US Economics Analyst

Issue No: 11/36 September 9, 2011

FOR THOSE PERMISSIONED: Goldman Sachs Global ECS Research at https://360.gs.com

The Likelihood of Additional Fed Easing

Given the economys lack of momentum and dovish Fed communication, we expect the Federal Open Market Committee (FOMC) to provide additional monetary easing at the September 20-21 meeting. We think that the most likely easing step is a lengthening of the average maturity of the Feds balance sheet with sales of relatively short-dated Treasuries and purchases of relatively long-dated Treasuries (the twist, for short). Our Taylor rule has been a useful tool in forming our view that additional monetary stimulus is needed. However, it offers little guidance for the likelihood of easing at a particular meeting. To obtain such an estimate we construct a new model that explains the meeting-by-meeting probability of federal funds rate changes using the Fed staff forecasts and a number of details from official Fed communication. We find that that past FOMC decisions are well explained with the following six factors: (1) the policy move at the previous meeting; (2) the extent to which the policy stance deviated from the Taylor rule going into the meeting; (3) whether the FOMC adopted a policy bias at the previous meeting; (4) whether the meeting spanned one day or two; (5) the Fed staff forecast for the unemployment rate; and (6) the number (and direction) of dissents at the previous meeting. Additional easing at the September meeting is very likely. Our model suggests that a small easing step such as a moderate version of the twist is the base case, but a more aggressive step is also possible. We revised up our Q3 GDP growth forecast from 1% to 2% (annual rate) on the back of better-than-expected trade and consumer spending data. Our New Model Captures Most Fed Decisions
Probability of Fed Tightening or Easing at the Next Meeting (percent): Actual 100 80 60 40 20 0 -20 -40 -60 -80 -100 94 95 96 97 98 99 00 01 02 03 04 05 Source: Federal Reserve Board. GS Global ECS Research. Tightening (lef t) Easing (right, inverted) 100 80 60 40 20 0 -20 -40 -60 -80 -100

Jan Hatzius jan.hatzius@gs.com 212 902 0394 Zach Pandl zach.pandl@gs.com 212 902 3393 Alec Phillips alec.phillips@gs.com 202 637 3746 Sven Jari Stehn jari.stehn@gs.com 212 357 6224 Andrew Tilton andrew.tilton@gs.com 212 357 2619 Shuyan Wu shuyan.wu@gs.com 212 902 3053 Maria Acosta-Cruz maria.acosta-cruz@gs.com 212 902 6709

And Suggests Easing in September is Likely

Percent 100 90 80 70 60 50 40 30 20 10 0 Baseline And Policy Gap And Aug Easing And Easing Bias And 2And Day Mtg Dissents Probability of Fed Easing at the September Meeting: Small Easing Large Easing Percent 100 90 80 70 60 50 40 30 20 10 0

Source: Federal Reserve Board. GS Global ECS Research.

Important disclosures appear at the back of this document.

GS Global ECS US Research

US Economics Analyst

I. The Likelihood of Additional Fed Easing

Given the economys lack of momentum and dovish Fed communication, we now expect the Federal Open Market Committee (FOMC) to provide additional monetary easing at the September 20-21 meeting. We think that the most likely easing step is a lengthening of the average maturity of the Feds balance sheet with sales of relatively short-dated Treasuries and purchases of relatively long-dated Treasuries (the twist for short).1 Larger stepssuch as a sizable expansion of the Feds balance sheet or communication changes signaling that rates are on hold for even longer than mid-2013appear less likely for now but might well be required eventually.2 Exhibit 1: Our Taylor Rule Suggests More Easing is Needed
Percent 12 10 8 6 4 2 0 -2 -4 -6 -8 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 Source: Federal Reserve Board. Our calculations. Actual Baseline (GS f orecasts) With Unconventional Policy (GS f orecasts) Warranted Federal Funds Rate Implied by "f orward-looking" GS Taylor Rules: Forecast Percent 12 10 8 6 4 2 0 -2 -4 -6 -8

Sluggish Growth + Dovish Fed = More Easing

Our motivation for this expectation is twofold. First, our version of the Taylor rule continues to call for additional easing under our economic outlook, even when adjusted for the effects of the Feds unconventional policies (Exhibit 1). Specifically, our rule suggests that the federal funds rate should be around 175 basis points (bp) lower than it currently is, even if we take into account the Feds unconventional measures already in place. Second, recent Fed communication suggests that the FOMC is in easing mode. At the August meeting the committee announced the conditional commitment to keep the funds rate on hold until at least mid-2013. The statement furthermore contained an easing bias which noted that the FOMC is prepared to employ [easing] tools as appropriate. Since then Chairman Bernanke announced at the Jackson Hole conference that the September meeting would span two days to to allow a fuller discussion of the easing options. Taken together, our Taylor rule and Fed communication provide a reason to expect additional easing steps at the September meeting. But can these factors give us a sense as to how likely additional easing is? And can we say anything about the likelihood that the FOMC will take a small step (e.g. twist) or opt for a larger step in September? In todays comment we construct a framework to attempt to answer these questions. Our analysis suggests that easing at the September meeting is very likely (with an estimated probability of around 75%), but that the likelihood of a small easing step (around

60%) is much higher than the probability of more aggressive action (around 15%).

From Taylor Rules to the Likelihood of Easing

Our QE-adjusted forward-looking Taylor rule has been a useful tool for thinking about the appropriate stance of monetary policy and the need for additional easing to realign the actual with the appropriate policy stance.3 However, by describing what monetary policy should do given historical behavior, the Taylor rule offers little guidance for the timing of policy changes and thus for the likelihood of easing at a particular meeting. This is especially true when thinking about policy moves immediately ahead, as our Taylor rule is estimated at a quarterly frequency rather than on a meeting-by-meeting basis. We therefore build a new framework that allows estimating the probability of Fed easing at the next meeting. We proceed in three steps. 1. We compile a meeting-by-meeting dataset for funds rate changes. To track monetary policy decisions meeting by meeting, we consider changes to the target federal funds rate at each of the eight annual scheduled FOMC meetings since 1994. We choose to start our sample in 1994, because the federal funds rate was previously commonly changed in between meetings. As all funds rate changes have been in multiples of 25bp since 1994, we classify Fed policy
3

For our latest thoughts on the twist see Zach Pandl, Doing the Twist, US Daily, September 8, 2011. For a discussion of the second of these options see Jan Hatzius, The Evans Proposal to Rebalance the Feds Dual Mandate, US Daily, September 7, 2011.

For details of our rule see Sven Jari Stehn, What Do Taylor Rules Say About the Threshold for Monetary Easing? US Daily, July 25, 2011.
September 9, 2011

Issue No: 11/36

GS Global ECS US Research

US Economics Analyst

Exhibit 2: Bucketing Fed Policy Moves

100 90 80 70 60 50 40 30 20 10 0 Large Easing (50bp) Small Easing (25bp) No Change Small Tightening (25bp) Large Tightening (50bp) Number of Funds Rate Changes at FOMC Meetings since 1987: 2006-2011 1994-2005 100 90 80 70 60 50 40 30 20 10 0
-1.0 -1.5 -2.0 94 0.0 -0.5 1.0 0.5

Exhibit 3: Staff Forecasts Are No Worse than the FOMCs

Percentage points 1.5 Percentage points 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 96 98 00 02 04 06

Unemploymert Rate Forecast Error*: Staf f FOMC

Source: Federal Reserve Board. GS Global ECS Research.

* Actual minus f our-quarter-ahead unemployment rate f orecast. Source: Federal Reserve Board. GS Global ECS Research.

moves into five categories: large easing (50bp cuts or more), small easing (25bp), no change, small tightening (25bp hike) and large tightening (50bp or more). At the 141 scheduled meetings since 1994, the Fed left the funds rate unchanged 88 times (Exhibit 2). The funds rate was cut at 23 scheduled meetingsin 12 small steps and 11 large steps. Policy was tightened 30 times, through 25 small and 5 large hikes in the funds rate. 2. We use the Fed staff forecasts to capture the economic outlook. Our forward-looking Taylor rule suggests that expected economic conditions are the key driver of monetary policy. Unfortunately, we cannot link meeting-by-meeting decisions to the FOMC forecasts, because these are currently only published four times a year (and twice a year before 2007). We therefore turn to the staff forecasts which are prepared in the Greenbook for each FOMC meeting. Drawbacks of using the staff forecasts are that (1) policy decisions are taken by the FOMC and not the staff; and (2) the Greenbooks are published with a five-year delay, forcing us to restrict our analysis of meeting-by-meeting funds rate changes to end in 2005. That said, it is not clear that the staff forecasts are any less accurate than the FOMCs. The economists Christina and David Romer have shown that the FOMC forecasts do not contain useful information relative to the staffs projections (which are available to FOMC members at the time).4 They conclude that Someone wishing to predict inflation and unemployment who had access to both the FOMC and staff forecasts would be well served by discarding the
4

FOMC forecast and just using the staff predictions. As a simple test of this view Exhibit 3 plots the forecast error in predicting the four-quarter-ahead unemployment rate for the FOMC and the staff.5 Consistent with the discussion above, there is not much difference in forecasting accuracy. As a check of how much information the staff forecasts hold for policy actions, we estimate a meeting-by-meeting version of our Taylor rule between 1994 and 2005 using the staff projections for the unemployment gap and core inflation. We find that the meeting-by-meeting rule looks very similar to our forward-looking Taylor rule (Exhibit 4), suggesting that the staff forecasts hold useful information for gauging the appropriate stance of monetary policy.6 Moreover, the meeting-by-meeting rule confirms that the Feds current stance of monetary policy is too tight: adjusting for the Feds unconventional policies, and feeding in what we think are reasonable guesses for the staff forecasts ahead of the September meeting, suggests that the policy gap is currently around 125bp. (This is similar to but a bit less than the 175bp policy gap implied by our forward-looking Taylor rule under our forecasts.)

See The FOMC versus the Staff: Where Can Monetary Policymakers Add Value? American Economic Review Papers & Proceedings, 2008.

Note that we needed to interpolate the FOMC forecasts as only end-year forecasts are published. The meeting-by-meeting rule differs in two regards. First, the unemployment gap matters most at its current perceived value, while the core inflation rate has most explanatory power four-quarters ahead (in our forward-looking Taylor rule both enter fourquarters-ahead). Second, the coefficients in the meeting-by-meeting rule are smaller in absolute value, consistent with the idea that the FOMC acts more aggressively in response to its own forecasts. September 9, 2011

Issue No: 11/36

GS Global ECS US Research

US Economics Analyst

Exhibit 4: Staff Forecasts Help Predict Funds Rate Changes

Percent 7 6 5 4 3 2 1 0 94 96 98 00 02 04 06 Source: Federal Reserve Board. GS Global ECS Research. Actual Implied by Meeting-by-Meeting Rule 0 -0.75 94 Percent 7 6 5 0.25 4 0.00 3 -0.25 2 1 -0.50 0.75

Exhibit 5: Policy Biases Signal Policy Changes

Percentage points

0.50

Policy Bias Funds Rate Change

95

96

97

98

99

00

01

02

03

04

05

Source: Federal Reserve Board. GS Global ECS Research.

3. We use information contained in official FOMC communication. Fed watchers typically look out for at least two additional pieces of information in FOMC statements: a policy bias and dissents. Until 2000 a history of policy biases is easily collected as it was explicitly expressed in the post-meeting policy directive written to convey the committees instructions to the manager of the System Open Market Account for the conduct of monetary policy.7 Between 2000 and 2005 we construct a policy bias ourselves from the discussion of economic risks from FOMC statements.8 Exhibit 5 shows the resulting time series alongside changes in the funds rate between 1994 and 2005. The expression of a policy bias is clearly strongly associated with corresponding moves in monetary policy: easing biases are typically followed by funds rate cuts (and vice versa). Dissents from the FOMC decision likewise tend to receive a lot of attention. To include these into our analysis we collect a history of dissents from the FOMC statements. Moreover, we track whether the dissents were favoring more policy tightening (hawkish dissents) or easing (dovish dissents) than the FOMC decision. Exhibit 6 shows that hawkish dissents occurred more frequently than dovish ones. While one dissent is common, we have only seen three dissents four times since 1987, most recently at the August meeting.

Exhibit 6: Dissents are Common

Number of dissents 3 Number of dissents 3 Number of Dissents: Dovish Hawkish 2

-1

-2

-3 87 89 91 93 95 97 99 01 03 05 07 09 11 Source: Federal Reserve Board. GS Global ECS Research.

Gauging the Determinants of Fed Moves

Using these inputs we then estimate a model which explains the probability of Fed policy moves between 1994 and 2005.9 We identify six factors that contain statistically significant information for explaining Fed moves. These include: (1) the policy move at the previous meeting; (2) the policy gap going into the meeting; (3) whether the FOMC had a policy bias at

We take these data from David Payne, Anticipating Monetary Policy with the Federal Reserves Beige Book, http://www.nabe.com/am2000/payne.pdf. For example, we interpret the FOMCs outlook description in 2001 (the risks are weighted mainly toward conditions that may generate economic weakness in the foreseeable future) as an easing bias.

Specifically, we construct an ordered probit model which models the probability of the five discrete changes in the funds rate in Exhibit 2. This approach follows Michael Dueker, Measuring Monetary Policy Inertia in Target Fed Funds Rate Changes, Federal Reserve Bank of St. Louis Review, 1999. September 9, 2011

Issue No: 11/36

Easing Bias
1 0 -1 -2 -3

Tightening Bias

GS Global ECS US Research

US Economics Analyst

Exhibit 7: The Model Captures Most Policy Moves, Especially Easing Steps
Probability of Fed Tightening or Easing at the Next Meeting (percent): Actual 100 80 60 40 20 0 -20 -40 -60 -80 -100 94 95 96 97 98 99 00 01 02 03 04 05 Source: Federal Reserve Board. GS Global ECS Research. Tightening (lef t) Easing (right, inverted) 100 80 60 40 20 0 -20 -40 -60 -80 -100

Exhibit 8: Easing in September is Likely

Percent 100 90 80 70 60 50 40 30 20 10 0 Baseline And Policy Gap And Aug Easing And Easing Bias And 2And Day Mtg Dissents Probability of Fed Easing at the September Meeting: Small Easing Large Easing Percent 100 90 80 70 60 50 40 30 20 10 0

Source: Federal Reserve Board. GS Global ECS Research.

the previous meeting; (4) the length of the meeting;10 (5) the staff forecast for the four-quarter ahead change in the unemployment rate; and (6) the number (and direction) of dissents at the previous meeting.11 Before turning to the implications of the model for the September meeting, Exhibit 7 presents an evaluation of the forecasting performance of the model between 1994 and 2005. The shaded areas denote policy action (tightening when above the horizontal axis and easing when below) and the lines plot the estimated probability from our model. First, we see that the model does a decent job in capturing the main policy moves. The model provides a few false alarms (e.g. called for a tightening in July 1997) and fails to predict some moves (e.g. the February 2000 hike). But the model correctly predicts 25 out of 41 policy moves that occurred. Second, we note that the model is better at capturing easing moves than funds rate increases. Specifically, the model correctly predicts 10 out of 15 funds rate cuts (and 4 out of 6 large cuts) but only 15 out of 26 tightening steps (none of the 5 large ones, for some of which it predicted small hikes).

unchanged unemployment rate and 1.6% core PCE year-on-year inflation in four quarters time. These assumptions imply that the staff currently perceives a policy gap of 125bp. (Our own forecasts are for a 9.4% unemployment rate and 1.4% core PCE inflation in 2012Q3, implying a larger policy gap.) Combining our model with these assumptions suggests that the probability of additional easing at the September meeting is 75%. Our model further suggests that a small easing step is much more likely (with an estimated probability of 60%) than a more aggressive step (with a probability of 15%). We explain in two steps how our model constructs this estimate. First, we start from a baseline scenario which assumes an unchanged unemployment rate, and no policy gap, no easing bias etc., and explore how the probability of easing is affected by adding each of the factors discussed above (Exhibit 8; each bar displays the resulting probability of easing):12 1. Policy gap. As we have argued with our Taylor rule for a while, a gap between the warranted and actual funds rate raises the probability of additional easing. For a 125bp policy gap, the probability of a small easing step rises to 20%. 2. August policy move. Assuming that the FOMCs strengthening of the guidance language at the last meeting was worth as much as a 25bp funds rate cut, the probability of a small easing step rises to about 45% and the likelihood of a large easing step emerges (although with a likelihood of 4%).
12

Our Model Suggests

With this model at hand we turn to gauging the probability of easing at the September meeting. To do so, we need to make an assumption for the staff forecast of unemployment and inflation. In our view a reasonable guess is that the staff currently expects an
10

11

We model the effect of a two-day meeting by interacting a dummy variable with the policy gap. The idea is that two-day meetings allow for a fuller discussion of policy moves that close the policy gap. The estimated model is qualitatively robust to (1) extending the sample back to 1987 and (2) including inter-meeting funds rate changes into the analysis.

Note that the changes in the easing probability we cite below only apply if the explanatory variables are added to the baseline scenario in this order (because the model is non-linear). September 9, 2011

Issue No: 11/36

GS Global ECS US Research

US Economics Analyst

3. August easing bias. The adoption of the easing bias at the last meetingstating that the committee is prepared to employ [easing] tools as appropriatesharply raises the probability of action at the September meeting to 55% for a small easing step and 35% for a large move. 4. Two-day meeting. Our model suggests that twoday meetings make it more likely that the FOMC closes any remaining policy gap. The intuition is that two-day meetings allow for a fuller discussion of the policy options. The FOMCs decision to spread the September meeting over two days dramatically raises the prospect of a large easing step. 5. Three dissents. Taking into account the three hawkish dissents at the August meetingFed presidents Plosser, Kocherlakota and Fisher lowers the prospects of a large easing move. Our simulations suggest that Fed communication contains key information for anticipating future Fed moves (or that the Fed walks the talk). This is of course exactly the point of conducting monetary policy in a transparent fashion. Fed watchers are therefore right to pay close attention to the details of FOMC communication. Second, we explore the sensitivity of the model estimate to the staff forecast of the unemployment rate (Exhibit 8).13 For example, the model suggests that the likelihood of further easing rises to 90% assuming a 0.3 point increase in the unemployment rate over the coming year (as we currently do). In particular it is the likelihood of large easing steps that rises sharply when the economic outlook is expected to deteriorate. But even for expected declines in the unemployment rate, the probability of easing remains large.

Exhibit 9: Especially With Our Forecasts

Percent 100 Probability of Fed Easing at 90 the September Meeting: 80 70 60 50 40 30 20 10 0 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 Projected Increase in Unemployment Source: Federal Reserve Board. GS Global ECS Research. Fed Staf f (GS est.) Small Easing Large Easing Current GS Forecast Percent 100 90 80 70 60 50 40 30 20 10 0

probability easing by placing too much weight on the three hawkish dissents at the August meeting. The reason is that we suspect that forward-looking dissents (which favor policy action which was not taken) matter more for the future course of policy than backward-looking dissents (which oppose action that was taken, like the dissents at the August meeting).15 By not making this distinction the model might place too much weight on the August dissents. In the current instance we would put less weight on the dissents and believe that the probability of easing at the September meeting is higher than 75%. Our analysis further predicts that a small easing step at the September meeting is likely, while a large move is a possibility. What form could these steps take? As our model treats small (large) easing steps as 25bp (50bp) cuts in the funds rate, we need to map these conventional steps into unconventional ones. Our previous work in this area suggests that a small(ish) step could either consist of a twist or an equivalent expansion of the balance sheet.16 A large step would require a more sizable expansion of the balance sheet (to the tune of QE2 or more).

Easing in September is Very Likely

The bottom line from our analysis is that easing at the September meeting is very likelyaround 75% according to our model. How reasonable is this model prediction? On the one hand, we see reason to believe that the model might over-predict the probability of unconventional easing because it is constructed to predict changes in the funds rate. As unconventional policies are perceived to carry larger costs than conventional ones, the true likelihood of action in September might be smaller than our model predicts.14 On the other hand, our setup might under-predict the
13

Sven Jari Stehn

15

14

Note that we hold constant the four-quarter ahead core inflation rate for simplicity in this simulation. For a discussion see No Rush to the Exit, Global Economics Paper, No. 200.

16

Also, the model is estimated during a period (19942005) which never actually saw three dissents (two was the maximum number). See For More Easing, Will Fed Go Big or Go Long? US Daily, August 15, 2011. September 9, 2011

Issue No: 11/36

GS Global ECS US Research

US Economics Analyst

THE US ECONOMIC AND FINANCIAL OUTLOOK

(% change on previous period, annualized, except where noted) 2010 OUTPUT AND SPENDING
Real GDP Year-to-year change Consumer Expenditure Residential Fixed Investment Business Fixed Investment Federal Government State and Local Government Net Exports ($bn, '05) Inventory Investment ($bn, '05) Industrial Production, Mfg 3.0 2.0 -4.3 4.4 4.5 -1.8 -421.8 58.8 5.4 1.6 1.0 1.4 -2.0 9.6 0.18 0.30 0.62 1.93 3.29 4.42 27.5 -1,294 -3.2 1.32 83 1.6 2.1 -1.8 7.2 -1.9 -2.3 -413.5 47.7 4.1 3.1 1.6 1.5 1.5 9.1 0.10 0.30 0.30 1.25 2.75 3.75 5.8 -1,275 -3.2 1.50 76 2.0 1.7 5.1 3.1 -1.1 -0.5 -351.2 72.5 1.7 2.0 1.6 1.5 0.0 9.4 0.10 0.30 1.00 2.50 3.50 4.25 2.5 -1,025 -2.9 1.55 74 0.4 2.2 2.1 -2.4 2.1 -9.4 -3.4 -424.4 49.1 7.1 2.2 1.1 1.1 1.4 8.9 0.14 0.31 0.70 2.11 3.41 4.51 10.6
_

2011 (f)

2012 (f)

Q1

2011 Q2 Q3
1.0 1.5 0.4 3.4 9.9 2.0 -2.8 -421.3 40.6 0.9 3.3 1.5 1.3 1.9 9.1 0.09 0.25 0.41 1.58 3.00 4.23 9.2
_

Q4
1.5 1.2 1.5 2.5 2.5 -1.0 -1.0 -400.6 54.0 1.0 3.4 2.1 2.0 1.4 9.3 0.10 0.30 0.30 1.25 2.75 3.75 2.0
_

Q1
2.0 1.6 1.5 5.0 0.0 -1.5 0.0 -379.0 66.0 1.5 2.6 2.1 1.9 0.7 9.4 0.10 0.30 0.40 1.50 3.00 4.00 2.5
_

2012 Q2 Q3
2.5 2.0 2.0 5.0 2.5 -1.5 0.0 -356.7 74.0 2.0 2.0 1.8 1.7 -0.1 9.4 0.10 0.30 0.50 2.00 3.00 4.10 2.5
_

Q4
2.5 2.4 2.0 10.0 5.0 -1.5 1.0 -327.0 74.0 2.0 1.7 1.3 1.2 -0.3 9.4 0.10 0.30 1.00 2.50 3.50 4.25 2.5
_

2.0 1.4 1.8 3.0 5.0 -1.0 -2.5 -407.9 47.0 2.5 3.5 1.9 1.7 1.4 9.1 0.10 0.30 0.20 1.00 2.50 3.60 2.0
_

2.5 2.1 2.0 10.0 5.0 -1.5 1.0 -342.1 76.0 2.0 1.9 1.4 1.3 -0.3 9.4 0.10 0.30 0.75 2.25 3.25 4.20 2.5
_

INFLATION (% ch, yr/yr)

Consumer Price Index (CPI) Core CPI Core PCE* Unit Labor Costs

LABOR MARKET
Unemployment Rate (%)

FINANCIAL SECTOR
Federal Funds** (%) 3-Month LIBOR (%) Treasury Yield Curve** (%) 2-Year Note 5-Year Note 10-Year Note 30-Year Bond Profits*** (% chg, yr/yr) Federal Budget (FY, $ bn)

FOREIGN SECTOR
Current Account (% of GDP) Euro ($/)** Yen (/$)** -3.2 1.40 82 -3.2 1.44 80 -3.1 1.45 77 -3.2 1.50 76 -3.0 1.53 75 -2.9 1.55 74 -2.9 1.55 74 -2.8 1.55 74

* PCE = Personal consumption expenditures. ** Denotes end of period. *** Profits are after taxes as reported in the national income and product accounts (NIPA), adjusted to remove inventory profits and depreciation distortions. NOTE: Published figures are in bold
We, Jan Hatzius, Zach Pandl, Alec Phillips, Sven Jari Stehn and Andrew Tilton hereby certify that all of the views expressed in this report accurately reflect personal views, which have not been influenced by considerations of the firms business or client relationships. Global product; distributing entities The Global Investment Research Division of Goldman Sachs produces and distributes research products for clients of Goldman Sachs, and pursuant to certain contractual arrangements, on a global basis. Analysts based in Goldman Sachs offices around the world produce equity research on industries and companies, and research on macroeconomics, currencies, commodities and portfolio strategy. This research is disseminated in Australia by Goldman Sachs & Partners Australia Pty Ltd (ABN 21 006 797 897) on behalf of Goldman Sachs; in Brazil by Goldman Sachs do Brasil Banco Mltiplo S.A.; in Canada by Goldman Sachs & Co. regarding Canadian equities and by Goldman Sachs & Co. (all other research); in Hong Kong by Goldman Sachs (Asia) L.L.C.; in India by Goldman Sachs (India) Securities Private Ltd.; in Japan by Goldman Sachs Japan Co., Ltd.; in the Republic of Korea by Goldman Sachs (Asia) L.L.C., Seoul Branch; in New Zealand by Goldman Sachs & Partners New Zealand Limited on behalf of Goldman Sachs; in Russia by OOO Goldman Sachs; in Singapore by Goldman Sachs (Singapore) Pte. (Company Number: 198602165W); and in the United States of America by Goldman Sachs & Co. Goldman Sachs International has approved this research in connection with its distribution in the United Kingdom and European Union. European Union: Goldman Sachs International, authorized and regulated by the Financial Services Authority, has approved this research in connection with its distribution in the European Union and United Kingdom; Goldman Sachs AG, regulated by the Bundesanstalt fur Finanzdienstleistungsaufsicht, may also distribute research in Germany. General disclosures This research is for our clients only. Other than disclosures relating to Goldman Sachs, this research is based on current public information that we consider reliable, but we do not represent it is accurate or complete, and it should not be relied on as such. We seek to update our research as appropriate, but various regulations may prevent us from doing so. Other than certain industry reports published on a periodic basis, the large majority of reports are published at irregular intervals as appropriate in the analyst's judgment. Goldman Sachs conducts a global full-service, integrated investment banking, investment management, and brokerage business. We have investment banking and other business relationships with a substantial percentage of the companies covered by our Global Investment Research Division. SIPC: Goldman, Sachs & Co., the United States broker dealer, is a member of SIPC (http://www.sipc.org). Our salespeople, traders, and other professionals may provide oral or written market commentary or trading strategies to our clients and our proprietary trading desks that reflect opinions that are contrary to the opinions expressed in this research. Our asset management area, our proprietary trading desks and investing businesses may make investment decisions that are inconsistent with the recommendations or views expressed in this research. We and our affiliates, officers, directors, and employees, excluding equity and credit analysts, will from time to time have long or short positions in, act as principal in, and buy or sell, the securities or derivatives, if any, referred to in this research. This research is not an offer to sell or the solicitation of an offer to buy any security in any jurisdiction where such an offer or solicitation would be illegal. It does not constitute a personal recommendation or take into account the particular investment objectives, financial situations, or needs of individual clients. Clients should consider whether any advice or recommendation in this research is suitable for their particular circumstances and, if appropriate, seek professional advice, including tax advice. The price and value of investments referred to in this research and the income from them may fluctuate. Past performance is not a guide to future performance, future returns are not guaranteed, and a loss of original capital may occur. Fluctuations in exchange rates could have adverse effects on the value or price of, or income derived from, certain investments. Certain transactions, including those involving futures, options, and other derivatives, give rise to substantial risk and are not suitable for all investors. Investors should review current options disclosure documents which are available from Goldman Sachs sales representatives or at http://www.theocc.com/publications/risks/riskchap1.jsp. Transactions cost may be significant in option strategies calling for multiple purchase and sales of options such as spreads. Supporting documentation will be supplied upon request. All research reports are disseminated and available to all clients simultaneously through electronic publication to our internal client websites. Not all research content is redistributed to our clients or available to third-party aggregators, nor is Goldman Sachs responsible for the redistribution of our research by third party aggregators. For all research available on a particular stock, please contact your sales representative or go to www.360.gs.com. Disclosure information is also available at http://www.gs.com/research/hedge.html or from Research Compliance, 200 West Street, New York, NY 10282. No part of this material may be (i) copied, photocopied or duplicated in any form by any means or (ii) redistributed without the prior written consent of The Goldman Sachs Group, Inc. Copyright 2011, The Goldman Sachs Group, Inc. All Rights Reserved.

Issue No: 11/36

September 9, 2011

US Calendar
Focus for the Week Ahead

Retail sales likely held up well last month despite weak consumer confidence. We forecast an increase of 0.4% (month-over-month) in total sales and a gain of 0.2% in core sales (September 14). The early-September business surveys deserve close attention, due to the marked deterioration in these measures during August. We look for a rebound in the Philadelphia Fed index (September 15). The core CPI likely rose by 0.17% (month-over-month) in August. We look for another sizable increase in rent-related prices, partly offset by lower prices for cars and trucks (September 15).
Time (EST) 16:00 20:00 8:30 14:00 8:30 8:30 Estimate GS Consensus Last Report

Date Mon Sep 12 Tue Sep 13

Wed Sep 14

Thu

Sep 15

8:45 10:00 8:30

Fri

Sep 16

8:30 8:30 8:30 8:30 8:45 9:15 9:15 10:00 13:45 9:00 9:55 12:00

Indicator Dallas Fed Pres Fisher spks at 2011 NABE annual meeting St Louis Fed Pres Bullard spks at Fed event; St Louis Import & Export Prices (Jul) Federal Budget Balance (Aug) Producer Price Index (Aug) Ex Food & Energy Retail Sales (Jul) Ex Autos Ex Autos, Bldg Materials & Gas Bernanke gives opening remarks at Fed Boards risk conf Business Inventories (Jul) Consumer Price Index (Aug) Ex Food and Energy Consumer Price Index NSA Current Account Balance (Q2) Empire Manufacturing Survey (Sep) Initial Jobless Claims Continuing Claims Bernanke gives opening remarks at Fed Boards risk conf Industrial Production (Aug) Capacity Utilization (Aug) Philadelphia Fed Survey (Sep) Fed Gov Tarullo spks at Fed Boards risk conference Net Long-Term TIC Data (Jul) Reuters/U. Mich Consumer SentimentPrel (Sep) Flow of Funds

n.a. -0.8% n.a. -$132.0bn Flat -0.1% +0.2% +0.2% +0.3% +0.2% +0.4% +0.2% +0.4% +0.3% n.a. +0.22% +0.17% 226.140 n.a. n.a. n.a. +0.2% 77.6% -5.0

+0.3% -$90.5bn +0.2% +0.4% +0.5% +0.5% +0.3%

+0.5% +0.3% +0.2% +0.5% +0.2% +0.2% 226.290 225.922 -$122.9bn -$119.3bn -4.0 -7.7 411,000 414,000 3,705,000 3,717,000 +0.1% 77.5% -15.0 +0.9% 77.5% -30.7 $3.7bn 55.7

n.a.

56.6

Issue No: 11/36

September 9, 2011