122 Chapter 3. Foreign Exchange Determination and Forecasting 18, 1 MB46¢ zs 0 2 14300 43958 a 3152750 158600151 4 osm 494008 5 120800 120700 130 6 13950 9800 TST Solutions 1 2 . Which forecast(s) correetly predicted the depreciation of Japanese yen relative to the dollar? . David Brock and Brian Lee are speculators who buy and sell currencies forward David Brock took a position based on Commerzbank’s forecast for the yen io dollar ‘exchange rate while Brian Lee took a position based on Harri Bank's forecast. Who tummed out:o be better off? 4. Discussif there is any conflict in your answers to parts (a) and (c) ‘The following table contains Japanese yen ‘0 dollar exchange rate data that were pub- lished in issues of the Risk magazine. Each period is three months. Spol rates the actual spot exchange rate prevailing atthe startof a period. Forward rates the three-month for- ward exchange rate prevailing at:he start of a period. Forecast rates the forecast made by the Industrial Bank of fapan at the start ofa period for the spot exchange rate at the start ofthe next period (ihat i, the forecast for three months later) To illustrate, at the beginning of the third period, the actual spot exchange rate was 152.750, the three- month ahead forward rate was 153.600, and the rate forecast by the Industrial Bank for the start of the fourth period was 151. The actual spot exchange rate that was real- ized at the start of the fourth period was 149,400, Based on :he root mean squared error, was the Industrial Bank of Japan able to outperform the forward rate (do the zal- culations for the percentage forecast error)? You are also given that the spot rate real ized three months after the last forecast given in the table was 159.25, Period Spot Forward Forecast Rate Rate Rate Applying expansionary macroeconomic policy, which results in higher goods prices and lower rea) interest rates, wll not reduce the balance of payments deficit. Higher prices will make the country’s goods less competitive internationally, and lower interest rates will discourage foreign capital. Thus, the balance of payments deficit will Worsen instead ‘of improve. On the ather hand, ‘a),(b), and ‘d) will help in remedying the balance af payments deficit. Accordingly the answer isc), 4. One advantage a wider bund is “emotional.” France could claim that it diel non Ae. value is €urFene Another advantage is flexibility If there Wf long-term fund mental FeasonS (inflation, balance of payments deficit, ete.) for devaluation, the temporary prewure on the franc could #88@ and the franc could later revert its previous level, The disadvantage ofa wider band is exchange rate uncertainty for all firms A ‘credible" small band s preferable for firms conducting international trade b, Awider band makes speculation less attractive, because there is ho guarantee shat a central bank wil defend is currency until the wide fluctuation margin is reachedSolutions 123 3. Remember that the Eurozone is made up of those countries in the EU that have adopted the curo as common currency. Statements ( and UI ure clearly correct, Statement Ili clearly not correct, because there isa possibilty that more countries may join the Eurozone in future, Kor example, the British are debating whether to join the Eurozone. 4, The statement is rue. Because ofriskless arbitrage, interest rate parity between two cur rencies would hold if the markets for both are free and deregulated. Developed finan= cial markets tend to be more free and deregulated. A developing country is more likely to impose various forms of capital controls and taxes that impede arbitrage. A develop ing country is economically not as well integrated with the world Fmancial markets #8 the developed markets are, Also, some smaller currencies can only be borrowed and lent domestically, and the domestic money markets of developing countries are more ‘kely to be subjected t0 political risk. 5. a Using the firscorder approximation of PPP relationship, the variation in rupee to dollar exchange rate should equal the inflation differential between rupee and do lar, So, the rupee to dollar exchange rate should increase by 5 ~ 2.5 = 3.5%. That's, rupee should depreciate by 3.5 pereent relative to the dollar b. The nominal dollar retum for the U.S. institutional investor is approximately 12 ~ 5 = 7%, The real rewrn forthe U.S. investoris approximately 7 ~ 2.5 ~ 4.5%, The real return for the Indian investor is approximately 12 ~ 5 = 6%. Thus, the US. in vestor has a lower real return than the Indian investor. This is so because the rupee depreciated with respect tothe dollar by more than what the PPP relationship would indicate, Indeed, the difference between the real returns is 5 = 4.5 = 1.5%, which is the same as the difference between the actual depreciation ofthe rupee and the de- preciation that should have occurred as per PPP (5 ~ 3.5 = | Ifa countey’s currency is undervalued, it means that the real prices of assets in this country are low compared with other countries. Also, the wages are lower in real terms than in other countries, Thus, investors from other countries would invest in this court tay to take advantage of low prices and wages. This action would help in the apprecia tion of the undervalued currency and restoration of the PPP in the long run, fa terms of foreign trade, an undervalued currency implies that exports from this country woul «ta boost while imports would become less attractive. This would also help in the ap- preciation of the unclervalued curreney and restoration of the PPP in the long run. In riskneutzal efficient foreign exchange markets, the forward rate is the expected value ofthe future spot rate, The forward rate can be computed! using the interest rate parity relation. Because the exchange rates given in SFr/$ terms, the appropriate ex- pression for the interest rate parity relation is B14 Saag {thatis, sn isa part of the numerator and ‘isa part of the denominator). Accordingly, the three-month forward rates 1+ 0.0095 Be 1428 Oot80 4600 Thus, the implied market prediction for the three-month ahead exchange rate is SFr1.8600/8,na Chapter 3. Foreign Exchange Determination and Forecasting 8 AS per the model, one € would be worth $).978I six months later. Based on che forward rate, one € would be worth $).9976 six months later. Therefore, the market farticr fants, who believe that the model is quite good, would tuy the dollar m the Forward rrarket (se eatos). Consequently, the price of eurg forward would decrease (ihe dollar forward would increase) and the forward rate would become equal to $0.9751/€. The spot exchange rate and the dollar and euro iaterest rates would change su as to be coe sistent with this forward rate. A look at the interest rate parity relationship—wrtten in Be form, gllts s"T+r as and S are in $/€ terms—suggests that with the Gecrease ta forward rte trom $0.9976/€ 19 $0.9781/€, the spot rate and interest rate m Gollats ae likely to gu own and intevest rate ia euros i Lkely to-go up. a The forward rate can be computed using the interest rate parity refation. Because the exchange rate is given in $/£ terms, the appropriate expression for the interest rate party relation is pr lts (hat is, mie a part of Ge numerator, and ri a part of the denominator), Accordingly, tue one-year forward rate is 1.0200 a 1085 L 620 = $ 1s2esve b, Based on tae forward rate, cme pound would be worth $1.5285 one year Later; The ‘model predicts that cne pound would be worth $1,5815 ene year later, Thus, as per the nodel, the pound is underpriced in the forward market. Accordingly, Dustin Green would buy pounds forward at $1.5288/¢. everyone were 1 buy pounds forward, the price of pounds lorward would increase and become equal to $1.5815/£, The spot exchange rate and the dollar and pound jatetest rates would change s0 38 tbe consistent with this forward rate. A look at the interest rate parity relationship suggests that the spot tate and the anterest rate ia collars are likely to go up and the interest rate in pounds is Ikely 10 go down, to be consistent with tae increase in the forward rate. 10, a, tre market participants are risk-neutral, the expected future spot exchange rate would be the same 4s the current forward rate. The forward rate i determined Ihased on the current spot exchange rate and the interest rate differential between the two currencies. Thus, the expected future spot exchange rate would depend on the current spot exchange rate and the interest rate cifferentil . Ifthe market participants zre riskaverse, the forward rate would differ from the ex pected future spot exchange rate by the tisk premium. The tisk premium, based on the extent of risk aversion of the market participants, could be positive or negative. So, the expected future spot exchange rate s the forward rate less the tisk premiuin Since the forward rate s based on the current spot exchange rate and the interestSolutions 125 rate differential between the two currencies, the expected future spot exchange rate would depend on the current spot exchange rate, the interest rate differential, and the risk premium, 11, There is some evidence of positive serial correlation in exchange rate movements ‘real and nominal). Hence, when a currency is going up, a reasonable forecasts that i wil continue going up. Similarly, when a currency is going down, a reasonable forecast is that it will continue going down, However, at some point in time, the trend would re- verse, and the problem is that a trendbased forecasting model cannot forecast when this cumming point would occur. Although these turning points may be infrequent, they can be the occasion of a huge swing. The Mexican peso is a good example, For a few years until the end of 1994, the real value ofthe peso appreciated steadily. So, forecast ers using (rend models for the real exchange rates were quite successful for many ‘months. However, in December 3994, he peso suddenly crashed and lost aroundhalf of its value, 12, Statements (a), (€), and (d) are true. Statement (b) isnot true, because the objective of central bank activity in the foreign exchange market s not to profit from trading activ- ties, but to implement monetary policy and exchange rate targets. 13, A mean reverting time series is one that may diverge from its fundamental value in the short run but reverts to its fandamental value in the long run. Empirical evidence sug. sgests that exchange rates are mean reverting. The real exchange rates ‘observed ex change rate minus inflation) do deviate from the fundamental value implied by PPP in. the short run, but lend to revert to the fundamental value in the long run. 14, Most econometric models are unsuitable for short-term exchange rate forecasts, as they ‘model long-term structural economic relationships, For long-term exchange rate fore- casts, the use of econometric models has some problems. First, most of them rely On. predictions for certain key variables, such as money supply and interest rates Itis not ‘easy to forecast these variables, Second, the structural correlation estimated by the para ‘meters of the equation can change over time, so that even if al causative variables are correctly forecasted, the model can stil yield poor exchange rate predictions. In peri= ‘ods when structural changes are rapid compared with the amount of time-series data re Quired to estimate parameters, econometric models are of litle help. 15, Technical analysis is more likely to be used for short-term exchange rate movements, while dhe econometric approach is more likely o be used for longterm exchange rate movements, The manager of a currency hedge fund and currency traders change their foreign exchange positions quite quickly, and are interested in shoreterm changes in exchange rates. On the other hand, the manager of an international stock portfolio is unlikely to change his foreign exchange positions quickly, because the transaction costs of buying and selling stocks are quite high. Therefore, the manager of an international stock portfolio is more interested in the long-term movements in exchange rates, Similarly the long-term strategic planner ofa corporation is interested in the long-termn movements. Accordingly the answers are a. Technical analysis b. Econometric approach . Technical analsis d. Econometric approach126 Chapter 3. Foreign Exchange Determination and Forecasting 6. "7 18, a The absolute values of prediction errors are as follows. Forward rate: 1440 — 1.308 = 0132; Analyst A: 1410 — LSU8 = 0.102; and Analyst B: L980 — 1.308) 272, Thus, the forecast by Analyst A was the most accurate. b, ‘The forward rate and Analyst B erroneously predicted that the SFr to $ exchange rate would go up from the then spot rate of SFr1.420/8; that is, the SFr would de preciate, Only Analyst A correctly predicted that the Swiss franc would appreciate, 44 The absolute values of forecast errors are as follows. Forward rate: "48.148 144.697 = 3.451; Commerabank: 148.148 — 142 = 6.148; and Harts Bank: 156 ~ 148.148 ~ 7.852. Thus, the forecast as per the forward rate was the most accurate, followed by Commerzbank, and then by Harts Bank, b, Although the forward rate and Commerzbank were wore accurate than Harts Bank, both of them erroneously predicted that the yen would appreciate relative 10 the dollar. Only Harris Bank correctly predicted thatthe yen would depreciate ©. Commersbank’s forecast was ¥I42/S, which was less than the forward rate Therefore, David Brock bought yen forward (sold dollars) a the rate of ¥144.097/5. Beciase the actual rate turned out to be ¥I48.148/8, buying yen forward did not tum out to be the right suategy. On the other hand, Hartis Bank's forecast was ¥156/S, which was moe dhan the forward rate. Therefore, Brian Lee sold yen for ward (bought dollars) atthe rate of FLH.697/5, Because the actual ate turned out tobe ¥148.148/, sling ye forward turned out to be the ight strategy 4. Commersbank’s forecast had Tower forecast error than Harris Bank's in part (3. Hoeven in part), it was right to buy and sell based on Harts Bank’ forecast and snot Commer*bank’s, The reason is that Harris Bank's forecast turmed out to be an the right side of the forward rte, while Commersbank's did not Let the forecasts made by the Industrial Bank of Japan at the beginning of period # for the beginning of period {+ { be £(S.1l). Let the forward rates quoted atthe begin hing of period r for the beginning of period r+ { be F, and the actual spotrates atthe beginning of periods f and ¢~ | be Syand S,., respectively. The percentage forecast t= rors (e and #) for each forecast made by the Industrial Bank of Japan and the forward rate are computed 48 €jy) = [Sot Eli alb)1/S, and es, = Wat ~ H/ Sn eespec tively: The following able details the computations. Pa, FBS) Son te 1 tos 18164 4251) 140144300 0.0300 0.9009 9.0002 2 130 S968 LAL 152.750 00814 20066 0.0087 3 152750 153500151 49.400 -0.0105 0001 9,0008 4149400 199.400 43——129.600. ~0.0897 008090175 5129600 9.700130 19.500 0039 —0.0015 3.0000 0.0000, 6 129300 1800131189250 Ons? 00730 andl L005 Msi 2.0033 2.0046 RMSE 00573 0.0578 The RMSE for the Industrial Bank of Japanis lower than that for the forward cate, Thus, as per the limited data set in this problem, the Industrial Bank outperformed the for ward rate in terms of accuracy of the forecasts, as measured by the RMSE. We have not tested whether the ffference in forecast performances is statistically significant