently appealing. For semi-annual conventions the approximation is notably closer. the real yield is not pure. along with the degree of inflation uncertainty. ie.

unlike at the long end of the bond curve where there can be contradictory and potentially unstable convexity issues for real and nominal curves.0 3. even if changes in relative bond market liquidity may show up on this measure. A potential distortion that is more commonly an issue shorter on the curve is that in many countries inflation-linked bonds have deflation floors that can distort the meaning of real yield and breakeven valuations when implied inflation is very low or negative. it is not a safe assumption that the difference between bond and swap valuations accurately measures the liquidity discount of cash linkers. eg. However.0 2. in a relatively functional market we would suggest that inflation swaps offer a more realistic starting point for the assessment of inflation expectations than government bond spreads. Despite the measurement advantage of inflation swap rates compared to bond breakevens.5 3. indeed in markets where swaps are less liquid than bonds.0 1998 2000 2002 2005 2007 2009 Source: Barclays Capital Source: University of Michigan. inflation swaps were less distorted by deleveraging pressures than bond breakevens. the distortions can be more extreme and less stable. Other than any indexation lag the inflation swap rate is a pure measure. although liquidity in inflation swaps was particularly limited in most markets. During late 2008.0 0.5 2. whereas the default zero-coupon inflation swap quote contains no optionality.0 1. By way of example we consider the market dislocation from September 2008. This is only one reason why inflation swaps may be a better measure for observing changes in inflation expectations than bond breakevens. as measured by forecasts and Figure 194: Relative asset swap valuations may highlight moves in bond liquidity premia. Supply and demand factors can distort inflation swap rates away from expectations just as easily as in bonds. Barclays Capital 15 March 2010 221 . Federal Reserve Bank of Philadelphia. Full-year inflation swap rates do not have to be adjusted to account for seasonality as bond breakevens do and there are no convexity distortions in spot starting zero coupon swaps. with the final value of an inflation swap only determined by the difference between realised inflation and the traded level. Practically separating out these factors is virtually impossible even when in extreme periods when we can be confident of the direction of moves in individual variables. Conceptually. However. Arguably swap rates do not include the same element of liquidity discount evident in most inflation-linked bond markets. UK spiked on nominal QE 160 140 120 100 80 60 40 20 0 -20 Jan-07 OAT€i20 UK IL17 US 10y TIPS Figure 195: US breakevens are poorly correlated with inflation expectation surveys 4.5 4. it is not clear that it is more feasible to disentangle risk premium elements from inflationary expectations in swaps than in bonds.Barclays Capital | Global Inflation-Linked Products – A User’s Guide factoring in liquidity.5 1. particularly at shorter maturities. when liquidity premia in all asset class increased sharply and expected inflation. the inflation rate implied by a zero coupon inflation swap should be the cleanest measure of breakeven inflation available.5 Michigan survey 5-10y median inflation expectations Expected 10y CPI: Livingston survey of professionals !0y TIPS breakeven Jul-07 Jan-08 Jul-08 Jan-09 Jul-09 0.

7.2 1.3% implied by the relative asset swap move of the off the run 5y. with no change in 10y inflation expectations between June and December 2009 and a 1.4% on a median basis. However.6%. off-the-run TIPS breakevens in the 5y sector cheapened 3. compared to a 2. usually using a historical beta relationship.1% fall for the on-the-run issue.2% fall in the 5y CPI swap. and a 2.0 -0. During the period from mid September to the end of November.0 0. if they were then linkers would not be an independent asset class.2 0. where at the end of 2009 the real modified duration of the IL 1.4 0.25% 2055 of 20. whose floor value suddenly became much more valuable.2 May-06 UK IL17 Sweden 2020 Brazil 2017 S Africa 2023 Figure 197: Effective IL55 duration with different beta methods 50 45 40 35 30 25 20 15 10 5 Real duration 3m returns beta 3m yield beta 12m weekly yield beta 12m weekly returns beta Jan-07 Jan-08 Jan-09 May-07 May-08 May-09 0 Jan-06 Source: Barclays Capital Source: Barclays Capital 15 March 2010 222 . compared with the nominal modified duration of the nominal UKT 4.8%. Figure 196: 3 month yield betas in various countries 1. the duration of a linker is likely to be longer than that of a standard nominal coupon bond. The fundamental problem with any approach such as this though is that betas are not stable. For the same maturity. Assigning an effective nominal duration to inflation-linked bond thus requires an arbitrary assumption of the sensitivity of movements in the inflation-linked bond to changes in nominals.25% 2055 was 36. Inflation risk premium in short-dated bonds likely turned negative temporarily due to investors who had nominal benchmarks being forced to cut off benchmark breakeven positions. An extreme example of this is in the UK. as the yield and coupon of the linker are likely to be lower. Calculating the real duration of a Canadian model inflation-linked bond with respect to changes in real yield is done in the same way as calculating the nominal duration of a nominal bond.5% move in 1y forecasts.3%. also fell. the University of Michigan consumer survey of 5-10y inflation expectations fell only 0. to -2. Over the period from September to December. While market expectations of inflation may have adjusted more dramatically than those of consumers. This implies that the inflation expectations component was the least important factor of the 5y breakeven collapse. but that the most important driver was probably the temporary.6 0. position-driven move to a negative inflation risk premium. liquidity could have been worth up to the 1.1. there is no such formulaic method to measure the nominal duration of an inflation-linked bond. with 1y expectations falling 2.Barclays Capital | Global Inflation-Linked Products – A User’s Guide surveys.8 0. the Livingston Survey of professional forecasters suggested not. Effective nominal duration of inflation-linked bonds It is mathematically straightforward to calculate the real duration of an inflation-linked bond using the relevant bond conventions.

as the real duration of the linker is about 1. On the other hand. For longer-term total return investors. Until the final principal value is fixed. The estimate of beta is sensitive to the methodology used. real yield investors are predominant.5 is roughly equivalent to a returns beta of 0. weekly or monthly data. Historically these have been most prominent at longer maturities. as appropriate). as stable yield trend regimes typically last three months to a year. may make sense. For a trader looking for a short-term hedge for linker exposure with nominal futures or bonds. and this is also the time horizon over which yield level betas are most commonly used. This has the benefit of directly including inflation carry data.8x the nominal duration of the nominal 50yr. whereas the returns beta provides a direct estimate of the relative volatility of the two bonds (or indices. for very short-dated bonds. based on daily changes over one to three months. anyway. duration should arguably be the more appropriate measure for assessing risk. where inflationary trends can be an important factor. but the type of beta that is most appropriate depends on its use. inflation uncertainty becomes an increasingly significant factor relative to the decline in nominal price volatility. often leading to lower betas at the long end of the curve. Some participants in the inflation-linked market use the level rather than the change in real versus nominal yields to estimate beta. a two-year or longer monthly total return volatility beta is a logical starting point for asset allocation. where the bias can become extreme. a three. Decomposing volatility and the impact on betas The only mathematically correct way to report duration for a mixed portfolio of nominals and linkers. Beta should never be considered a stable relationship. betas will often be significantly above 1 as a result. it has the advantage of picking up potentially important trend data that are lost in pure volatility analysis. An alternative measure of beta that is arguably preferable to using yields is to calculate based on the volatility of relative returns. Calculating effective nominal duration from a returns beta is different from using a yield beta analysis. is to drop the standard duration figure and instead show two new numbers: duration with respect to real yield and with respect to inflationary expectations. For active money managers. real. At the short end. Beta estimation provides a useful starting point for relative analysis of inflation-linked versus nominal duration. a particularly important factor at shorter maturities but also for longer-term analysis. For those with long-term real return aims or inflation-linked liabilities. While this is a statistically-biased method. The yield beta provides an estimate of the multiple that should be applied to real duration to get an equivalent nominal duration. in the extreme example of the UK IL55. Taking the variances of both sides of the simplified Fisher equation produces the result: Variance (n) = Variance (r) + Variance (bei) + {2 x Covariance (r.or six-month yield daily or weekly change beta may be the most representative. a returns beta may be more advisable. but one of the most significant is the type of investor in the asset class. a short-term yield beta estimation. although in periods of extreme carry. eg. in a way that adds some useful information. but is not appropriate for the long term. indeed.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Most commonly estimating a beta value to determine effective nominal duration of inflation-linked bonds estimates the sensitivity of real yields to changes in nominal yields. the nominal volatility of a sub-1y bond is likely to be significantly higher than 1. the partial derivatives of the Fisher equation. but also to the time period assessed. with a returns beta to compare the two. ie. In addition results can be very different based on whether the analysis uses daily. the time period to assess this sensitivity is far from clear. However. For instance. not nominal. the effective nominal duration of an inflation-linked bond should be lower than its real duration.9. In most circumstances. Betas will vary due to many factors. a yield beta of about 0. using yield beta as a shorthand way to convert real yield duration into nominal space is useful as long as its limitations are remembered. Betas are usually lower when real money. bei)} 15 March 2010 223 . though.

the lower dependence of TIPS yields to nominals has become clearer. However.2 0.14 0. In other words.6 0. at this stress point the covariance moved sharply negative and real yield volatility spiked notably higher than nominal vol. For most of this period.4 0. with real yield volatility similar to that of nominals. The market is not back to the pre-2002 era when real yields were extremely stable and breakevens were almost as volatile as nominals.8 0.0 0.0 10y TIPS 3m yield change beta -0.04 1. of an inflation-linked bond to a change in the equivalent nominal yield will usually be less than one.10 0. We expect this tendency for lower real yield than nominal volatility to continue.06 0.02 0. there was an illiquid period in which breakeven volatility fell sharply. real yields now track nominal movements much less closely than in previous years – a situation that in our opinion is likely to continue. or ‘beta’. but we expect that breakevens are likely to be more volatile than between 2003 and mid-2008. the yield sensitivity.Barclays Capital | Global Inflation-Linked Products – A User’s Guide This formula shows that provided the covariance between the real yield and breakeven inflation is not sharply negative. As liquidity improved. including in some cases the same investors previously forced out of offbenchmark positions that now include explicit real yield allocations. The volatility of linkers spiked above that of nominal bonds in the 2008 turmoil.00 Jan-99 Jan-01 Jan-03 Jan-05 Jan-07 Jan-09 Source: Barclays Capital Source: Barclays Capital 15 March 2010 224 . In US TIPS. betas briefly rose back to previous levels. despite linker markets once again becoming fully functional. real yields will be less volatile than nominal yields. During this time. Figure 198: TIPS yield betas appear to have fallen structurally Figure 199: Realised volatilities have fallen from peak but still high 0. The three months from September 2008 forced out almost all breakeven investors – not just those with leveraged positions but also off-benchmark positions versus nominal indices. albeit with a relatively low statistical significance since the market was still poorly defined. US and euro betas fundamentally altered by late 2008 dislocations Prior to September 2008. but trading on this basis has only recovered to become a relatively small percentage of overall turnover. Hence.16 0.12 10y TIPS 3m real yield realised daily bp vol 10y nominal 3m realised daily bp vol 10y breakeven 3m realised daily bp vol 0. breakeven investors (ie. while the covariance between real yields and breakevens had been small over most of the past decade in the US and Europe. Now the vast majority of the market is held by outright investors. realised breakeven vol rose to similar levels. Breakevens are still widely considered as a valuation metric. this reflects a combination of the new market dynamic and increased inflation uncertainty.08 0. after the liquidation phase in Q1 09. TIPS real yields had a period of over five years in which they were only slightly less volatile than nominal Treasuries. and with correlations disappearing. In our view. those buying TIPS versus nominals or vice versa) were the marginal pricers of the 10y sector and below. with 10y real yield volatility falling much more rapidly than in nominals. indeed close to zero in the five years until mid 2008.2 Jan-99 R-squared of beta relationship Jan-01 Jan-03 Jan-05 Jan-07 Jan-09 0.

In the UK. bond and swap breakevens started to trade together again.8-1.08 0. but the difference became more extreme in 2009 than in previous years.8 0. Betas at longer maturities have tended to be lower than at 10y. The major market dynamic behind this is likely to be that among the largest buyers of euro linkers in 2009 were asset swap investors. The IL17 3 month yield change beta stayed at 0. akin to that in TIPS in the middle part of the decade. with yield betas averaging close to 0. volatility trends during this time were very similar. however. Breakevens were in a tight range for most of this time.10 0.0 0. which caused far less breakeven position cutting than elsewhere as less of the market was held by breakeven investors to start with. As the euro linker market once again became fully functional. 10y real yield vol has settled back at a similar level to nominals. and the statistical significance of this beta returned above 80% from mid-2009. This is despite euro bond breakeven trading and investment remaining relatively limited as in the US. having a central bank with an explicit inflation target. The dislocations in the three months from September 2008 were less extreme than in TIPS.9.02 0.00 Jan-02 10y OAT€i 3m real yield realised daily bp vol 10y OAT 3m nominal realised daily bp vol 10y OAT€i 3m breakeven realised daily bp vol 1. This is addition to the structural factors that make euro breakevens more stable – ie.6 0. it settled into a beta regime that while notably less stable and well defined than previously.4 0. with breakeven vol almost equal to that of nominals and real yield vol higher still. there does not appear to have been a fundamental change after the late 2008 dislocations. meaning pure inflation activity once again often being the marginal driver of €i prices. Figure 200: €i yield betas have been less volatile than TIPS Figure 201: Euro real yield vol remains relatively close to nominals 0.2 1. was still relatively high compared with the US – indeed.Barclays Capital | Global Inflation-Linked Products – A User’s Guide The beta and volatility trends for linkers in the euro area are in many ways similar to those in US TIPS. 3mth yield betas briefly spiked higher during this period.12 0. real yields displayed a similar volatility to nominals for the five-year period until mid-2008.2 0.04 0. As relative asset swap levels returned to normal ranges. with breakeven vol an order of magnitude lower. and with euro governments notably more sensitive than the US Treasury to breakeven levels with regards to how heavily they issue euro linkers. largely due to the lower share of the euro linkers held by breakeven investors. but this corresponded with the least statistical significance of the relationship.0 Apr-04 OAT€i20 3m real yield beta R-squared of beta relationship Apr-05 Apr-06 Apr-07 Apr-08 Apr-09 Jan-04 Jan-06 Jan-08 Source: Barclays Capital Source: Barclays Capital 15 March 2010 225 .06 0.0 even through 2008. Even more so than in the US.

Here. we define the real interest rate on an ex post basis – for instance in the case of the US. Figure 202: Annual real return and risk of real return assets since 1987 30% 25% Real Return Real Risk 20% 15% 10% 5% 0% Global Real estate Equities Treasuries linkers* Note: *Hedged to USD. History has shown that during periods of high inflation it is essential to focus allocations in the narrow band of assets which offer effective inflation protection. including commodities. as many investors seek to match future real liabilities that are based on real interest rates rather than just inflation. which in this chapter we consider as consumer inflation. even fewer to real rates. global linkers and real estate provide the most attractive returns relative to risk. However. which are likely to bias risk lower. dividend swaps. and only inflation-linked bonds and inflation swaps have the ability to offer long. Indeed. despite also being affected by the deleveraging. Source: GSCI. we also seek to determine whether these assets hold a strong correlation to the level of real interest rates. the real estate and timber real total return data used here are based on appraisal. Meanwhile. UK linkers. NCREIF. Global linkers from 1997. rather than transaction data. as simply relying on portfolio diversification is insufficient to ensure a positive real return. UK linkers shown due to long sample. Barclays Capital Commodities Timberland UK linkers* The extent to which the assets in this study are capable of hedging real rates or inflation depends crucially on the holding period as many of these assets must be held for much longer than five years to generate consistently positive inflation-adjusted returns.and short-term protection from both. commercial property and volatility to determine their inflation hedging properties. timberland. We find that very few assets classes offer appropriate protection from inflation.com We examine a range of assets typically thought to provide a return linked to inflation and compare their merits as a hedge against inflation and real rates.bettiss@barcap. Each real return asset offers a different real risk/return profile across the business cycle. although resource-based equity sectors may offer protection from commodity-related inflation spikes. Commodities and equities still offer the least attractive risk/reward profile. equities. the volatility of inflation-linked bonds increased the least of all the asset classes surveyed in this chapter since the previous Global InflationLinked Products: A User’s Guide was published in February 2008. the 1-year realised Fed Funds rate minus the annualised rate of US headline CPI. We examine a range of assets. Equities failed to provide sustained short or long-term inflation protection.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Real return assets Chris Bettiss +44 (0) 20 7773 0836 chris. Real return assets are those whose real returns are not typically eroded by inflation. with their volatility increasing significantly as a result of the deleveraging moves that occurred in H2 08. However. Dividends offer an attractive alternative 15 March 2010 226 .

69 0. Barclays Capital In some cases. the returns on each asset class can also be split into a price return and an income return. In this study we examine each asset class with reference to the extent to which each element is exposed to inflation. For example. Source: NCREIF. there are significant divergences in the payout ratios across equity sectors and dividends prove an imperfect tool for accessing earnings streams and are less efficient as an inflation hedge than they first appear.32 -0. Overall.60 0.29 0.20 -0.15 0. but the traded price of the linker fluctuates based on expectations of real interest rate levels. commercial property provides little exposure to inflation over shorter periods. although this is over a relatively short data sample. providing the strongest performances when growth is strongest. Figure 204: Fraction income return and source of income return Type of income return Equities (S&P 500) Property* Timberland Commodities Nominal Bonds UK Inflation-Linked Bonds Dividends Rent Timber harvesting Fixed coupons Inflation-linked coupons Fraction of total return from income 39% 84% 24% 0% 5% 41% Note: *NCREIF US National Property Index. which contains the NCREIF Residential.11 0.64 -0.84 Note: *NCREIF US National Property Index. Figure 203: Correlations between asset total returns and headline US CPI (UK RPI for UK bonds) since 1987. Industrial.11 -0.05 0. timberland and shortdated inflation-linked bonds hold the most positive correlations with inflation over periods of three years and less.30 Yearly 0. but not necessarily both. with the rolling 20y correlation between the total 15 March 2010 227 . However.Barclays Capital | Global Inflation-Linked Products – A User’s Guide to pure equity exposure. Retail and Hotel Property Indices. which contains the NCREIF Residential. despite relatively low volatility. Ecowin. 1-5y -0. Ecowin. However. Commodities. under different holding periods Monthly Equities (S&P 500) Dividends (S&P 500) Commodities (GSCI) US Property* US Timberland UK nominal gilts UK Inflation-Linked Bonds UK Inflation-Linked Bonds. as profit growth tracks inflation.45 -0. Barclays Capital Equities Equities have been considered a real asset because the investor receives dividend payments plus any capital appreciation of the stock. the coupon and principal of inflation-linked bonds are directly exposed to inflation.12 0. Apartment. Source: NCREIF.15 0.51 0.04 0.08 -0.07 Quarterly 0. However.20 0.01 0.15 0. Timberland offers broadly robust inflation and real rate protection. Industrial. Retail and Hotel Property Indices. inflation-linked bonds remain the most attractive because they can provide a complete inflation hedge if held to maturity or exposure to expectations of future real rates if held over a shorter period. Commercial property and timberland display more traditional cyclical movements across the business cycle. Office. Depending on the asset.50 0.41 3 Year 0.12 0. Office. and show a solid short-term positive correlation with inflation.80 -0.05 0.12 0. Apartment. However.28 0. the volatility of commodity returns means they do not provide a reliable inflation hedge for investors with either short or long real liabilities. either component may be exposed to inflation.14 0. equities themselves do not provide a good long-run hedge against inflation.

Figure 205: Long-term real total returns of US and UK equities 45% 35% 25% 15% 5% -5% -15% 1909 1919 1929 1939 1949 1959 1969 1979 1989 1999 200 Source: Barclays Capital Figure 206: Equities are de-rated by inflation in the short term 16% 14% 12% 10% 8% 6% 4% 2% 0% -2% -4% 1955 1961 1967 1973 1979 1985 1991 1997 2003 2009 Source: Barclays Capital US equity 10y annualised real UK equity 10y annualised real US equity 20y annualised real UK equity 20y annualised real total total total total returns returns returns returns S&P 500 Earnings yield US CPI y/y (RHS) Although equities underperform gilts substantially in periods of high inflation. Bonds underperform equities during broader inflationary periods because a rise in profits can offset some downside to equities. particularly during periods of high inflation. the strongest real returns on US equities and nominal bonds are during periods that begin with very high inflation. Not all equity sectors underperform in periods of high inflation. while falls in bond prices persisted until inflation subsided. they also tend to be correlated with bond yields. Equity returns in the US and UK were narrowly distributed in the 1970s and real returns were negative across most sectors and positive equity total returns were concentrated only in sectors which directly benefitted from inflation.Barclays Capital | Global Inflation-Linked Products – A User’s Guide return on US or UK equities and inflation negative for most of the past century. leading to notable underperformance. the falls in equity values during these periods tend to be temporary. utilities and technology sectors. Short-term equity total returns are negatively correlated with inflation. Therefore. as oil prices outperformed industrial metals and other commodity prices. While utilities provide the highest fraction of return from dividends among all equity sectors. However. Hence. investors can protect themselves in periods of sustained inflation by selling nominal bonds versus equities. and the volatility of equity capital returns makes equities highly risky for investors with short duration real liabilities. such as the oil and gas and industrials sectors. as the rapid decline in inflation following the initial spike leads to a sharp pick-up in the value of these assets. healthcare. historically. 15 March 2010 228 . a means to protect against inflation appears to be to buy equities in the oil and gas and industrial goods sectors versus equities in the consumer goods services. The basic materials sector performed poorly.

while equity returns do not typically offer protection in periods of high inflation. They enable investors to receive exposure purely to dividends. it is logical that profit growth should track inflation. Dividend swaps are over-the-counter derivatives used to trade dividends on a forward basis. Barclays Capital However. While equities failed to provide positive real returns in the stagflationary 1970s. Barclays Capital Source: Datastream.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 207: Equities do not provide an effective real rate hedge 100% 80% 60% 40% 20% 0% -20% -40% S&P 500 annualised total returns US real interest rate (1y realised US Fed Funds rate minus y/y US CPI) 14% 12% 10% 8% 6% 4% 2% 0% -2% -4% Figure 208: Total return for US equity sectors (1969-79) Oil & Gas Industrials Market Utilities Telecoms Consumer goods Basic materials Financials Technology Healthcare Consumer services -10% -5% 0% 5% 10% -6% -60% 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 Source: Bloomberg. rising as the holding period is extended to 5-10y and dividend growth generally became elevated when inflation spiked above trend levels. 15 March 2010 229 . Over the past century. Dividend swaps Until recent years it was only possible to gain exposure to corporate profits via a traditional equity investment. Dividend distributions have mostly proven resilient to modest market downturns as corporates have generally been reluctant to cut dividends except as a last resort. dividends have a modest positive correlation with inflation. it is now possible to access corporate profits almost directly by investing in dividend swaps. with the correlation between S&P 500 and US inflation dividend growth almost 60% over the decade. rather than dividend plus book value plus market capitalisation of the long-run flow of dividends. the dividend income component may be used as an alternative investment. However. this is usually caused by de-rating rather than weak earnings. while equity capital returns are volatile. Since a key component of corporate profits is output pricing. The stable long-run performance of dividends also offered an effective real rate hedge. Therefore. leading us to examine dividend swaps as an inflation hedge. dividends provided strong inflation protection.

Figure 211: Change in dividend payout ratios. Thus. the price of energy.7% Health care 7.8% Industrials -24. the historical performance of dividends is deceptively mixed. 70% of which is due to energy. Meanwhile. they may provide imperfect access to earnings streams. sectors whose earnings would have provided the most effective inflation hedge during the 1970s reduced their payout ratios the most. end 1969-end 1980 Market -19.5% Consumer goods -7.9% Oil & Gas -22. While 16% of the US headline CPI basket is linked to energy. while dividend exposure offers an attractive alternative to pure equity exposure from the point of view of hedging inflation and real rates in a general sense. Therefore. food and energy are two important components in most headline CPI baskets. 26 June 2008.3% Financials -40. energy has accounted for the majaority of total inflation volatility over the past ten years.1% Basic materials 5. hence. namely liquidity and the credit risk of the counterparty. Commodities During inflationary periods it is useful to focus on the resources that are causing inflation. food and energy collectively represent 23% of the CPI basket. Dividend payout ratios in the US and UK fell over the 1970s.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 209: Dividends modestly correlated to inflation 20% 15% 10% 5% 0% -5% -10% -15% S&P 500 Dividend per share growth US CPI y/y (RHS) 8% 6% 4% 2% 0% -2% -4% -6% Figure 210: Dividends mostly offer protection against real rates 20% 15% 10% 5% 0% -5% -10% -15% -20% 1977 1981 1985 1989 1993 1997 2001 2005 2009 Source: Robert Shiller. have provided the best overall performance and remained mostly stable. However. In other words. 15 March 2010 230 . Barclays Capital S&P 500 Dividend per share growth US real rate (Fed Funds minus y/y US CPI. dividends were an imperfect tool for accessing earnings streams. While commodities do not provide direct income return. implying that their specific inclusion in a portfolio as a hedge against inflation or real rates should remain modest. particularly in the oil & gas and industrials sectors.3% Source: Thomson Datastream There are also risks to dividend swaps that are not reflected in earnings streams. In US headline CPI. RHS) -8% -20% 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008 Source: Robert Shiller. for instance. Barclays Capital Dividends from commodity-related sectors. with significant divergences in the payout ratios across equity sectors.4% Consumer services 6. particularly energy and materials.and energybased commodities make up 82% of the S&P GSCI Index. food.and food-based commodities logically hold some correlation with inflation. UK market. as Tim Bond highlights in Nine steps to profit from inflation.

8% 2. Barclays Capital Examining optimal holding periods reveals that investors have needed to hold commodities for at least 15 years to ensure they achieve positive real returns.7% -3. However. but in the ability to protect investors from unexpected spikes in inflation.0 1976 1980 1984 1988 1992 1996 2000 2004 2008 Source: GSCI.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 212: Long-term commodities real total returns 30% 25% 20% GSCI 5y annualised real total returns GSCI 10y annualised real total returns GSCI 15y annualised real total returns Figure 213: Real geometric annual returns by asset class.2 0.7% -0.6% 0. Barclays Capital Foodstuffs (CRB) Property Equities Government bonds Cash Source: Haver Analytics.1970s Real returns. The main disadvantage of commodities is their volatility.9% UK 16.0 -0.0% -4.4 0.8 0. but not industrial metals or foodstuffs. Figure 214: Correlation of commodity real total returns and RPI 1.7% -0. commodities may have a role in inflation hedging portfolios over a long-term horizon. geometric annual. Barclays Capital Figure 215: Commodities are not an effective real rate hedge 140% 120% 100% 80% 60% 40% 20% 0% -20% -40% GSCI total return (y/y) US real interest rate (1y realised US Fed Funds rate minus y/y US CPI. unless held on a purely tactical basis.6 -0. their extreme volatility suggests that even this should remain modest.4 -0.2 -0. Datastream. There is virtually no correlation between the level of commodity price and real interest rates. namely the energy sector and specifically oil. Ecowin. as witnessed by the 15-year holding period that has historically ensured a positive real return. they do not offer protection from a real rate hedging perspective. 1970-80 Oil Industrial metals (fibre INDEX) US 11.9% 15% 10% 5% 0% -5% -10% 1975 1979 1983 1987 1991 1995 1999 2003 2007 Source: GSCI.5% -4.8% -2.4% 0. However.9% 1.8 -1.3% -2. while commodities benefit from unexpected spikes in inflation. While.7% 2. As highlighted by their performance in the 1970s. positive real returns were focussed on a narrow band of commodities.6 0. their high shortterm volatility makes them an impractical inflation hedge for investors with short duration real liabilities. Commodities also have the advantage of being negatively correlated with most other assets in this study and thus act a strong diversifier within a real asset portfolio. RHS) 14% 12% 10% 8% 6% 4% 2% 0% 5y correlation of GSCI annual returns and y/y UK RPI -2% -60% -80% -4% 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 Source: GSCI. commodities provided significantly positive real returns on the decade and dramatically outperformed other asset classes in this study. from a diversification perspective. hence. the main benefit of investing in commodities lies not in the long-run returns. Datastream. Barclays Capital Therefore. 15 March 2010 231 .0 0.

In addition. the price of timber is the single most important return driver and we examine the inflation and real rate hedging characteristics of the asset class on this basis. US timberland has provided sustained positive real returns over a relatively short holding period of five years and broadly positive returns over the past twenty years. as is the case for UK RPI. Commercial property Annual returns for real estate. particularly commercial property. thus. house prices tend to lead consumer spending. such as the S&P Global Timber & Forestry Index and its constituents. RHS) US Timberland Index annualised returns 1990 1993 1996 1999 2002 2005 2008 -10% 1987 1990 1993 1996 1999 2002 2005 2008 Source: NCREIF. Barclays Capital Source: NCREIF. with such a short sample – from 1987 in the case of the NCREIF US Timberland Index – we do not have sufficient data to conclusively determine the long-term inflation hedging capabilities of the asset class. Rising commercial property prices tend to benefit the corporate sector and affect economic activity through business investment. residential housing is usually the largest component of household wealth. this is especially so if there is a direct house price component in the inflation basket. There are also notable divergences in timber prices between regions. bank’s balance sheets are exposed to housing through loans and mortgages. real estate price movements have a major effect on economic conditions and hence inflation. Poor liquidity and maintenance costs are not captured in these data. For inflation. However. which are based on appraisal rather than transaction values and hence may understate the risks of investment.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Timberland Another natural resource that has been considered correlated with inflation is timberland. illustrate how closely its capital growth and therefore total returns are related to the general business cycle. The advantage of timberland is that it shows a relatively low dispersion across the business cycle. As has been particularly notable since the start of the credit crisis. Barclays Capital An alternative source of timberland returns is timber equities. Although timber can be put to many uses. Meanwhile. this analysis uses the NCREIF US timberland data. 15 March 2010 232 . The real risk/return on timberland appears far more attractive than on other commodities. Furthermore. Figure 216: US timberland hedges inflation over 5y period 50% 40% 30% 20% 10% 0% -10% 1987 US Timberland Index annualised real returns US Timberland Index 5y annualised real returns Figure 217: US timberland provides robust real rate hedge 80% 70% 60% 50% 40% 30% 20% 10% 0% US real interest rate (1y realised US Fed Funds rate minus y/y US CPI. but these are subject to the same problems as a typical equity investment and hence do not offer a significant advantage over a timberland investment.

Barclays Capital US Commercial Property 5y annualised real total return US Commercial Property 10y annualised real total return Total return Income return Capital growth A large proportion of the return on commercial property comes from rent. Indeed. Investment in commercial property is also subject to other constraints that are not captured in these data. US commercial property appears to offer a more attractive inflation-adjusted risk level than other real asset classes. they are likely to bring the Sharpe ratio on commercial property closer in line with riskier assets in this study. Barclays Capital Figure 219: UK real property returns annual % change 30% 20% 10% 0% -10% -20% -30% -40% 1971 1975 1979 1983 1987 1991 1995 1999 2003 2007 Source: IPD. Retail and Hotel Property Indices. which is mostly very stable over time. Office. Income returns on commercial property have generally been relatively stable and far less volatile than equity income returns. Investors could have taken exposure to US commercial property for the past thirty years and gained positive real returns over any 9-year period. Source: NCREIF. However. in reality. US commercial property provide little exposure to inflation over shorter periods and returns are only capable of occasionally outpacing US real interest rates despite a modest long-run correlation. such as very poor liquidity. Indeed. Office. RHS) Note: US Commercial Property is an aggregate of the NCREIF Industrial.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 218: Real total returns on US commercial property 16% 14% 12% 10% 8% 6% 4% 2% 0% -2% -4% 1983 1986 1989 1992 1995 1998 2001 2004 2007 Note: US Commercial Property is an aggregate of the NCREIF Industrial. Retail and Hotel Property Indices. Source: NCREIF. with only a few periods of outright negative real returns. the appeal of commercial property as a real return asset is in its stable real performance. as the NCREIF US commercial property returns data are also based on appraisal rather than transaction values. tenant default risk and maintenance costs. Figure 220: US commercial property is generally insufficient to hedge US real rates 10% 8% 6% 4% 2% 0% -2% -4% -6% -8% -10% 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008 US Commercial Property annualised real total return US real interest rate (1y realised US Fed Funds rate minus y/y US CPI. Barclays Capital 15 March 2010 233 . commercial property returns are likely to be more volatile than these data would suggest. When these effects are adjusted for. Despite relatively low volatility.

economic and demographic growth. Therefore. Equity earnings yields and the volatility of profits are very well correlated and. Thus. by their very nature. water management and power networks have also been noted as real return assets. Indeed. sharp rises or falls in the rate of inflation tend. a means to hedge against inflation is to be long nominal bond and equity volatility.Barclays Capital | Global Inflation-Linked Products – A User’s Guide An alternative way to source property market returns is via property derivatives. Infrastructure assets are excluded from our analysis given their short data sample. it should be noted that utilities equities provided the highest fraction of return from dividends among all US equity sectors since 1989. for instance. volumes and liquidity in property derivatives are low compared with inflation derivatives. There are also considerable lags between the valuation and publication dates for the property indices. it is often a principle contributing factor. the economic and profit climate becomes far more volatile and uncertain during inflationary periods. Figure 221: US 20y equity volatility versus inflation. 1y standard deviations of quarterly return US CPI. while the recent dramatic falls in residential and commercial property prices may have dampened investor demand for the asset class despite its scope to develop. However. Investors can now take positions in actual and implied asset volatility as a separate asset class. given that their values exhibit positive correlations with inflation. % y/y (RHS) 14% 12% 10% 8% 0% 0% 19611963 19651967 1969 19711973 1975 19771979 1981 Source: Haver Analytics 1965 1968 1971 1974 1977 Source: Haver Analytics 15 March 2010 234 . the volatility of profits and economic growth are also tightly correlated with inflation. such as the VIX and MOVE indices. and while inflation is certainly not the sole cause of volatility. airports. % y/y (RHS) 14% 12% 10% 8% 6% 6% 4% 2% 0% 1980 Figure 222: US 20y rate volatility versus inflation. Infrastructure assets such as roads. in turn. Meanwhile. to demand for higher risk premia to compensate for the increased volatility of profits. with data on the UBS US Infrastructure index. Equity and interest rate volatility typically rise with inflation. to be unexpected and so also produce unexpected changes in interest rates and as a result interest rate volatility also tends to be correlated with inflation. only available since 1995. whose returns are linked to the performance of an index of property prices such as the UK IPD AllProperty Total Return Index or the US NCREIF index. Volatility An asset class that did not exist during the inflationary 1970s is volatility. the broad rise in earnings yields that accompanies periods of high inflation is due. 1962-80 20% 18% 16% 14% 12% 10% 8% 6% 4% 2% 6% 4% 2% 4% 2% 0% 1962 S&P 500 1y standard deviation of quarterly returns US CPI. However. 1962-80 12% 10% 8% US 20y nominal swap. in part.

an outright position in TIPS still offers a good inflation hedge relative to other fixed income products. Therefore.bettiss@barcap.com Chirag Mirani +1 212 412 6819 chirag. Essentially. Figure 223: Y/y return correlations show TIPS have a low correlation with other assets. Using Barclays TIPS Index 15 March 2010 S&P Oil . with US NSA CPI.pond@barcap. investors must look at breakeven inflation positions or TIPS returns relative to nominals. before getting to an asset allocation decision. rather than purely to TIPS. The low correlation with CPI makes sense because TIPS provide real rate exposure.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Linkers in a portfolio context Michael Pond +1 212 412 5051 michael. We note that even if investors buy TIPS on an outright basis.com Chris Bettiss +44 (0)207 7730836 chris. The diversification hypotheses outlined above and potential inflation hedge are good reasons to add TIPS to a diversified portfolio. we first show how TIPS total returns have held up compared with other assets. not inflation exposure.mirani@barcap. a diverse portfolio containing TIPS would be expected to have lower risk versus a comparable portfolio which does not contain TIPS. offers even more attractive risk/reward. Our analysis confirms that TIPS should indeed be included in a portfolio. if the decision is whether to purchase TIPS or another fixed income product. In order to strip out real rate exposure. perhaps most surprisingly. but that including exposure to global inflation-linked bonds. However. A correlation matrix of TIPS returns versus other assets reveals that TIPS have a low correlation with these assets and. then this is essentially a breakeven position. We also find that inflation-linked bonds do not offer effective currency protection. March 1997 to December 2009. suggesting diversification benefits Short Tsy BEI Index AUDCAD NZDNOK Global Inflation 1-3y BEI CPINSA 100% 235 GSCI Gold TIPS CRB Asset TIPS Global Inflation 1-3yBEI BEI Index Gold Oil CRB GSCI S&P Short Tsy AUDCADNZDNOK CPINSA 100% 74% 72% 31% 13% 28% 57% 50% -9% -54% 53% 31% 100% 83% 46% 36% 38% 64% 43% 34% -14% 86% 26% 100% 96% 56% 81% 83% 72% 73% 43% 81% 60% 100% 27% 68% 64% 59% 58% 61% 57% 39% 100% 25% 47% 20% 6% 16% 52% 25% 100% 80% 86% 37% 36% 41% 62% 100% 90% 32% 13% 62% 74% 100% 27% 11% 38% 75% 100% 60% 39% 25% 100% 13% 8% 100% 20% Source: Barclays Capital.com We examine whether inflation-linked assets such as US TIPS belong in a diversified portfolio. over time. The relatively low correlation TIPS hold with most other assets is beneficial to TIPS owners from a diversification perspective.

3% 4.06 0.2% 6.2% 3.5% 6.5% 22.5% 5.6% 0.0% 5.5% -5.8% 3. then benefited as energy prices rallied significantly.5% 31.0% 5.04 1.5% 6.2% 6.9% 16.8% 9.7% 4.6% -0.2% 25.7% 6.29 GSCI 13.65 1.6% 8. TIPS initially had relatively low volatility.4% 0.26 Corp 18. the S&P 500 and the Goldman Sachs commodities indices on a total return basis over this period. Only the unhedged Global Inflation-linked Bond Index had a higher return among the assets shown in Figure 225.9% 2.8% 6. Source: Bloomberg.59 0.5% 6.91 0. Barclays Capital We expect TIPS to continue to become more liquid as the Treasury is taking steps to improve the programme.5% 3.5% 10.9% 6.7% 4.4% 6. MBS.7% 7.20 0.6% 6.2% 3.11 Note: March 1997 to December 2009. We note that the volatility of TIPS returns during this time was lower than other assets which are seen as an inflation hedge. such as commodities.0% 6. 13.8% 6.7% 7.9% 7.15 1.0% 2.7% 1.6% 4.27 US Agg 5.8% 1.22 1.4% 6.3% 6.7% 5.5% 11. In fact they have returned the highest among a portfolio of Treasuries.0% 7.8% 9.1% 16.1% 4. Figure 225: TIPS have delivered high returns with less volatility High Yield 58.3% 5.9% 1.8% 7.9% 8.6% 6.8% 5. Barclays Capital Figure 224 shows that since their inception (March 1997).3% 3.7% 6. TIPS underperformed nominals along with other riskier assets during the financial crisis as there was a significant flight to liquidity to nominal Treasuries.0% 1.10 Tsy -3.67 Global Inf.8% 2.5% 11. but 15 March 2010 236 .Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 224: Total returns of all assets in a diversified portfolio since inception of TIPS 240% 210% 180% 150% 120% 90% 60% 30% 0% -30% -60% Mar 97 Sep 98 TIPS Mar 00 Tsy Sep 01 HY Mar 03 Sep 04 Mar 06 MBS Sep 07 SP Mar 09 GSCI Corp Note: March 1997 to December 31.60 S&P 500 TR 26. Source: Bloomberg.5% -6.0% 16.0% 17.1% 4.7% 6.2% 13.4% -0. High Yield and Investment Grade corporate debt.0% 0.1% 28.9% 7.1% 2.9% -3. Most recently the Treasury solicited market suggestions and has not only acted to increase the size of TIPS auctions and reintroduce 30y TIPS issuance.3% 19. 2009.6% 24.6% 7.3% 13.7% 3.6% 4.23 2.7% 10.96 TIPS 1y 3y Annualized Returns 5y 10y Since inception 1y 3y Annualized Vol of monthly returns 5y 10y Since inception Return/Risk 10y Since inception 10.6% 24.9% 4.0% 5.6% 6.0% 6.2% 6.7% 4.8% 8.6% 0.4% 2. TIPS have held up fairly well in the volatile trading environment. Using Barclays TIPS Index.07 MBS 5.1% 5.

8% 20% 2.1% Figure 227: Optimal portfolio includes TIPS across risk appetites. we know that historical returns are no indication of future returns. 1974). supply equals demand.5% 3. Lastly. Alternatively. In our view. Bills < 0. Constraint: Weights < 0. then we can back-out maximum returns (and minimum risk) that would justify the market portfolio.05% 3. then we should hold the market portfolio as this is the most efficient portfolio (meaning. According to this analysis. If we use historical covariance as an assumption and assume market capitalization portfolio weights to be optimum. a fundamental feature of TIPS is that they provide increased certainty of real returns to maturity. market capitalization-based weights are the optimal portfolio weights as this is where supply equals demand). However.1 Source: Barclays Capital. We use the Black-Litterman model to determine. optimal weight allocation to TIPS for a diverse portfolio consisting of assets shown in Figure 2 and 3m T-bills.25. starting with Market Equilibrium Portfolio Historical excess returns based mean-variance portfolios are notorious for being unstable in that a slight variation in expected returns can skew weights assigned to each asset.8% Corp 3m T-Bill Source: Barclays Capital. Bloomberg 15 March 2010 237 . ie. using historical returns Optimal Weights 100% 80% 3. protecting purchasing power of investors and are therefore unlike any other securities.65% 3. TIPS continue to improve the efficient frontier. Figure 226and Figure 227 show that using historical data.25.0% 60% 2.8% TIPS SP Diversified Portfolio Diversified Portfolio + TIPS 4.3% Risk Treasury HY GSCI MBS 4.6% 3. and to some degree confirm. Figure 226: TIPS push out the efficient frontier when using historical returns as expected returns Excess return 3. We see these actions as a positive for TIPS in a market with improving economic fundamentals. These returns are known as excess market equilibrium returns and this extraction process is referred to as reverse optimization (Sharpe. most investors agree with this assessment but wonder how much they should allocate to TIPS. Bloomberg. these same market equilibrium returns can be retrieved using the following method (this is the solution to above optimization problem). a diverse portfolio should include more than a 10% weight in TIPS.9% 40% 2. Bills<0.7% 2. We incorporate our view of 1y ahead total returns in this model to arrive at optimal allocation weights. Thus. we view them as a legitimate asset class that should be included in a diversified portfolio. Black-Litterman Model. although historical covariance can be a guide to future covariance.85% Risk 4.45% 0% 3. If we assume that the market is efficient and the market portfolio is at equilibrium. The Black-Litterman Model starts out with a market equilibrium portfolio before incorporating any views.Barclays Capital | Global Inflation-Linked Products – A User’s Guide has also indicated consideration for increasing the frequency of TIPS auctions.1 Note Analysis done in March 1997 to December 2009. Constraint: Weights < 0.

it will tune the magnitude of returns.40% 2. although mathematically sound. Black-Litterman Model.4% 3. Figure 228summarizes our market portfolio. derived using historical covariance matrix and noted excess equilibrium returns below.3% 10.8% 24.9% 1.5% 3. TIPS should be allocated 3% (market cap weight) if your asset allocation choices are the same as in this table. Hence under market equilibrium assumptions. In line with this. Bills < 0. Letting weights vary freely. the optimal weights (maximum Sharpe ratio). incorporating a view The useful element of the Black-Litterman model is that it allows users to incorporate views along with respective confidence levels to build their own expected excess returns vector for all assets under consideration. We have forced this condition by defining the relationship in the equation above. TIPS are allocated as much as 12% at the least risk optimal portfolio.699 $350 $5.05% 4.60% 2.20% 2.4% 1. investors should allocate TIPS according to the relative market caps of their investable asset universe.40% 3.60% 1.8% 41.5% 0. The risk aversion coefficient we arrived at is roughly 9.500 $9.00% 2. Excess equilibrium returns for equities are quite high relative to others as this return is needed to justify higher weight held for this asset in an optimum market portfolio.80% 2.80% 1.45% Asset TIPS Treasury HY Corp SP Comdtys MBS 3m Bill Investable market size (bn) $600 $3. Barclays Capital 15 March 2010 238 . we used historical returns and expected 1y future risk-free rate (50bp). This rather high coefficient will not determine relative weights of each asset in an optimal portfolio.25. are the same as relative market weights of all assets.Barclays Capital | Global Inflation-Linked Products – A User’s Guide ∏ = λ ∑ω ∏ = Excess Market Equilibrium Returns. Under this constraint.4% 1. Source: Bloomberg. λ = Risk Aversion Coefficient (Expected risk-return trade off).45% 3.1% -0.1% 0.10. is not very practical. ω = Proportional market weight vector for various assets To estimate the risk aversion coefficient.6% 16.022 $400 Excess equilibriu m returns 1.5% 21.0% 3.800 $762 $2.65% 3. TIPS push the efficient frontier of the constrained market portfolio to the left as they offer higher returns compared with nominals and bills while maintaining only slightly higher volatility.9% 2.1% Source: Bloomberg. An efficient frontier of these equilibrium returns and covariance matrix has one optimal portfolio which assigns as much as 7% weight to TIPS.0% 9.25% 4. These views are built on top of relative views implied by Figure 228: Market equilibrium portfolio weights Market or equilibriu m portfolio weights 2.85% Risk Diversified Portfolio Diversified Portfolio + TIPS 4.00% 1. ∑ = Variance/Covariance Matrix (we used historical var/covar).1% 0. Barclays Capital Note: Constraint: Weights < 0. Below we have shown the equilibrium market return frontier with a constraint of each weight being below 25% of total allocation (Bills are limited to a 10% allocation).7% Figure 229: Equilibrium Market Portfolio Efficient Frontier Excess returns 3.1.7% 1.0% View updated excess returns 2.4% 0.9% 10.2% 4. As a side note.

Again. The most significant change comes from view 3. Figure 230: TIPS push out the efficient frontier left. View 2: Treasury should outperform MBS by 1%. We notice that our first view regarding TIPS is not far away from what is implied by relative equilibrium returns.0% 2.65% 3.5% 3. Barclays Capital holds the following six views about one year ahead excess returns of the assets in consideration (non-rates views are taken from the respective Barclays Capital strategy group in each asset class).45% 40% 20% 0% 3. using our 2010 asset views (Black-Litterman Model) Optimal Weights 100% 80% 4.5%.0% 3. Fischer. Bills < 0.48% 3. for a relative view (such as view 1.10. The commodity view return expectations are well above what is priced in by equilibrium returns.5% Excess Returns Figure 231: Optimal Portfolio Includes TIPS at lower risks. we do not expect significant change in our efficient frontier owing to view 1.70% Corp 3m T-Bill Diversified Portfolio RiskDiversified Portfolio + TIPS Risk Treasury HY GSCI MBS Note: Constraint: Weights < 0.0% 4.Source: Bloomberg. and Robert Litterman.05% 4. View 6: S&P should deliver a total excess return of 2. Barclays Capital Note Analysis done in March 1997 to December 2009.” Financial Analysts Journal.40% of return change to our equilibrium view.25. Black.5%.85% 4.). using our 2010 asset views (Black-Litterman Model) 5. Constraint: Weights < 0.5% 2.80% 4. View 1: TIPS should outperform nominals by 1.25% 3. Using the Black-Litterman model (See paper for implementation details. This exercise corroborates the view that TIPS should be included in a diversified portfolio.9% in a market equilibrium portfolio. View 5: Corporates should outperform Treasuries by 2. “Global Portfolio Optimization.10. Figure 6 shows that TIPS outperform nominals by 0. we included these views to form our expected excess returns vector (Figure 228. In other words.Source: Bloomberg. derived absolute return levels are not a concern because we have only identified views of TIPS versus Treasuries and Treasuries versus MBS. We recommend that investors start with the market portfolio (relative market caps) to determine specific weight allocated to TIPS and adjust allocation based on their views.25. thus view 1 only adds 0. View 4: High Yield should outperform Treasuries by 9.Barclays Capital | Global Inflation-Linked Products – A User’s Guide the market portfolio. September/October 1992: 28–43. Bills < 0.40% 4. 2).20%.50% TIPS SP 3. Barclays Capital 15 March 2010 239 . Figures 8 and 9 confirm that the allocation of TIPS at more than 10% lowers risks. for commodities. View 3: Commodity indices should have an excess return of 24.0% 60% 3. reducing risk.10% 4. last column) and derive view based efficient portfolios (Figure 230).10%.5%.

In isolation.e. 15 March 2010 240 . from deleveraging and the economic slowdown. and their subsequent re-widening were much greater in magnitude than the movement in USD. Datastream Source: Bloomberg. enhanced returns and low risk. Barclays Capital Global inflation-linked bonds The benefits of including global inflation-linked bonds within a portfolio are evidenced by the large number of institutional and retail investors that have made explicit allocations to the asset class within their portfolios. a fall in the currency should precipitate a rise in inflation. Using the performance of TIPS breakevens as a guide. In practice. the best way to diversify a portfolio by including linkers is to invest in global inflation-linked bonds hedged to the local currency.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation-linked bonds in a non-currency hedged portfolio In theory. In our view. especially in periods of economic volatility. however. to move more in response to currency movements than US CPI and hence monthly inflation-linked bond returns seem to be better correlated with currency over the medium-term. with the monthly correlation between US CPI and the Broad Trade Weighted USD showing an r-squared of only 6% from 2000-2009. global inflation-linked bonds add a further diversification benefit to a portfolio. especially over the long term. However. A broad directional relationship was still clearly intact. specifically. the correlation fell sharply from Q3 08 as the fall in TIPS breakevens. but the weak correlation between TIPS breakevens and the broad Trade Weighted USD suggests that a portfolio of domestic inflation-linked bonds offers very limited value as a currency hedge. This was notably higher from late 2004 to mid 2008. as inflation and currency movements should be well correlated. inflation expectations appear. the long-term link between short-term currency movements and realised inflation is very weak. as a greater number of central banks began buying TIPS. the correlation between the total return on TIPS breakevens and the Broad Trade Weighted USD is statistically still very weak. i. Figure 232: US CPI not correlated with Trade Weighted USD 150 140 130 120 110 100 90 1997 US CPI index Broad Trade Weighted USD (RHS. inv) 80 90 100 110 120 130 140 1999 2001 2003 2005 2007 2009 Figure 233: TIPS breakevens not an effective currency hedge 115 110 105 100 95 90 85 80 75 70 65 60 1997 TIPS index performance versus nominal Treasuries 90 Broad Trade Weighted USD (RHS. a currency-hedged global linker portfolio maintains the most attractive features of a domestic inflation-linked bond portfolio. This sample period allows us to strip out initial liquidity-based distortions in TIPS in the early stages of the TIPS program. However. with an r-squared of only 18% from 20002009. optically at least. inv) 100 110 120 130 140 150 1999 2001 2003 2005 2007 2009 Source: Bloomberg. at 78%. better diversification. owning inflation-linked bonds as part of a non-currency hedged portfolio of other assets reduces returns volatility from currency fluctuations. However.

the impact of macroeconomic shocks and financial market volatility should be more evenly spread than for domestic linkers. The difference in volatility between the returns on a currency hedged index of global linkers and on an index of domestic inflation-linked bonds is that.21% 1.26% 1.20% 1.54% 0. There is also a broad similarity between the expected paths of developed country interest rates. leading us to see a global linker portfolio as the best way to capture the strategic benefits of owning linkers.7% peak-to-trough through the deleveraging of financial assets from October to December 2008. However.23% 1.29% 1.22% 1. using historical mean and covariance 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1.27% Diversified Portfolio Diversified Portfolio + Global Linkers Figure 235: Global linkers in an optimized portfolio of US assets.6% over this period.35% 1. for a portfolio of global linkers.25% 1. The slight divergence is explained by the fact that returns on TIPS are purely sensitive to changes in expectations for the future path of US real rates while the global inflation-linked bond index provides exposure to future real rate expectations over a variety of developed countries. When examining the properties of the Barclays Capital World Government Inflation-Linked Bond Index versus Treasuries. For instance. 15 March 2010 241 .48% 1. using Historical mean and covariance 0. The total return on our global inflation-linked bond index also holds a low correlation to riskier assets such as commodities. Thomson DataStream. we find.56% 0.26% 1.55% 0. the performance of our global inflation-linked bond index holds a high correlation to the total return on US investment grade corporate debt. Barclays Capital The return volatility on our index of global inflation-linked bonds has been considerably lower than on a pure exposure to TIPS.Barclays Capital | Global Inflation-Linked Products – A User’s Guide even if this already includes domestic inflation-linked bonds. hence the correlation our global inflation-linked bond index holds to US fixed income assets. while the total return on the currency hedged Barclays Capital World Government Inflation-Linked bond index fell by 7.23% 1.38% Global IL Bonds MBS GSCI S&P 500 US Investment Grade US High Yield US Treasuries Source: Bloomberg. the total return on the Barclays Capital TIPS 1y+ index fell by 10.49% 0. the S&P 500 and the GSCI. The explanation for this is intuitive. we also find that the total return on our global inflation-linked bond index holds a stronger correlation than TIPS to the total return on the GSCI and S&P 500.32% 1. mortgagebacked securities. but a weaker correlation to the performance of US fixed income assets. Barclays Capital Source: Bloomberg. For example. US high yield and investment grade corporate debt. unsurprisingly.50% 0. from a mean-variance perspective at least.51% 0. history indicates that a portfolio of global linkers is the fixed-income asset of choice.24% 1. Thomson DataStream. A global portfolio of currency-hedged linkers maintains almost the same correlation to US inflation as a pure TIPS portfolio and currency hedging a global inflation-linked bond portfolio generally compensates for any differences in local inflation rates.53% 0.20% 1. that the relationship global linkers hold with these assets is very similar to the relationship of TIPS to these assets. Figure 234: Global linkers greatly enhance the efficient frontier. like TIPS.52% 0. suggesting that.

we do not include separate exposure to TIPS in addition to our global inflation-linked bond index as TIPS already make up the largest domestic allocation in our global inflation-linked bond index. investment grade corporate bonds. risk/return. The volatility advantage which global linkers hold over the other asset classes in our study is even greater when considering real. Global linkers were selected at a negligible weighting when the risk on the portfolio was below 1% and occupied less than 10% of the portfolio until risk was increased to just above 1. 39% at the end of 2009. The inclusion of global linkers was to the detriment of equities and US fixed income assets. thereby pushing out the efficient frontier. at which point this rose rapidly to the maximum allocation of 25%. the inclusion of global inflation-linked bonds within our mixed portfolio of US assets lowered risk and sharply increased returns.Barclays Capital | Global Inflation-Linked Products – A User’s Guide As a result. rather than nominal. hedged to USD. the benefits of exposure to inflation-linked bonds from a variety of developed markets holds positive implications for a portfolio of assets. MBS. Therefore. The example we present here is when exposure to the total return on the Barclays Capital World Government Inflation-Linked Bond Index. Naturally. we limited the contribution of global linkers to a maximum of 25%. even for investors who do not have explicit inflationlinked liabilities. is added to our portfolio of US Treasuries. 15 March 2010 242 . in particular investment grade corporate debt.2%. the S&P 500 and GSCI. As the Sharpe ratios on global inflation-linked bonds are much higher than on the other asset classes in this study.

3 5. we aim to include variables that are both consistently statistically significant through the observation period and economically logical. Modelling the dynamics of real yields and breakevens has long been both an academic and a market goal.82 t-stat 13.5 R2 Model stats 0. there is a significant danger of spurious regression relationships where coincidentally similar trends create a strong statistical link.182 0.8 -12. provide sufficient extra statistical rigour to compensate for their lower transparency.0 2. We attempt to model 10y real yields and breakevens as a sector that is likely to be relatively liquid but also more fundamentally driven than elsewhere on the curve. This is a particular danger for real yields. with a trend towards falling real yields over the course of the last decade. For breakevens. There are significant shortcomings to this approach.0 1. Even so. as we have previously pointed out.296 0. euro area and UK.923 Jan 00 Jan 02 Jan 04 Jan 06 Jan 08 Jan 10 Source: Haver Analytics.5 2. but with some useful insights for Europe and the UK.110 -1.skeoch@barcap. Figure 236: 10y TIPS real yield versus fair value model 4.6 5. Many of the variables within the models for different countries are similar and.0 0. with breakevens higher in environments when the front end is pricing in significant rate hikes.5 -1. in particular the instability of regression components over time which leaves this type of model inappropriate to form the basis of a trading strategy. The long end is more likely to be significantly distorted by supply/demand considerations. Our TIPS real yield model has proved surprisingly robust through the turmoil of recent years.0235 -0.james@barcap.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Modelling real yields and breakevens Alan James +44 (0) 20 7773 2238 alan. we have not found that more complex time series methods. the factors are almost unchanged from those that were shown in previous editions of this publication. Hence. 10y real yields are understandably highly correlated with policy rates in the US. However. where statistically the series in all three markets appear to be non-stationary. The front end is liable to be particularly sensitive to short-term carry and seasonality factors. such as ARIMA and GARCH-based modelling.5 4. euro area and UK. finding them most useful for the TIPS market.27 0. it is typically the slope of front end of the cash curve that is a more significant variable.com We update our ‘fair value’ models for real yields and breakevens. the topic is something of a notorious analytical quagmire. Bloomberg. in some cases.5 1. Barclays Capital Source: Haver Analytics. In this section. including in earlier editions of this guide.com Henry Skeoch +44 (0) 20 7773 7917 henry. Bloomberg. we put forward linear regression fair value models for the US.5 Jan 98 10y TIPS real yield Model Figure 237: 10y TIPS real yield model coefficients and statistics Coefficient Change in non-financial business credit (1y lag) Gasoline (log) Fed Funds futures slope (6m minus front) Fed funds target Trade-weighted dollar Constant 0.5 3.0 3.9 14. Barclays Capital 15 March 2010 243 . Nonetheless.

The importance of the two other drivers of real yields within our fair value model. The standard deviation of the model over this time was 43bp. double that in the previous ten years and with the largest deviation between actual and model yields both to the high side and low side occurring during this period. The uncertainty over model estimation has obviously increased during the period of elevated market and economic volatility in 2008 and 2009. Retail gasoline prices appear to impact real yields due to the direct pass through into carry. The trade-weighted dollar has on average tended to be positively correlated with real yields. Nonetheless. Testing out of sample after August 2007. This is for two main reasons. the model itself has proved surprisingly resilient.Barclays Capital | Global Inflation-Linked Products – A User’s Guide 10y TIPS real yield A significant element of TIPS real yields can be described by the fed funds target rate. While there is strong statistical evidence suggesting changes in loans growth lead real yields the length of this lead has not been stable. The other major structural driver of real yields that is apparent is business loan growth. with the slope of the Fed Funds futures strip out to six months a lesser factor. but in these it has been hard to disentangle measurement issues due to the change in status of financial entities from August 2007 onwards. which proves to be an effective leading indicator of supply pressures. energy prices and the currency. and the factor was also unstable during the sharp run-up in energy prices to all-time highs in mid-2008. but was a relatively insignificant variable prior to central bank buying of TIPS picking up from 2005. we include both factors within our model as over shorter time periods they have been pivotal drivers of the market. varying from as short as six months to as long as eighteen. Previously we have used pure business loans series. the real yield model is always less than 10bp away from that estimating over the whole sample. when the first strong indication of the end of the credit boom emerged with the dislocation of cash and Libor rates. Energy was relatively insignificant prior to 2003 though. has also been somewhat unstable over time. The first of these is the uncertainty as to the status of CPI measurement during 1997 following the publication of the Boskin Commission report in December 1996. Real yields were exceptionally cheap versus the model between October and Thanksgiving 2008 when the TIPS market dislocation due to breakeven deleveraging was at its extreme. with energy price increases driving real yields down in the expectation of significant positive carry. Despite this caveat we use a 1y lag of annual growth of business lending within our model. The most dramatic overshoots versus the model are also entirely understandable in hindsight. However. despite the volatility of real yields and many of the inputs. Coupled with a relative illiquid and poorly defined market for the new asset class. In our model we use the Federal Reserve flow of funds series for all credit market debt of non-financial businesses. Real yields by contrast moved very rich versus the model in March 2009 following initial the FOMC announcement of direct purchases. we see 1997 as a poor starting point for estimating TIPS fair value. 15 March 2010 244 . Note that we estimate our 10y TIPS models from the start of 1998 rather than the first TIPS issuance in February 1997.

0 Jan 98 10y TIPS Breakeven Model Figure 239: TIPS breakeven model coefficients and statistics Coefficient Fed Funds slope Global Industrial Confidence (1m lag) Gasoline (log) Nominal Liquidity Spread Constant 0. Barclays Capital Source: Bloomberg. the model still significantly underestimates the degree to which breakevens collapsed in late 2008. while retail gasoline prices have always been a major driver of breakeven valuations. While an entirely logical factor. To partly offset this instability we introduce a liquidity variable.0 1.0131 2. which we would expect to coincide with a rise in inflation expectations. However. Nonetheless. the spread between a fitted Treasury curve which includes old bond issues and the on-the-run nominal 10y. whose importance has also been unstable and was not significant during 2008 and 2009. Figure 238: 10y TIPS breakeven versus fair value model 3.02 t-stat 5.5 0. but the model has not proved stable through the extreme breakeven volatility of 2008 and 2009.5 2. The final variable in the model is the slope of the Fed Funds futures strip. Over most of the sample period the most important factor driving the breakeven model is global industrial confidence.210 0. We see such a deviation from fair value as understandable. Even with the inclusion of this factor. Barclays Capital 15 March 2010 245 .5 1.Barclays Capital | Global Inflation-Linked Products – A User’s Guide 10y TIPS breakeven Previously we have had a relatively parsimonious three factor model for 10y TIPS breakeven fair value. the sensitivity to energy movements has varied considerably over time.8 6.4 -5. as much of the valuation of the 10y breakeven comes from the degree of benchmark premium. but most of the major deviations can be explained by flow factors. indeed almost inevitable.9 R2 Model Stats 0.395 0.376 -0. Similarly. Notably this series tends to dominate currency and broad commodity variables. We use a proprietary series that Barclays Capital economists have developed based on combining major national business confidence surveys weighted by purchasing power parity adjusted industrial production. It also tends to lead breakeven movements modestly. for a period in which the vast majority of breakeven positions were closed out by a combination of deleveraging pressure and real money investors being forced out of off-benchmark positions in TIPS. while this series has been consistently statistically significant its importance has varied substantially over time. hence. we see this as an appropriate measure to capture the degree of risk premium within the market and expect that it will quickly become an important factor again as rate hike expectations rise.1 7.814 Jan 00 Jan 02 Jan 04 Jan 06 Jan 08 Jan 10 Source: Bloomberg. which implies a degree of under specification. so we are relatively comfortable with the formulation of this as a fair value model.0 0.0 2.3 36. it was barely statistically significant prior to 2008. Statistically the constant is the most important variable in the model. it is in the model with a one month lag.

Currency variables do not show up as a significant factor for euro real yields.Barclays Capital | Global Inflation-Linked Products – A User’s Guide 10y Euro real yield Of our real yield models. with significant increases in money supply tending to push real yields higher. Barclays Capital Source: Bloomberg.4 17. the log of Brent crude prices and global industrial confidence. The ECB policy rate and energy were highly significant for French real yields during this period though.92 4. but there is no stability in their degree of sensitivity.0 2.5 2. is lagging developments in real yields in contrast to the leading tendency of the higher-frequency data. if anything. Coupled with the fact that for more than half of the sample period since 2002 euro breakevens were in a very tight range there is no out of sample stability for a two factor model even excluding the volatility of the second half of 2008. rather than using OATi data to extend the time period back to 1998.43 t-stat -7.622 -0.4 22. The most important drivers are the ECB refinancing rate and the log of Brent crude prices. it is hard to know if inflation expectations were consistent. M3 is a significantly less reliable variable through the initial introduction of the euro and was again distorted into the start of 2001 with the introduction of the physical currency alongside the entry of Greece into the monetary union. our preferred model uses the three monthly change in seasonally adjusted M3. The model underestimates the spike higher in real yields in late 2008 only modestly. In addition. 10y Euro breakeven We do not present full details of a euro breakeven model as we cannot find variables to provide a sufficiently well specified model since 2002. Liquidity prior to the launch of the OAT€i12 was limited and whilst the trend for French inflation over this time was relatively similar to that of euro HICPx.0 -17. Barclays Capital 15 March 2010 246 . Figure 240: 10y Euro real yields versus fair value model 4.0 0.5 0.0 Jan 02 10y Euro real yield Model M3 3m change (1m lag) ECB refi rate Brent crude (log) Constant Figure 241: Euro real yield model coefficients and statistics Coefficient -0. Specifically. which impacts real yields with a lag of one month. There are two variables which are consistently significant drivers of breakevens.0 1.5 3. this variable. Note that we estimate our euro real yield model since the start of 2002. The slope of the front end of the euro curve was statistically significant until 2007 but has not been since and does not help improve the stability of a model including the energy and global confidence variables.286 0. but the sensitivity of both these factors varies over time. the 10y Euro model appears the least stable. While the statistics of the equation are improved slightly by using the y/y change in M3. The other significant element is broad money supply growth.844 Jan 04 Jan 06 Jan 08 Jan 10 Source: Bloomberg.5 1.0 3.2 R2 Model Stats 0. we are aware of the higher risk of a spurious statistical link in the smoother and more trending annual change data.

The level of the FTSE 100 share index is an important determinant of gilt linker valuations. However.000258 -0. from mid 2007. We model UK real yields since the start of 1998. When share prices are high it accelerates the ongoing trend since the late 1990s for schemes to switch out of equity and into fixed income.4 9. Barclays Capital 15 March 2010 247 . the spike yields in Q4 2008 is not picked up.5 0.744 Jan 00 Jan 02 Jan 04 Jan 06 Jan 08 Jan 10 Source: Bloomberg. with real yields tending to fall as equity prices rise. in aggregate pension funds tend to reduce their asset versus liability mismatch when they can afford to do so.0 Jan 98 Model Figure 243: UK real yield model coefficients and statistics Coefficient Base Rate 4th-1st Short Sterling Contract FTSE 100 Global Industrial Confidence (1m lag) Constant 0.0 -7. with real yields being particularly sensitive to this flow. when the linker market started to become significantly dislocated and dysfunctional. in our real yield model we take the spread between the fourth and first Short Sterling contracts. As in our breakeven models which include this variable.5 1. leaving the positive relationship between falls in confidence and breakevens tending to push real yields around. It is surprising that this variable appears with a negative sign – ie. This negative sign is explained by nominal gilt yields having very little sensitivity to global cyclical variables when monetary policy variables are accounted for.1 10. The slope of the Short Sterling curve is very significant in conjunction with the policy rate. but its statistical importance is relatively stable.5 10y UK Real Yield 3. Figure 242: 10y UK real yields versus fair value model 3. we find that it works best when leading yields by a month.469 -0.321 1. a fall in confidence tends to push up real yields. In common with real yield models in other countries.5 R2 Model Stats 0.0 1. We also find that including our global industrial confidence series improves the stability of the model. Unsurprisingly.Barclays Capital | Global Inflation-Linked Products – A User’s Guide 10y UK real yield The Bank of England Base Rate has proved to be the major driver of 10y real yields in the UK.393 0.9 -8. with the model dealing poorly with the period prior to this.0 0. the long-term sensitivity of 10y nominal yields to share prices is low and the level of the equity market is a significant factor in determining asset allocation decisions from pension funds into gilt linkers. This is in order to avoid misspecification stemming from the structural break following the decision in May 1997 to give the Bank of England monetary policy independence.41 t-stat 19.5 2. Barclays Capital Source: Bloomberg.0 2. Including this global industrial confidence factor tends to dominate both the currency and commodity variables that we have used in previous gilt real yield models. The dynamics for linkers were also significantly changed around this time with the implementation of the minimum funding requirement for pension funds. too. While the direction of this relationship is seemingly illogical. the dynamics of yields in 2009 are tracked closely even estimating out of sample from the period prior to mid-2007.

703 Jan 00 Jan 02 Jan 04 Jan 06 Jan 08 Jan 10 Source: Bloomberg. Movements in front end yields also feed into RPI inflation via mortgage interest costs. either of these two would show up as the most statistically significant factor within the model.923 t-stat 4. given the difficulty of defining the significance of the cyclical coefficients stemming from their statistical overlap.4 9.5 9. Figure 244: 10y UK breakeven versus fair value model 4. UK breakevens are not highly responsive to oil prices. Barclays Capital 15 March 2010 248 .5 1. Without these two variables. The correlation between these two cyclical variables masks the statistical importance in the model of each of the variables and leaves the value of each relatively unstable. Testing the model out of sample using an estimation to mid-2007 highlights that the sensitivity to various components has been unstable over time.Barclays Capital | Global Inflation-Linked Products – A User’s Guide 10y UK breakeven 10y UK breakevens. Unlike in either the US or euro area.5 2.5 Jan 98 UK 10y Breakeven Model Figure 245: UK breakeven model and coefficients Coefficient Global Industrial Confidence (1m lag) CRB Raw Industrials Short Sterling (4th vs.00304 0. On their own. while the relative significance of the industrial confidence and industrial materials prices coefficients is reversed. Broader cyclical and commodity price pressures remain significant though. with the out of sample model implying breakevens 30-50bp lower than the whole period estimation. the level of sterling would also show up as an important component of breakevens. 1st) UK Base Rate Constant 0.8 5. but not including both notably reduces the effectiveness of the overall model. The former may be more concerning than the latter.0 0. The importance of the interest rate variables is notably lower over the whole sample period than the pre July 2007 period.2 R2 Model Stats 0.5 3.180 0. This is due to the combination of a notably lower pass through of energy price movements into RPI inflation than elsewhere and with most of the drivers of the 10y sector coming from the supply/demand dynamic at longer maturities. but it is dominated by the two factors that we use.0 2.0 1.7 7. are heavily influenced by monetary policy. Barclays Capital Source: Bloomberg.0 3. hence our model includes variables of global industrial confidence and the level of CRB raw industrials prices.376 0. like real yields. The importance of these factors is not just due to real yields being naturally stickier than nominals.182 0. with a strong positive correlation with both the level of the Bank Rate and the slope of the Short Sterling strip. despite all components remaining highly significant.5 4.

ie. which the vector needs to account for. This influences the level of breakevens across the year and micro valuations on the curve. or as an average month-on-month seasonal factor. persistent deviations from trend at specific times of the year.com Stefan Liiceanu +81 3 4530 1554 stefan. while TRAMO/SEATS is widely used in Europe. Four methods of seasonal adjustment are common: ratio versus moving average (or sum). and the amplitude of seasonal adjustments constant. while from a practical perspective there is evidence in many countries that there has been a tendency for the magnitude of seasonality to increase in recent years. Seasonal adjustment attempts to isolate seasonal patterns from a time series by detrending the series.com Henry Skeoch +44 (0) 20 7773 7917 Henry. There are two ways of constructing a seasonal vector – additive and multiplicative. Seasonality is most commonly quoted either in vector form.mirani@barcap. with a seasonal model needed to build an inflation swap infrastructure. The vector form is most useful in assessing relative seasonal value between different maturities. X-12 is the most widely used by statistics agencies. However. with fitting an ARIMA model an optional extra (full details are available on the US Bureau of the Census Website). To the extent that seasonal trends are predictable an efficient market ought to price these in. there is an argument that certainty over the seasonal vector declines over time so any forecasted seasonal should contain a decay component. We see it as appropriate to assume a static seasonal vector and adjust this as more information becomes available. 15 March 2010 249 . the degree to which the underlying index has to be adjusted to create a seasonally adjusted series. variants of US Census Bureau’s X-12-ARIMA.Skeoch@barcap.com Chirag Mirani +1 212 412 6819 chirag. whereas the core seasonal adjustment method underlying X-12 uses a series of non-parametric moving average filters. whereas the m/m factors highlight the seasonal trends that will be reflected in bond carry. ie. Even in countries where statistics agencies publish full seasonality estimates for market participants to observe. This gives X-12-ARIMA the upper hand when constructing a seasonal vector including individual components that do not exhibit highly statistically significant seasonality. The main difference between TRAMO/SEATS and X-12 is that TRAMO/SEATS estimates and fits an ARIMA model to the data in order to derive an estimate of seasonality. meaning that multiplicative adjustment should be used – that is. It is also a very important feature for valuing non-annual inflation swap points. and the Bank of Spain’s TRAMO/SEATS procedure. having existed in some form since the late 1960s (although many agencies use the less complex X-11 variant). While this can be confusing.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Seasonality and inflation markets Alan James +44 (0) 20 7773 2238 alan. However.com Seasonality is an important factor for several areas of inflation-linked markets because inflation bonds and swaps are almost always linked to non-seasonally adjusted inflation indices. incorporating some form of statistical diagnostics. regression using dummy variables. Most price indices have a marked long-run upward trend. filtering any seasonality to leave residual ‘white noise’. Inflation indices usually show evidence of seasonality. Typically both bond and swap markets price in significantly less seasonality for future years than statisticians would estimate given current data.liiceanu@barcap. calculating seasonal factors which seasonally adjust the index by multiplying it. a direct transformation between these two formats is mathematically straightforward. it is not always straightforward for market participants to observe seasonality. Additive seasonality is used when the index is stationary. A counter-argument to this is that there is no reason to dampen an unbiased estimation.james@barcap. but still show discernible patterns. there is uncertainty as to how this seasonality will develop over time. The seasonal adjustment process will typically employ an algorithm or model to search for these seasonal patterns.

be circumvented by estimating a seasonal vector for each component of the index. however. 15 March 2010 250 . which attempts to adjust for monthly distortions. This can.1% 0. Where appropriate an intervention analysis is used.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Generally we prefer to use seasonality estimates from statistics agencies if they are published in sufficient detail.4% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Source: Bureau of Labor Statistics.3% -0. The NSA CPI tends to rise most notably between February and May. One way to visualise the typical seasonal pattern is shown in Figure 246.1% -0. but can bias the end result (due to the Cauchy-Schwarz inequality). Haver Analytics. Estimating on an aggregate series on average tends to produce a slightly higher seasonality than considering sub-components.3% 0. Problems can arise in adjusting an inflation index constructed from unchained series using a set of weights.4% 0. Seasonal factors are reestimated and published at the start of each year. The adjustment is carried out on subcomponents of the index and then aggregated. whether or not the sub-component approach estimates all variables or only those with significant seasonal components. which tends to produce a slightly more conservative estimate of seasonal factors than top-down estimation. Successive unchaining and chaining of series is not only laborious. follows a regular seasonal pattern. which adjusts for sharp and permanent shifts in the underlying trend. 2001-08 (difference between m/m NSA and SA% chg) 0. Figure 246: CPI seasonals. but where this does not exist we suggest using an X-12 estimation at as low a level of aggregation as feasible. then aggregating these individual vectors by the index weights to give an overall seasonal vector for the index. United States Consumer inflation. and are accounted for via regression-ARIMA models.2% 0. even before the dampening impact of the intervention analysis is considered. Barclays Capital Average: 2001-2008 2009 The Bureau of Labor Statistics (BLS) estimates the seasonal factors by applying the X-12ARIMA seasonal adjustment model developed by the US Bureau of the Census. with seasonality negative from July onwards. at an individual component level. then chained at the index level. as measured by the Consumer Price Index.2% -0. coinciding with the January data release and causing a change in the historical seasonally-adjusted series. While seasonals vary over time there has been a consistent tendency for prices to rise more quickly in the early part of each year. which plots the average difference between the Not Seasonally Adjusted (NSA) and Seasonally Adjusted (SA) monthly percentage change in the CPI. These have the potential to distort the results of the seasonal adjustment.0% -0. but decline significantly in November and December.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 247 shows a breakdown of the estimated seasonal contributions from the main CPI components. Shelter costs begin to rise early in the year as travel patterns start to normalise.20% 0. They fall off again in June-July as summer merchandise is cleared ahead of the backto-school season. fall off in May-July. Haver Analytics.30% 0. NSA core inflation tends to rise significantly in February and March. In contrast.00% -0. while the seasonality of energy is larger than other components it is also relatively less stable. are relatively stable throughout the year. Seasonality in shelter is mainly driven by out-of-town lodging. They are calculated as the implied difference between the NSA and SA percentage changes for these components.10% -0. However. The main contributors are shelter and apparel costs. energy prices tend to fall off in the fourth quarter. December marks the low point of the year for out-of-town lodging costs. then rise notably in February and March. 15 March 2010 251 . average 2001-08 (difference between m/m NSA and SA % change of CPI components) 0. These prices fall off in November-January as holiday discounting dominates. Energy prices tend to rise most significantly in April-June.20% -0. which tends to soften in September-December. weighted by their relative importance in the headline CPI. as demand falls off when the summer travel season ends. Seasonal fluctuations in the headline index are driven mainly by movements in core and energy prices during the year. Barclays Capital As can be seen in Figure 247. when they pick up again. presumably because travellers tend to stay with family rather than in hotels during the holiday season. The seasonal pattern of home heating oil is generally the reverse of gasoline but has a much smaller weight. as the anticipation and onset of the summer driving season tends to put upward pressure on retail gasoline prices in those months. so gasoline is the dominant factor. food prices. Figure 247: Contributions of components to headline CPI seasonal pattern.30% Jan Feb M ar Apr May Jun Jul Aug Sep Oct Nov Dec Core Energy Food Source: Bureau of Labor Statistics.10% 0. in contrast. rise in October. Figure 248 shows the most important factors behind this.40% 0. and fall sharply in November and December. Another important contributor to core seasonality is apparel.

10% 0.05% 0. going back to March 1997.80% 0.40% -0.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 248: Contributions of core components to headline CPI seasonal pattern.15% -0.10% 0.20% -0.40% 0.20% 0. It is notable that this tendency towards seasonal performance is much stronger in TIPS than in other inflation markets. This pattern in the US is due primarily to the CPI seasonality discussed above.30% -0.40% Breakeven Returns (Monthly.20% Jan Feb M ar Apr May Jun Jul Aug Sep Oct Nov Dec Apparel Shelter Source: Bureau of Labor Statistics. While divergences certainly occur.30% 0.60% -0.20% 0.00% -0. Figure 249 shows that breakeven returns have traditionally risen in the first half of the year and fallen in the latter. we can detect seasonal patterns in breakeven returns in general and examine the relative value impact on issues that will mature in different months. Haver Analytics.00% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 0. Figure 249: Average monthly breakeven returns are higher in the first half of the year 0.20% -0. LHS) Average CPI Seasonals: 2001-2008 Data Period: March 1997 to December 2009 Source: Barclays Capital 15 March 2010 252 .15% 0. Barclays Capital Seasonality and TIPS The patterns discussed above play a very important role in the traded inflation market.60% 0.10% -0.50% 0.05% -0. We calculate these averages by looking at duration weighted breakeven return averages by month. despite seasonality itself not being significantly larger.40% 0.00% -0. average 2001-08 (difference between m/m NSA and SA % change of CPI components) 0.10% -0.20% 0. Specifically.00% -0.80% -1. One of the most noticeable effects that inflation seasonality can have on the tradable inflation market is the seasonal pattern of breakeven returns.

5 Jul12MaturityCPI = Jan12MaturityCPI * EXP (Trend * (T) + Seasonality) For this example. The Jan12 and Jan14 maturity NSA CPI. and investors need to account for curve. The forward breakeven is assumed to be the general trend of an NSA CPI path between them. The cumulative seasonality between the two issues is 71bp. Then. as the seasonal effect tends to be this way during the first half of a year. 2009).18%).77 (breakeven: 1. are 220.Barclays Capital | Global Inflation-Linked Products – A User’s Guide While the carry-based seasonality buying effect may retreat in future as efficiency increases in the TIPS markets. Methodology The general approach to calculate appropriate issue-specific breakeven levels based on the current market-implied forward breakeven trend and seasonal assumptions uses the trend and adding cumulative seasonality to project maturity NSA CPIs for specific issues. Based on this approach and using 2008 seasonal assumptions.49 (implied breakeven: 1. 15 March 2010 253 . All breakevens should be adjusted for respective floor values. as implied by the markets. depending on your seasonal assumptions.50%). all else being equal. for the Jul12s. As an example. Jul12 market implied (with our seasonal assumption) Maturity NSA CPI should be 225. April maturity issues should trade cheap to those that mature in January because the April issues picks up one extra December print. In other words. Below we discuss a methodology by which it is possible to use a seasonal assumption to detect value among TIPS issues that are less than one to two years apart. Similarly. Example: We pick two issues that are exactly one or two years apart in maturity and calculate their implied forward breakeven.83%. the market consistently underpriced breakeven levels for July issues relative to nearby January issues throughout 2009. we proceed to calculate the cumulative seasonality between the Jan12 and Jul12. All else is not equal though. here it is positive. 15 October 2009. when seasonality is consistently and significantly negative due to holiday discounting and other factors. To calculate Jul12s maturity NSA CPI. We use seasonality assumptions similar to that of 2008 (CPINSA-CPISA).14 (as of 18 November. A quick way to estimate floor value is by taking the difference between Traded ASW versus calculated ASWs levels. Jul12 breakevens should be 32bp higher if one assumes seasonality similar to 2008.83% Seasonality = Cumulative Seasonality between Jul12 (April15) and Jan12 (Oct15) = 71bp T = (Jul12Maturity – Jan12Maturity)/365 = 0. for related notes). we use following equations: Trend (Annualised) = (Jan14MaturityNSA-CPI/Jan12MaturityNSA-CPI)^(1/2) = 1. while the market is implying it should be 222. we also find that the market misprices issue-specific seasonal differences. July issues should trade rich to January and April TIPS because they are expected to benefit from the typically good spring carry period.’ Global Inflation-Linked Monthly. a continuation of an ongoing theme. Remember the seasonality should be calculated between the lagged maturity dates. For example. floor value and other factors that can drive relative value to isolate seasonal effects. respectively. These projected NSA CPIs can then be used to calculate appropriate breakeven levels (see ‘Marketimplied breakeven curve.97 and 229. giving a forward annualised breakeven trend of 1. we first calculate the forward breakeven trend using Jan12 and Jan14 bonds.

The Insee statistics institute estimates French inflation seasonality at a headline level rather than within components or sectors so is not directly comparable to the equivalent ECB series despite using a similar X-11 approach.4% -0. Barclays Capital Figure 251: French CPIx m/m seasonality 2009 2008 2007 0. but there has been a tendency towards increasing volatility of retail prices in other countries.5% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Source: INSEE. Eurostat has encouraged a standardisation of processes across Europe.0% -0. An X-12 analysis of the energy series suggests that any seasonality is highly unstable and unlike the US there are few months in which the sign of the month-on-month seasonal has stayed the same over the long term. Italy and Spain did not include apparel sales prices in their price data. Figure 250 and Figure 251 show monthly seasonality Figure 250: Euro HICPx m/m seasonality 0. too. but given the exclusion of energy components we would caution against comparison with seasonals in other countries that include this sub-sector. the exclusion of energy components is somewhat controversial and tends to reduce aggregate seasonality.4% 0. offset by a similar fall in the three months to January. mainly due to changes in measurements and the deregulation of retail prices in several countries. However. leading to more seasonality in the aggregate series due to the timing of distortions being more consistently measured. for example. using the latter to seasonally adjust bond breakevens in its macro analysis.0% -0.8% -1. with petrol prices increasing an average 5%. Relative to the BLS estimation at the lowest possible level of aggregation and with more intervention. Barclays Capital 2009 2008 2007 p 15 March 2010 254 . Seasonality in euro area HICPx increased in the early part of the decade.6% -0. On the other hand. but conducted at a higher level of aggregation (only five sub-sectors) and excluding energy.2% 0.4% 0.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Euro area Eurostat does not publish official seasonally adjusted HICP series. The clearest example of this is that until 2001.0% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Source: ECB.4% -0. but the ECB does publish seasonally adjusted series for both headline HICP and HICPx. in practice there is not a large difference between the relative seasonality using the reported series and adjusting each on a consistent basis.2% 0. However.1% -0. Eurostat. as the ECB argues that there is no evidence of seasonality in this series while estimation is volatile. This has lessened the tendency for the impact to the aggregate index to be smoothed out.2% -0.3% 0.1% 0. The ECB series indicate that there is now more seasonality in the euro area than in the US. there has been a strong tendency for petrol prices to rise in the three months to June for the past 10 years. These series use an X-12-ARIMA model similar to that in the US.3% -0.2% -0.6% 0. We are relatively comfortable to use the ECB data for HICPx seasonality for considering valuations within the euro inflation market. the ECB methodology may produce a slightly higher estimate.

We avoided deploying an ARIMA model as part of the X-12 process. for the past three full years of data. and there is a borderline statistical rejection of Petrol and Oil. In terms of micro valuations it is notable that OAT€is maturing in July mature at a favourable time to benefit from the upswing in prices during the spring. United Kingdom For the UK. the seasonal estimate we obtained from this disaggregated level of RPI produced slightly more seasonality that running X-12 on the aggregate RPIX series. as it is very hard to trade swaps without using a seasonal vector to price non full-year positions. respectively. Council Tax is only raised in April. whereas German issues redeem close to the least favourable time in April. However. Statistically significant seasonality was detected in Seasonal Food. This may be due in part to more active trading in swaps than elsewhere. Other Goods and Services ex-Utilities. The trends in Euro and French seasonality have become relatively similar in recent years. Euro seasonality is generally priced more efficiently than in other markets. 15 March 2010 255 . the spread of more recent German issues to the OAT€i breakeven curve have on average priced seasonality in line with the ECB estimate of seasonality – 0. but this is no longer significant. While it is true that the seasonality in oil is somewhat unstable. Unusually. there is still a clearer pattern than in the euro area. Alcohol and Tobacco. seasonal adjustment of the overall RPI Index derives statistically significant seasonality. To gauge the stability and drivers of seasonality. but the Office for National Statistics (ONS) does not produce a seasonally adjusted RPI series. On average. we created a seasonal vector by using X-12 to adjust the key components of the RPI. which would mean an inconsistent seasonality estimate. In the cash linker market there is very little evidence of seasonality biasing performance across the year. However. since the DBR€i16 was first issued in 2006 this bond has priced around two thirds of the negative seasonality versus the OAT€i breakeven curve. partly due to seasonality of petrol tax increases. We found that a more consistent seasonal vector was generated when seasonally adjusting all components other than mortgage interest payments. then aggregating these vectors using the official index weights. Historically there was a correlation between the performance of OATis versus €is and the deviation of French and euro seasonality.9% based on 2008 full-year seasonals. so is clearly seasonal. as for some components there was no evidence of an underlying ARIMA process.Barclays Capital | Global Inflation-Linked Products – A User’s Guide estimates for Euro HICPx and French CPIx. The differences stem mainly from varying sales periods and a greater upward bias for euro (particularly German) prices for Christmas and Easter that are subsequently unwound.

6% 0.4% 0. The degree of seasonality priced into the UK inflation market is harder to observe than in the US or euro area due to irregularly dated old inflation-linked gilts and the inconsistency of these with new style issues. while alcohol and tobacco taxes are consistently raised immediately following the budget. while historically seasonally adjusting breakevens has done little to smooth either the curve or time series of breakevens. so almost always showing up in April RPI.1% -0. this is also seen in July. which indicates that the seasonal vector is reasonably stable. Petrol and oil has consistently shown a downward contribution in November. creates a relatively stark differential versus the most common 22 November redeeming bonds.8% 0. However. Hence. Alcohol and Tobacco Seasonal Food Services Petrol and Oil Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Source: Barclays Capital. This is understandable as council tax has always been increased at the start of each fiscal year since it was introduced.7% Others Council Tax. while the strong trend towards April price rises from taxation has eased. The relatively long sample period from 1987-2009 allows for a more accurate estimation of the underlying trend and seasonal factors than in the euro area. -0. ONS Figure 253: Key contributions to 2009 RPI seasonality vector 0. Figure 252: Barclays Capital estimated UK RPI m/m seasonality 0. Relatively poor inflation swap liquidity at shorter maturities does not help.2% -0. which excludes indirect taxation as well as mortgage interest payments.3% 0. but the introduction of 22 March maturing new-style issues. making it harder to accurately measure monthly swap rolls that could otherwise be used to back out implied seasonality. even as market efficiency was curtailed in early 2009. but with a significant element of UK seasonality driven by tax factors we prefer to use our own series. there is a strong upwards April effect due to council tax. The ONS produces a seasonally adjusted series for RPIY. seasonality appears to be also more stable than the US and most large euro countries.5% 0. the instability of any seasonality estimates based on headline RPI and hence distorted by the impact of aggressive rate cuts on mortgage interest costs has led to an increased need to accurately estimate RPI seasonality. ONS 2009 2008 2007 15 March 2010 256 . it is more likely to be an important factor.5% -0. although an estimation without stripping out mortgage interest payments would be notably less stable.6% -0. Goods show a strong downwards January effect due to seasonal sales.7% 0. Figure 253 shows the contribution of the various key components of RPI to the month-on-month seasonality.3% -0. even though the aggregate trend towards higher prices in Q2 is undimmed.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 252 shows the estimated seasonality for the past three years. In recent years there has been little to choose between the seasonality of gilt linkers. As can be seen. Even so.75% on our seasonality estimate.2% 0.0% -0.8% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Source: Barclays Capital. alcohol and tobacco. which capture most of the January negative seasonal.4% -0.1% -0.

a development that pushes consumer prices significantly above the year’s average in the last quarter of the year. ie. while the month of August exhibited negative seasonality as opposed to it currently being seasonally positive (ie. Clothing is by far the most volatile item in terms of seasonality. as fees for entertainment facilities. especially around Q1 and Q4 and over shorter sample periods (eg. On the other hand. Similarly. In contrast with the first quarter of the year. Historically. seasonality has been “most negative” in the first quarter of the year. but some changes did occur over time with regards to the strength of seasonality in Q2 and around the summer period – in the early 1990s the month of May for example featured stronger positive seasonality than it currently does (the 5y average ratio of non-seasonally adjusted core CPI to the seasonally adjusted core CPI was 1. flat or negative inflation rates. determining the bulk of the core CPI’s seasonal patterns. Prices of services also tend to decline overall in Q1. and 3) transportation and communications. while in the September-December period clothing explains nearly 80% of that period’s positive seasonality. Clothing prices subsequently rebound with winter lines coming in and prices of goods such as food and rice tend to rise as preparations for the New Year holidays get underway. as well as the fact that during the summer vacation season of July-August.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Japan 1) Seasonal features of the Japanese core CPI As is the case with consumer prices elsewhere. recreation/entertainment accounts for roughly 20% of the negative seasonality observed each year in Q1 and it explains about 40% of the positive seasonality around the summer period (when it is offset by negative clothing seasonality). leisure-related service prices. Historically Japanese seasonal patterns have been generally stable. at the same time as the release of headline and core CPI figures and past data is available to 15 March 2010 257 . The primary reason behind this is that in the January-March period. for further details please refer to Figure 256). 2) recreation and entertainment.0016 under the 2005-base core CPI. 2) Seasonal factors The seasonal patterns of the Japanese core CPI described above can be visualised using official seasonally-adjusted core CPI numbers published by the Ministry of Internal Affairs and Communications Statistics Office. This reflects price hikes in clothing from April. this item accounts on average for nearly 60% of the CPI’s negative seasonality (based on monthly data starting from 1993). the Japanese core CPI (CPI excluding fresh food) exhibits fairly regular seasonal patterns that reflect corporate sales practices. tend to increase significantly. five years or shorter). prices of goods tend to decline due to the impact of sales following summer bonus payments and due to clearing sales on summer clothing. In Q1 for example. decrease. seasonality tends to improve from the second quarter of the year and actually becomes “most positive” (or strongest) during the AugustDecember period.0036 versus 1. where usage declines in winter. as well as consumer habits and preferences.0010 under the 2005-base core CPI). on average. Over the past two decades the major items accounting for the index’ seasonality have been: 1) clothing/apparel. The seasonal patterns followed by Japanese consumer prices are fairly different from those observed in the US (a m/m correlation of just 24% since 2000) but similar to those seen in the euro area (an 82% correlation). Japanese consumer prices in Q1 tend to be lower than prices in other months regardless of whether the year as a whole exhibits positive. such as travel. prices of Japanese goods (especially clothing) tend to fall sharply due to the impact of New Year and year-end inventory clearance sales. a ratio of 0.9974 compared to 1. These are disseminated on a monthly basis.

9996 1.9963 0.9964 FEB 0.10% 2000-Base CPI 2005-Base CPI 0.9964 0.0030 1.0015 1.9946 MAR 0.18% 0.24% 0.63% Jan -0.12% 0.0036 JUN 1.4% 0. Furthermore.0014 1. This suggests that the seasonal adjustment of JGBi real yields and breakevens can employ longer-run averages.9995 AUG 1. Barclays Capital Figure 256: Japanese core CPI average monthly seasonal factors JAN 2000-Base 2005-Base 2000-2009 Avg.6% -0.000 0.22% -0.006 1.9962 0.0016 1.0012 1. Figure 254 illustrates the time series of the ratio between the non-seasonally adjusted core CPI and the seasonally adjusted core CPI.0010 1. Source: MIAC.002 1.9972 0.0009 0.998 0.00% -0.14% -0.2% -0. The seasonality adjustment applied by the Ministry to original core CPI time series is based on the X-11 ARIMA method originated by the US Census Bureau 18.0024 1.Barclays Capital | Global Inflation-Linked Products – A User’s Guide January 2000.0018 1.0002 1.14% -0.0000 0. .0019 1.0020 NOV 1. such as 5y or 3y averages.0024 1.06% 0. over the period between 2000 and 2009. the semi-durable goods index and the index of goods excluding fresh food.0006 1.8% 2000-Base core CPI 2005-Base core CPI May Mar Aug Dec DEC 1.0014 258 0. CPI excluding imputed rent and fresh food. The ratio highlights stable seasonal patterns.9971 0.0028 1.47% -0. Barclays Capital Note: 2000-base CPI data covers period from Jan 2000 until Dec 2004.0000 1.0% 0.994 February is usually the mo nth with mo st negative seaso nality .0022 1.10% -0. The average m/m changes in the seasonal factors using both 2000-base CPI and 2005-base CPI data are shown in Figure 255.9974 SEP 1.992 2000 2002 2004 2006 2008 2010 Source: Ministry of Internal Affairs and Communications Statistics Office.0029 1.34% 0.06% .9942 0.0030 JUL 0.0021 MAY 1.9945 0.0010 1. the goods index.06% 0. making onemonth JGBi carry from 10 March to 10 April most negative (under the 2005 base it declined 18 Although the Japanese core CPI aggregates prices of more than 550 items. Source: MIAC.0010 1.9970 0. the MIAC publishes time series adjusted for seasonality only for the headline and core CPI along with six other broader sub-indices: CPI excluding imputed rent. 3) Implications for JGBi carry While the ratio of the NSA core CPI to the SA Index as a time series portrays a reliable image of seasonal patterns.9998 0.0010 1.16% 0. *1990-1995 0.0003 1.9947 0.04% Octo ber is usually the mo nth with mo st po sitive seaso nality .20% 0. Barclays Capital. CPI excluding food and energy (known as the “core core” CPI).08% -0. from a practical perspective the m/m growth rate in the seasonal factors provides a better picture of the impact of CPI seasonality on carry. and 2005base CPI data is for Jan 2005 – Nov 2009.4% -0.0010 OCT 1.9970 APR 1. the graph also reveals no major change in seasonality due to the rebasing of the core CPI from the year 2000 to the year 2005.10% -0. The largest m/m drop occurs in the month of January. 15 March 2010 Nov Jun Sep Feb Apr Oct Jul -0.28% 0. Figure 254: Time series of ratio between NSA core CPI & SA core CPI 1.0018 1.30% 0.00% 0. . without the risk of overstating or understating the impact from core CPI seasonality. . 0.0012 1.10% 0. Figure 255: Average m/m change in monthly seasonal factors under 2000-base and 2005-base core CPI 0.2% 0.004 1.0026 1. .0021 Note: 2000-base and 2005-base core CPI data sample from Jan 2000 to Nov 2009. with February being the month with most negative cumulative seasonality and October/November featuring the most positive seasonality.996 0.

but historically it has never been adequately priced. as the m/m change in seasonal factors is most positive in April (+0. The reason behind this discount is naturally the fact that the cash flows of weak-seasonality linkers. whose base CPI is taken from March. are adjusted for inflation based on core CPI index levels taken from March and September. In the more recently launched inflation-linked markets. Apart from the short-term inflation carry.47% seasonal implies negative carry of over 9bp and a positive seasonal of 0. Most issues bonds redeem with a seasonally favourable point. 8 and 12 have been trading on average between 2bp and 4bp cheaper versus our parametric real yield curve model. a -0. in accordance with the three-month indexation lag.14%. We estimate the difference in the seasonal between these two months to be worth 1. For instance. The basis point impact on 1mth carry will depend on the residual duration of JGBis. the final principal repayment are adjusted for inflation based on March CPI numbers (an average seasonal factor of 0. The Latin American country with the most extreme seasonality is also the one with the lowest and most stable inflation trends. For instance. this also demonstrates very limited seasonality. becoming larger the shorter the duration of the bond. the largest positive inflation carry due to seasonality occurs during the 10 July-10 August period. For a 5y duration bond. have the poorest seasonality among JGBis’ base CPI months. In general. with an X-12 estimation of seasonality across the year varying by less than 0. Chile.9% better seasonal than a bond maturing in April. the other half and much more importantly. Despite measurement changes in Argentinean CPI. JGBi4. However. In contrast. Here the seasonal trend is for prices to rise from February through September before seasonals turn negative for each of the following five months. issues JGBi2. The difference in the seasonality of each JGBis’ base CPI is generally consistently discounted by the market. hence based on September CPI. the 2012 issue and the now small old 2014 bond redeem on 1 April and thus capture the worst seasonal of the year for January.30% implies a 6bp positive carry. while the seasonality in Mexico and Colombia is of a similar order of magnitude to that in the US despite having inflation that is notably more volatile.47% m/m). June bonds (issues JGBi2. seasonality also has an impact on the relative pricing of individual JGBis. while other issues have generally been trading fair to the curve. In general seasonality is less important in emerging markets due to more unstable inflation and less regular distortions. seasonality has been a factor encouraging investment during positive carry periods but has not been that important relative to the underlying volatility of the inflation indices. maturing on1 December. In contrast.9970). in South Africa the dispersion of seasonality is the same order of magnitude as in the US despite inflation being 15 March 2010 259 . JGBi8.30% on average under the 2005-base). While half of each June bond’s coupons are paid with a relatively favourable seasonal.4%. In particular in Brazil we see less statistically significant seasonality in IPCA inflation than in any other major economy.Barclays Capital | Global Inflation-Linked Products – A User’s Guide 0. Hence a UF bond redeeming in November would have around 1. 4. including the principal at redemption. which consistently sees a strong increase due mainly to clothing price increases that are larger than in most other countries. inflation-linked JGBs whose base CPI is taken from a month with poor seasonality tend to trade at a discount relative to neighbouring bonds. such that since 2003 prices have on average fallen in this negative seasonal period even with inflation averaging almost 4%. JGBi12 and JGBi16). linkers issued in December (and thus featuring base CPI from September) enjoy the strongest seasonal factors. Other inflation markets Sweden is a market where seasonality is relatively extreme for inflation-linked bonds. whether in terms of inflation-adjusted prices/ yields or breakevens.

4% 0.6% 0. August maturing bonds ought to have the most favourable seasonality.2% 0. whereas if a bond were issued with a maturity at the start of December its seasonality would be up to 2% worse.0% -0.5% -1.5% 0. There is also a higher degree of uncertainty of seasonality due to the significant change in inflation measurement that occurred from the start of 2008.4% -0. Barclays Capital Source: Bloomberg. Figure 257: Estimated Latin American m/m seasonality 0. although February is almost as positive. Nonetheless.5% Turkey Poland Israel South Africa Sweden 1.0% -0. Prices tend to fall in summer as elsewhere in Europe such as Poland.6% -0.2% -0.8% Jan Mar May Jul Sep Nov Argentina Colombia Brazil Mexico Chile Figure 258: Estimated m/m seasonality selected EMEA 1. December maturing bonds still have a better seasonality than those redeeming at the end of March by more than 1%.0% Jan Mar May Jul Sep Nov Source: Bloomberg. Turkish seasonality is relatively strong.0% 0. which we attempt to minimise by considering seasonality in CPI excluding housing. but more dramatically due to the particularly heavy weight of food in the Turkish CPI basket and with a sharp rebound in October.8% 0. though not especially stable. Barclays Capital 15 March 2010 260 . With Turkish linkers being short dated the sensitivity to this seasonality is relatively high.Barclays Capital | Global Inflation-Linked Products – A User’s Guide around four times as volatile.

35% 2. in the same way that a DV01-weighted curve trade in nominal bond space would be constructed.pond@barcap.5% 1. A breakeven rate is the difference between the real yield of an inflation-linked bond and the nominal yield of its nominal comparator. Investors largely focus on spot real yields and breakevens or near-term carry-adjusted ones. that the DV01 of a linker is a “real” DV01.com Michael Pond +1 (212) 412 5051 michael.sooben@ barcap. We highlight.85% 3. Figure 261shows the cash flow and weighting details of a long breakeven trade on the TIIJul14. however.35% Figure 260: Forward TIPS breakeven curve 4. Spot breakeven rate = nominal yield – real yield Essentially. Figure 259: Our conceptual measure tracks spline-based 5y5y 4.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Forwards and their usage Khrishnamoorthy Sooben +44 (0) 20 777 37514 khrishnamoorthy. Forwards are best suited to relative value strategies but are also relevant for fundamentallydriven ones. A pure breakeven trade is a DV01-weighted one.0% 1. given the emphasis of central bankers on market-based measures of forward inflation expectations. The ratio of notionals of the linker and nominal legs will be the inverse of their DV01 ratio. but the importance of forwards between maturities on the curve is significant in the analytical framework for developed inflation markets.0% 3.5% 2. ie. ie. In this article. this represents the market’s expectation of inflation over a given period. less a liquidity premium differential (for a detailed discussion of these concepts.5% 3. we explain how forward breakeven trades can be built in both the cash and derivatives market.85% 3.5% Jul-03 Jul-04 Jul-05 Jul-06 Jul-07 Jul-08 Jul-09 5y5y Spline Zero-Breakeven (Nominal-Real.0% 1.mirani@barcap. It is a well known fact that central banks look at market-implied measures of inflation expectations.com Chirag Mirani +1 212 412 6819 chirag.35% 1. While trading the forward curve in inflation markets may not be widespread. We first look at the construction of a forward breakeven trade in bond space and use TIPS in our illustration. please see Breakeven inflation and the relationship between real and nominal yields earlier in this publication). breakevens.com Forward trading is not widespread in inflation markets. For the sake of clarity. expresses the change in the linker’s full price following a 1bp move in its real yield. LHS) ModDur Weighted 5y5y BEI (RHS) Source: Barclays Capital 15 March 2010 261 .0% 3. we first define what a breakeven is and how a spot breakeven trade is constructed. with the 5y5y forward measure explicitly mentioned by central bankers in the US and euro area. San Francisco Fed President Janet Yellen’s reference to forward inflation in 2004 initially contributed to highlight the concept of forwards and a 5y5y forward TIPS breakeven series is even available from the Fed.5% 2.0% 0.0% Jan-10 Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Forward 1yr breakevens Source: Barclays Capital 1. plus an inflation risk premium.0% 2.0% 2.5% 2. it is not uncommon either.5% 0. although forward analysis is extensively used within the analytical toolkit available to market participants.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Figure 261: Details of a long breakeven trade on TIIJul14
Bond Type TIPS (Buy) NOMINAL (Sell) Issue TII 2Jul 14 T 2.625Jul 14 Price (decimal) 122.72 102.59 Breakeven
Note: As of 14 January 2010. Source: Barclays Capital

Yield 0.43% 2.30% 1.87%

DV01 5.31 4.33

Index ratio 1.14721

Par weights 100.0 122.7 Cash residual

Cash value ($) (122.7) 125.9 3.1

There will be some residual starting cash flows (due to dirty price mismatch), as well as coupon accrual differential depending on the path of inflation and reinvestments. The return attribution to this residual when compared with realised – expected breakeven returns will be small. A breakeven position can really be thought of as a synthetic par breakeven bond. The risk measure of such a bond would be: Breakeven_DV01 = Breakeven_ModDur * Index_ratio A long forward breakeven rate position is essentially a position on a forward starting breakeven contract at the present time. The P&L of such a trade will be realized as the market’s expectation of this forward breakeven rate changes. Here, we go through the steps to create such a forward. Forward breakeven contract = ForwardBEIRate + residuals (bp) Residuals (bp) is the offsetting residual costs of two simultaneous breakeven positions (two needed to construct a forward breakeven rate) described above. These residuals exist because we are approximating forwards using cash coupon bonds, whereas bootstrapping spot rates to create forwards is only accurate using zeros. An approximate ForwardBEIRate can be derived in the same way as forward nominal rates using zeros. Here, we assume zero breakeven rates (BEI) to be spot breakeven rates of equivalent maturity bonds. We further assume constant rates and continuous compounding to arrive at a ForwardBEIRate. EXP (ForwardBEIRate * (T2 – T1) )=

EXP( BEI 2 * T2 ) = EXP ( T2 * BEI 2 − T1 * BEI1 ) EXP( BEI1 * T1 )

ForwardBEIRate =

(T2 * BEI 2 − T1 * BEI1 ) , T2>T1 (T2 − T1 )

For zero-coupon bonds, the modified duration (D) is roughly equal to its time to maturity. So the time component in the above formula can be replaced with modified duration (breakeven modified duration). We use TIPS modified duration for approximation purposes. ForwardBEIRate =

( D2 * BEI 2 − D1 * BEI1 ) ( D1 ) , W = ( D2 ) , W1 = 2 ( D2 − D1 ) ( D2 − D1 ) ( D2 − D1 )

forwardBEIRate = W2*BEI2 - W1*BEI1,

15 March 2010

262

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Figure 262: Long 5y5y cash breakeven trade
Bond Type TIPS Nominal Price (decimal) 122.72 102.59 Breakeven TIPS Nominal TII 1.875 Jul 19 T 3.625 Aug 19 106.84 100.28 Breakeven Fwd BEI
Source: Barclays Capital

Issue TII 2 Jul 14 T 2.625 Jul 14

Yield 0.43% 2.30% 1.87% 1.34% 3.74% 2.40% 2.92%

DV0 1 5.31 4.33

Index ratio 1.1472 1

Par weights 100.0 122.7 Residual

Cash value ($) 122.7 125.9 3.1 121.0 132.0 (11.0) (7.9)

TIPS Modified duration, (BEI_DV01) 4.32, (4.96) 4.22

9.24 7.95

1.0127 8

113.3 131.6 Residual Net (Residual)

8.71, (8.82) 7.90

For a 5y5y fwd breakeven rate, with respective modified durations for 5y and 10y of 4.33 and 8.73, the above weights (W1 and W2,) would have roughly a 2-to-1 ratio, meaning the Breakeven_DV01 of the longer security has to be twice that of the shorter security. This forces the inflation-adjusted notionals of the two TIPS securities involved in forward breakevens to be the same. This is an important step for a forward position because equal TIPS notionals force the accrual inflation gains/losses prior to the forward start date to be zero. For a forward breakeven position, we want inflation exposure to be defined from the forward start date and not prior. Inflation Sensitivity = Jul19Notional * Jul19_IdxRatio - Jul14Notional * Jul14_IdxRatio = 113.3*1.01278 – 100*1.14721 = 0, In other words, till the maturity of TIIJul14, the position will not incur any P&L due to net notional inflation accretion. In the example below, we show how this translates to notionals on a 5y5y fwd breakeven trade. In Figure 263, we look at how this position performs in various breakeven curve shifts and changes in forward rate. Given the index ratio-weighted positions ($100mn to $113.3mn) on each breakeven leg with modified durations of 4.32 and 8.71, if breakevens across the curve rise in parallel by 10bp (fwd breakeven will rise by 10bp), the long position in 10y will rise in value by $1mn, and the short position in 5y will lose $0.5mn, netting $0.5mn of positive P&L. If the 5y breakeven falls by 10bp while the 10y breakeven is unchanged, the forward breakeven will also rise by 10bp (breakeven curve steepens), and the 5y leg incurs positive P&L of $0.50mn. This P&L matches the net P&L of a 10bp parallel shift in forward breakeven position. We also show the declining forwards (-10bp) scenario through parallel curve shifts, as well as curve flattening. In both 10bp fwd breakeven narrowing scenarios, the trade loses $0.50mn. Therefore, we feel confident that the above-described weighting method approximates forward rate exposure, even though we are using cash bonds rather than zeros.

15 March 2010

263

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

Figure 263: Index ratio-weighted breakeven position gives inflation exposure only
$100mn short TII Jul14 position (BEI_DV01: 4.96) At entry + 10bp parallel BEI shift 5y BEI falls 10yr unchanged - 10bp parallel BEI Shift 10y falls by 5bp, 5y BEIs unchanged
Source: Barclays Capital

P/L ($mn)

$113.3mn short TII Jul19 position (BEI_DV01: 8.82) 2.40%

P/L ($mn)

Fwd BEI rate 2.92%

Overall P/L ($mn)

1.87% 1.97% 1.77% 1.77% 1.87% (0.50) 0.50 0.50 -

2.50% 2.40% 2.30% 2.35%

1.00 0.00 -1.00 (0.50)

3.02% 3.02% 2.82% 2.82%

0.50 0.50 (0.50) (0.50)

Creating a specified DV01 trade using the above approach
We would use the following approach to find relative notionals to create a specific DV01 forward breakeven rate trade. July14 has DV01 of 49,600 per $100mn July19 has DV01 of 88,200 per $100mn July14 index ratio is 1.14721 July19 index ratio is 1.01278 Relative index ratio (July19/July14) is 0.8828
− − − −

50k DV01 trade = 88,200 x July19 Notional – 0.8828 x July19Notional x 49600 Jul19Notional = $112.58mn Jul14Notional = 0.8828 x Jul19Notional = $99.38mn Check that the relative index ratio weight is maintained = $99.38mn/$112.58mn = 0.8828. So that the inflation sensitivity of this trade will be zero until the maturity of the first bond. Check that the DV01 of the trade adds up to $50k DV01 50k = 88,200 x 1.1258 – 49600 x 0.9938

− −

Swaps are a more natural instrument to look at forward inflation pricing than bonds. This is because the commonly traded structure in inflation swaps is a zero coupon one, such that a ‘clean’ forward rate can be bootstrapped from the curve without any distortion from coupon payments. The generic formula to calculate an inflation swap rate with a tenor S, at a forward date F is:

S

(1 + Y )F +S (1 + X )F

−1

Where: Y is the zero-coupon rate for a spot starting swap of tenor F+S X is the zero-coupon rate for a spot starting swap of tenor F Alternatively, expressed in terms of CPI Index values, the above formula can be rewritten as:
15 March 2010 264

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

S

E (CPI F + S ) −1 E (CPI F )

Where: E(CPIF+S) is the expected value of the CPI Index in F+S years E(CPIF) is the expected value of the CPI Index in F years However, the above formula is not equivalent to the strict theoretical computation for a forward rate. If expressed in CPI values, the forward rate is theoretically expressed as:

S

⎛ CPI F + S E⎜ ⎜ CPI F ⎝

⎞ ⎟ −1 ⎟ ⎠

In other words, the theoretical computation of a forward inflation swap rate is related to the expectation of the ratio of future CPI values. The generic formula, on the other hand, computes a ratio of expectations. Mathematically, they are not equivalent given that the future CPI values are not independent variables. A convexity adjustment therefore needs to be applied to the ‘naïve’ forwards calculated from the generic formula. The need for this adjustment is evident when pricing y/y structures that are effectively portfolios of instruments priced from forward starting swaps. Caps and floors on inflation are such instruments. Inflation market models such as those by Belgrade, Benhamou and Koehler provide a convexity adjustment formula when computing the forward value of the CPI ratio. For an intuitive grasp of the notion of convexity in forward inflation swaps, we look at the dynamic hedging of a forward swap position. We assume a 5y5y forward swap trade in which an end user goes long inflation, ie, pays the compounded quoted fixed rate and receives accrued inflation at maturity. The other counterparty (typically a bank) will hedge the 5y5y forward position through a combination of 5y and 10y zero coupon inflation swaps at the inception of the trade; the hedge will be a long position in 10y and a short position in 5y swaps. By definition, a forward starting swap should have no accretion from realised inflation before the forward date. Therefore, the notional on the 5y and 10y swaps should be equal to immunise the inflation accretion before the forward date. While this may seem straightforward, the delicate element is setting the initial notional on the swaps. Assuming a €100mn notional on the 5y5y trade, and assuming positive inflation over the five years preceding the forward start date of the swap, then setting the notionals on the 5y and 10y at €100mn is likely to result in an overexposure to the 5y5y forward rate. After five years, the initial 5y swap would have expired but the initial 10y one will effectively be a swap with a remaining term of five years and its notional will have accrued with actual inflation and risen above the €100mn that needs to be hedged. Hence, if positive inflation is expected over the first five years, the notional on the swaps in the hedge portfolio should be set at less than €100mn. The €100mn notional of the forward swap therefore needs to be discounted with inflation that is expected over the initial 5y forward horizon. In practice, the 5y zero coupon inflation rate at inception serves as the best guess for this expected inflation. Yet, unless actual inflation proves to be equal to the expected inflation, the notional will still need to be rebalanced if inflation overshoots or undershoots initial expectations. This rebalancing mechanism illustrates the convexity effect at work. Let us assume that after inception of the hedging strategy, inflation is significantly higher than initially discounted. This would point to a potential over-hedge at the end of the forward horizon. The hedger will therefore need to reduce the notional on the 5y and 10y swaps by effectively unwinding part of the hedge. The likelihood is that the unwind will be done at a profit, excluding transaction costs, given that breakevens tend to rise with higher15 March 2010 265

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

than-expected inflation. Assuming a higher breakeven curve, a profit will be made on the partial unwind of the 10y swap, but a loss will be made on the 5y leg. However, given the higher inflation PV01 of the 10y leg, the net effect should be a profit, unless the curve flattens significantly and/or the discount factor on the longer swap is significantly lower. On the other hand, if actual inflation undershoots initial expectations, then the notional on the hedge will need to be increased. In this case of lower-than-expected inflation, the breakeven curve will tend to move lower, which implies that the increase in the hedge notional is likely to be done at cheaper levels than initially. In this case too, this is positive for the hedger. Hence, there is a positive convexity effect in hedging a short position in forward inflation swaps. This effect comes from the likely positive correlation between actual inflation and expected inflation, ie, breakevens. If this convexity is passed on to the buyer of forward inflation, then the quoted forward rate should be lower than what is implied by the ‘naïve’ forward. Obviously, this convexity effect will work the other way round for a seller of forward inflation. It is, however, difficult to re-adjust a hedge portfolio frequently to take into account the convexity effect, given that the magnitude of the latter tends to be relatively small compared to bid/offer spreads on swaps. Typically, because of transaction costs, rebalancings due to convexity are rarely done more than once a year. Although the hedging of a forward swap position may not be straightforward conceptually, taking positions is relatively easy in the developed swaps markets. Forward positions are directional ones, with an added exposure to the curve, as in nominal rates space. This is obvious from the generic forward formula above. Other things being constant, an inflation swap with a tenor of S years, starting in F years has a positive sensitivity to the spot swap rate with a S+F year tenor. This is the directional element. Although the sensitivity to the curve is less straightforward formulaically, a steepening/flattening curve move will tend to increase/decrease the forward rate. Hence, the best configuration to enter a long/short forward trade is when bullish/bearish expectations are blended with a steepening/flattening view. However, given that the slope of the curve tends to be inversely related to the level of breakevens, the directional element has historically been the strongest driver of forward swaps. Forward strategies therefore seem better suited to fade unjustified relative value distortions on the curve, rather than targeting curve slope moves.

Figure 264: Forward curve singles out micro distortions
450 400 350 300 250 200 150 100 50 0 0
Source: Barclays Capital

Structure of 1Y forwards (bp)

euro HICPx

US CPI

UK RPI

5

10

15

20

25

30

15 March 2010

266

Barclays Capital | Global Inflation-Linked Products – A User’s Guide

The analysis of inflation swaps can be broken down in 1y forwards, an insight that is not possible in bond markets given the limited number of maturity points. Micro distortions on the swap curve and cheap segments are therefore more easily identified: for example, Figure 264 shows that 1y forwards between the 10y and 15y on the US CPI swaps curve tend to be low. While such granularity can be useful to fine-tune relative value strategies, some distortions may sometimes be justified. For instance, the relative cheapness of the segment between the 10y and 15y points on the US CPI swaps curve is largely explained by illiquidity in that part of the curve, which itself stems from the lack of corresponding TIPS in this gap. In the euro and UK markets, where asset swapping is an important feature, heavy asset swapping on a specific issue could explain the cheapness of the corresponding maturity on the swaps curve. The most obvious use of inflation swap (and bond) forwards is within fundamentally-driven strategies. Inflation is a tangible economic phenomenon, on which most central banks around the world have more or less official target levels or ranges. Within developed economies, the UK and the euro area have explicit targets/ranges on the desired inflation levels, which imply that expectations of inflation beyond the short term should remain anchored unless the central bank’s policy is not credible. The deflationary period amid the crisis has some interesting elements with regards to this; 5y5y forwards have certainly been volatile during this period but significantly less than spot 10y rates. Inflation forwards can indeed be related to a policy-related “fair value”, which reduces the probability for extreme valuations. Strategies to fade sharp divergences from the perceived fair economic value of forward inflation swaps are relatively common. Towards the end of 2009/start of 2010, 5y5y forwards stood at relatively elevated levels, probably pricing in the probability that beyond the short term, extreme simulative economic policies implemented during the crisis will generate inflation above desired levels. This can be seen as a justified risk premium, in which case fading elevated levels may not have a good rationale.

Figure 265: Forward inflation rates less volatile than spot rates
3.0 10y euro HICPx swaps 2.7 5y5y euro HICP swaps 2.4 2.1

1.8

1.5 Jun 05

Mar 06

Dec 06

Sep 07

Jun 08

Mar 09

Dec 09

Source: Barclays Capital

15 March 2010

267

It is true that the inflation and deflation scare of 2008-09 failed to kick-start a selfsustaining market in inflation options. estimates 15 March 2010 268 . in LPI swaps) finds support from the regulatory framework for the pension fund industry and even here activity has been discouraged by the realised volatility. and inflation rates. while the volatility of the inflation rate is derived from the CPI time series. Volatility parameters for the nominal and real forward rates are computed from historical data on TIPS and nominal US Treasuries. The Jarrow-Yildrim model provides arbitrage-free conditions between the three components. and CPI is the exchange rate between the nominal and real yields. The model consists of a three-factor framework in which the three components are the nominal.com Michael Pond +1 (212) 412 5051 michael. the increased issuance in structured notes and concerns about the tail risks of extreme inflation or deflation led to a surge in inter-dealer activity. The nominal and real interest rates are assumed to follow an HJM diffusion process. an analogy with the foreign exchange market seems to be an obvious starting point when in inflation models: the nominal yield corresponds to the domestic currency.pond@barcap. As inflation started to normalise within the economic recovery and structured note activity dried down thereafter. In their paper. Overview of pricing models and practices Following the Fisher equation. the real yield is analogous to the foreign currency. but the valuation of floors embedded in bonds is very relevant after the deflation scare of 2009. the authors obtain the nominal and real rate term structures by applying standard stripping techniques to nominal US Treasuries and TIPS.sooben@ barcap. rather than simply on a calculation of realised real yield or breakeven volatility which would entail much more subjective elements. while in Europe. The CPI is driven by an instantaneous inflation rate defined as the difference between the nominal and the real interest rates. it is now much easier to obtain implied volatility data on which pricings can be done. The increased volatility in realised inflation during 2008 pushed non linear products to the forefront of the inflation market. The obvious drawback is that historical volatilities are just one component of implied volatilities. This analogy is the starting point of the model developed by Jarrow and Yildrim. More readily observable implied volatilities in the caps and floors market are valuable inputs in other inflation volatility products. Inflation vol surfaces are now relatively well defined though across the most developed markets. Although a market for y/y caps and floors existed in Europe and the US prior to that. Finally. and sometimes very poor. The general approach before consisted mainly in looking at historical volatilities as a starting point to estimate implied volatilities. trading in caps and floors also fell. The UK remains the only market where non-linear activity (ie. activity remains episodic. For instance. Trading in caps and floors in Europe was boosted in 2008 on the back of structured note issuance. liquidity in the US and euro inflation volatility markets. real. can help create a self-sustaining market. in the long term. we highlight that despite limited. This should tend to encourage more consistent pricing which.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Inflation volatility and deflation floors Khrishnamoorthy Sooben +44 (0) 20 777 37514 khrishnamoorthy. which helped define the implied volatility surface. the pricing of TIPStions and options on breakevens can then be based partially on a market-related volatility input. episodic activity mainly linked to the hedging of structured note issuance continued to prevail. a risk premia based on flow/hedging consideration further needs to be added. However.com Liquidity in inflation vol markets remains limited. but in the absence of the latter. with the US seeing scarce activity.

Traded optional instruments are used to define the nominal volatility structure. Hence. The simulation of nominal rates can be implemented through a standard HJM framework. hence inflation volatility is supported and goes up. in the particular case where the CPI process is linked to an instantaneous inflation rate. where the more natural curve for data calibration would be the inflation swap curve. assuming the normal-like distribution for inflation rates). an HJM-type diffusion is assumed. Benhamou and Koehler. long-dated caps to hedge their liabilities. whether they have a zero coupon or a y/y format. positive correlations make sense for longer-dated options (10yrs and out). On the other hand. other models which fall under the ‘market models’ category have emerged as more suitable candidates. there is more demand for low-strike. The calibration to market data is done using money market and swap prices for the nominal zero coupon term structure. This is especially problematic for the euro area inflation options market. pension funds should be natural buyers of high-strike. Sophisticated models are needed to cater for the complexity in volatility products. To this end. to value the product a simulation of annual inflation rates is needed up to the maturity of the swap. Also. real. the availability of the CPI forward is considered to be sufficient such that real rates are not used as an input. Advantages of this model are its simplicity and intuitiveness and the fact that its framework is easy to implement. Inflation volatilities implied from option prices are then calibrated in the SABR model. Typically. 15 March 2010 269 . Forward values of the CPI are modelled. a common but complex (in terms of pricing) product is the LPI swap. As inflation can become deflation and drop below the 0% level (as happened in several countries in 2009). When inflation goes up. For example. An LPI swap is an interest rate product dependent on the path of inflation. They link the zero-coupon and the y/y inflation derivatives in the European inflation swap and options markets through a market model. Unlike in the Jarrow-Yildrim framework. However. As for the nominal curve. and nominal interest rate data. the most liquid non-linear inflation instruments are caps and floors. the forward zero coupon bond and the correlations between them. therefore the SABR parameter beta is set close to zero (ie. The authors provide a convexity adjustment formula when computing the forward value of the CPI ratio. in which the correlation (an inflation rate versus its volatility) and the vol-of-vol (volatility of inflation volatility) parameters determine the “skewness” (asymmetry) and the “smileness” (curvature) of vol smiles. The SABR (stochastic volatility) model seems to be a natural choice as a calibration tool for inflation smiles. In the euro inflation market. the correlation parameter tends to be negative up to the 10y expiry. They show that the forward value of the ratio is the respective forward CPI multiplied by an adjustment that is an explicit function of the forward CPIs. Traders build the inflation volatility term structure (vol as a function of expiry) for individual ATM caplets/floorlets (not directly quoted in the market) so that they can match ATM cap/floor straddles (quoted in the market). The simulation process needs to be carried out within an inflation model.Barclays Capital | Global Inflation-Linked Products – A User’s Guide of the correlation parameters between the three components are calculated through sample moments using historical inflation. it provides closed form solutions for inflation swaps and BlackScholes formulas to evaluate inflation options. shorter-dated floors as a protection/hedge. The authors give closed formulas for the valuation of breakeven swaptions and numerical integration for options on real yields. its main drawback is that it is particularly suited to markets in which calibration needs to be done from data on the bond market. One such approach is the one developed by Belgrade. When inflation goes down.

the TIIApr13 Index ratio.99858. including the US. A fall in the bond’s breakeven brings the floor closer to the money and its theoretical value can increase substantially. at one point the real yield spread between TIIApr13s and TIIJul13s was 200bp. have an embedded par floor such that at maturity. 3) The level of breakevens – This determines the extent to which the floor is considered to be in the money. It is important to remember that the “strike” on the floor is at par. This means that for short maturity bonds. only those with little inflation accretion (likely those that have recently been issued) will tend to have floors with significant value. This element is likely the most important one. The effect on the coupon can vary somewhat across issues with different coupons in significant deflation.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Deflation floors In the previous Inflation-Linked Users’ Guides that Barclays Capital has published.02332. we explain the details of the floor and discuss its effect on potential returns to maturity. As an example. The pay-off on the principal amount at maturity can be written as: Max{Par . for a trade that settles on 25 January 2010. there has been usually only a passing mention of the embedded deflation floor. the closer the strike is to zero.2 years left. However. For example. the index ratio is floored at 1 as it applied to the principal. The floor would kick in if the index ratio fell below 1. With about 3. Since the inflation-adjusted principal is the par amount times the index ratio (which is the ratio of the reference CPI to the base CPI). is 1. Par × IndexRatio} The floor applies only to the principal amount at maturity – coupon payments can be and have been paid off an inflation-adjusted principal amount less than par. This is primarily because the value of the floor was seen as miniscule and only very rarely were there pactical questions. during the financial crises when US breakevens out to the 9y turned negative and investors were increasingly risk averse. Four elements determine the value of the floor on the bond: 1) The inflation that has already accrued on the bond – The higher the current index ratio. the investor gets the greater of par or the inflation-adjusted principal. there needs to be 0. In the following. or an index ratio of 1. euro area and newer issues in Sweden. the first coupon payment on the TIIJul16s was paid off an index ratio of 0. The fact that there is no floor on the coupon payments complicates the calculation profile of TIPS somewhat. the lower is the strike on the embedded floor. the floor value of newer bonds tends to be more valuable because the index ratio is typically lower than seasoned TIPS. at maturity. so the inflation accrued since issuance would first need to be fully reversed out in a period of deflation. 2) The remaining time to maturity of the bond – The higher the maturity. This means that there has been 2. The remaining time to maturity is a known parameter and inflation accrual changes 15 March 2010 270 . The non-linear inflation market began quoting cumulative caps and floors by early 2009 and the value of the embedded option could then be priced separately from TIPS. this is another way of saying that. but the deflation effect on coupons becomes significant only in severe deflation environments given their relatively small size compared to the principal. For this reason. not where the index ratio is at the time of purchase. with most of this difference explained by the floor. Several inflation-linked markets.33% cumulative inflation accrued since issuance in April 2008.71% annualised deflation to maturity for the floor to be at the money at maturity. the value increased significantly: for example.

however. especially compared with realised inflation volatility. 4) The market’s perception of implied volatility – This is also an important parameter. breakevens can be subject to big changes over a relatively short period. the implied volatility is important because it plays into options time value. provides an indication about the value that the market gives to embedded floors in linkers. Figure 266: Realised CPI swap volatily is generally much higher than realised CPI NSA Vol 50 1y CPI SWAP Realized CPI NSA Rlzd 40 bp/day 30 20 10 Feb 04 Source: Barclays Capital Feb 05 Feb 06 Feb 07 Feb 08 Feb 09 In practice. as quoted in the market.Barclays Capital | Global Inflation-Linked Products – A User’s Guide relatively slowly. hence. if the asset swap is calculated in the usual way by taking the quoted price and without adjusting for the value of the floor. Many dealers were left short inflation volatility in Q4 08. exposed to inflation market volatility more than realised inflation (Figure 266). On the other hand. and the market became temporarily dislocated. This can also be a reflection of the market’s appetite for risk. especially in an environment in which implied inflation volatility is at significantly rich levels. This is because the price of a linker. but the limited depth of the inflation options market means that such an exercise is not straightforward. then the resulting asset swap will tend to be richer than the asset swap quoted in the market. The market-quoted asset swap level will be based on the quoted linker price but will also take into account the floor value that may. should be concerned much more with the volatility of actual inflation. gauging the fair value of floors on linkers is not trivial. Figure 268 shows the decline in estimated floor premium levels with the rise in breakeven levels. Specifically. In the US for instance. One may have noticed that implied volatilities priced by inflation vol market makers tend to be much higher than realised NSA CPI Index Vol. On the other hand. however. An option-pricing framework would be the natural starting point for valuation purposes. For this reason this relative richness is an indication of the value of the embedded floor (Figure 267). We believe the reason behind the high implied volatility levels. Inflation vol Investors who can hold the position to maturity. be neutral. poor liquidity in inflation volatility products means that implied volatility levels are not easily observable. at least theoretically. 15 March 2010 271 . even as the rest of the inflation markets have largely normalised. be considered as an up-front option premium. The effect of the floor on the quoted asset swap valuation should therefore. The asset swap market. will incorporate the market’s valuation of the floor. is that traders need to hedge their positions with inflation swaps and are. for example.

most notes. breakevens surged on higher inflation and heavy structured note issuance. and sold primarily to retail investors. the issuance of such notes would have required offsetting long positions in inflation floors. and the leverage amplified the implicit notional on embedded floors. Although implied inflation vols increased substantially. During H1 08.625% Apr 13 Traded ASWs Figure 268: Estimated floor premium declines as BEIs rise 180 160 140 120 100 80 60 40 20 0 -20 May 08 3 2. in 2-5y maturities. However. The impact on short floor positions was not heavy until when short-dated swaps fell below 1% in October. the mark-to market value of low-strike floors was then negligible.5 0 -0. However. Dealers often tackle the vega exposure in inflation options using nominal vols as a 15 March 2010 272 . This trend had important implications for zero coupon inflation swaps. There were significant dealer shorts in Euro HICPx floors at that time from past inflation-linked structured note issuance. the end of 2008 saw a powerful impact of the volatility market on the dynamics of zero-coupon swaps and bond breakevens. even those with above zero strikes. with breakevens relatively stable and around 2% prior to 2008. almost all notes issued had some form of embedded inflation floor because coupons cannot be negative. however. have been short-dated during that year. These notes left dealers structurally short inflation floors. pushing floor vols sharply higher. the depth of the inflation options market relative to outstanding embedded exposures is small. Indeed.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 267: Calculated and traded TIPS ASWs difference is roughly floor premium 100 50 0 -50 -100 -150 -200 May 08 TII 0. While long-dated issuance has been common in 2008 to cater for demand for investors like pension funds. products issued in 2006 and 2008 were leveraged. Dealers tried to cover their short positions by buying mostly 5y 0% floors in the broker market. as a perfect hedge. Given that most embedded floors carried a 0% strike. higher strikes were not uncommon. However.5 -1 Aug 08 Nov 08 Feb 09 May 09 Aug 09 Nov 09 Feb 10 Aug 08 Nov 08 Feb 09 May 09 Aug 09 Nov 09 Feb 10 Apr13 Deflation Floor (LHS) Source: Barclays Capital Apr13 Breakeven (RHS) Source: Barclays Capital The 2008 ‘Gamma trap’ In contrast to the US market. the sharp bearish momentum in short-dated swaps increased their negative mark-to-mark valuations as levels drifted towards zero.5 2 1. most embedded floors were deeply out of the money.625% Apr 13 Calculate ASWs TII 0. or short positions in inflation swaps if only delta hedging is considered.5 1 0. From a hedging perspective. negative breakevens over long maturities have not been seen in Europe and floor valuations in European linkers never accounted for a significant proportion of their value. vega and delta risk on floors can only be hedged perfectly simultaneously with other floors. Although the bulk of such positions were likely in 0% strikes. Given that inflation can be negative. The tide however turned as breakevens collapsed from the levels reached in summer that year. In particular.

5 delta hedging of short floor positions 1. the delta exposure can be more easily hedged using swaps. Indeed. given the greater concentration of structured notes there. Given the illiquidity of the inflation swaps market towards the end of 2008 and the fact that most dealers had similarly short exposures in inflation floors from structured notes issuance. given that a typical floor would then be increasingly sensitive to moves in the underlying (ie. the steepness of the breakeven curve and the greater gamma on shorter-dated floors.0 5y euro HICPx swap Figure 270: Extreme richening in implied vols in 2008-09 200 2. the embedded floors also moved closer to the strike. Figure 269: Impact of delta hedging on euro HICPx swaps 3.0 50 5y Inflation YoY euro HICPx Capvol 10y Inflation YoY euro HICPx Capvol 0.5 Jun 05 Mar 06 Dec 06 Sep 07 Source: Barclays Capital Jun 08 Mar 09 Dec 09 0 Jan 08 Jun 08 Nov 08 Apr 09 Sep 09 Feb 10 Source: Barclays Capital 15 March 2010 273 . the swap breakeven rate).5 150 2. A dealer delta hedging a position therefore needed to increase its shorts in inflation swaps as inflation breakevens moved lower. this put further substantial downward pressure on euro HICPx swaps. On the other hand. an effect that had already been seen a few weeks earlier in the US where the initial shock was larger despite the size of US CPI floor exposures being smaller. The need to delta hedge the short exposures in inflation floors created a self-reinforcing bearish momentum in breakevens.Barclays Capital | Global Inflation-Linked Products – A User’s Guide proxy hedge. This was especially true in short maturities.0 100 1. as swaps breakevens moved lower.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide Comparing baskets: The elements driving inflation Nick Verdi +44 (0) 20 7773 2173 nick.com This article examines the key components of inflation across economies. Second. with the price of cereals reaching their highest levels in almost 30 years in 2008. Ecowin. Haver Analytics. perishables UK RPI Sweden CPI Figure 272: Emerging economies’ CPIs. there are important idiosyncratic factors too. Barclays Capital 15 March 2010 274 . There are striking similarities across countries. increased biofuel feedstock demand. resulting in greater competition at a global level. For a number of countries.verdi@barcap. and also related to the increasing importance of the emerging economies. Among these countries. First. Figure 273 shows that the widest dispersion of weights across countries is in the food and housing sub-categories. an important consideration for investors in index-linked markets. however. which states that a smaller share of expenditure is devoted to food and income increases. The Organisation for Economic Co-operation and Development (OECD) and the UN’s Food and Agriculture Organization (FAO). high oil prices and US dollar depreciation. Looking ahead. Thereafter. % y/y 7 6 5 4 3 2 1 0 -1 -2 -3 97 99 01 03 05 07 09 US CPI E13 HICP ex. A useful reference here is Engel’s law. % y/y 45 40 35 30 25 20 15 10 5 0 -5 00 02 04 06 08 Brazil Argentina Mexico Israel South Africa Source: Thomson Datastream. housing costs carry the largest weight in the CPI. particularly developed ones. Price trends in the past decade have followed two discernible patterns. 2007-08 is the most obvious example of rising food prices at an international level. for example. These issues include draughts in key grain-producing regions. moves in food prices have tended to have the largest impact on CPI inflation in emerging economies (see Figure 277 to Figure 289). tobacco Japan CPI ex. Ecowin. it is important to look at the major sub-category weights of each basket. Major sub-categories of inflation In analysing the key drivers of CPI inflation across countries. Barclays Capital Source: Thomson Datastream. have argued that some of the factors behind the 2008-09 food-price spike could sustain higher prices over the next decade. food prices are likely to remain an important consideration for inflation investors. Haver Analytics. a step-up in demand for food and energy has meant that these factors had a more substantive bearing on global consumer-price indices in the latter part of the decade. where attempts have been made to incorporate Figure 271: Developed economies’ CPIs. international food prices declined as the latter part of 2008 saw a fall in oil prices and a reduction of food demand related to the global recession. As a result. goods price deflation was evident in the early part of the 2000s as producers in emerging economies raised their production capacity.

In the UK. notably the US.1 Note: * Refers to core CPI.6 4.6 19. Specifically.7 30.4 12.9 28.7 5.9 3.5 5. also have a major weighting across countries. Barclays Capital Energy costs. Figure 273: Key CPI sub-category weightings across economies CPI weights Apparel US Euro area Japan* UK Sweden Argentina Brazil Chile Colombia Mexico Poland S. These factors have been a source of volatility for consumer-price indices.7 30. In the euro area.7 11.2 22.5 1. The majority of countries have changed their methodologies for housing costs in the last three decades.2 5. where gasoline duties are higher.2 18. whether manifest in gasoline prices or household utility costs. a given 15 March 2010 275 .1 5.9 22.5 2. for example.5 18.4 8.3 6.6 23.2 4. see that housing overall comprises a significant share of their price baskets and has made significant contributions to CPI inflation over the past decade.4 19. Ecowin. with the key issue that a consumer-price basket would view housing as an investment and not a consumption good.2 16.0 32. this factor is accounted for by owners’ equivalent rent of primary residence (OER) sub-component (which has a weight of 24.1 26. In Europe. This can be explained by the fact that the US has the lowest gasoline tax among industrialised countries.4 4.0 4.9 5.6 5.4 4.6 27.2 2.7 15. reversing the trend that prevailed through 2008.3 22. The depreciation element represents expenditure by owner-occupiers necessary to maintain their dwelling at a constant quality.2 13.4 4. There are different approaches to accounting for owner-occupied housing costs.2 Education 3.1 14.6 2.5 7.6 21.5 0. In 2009. HICP series do not include any component for only occupied housing despite Eurostat studying include a factor to add them for at least seven years. most recently by South Africa at the start of 2009. Korea Israel 3. Africa Turkey S.1 6.9 5.2 37.9 3.0 17.7 6. with the OER approach most widely adopted when this has occurred. weight: 4.2 5.1 3.3 15. Countries that have taken into account these costs. Source: Thomson Datastream. this refers to the perceived rental value of property. In the US CPI basket.2 19.4 15. this has been a source of volatility in their overall CPIs.2 5.1 5.1 12.1 7.0 4.9 1. UK and Sweden.6 15.5 19.9 21.6 10. gasoline price inflation was very negative across countries.4%).Barclays Capital | Global Inflation-Linked Products – A User’s Guide owner-occupied housing costs. commensurate with the housing booms witnessed in those economies.7 Transport 15.1%) and “depreciation” (cf. Haver Analytics.4 16.7 18.3 7.4 14. the housing component of the RPI includes mortgage interest payments (MIPS: cf.0 20.7 14.8 13. In its last published assessment of this project in March 2009 it was suggested that a net cost approach be used.1 1. which if implemented would effectively link HICP to new home prices only.8 15.1 13.0 23.2 2.1 13. national statistical agencies.5 11. weight: 5%).8 5.6 17.2 Housing 43.2 5.2 2.4 15. An interesting aspect of these moves was that swings in US gasoline price inflation more closely reflected changes in the oil price.4 Health 3.2 5.2 Food & bev 15.9 16.

the administered part of the CPI basket carries an element of risk that is more political. directly regulated by the government partly regulated by the government quasi-regulated. For the euro area. are not particularly valuable in predicting sub-components of the basket that are not subject to demand-supply forces. in countries where such prices exist. A notable exception. This is in the form of council tax (4. In Latin America. for example. while the HICP as a whole increased only 2. Eurostat estimates that 10. In the euro area and South Africa.5% of the CPI basket in South Africa subject to such controls. for both countries. tobacco and petrol. where governments guarantee a minimum price or cap Applying these definitions yields the weights of administered goods and services. than economic. While this effect is largely absent in Sweden.Barclays Capital | Global Inflation-Linked Products – A User’s Guide percentage change in the oil price translates to a less-than-equivalent percentage move in gasoline prices. In other words. electricity and gas prices are subject to quite different price changes across countries. Israel and the US. which can be influenced or set by governments and their agencies without reference to market forces. public transport and road and air travel. there are differing views as to which sub-categories pertain to such a classification. First. however. for example. Prices not covered by these definitions would include goods and services that are subject to excise duties and indirect taxes – alcohol. the price of a good or service that is determined directly by producers. steps have been taken by the statistical agencies to account for these effects. This disparity can be largely explained by healthcare reform that led to a spike in such costs. when administered prices were raised 3. Across economies. in contrast to most other economies.0% of the RPI) – a system of local taxation that is based on the value of dwellings – that is set by local authorities. housing costs include a notable administered component in the UK. administered prices can be defined as follows. treatment and weight of administered prices differ across countries. excluding this portion of the basket yields inflation series that have broadly similar trends to the ‘headline’ series over time. broadly taking the form of gasoline and public transport fares. Second.7% of the HICP basket is administrated. a significant portion of the administered basket is taken by transport costs. goods that are subject to specific taxes indirectly regulated. governments and their agencies have to ability to apply price changes to these goods and services in an extemporaneous manner. This factor is an important consideration for students of consumer-price indices for three reasons. is in the euro area in 2004. with 14. A key reason for this is the presence of regulated price mechanisms.1%. 15 March 2010 276 . as outlined in Figure 274. Energy prices include gas and electricity prices. That is. CPI modelling techniques related to business cycle dynamics. Administered prices Another issue to consider when comparing the structure of consumer-price baskets across economies is that of price changes that are “administered”. ie. Meanwhile.6%. For the purposes of this article. while transport costs include items such as costs pertaining to owning and running automobiles. ie. Figure 275 and Figure 276 show that. a topic discussed in more detail below. The presence. In terms of household utility costs. energy and transport emerge as the two most heavily administered sub-components of the price basket.

2 3.7 13. Haver Analytics.4 0. Barclays Capital Figure 277: US CPI sub-category contributions.9 Recreation & personal services 0. national statistical agencies.0 2.9 Misc.0 6.5 2.1 3.6 9. 3.9 29. admin.0 0. Barclays Capital 15 March 2010 277 .9 0.4 4.5 Source: Haver Analytics. Barclays Capital Source: Haver Analytics.1 1.6 4. admin.7 11.0 3.2 14.6 0.6 0.0 Energy 2.9 3. y/y) 5 4 3 2 1 0 -1 -2 02 HICP HICP .8 1.4 0.0 0.87 Housing 1.3 17.0 Transport 1.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 274: Proportion of CPI basket that is administered (%) Non-energy industrial goods Euro area Japan UK Brazil Colombia Mexico South Africa 2.1 2.0 0.3 0.0 0.7 3.5 0.administered prices 15 CPI ex.4 15. prices 04 06 Figure 276: South Africa CPI (%.1 4.2 0.administered prices HICP ex.0 0. Ecowin.5 0.0 4.8 4. y/y) 20 CPI CPI . Barclays Capital Figure 275: Euro area HICP (%.0 0.2 0.31 Communication 0. Haver Analytics.0 0.1 1.0 0.5 Sum 10. prices 10 5 0 -5 08 02 04 06 08 Source: Eurostat.5 3.1 4. pp 5 4 3 2 1 0 -1 -2 -3 -4 97 Other Transport Food & beverages 99 01 03 05 07 09 Education Apparel Medical care Housing Source: Thomson Datastream.7 10.0 8.4 0.0 4.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 278: E13 HICP ex. Haver Analytics.0 -3.0 97 99 01 03 05 07 09 Source: Thomson Datastream. goods & services Transport Housing Figure 279: Japan CPI ex. Barclays Capital Other Communication Health Food & beverages Misc.0 -2.0 1. pp 6 5 4 3 2 1 0 -1 -2 -3 -4 97 Food & catering Alcohol & tobacco Housing & household expenditure Personal expenditure Travel & leisure 99 01 03 05 07 09 Source: Thomson Datastream.0 2. Haver Analytics. pp 5 4 3 2 1 0 -1 -2 97 99 01 03 05 07 09 Source: Thomson Datastream. Haver Analytics. tobacco sub-category contributions. light & water charges Food Transport & communication Clothing & footwear Housing Figure 280: UK RPI sub-category contributions. perishables sub-category contributions. Barclays Capital Other Medical care Fuel.0 0. Ecowin. Barclays Capital 15 March 2010 278 .0 -1. Ecowin. Ecowin. pp 3.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 281: Sweden CPI sub-category contributions. Haver Analytics. Haver Analytics. Haver Analytics. pp 6 5 4 3 2 1 0 -1 -2 -3 -4 97 99 01 03 05 07 09 Source: Thomson Datastream. Barclays Capital Other Education & recreation Housing Transport & communication Clothing Food & beverages 15 March 2010 279 . pp 10 8 6 4 2 0 -2 -4 01 02 03 04 05 06 07 08 09 Source: Thomson Datastream. Barclays Capital Other Recreation & culture Transport Food Restaurants & hotels Communication Housing Figure 282: Brazil IPCA sub-category contributions. Ecowin. pp 20 18 16 14 12 10 8 6 4 2 0 -2 01 02 03 04 05 06 07 08 09 Source: Thomson Datastream. Ecowin. Ecowin. Barclays Capital Clothing Communication Domestic goods Food & beverages Housing Transport Other Figure 283: Chile CPI sub-category contributions.

pp 25 20 15 10 5 0 -5 04 05 06 07 08 09 Source: Thomson Datastream. Haver Analytics. Haver Analytics. Haver Analytics. Barclays Capital Food & beverages Clothing Housing Health Transport & communication Education Other Figure 285: Colombia CPI sub-category contributions. pp 16 14 12 10 8 6 4 2 0 -2 01 02 03 04 05 06 07 08 09 Source: Thomson Datastream. Ecowin. Barclays Capital Food & beverages Clothing Housing Health Transport & communication Education Other 15 March 2010 280 .Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 284: Mexico CPI sub-category contributions. Ecowin. pp 10 8 6 4 2 0 -2 01 02 03 04 05 06 07 08 09 Source: Thomson Datastream. Ecowin. Barclays Capital Food Clothing Housing Health Transport & communication Education Other Figure 286: Turkey CPI sub-category contributions.

Ecowin.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 287: South Korea CPI sub-category contributions. Ecowin. pp 8 6 4 2 0 -2 -4 01 02 03 04 05 06 07 08 09 Source: Thomson Datastream. pp 14 12 10 8 6 4 2 0 -2 -4 03 04 05 06 07 08 09 Source: Thomson Datastream. pp 10 8 6 4 2 0 -2 -4 01 02 03 04 05 06 07 08 09 Source: Thomson Datastream. Barclays Capital Food Clothing Housing Health Transport & communication Education Other Figure 289: South Africa CPI sub-category contributions. Haver Analytics. Barclays Capital Food Clothing Housing Health Transport Education Other 15 March 2010 281 . Barclays Capital Food & beverages Clothing & footwear Housing Transport & communication Education Entertainment Other Figure 288: Israel CPI sub-category contributions. Ecowin. Haver Analytics. Haver Analytics.

can still offer some insights into the future behaviour of core inflation. particularly with regard to change in the underlying rate. we seek to show that output gap estimates. while vulnerable to significant retrospective revision. 15 March 2010 282 . In so doing. given the likely continuation of commodity prices and government indirect tax increases outpacing core inflation rates.5 0 -0. The reduced sensitivity of the change in core inflation to the output gap seen in many countries since the early 1980s is associated with a decline in published inflation rates and an associated decline in the variance of inflation expectations. since the revisions can be significant and even of sufficient magnitude to alter the sign.5 -1 -1. Figure 290: The relationship between the change in G3 inflation and output gap estimates 1 0. we can form stronger conclusions about the output gap. when both measures signal a very wide negative output gap.5 82 86 BarCap survey based output gap (4 qtr ma) Underlying inflation change (4 qtr ma of 4 qtr change in annual % rate) Output gap (composite measure 4 qtr ma. % (survey-based output gap) % (composite output gap) 2 1 0 -1 -2 -3 -4 -5 Source: Haver Analytics. and hence contribute to inflation forecasts.com We consider the relationship of output gaps to underlying inflation using data for the euro area. continue to exhibit a significant relationship to inflation. by supplementing conventional (retrospectively revised) output gap estimates with survey-based estimates (which are not revised retrospectively. examining data from the US. which get revised retrospectively. The currently large negative output gaps across all economies imply downward pressure on core inflation rates. That said. Barclays Capital In this article we seek to focus on the relationship between core inflation and the output gap.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Why output gaps still matter for inflation Julian Callow +44 (0) 20 7773 1369 julian. euro area and Japan. RHS) 90 94 98 02 06 10 change pp (inflation). Datastream. this may also reflect a stronger role for inflation expectations in determining price setting behaviour (as can be seen by contrasting the relatively stable relationship of the output gap to underlying inflation in Germany since the early 1970s with that of the US and rest of euro area). and so have better “real time” information). The conventional output gap estimates. along with unrevised survey estimates.callow@barcap. particularly at a time such as the present. In the context of a move towards more independent central banks and more visible inflation targets. US and Japan and find that retrospective estimates of the output gap. which means deflation risks have not disappeared. headline CPI measures are likely to continue to have a positive differential over core inflation in future years.5 -2 -2. are problematic for real-time forecasting and policymaking. However.

note that other research we have published suggests that overall rents are particularly sensitive to the output gap (see “The output gap is likely smaller than many think” (P Newland) . certain financial services. While taking account of these differentials.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Constructing demand-sensitive core consumer price indices As part of our analysis. By using chain weighting (based on nominal consumption weights). we find that in the past ten years the mean differential has averaged 0. based on consumer expenditure weights and price indices (which helps to avoid breaks in the headline CPI series). The series we have constructed also excludes the effects of the major changes in indirect taxes and administered prices (such as the sales tax increase in Germany and Japan. euro area and Japan (Figure 291). It is important to appreciate from the outset that core inflation measures typically reflect 50-70% of CPI measures. IT and photographic products. and exogenous assumptions may need to be made with respect to government-driven price changes. we use the ECB’s seasonally adjusted series for core inflation excluding food. it is likely that headline inflation measures will tend to outpace core inflation measures. We have done so primarily to obtain a series that represents prices that are likely to be particularly influenced by demand factors. While we exclude imputed rent among components for the purposes of this analysis of US inflaton. which we have adjusted for particularly large changes in administered prices and value-added taxes (this series is chainweighted by construction and tends to correspond to the US core PCE deflator. euro area and Japan with our core measures. its relationship to measures of economic slack. For the euro area. if we compare the difference of “headline” annual inflation measures in the US. As it is.78pp for the US. as well as tobacco 20). as opposed to supply shocks (commodities) and government interventions. and used cars). as has been happening. we try to avoid discontinuities associated with CPI re-basing. This is backcast pre-1990 by using the OECD’s estimate for the euro area CPI excluding food and energy. Our series are chain-weighted and. for the US and Japan. when it covered 48%. Another reason to expect headline inflation to outpace core inflation is the record dimensions of budget deficits in all major economies: it is very likely that governments will continue to rely heavily on indirect taxation as a source of additional revenue generation. This note also explains the calculation of the BarCap production function based US output gap 20 Overall. we have constructed our own series of core/underlying consumer price inflation. those price series that are estimated by statisticians (such as imputed rent 19. our “Underlying” US PCE deflator still covers 54% of US private consumption – this ratio has been relatively stable since 1960. As well.49 for the euro area and 1. in a highly globalised and integrated economy characterised by excess labour supply and pressure on limited commodity stocks. in our view. by seeking to identify the relationship between “core” inflation and economic slack we at least can have a basic building block for linking projections of economic activity to longer-term inflation projections. 19 15 March 2010 283 . in particular. This is also backcast (before 1980) using the OECD’s series for Japan’s CPI excluding food and energy. This can also be seen in the comparison of our “core” inflation rates with headline measures for the US. Global Economics Weekly 19 February. it excludes imputed rents). The underlying inflation measures that we have constructed and that we use in the following analysis exclude not just commodity prices but also. by no means accounting for all inflationary pressure. tobacco and energy.00pp for Japan. 0. drinks. Even though it can be hard to predict commodity price movements. it is still instructive to focus on core inflation and. we have constructed a chain-weighted index based on the private consumption deflators for durable and semi-durable goods. For Japan. with economic growth tending to be led by emerging economies. in the case of the US. plus services.

Thomson Datastream. % y/y) Euro area HICP less BarCap core rate. 5 yr MA 75 79 83 87 91 95 99 03 07 11 Japan Japan BC core PCE annual inflation (OECD core CPI pre 1980) Japan CPI ex fresh food % y/y Japan CPI ex fresh food less BarCap core PCE rate. % y/y) Euro area HICP (Datastream calculated pre 1990. 5 yr MA 15 12 9 6 3 0 -3 71 24 21 18 15 12 9 6 3 0 -3 71 75 79 83 87 Euro area Euro area core inflation (BarCap/ECB/OECD measure. Barclays Capital 15 March 2010 284 . 5 yr MA 91 95 99 03 07 11 Source: Haver Analytics.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 291: Comparing headline CPI indices with core inflation measures US 15 12 9 6 3 0 -3 63 67 71 75 79 83 87 91 95 99 03 07 11 US PCE core annual inflation % y/y US BarCap 'Underlying' annual inflation % y/y US CPI (NSA) % y/y US CPI less BarCap Underlying inflation rate.

2 0. whereas the CBO’s estimate for 1971 is currently -1.9 0. 1999).3 0.2 1. latest estimate vs.7 0.4 -0.3 0.3 2000 -0.1 2006 -0. casting doubt on the validity of the output gap as a relevant indicator for making policy decisions. 15 March 2010 285 . For example.9 0.6 … -4.0 -0. A particularly famous case concerns the real time estimates of the US output gap during the early 1970s.2 1. it is fairly common for such revisions to even cause the sign on the output gap.3 -0.8 1.3 -0. latest available estimates 2 1 0 -1 -2 -3 -4 latest estimate (Oct 09 WEO) "real time" estimate.2 0.9 0. the “real time” estimate made in the same year Source: IMF World Economic Outlook database archives 0. in Figure 292 we compare the IMF’s real time estimates of the output gap of “advanced economies” with its latest estimates (all data taken from various issues of the World Economic Outlook). Indeed. The very strong productivity growth in the late 1960s led economists at the time to extrapolate this into the 1970s.4 0.1 2001 -0. based on spring WEOs Estimate for year.7 -1.8 2007 -0.0 1.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Output gap considerations During the past decade it has become increasingly apparent that estimates of the output gap made in “real time” are subject to large revisions as information emerges in later years about the trend (potential) rate of GDP expansion.8 2003 -1. recently the “real time” estimates of output gaps have continued to be revised significantly.9 -0.4 -0. hence “real time” output gap estimates during the early 1970s signalled much greater economic slack than is now estimated (particularly given the subsequent inflation). difference.6 -0.2 2008 -0.4 0. Figure 292: IMF estimates of the output gap for “advanced economies”: “real time” vs.1 -0.2 0.1 2002 -1. the US Council of Economic Advisers estimated that in 1971 the US output gap was -6%.0 0.1 2004 -0. one year later Latest estimate for that year (Oct 09 WEO) average -1. “The reliability of output gaps in real time” (A Orphanides & Simon van Norden.0 2009 -4.7 -1.2 2005 -0. spring WEO for that year -5 1996 OG Dec Dec 94 96 Dec 98 Figure 293: How the OECD’s estimates for Japan’s output gap for the years 1994-98 were revised over time 3 2 1 0 -1 -2 -3 -4 1995 OG 1998 Dec 00 Dec 02 Dec 04 Dec 06 Dec 08 1994 OG % 1997 OG % -5 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: IMF World Economic Outlook issues Source: OECD Economic Outlooks Figure 294: Comparing IMF estimates of advanced economies’ output gap in real time with the latest IMF estimate % of potential GDP “Real time” estimate of output gap for that year. For example.5 21 This issue has been researched extensively by academics.3 0.1 Abs.0% 21.9 -0. See.5 1. Moreover.4 1. for example.

for example. as well as survey-based measures (which are not revised). but by its December 1999 Economic Outlook this had been revised to +0. Barclays Capital Japan output gap (OECD) US output gap (composite of BarCap production and survey based.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 295: Output gap measures Euro area output gap (OECD series. real time estimates of output gaps are particularly susceptible to retrospective revision when a major productivity shock is experienced. in our view this ignores the significance of both retrospectively estimated output gap measures. the IMF’s “real time” estimates of advance economies were persistently biased downwards in relation to the current. retrospective perspective (see also Figure 293 for more information behind Figure 291). the OECD estimated that Japan’s output gap for that year was -3. 15 March 2010 286 . such projections would risk generating inflation projections with a downward bias. Hence.8%. both approaches signal a large negative output gap. particularly when. However. as the experience of the 1970s suggests. if used in “real time” for economic projections. Haver Analytics. Given the propensity for significant retrospective revision of output gap estimates.9%). plus OECD series) As can be seen. This was also apparent from considering “real time” versus retrospective estimates of Japan’s output gap during the 1990s. Meanwhile. in December 1996. with respect to the evolution of core inflation. there can be a tendency to downplay the output gap as a concept of use for practical policymaking. using BarCap GDP projection) 8 6 4 2 0 -2 -4 -6 62 65 68 71 74 77 80 83 86 89 92 95 98 01 04 07 Source: OECD. Figure 293 shows how the OECD’s estimates of Japan’s output gap swung markedly as time went on. as at now.

Barclays Capital 15 March 2010 287 . 4qtr ma.5 0 -0.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 296: Comparing output gap measures with the change in the underlying consumer inflation rate US 8 6 4 2 0 -2 -4 -6 -8 62 66 70 74 78 82 86 90 94 98 02 06 10 Change in underlying inflation(4 qtr ma of 4 qtr change in annual % rate) BarCap production-based OG (4 qtr ma) OECD production-based OG (4 qtr ma) BarCap survey-based OG (4 qtr ma) Germany Underlying inflation change (4 qtr ma of 4 qtr change in annual % rate) 4 3 2 1 0 -1 -2 -3 71 75 79 83 87 91 95 99 03 07 11 Output gap (OECD measure. Haver Analytics. 4qtr ma) 5 4 3 2 1 0 -1 -2 -3 -4 1 0.5 1 0. RHS) 5 4 3 2 1 0 -1 -2 -3 -4 -5 5 4 3 2 1 0 -1 -2 -3 -4 -5 71 75 Euro area Underlying inflation change (4 qtr ma of 4 qtr change in annual % rate) Output gap (OECD measure.5 -1 -1. RHS) 2 1 0 -1 -2 -3 -4 -5 98 02 06 10 Source: OECD.5 82 86 90 94 98 02 06 10 Output gap (OECD measure. ECB data. 4qtr ma) BarCap survey-based OG (4 qtr ma) 79 83 87 91 95 99 03 07 11 Japan Underlying inflation change (4 qtr ma of 4 qtr change in annual % rate) 1.5 82 86 90 94 G3 Underlying inflation change (4 qtr ma of 4 qtr change in annual % rate) BarCap survey based OG (4 qtr ma) Output gap (composite measure 4 qtr ma.5 0 -0.5 -1 -1.5 -2 -2.

thereby solving the problem of retrospective revisions.0%. our analysis suggests the sensitivity of our underlying inflation measures to the survey-based output gap series has been somewhat less than that of the OECD’s production based series. 23 For the US. the OECD estimate of the 2006 output gap.34). The linkage between actual inflation and inflation expectations in two years’ time is apparent from the Livingstone published by the Philadelphia Fed (Figure 298).22. the German slope was already quite flat during the earlier episode (at 0. There are a couple of points of interest: Comparing the 1985-2009 episodes with that of 1972-1984. the OECD estimated that the 2006 euro area output gap was -0. However. this measure is still prone to revision – indeed. we use the ISM semi-annual capacity utilisation data. though this is offset to a considerable extent by the more reliable “real-time” aspect of survey data. whereas in December 2009 the estimated was +1. Explaining the “flatter Phillips curve” We explore the “flatter Phillips curve” in Figure 297. 24 While the Phillips curve was originally specified as the relationship of the inflation rate and the unemployment rate. because it is not historically revised and is constructed on a similar basis to the European Commission and Bank of Japan Tankan measures. For example. to 0. we see that the slopes of the best fit lines for the change in the underlying inflation rate for the US and euro area flattened and converged to a similar coefficient on the output gap of around 0. 15 March 2010 288 . while the EC and BoJ measures identify labour constraints within their quarterly surveys. which shows the ongoing significance of the relationship between the output gap and the change in underlying inflation rate. In our opinion.2%. Some results of our analysis can be seen in Figure 296. Similarly. That said. we refer to it here to describe the relationship of the change in the inflation rate to the output gap. for Japan. Another approach to calculating output gaps that is not prone to such major revisions is to base the estimate upon survey data. which are derived from the independence and credibility of the Deutsche Bundesbank.9%. and consequently flattened much less. was -0. the December 2007 estimate of the US 2006 output gap was +0.25. we use the NFIB labour shortages diffusion balance. and of their determination to enforce this. this reflects in part the stability of German inflation expectations.7%. Overall. whereas in its December 2009 Economic Outlook it pegged the 2006 output gap at +1. The flattening of the Phillips curve in the US and much of Europe during the 1970s and 1980s is associated with a significant decline in inflation rates. It also reveals that the sensitivity of changes in core inflation with respect to the output gap has changed over time. This has the advantage of using data that are not revised.7%. rather than the FRB series.Barclays Capital | Global Inflation-Linked Products – A User’s Guide In our analysis we have found that the best single retrospective predictor of core inflation series for the G3 countries is the OECD’s standard production function based output gap estimate. the OECD undertook a fundamental revision of its estimation procedure for potential GDP during the past year. compared with +1. This raises the question of why the German “Phillips curve” slope has historically been flatter and more stable than that of the rest of the euro area and of the US. 22 These revisions affected the OECD’s estimates of the output gap for the euro area and Japan more than those of the US. the US and euro area charts show that this elasticity since the mid-1980s has been much weaker than during the 1970s and early 1980s. However. Often referred to as a “flatter Phillips curve” 24. as central banks and governments gave much clearer signals as to what they considered an acceptable inflation rate. For US labour shortages. which has resulted in substantial differences from its previous estimates 22. as published in December 2007. particularly the difficulty of obtaining reliable estimates of labour and capacity shortages in the service sector – our estimates tend to reflect developments in just the manufacturing sector 23. in its December 2007 Economic Outlook.1% in the latest (December 2009) estimate. The drawback is that such an approach is constrained by data limitations.

1083 4 2 0 -2 -4 -6 R = 0.2219x .0.5281 2 Germany Q1 72 to Q4 84 5 4 change in inflation rate Germany Q1 85 to Q4 09 5 4 change in inflation rate 3 2 1 0 -1 -2 -3 -4 -5 3 2 1 0 -1 -2 -3 -4 -5 -6 y = 0.8564 2 4 R = 0.4154 3 2 1 0 -1 -2 -3 -4 -5 -6 -3 0 3 6 output gap (OECD estimate) 5 Euro area Q1 85 to Q4 09 5 change in inflation rate 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -3 0 3 6 output gap (OECD estimate) change in inflation rate y = 0.4304 R = 0.2403x . Thomson Datastream.3406x .0. then for pan-Germany from 1994 onwards).3342 R = 0.7643x + 0.5035 2 6 y = 0.6693x + 0.2076 2 -3 0 3 6 -6 -3 0 3 6 output gap (OECD estimate) output gap (OECD estimate) Note: Inflation rates are BarCap underlying series except for Germany (OECD series for core CPI ex food and energy for west Germany from 1972 to 1993.5297 2 y = 0.0368 R = 0.0.267x . Barclays Capital 15 March 2010 289 .0.0414 R = 0. Source: Haver Analytics.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 297: Comparing the relationship between the change in underlying inflation rate and output gap US Q1 72 to Q4 84 8 change in inflation rate change in inflation rate US Q1 85 to Q4 09 8 6 4 2 0 -2 -4 -6 -6 y = 0.1729 2 5 -3 0 3 6 -6 -3 0 3 6 output gap (composite measure) output gap (composite measure) Euro area Q1 72 to Q4 84 5 y = 0.

year after next (Livingston Dec surveys) US CPI: next 10 years (Livingstone survey) Dec 80 6 5 4 3 2 1 0 0 1 2 Standard deviation Dec 03 Dec 90 Dec 86 Dec 09 Dec 74 3 Note: This chart shows the deviation of individual responses for the projected US CPI annual inflation rate two years ahead. we are not aware of an equivalent survey with a time history as long as the Livingstone survey (the nearest equivalent is the ECB’s Survey of Professional Forecasters. Figure 298: US inflation expectations have moderated in the past decade 10 9 8 7 6 5 4 3 2 1 0 1969 1974 1979 1984 1989 1994 1999 2004 2009 Source: Federal Reserve Bank of Phladelphia. to close to German levels. Germany 7 6 5 4 3 2 1 0 1985 Euro area (ECB SPF: year after next.ahead HICP forecast) Euro ex Germany (OECD Nov EO 2yr. the projections published in successive December Economic Outlooks). which only dates back to 1999). as contained in successive vintages of the December Livingstone survey. so did the variation of individual forecasts. However. German CPI outturns and OECD forecasts German CPI (% y/y) Bundesbank "medium term inflation assumption" (ECB price stability definition after 1999) Germany (OECD Nov. Nov.ahead HICP fc) Germany (OECD Nov. Substantial outliers have been excluded from the calculations. the deviation of expected outturns has narrowed 10 9 Dec 79 8 7 Mean US core PCE deflator (% y/y) US CPI inflation rate. We can see how the OECD’s projections for future CPI inflation within the euro area countries excluding Germany declined significantly from 1985 to 1999. Figure 299 illustrates this point by showing the standard deviation of forecasts for the US annual CPI inflation rate in two years’ time. in Figure 300 we show OECD projections for inflation in Germany and the euro area (again. Haver Analytics Figure 299: As US inflation expectations moderated. surveys) Euro area (OECD Nov EO 2yr. ECB 15 March 2010 290 . EO 2 yr ahead CPI forecasts) German series boosted by re-unification in 1991 Figure 301: Bundesbank inflation projection vs. Haver Analytics Source: OECD (Economic Outlooks). EO 2 yr ahead CPI forecasts) 8 7 6 5 4 3 2 1 0 1988 1991 1994 1997 2000 2003 2006 2009 -1 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 Source: Bundesbank Discussion Paper 12/2009. For the euro area. Source: FRB of Philadelphia (Livingstone survey) Figure 300: Inflation forecasts two years ahead: euro area vs.Barclays Capital | Global Inflation-Linked Products – A User’s Guide As mean expectations for US inflation came down during the 1980s and into the 1990s. OECD.

Moreover. the relationship between the output gap and underlying inflation still appears to be at work.0 -1.Barclays Capital | Global Inflation-Linked Products – A User’s Guide In our view. RHS) 3 US output gap (composite measure. 4 qtr ma.5 -1 -1. this is often considered to have helped explain the smaller degree of second-round price increases in Germany during the 1970s and 1980s compared with other economies 25.5 0. series. then the deeply negative output gaps are likely to continue to exert downward pressure on underlying inflation rates.5 1 0. RHS) 3 2 1 0 -1 -2 -3 -4 -5 1. The experience of Japan – and. more recently.5 -2 90 94 98 02 06 10 BarCap survey-based OG (4 qtr ma. of countries in Europe. RHS) 1.5 0 -0. suggests that once deflation occurs. 2 1 0 -1 -2 -3 -4 90 94 98 02 06 10 Euro area Underlying inflation change (4 qtr ma of 4 qtr change in annual % rate) Output gap (OECD measure. if we consider developments since 1990. a cross check of production function-based estimates of inflation with “real time” survey measures signals that both are at very negative.5 1. which saw the consumer expenditure deflator and GDP deflator hit record lows of close to -3% y/y in Q4 09. it can be very difficult to extricate back into positive inflation. Bundesbank Discussion Paper 12/2009). Japan’s experience.5 -1. Figure 301 shows how the Bundesbank began to lay out a desired inflation estimate from 1976 onwards. suggesting that core inflation rates are likely to continue to decline in response to the large amount of economic slack. Figure 302: Comparing the recent relationship of the output gap to the change in the underlying inflation rate US Change in underlying inflation(4 qtr ma of 4 qtr change in annual % rate) Output gap (BarCap prod. These charts illustrate that it is a pressing need for economies to recover sufficiently so that their output gaps return to positive territory in a relatively short period of time.0 Source: Haver Analytics 25 See “Opting out of the great inflation: German monetary policy after the breakdown of Bretton Woods” (A Beyer. 15 March 2010 291 . which reflects the credibility and independence of the Bundesbank.5 -2. Moreover. 4qtr ma. C Gerberding and O Issing. suggesting that the signs on the production-function based estimates are unlikely to be revised. Along with a pragmatic approach to money targeting and the institutional framework for German wage setting. func. Recent developments As illustrated in Figure 302.0 -0.0 0. 4qtr ma. V Gaspar. a significant part of the explanation for the flatter Phillips curve in Germany during the 1970s and 1980s can be ascribed to the greater prominence in price setting paid by inflation expectations. If this does not happen. given the importance of expectations. such as Ireland – shows that it can be relatively straightforward to slip into deflation.

as unemployment can fall by 1. the main creditors of the G4 economies. Undeniably. As the majority of government expenditures are not indexed to inflation and taxes rise one-to-one with inflation. inflation remains very low. while eluding the political hurdles that typically prevent expenses from falling. Central bankers. inflation can achieve what no congress can. Quantitative easing (QE) has shown that the zero interest rate bound does not mean central bankers are left without ammunition against deflation.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Run inflation run Christian Broda +1 212 526 8536 christian.yang@barcap. especially the Fed and the BoE. which gives them another reason to be ‘behind the curve’ at the start of their tightening cycle. fast reductions in fiscal deficits. UK and Japan are considerable. Figure 303: 5y breakeven inflation rates % 4 3 2 1 0 -1 -2 -3 -4 Jan-06 UK EUR Japan US 5yr breakeven inflation rates Jan-07 Jan-08 Jan-09 Jan-10 Note: Breakeven inflation based on inflation indexed bonds with around 5yrs maturity. are likely to tolerate higher inflation in coming years. We argue that the benefits of temporarily higher inflation (eg. While policymakers should worry about another asset bubble.com Mimi Yang +1 212 526 8536 mimi. two years of 5% inflation) for the US. regulation rather than tight monetary policy should tackle excessive leverage. so the best policy prescription today is not to increase inflation targets – as was recently suggested by the IMF – but rather to have above-target inflation. But contrary to popular belief. we estimate that 3pp higher inflation can reduce fiscal deficits by 1pp of GDP. inflation is an undercover fiscal reform. important reasons to increase inflation or inflation targets. and the redistributive loss from higher inflation would hit the balance sheets of ailing domestic financial sectors.5pp in one year solely due to this higher inflation. in our view. On the fiscal front. In this context. as most emerging markets can testify. we believe this would be a positive development for the global economy. Combating deflation and reducing real debt burdens are not. Central bankers know this. The principal merits of inflation disappear if inflation is anticipated. it is surprising that markets are not expecting inflation to rise above their targets (at least temporarily) in coming years.mondino@barcap.broda@barcap. Source: BarCap Live 15 March 2010 292 .com Guillermo Mondino +1 212 526 8536 guillermo.com Contrary to previous periods of high unemployment and fiscal deficits. We estimate that having 5% inflation rather than 2% can dramatically accelerate the decline in unemployment. Inflation has historically been a powerful force to achieve fiscal and real adjustments. making the adjustment process harder and slower.

Contrary to previous times of high unemployment (eg. Barclays Capital Source: Haver. A February 2010 IMF report by Blanchard. Especially for countries/regions with highly inflexible labor markets. This illustrates the main point of this piece. to quantify the cost of having temporarily high inflation. at times where unemployment is high – that is. QE can side-step the banking system and provide direct credit support. We do not attempt. inflation allows for a politically costly fiscal consolidation to occur without the need for the political support needed to cut government expenditure. If alive. and effectively reduce the yield curve in ways that lowers the effective credit costs. today inflation rates are substantially lower making the adjustment process harder and slower. making today’s situation more complicated to resolve 14 12 10 8 6 4 2 0 -2 -4 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 US UK Japan Euro Area G4 inflation rates % y/y 14 12 10 8 6 4 2 0 80 85 US G4 unemployment rates % 90 UK 95 Japan 00 05 Euro Area 10 Source: Haver. in our view. The new blueprint for QE that the Fed and BoE have effectively established is more than enough ammunition for central bankers to fight deflation when hitting the zero-interest rate bound. inflation can be your biggest ally. at times where the real wages are too high –higher inflation helps reduce real wages and. The main contribution of this note is to quantify the gains from moderately higher inflation. but if these costs are not highly variable with unemployment and debt – as they are likely not to be – the time for high inflation is now. For reasons that keep behavioral economists busy. Contrary to the view spelled out in that work. Dell’Ariccia and Mauro reignited a debate about whether inflation targets should be higher. For this reason. But there are other reasons that provide a much larger benefit of having higher inflation. A modest increase in inflation can produce large gains in terms of falls in unemployment. Inflation acts as an undercover real and fiscal reform as inflation can accelerate the fall in unemployment and the consolidation of fiscal accounts.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Rudiger Dornbusch once said that Argentina needed an Austrian central banker to impose credibility and reduce inflation. Since a large portion of government expenditures are not indexed to inflation. Moreover. Central bankers know this and they are likely to tolerate higher inflation in coming years rather than change their inflation targets. people don’t accept nominal wage cuts but are insensitive to inflation eating into their purchasing power. Barclays Capital 15 March 2010 293 . that the benefits from higher inflation in the context of a developed world with high unemployment. We estimate that the gains from higher inflation could be as high as 1pp of GDP per year of reduction in fiscal deficits purely due to a change in inflation from 2% to 5%. unemployment. he would probably be saying today that G4 central bankers would benefit from having an Argentine on their board. 80s or early 90s). inflexible institutions – except in the US – and high fiscal deficits are considerable. however. we do not think that higher inflation targets are warranted. Figure 304: Unemployment is close to previous highs… Figure 305: … but inflation is not nearly close to its previous highs. inflation acts as an undercover fiscal reform. in turn.

This makes the inflation rate the upper bound of any fall in real wages. At today’s real wages (eg. Moreover.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Contrary to popular belief. which mean that the income losses are not borne by foreigners. (eg. Formally. which gives rise to the currently high levels of unemployment. Unanticipated Inflation can reduce real wages. The higher is inflation in a given year. the larger is the potential fall in real wages in that year. unemployment can fall by around 0. and inflation rates are very low (Figure 305). The domestic financial market is the largest creditor of debt instruments in most of these economies. Inflation: An efficient way of reducing unemployment There are two key macro building blocks that suggest why inflation can be good to reduce unemployment. and so the income loss from inflation to creditors would be an additional hurdle in the rebuilding of their balance sheets. the redistributive power of inflation from debtors to creditors is not a large plus to G4 economies. The faster that real wages are reduced. These two facts combined make inflation more desirable in periods like today. Conservative estimates of both these elasticities suggest that for every 1 percent fall in real wages achieved via a higher inflation rate. First is the simple (and undisputed) fact that employers typically do not reduce nominal wages (Figure 4). We estimate that the gains from reducing the real burdens of debt are small for all G4 countries economies. Real wages do not fall fast enough to reduce unemployment because nominal wages are ‘sticky’ downwards. rw0) in the figure. where unemployment is high (Figure 304). the gap between labor demand (how many people firms want to employ) and labor supply (how many people want to work) is positive. to rw1).5%. But how much can inflation really help reduce unemployment? Figure 307 helps understand what variables matter to understand the quantitative impact of higher inflation on lower unemployment. Barclays Capital rw rw' Labor demand Quantity demand of labor: L Source: Barclays Capital 15 March 2010 294 . Higher inflation essentially allows real wages to fall and thus reduce unemployment faster. This Figure 306: Nominal wage growth doesn’t turn negative… Figure 307: … even though a fall in real wages can reduce unemployment Real wage: w/p unemployment without inflation unemployment with inflation Labor supply Wages (wkly earnings) y/y 10 9 8 7 6 5 4 3 2 1 0 65 68 71 74 77 80 83 86 89 92 95 98 01 04 07 Source: BLS. Second. and concomitantly reduce the levels of unemployment. the extent of the decline in unemployment from a fall in real wages is given by the elasticity of labor supply minus the elasticity of demand. and with low rates of inflation the economy is left hanging in a high unemployment range for a long time. unemployment is the result of real wages being too high. Europe and Japan are net foreign creditors. the faster unemployment falls.

0 US Source: Barclays Capital Figure 309: Reduction in fiscal deficits due to inflation when nominal wages hit the zero bound % GDP 1. fiscal expenditures are hard to cut.2 1. Thus in two years. some government expenditures are indexed to inflation. Barclays Capital 15 March 2010 295 . but first we highlight the benefits of inflation to improve fiscal deficits. inflation can help make the most unfeasible fiscal deficit reductions a reality.0 Euro area UK Japan Note: Source: OECD. encompassing on average around 14% of GDP. Just like wages. the outlook for higher inflation expectations would probably increase as the temptations will be substantial. This means that while tax revenues grow at the rate of inflation.0 0. Social security benefits are typically around one-third of all government expenditures. Higher inflation helps reduce fiscal deficits as governments can collect higher nominal revenues.8 0. unemployment can fall by 3% only as a result of having temporarily higher inflation for two years.Barclays Capital | Global Inflation-Linked Products – A User’s Guide implies that having 5% inflation in a given year can reduce unemployment by 1. But without certainty about whether inflation will be higher in the future. as political reasons make reducing nominal expenditures a hard sell.8 Reduction in government expenditures from a 3% increase in inflation Share of government expenditures by type in 2009 Government wages Socical security benefits Other 52 0. Inflation expectations have remained steady through this period and policymakers know the power of inflation surprises at this juncture.2 0. While it is hard to generalize. it is unlikely that wage negotiations will incorporate higher inflation so rapidly. fast reductions in fiscal deficits as revenues grow with inflation.6 1. Figure 308 shows the share of spending for different expenditure categories. and given the current high unemployment rates. Assume that. However. just as Figure 308: Social security benefits are mostly indexed to inflation % 60 50 40 30 20 27 10 0 US Euro area UK Japan 46 38 35 21 33 23 28 16 49 33 0. Nominal wages for public servants can remain flat for a year or two while inflation runs higher. If inflation rates are low. If unemployment doesn’t fall in coming years. so do all the expenditures that are indexed to inflation. We will discuss this potential benefit in the next section (again.1 0. with an eye on quantifying its impact).7 1. politicians face the barrier of having to agree on wage cuts for public servants or pensions reductions to the elderly. To understand the impact that inflation has on fiscal savings we provide a simple illustration. most social security spending in G4 economies are adjusted according to a cost of living price index. It is likely that after some time the benefits of having higher inflation would dissipate. Inflation: An undercover fiscal reform When investors discuss the positive aspects of inflation the focus typically is on the benefits of inflating away high levels of debt. This can achieve what no congress can.5% more than in a world with 2% inflation. Inflation helps elude political barriers to achieve fiscal consolidation. But as many emerging markets can testify. as unions internalize that they need to bargain for larger increases in nominal wages and the benefit of falling real wages gets eroded.4 0.

and the benefits from inflation as an implicit fiscal reform disappear. the benefits of inflation to reduce debt burdens and improve aggregate demand are more nuanced than what appears at first sight. The non-financial domestic sector comprises households. Fiscal deficits would fall by between 0. Once again. There are two main reasons why the impact of inflation to reduce debt burdens may not be as beneficial as most investors expect. all other net credit liabilities calculated from Flow of Funds data.1 … ….1pp of GDP (UK) solely from higher inflation – that is. implying that deflation would make the real burden of the debt servicing larger and drag aggregate demand downwards. However. Barclays Capital Note: Average maturity of household debt is for mortgages. Considering the magnitude of today’s deficit in all G4 economies.4 73. for some time. Nuances about inflating away debt Irving Fisher has popularized the fears about debt deflation – the notion that payments to service outstanding debts may be fixed.7pp of GDP (US) and 1.8 13.7 10.2 78.2 55.3 6. ignoring any explicit effort to reduce real wages or increase tax rates.2 4. inflation runs at 5% rather than 2% for one year. this is not a permanent benefit from inflation.5 28. When budgetary discussions start to internalize the higher inflation this would prevent government expenditures from falling in real terms. The domestic financial sector is the main creditor of credit assets. net foreign debtors do not bear a large part of the costs of inflation.Barclays Capital | Global Inflation-Linked Products – A User’s Guide in the previous section. Figure 310 hints to why this is small. this transfer happens at the expense of creditors. which puts holders of existing debt at a loss. *Japan general government net debt from OECD.0* 4. Only a small fraction of outstanding debt is held abroad. Source: Flow of Funds. As such. Barclays Capital 15 March 2010 296 . However.0 96. inflation could provide a much needed acceleration in fiscal consolidation without confronting political barriers.0 46. so even in the case of the US and UK.8 100. If all debt is held domestically in a country. and although inflation effectively is an income transfer to debtors.9 67. If the debt servicing burdens are fixed in nominal terms.3 4. *Japan general government net debt as reported by the OECD. above-target inflation could imply lower real burdens for governments. We will examine this impact quantitatively below.9 5. for each debtor there is a creditor.0 43. … 75. Figure 310: Debtors and creditors in the credit market Nonfinacial domestic sector US -181 Financial Sector 156 RoW Figure 311: Net credit assets by sector Net credit liabitlies of the nonfinancial sector Households Businesses General Govt Composition of debt % GDP 41 EA -2 -187 189 US Euro UK Japan 14 65. as high real debt servicing is a reduction in disposable incomes. Assume also that social security benefits and taxes increase at 5% and that all non-indexed expenditures grow at only 2% (like the typical inflation target). maybe a couple of years. this means that the impact of inflation on aggregate demand is determined by the difference between the propensities to consume of debtors versus creditors times the effective income transfer. businesses and general government.5 6. Source: OECD. How large would the fiscal savings be in this case? Figure 07 provides the answer for each country.2 84. but it is much smaller than most believe. Flow of Funds.3 UK -239 225 Average maturity of debt (in years) Japan* -20 -200 -218 213 -300 -100 0 100 200 300 US Euro UK Japan Note: Shows only net credit asset positions. higher inflation reduces the real burden of interest payments.2 55. First. inflation could have the opposite effect.

5 189.1 -187. Figure 9 shows the average maturity and overall debt composition of each component of domestic debt in G4 economies.5. they consume 1 dollar – while that of creditors is 0. Source: Barclays Capital 15 March 2010 297 .12 Financial sector -0.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Second.5 -178.17 -0. undoing any of the benefits of higher inflation.0 -3. The figure shows that the increase in consumption due to a 3pp higher inflation rate is around 0.1 -235.3 225.2 -2.18 0.7 Financial sector 213.6 Financial sector -3. In the context of the need for the financial sector to start lending and improve their balance sheets.3 -239.12 Financial sector -0.2 186. This highlights the benefits for this sector to increase inflation.1 3.5.2 -180.02 0.11 0.11 0. it is enough to illustrate how small these benefits are relative to those previously quantified. 2010) highlighted that one of the one main implications of this was the need for more reliance on fiscal policy and for larger deficits than would have been the case absent the binding zero interest rate constraint. The general government debt in the UK is also of particularly long maturity.18 0. it is likely that short-term debt would be rolled over at higher nominal interest rates.4 2. Blanchard et al (IMF. corporate debt and mortgages weighted by share of all nonfinancial debt. while credit market assets are half fixed in nominal terms.1 156.3 222. in most countries the main creditors of debt instruments are the financial sector and this implies that while inflation is probably a benefit for most. it is a drag on financial institutions. In (5) we use simple model where marginal propensity for net debtors (non-financial sector) to consume is 1.02 -0.04 0.04 0.05 0.06 percent of GDP. the real losses from inflation would not help this process.0 (3) Reduction in value of real debt Nonfincial sector 3. and only half effectively indexed to inflation. Assume that the propensities to consume of debtors are 1 – ie. not all debt is subject to fixed interest payments. out of an increase of 1 dollar of income. as nominal interest rates are likely to incorporate the higher inflation expectations. However.7 -2. But the zero nominal interest rate bound prevented them from being so low. After a bout of inflation.06 -0. they are real assets). the IMF Figure 312: Back of the envelope calculation – benefit of inflating away debt % of GDP Japan UK EA US (1) Net debt assets of domestic economy (excludes the foreign sector) Nonfincial sector -218. This is essentially the reason why inflating debts is not a big boost to demand. While of course this assumption is important to quantify the precise benefits of reducing the real burden of debt. Household debts in the US – mostly mortgages – stand out as having long maturities (and mostly tied to fixed interest rates).4 154.2 (4) Reduction in debt service (reduction in real debt*average interest rate) Nonfincial sector 0.05 Net consumption gain 0.10 0.07 Note: In (4) the interest rate is used is the average on government bonds. As a result.09 -0.2 (2) Value of debt after 5% inflation (assuming 1/2 of assets grow 1-1 with inflation) Nonfincial sector -215. Combating deflation Based on a simple Taylor rules nominal interest rates should be between -3 to -5 percent in several of the G4 countries.04 -0. making the UK Treasury a primary beneficiary of potentially higher inflation rates. a negligible amount.7 -184.10 (5) Estimated impact on consumption from reduction in debt servicing cost* Nonfincial sector 0. Also assume that all housing and equity assets increase in value one-for-one with inflation (ie. (The gap is large to highlight that the small results are particularly stark). How to quantify this potential benefit of inflation? Figure 312 shows a simple back-of-theenvelope calculation.7 2.11 -0.0 and for net creditors (financial sector) is 0.2 Financial sector 210.

Barclays Capital As of today Euro UK Japan 15 March 2010 298 .0 -0. Figure 313 shows how much the relevant interest rates for the private sector have fallen since short-term rates have hit their lows.” We believe this conclusion assumes that short-term nominal interest rates are the only instrument available for central bankers.Barclays Capital | Global Inflation-Linked Products – A User’s Guide suggests: “It appears today that the world will likely avoid major deflation and thus avoid the deadly interaction of larger and larger deflation. But it is clear that the zero nominal interest rate bound has proven costly.6 -0. QE is enough ammunition for central bankers to fight deflation when hitting the zero-interest rate bound. Through purchases of assets of different maturity they have been able to affect the entire yield curve without changing the overnight rates. higher and higher real interest rates. central bankers have been able to influence other interest rates beyond the overnight policy rate. It is hard to quantify how much the option value is of having more room to manouvre in terms of short-term rates to prevent deflation. existing labor contracts and expenditure plans are likely not to incorporate the higher inflation for some time.2 US Source: Haver. Figure 313: Change in average interest rates since last central bank rate cut Change in average* interest rate since last offical rate cut As of 6mos after last rate cut 0.4 -0. As has been particularly clear in 2008-09. this value has fallen considerably and may now be negligible. Higher average inflation and thus higher nominal interest rates to start with. If central bankers announced an immediate increase in targets and inflation moved beyond previous targets. While it is true that long-term rates have moved for reasons other than policy. However.2 -0. it is undeniable that QE has had an impact on the relevant funding rates for the private sector. This could imply some short-term benefits from the nominal stickiness of many of the existing real and fiscal conditions. we do not believe that it is warranted to modify the inflation target to combat deflation. would have made it possible to cut interest rates more.0 -1. some of the benefits from above-target inflation that we highlighted in previous sections could also be achieved with an immediate change in inflation targets. and a larger and larger output gap. For this reason. as argued in our previous discussion. typically until the contracts are renegotiated. but we believe that given the role played by QE. thereby probably reducing the drop in output and the deterioration of fiscal positions.8 -1.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide APPENDICES 15 March 2010 299 .

maliye.fazenda.co www.treas.gov www.br www.gov.ca www.ar www.gov noticias.mecon.gov www.bundesbank.gob.com Barclays Capital Barclays Capital Live (Publications & Analytics) https://live.int www.gov.mof.kr Australian Office of Financial Management Bank of Canada Agence France Tresor (AFT) German Bundesbank Public Debt Division.gov.cl www.aft.bankofengland.mx www.gov. ecb.federalreserve.go.bankofcanada.gov.org.aofm.bls.hacienda.uk www.tr english.gov.gov.com live.com www.uk www.gouv.rgk.Barclays Capital | Global Inflation-Linked Products – A User’s Guide KEY INFORMATION SOURCES Figure 314: Barclays Capital links www.gov.it/en www.dt.rba.pl www. Italian Treasury Japanese Ministry of Finance South African National Treasury The Swedish National Debt Office UK Debt Management Office US Treasury US Treasury Direct Argentinian Finance Ministry Brazilian National Treasury Mexican Ministry of Finance and Public Credit Chilean Ministry of Finance Polish Ministry of Finance Colombian Ministry of Finance Turkish Finance Ministry South Korean Ministry of Finance Figure 316: Useful links www.mosf.treasury.tesoro.or.br/english www.barcap.go.gov.se www.shcp.mf.barcap.minhacienda.de www.treasurydirect.fr www.au www.jp www.riksbank.gov www.mx www.cl www.bcentral.tesouro.boj.uk www.barcap.au www.za www.gov.pl US Federal Reserve Bank of Japan European Central Bank US Bureau of Labor Statistics The Bank of England UK National Statistics Office Riksbank Reserve Bank of Australia Brazilian Ministry of Finance Banco de Mexico Chilean Central Bank National Bank of Poland 15 March 2010 300 .bcb..com/BC/barcaplive?menuCo Barclays Capital Inflation-Linked Analytics de=MENU_AR_IR_AT_ILAT Figure 315: Issuer links www.nbp.gov.co.jp/english www.banxico.statistics.dmo.

Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 317: Bloomberg pages BINF BILR BCIN BBEM BCAP1 BCAP2 BCAP3 BCEQ9 BCAP7 BXEI BXGL BISW1 BISW2 BISW3 BISW4 BISW5 BISW6 BISW7 BISW8 BISW9 BISW10 BXPB3 BXEY BXEZ BXJB11 BEMI BXIX2 SWIL ILB ILBE ILBI DMO1 Barclays Capital Inflation-Linked Menu Barclays Capital Inflation-Linked Research Barclays Capital Inflation-Linked Indices Barclays Capital Inflation-Linked EM Indices Barclays Capital TIPS Prices Barclays Capital US Actives and Forwards Barclays Capital US Zero Coupon CPI Swaps Barclays Capital Structured Inflation Notes Barclays Capital Inflation Forecasts Barclays Capital Euro Inflation-Linked Prices Barclays Capital UK Index Linked Prices Barclays Capital Euro ZC HICPx Swaps Barclays Capital French ZC CPIx Swaps Barclays Capital UK RPI Swaps Barclays Capital Italian ZC CPIx Swaps Barclays Capital Spanish ZC CPI Swaps Barclays Capital Euro Govt Asset Swaps Barclays Capital UK Linker Asset Swaps Barclays Capital European Real Rate Swaps Barclays Capital French Real Rate Swaps Barclays Capital UK Real Rate Swaps Barclays Capital UK Corporate Linkers Barclays Capital OAT€i Strips Barclays Capital OATi Strips Barclays Capital Japan Inflation-Linked Prices Barclays Capital EM Inflation-Linked Products iShares Fixed Income Inflation Bond/Swap Defaults Global Inflation-Linked Bonds Bloomberg World Inflation Breakeven Rates Inflation-Linked CPI Indices UK DMO Screen Announcements 15 March 2010 301 .

235bn January 97 CPI All urban nsa $341.692bn September 98 French CPI ex-tobacco Euro HICP extobacco FRCPXTOB Index.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Summary sovereign table Figure 318: Developed markets overview US Treasury Inflation Indexed Securities.767bn Japan Generic name OATi.12bn 5 CAD45.239bn March 06 Euro HICP ex-tobacco $131.130bn March 03 Euro HICP ex-tobacco $33.235bn UK United Kingdom Index-Linked Gilts 16 £211. TIIS.583bn July 85 $69. TIPS 28 $563. 3 without Annual or zero coupon 2-3 months 6 months Coupon and principal par floor Quarterly Floor? Coupon frequency Note: * At end of 2009 Source: Barclays Capital Par floor No floor Par floor No floor Par floor Par floor No floor Semi-annual Semi-annual Annual Annual Semi-annual Annual Semi-annual Semi-annual 15 March 2010 302 .474bn December 91 CPI All Items nsa $10.350bn April 94 $43.758bn 2 €12. CPTFEMU Index 2-3 months $43.074bn March 81 $218.211bn France Germany Italy Greece Sweden Swedish Government Index-Linked Canada Canadian Real Return Bonds Australia Australian Capital Indexed Bonds 3 AUD11.636bn 5 SEK238.425bn 3 €30.576bn 16 ¥6.650bn September 03 Euro HICP ex-tobacco $18.012bn March 04 Nationwide CPI General ex-Fresh Food JCPNGENF Index 2-3 months to 10th of month No floor Linking Index RPI CPI nsa All groups CPI Linking Index Bloomberg ticker CPURNSA Index UKRPI Index CPTFEMU Index CPTFEMU Index CPTFEMU Index SWCPI Index CACPI Index AUCPI Index Indexation lag 2-3 months 8 months or 2-3months 2-3 months 2-3 months 2-3 months 2-3 months 3 with par floor. BTAN€i OBL€i DBR€i BTP€i GGB€i JGBi No bonds Outstanding* Market value outstanding bn* Market value outstanding $bn* First issue date 10 €152.424tn $563.137bn 7 €91. OAT€i.

149bn $18.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 319: Emerging markets overview Brazil Generic name No.699bn $7.613bn TRY37.483bn PLN10. IBREIGPM Index Up to 4 weeks. bonds outstanding* Market value outstanding bn* Market value outstanding $bn* First issue date in current format Linking Index Linking index Bloomberg ticker NTN-Bs.534bn $1.113bn KRW1.383bn $27.412bn COP15. includes forecast Turkish CPI Polish CPI Korean CPI MXUDI Index ARCPI Index CLUFUF Index COCPI Index SACPI Index TUCPI Index POCPIYOY Index KOCPI Index Indexation lag Up to 2 weeks T-5.540bn CLP4.176bn MXN360.5 months.619bn $15.726bn $38. No floor (ILCPI) Annual March 00 South Africa CPI NSA February 07 September 03 February 07 IPCA.678bn May 00 May 96 Unidas de Inversion (UDI) December 03 CER Consumer Price Index September 02 UF Consumer Price Index October 02 UVR Consumer Price Index June 06 Israel CPI ISCPINM Index Up to 1.784bn ILS140bn ZAR139.129bn $8. T-10 to ACERCER Index 1 month to 9 of month th 1 month to 15 of month th 3-4 months 2-3 months 2-3 months 2-3 months Floor? No floor No floor No floor No floor No floor Par floor Par floor Par floor No floor Coupon frequency Note: * At end of 2009 Source: Barclays Capital Semi-annual Semi-annual Monthly or semi-annual Semi-annual Monthly Semi-annual Semi-annual Annual Semi-annual 15 March 2010 303 . ILCPI South Africa South Africa Index-Linked 4 Turkey TURKGB Poland POLGB South Korea KTBi 9 16 6 10 4 1 1 BRL253.960bn $25. NTNCs 11 Mexico Udibonos Argentina Argentinean Government Inflation-Linked 5 Chile BCU Colombia TES Israel Galil.579 bn (liquid NTN-Bs) ARS57. IGPM BZPIIPCA Index. adjusted on inflation release Coupon and principal par floor (Galils).954bn $134.006bn $3.

0% 2002 1.5% 0.5% 4.0% OAT€i 10yr CM real yield OATi 10yr CM real yield Figure 323: Germany DBR€i 10y real yield 3.5% 0.0% 0.Barclays Capital | Global Inflation-Linked Products – A User’s Guide REAL YIELD HISTORIES Figure 320: US 10y TIPS real yield 5.5% 4.0% 3.5% 1.0% 1.5% 2.5% 0.5% 1.0% 2.5% 2.5% 0.0% 0.0% 3.5% Germany 10yr CM real yield 1.5% 2.5% 3.0% 2004 BTP€I 10yr CM real yield Figure 325: Greece GGB€i 2025 real yield 6 GGB€I 2025 5 4 3 2 2005 2006 2007 2008 2009 2010 1 2003 2004 2005 2006 2007 2008 2009 Source: Barclays Capital 15 March 2010 .5% 1.5% 2.5% 3.5% 1.0% 2.0% 3.5% 3.0% 1.0% 2.5% 3.0% 2.0% 0.0% 1.0% 1997 1999 2001 2003 2005 2007 2009 US TIPS 10yr CM real yield Figure 321: UK 10y real yield 4.0% 1997 1999 2001 2003 2005 2007 2009 2.0% 4.0% 0.0% 0.0% 2006 2003 2004 2005 2006 2007 2008 2009 2007 2008 2009 Figure 324: Italy 10y BTP€i real yield 4.5% 0.0% 2.0% 1.5% UK 10yr CM real yield Figure 322: France OAT€i and OATi 10y real yields 3.

5% 2.0% 1.0% 0.0% 4.5% 1.0% 1.5% 2.5% 0.0% 1.5% 3.0% 1.0% 2.5% 3.0% 1997 1999 2001 2003 2005 2007 2009 Figure 328: Japan 10y real yield 5.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 326: Australia 10y real yield 4.5% 3.0% Australia 10yr CM real yield Figure 327: Canada 2021 real yield 5.0% 3.0% 2.0% 4.5% 0.0% 4.0% 2.0% 1.0% 5.0% 2.5% 5.0% 2004 2005 2006 2007 2008 2009 JGBi 10yr CM real yield Figure 329: Sweden 10y real yield 5.0% 2.5% 2.0% 4.5% 4.0% 3. Bloomberg 15 March 2010 .5% 3.5% 2005 2006 2007 2008 2009 Source: Barclays Capital.5% 6.0% 3.5% 2.0% South Africa 10yr CM real yield 4.5% 2.0% 0.0% 2004 2005 2006 2007 2008 2009 2010 2.0% 1997 1999 2001 2003 2005 2007 2009 Sweden 10yr CM real yield Figure 330: Israel 10y real yield 6.5% 1.5% 2002 2003 2004 2005 2006 2007 2008 2009 0.5% Israel 10yr CM real yield Figure 331: South Africa 10y real yield 4.5% 2004 2.0% 3.5% 3.5% 3.5% CANRR 2021 3.0% 0.0% 3.5% 4.5% 1.5% 4.5% 4.

0% 3.5% 2004 Mexico 10yr CM real yield Figure 337: Turkey 5y real yield 25% Turkey 5yr CM real yield 20% 15% 10% 5% 2005 2006 2007 2008 2009 2010 0% 2007 2008 2009 2010 Source: Barclays Capital 15 March 2010 .5% 3.5% 5.0% 5.5% 3.5% 4.0% 4.0% 3.5% 6.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Figure 332: Argentina 10y real yield 70% 60% 50% 40% 30% 20% 10% 0% 2004 Argentina 10yr CM real yield Figure 333: Brazil 10y real yield 12% 11% 10% 9% 8% 7% 6% 5% 2004 Brazil 10yr CM real yield 2005 2006 2007 2008 2009 2010 2005 2006 2007 2008 2009 2010 Figure 334: Chile 10y real yield 5.0% 4.0% 2004 Chile 10yr CM real yield Figure 335: Colombia 10y real yield 10% 9% 8% 7% 6% 5% 4% 3% 2% 2004 Colombia 10yr CM real yield 2005 2006 2007 2008 2009 2010 2005 2006 2007 2008 2009 2010 Figure 336: Mexico 10y real yield 6.0% 2.0% 2.5% 4.5% 2.

5% 0.0% 1999 2001 2003 2005 2007 2009 OATi 10yr CM breakeven FRFCPx y/y Rolling 3yr average inflation Figure 342: Canada 10y breakeven vs realised inflation 5% 4% 3% 2% 1% 0% CANRR2021 breakeven -1% -2% 1998 Canada CPI y/y Rolling 3yr average inflation 2000 2002 2004 2006 2008 2010 Figure 343: Sweden 10y breakeven vs realised inflation 5% 4% 3% 2% 1% 0% -1% -2% 1997 Sweden 10yr CM breakeven Sweden CPI y/y 3yr rolling average inflation 1999 2001 2003 2005 2007 2009 Source: Barclays Capital 15 March 2010 .5% -1.0% 1.5% 3.0% 0.5% 2.5% 1.0% -0.0% 3.Barclays Capital | Global Inflation-Linked Products – A User’s Guide BREAKEVEN INFLATION HISTORIES Figure 338: US 10y TIPS breakeven vs realised inflation 6% 5% 4% 3% 2% 2% 1% 0% -1% -2% -3% 1997 TIPS 10yr CM breakeven US CPI-U NSA y/y Rolling 3yr average inflation 1999 2001 2003 2005 2007 2009 1% 0% -1% -2% 1997 Figure 339: UK 10y breakeven vs realised inflation 6% 5% 4% 3% UK 10yr CM breakeven UK RPI y/y Rolling 3yr average inflation 1999 2001 2003 2005 2007 2009 Figure 340: France 10y OATi breakeven vs realised inflation 5% 4% 3% 2% 1% 0% -1% -2% 2001 2002 2003 2004 2005 2006 2007 2008 2009 €i 10yr CM breakeven EUHICPx yy Rolling 3yr average inflation Figure 341: Euro €i 10y breakeven vs realised inflation 4.0% 2.

com Marcelo Salomon Chief Brazil Economist +54 (0)11 5509 3295 marcelo.com Xiaonan Jiang +44 (0) 20 7773 0822 jenny.com Jeff Gable Head of ABSA Capital Research +27 11 895 5368 jeff.com Anand Venkataraman +44 (0) 20 7773 0852 anand.callow@barcap.sen@barcap.com Arko Sen EMEA Rates Strategist +44 (0)20 3134 2839 arko.com Marcela Barreto +44 (0)20 3134 0750 marcela.jiang@barcap.com Christian Broda Head of International Economic Research +1 212 526 8536 christian.com Stefan Liiceanu Japanese Strategy +81 3-4530 1554 stefan.com Roberto Melzi Senior Latam Strategist +52 55 5241 3260 roberto.keller@barcap.com Christian Keller Chief Economist – Emerging Europe +44 (0)20 7773 2031 christian. Malaysia.huang@barcap.gable@absacapital.com Nick Verdi International Economist +44 (0) 20 7773 2173 nick.com Daniel Hewitt Emerging EMEA Economist +44 (0)20 3134 3522 daniel.com Dean Maki Chief US Economist +1 212 526 1731 dean.liiceanu@barcap.sharma@barcap.com Rahul Sharma +65 6308 3982 rahul. Taiwan +65 6308 3292 waiho.com Yuan Tian +44 (0) 20 7773 1436 yuan.com Mimi Yang International Economist +1 212 526 5194 mimiyang@barcap.com Bulent Badsha South Africa Strategist +27 (0)11 895 5323 bulent.bettiss@barcap.verdi@barcap.leong@barcap.Barclays Capital | Global Inflation-Linked Products – A User’s Guide GLOBAL INFLATION-LINKED USER'S GUIDE CONTRIBUTORS AND CONTACTS Barclays Capital 5 North Colonnade London E14 4BB Strategy Alan James Global Inflation-Linked Strategy +44 (0)20 7773 2238 alan.melzi@barcap.salomon@barcap.mondino@barcap.upbin@barcap.tian@barcap.vargas@barcap.com Jenna Myers +1 212 526 0500 jenna.christodoulou@barcap . Colombia.com Matthew Huang Strategist – North Asia +65 6308 3093 matthew.com Chris Bettiss Inflation-Linked Strategy +44 (0)20 7773 0836 chris.com Brian Upbin +1 212 526 6981 brian.maki@barcap.com Michalis Christodoulou +44 (0)20 3134 2751 michalis.sooben@ barcap.com Guilherme Loureiro Economist – Brazil +52 55 5509 3372 guilherme.com Index Products Scott Harman +44 (0) 20 7773 1775 scott.com Jimena Zuniga Economist – Chile.mazoy@barcap.com Tim Bond Head of Global Asset Allocation +44 (0)20 7773 2242 tim.com Jose Mazoy +44 (0) 20 3134 0998 jose.zuniga@barcap.pond@barcap.venkataraman@barcap.com 15 March 2010 .com Henry Skeoch Inflation-Linked Strategy +44 (0)20 7773 7917 henry.com Chirag Mirani US Fixed Income Strategy +1 212 412 6819 chirag.com Khrishnamoorthy Sooben Inflation-Linked Strategy +44 (0)20 7773 7514 khrishnamoorthy.skeoch@barcap.com Wai Ho Leong Economist – Korea. Mexico +1 212 412 5361 jimena.loureiro@barcap.myers@barcap.bond@barcap. Uruguay +54 (0)11 4850 1230 sebastian.badsha@absacapital.hewitt@barcap. Singapore.barreto@barcap.com Michael Pond US Fixed Income Strategy +1 (212) 412 5051 michael.com Guillermo Mondino Head of Latin American Research +1 212 412 7961 guillermo.harman@barcap.mirani@barcap.broda@barcap.james@barcap.com Sebastian Vargas Economist – Argentina.com Economics Julian Callow Chief European Economist +44 (0) 20 7773 1369 julian.

15 March 2010 . Recommendations contained in one type of research product may differ from recommendations contained in other types of research products. generally deals as principal and generally provides liquidity (as market maker or otherwise) in the debt securities that are the subject of this research report (and related derivatives thereof). and the potential interest of the firms investing clients in research with respect to. the firm's fixed income research analysts regularly interact with its trading desk personnel to determine current prices of fixed income securities.pl or call 212-526-1072. Where permitted and subject to appropriate information barrier restrictions. which may pose a conflict with the interests of investing customers. the profitability and revenues of the Fixed Income Division and the outstanding principal amount and trading value of. 17th Floor. NY 10019 or refer to https://ecommerce. Nick Verdi. some or all of which may have changed since the publication of this document. Jose Mazoy. but not limited to. the overall performance of the firm (including the profitability of the investment banking department). Chris Bettiss. hereby certify (1) that the views expressed in this research report accurately reflect our personal views about any or all of the subject securities or issuers referred to in this research report and (2) no part of our compensation was.com/research/cgibin/all/disclosuresSearch. Roberto Melzi. the profitability of. prices or spreads. Alan James. Wai Ho Leong. Any reference to Barclays Capital includes its affiliates. Marcela Barreto Rivera. Jenna Myers. Henry Skeoch. and trade ideas. the asset class covered by the analyst. Marcelo Salomon. the quality of their work. Julian Callow. the firm makes no representation that it is accurate or complete. whether as a result of differing time horizons. investors should be aware that Barclays Capital may have a conflict of interest that could affect the objectivity of this report.Barclays Capital | Global Inflation-Linked Products – A User’s Guide Analyst Certification(s) We. but not limited to. Bulent Badsha. Xiaonan Jiang. Yuan Tian. Michael Pond. Important Disclosures For current important disclosures regarding companies that are the subject of this research report. Jeff Gable. Barclays Capital does and seeks to do business with companies covered in its research reports. Christian Broda and Mimi Yang. Sebastian Vargas. Matthew Huang. equity-linked analysis. As a result. The firm's proprietary trading accounts may have either a long and / or short position in such securities and / or derivative instruments. Chirag Mirani. is or will be directly or indirectly related to the specific recommendations or views expressed in this research report. Jimena Zuniga. Arko Sen. prices and spreads are historical and do not represent current market levels.barcap. quantitative analysis. Guilherme Loureiro. Michalis Christodoulou. Barclays Capital produces a variety of research products including. Khrishnamoorthy Sooben. Barclays Capital and/or an affiliate thereof (the "firm") regularly trades. Christian Keller. The firm's fixed income research analyst(s) receive compensation based on various factors including. To the extent that any historical pricing information was obtained from Barclays Capital trading desks. All levels. please send a written request to: Barclays Capital Research Compliance. New York. Stefan Liiceanu. Anand Venkataraman. Rahul Sharma. Brian Upbin. Scott Harman. 745 Seventh Avenue. methodologies. Guillermo Mondino. or otherwise. Daniel Hewitt. fundamental analysis.

nor any of their respective officers. Barclays Bank PLC is registered in England No.. New York 10019. including persons involved in the preparation or issuance of this document. brokers or commercial and/or investment bankers in relation to the securities or related derivatives which are the subject of this publication. persons should contact and execute transactions through a Barclays Bank PLC branch or affiliate in their home jurisdiction unless local regulations permit otherwise. is distributing this material in South Africa. security or instrument discussed in this publication and consult any independent advisors they believe necessary. is distributing this material in the United States and. © Copyright Barclays Bank PLC (2010). partners. This publication is not. Subject to the conditions of this publication as set out above. and cannot be used. Please be advised that any discussion of U. No part of this publication may be reproduced in any manner without the prior written permission of Barclays Capital or any of its affiliates. including those of Barclays Capital and/or its affiliates. advice as defined and/or contemplated in the (South African) Financial Advisory and Intermediary Services Act. Barclays Capital recommends that investors independently evaluate each issuer. Tokyo 100-0004. Barclays Bank PLC Frankfurt Branch is distributing this material in Germany under the supervision of Bundesanstalt fuer Finanzdienstleistungsaufsicht (BaFin). Any U. deal in. Any South African person or entity wishing to effect a transaction in any security discussed herein should do so only by contacting a representative of Absa Capital in South Africa.S.org). the investment banking division of Barclays Bank PLC. legal. or any other financial. and Barclays Capital has no obligation to update its opinions or the information in this publication. nor any affiliate. The investments to which it relates are available only to such persons and will be entered into only with such persons.finra. partners and employees. Registered office 1 Churchill Place. New York. Additional information regarding this publication will be furnished upon request. foreign exchange research reports are prepared and distributed by Barclays Bank PLC Tokyo Branch. Sandton. Otemachi. and therefore should only be relied upon by. Registered Number: Kanto Zaimukyokucho (kinsho) No. co-manager or underwriter of a public offering or otherwise. and such opinions were prepared independently of any other interests. The information in this publication is not intended to predict actual results. E14 5HP. In Japan. The analyst recommendations in this report reflect solely and exclusively those of the author(s). All rights reserved. the information contained in this publication has been obtained from sources that Barclays Capital believes to be reliable.S. 37 of 2002. Japan. nor is it intended to be.: 1986/004794/06). This material is distributed in Malaysia by Barclays Capital Markets Malaysia Sdn Bhd. directors. Absa Capital. tax. Barclays Capital Inc.S. accounting. This communication is being made available in the UK and Europe to persons who are investment professionals as that term is defined in Article 19 of the Financial Services and Markets Act 2000 (Financial Promotion Order) 2005. Neither Barclays Capital. in connection therewith accepts responsibility for its contents. This publication is provided to you for information purposes only. but Barclays Capital does not represent or warrant that it is accurate or complete. Johannesburg. Chiyoda-ku. or employees accepts any liability whatsoever for any direct or consequential loss arising from any use of this publication or its contents. The views in this publication are those of Barclays Capital and are subject to change.S. directors. actuarial or other professional advice or service whatsoever. The value of and income from any investment may fluctuate from day to day as a result of changes in relevant economic markets (including changes in market liquidity). an authorised financial services provider (Registration No. the Investment Banking Division of Absa Bank Limited. Gauteng 2196. Prices shown in this publication are indicative and Barclays Capital is not offering to buy or sell or soliciting offers to buy or sell any financial instrument. and (ii) was written to support the promotion or marketing of the transactions or other matters addressed herein. Barclays Capital and its affiliates and their respective officers. Other research reports are distributed to institutional investors in Japan by Barclays Capital Japan Limited. 143. in the U. London. investment.Barclays Capital | Global Inflation-Linked Products – A User’s Guide This publication has been prepared by Barclays Capital. in the capacity of principal or agent. tax-related penalties. persons who have professional experience in matters relating to investments. It is directed at.S. may from time to time act as manager. 15 March 2010 . tax matters contained herein (including any attachments) (i) is not intended or written to be used. Past performance is not necessarily indicative of future results. The securities discussed in this publication may not be suitable for all investors. US registered broker/dealer and member of FINRA (www. Absa Capital is an affiliate of Barclays Capital. Barclays Capital is authorized and regulated by the Financial Services Authority ('FSA') and member of the London Stock Exchange. Barclays Capital Japan Limited is a joint-stock company incorporated in Japan with registered office of 2-2-2. retirement. by you for the purpose of avoiding U. 15 Alice Lane. Non-U. Accordingly. at 745 Seventh Avenue. person wishing to effect a transaction in any security discussed herein should do so only by contacting a representative of Barclays Capital Inc. you should seek advice based on your particular circumstances from an independent tax advisor. hold or act as market-makers or advisors. and/or one or more of its affiliates as provided below. It is a subsidiary of Barclays Bank PLC and a registered financial instruments firm regulated by the Financial Services Agency of Japan. Other than disclosures relating to Barclays Capital. IRS Circular 230 Prepared Materials Disclaimer: Barclays Capital and its affiliates do not provide tax advice and nothing contained herein should be construed to be tax advice. 1026167. trading. which may differ substantially from those reflected. Absa Bank Limited is regulated by the South African Reserve Bank.