Statistical and Economic Methods for Evaluating Exchange Rate Predictability Pasquale Della Corte Imperial College London Ilias Tsiakas University of Guelph October 2011 Abstract This chapter provides a comprehensive review of the statistical and economic methods used for evaluating out-of-sample exchange rate predictability. We illustrate these methods by assessing the forecasting performance of a set of widely used empirical exchange rate models using monthly returns on nine major US dollar exchange rates. We …nd that empirical models based on uncovered interest parity, purchasing power parity and the asymmetric Taylor rule perform better than the random walk in out-of-sample forecasting using both statistical and economic criteria. We also con…rm that conditioning on monetary fundamentals does not generate out-ofsample economic value. Finally, combined forecasts formed using a variety of model averaging methods perform better than individual empirical models. These results are robust to reasonably high transaction costs, the choice of numeraire and the exclusion of any one currency from the investment opportunity set. Keywords: Exchange Rates; Out-of-Sample Predictability; Mean Squared Error; Economic Value; Combined Forecasts. JEL Classi…cation: F31; F37; G11; G17. This paper is forthcoming as a chapter in the Handbook of Exchange Rates. The authors are grateful to Lucio Sarno and an anonymous referee for useful comments. Contact details: Pasquale Della Corte, Finance Group, Imperial College Business School, Imperial College London, 53 Prince’s Gate, London SW7 2AZ, UK. Email: [email protected]; Ilias Tsiakas, Department of Economics and Finance, University of Guelph, Guelph, Ontario N1G 2W1, Canada. Email: [email protected]. 1 Introduction Exchange rate ‡uctuations are regularly monitored with great interest by policy makers, practitioners and academics. It is not surprising, therefore, that exchange rate predictability has long been at the top of the research agenda in international …nance. Starting with the seminal contribution of Meese and Rogo¤ (1983), a large body of empirical research …nds that models which condition on economically meaningful variables do not provide reliable exchange rate forecasts. This has lead to the prevailing view that exchange rates follow a random walk and hence are not predictable, especially at short horizons. Several well known puzzles in foreign exchange (FX) are responsible for this view. First, the “exchange rate disconnect puzzle” concerns the empirical disconnect between exchange rate movements and economic fundamentals such as money supply and real output (e.g., Mark, 1995; Cheung, Chinn and Pascual, 2005; Rogo¤ and Stavrakeva, 2008). Second, the “forward premium puzzle” implies that on average the interest di¤erential is not o¤set by a commensurate depreciation of the investment currency, which is an empirical violation of uncovered interest rate parity. As a result, borrowing in low-interest rate currencies and investing in high-interest rate currencies forms the basis of the widely used carry trade strategy in active currency management (e.g., Fama, 1984; Burnside, Eichenbaum, Kleshchelski and Rebelo, 2011; Brunnermeier, Nagel and Pedersen, 2009; and Della Corte, Sarno and Tsiakas, 2009). Third, there is extensive evidence that purchasing power parity holds in the long run (e.g., Lothian and Taylor, 1996). A recent contribution by Engel and West (2005) provides a possible resolution to the di¢ culty of tying exchange rates to economic fundamentals. Speci…cally, Engel and West (2005) show analytically that exchange rates can be consistent with present-value asset pricing models and still manifest nearrandom walk behaviour if two conditions are met: (i) fundamentals are integrated of order one, and (ii) the discount factor for future fundamentals is near one.1 A model that is nested by the Engel and West (2005) present value relation is a variant of the Taylor (1993) rule used for exchange rate determination. The Taylor rule postulates that the central bank adjusts the short-run nominal interest rate in response to changes in in‡ation, the output gap and the exchange rate. Using alternative speci…cations of Taylor rule fundamentals, Molodtsova and Papell (2009) provide strong evidence of short-horizon exchange rate predictability, and hence o¤er renewed hope for empirical success in this literature. In short, one way to summarize the state of the literature is that it has come full circle: from the Meese and Rogo¤ (1983) “no predictability at short horizons,” to the Mark (1995) “predictability at long but not at short horizons,” to the Cheung, Chinn and Pascual (2005) “no predictability at any horizon,”to …nally the Molodtsova and 1 The assumption of integrated fundamentals of order one is widely accepted in the literature. The assumption that the discount factor is close to one has been empirically validated by Sarno and Sojli (2009). 1 Papell (2009) “predictability at short horizons with Taylor rule fundamentals.” This chapter aims at connecting these related literatures by providing a comprehensive review of the statistical and economic methods used for evaluating exchange rate predictability, especially out of sample. We assess the short-horizon forecasting performance of a set of widely used empirical exchange rate models that include the random walk model, uncovered interest parity, purchasing power parity, monetary fundamentals, and symmetric and asymmetric Taylor rules. Our analysis employs monthly FX data ranging from January 1976 to June 2010 for the 10 most liquid (G10) currencies in the world: the Australian dollar, Canadian dollar, Swiss franc, Deutsche markneuro, British pound, Japanese yen, Norwegian kroner, New Zealand dollar, Swedish kronor and US dollar.2 The vast majority of the FX literature uses a well established statistical methodology for evaluating exchange rate predictability. This methodology typically involves statistical tests of the null hypothesis of equal predictive ability between the random walk benchmark and an alternative empirical exchange rate model. The tests are based on the out-of-sample mean squared error (MSE) of the forecasts generated by the models. In this chapter, we discuss the main recent contributions to this methodology. The most popular method for testing whether the alternative model has a lower MSE than the benchmark is using the Diebold and Mariano (1995) and West (1996) statistic. By design, however, all the models we estimate are nested and this statistic has a non-standard distribution when comparing forecasts from nested models. Therefore, we focus on the recent inference procedure by Clark and West (2006, 2007), which accounts for the fact that under the null the MSE from the alternative model is expected to be greater than that of the RW benchmark because the alternative model introduces noise into the forecasting process by estimating a parameter vector that is not helpful in prediction. For a comprehensive statistical evaluation, we also implement the encompassing test of Clark and McCracken (2001) and the F -statistic of McCracken (2007) using bootstrapped critical values. Finally, following Campbell and Thompson (2008) and Welch and Goyal (2008) we also 2 measure and a root MSE di¤erence statistic. report the out-of-sample Roos In addition to the extensive literature on statistical evaluation, there is also an emerging line of research proposing a methodology for assessing the economic value of exchange rate predictability. A purely statistical analysis of predictability is not particularly informative to an investor as it falls short of measuring whether there are tangible economic gains from using dynamic forecasts in active portfolio management. We review this approach based on dynamic asset allocation that is used, among others, by West, Edison and Cho (1993), Fleming, Kirby and Ostdiek (2001), Marquering 2 Note that we will not be discussing two recent approaches to predicting movements in exchange rates: the microstructure approach that conditions on order ‡ow as a measure of net buying pressure for a currency (e.g., Evans and Lyons, 2002, and Rime, Sarno and Sojli, 2010); and (ii) the global imbalances approach (e.g., Gourinchas and Rey, 2007, and Della Corte, Sarno and Sestieri, 2011). 2 and Verbeek (2004), Abhyankar, Sarno and Valente (2005), Bandi and Russell (2006), Han (2006), Bandi, Russell and Zhu (2008), Della Corte, Sarno and Thornton (2008) and Della Corte, Sarno and Tsiakas (2009, 2011). We …rst design an international asset allocation strategy that exposes a US investor purely to FX risk. The investor builds a portfolio by allocating her wealth between a domestic and a set of foreign bonds and then uses the exchange rate forecasts from each model to predict the US dollar return of the foreign bonds. We evaluate the performance of the dynamically rebalanced portfolios using mean-variance analysis, which allows us to measure how much a risk averse investor is willing to pay for switching from a portfolio strategy based on the random walk benchmark to an empirical exchange rate model that conditions on economic fundamentals. In contrast to statistical measures of forecast accuracy that are computed separately for each exchange rate, the economic value is assessed for the portfolio generated by a model’s forecasts on all exchange rate returns. This contributes to our …nding that even modest statistical signi…cance in out-of-sample predictive regressions can lead to large economic bene…ts for investors. Our review also includes an assessment of the economic value of combined forecasts. We use a variety of model averaging methods, some of which generate forecast combinations in a naive ad hoc manner, some exploit statistical measures of past out-of-sample forecasting performance, and some that use economic measures of past predictability. All forecast combinations we explore are formed ex ante using the full universe of individual forecasts of each model for each exchange rate. It is important to note that the combined forecasts do not require a view of which model is best at any given time period and therefore provide a way for resolving model uncertainty. To preview our key results, we …nd strong statistical and economic evidence against the random walk benchmark. In particular, empirical exchange rate models based on uncovered interest parity, purchasing power parity and the asymmetric Taylor rule perform better than the random walk in outof-sample prediction using both statistical and economic criteria. We also con…rm that conditioning on monetary fundamentals does not generate out-of-sample economic gains. The worst performing model is consistently the symmetric Taylor rule. Finally, combined forecasts formed using a variety of model averaging methods perform even better than individual empirical models. These results are robust to reasonably high transaction costs, the choice of numeraire and the exclusion of any one currency from the investment opportunity set. The remainder of the chapter is organized as follows. In the next section we brie‡y review the empirical exchange rate models we estimate and their foundations in asset pricing. Section 3 describes the statistical methods we use for evaluating exchange rate predictability. In Section 4 we present a general framework for assessing the economic value of forecasting exchange rates for a risk averse investor with a dynamic mean-variance portfolio allocation strategy. Section 5 explains the 3 construction of combined forecasts using a variety of model averaging methods. Section 6 reports our empirical results and, …nally, Section 7 concludes. 2 Models for Exchange Rate Predictability In this section we review the empirical models we use for evaluating exchange rate predictability. We begin by describing the Engel and West (2005) present value model that nests and motivates many of the predictive regressions we estimate. 2.1 A Present Value Model for Exchange Rates The Engel and West (2005) model relates the exchange rate to economic fundamentals and the expected future exchange rate as follows: st = (1 b) (f1;t + z1;t ) + b (f2;t + z2;t ) + bEt st+1 ; (1) where st is the log of the nominal exchange rate de…ned as the domestic price of foreign currency, fi;t (i = 1; 2) are the observed economic fundamentals and zi;t are the unobserved fundamentals that drive the exchange rate. Note that an increase in st implies a depreciation of the domestic currency. This is a general asset pricing model that builds on earlier work on pricing stock returns by Campbell and Shiller (1987, 1988) and West (1988). Iterating forward and imposing the no-bubbles condition leads to the following present-value relation: st = (1 b) 1 X bj Et (f1;t+j + z1;t+j ) + b j=0 1 X bj Et (f2;t+j + z2;t+j ) : (2) j=0 Engel and West (2005) show that the exchange rate will follow a random walk if the discount factor b is close to one and either: (1) f1;t + z1;t I (1) and f2;t + z2;t = 0; or (2) f2;t + z2;t I (1). Some other well-known exchange rate models take the general form of Equation (1), and in what follows, we discuss two examples. 2.1.1 Monetary Fundamentals Consider …rst the monetary exchange rate models of the 1970s and 1980s, which assume that the money market relation is described by: mt = pt + yt it + vm;t ; where mt is the log of the domestic money supply, pt is the log of the domestic price level, is the income elasticity of money demand, yt is the log of the domestic national income, (3) >0 > 0 is the interest rate semi-elasticity of money demand, it is the domestic nominal interest rate and vm;t 4 is a shock to domestic money demand. A similar equation holds for the foreign economy, where the corresponding variables are denoted by mt , pt , yt , it and vm;t . We assume that the parameters f ; g of the foreign money demand are identical to the domestic parameters. The nominal exchange rate is equal to its purchasing power parity (PPP) value plus the real exchange rate qt : s t = pt p t + qt : (4) Finally, the interest parity condition is given by: Et st+1 where t st = it it + t; (5) is the deviation from the uncovered interest parity (UIP) condition that is based on rational expectations and risk neutrality. Hence t can be interpreted either as an expectational error or a risk premium. Using Equations (3) to (5) for the domestic and foreign economies and re-arranging, we get: st = 1 1+ mt mt (yt yt ) + qt vm;t vm;t t + 1+ Et st+1 : (6) This equation takes the form of the original model in Equation (1), where the discount factor is given by b = =1 + , the observable fundamentals are f1;t = mt fundamentals are z1;t = qt 2.1.2 vm;t vm;t and z2;t = mt (yt yt ), and the unobservable t. Taylor Rule The second model to be nested by the Engel and West (2005) present value relation is the Taylor (1993) rule, where the home country is assumed to set the short-term nominal interest rate according to: it = i + g 1 yt + 2( t ) + vt ; (7) where i is the target short-term interest rate, ytg is the output gap measured as the % deviation of actual real GDP from an estimate of its potential level, t is the in‡ation rate, is the target in‡ation rate and vt is a shock. The Taylor rule postulates that the central bank raises the short-term nominal interest rate when output is above potential output and/or in‡ation rises above its desired level. The foreign country is assumed to follow a Taylor rule that explicitly targets exchange rates (e.g., Clarida, Gali and Gertler, 1998): it = 0 (st st ) + i + 1 yt 5 g + 2( t ) + vt ; (8) where 0 < 0 < 1 and st is the target exchange rate. For simplicity, we assume that the home and foreign countries target the same interest rate, i, and the same in‡ation rate, . The rule indicates that the foreign country raises interest rates when its currency depreciates relative to the target.3 We assume that the foreign central bank targets the PPP level of the exchange rate: s t = pt pt : (9) Taking the di¤erence between the home and foreign Taylor rules, using interest parity (5), substituting the target exchange rate and solving for st gives: st = 0 1+ (pt 1 1+ pt ) 0 1 0 ytg yt g + 2( t t) + vt vt + t + 1 1+ Et st+1 : (10) 0 This equation also has the general form of the present value model in Equation (1), where the discount factor is b = 1=1 + 2.2 0, f1;t = pt pt and z2;t = 1 ytg yt g + 2( t t) + vt vt + t . Predictive Regressions Our empirical analysis is based on six predictive regressions for exchange rate returns, many of which are nested and motivated by the Engel and West (2005) present value model. All predictive regressions have the same linear structure: st+1 = where st+1 = st+1 st , and + xt + "t+1 ; (11) are constants to be estimated and "t+1 is a normal error term. The empirical models di¤er in the way they specify the economic fundamentals xt that are used to forecast exchange rate returns. 2.2.1 Random Walk The …rst regression is the random walk (RW) with drift model that sets = 0. Since the seminal work of Meese and Rogo¤ (1983), this model has become the benchmark in assessing exchange rate predictability. The RW model captures the prevailing view in international …nance research that exchange rates are not predictable when conditioning on economic fundamentals, especially at short horizons. 3 The argument still follows if the home country also targets exchange rates. It is standard to omit the exchange rate target from Equation (7) on the interpretation that US monetary policy has essentially ignored exchange rates (see, Engel and West, 2005). 6 2.2.2 Uncovered Interest Parity The second regression is based on the UIP condition:4 xt = it it : (12) UIP is the cornerstone condition for FX market e¢ ciency. Assuming risk neutrality and rational expectations, it implies that = 0, = 1, and the error term is serially uncorrelated. However, numerous empirical studies consistently reject the UIP condition (e.g., Hodrick, 1987; Engel, 1996; Sarno, 2005). As a result, it is a stylized fact that estimates of tend to be closer to minus unity than plus unity. This is commonly referred to as the “forward premium puzzle,” which implies that high-interest currencies tend to appreciate rather than depreciate and forms the basis of the widely used carry trade strategy in active currency management.5 2.2.3 Purchasing Power Parity The third regression is based on the PPP hypothesis: xt = p t pt st : (13) The PPP hypothesis states that national price levels should be equal when expressed in a common currency and is typically thought of as a long-run condition rather than holding at each point in time (e.g., Rogo¤, 1996; and Taylor and Taylor, 2004). 2.2.4 Monetary Fundamentals The fourth regression conditions on monetary fundamentals (MF): xt = (mt mt ) (yt yt ) st : (14) The relation between exchange rates and fundamentals de…ned in Equation (14) suggests that a deviation of the nominal exchange rate st+1 from its long-run equilibrium level determined by the fundamentals xt , requires the exchange rate to move in the future so as to converge towards its longrun equilibrium. The empirical evidence on the relation between exchange rates and fundamentals is mixed. On the one hand, short-run exchange rate variability appears to be disconnected from the underlying fundamentals (Mark, 1995) in what is commonly referred to as the “exchange rate disconnect puzzle.” On the other hand, some recent empirical research …nds that fundamentals and nominal exchange rates move together in the long run (Groen, 2000; and Mark and Sul, 2001). 4 An alternative way of testing UIP is to estimate the “Fama regression” (Fama, 1984), which conditions on the forward premium. Note that if covered interest parity (CIP) holds, the interest rate di¤erential is equal to the forward premium and testing UIP is equivalent to testing for forward unbiasedness in exchange rates (Bilson, 1981). For recent evidence on CIP see Akram, Rime and Sarno (2008). 5 Clarida, Sarno, Taylor and Valente (2003, 2006) and Boudoukh, Richardson and Whitelaw (2006) also show that the term structure of forward exchange (and interest) rates contains valuable information for forecasting spot exchange rates. 7 2.2.5 Taylor Rule The …nal two regressions are based on simple versions of the Taylor (1993) rule. We estimate a symmetric Taylor rule (TRs ): xt = 1:5 ( t t) + 0:1 ytg yt g ; (15) as well as an asymmetric Taylor rule (TRa ) that assumes that the foreign central bank also targets the real exchange rate: xt = 1:5 ( t t) + 0:1 ytg yt g + 0:1 (st + pt pt ) : (16) The domestic and foreign output gaps are computed with a Hodrick and Prescott (1997) (HP) …lter.6 The parameters on the in‡ation di¤erence (1:5), output gap di¤erence (0:1) and the real exchange rate (0:1) are fairly standard in the literature (e.g., Engel, Mark and West, 2007; Mark, 2009). Alternative versions of the Taylor rule that we do not consider in this chapter may also account for smoothing, where interest rate adjustments are not immediate but gradual, and heterogeneous coe¢ cients for (i) the US versus foreign in‡ation, and (ii) the US versus foreign output gap (e.g., Molodtsova and Papell, 2009). 3 Statistical Evaluation of Exchange Rate Predictability The success or failure of empirical exchange rate models is typically determined by statistical tests of out-of-sample predictive ability. Our statistical analysis tests for equal predictive ability between one of the empirical exchange rate models we estimate (UIP, PPP, MF, TRs or TRa ) and the benchmark RW model. In e¤ect, we are comparing the performance of a parsimonious restricted null model (the RW, where model (where = 0) to a set of larger alternative unrestricted models that nest the parsimonious 6= 0).7 We estimate all empirical exchange rate models using ordinary least squares (OLS), and then run a pseudo out-of-sample forecasting exercise as follows (e.g., Stock and Watson, 2003). Given today’s known observables f st+1 ; xt gTt=11 , we de…ne an in-sample (IS) period using observations T 1 f st+1 ; xt gM t=1 , and an out-of-sample (OOS) period using f st+1 ; xt gt=M +1 . This exercise produces P = (T 1) M OOS forecasts. Our empirical analysis uses T 1 = 413 monthly observations, M = 120 and P = 293.8 6 Note that in estimating the HP trend in sample or out of sample, at any given period t, we only use data up to period t 1. We then update the HP trend every time a new observation is added to the sample. This captures as closely as possible the information available to central banks at the time decisions are made. 7 For a review of forecast evaluation see West (2006) and Clark and McCracken (2011). 8 The IS period for xt ranges from January 1976 to December 1985. The …rst OOS forecast is for the February 1986 value of st+1 that conditions on the January 1986 value of xt . The last forecast is for June 2010. 8 In what follows, we describe a comprehensive set of statistical criteria for evaluating the OOS predictive ability of empirical exchange rate models. First, we compute the Campbell and Thompson 2 , that compares the unconditional forecasts of the benchmark RW (2008) OOS R2 statistic, Roos model to the conditional forecasts of an alternative model. Let st+1jt denote the one-step ahead sbt+1jt be the one-step ahead conditional forecast from the unconditional forecast from the RW and 2 statistic is given by: alternative model represented by one of Equations (12) to (16). Then, the Roos 2 Roos PT 1 t=M +1 PT 1 t=M +1 =1 2 sbt+1jt st+1 st+1 2: st+1jt 2 statistic implies that the alternative model outperforms the benchmark RW by having A positive Roos a lower mean squared error (MSE). Second, we compute the OOS root MSE di¤erence statistic, RM SE, as in Welch and Goyal (2008): RM SE = A positive s PT 1 t=M +1 st+1jt st+1 P s 2 PT 1 t=M +1 sbt+1jt st+1 P 2 : RM SE denotes that the alternative model outperforms the benchmark RW by having a lower RMSE. The most popular method for testing whether the alternative model has a lower MSE than the benchmark is using the Diebold and Mariano (1995) and West (1996) statistic, which has an asymptotic standard normal distribution when comparing forecasts from non-nested models. However, as shown by Clark and McCracken (2001) and McCracken (2007), this statistic has a non-standard distribution when comparing forecasts from nested models and is severely undersized when using standard normal critical values. Clark and McCracken (2001) and McCracken (2007) account for this size distortion by deriving the non-standard asymptotic distributions for a number of statistical tests as applied to nested models. We report the two tests with the best overall power and size properties: the EN C-F encompassing test statistic proposed by Clark and McCracken (2001) de…ned as follows: EN C-F = PT 1 t=M +1 st+1 P st+1jt PT 1 2 st+1 1 t=M +1 st+1 st+1jt st+1jt sbt+1jt st+1 2 ; and the M SE-F test of McCracken (2007): M SE-F = PT 1 t=M +1 P st+1 PT 1 st+1jt 1 t=M +1 2 st+1 st+1 sbt+1jt 2 sbt+1jt 2 : When the models are correctly speci…ed, the forecast errors are serially uncorrelated and exhibit conditional homoskedasticity. In this case, Clark and McCracken (2001) and McCracken (2007) numerically generate the asymptotic critical values for the EN C-F and M SE-F tests. When the 9 above conditions are not satis…ed, a bootstrap procedure must be used to compute valid critical values, which we discuss later. Finally, we also apply the recently developed inference procedure by Clark and West (2006, 2007) for testing the null of equal predictive ability of two nested models. This procedure acknowledges the fact that under the null the MSE from the alternative model is expected to be greater than that of the RW benchmark because the alternative model introduces noise into the forecasting process by estimating a parameter vector that is not helpful in prediction. Therefore, …nding that the RW has smaller MSE is not clear evidence against the alternative model. Clark and West (2006, 2007) suggest that the MSE should be adjusted as follows: M SEadj = 1 XT 1 ( st+1 t=M +1 P sbt+1jt )2 1 XT 1 ( st+1jt t=M +1 P Then, a computationally convenient way of testing for equal MSE is to de…ne fbt+1jt = ( st+1 st+1jt )2 [( st+1 sbt+1jt )2 ( st+1jt sbt+1jt )2 : sbt+1jt )2 ]; (17) (18) and to regress fbt+1jt on a constant, using the t-statistic for a zero coe¢ cient, which we denote by MSE-t. Even though the asymptotic distribution of this test is non-standard (e.g., McCracken, 2007), Clark and West (2006, 2007) show that standard normal critical values provide a good approximation, and therefore recommend to reject the null if the statistic is greater than +1:282 (for a one sided 0:10 test) or +1:645 (for a one sided 0:05 test).9 The above statistical tests compare the null hypothesis of equal forecast accuracy against the one-sided alternative that forecasts from the unrestricted model are more accurate than those from the restricted benchmark model. Asymptotic critical values for these test statistics, whenever available, tend to be severely biased in small samples. In addition to the size distortion, there may be spurious evidence of return predictability in small samples when the forecasting variable is su¢ ciently persistent (e.g., Nelson and Kim, 1993; Stambaugh, 1999). In order to address these concerns, we obtain bootstrapped critical values for a one-sided test by estimating the model and generating 10; 000 bootstrapped time series under the null. The procedure preserves the autocorrelation structure of the predictive variable and maintains the cross-correlation structure of the residual. The bootstrap algorithm is summarized in Appendix A. 4 Economic Evaluation of Exchange Rate Predictability This section describes the framework for evaluating the performance of an asset allocation strategy that exploits predictability in exchange rate returns. 9 This approximation tends to perform better when forecasts are obtained from rolling regressions than recursive regressions. 10 4.1 The Dynamic FX Strategy We design an international asset allocation strategy that involves trading the US dollar and nine other currencies: the Australian dollar, Canadian dollar, Swiss franc, Deutsche markneuro, British pound, Japanese yen, Norwegian kroner, New Zealand dollar and Swedish kronor. Consider a US investor who builds a portfolio by allocating her wealth between ten bonds: one domestic (US), and nine foreign bonds (Australia, Canada, Switzerland, Germany, UK, Japan, Norway, New Zealand and Sweden). The yield of the bonds is proxied by eurodeposit rates. At the each period t + 1, the foreign bonds yield a riskless return in local currency but a risky return rt+1 in US dollars, whose expectation at time t is equal to Et [rt+1 ] = it + st+1jt . Hence the only risk the US investor is exposed to is FX risk. Every period the investor takes two steps. First, she uses each predictive regression to forecast the one-period ahead exchange rate returns. Second, conditional on the forecasts of each model, she dynamically rebalances her portfolio by computing the new optimal weights. This setup is designed to assess the economic value of exchange rate predictability by informing us which empirical exchange rate model leads to a better performing allocation strategy. 4.2 Mean-Variance Dynamic Asset Allocation Mean-variance analysis is a natural framework for assessing the economic value of strategies that exploit predictability in the mean and variance. Consider an investor who has a one-period horizon and constructs a dynamically rebalanced portfolio. Computing the time-varying weights of this portfolio requires one-step ahead forecasts of the conditional mean and the conditional variancecovariance matrix. Let rt+1 denote the K 1 vector of risky asset returns; the conditional expectation of rt+1 ; and = Et t+1jt rt+1 rt+1 t+1jt t+1jt 0 t+1jt = Et [rt+1 ] is is the K K conditional variance-covariance matrix of rt+1 . Mean-variance analysis may involve three rules for optimal asset allocation: maximum expected utility, maximum expected return and minimum volatility. Following Della Corte, Sarno and Tsiakas (2009, 2011) our empirical analysis focuses on the maximum expected return strategy as this is the strategy most often used in active currency management. For details on the maximum expected utility rule and the minimum volatility rule see Han (2006). The maximum expected return rule leads to a portfolio allocation on the e¢ cient frontier for a given target conditional volatility. At each period t, the investor solves the following problem: max wt n p;t+1 s.t. = wt0 2 p t+1jt = wt0 11 + 1 wt0 t+1jt wt ; rf o (19) (20) where p is the target conditional volatility of the portfolio returns. The solution to the maximum expected return rule gives the following risky asset weights: wt = p where Ct = t+1jt rf 0 1 t+1jt t+1jt p Ct 1 t+1jt t+1jt rf ; (21) rf + wt0 rt+1 : (22) rf . Then, the gross return on the investor’s portfolio is: Rp;t+1 = 1 + rp;t+1 = 1 + 1 Note that we assume that t+1jt = , where wt0 is the unconditional covariance matrix of exchange rate returns. In other words, we do not model the dynamics of FX return volatility and correlation. Therefore, the optimal weights will vary across the empirical exchange rate models only to the extent that the predictive regressions produce better forecasts of the exchange rate returns.10 4.3 Performance Measures We assess the economic value of exchange rate predictability with a set of standard mean-variance performance measures. We begin our discussion with the Fleming, Kirby and Ostdiek (2001) performance fee, which is based on the principle that at any point in time, one set of forecasts is better than another if investment decisions based on the …rst set lead to higher average realized utility. The performance fee is computed by equating the average utility of the RW optimal portfolio with the average utility of the alternative (e.g., UIP) optimal portfolio, where the latter is subject to expenses F. Since the investor is indi¤erent between these two strategies, we interpret F as the maximum performance fee she will pay to switch from the RW to the alternative (e.g., UIP) strategy. In other words, this utility-based criterion measures how much a mean-variance investor is willing to pay for conditioning on better exchange rate forecasts. The performance fee will depend on , which is the investor’s degree of relative risk aversion (RRA). To estimate the fee, we …nd the value of F that satis…es: T X1 t=0 Rp;t+1 F 2 (1 + ) Rp;t+1 F 2 = T X1 Rp;t+1 t=0 2 (1 + ) 2 Rp;t+1 ; (23) where Rp;t+1 is the gross portfolio return constructed using the forecasts from the alternative (e.g., UIP) model, and Rp;t+1 is the gross portfolio return implied by the benchmark RW model. We also evaluate performance using the premium return, which builds on the Goetzmann, Ingersoll, Spiegel and Welch (2007) manipulation-proof performance measure and is de…ned as: " T 1 # " T 1 # 1 1 1 X Rp;t+1 1 1 X Rp;t+1 1 ; P= ln ln (1 ) T Rf (1 ) T Rf t=0 (24) t=0 10 See Della Corte, Sarno and Tsiakas (2012) for an economic evaluation of volatility and correlation timing in foreign exchange. 12 where Rf = 1 + rf . P is robust to the distribution of portfolio returns and does not require the assumption of a particular utility function to rank portfolios. In contrast, the Fleming, Kirby and Ostdiek (2001) performance fee assumes a quadratic utility function. P can be interpreted as the certainty equivalent of the excess portfolio returns and hence can also be viewed as the maximum performance fee an investor will pay to switch from the benchmark to another strategy. In other words, this criterion measures the risk-adjusted excess return an investor enjoys for using one particular exchange rate model rather than assuming a random walk. We report both F and P in annualized basis points (bps). In the context of mean-variance analysis, perhaps the most commonly used measure of economic value is the Sharpe ratio (SR). The realized SR is equal to the average excess return of a portfolio divided by the standard deviation of the portfolio returns. It is well known that because the SR uses the sample standard deviation of the realized portfolio returns, it overestimates the conditional risk an investor faces at each point in time and hence underestimates the performance of dynamic strategies (e.g., Marquering and Verbeek, 2004; Han, 2006). Finally, we also compute the Sortino ratio (SO), which measures the excess return to “bad” volatility. Unlike the SR, the SO di¤erentiates between volatility due to “up” and “down” movements in portfolio returns. It is equal to the average excess return divided by the standard deviation of only the negative returns. In other words, the SO does not take into account positive returns in computing volatility because these are desirable. A large SO indicates a low risk of large losses. 4.4 Transaction Costs The e¤ect of transaction costs is an essential consideration in assessing the pro…tability of dynamic trading strategies. We account for this e¤ect in three ways. First, we calculate the performance measures for the case when the bid-ask spread for spot exchange rates is equal to 8 bps. In foreign exchange trading, this is a realistic range for the recent level of transaction costs.11 We follow the simple approximation of Marquering and Verbeek (2004) by deducting the proportional transaction cost from the portfolio return ex post. This ignores the fact that dynamic portfolios are no longer optimal in the presence of transaction costs but maintains simplicity and tractability in our analysis.12 The second way of accounting for transaction costs acknowledges the fact that for long data samples the transaction costs will likely change over time. Neely, Weller and Ulrich (2009) …nd that 11 In recent years, the typical transaction cost a large investor pays in the FX market is 1 pip, which is equal to 0:01 cent. For example, if the USD/GBP exchange rate is equal to 1:5000, 1 pip would raise it to 1:5001 and this would roughly correspond to 1=2 basis point proportional cost. 12 Our empirical analysis uses the full bid-ask spread. Note, however, that the e¤ective spread is generally lower than the quoted spread, since trading takes place at the best price quoted at any point in time, suggesting that the worse quotes will not attract trades. For example, Goyal and Saretto (2009) and Della Corte, Sarno and Tsiakas (2011) consider e¤ective transaction costs in the range of 50% to 100% of the quoted spread. Assuming that the e¤ective spread is less than the quoted spread would make our economic evidence stronger. 13 the transaction cost for switching from a long to a short position in FX has on average declined from about 10 bps in the 1970s to about 2 bps in recent years. If we were to keep transaction costs constant over our sample period, we would spuriously introduce a decline in performance by penalizing more recent returns too heavily relative to those early in the sample period. Therefore, we follow Neely, Weller and Ulrich (2009) in estimating a simple time trend that assumes that the bid-ask spread was 20 bps at the beginning of our data sample and declined linearly to 4 bps by the end of the sample. The actual one-way transaction cost is half of the bid-ask spread and hence declines from 10 bps to 2 bps. Speci…cally, the net return from buying a currency at the spot exchange rate at time t and selling at time t + 1 is equal to sbid t+1 ask ct+1 = 0:5 St+1 sask = smid t t+1 smid t t+1 , where t+1 ct+1 ) = ln (1(1+c and t) bid =S mid is the one-way transaction cost (e.g., Neely, Weller and Ulrich, 2009). St+1 t+1 Upper case St is the spot exchange rate and lower case st is st = ln St . In the …rst case we assume a …xed bid-ask spread and hence t = , whereas in the second case t is time-varying.13 Third, we also calculate the break-even proportional transaction cost, indi¤erent between two strategies (e.g., Han, 2006). We assume that be , that renders investors is a …xed fraction of the value traded in all assets in the portfolio. Then, the cost of the dynamic strategy is jwt each asset j 1 (1+rj;t ) j 1+rp;t for K. In comparing a dynamic strategy with the benchmark RW strategy, an investor who pays transaction costs lower than be will prefer the dynamic strategy. Since cost paid every time the portfolio is rebalanced, we report 5 wt be be is a proportional in monthly basis points.14 Combined Forecasts Our analysis has so far focussed on evaluating the performance of individual empirical exchange rate models relative to the random walk benchmark. Considering a large set of alternative models that capture di¤erent aspects of exchange rate behaviour without knowing which model is “true”(or best) inevitably generates model uncertainty. In this section, we resolve this uncertainty by exploring whether portfolio performance improves when combining the forecasts arising from the full set of predictive regressions. Even though the potentially superior performance of combined forecasts is known since the seminal work of Bates and Granger (1969), applications in …nance are only recently becoming increasingly popular (e.g., Timmermann, 2006). Rapach, Strauss and Zhou (2010) argue that forecast combinations can deliver statistically and economically signi…cant out-of-sample gains for two reasons: (i) they reduce forecast volatility relative to individual forecasts, and (ii) they are 13 The derivation is as follows: bid St+1 Stask = mid St+1 0:5 Stmid +0:5 ask St+1 Stask ( ( bid St+1 Stbid ) = ) Stmid (1 ct+1 ) mid ln sbid ln sask ln smid ln (1+ct ) . t+1 t+1 = ln st+1 t+1 14 For a slightly di¤erent calculation see Jondeau and Rockinger (2008). 14 bid St+1 S mid t+1 ask S bid 0:5 St t 1+ mid St mid 1 St+1 ask 0:5 St+1 ( ( ) ) ! ! = mid St+1 (1 ct+1 ) Stmid (1+ct ) . Then, linked to the real economy.15 Recall that we estimate N = 6 predictive regressions each of which provides an individual forecast sbi;t+1 for the one-step ahead exchange rate return, where i N . We de…ne the combined forecast sbc;t+1 as the weighted average of the N individual forecasts sbc;t+1 = XN i=1 ! i;t sbi;t+1 ; sbi;t+1 : (25) where f! i;t gN i=1 are the ex ante combining weights determined at time t. We form three types of combined forecasts. The …rst one uses simple model averaging that in turn implements three rules: (i) the mean of the panel of forecasts so that ! i;t = 1=N ; the median of the f sbi;t+1 gN i=1 individual forecasts; and (iii) the trimmed mean that sets ! i;t = 0 for the individual forecasts with the smallest and largest values and ! i;t = 1= (N 2) for the remaining individual forecasts. These combined forecasts disregard the historical performance of the individual forecasts. The second type of combined forecasts is based on Stock and Watson (2004) and uses statistical information on the past OOS performance of each individual model. In particular, we compute the discounted MSE (DM SE) forecast combination by setting the following weights: where DM SEi;t1 ! i;t = PN ; 1 j=1 DM SEj;t DM SEi;t = XT 1 T 1 t t=M +1 ( st+1 sbi;t+1 )2 ; (26) is a discount factor and M are the …rst in-sample observations on which we condition to form the …rst out-of-sample forecast. For < 1, greater weight is attached to the most recent forecast accuracy of the individual models. The DM SE forecasts are computed for three values of = f0:90; 0:95; 1:0g. The case of no discounting ( = 1) corresponds to the Bates and Granger (1969) optimal forecast combination when the individual forecasts are uncorrelated. We also compute simpler “most recently best” M SE ( ) forecast combinations that use no discounting ( = 1) and weigh individual forecasts by the inverse of the OOS MSE computed over the last months, where = f12; 36; 60g. The third type of combined forecasts does not use statistical information on the historical performance of individual forecasts. Instead it exploits the economic information contained in the Sharpe ratio (SR) of the portfolio returns generated by an individual forecasting model over a prespeci…ed recent period. We compute the discounted SR (DSR) combined forecast by setting the following weights: DSRi;t ; ! i;t = PN j=1 DSRj;t DSRi;t = XT 1 t=M +1 T 1 t SRt+1 ; (27) Finally, we also compute simpler “most recently best” SR ( ) forecast combinations that use no discounting ( = 1) and weigh individual forecasts by the OOS SR computed over the last 15 months, For a Bayesian approach to forecast combinations see Avramov (2002), Cremers (2002), Wright (2008), and Della Corte, Sarno and Tsiakas (2009, 2012). 15 where = f12; 36; 60g. We assess the economic value of combined forecasts by treating them in the same way as any of the individual empirical models. For instance, we compute the performance fee, F, for the DM SE one-month ahead forecasts and compare it to the RW benchmark. Finally, note that where possible we use these forecast combination methods not only for the OOS mean prediction but also for the OOS variance covariance matrix that enters the weights in mean-variance asset allocation. 6 Empirical Results 6.1 Data on Exchange Rates and Economic Fundamentals The data sample consists of 414 monthly observations ranging from January 1976 to June 2010, and focuses on nine spot exchange rates relative to the US dollar (USD): the Australian dollar (AUD), Canadian dollar (CAD), Swiss franc (CHF), Deutsche markneuro (EUR), British pound (GBP), Japanese yen (JPY), Norwegian kroner (NOK), New Zealand dollar (NZD) and Swedish kronor (SEK). The exchange rate is de…ned as the US dollar price of a unit of foreign currency so that an increase in the exchange rate implies a depreciation of the US dollar. These data are obtained through the download data program (DDP) of the Board of Governors of the Federal Reserve System.16 Table 1 provides a detailed description of all data sources we use. For interest rates, we use the one-month euro deposit rate taken from Datastream with the following exceptions. For Japan, the euro deposit rate is only available from January 1979 and hence before this date we use Covered Interest Parity (CIP) relative to USD to construct the no-arbitrage riskless rate. The one-month forward exchange rate required to implement CIP is taken from Hai, Mark, and Wu (1997). For Australia, Norway, New Zealand and Sweden, euro deposit rates are only available from April 1997. For Australia and New Zealand, we combine the money market rate from January 1976 to November 1984 taken from the IMF’s International Financial Statistics (IFS) and CIP relative to USD from December 1984 to March 1997 using one-month forward exchange rates taken from Datastream. For Norway and Sweden, we use CIP relative to GBP from January 1976 to March 1997, using spot and one-month forward exchange rates from Datastream. Turning to macroeconomic data, we use non-seasonally adjusted M1 data to measure money supply. For the UK, we use M0 due to the unavailability of M1 data. To construct these times series, we combine IFS and national central bank data from Ecowin.17 We deseasonalize the money supply data by implementing the procedure of Gomez and Maravall (2000). 16 Before the introduction of the euro in January 1999, we use the US dollar-Deutsche mark exchange rate combined with the o¢ cial conversion rate between the Deutsche mark and the euro. 17 For Germany regarding the period of January 1976 to December 1979, we construct the money supply using data on currency outside banks and demand deposits from IFS. 16 The price level is measured by the monthly consumer price index (CPI) obtained from the OECD’s Main Economic Indicators (MEI). For Australia and New Zealand, CPI data are published at quarterly frequency and hence monthly observations are constructed by linear interpolation. For the in‡ation rate we use an annual measure computed as the 12-month log di¤erence of the CPI. We de…ne the output gap as deviations from the HP …lter. Since GDP data are generally available quarterly, we proxy real output by the seasonally adjusted monthly industrial production index (IPI) taken from IFS. For Australia, New Zealand, and Switzerland, however, IPI data are only released at quarterly frequency and hence we obtain monthly observations via linear interpolation.18 Orphanides (2001) has recently stressed the importance of using real-time data to estimate Taylor rules for the United States, which are data available to central banks when the policy decisions are made. Since real-time data are not available for most of the countries included in this study, we mimic as closely as possible the information set available to the central banks using quasi-real time data: although data incorporate revisions, we update the HP trend each period so that ex-post data is not used to construct the output gap. In other words, at time t we only use data up to t 1 to construct the output gap. Using a number of detrending methods, Orphanides and van Norden (2002) show that most of the di¤erence between fully revised and real-time data comes from using ex post data to construct potential output and not from the data revisions themselves.19 We convert all data but interest rates by taking logs and multiplying by 100. Throughout the rest of the chapter, the symbols st , it , mt , pt , t, yt and ytg refer to transformed spot exchange rate, interest rate, money supply, price level, in‡ation rate„real output and output gap, respectively. We use an asterisk to denote the transformed data (it , mt , pt , t, yt and ytg ) for the foreign country. Table 2 reports the descriptive statistics for the monthly % FX returns, tween domestic an foreign interest rates, it it ; the di¤erence in % change in price levels, the di¤erence in % change in money supply, output, (yt st ; the di¤erence be- (mt (pt pt ); mt ); and the di¤erence in % change in real yt ). For our sample period, the monthly sample means of the FX returns range from 0:138% for SEK to 0:296% for JPY. The return standard deviations are similar across all exchange rates at about 3% per month. Most FX returns exhibit negative skewness and higher than normal kurtosis. Finally, the exchange rate return sample autocorrelations are no higher than 0:10 and decay rapidly. For the economic fundamentals the notable trends are as follows: (i) it persistent with long memory; (ii) and (yt (pt pt ) are always negatively skewed; and (iii) it are highly (mt mt ) yt ) have occasionally high kurtosis. 18 For New Zealand, IPI data are only available from June 1977. We …ll the gap using quarterly GDP data. The output gap for the …rst period is computed using real output data from January 1970 to January 1976. In the HP …lter, we use a smoothing parameter equal to 14,400 as in Molodtsova and Papell (2009). 19 17 6.2 Predictive Regressions We test the empirical performance of the models by …rst estimating the six predictive regressions for nine monthly exchange rates. The regressions include the random walk (RW) model, uncovered interest parity (UIP), purchasing power parity (PPP), monetary fundamentals (MF), symmetric Taylor rule (TRs ) and asymmetric Taylor rule (TRa ). Table 3 presents the OLS estimates with Newey and West (1987) standard errors. We focus primarily on the signi…cance of the slope estimate of the predictive regressions since this would be an indication that the RW benchmark is misspeci…ed. Consistent with the large literature on the forward premium puzzle, the UIP is predominantly negative. The PPP is always positive and for TRa it is always negative. For these three cases (UIP, PPP and TRa ), the estimates are signi…cant for half of the exchange rates. The least signi…cant slopes are for MF revolving around zero and for TRs for which they are always negative. Finally, the 2 of the predictive regressions is as high as 2:4% but in most cases it is below 1%. In conclusion, Roos the predictive regression results demonstrate that the empirical exchange rates models with the most signi…cant slopes are the UIP, PPP and TRa . 6.3 Statistical Evaluation We assess the statistical performance of the empirical exchange rate models (UIP, PPP, MF, TRs and TRa ) by reporting out-of-sample tests of predictability against the null of the RW. We focus on the 2 following statistics: (i) the Roos statistic of Campbell and Thompson (2008). Recall that a positive 2 Roos value implies that the alternative model has lower MSE than the benchmark RW. However, 2 may be consistent with a better performing alternative because the even a slightly negative Roos 2 calculation of the Roos does take into account the adjustment in the MSE proposed by Clark and West (2006, 2007) to account for the noise introduced in forecasting by estimating a parameter that is not helpful in prediction; (ii) the RM SE statistic, a positive value for which denotes superior 2 ; (iii) OOS performance for the competing model but is subject to the same criticism as the Roos the Clark and McCracken (2001) EN C-F statistic; (iv) the McCracken (2007) M SE-F statistic; and (v) the Clark and West (2006, 2007) M SE-t statistic. The null hypothesis for the EN C-F , M SE-F and M SE-t statistics is that the MSEs for the random walk and the competing model are equal against the alternative that the competing model has lower MSE. One-sided critical values are obtained by generating 10,000 bootstrapped time series as in Mark (1995) and Kilian (1999). The OOS monthly forecasts are obtained in two ways: (i) with rolling regressions that use a 10-year window that generates forecasts for the period of January 1986 to June 2010; and (ii) with recursive regressions for the same forecasting period that successively re-estimate the model parameters every time a new observation is added to the sample. 18 Table 4 shows that most of the statistics tend to be negative and hence provide evidence against the alternative model. In many cases, however, the results are not statistically signi…cant. If instead we focus on the Clark and West (2006, 2007) M SE-t statistic, which makes the adjustment to the MSE and is hence more reliable, a di¤erent picture emerges. For rolling regressions, the UIP and PPP models have a positive M SE-t statistic for seven of the nine exchange rates, whereas the MF and TRa models for six. The model that is most often signi…cantly di¤erent from the RW is the TRa . The worse performing model is the TRs . The results are very similar for recursive regressions. In short, therefore, a careful examination of the empirical evidence reveals that many of the models perform well against the RW with the clear exception of the TRs . It is important to note that in out-of-sample predictive regressions, lack of statistical signi…cance does not imply lack of economic signi…cance. Campbell and Thompson (2008) show that a small R2 can generate large economic bene…ts for investors. They use a mean-variance framework to demonstrate that a good way to judge the magnitude of R2 is to compare it to the square of the Sharpe ratio (SR2 ). Even a modest R2 can lead to a substantial proportional increase in the expected return by conditioning on the predictive variable xt . Indeed, regressions with large R2 statistics would be too pro…table to believe, which is equivalent to the saying: “if you are so smart, why aren’t you rich?” In the limit, an R2 close to 1 should lead to perfect predictions and hence in…nite pro…ts for investors. Furthermore, dynamic asset allocation is by design multivariate thus exploiting predictability in all exchange rate series. In the following section, we discuss in detail whether the predictive regressions can generate economic value. 6.4 Economic Evaluation We assess the economic value of exchange rate predictability by analyzing the performance of dynamically rebalanced portfolios based on one-month ahead forecasts from the six empirical exchange rate models we estimate. The economic evaluation is conducted both IS and OOS, but again the main focus of our analysis is OOS. The OOS results we present in this section are based on forecasts constructed according to a recursive procedure that conditions only upon information up to the month that the forecast is made. The predictive regressions are then successively re-estimated every month. Our empirical analysis focuses on the Sharpe ratio (SR), the Sortino ratio (SO), the Fleming, Kirby and Ostdiek (2001) performance fee (F), the Goetzmann, Ingersoll, Spiegel and Welch (2007) premium return measure (P) and the break even transaction cost be . The F and P performance measures are computed for three cases: (i) zero transaction costs; (ii) a bid-ask spread of 8 bps; and (iii) a bid-ask spread of 20 bps at the beginning of the sample that linearly decays to 4 bps at the end of the sample as suggested by Neely, Weller and Ulrich (2009). Following Della Corte, Sarno 19 and Tsiakas (2009, 2011) our empirical analysis focuses on the maximum expected return strategy as this is the strategy most often used in active currency management. We set a volatility target of p = 10% and a degree of RRA = 6. We have experimented with di¤erent p and values and found that qualitatively they have little e¤ect on the asset allocation results discussed below. Table 5 reports the IS and OOS portfolio performance and shows that there is high economic value associated with some of the empirical exchange rate models. We …rst discuss the IS results, which demonstrate that all models outperform the RW, except for the symmetric Taylor rule (TRs ). For example, SR = 1:30 for PPP, 1:28 for TRa , 1:18 for MF, 1:14 for UIP, 1:08 for RW and 0:96 for TRs . The SO have higher values ranging from 1:26 for TRs to 2:00 for PPP. Switching from the benchmark RW to another model generates F = 285 annual bps for PPP, 202 bps for TRa ; 138 for MF and 83 bps for UIP. The P performance measure has similar value to F. Furthermore, both measures are largely una¤ected by transaction costs. This can be exempli…ed by the very large value of the monthly be , which are 586 bps for UIP, 328 bps for MF and 138 bps for PPP. The literature on exchange rate forecasting is primarily concerned with out-of-sample predictability and hence we turn our attention to the OOS results. The …rst thing to notice is that the value of the OOS SR is smaller than IS. The RW has an OOS SR = 0:54 and is outperformed only by the PPP (SR = 0:76), UIP (SR = 0:65) and TRa (SR = 0:65). Consistent with a very large literature in FX, monetary fundamentals models do not outperform the RW and neither does the TRs . The F values are 252 annual bps for PPP, 131 bps for UIP, 130 bps for TRa , 10 bps for MF and 384 bps for TRs . The P measure has slightly higher value than F. Transaction costs seem to be a bit more important OOS than IS. For example, the be are 173 bps for UIP, 161 bps for TRa and 70 bps for PPP. However, it seems that whether we assume …xed transaction costs or linearly decaying costs makes little di¤erence in the performance of the empirical exchange rate models. In short, our …ndings demonstrate that it is worth using the UIP, PPP and TRa empirical exchange rate models as their forecasts generate signi…cant economic value. By design, the dynamic FX strategy invests in nine foreign bonds and thus exploits predictability in nine exchange rates. Since we economically evaluate the performance of portfolios rather than individual exchange rates, it would be interesting to assess whether the superior portfolio performance of one versus another empirical model is driven by one particular currency. Table 6 reports the economic value of exchange rate predictability when we remove one of the currencies (and hence one of the bonds) from the investment opportunity set. For example, AUD in Table 6 denotes the dynamic allocation strategy that invests in all currencies, except for AUD. The results for excluding one currency at a time show that the best performing models are still the same as before. In sample, all models but the TRs outperform the RW, whereas out of sample the UIP, PPP and TRa are still the best models. Therefore, the empirical evidence suggests that our results are not driven by any 20 one particular currency. A unique feature of the FX market is that investors trade currencies but all prices are quoted relative to a numeraire. Consistent with the vast majority of the FX literature, we use data on exchange rates relative to the US dollar. It is of interest, however, to check whether using a di¤erent currency as numeraire meaningfully a¤ects the economic value of the empirical exchange rate models. This is a crucially important robustness check since it is straightforward to show analytically that the portfolio returns and their covariance matrix are not invariant to the numeraire. For example, consider taking the point of view of a European investor and hence changing the numeraire currency from the US dollar to the euro. Then, all previously bilateral exchange rates become cross rates and nine of the previously cross rates become bilateral. Furthermore, converting dollar FX returns into euro FX returns replaces the US bond as the domestic asset by the European bond. It also replaces all US economic and monetary fundamentals by Europe’s fundamentals. The main question, however, can only be answered empirically: if changing the numeraire also changes the portfolio returns, does the economic value of the empirical exchange rate models also change? Table 7 shows the IS and OOS economic value of exchange rate predictability from the perspective of each of nine countries other than the US. For example, using the AUD as numeraire means that all exchange rates are quoted relative to AUD, all predictive regressions are estimated using the new exchange rates and the mean-variance economic evaluation is done from the perspective of an Australian investor. The same holds when the numeraire changes to CAD, CHF, EUR, GBP, JPY, NOK, NZD and SEK. We …nd that our main results remain robust across all numeraires: the best IS and OOS models are consistently the UIP, the PPP and the TRa . In terms of Sharpe ratios and performance fees, IS the PPP and TRa outperform the RW for all nine numeraires and UIP does so six of nine times; OOS the PPP outperforms the RW seven of nine times, whereas the UIP and TRa …ve of nine times. To conclude, the economic value of exchange rate predictability of the best individual empirical exchange rate models remains robust regardless of the numeraire choice. In addition to the results associated with individual models, even stronger economic evidence is found for the combined forecasts reported in Table 8. In all cases, forecast combinations signi…cantly outperform the RW model. In fact, the best performing model averaging strategies are those based on the SR. For example, the SR( = 12) strategy generates: (i) SR = 0:76 compared to the RW where SR = 0:54, and (ii) F = 254 annual bps with be = 128 monthly bps. It is noteworthy that the simple model average strategy using the mean forecast also generates a high SR = 0:74 and F = 234 bps. Another trend worth mentioning is that the degree of discounting ( ) or the length of the most recently best period ( ) have little or no e¤ect on the performance of combined forecasts. In short, therefore, there is clear out-of-sample economic evidence on the superiority of combined forecasts relative to the RW benchmark that tends to be robust to the way combined forecasts are formed. 21 Finally, Figure 1 illustrates that the OOS Sharpe ratios for the three best performing individual models (UIP, PPP and TRa ) and the SR( = 60) forecast combination against the RW. 7 Conclusion Thirty years of empirical research in international …nance has attempted to resolve whether exchange rates are predictable. Most of this literature uses statistical criteria for out-of-sample tests of the null of the random walk representing no predictability against the alternative of linear models that condition on economic fundamentals. The results of these studies are speci…c to, among other things, the empirical model and the exchange rate series. An emerging literature has moved in a di¤erent direction by providing an economic evaluation of predictability. This second line of research takes the view of an investor who builds a dynamic asset allocation strategy that conditions on the forecasts from a set of empirical exchange rate models. The results of these studies are also speci…c to the empirical model, but instead of providing results for one exchange rate at a time, they evaluate predictability by looking at the performance of dynamically rebalanced portfolios. Finally, there is a third strand of empirical work that forms ex ante combined forecasts from a set of individual empirical models. The results of these studies are not particular to an empirical model but rather relate to forecast combinations that account for model uncertainty. This chapter reviews and connects these three loosely related literatures. We illustrate the statistical and economic methodologies by estimating a set of widely used empirical exchange rate models using monthly returns from nine major US dollar exchange rates. In line with Campbell and Thompson (2008), we show that modest statistical signi…cance can generate large economic bene…ts for investors with a dynamic FX portfolio strategy. We …nd three main results: (i) empirical models based on uncovered interest parity, purchasing power parity and the asymmetric Taylor rule perform better than the random walk in out-of-sample forecasting using both statistical and economic criteria; (ii) conditioning on monetary fundamentals or using a symmetric Taylor rule does not generate economic value out of sample; and (iii) combined forecasts formed using a variety of model averaging methods perform better than individual empirical models. These results are robust to reasonably high transaction costs, the choice of numeraire and the exclusion of any one currency from the investment opportunity set. A Appendix: The Bootstrap Algorithm This appendix summarizes the bootstrap algorithm we use for generating critical values for the OOS test statistics under the null of no exchange rate predictability against a one-sided alternative of linear predictability. Following Mark (1995) and Kilian (1999), the algorithm consists of the following steps: 22 T 1 1. De…ne the IS period for f st+1 ; xt gM t=1 and the OOS period for f st+1 ; xt gt=M +1 . We gen- erate P = (T 1) regression: M OOS forecasts f st+1jt ; sbt+1jt gTt=M1 +1 by estimating the predictive st+1 = + xt + "t+1 and then computing the test statistic of interest, b. 2. De…ne the data generating process (DGP) as st+1 = + xt + u1;t+1 xt = + 1 xt 1 + ::: + and estimate this model subject to the constraint that full sample of observations f st+1 ; xt gTt=11 . p xt p + u2;t ; in the …rst equation is zero, using the The lag order p in the second equation is determined by a suitable lag order selection criterion such as the Bayesian information criterion (BIC). 3. Generate a sequence of pseudo-observations st ; xt st+1 = b + u1;t+1 xt = b + b1 xt 1 T 1 1 t=1 + : : : + bp xt as follows: p + u2;t : The pseudo-innovation term ut = (u1;t ; u2;t )0 is randomly drawn with replacement from the set of observed residuals u bt = (b u1;t ; u b2;t )0 . The initial observations xt 1 ; : : : ; xt p 0 are randomly drawn from the actual data. Repeat this step B = 10; 000 times. 4. For each of the B bootstrap replications, de…ne an IS period for period for st+1 ; xt T 1 . t=M +1 st+1 ; xt M , t=1 and an OOS Then, generate P OOS forecasts f st+1jt ; sbt+1jt gTt=M1 +1 by estimating the predictive regression: st+1 = + xt + u1;t+1 both under the null and the alternative for t = M + 1; : : : ; T 1, and construct the test statistic of interest, b . 5. Compute the one-sided p-value of b as: p-value = B 1 X I(b > b); B j=1 where I ( ) denotes an indicator function, which is equal to 1 when its argument is true and 0 otherwise. 23 Table 1: Data Sources The table presents a detailed description of the sources of the raw data. The exchange rate data range from January 1976 to June 2010. The riskless rate and the money supply data range from January 1976 to May 2005. Data on real output range from January 1970 to May 2010 and are used to construct the output gap. Data on the price level range from January 1975 to May 2010 and are used to construct the inflation rate. The data are monthly but quarterly data are used to retrieve monthly observations via linear interpolation when monthly data are not available. The raw money supply is not seasonally adjusted but the raw real output is. Country Description Source Australia Canada Switzerland Germany Spot Spot Spot Spot Spot Spot Spot Spot Spot Spot Federal Federal Federal Federal Federal Federal Federal Federal Federal Federal UK Japan Norway New Zealand Sweden Australia Canada Switzerland Germany UK Japan Norway New Zealand Sweden US USD/AUD CAD/USD CHF/USD DEM/USD USD/EUR USD/GBP JPY/USD NOK/USD USD/NZD SEK/USD Money Market Rate Spot AUD/USD 1M Fwd AUD/USD 1M Euro Deposit Rate 1M Euro Deposit Rate 1M Euro Deposit Rate 1M Euro Deposit Rate 1M Euro Deposit Rate Spot JPY/USD 1M Fwd JPY/USD 1M Euro Deposit Rate Spot NOK/GBP 1M Fwd NOK/GBP 1M Euro Deposit Rate Money Market Rate Spot NZD/USD 1M Fwd NZD/USD 1M Euro deposit rate Spot SEK/GBP 1M Fwd SEK/GBP 1M Euro deposit rate 1M Euro deposit rate Range Nominal Exchange Rate Reserve Board 76:01-10:06 Reserve Board 76:01-10:06 Reserve Board 76:01-10:06 Reserve Board 76:01-98:12 Reserve Board 99:01-10:06 Reserve Board 76:01-10:06 Reserve Board 76:01-10:06 Reserve Board 76:01-10:06 Reserve Board 76:01-10:06 Reserve Board 76:01-10:06 Frequency Series Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly DDP [RXI$US_N.B.AL] DDP [RXI_N.B.CA] DDP [RXI_N.B.SZ] H.10 Historical Rates DDP [RXI$US_N.B.EU] DDP [RXI$US_N.B.UK] DDP [RXI_N.B.JA] DDP [RXI_N.B.NO] DDP [RXI$US_N.B.NZ] DDP [RXI_N.B.SD] Riskless Rate IMF IFS 76:01-84:11 Barclays Bank 84:12-97:03 Barclays Bank 84:12-97:03 Thomson Reuters 97:04-10:05 Thomson Reuters 76:01-10:05 Thomson Reuters 76:01-10:05 Thomson Reuters 76:01-10:05 Thomson Reuters 76:01-10:05 Hai, Mark and Wu (1997) 76:01-78:12 Hai, Mark and Wu (1997) 76:01-78:12 Thomson Reuters 79:01-10:05 Not Specified 76:01-97:03 Not Specified 76:01-97:03 Thomson Reuters 97:04-10:05 IMF IFS 76:01-84:11 Barclays Bank 84:12-97:03 Barclays Bank 84:12-97:03 Thomson Reuters 97:04-10:05 Not Specified 76:01-97:03 Not Specified 76:01-97:03 Thomson Reuters 97:04-10:05 Thomson Reuters 76:01-10:05 Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Ecowin [ifs:s19360b00zfm] Datastream [BBAUDSP] Datastream [BBAUD1F] Datastream [ECCAD1M] Datastream [ECAUD1M] Datastream [ECSWF1M] Datastream [ECWGM1M] Datastream [ECUKP1M] Nelson Mark’s website Nelson Mark’s website Datastream [ECJAP1M] Datastream [NORKRON] Datastream [NORKN1F] Datastream [ECNOR1M] Ecowin [ifs:s19660000zfm] Datastream [BBNZDSP] Datastream [BBNZD1F] Datastream [ECNZD1M] Datastream [SWEKRON] Datastream [SWEDK1F] Datastream [ECSWE1M] Datastream [ECUSD1M] (continued) 24 Table 1: Data Sources (continued) Country Description Australia Canada Switzerland UK Japan Norway M1 M1 M1 M1 Currency in Circulation Demand Deposits M1 M0 M1 M1 New Zealand M1 Sweden M1 US M1 Range Money Supply Reserve Bank of Australia 76:01-10:05 Bank of Canada 76:01-10:05 IMF IFS 76:01-84:11 Swiss National Bank 84:12-10:05 IMF IFS 76:01-79:12 IMF IFS 76:01-79:12 Deutsche Bundesbank 80:01-10:05 Bank of England 76:01-10:05 Bank of Japan 76:01-10:05 IMF IFS 76:01-86:12 Norges Bank 87:01-10:05 IMF IFS 76:01-77:02 Reserve Bank of New Zealand 77:03-10:05 IMF IFS 76:01-98:02 Sveriges Riksbank 98:03-10:05 Federal Reserve United States 76:01-10:05 Australia Canada Switzerland Germany UK Japan Norway New Zealand Sweden United States Industrial Production Index Industrial Production Index Industrial Production Index Industrial Production Index Industrial Production Index Industrial Production Index Industrial Production Index Gross Domestic Product Industrial Production Index Industrial Production Index Industrial Production Index IMF IMF IMF IMF IMF IMF IMF IMF IMF IMF IMF Australia Canada Switzerland Germany UK Japan Norway New Zealand Sweden United States Consumer Consumer Consumer Consumer Consumer Consumer Consumer Consumer Consumer Consumer OECD OECD OECD OECD OECD OECD OECD OECD OECD OECD Germany Source Frequency Series Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly Monthly EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin [ew:aus12045] [ew:can12042] [ifs:s14634000zfm] [ew:che12045] [ifs:s13434a0nzfm] [ifs:s13434b0nzfm] [ew:deu12990] [boe:lpmavaa] [ew:jpn12066] [ifs:s14234000zfm] [ew:nor12045] [ifs:s19634000zfm] [ew:nzl12045] [ifs:s14435l00zfm] [ew:swe12010] [ew:usa12010] 69:12-10:06 70:01-10:05 69:04-10:06 70:01-10:05 70:01-10:05 70:01-10:05 70:01-10:05 69:12-77:05 77:06-10:06 70:01-10:05 70:01-10:05 Quarterly Monthly Quarterly Monthly Monthly Monthly Monthly Quarterly Quarterly Monthly Monthly EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin [ifs:s1936600czfq] [ifs:s1566600czfm] [ifs:s1466600bzfq] [ifs:s1346600czfm] [ifs:s1126600czfm] [ifs:s1586600czfm] [ifs:s1426600czfm] [ifs:s19699b0czfy] [ifs:s19666eyczfq] [ifs:s1446600czfm] [ifs:s1116600czfm] 74:12-10:06 75:01-10:05 75:01-10:05 75:01-10:05 75:01-10:05 75:01-10:05 75:01-10:05 74:12-10:06 75:01-10:05 75:01-10:05 Quarterly Monthly Monthly Monthly Monthly Monthly Monthly Quarterly Monthly Monthly EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin EcoWin [oecd:aus_cpalcy01_ixobq] [oecd:can_cpaltt01_ixobm] [oecd:che_cpaltt01_ixobm] [oecd:deu_cpaltt01_ixobm] [oecd:gbr_cpaltt01_ixobm] [oecd:jpn_cpaltt01_ixobm] [oecd:nor_cpaltt01_ixobm] [oecd:nzl_cpalcy01_ixobq] [oecd:swe_cpaltt01_ixobm] [oecd:usa_cpaltt01_ixobm] Real Output IFS IFS IFS IFS IFS IFS IFS IFS IFS IFS IFS Price Level Price Price Price Price Price Price Price Price Price Price Index Index Index Index Index Index Index Index Index Index MEI MEI MEI MEI MEI MEI MEI MEI MEI MEI 25 Table 2. Descriptive Statistics The table presents descriptive statistics for nine major exchange rates and a set of economic fundamentals. s is the % change in the US dollar exchange rate vis-à-vis the Australian dollar (AUD), Canadian dollar (CAD), Swiss franc (CHF), Deutsche markneuro (EUR), British pound (GBP), Japanese yen (JPY), Norwegian kroner (NOK), New Zealand dollar (NZD) and Swedish kronor (SEK); i is the one-month interest rate; p is the % change in the price level; m is the % change in the money supply; y is the % change in real output; and the asterisk denotes a non-US value. The exchange rate is de…ned as US dollars per unit of foreign currency. l is the autocorrelation coe¢ cient with l lags. The data range from January 1976 to June 2010 for a sample size of 414 monthly observations. M ean AU D s i CAD i (p p ) (m m ) (y y ) s i CHF i (p p ) (m m ) (y y ) s i EU R i (p p ) (m m ) (y y ) s i GBP i (p p ) (m m ) (y y ) s i JP Y i (p p ) (m m ) (y y ) s i N OK i (p p ) (m m ) (y y ) s i N ZD i (p p ) (m m ) (y y ) s i SEK i (p p ) (m m ) (y y ) s i i (p p ) (m m ) (y y ) Std Skew Kurt 1 3 6 12 0:089 0:175 0:088 0:350 0:026 3:237 0:290 0:405 1:528 0:785 1:396 0:092 0:803 2:880 0:218 9:440 4:547 4:304 35:059 4:438 0:056 0:956 0:560 0:009 0:278 0:046 0:873 0:138 0:167 0:035 0:002 0:795 0:103 0:078 0:038 0:100 0:643 0:259 0:040 0:068 0:009 0:063 0:005 0:256 0:016 1:907 0:140 0:365 1:004 1:089 0:672 0:223 0:739 0:354 0:890 10:792 3:646 6:657 4:124 11:287 0:033 0:885 0:008 0:076 0:216 0:017 0:732 0:079 0:230 0:062 0:121 0:648 0:080 0:128 0:035 0:002 0:440 0:240 0:102 0:067 0:210 0:256 0:157 0:037 0:015 3:528 0:279 0:419 1:531 0:949 0:038 0:443 0:123 0:443 0:105 3:921 3:768 3:724 10:496 4:278 0:013 0:970 0:231 0:173 0:355 0:023 0:890 0:054 0:072 0:036 0:092 0:834 0:330 0:205 0:020 0:029 0:747 0:509 0:003 0:033 0:119 0:116 0:132 0:183 0:071 3:218 0:239 0:411 1:442 1:807 0:147 0:249 0:770 0:929 0:011 3:730 3:615 5:174 30:164 9:269 0:024 0:961 0:171 0:080 0:286 0:036 0:875 0:031 0:009 0:079 0:057 0:827 0:078 0:167 0:007 0:022 0:723 0:504 0:060 0:003 0:073 0:176 0:063 0:091 0:121 3:065 0:220 0:559 0:942 1:340 0:205 0:674 1:325 0:475 0:281 4:793 5:378 8:335 6:410 6:516 0:092 0:921 0:087 0:033 0:266 0:018 0:763 0:012 0:195 0:068 0:078 0:573 0:235 0:220 0:063 0:002 0:296 0:551 0:108 0:063 0:296 0:247 0:202 0:132 0:026 3:361 0:221 0:529 1:795 1:604 0:365 0:327 0:399 4:414 0:568 4:325 3:446 5:214 52:393 5:895 0:039 0:952 0:065 0:036 0:081 0:042 0:829 0:104 0:018 0:105 0:105 0:671 0:110 0:050 0:028 0:035 0:445 0:468 0:086 0:024 0:037 0:181 0:038 0:475 0:033 2:965 0:277 0:534 3:099 4:358 0:489 0:013 0:884 1:950 0:696 4:695 3:043 5:370 17:999 28:252 0:067 0:934 0:192 0:184 0:425 0:022 0:831 0:112 0:097 0:034 0:038 0:744 0:162 0:114 0:006 0:082 0:499 0:427 0:025 0:095 0:093 0:296 0:169 0:225 0:073 3:366 0:350 0:495 1:925 1:071 1:232 1:666 1:414 0:213 1:232 11:815 7:590 7:351 4:944 9:601 0:036 0:957 0:690 0:207 0:477 0:185 0:859 0:361 0:212 0:113 0:035 0:736 0:326 0:124 0:062 0:098 0:546 0:330 0:168 0:001 0:138 0:148 0:045 0:184 0:107 3:192 0:277 0:546 1:443 2:803 0:876 0:826 1:103 1:650 0:086 6:427 4:141 6:226 21:307 38:324 0:104 0:937 0:199 0:010 0:297 0:062 0:837 0:048 0:050 0:047 0:104 0:749 0:102 0:124 0:030 0:034 0:543 0:342 0:023 0:026 26 Table 3. Predictive Regressions The table reports the least squares estimates of the predictive regression st+1 = + xt +"t+1 for nine major exchange rates de…ned as US dollars per unit of foreign currency. st is the monthly % exchange rate return. The random walk (RW) model sets = 0; the uncovered interest parity (UIP) model sets xt = it it , which is the interest rate di¤erential between the home and foreign country; the purchasing power parity (PPP) model sets xt = pt pt st , where pt pt is the log price di¤erential; the monetary fundamentals (MF) model sets xt = (mt mt ) (yt yt ) st , where mt mt is the the log money supply di¤erential and yt yt the log g real output di¤erential; the symmetric Taylor rule (TRs ) sets xt = 1:5 ( t ytg , where t t ) + 0:1 yt t g is the in‡ation di¤erential and yt xt = 1:5 ( t t )+0:1 ytg ytg ytg the real output gap di¤erential; and the asymmetric Taylor rule (TRa ) sets +0:1 (st + pt pt ), where st +pt pt is the log real exchange rate. Newey- West (1987) standard errors are reported in parentheses. The superscripts a, b, and c indicate statistical signi…cance at the 10%, 5%, and 1% level, respectively. The sample period comprises monthly observations from January 1976 to June 2010. RW 0:089 AU D (0:170) 2 Roos (%) CAD 0:009 (0:097) U IP 0:154 PPP 0:047 0:368 (0:160) (%) CHF 0:210 (0:182) (%) EU R 0:119 (0:171) (%) GBP 0:073 (0:168) (%) 0:296 a JP Y (0:178) N OK 0:037 (0:161) (%) N ZD 0:093 (0:184) (%) SEK 0:138 (0:182) (0:092) 0:049 0:816 0:070 0:288 (0:116) (%) 0:640 (0:185) a 0:015 0:008 (0:929) a (0:097) 0:003 0:025 0:273 (0:172) a 0:139 (0:586) (0:008) (0:004) (0:440) (0:074) 0:512 0:828 0:328 0:001 0:790 0:498 b 1:735 (0:250) 1:632 (1:149) 0:023 0:322 (1:075) a 0:009 (0:236) a 0:483 1:673 (1:067) a 0:226 (0:720) (0:014) (0:005) (0:686) (0:129) 0:793 1:028 0:690 0:173 1:104 0:192 0:386 (0:190) (0:374) 0:016 0:225 (0:304) 0:004 0:214 (0:194) 0:458 0:338 (0:335) 0:157 (0:789) (0:011) (0:004) (0:606) (0:103) 0:218 0:691 0:531 0:182 0:732 0:342 a 2:079 (0:199) (1:387) a 0:028 1:620 (1:504) 0:005 0:106 (0:165) 0:302 1:896 (1:218) a 0:257 (0:887) (0:018) (0:005) (0:511) (0:155) 1:204 1:196 0:366 0:135 1:224 0:881 c 9:156 a (0:211) c (5:379) (0:655) (0:010) 2:431 0:977 0:018 a 0:128 3:466 (0:173) (2:295) 0:020 0:349 a (0:193) 0:487 b 9:685 a (0:194) a (5:145) (0:002) (0:399) (0:096) 0:332 0:396 1:150 0:003 0:065 (1:240) 0:001 0:654 0:046 (0:162) 0:253 0:186 a 3:076 (2:022) 0:174 (0:551) (0:013) (0:003) (0:503) (0:115) 0:225 0:779 0:001 0:161 0:800 0:387 a 0:041 (0:218) (0:274) a 0:009 0:590 (2:797) 0:001 0:165 (0:191) 0:284 0:018 (0:239) 0:095 (0:516) (0:011) (0:006) (0:318) (0:097) 1:073 0:222 0:029 0:212 0:295 2:619 a 0:130 (0:183) (1:579) 0:051 2 Roos (0:174) (0:486) 0:995 2 Roos 0:178 a 0:033 0:507 2 Roos 0:208 (0:154) (0:005) 2:369 2 Roos (%) 0:001 0:821 1:529 2 Roos 0:019 a (0:010) 0:630 2 Roos T Ra 0:076 0:108 1:126 2 Roos (1:380) T Rs 0:120 (0:433) 0:974 2 Roos (0:174) MF 0:251 0:015 1:862 (1:636) 0:008 0:145 (0:189) 0:125 2:668 (1:642) 0:150 (0:863) (0:009) (0:007) (0:465) (0:096) 0:002 0:807 0:493 0:024 0:849 27 Table 4. Statistical Evaluation of Exchange Rate Predictability The table displays out-of-sample tests of the predictive ability of a set of empirical exchange rate models against the null of a random walk (RW). In addition to RW, we form exchange rate forecasts using five alternative models: uncovered interest parity (UIP), purchasing power parity (PPP), monetary fundamentals (MF), symmetric Taylor rule (TR ) and asymmetric Taylor rule (TR ). The out-of-sample monthly forecasts are obtained in two ways: (i) with rolling regressions that use a 10-year window generating forecasts for the period of January 1986 to June 2010; and (ii) with recursive regressions for the same forecasting period that successively re-estimate the model 2 is the Campbell and Thompson (2008) statistic. ∆ is the root mean squared error difference parameters every time a new observation is added to the sample. between the RW and the competing model. - is the Clark and McCracken (2001) F -statistic, - is the McCracken (2007) F -statistic and - is the Clark and West (2006, 2007) t-statistic, all of which test the null hypothesis of equal mean squared error (MSE) between the RW and the competing model; the alternative hypothesis is that the competing model has lower MSE. One-sided critical values are obtained by generating 10,000 bootstrap samples as in Mark (1995) and Kilian (1999). Significance levels at 90%, 95%, and 99% are denoted by a, b, and c, respectively. 2 (%) ∆ (%) - - - 2 (%) ∆ (%) - - - 2 (%) ∆ (%) - - - −098 −002 187 −285 060 −334 −005 044 −946 011 −274 −004 −231 −780 −139 −188 −003 −186 −541 −143 −192 −003 −037 −551 −019 −138 −002 138 −399 060 −274 −004 253 −782 076 −133 −002 007 −385 005 −263 −004 −121 −751 −073 −214 −004 −117 −614 −044 −145 −002 −098 −420 −062 −309 −005 029 −878 009 −152 −003 176 −439 070 −319 −005 −090 −907 −037 053 001 289 156 173 −336 −005 254 −951 047 002 001 328 007 114 −081 −001 204 −235 070 Rolling Regressions −135 −169 −263 −001 −002 −003 −104 −040 −292 −391 −487 −751 −079 −024 −130 057 001 375 167 174 −162 −003 −142 −468 −126 −087 −001 262 −252 142 −180 −003 022 −518 009 −205 −003 −178 −589 −105 −084 −001 130 −243 069 −111 −001 −007 −321 −004 129 002 498 384 225 −206 −003 125 −593 068 −180 −003 435 −518 100 −039 −001 387 −114 135 −638 −010 −410 −176 −131 −103 −002 214 −298 092 −060 −001 533 −174 162 −017 001 306 −050 095 −148 −002 396 −426 085 −044 −001 092 −128 064 −051 −001 200 −149 087 −182 −003 −123 −524 −057 −375 −006 010 −106 003 −126 −002 077 −364 044 −378 −006 −069 −107 −020 −156 −003 −063 −451 −030 039 001 260 113 125 (continued) 28 Table 4. Statistical Evaluation of Exchange Rate Predictability (continued) 2 (%) ∆ (%) - - - −026 000 −025 −075 −049 034 001 136 100 096 2 (%) ∆ (%) - - - −155 −002 030 −447 010 074 001 144 219 166 2 (%) ∆ (%) - - - −050 −001 −005 −146 −004 068 001 134 202 146 −121 −002 −071 −350 −053 −073 −001 324 −212 110 −142 −002 −143 −411 −129 Recursive Regressions 030 −132 −041 001 −001 000 126 −050 −056 089 −381 −120 116 −022 −113 −034 −001 −039 −098 −104 091 002 229 270 176 006 000 140 018 121 −106 −002 −074 −306 −056 −171 −003 −013 −493 −005 −122 −002 091 −354 029 122 002 265 362 193 −033 000 160 −097 088 −153 −002 −148 −442 −088 101 002 220 300 200 011 000 167 033 085 −037 −001 −039 −107 −061 −065 −001 −054 −190 −062 29 009 001 096 027 068 −168 −003 243 −485 064 111 002 241 329 188 −035 −001 573 −102 178 −069 −001 −079 −200 −114 164 002 386 489 287 071 001 427 209 178 030 001 098 087 091 −167 −003 410 −482 107 −006 000 058 −018 061 115 002 252 341 210 −100 −002 −122 −291 −085 072 001 140 213 159 −028 000 150 −082 071 −081 −001 −030 −236 −017 −237 −004 112 −679 036 −085 −001 −071 −246 −088 −164 −003 −074 −473 −039 −060 −001 −078 −174 −131 077 001 168 227 169 Table 5. The Economic Value of Exchange Rate Predictability The table shows the in-sample and out-of-sample economic value of a set of empirical exchange rate models for nine nominal spot exchange rates relative to the US dollar. We form exchange rate forecasts using six models: the random walk (RW), uncovered interest parity (UIP), purchasing power parity (PPP), monetary fundamentals (MF), symmetric Taylor rule (TRs ) and asymmetric Taylor rule (TRa ). Using the exchange rate forecasts from each model, we build a maximum expected return strategy subject to a target volatility p = 10% for a US investor who every month dynamically rebalances her portfolio investing in a domestic US bond and nine foreign bonds. For each portfolio, we report the annualized % mean ( p ), % volatility ( p ), Sharpe ratio (SR) and Sortino ratio (SO ). F denotes the performance fee a risk-averse investor is willing to pay for switching from the benchmark RW strategy to a competing strategy. P is the premium return performance measure. F and P are computed for a degree of relative risk aversion equal to 6 and are expressed in annual basis points. be is the break-even proportional transaction cost that cancels out the utility advantage of a given strategy relative to the RW. be is only reported for positive performance measures and is expressed in monthly basis points. F (P ) denote the performance fee (premium return) reported net of the bid-ask spread, which is assumed to linearly decay from 20 bps in 1976 to 4 bps in 2010. F8 and P8 are computed for a …xed bid-ask spread of 8 bps. The in-sample analysis covers monthly data from January 1976 to June 2010. The out-of-sample analysis runs from January 1986 to June 2010. Strategy p p SR SO F be P In-Sample F P F8 P8 RW 18:0 10:8 1:08 1:46 U IP 19:1 11:1 1:14 1:68 83 92 586 83 91 83 91 PPP 21:5 11:6 1:30 2:00 285 297 138 285 297 285 297 MF 19:9 11:5 1:18 1:64 138 139 328 138 139 138 139 T Rs 16:9 11:1 0:96 1:26 127 128 127 128 127 128 T Ra 19:8 10:5 1:28 1:86 202 202 202 202 202 202 Out-of-Sample RW 10:8 11:4 0:54 0:73 U IP 11:9 11:1 0:65 0:99 131 143 173 138 153 141 163 PPP 13:3 11:3 0:76 0:97 252 247 70 230 227 231 208 MF 11:1 11:7 0:55 0:73 10 4 3 3 11 5 T Rs 7:1 11:6 0:21 0:23 384 433 395 445 392 451 T Ra 12:1 11:4 0:65 0:83 130 121 162 154 156 174 30 161 Table 6. The Economic Value of Exchange Rate Predictability when Removing one Currency The table shows the in-sample and out-of-sample economic value of a set of empirical exchange rate models when one of the nine foreign currencies is removed from the investment opportunity set. The nine exchange rates include the Australian dollar (AUD), Canadian dollar (CAD), Swiss franc (CHF), Deutsche mark\euro (EUR), British pound (GBP), Japanese yen (JPY), Norwegian kroner (NOK), New Zealand dollar (NZD), and Swedish kronor (SEK) relative to the US dollar (USD). For example, AUD denotes an investment strategy that invests in all currencies except for AUD. We form exchange rate forecasts using six models: the random walk (RW), uncovered interest parity (UIP), purchasing power parity (PPP), monetary fundamentals (MF), symmetric Taylor rule (TR ) and asymmetric Taylor rule (TR ). Using the exchange rate forecasts from each model, we build a maximum expected return strategy subject to a target volatility ∗ = 10% for a US investor who every month dynamically rebalances her portfolio investing in a domestic US bond and nine foreign bonds. For each portfolio, we report the Sharpe ratio (SR) and Sortino ratio (SO). F denotes the performance fee a risk-averse investor is willing to pay for switching from the benchmark RW strategy to a competing strategy. P is the premium return performance measure. F and P are computed for a degree of relative risk aversion equal to 6 and are expressed in annual basis points. is the break-even proportional transaction cost that cancels out the utility advantage of a given strategy relative to the RW. is only reported for positive performance measures and is expressed in monthly basis points. The in-sample analysis covers monthly data from January 1976 to June 2010. The out-of-sample analysis runs from January 1986 to June 2010. SR SO 111 115 125 119 096 121 153 174 194 171 130 178 101 107 119 110 097 127 141 162 185 152 131 183 098 105 115 112 087 115 124 150 176 153 108 171 F P SR SO 163 99 410 − 454 094 099 120 106 081 120 125 140 193 152 105 161 146 109 153 − 62 102 111 122 110 089 118 133 154 174 145 115 159 218 130 171 − 314 110 117 130 119 096 123 152 183 208 174 134 192 AUD 56 219 128 −148 119 64 234 130 −147 120 F P In-Sample CAD EUR 67 235 121 −43 240 69 319 153 −151 253 SR SO 443 169 271 − 176 094 102 118 102 089 119 128 149 188 125 118 175 526 157 216 − 103 096 098 118 107 084 116 130 146 184 148 112 172 76 249 118 −46 238 107 256 117 −134 155 4 149 393 − 88 104 106 124 118 093 128 145 154 187 176 127 192 31 85 278 141 −133 141 P 91 281 97 −58 245 100 296 83 −62 244 127 146 175 − 99 41 284 136 −129 208 41 103 259 − 162 JPY 113 265 117 −134 155 NZD 104 231 187 −120 190 F CHF GBP NOK 89 215 181 −123 182 60 304 150 −150 258 32 271 137 −129 207 SEK 95 291 143 −131 146 35 249 161 −119 213 41 37 259 200 168 404 −118 − 214 206 (continued) Table 6. The Economic Value of Exchange Rate Predictability when Removing one Currency (continued) SR SO 056 067 078 058 020 066 078 102 100 075 023 083 048 058 070 054 026 064 067 091 091 068 030 081 042 058 058 045 021 053 061 091 072 056 024 064 F P SR 158 70 − − 104 046 052 068 056 011 064 SO F P Out-of-Sample CAD 060 069 71 79 088 251 253 077 110 117 012 −397 −439 070 206 176 260 55 149 − 132 050 061 066 048 017 057 067 092 086 065 019 072 135 50 2 − 112 052 065 077 052 016 060 069 092 095 069 017 076 AUD 118 249 10 −416 102 129 242 −1 −465 89 EUR 114 238 50 −271 172 SR SO 169 89 357 − 136 046 057 071 047 023 068 065 087 093 059 026 085 178 59 − − 82 052 056 070 054 017 065 068 079 092 076 019 084 123 228 37 −306 161 127 176 −39 −376 75 162 78 − − 164 049 063 076 051 020 065 070 100 110 066 026 092 32 139 277 −12 −412 83 P 131 287 −1 −262 252 144 275 −21 −312 230 390 64 − − 120 64 213 37 −434 147 57 63 19 − 185 186 312 8 −335 190 2693 114 32 − 157 JPY 139 173 −41 −420 68 NZD 188 166 −14 −261 86 F CHF GBP NOK 178 176 10 −225 108 52 212 33 −392 151 SEK 150 270 −13 −471 77 179 314 27 −328 196 Table 7. The Economic Value of Exchange Rate Predictability for Alternative Numeraires The table presents the in-sample and out-of-sample economic value of a set of empirical exchange rate models for alternative numeraires other than the US dollar. The set of currencies includes the Australian dollar (AUD), Canadian dollar (CAD), Swiss franc (CHF), Deutsche mark\euro (EUR), British pound (GBP), Japanese yen (JPY), Norwegian kroner (NOK), New Zealand dollar (NZD), Swedish kronor (SEK) and the US dollar (USD). For example, AUD denotes an investment strategy using AUD as the domestic currency and expressing all exchange rates relative to AUD. We form exchange rate forecasts using six models: the random walk (RW), uncovered interest parity (UIP), purchasing power parity (PPP), monetary fundamentals (MF), symmetric Taylor rule (TR ) and asymmetric Taylor rule (TR ). Using the exchange rate forecasts from each model, we build a maximum expected return strategy subject to a target volatility ∗ = 10% for a US investor who every month dynamically rebalances her portfolio investing in a domestic US bond and nine foreign bonds. For each portfolio, we report the Sharpe ratio (SR) and Sortino ratio (SO). F denotes the performance fee a risk-averse investor is willing to pay for switching from the benchmark RW strategy to a competing strategy. P is the premium return performance measure. F and P are computed for a degree of relative risk aversion equal to 6 and are expressed in annual basis points. is the break-even proportional transaction cost that cancels out the utility advantage of a given strategy relative to the RW. is only reported for positive performance measures and is expressed in monthly basis points. The in-sample analysis covers monthly data from January 1976 to June 2010. The out-of-sample analysis runs from January 1986 to June 2010. SR SO 108 095 124 082 106 149 149 144 194 129 139 240 109 125 135 110 107 128 144 174 198 167 130 182 108 126 124 131 114 133 150 219 184 167 157 208 F P SR − 54 − − 212 108 110 132 107 093 117 F P In-Sample CAD 147 146 55 55 203 306 317 186 −11 6 137 −163 −155 164 100 103 170 245 8 1 290 107 118 123 104 103 122 147 173 185 158 147 181 827 127 139 8 166 107 105 115 098 106 155 153 146 185 162 147 261 AUD −141 242 −285 −13 370 −127 268 −269 −14 373 SO EUR 159 329 13 5 204 SR SO 114 116 − − 316 109 128 141 104 102 126 144 182 227 153 126 182 61 124 − − 147 109 116 138 108 102 124 147 201 201 151 142 168 162 342 24 1 208 133 222 −37 −50 157 − 28 − − 401 108 100 133 101 111 139 150 151 179 158 136 206 33 −19 129 −114 −11 452 P 200 394 −38 −72 190 207 412 −26 −78 194 137 378 − − 186 140 375 17 −80 183 92 251 23 − 277 JPY 145 236 −23 −49 157 NZD 150 248 239 74 265 F CHF GBP NOK 146 235 243 73 258 115 367 11 −86 180 SEK −14 152 −107 −7 452 −100 340 −62 45 292 −98 − 344 128 −46 − 37 5 294 21 (continued) Table 7. The Economic Value of Exchange Rate Predictability for Alternative Numeraires (continued) SR SO 054 025 054 029 042 046 075 036 083 038 058 063 054 064 068 042 036 033 073 094 097 051 043 039 053 050 065 060 052 067 074 080 092 070 077 094 F P SR 054 066 066 054 024 060 SO F P Out-of-Sample CAD 074 088 143 141 088 130 131 070 20 5 029 −313 −329 077 76 57 053 064 051 031 049 074 074 099 071 043 072 102 053 045 077 036 029 071 075 067 103 042 041 093 AUD −333 −15 −278 −111 −81 −326 −2 −280 −113 −71 − − − − − EUR 113 155 −145 −235 −254 SR SO 169 36 8 − 121 054 066 081 028 033 026 073 091 118 040 045 033 47 − − − 102 054 059 086 036 033 034 073 094 120 053 041 042 121 157 −183 −259 −284 89 56 − − − 122 −52 −250 −38 239 − 45 − − 109 054 052 061 017 047 073 074 077 085 023 054 089 − 37 39 − 278 34 −123 270 −173 −223 205 P 134 318 −320 −227 −333 136 323 −310 −220 −349 92 103 − − − 72 381 −184 −225 −252 23 104 − − − −14 73 −476 −150 198 − 22 − − 903 JPY 137 −47 −244 −34 238 NZD −20 126 49 2 163 F CHF GBP NOK −28 121 73 −3 164 49 379 −191 −222 −237 SEK −111 268 −207 −225 197 −21 70 −471 −110 217 Table 8. The Economic Value of Combined Forecasts The table presents the out-of-sample economic value of combined forecasts across a set of empirical exchange rate models. We form forecasts for the nine exchange rates using combinations of six models: the random walk (RW), uncovered interest parity (UIP), purchasing power parity (PPP), monetary fundamentals (MF), symmetric Taylor rule (TRs ) and asymmetric Taylor rule (TRa ). Simple Model Averaging denotes ex-ante combining methods that disregard the past out-of-sample performance of the individual models and use the mean, median and trimmed mean of the individual forecasts. Statistical Model Averaging denotes ex-ante combining methods based on the past out-of-sample mean-squared error (MSE) of the individual models. DM SE( ) use the inverse of the discounted MSE with as a discount factor. M SE( ) use the inverse of the MSE over the most recent months. Economic Model Averaging denote ex-ante combining methods which use the past out-of-sample Sharpe ratio (SR) of the individual models. DSR( ) use the discounted SR with as a discount factor. SR( ) uses the average SR over the most recent months. Using the forecast combinations, we build a maximum expected return strategy subject to a target volatility p = 10% for a US investor who every month dynamically rebalances her portfolio investing in a domestic US bond and nine foreign bonds. For each portfolio, we report the Sharpe ratio (SR) and Sortino ratio (SO ). F denotes the performance fee a risk-averse investor is willing to pay for switching from the benchmark RW strategy to a competing strategy. P is the premium return performance measure. F and P are computed for a degree of relative risk aversion equal to 6 and are expressed in annual basis points. be is the break-even proportional transaction cost that cancels out the utility advantage of a given strategy relative to the RW. be is only reported for positive performance measures and is expressed in monthly basis points. The in-sample analysis covers monthly data from January 1976 to June 2010. The out-of-sample analysis runs from January 1986 to June 2010. F P be Averaging 0:74 0:89 0:61 0:74 0:60 0:72 234 76 65 206 52 37 81 21 23 Statistical Model Averaging DM SE ( = 0:90) 13:2 11:7 0:72 0:87 DM SE ( = 0:95) 13:2 11:7 0:73 0:88 DM SE ( = 1:00) 13:3 11:6 0:74 0:89 213 218 232 186 191 204 71 74 81 M SE( = 60) M SE( = 36) M SE( = 12) 0:88 0:88 0:87 222 218 207 195 191 179 76 74 70 Economic Model Averaging 13:4 11:5 0:76 0:96 13:4 11:5 0:76 0:95 12:9 11:4 0:72 0:89 254 255 207 235 236 187 128 127 103 261 254 254 241 234 235 129 126 128 RW M ean M edian T rimmed M ean DSR ( = 0:90) DSR ( = 0:95) DSR ( = 1:00) SR( = 60) SR( = 36) SR( = 12) 10:8 11:4 Simple 13:3 12:0 12:0 Model 11:6 12:0 12:1 13:2 13:2 13:1 13:5 13:4 13:4 11:7 11:7 11:7 11:5 11:5 11:5 35 SR SO 0:54 0:73 0:73 0:73 0:72 0:76 0:76 0:76 0:96 0:95 0:96 Figure 1. Out-of-Sample Sharpe Ratios The figure displays the out-of-sample annualized Sharpe ratio (SR) for selected empirical exchange rate models. We show the results from forming exchange rate forecasts using uncovered interest parity (UIP), purchasing power parity (PPP), asymmetric Taylor rule (T Ra ), and the ex-ante forecast combination method that uses the out-of-sample SR over the past 60 months (SR(κ = 60)). All models (solid line) are displayed versus the random walk (RW) benchmark (dashed line). Using the exchange rate forecasts from each model, we build a maximum expected return strategy subject to a target volatility σp∗ = 10% for a US investor who every month dynamically rebalances her portfolio investing in a domestic US bond and nine foreign bonds. The SR is computed using the out-of-sample portfolio returns for one year. The out-of-sample period runs from January 1986 to June 2010. 36 References Abhyankar, A., L. Sarno, and G. Valente (2005). “Exchange Rates and Fundamentals: Evidence on the Economic Value of Predictability,” Journal of International Economics 66, 325–348. Akram, F., D. Rime, and L. Sarno (2008). “Arbitrage in the Foreign Exchange Market: Turning on the Microscope,” Journal of International Economics 76, 237–253. Avramov, D. (2002). “Stock Return Predictability and Model Uncertainty,” Journal of Financial Economics 64, 423–458. Bandi, F.M., and J.R. Russell (2006). “Separating Microstructure Noise from Volatility,”Journal of Financial Economics 79, 655–692. Bandi, F.M., J.R. Russell, and Y. Zhu (2008). “Using High-Frequency Data in Dynamic Portfolio Choice,” Econometric Reviews 27, 163–198. Bates, J.M., and C.W.J. Granger (1969). “The Combination of Forecasts,” Operations Research Quarterly 20, 451–468. Bilson, J.F.O. (1981). “The ‘Speculative E¢ ciency’Hypothesis,” Journal of Business 54, 435–451. Boudoukh, J., M. Richardson, and R.F. Whitelaw (2006). “The Information in Long-Maturity Forward Rates: Implications for Exchange Rates and the Forward Premium Anomaly,” working paper, New York University. Brunnermeier, M.K., S. Nagel, and L.H. Pedersen (2009). “Carry Trades and Currency Crashes” NBER Macroeconomics Annual 2008, 313–347. Burnside, C., M. Eichenbaum, I. Kleshchelski, and S. Rebelo (2011). “Do Peso Problems Explain the Returns to the Carry Trade?” Review of Financial Studies 24, 853–891. Campbell, J.Y., and R.J. Shiller (1987). “Cointegration and Tests of Present Value Models,”Journal of Political Economy 95, 1062–1088. Campbell, J.Y., and R.J. Shiller (1988). “Stock Prices, Earnings, and Expected Dividends.”Journal of Finance 43, 661–676. Campbell, J.Y., and S.B. Thompson (2008). “Predicting Excess Stock Returns Out of Sample: Can Anything Beat the Historical Average?” Review of Financial Studies 21, 1509–1531. Cheung, Y.-W., M.D. Chinn, and A.G. Pascual (2005). “Empirical Exchange Rate Models of the Nineties: are any Fit to Survive?” Journal of International Money and Finance 24, 1150–1175. Clarida, R., J. Gali, and M. Gertler (1998). “Monetary Policy Rules in Practice. Some International Evidence,” European Economic Review 42, 1033–1067. Clarida, R.H., L. Sarno, M.P. Taylor, and G. Valente (2003). “The Out-of-Sample Success of Term Structure Models as Exchange Rate Predictors: A Step Beyond,”Journal of International Economics 60, 61–83. Clarida, R.H., L. Sarno, M.P. Taylor, and G. Valente (2006). “The Role of Asymmetries and Regime Shifts in the Term Structure of Interest Rates,” Journal of Business 79, 1193–1225. Clark, T.E., and M.W. McCracken (2001). “Tests of Equal Forecast Accuracy and Encompassing for Nested Models,” Journal of Econometrics 105, 85–110. 37 Clark, T.E., and M.W. McCracken (2011). “Testing for Unconditional Predictive Ability,”in Hendry D., and M. Clements (eds.), Oxford Handbook on Economic Forecasting, Oxford University Press (forthcoming). Clark, T.E., and K.D. West (2006). “Using Out-of-Sample Mean Squared Prediction Errors to Test the Martingale Di¤erence Hypothesis,” Journal of Econometrics 135, 155–186. Clark, T.E., and K.D. West (2007). “Approximately Normal Tests for Equal Predictive Accuracy,” Journal of Econometrics 138, 291–311. Cremers, M. (2002). “Stock Return Predictability: A Bayesian Model Selection Perspective,”Review of Financial Studies 15, 1223–1249. Della Corte, P., L. Sarno, and G. Sestieri (2010). “The Predictive Information Content of External Imbalances for Exchange Rate Returns: How Much Is It Worth?,”Review of Economics and Statistics (forthcoming) doi:10.1162/REST_a_00157. Della Corte, P., L. Sarno, and D.L. Thornton (2008). “The Expectation Hypothesis of the Term Structure of very Short-Term Rates: Statistical Tests and Economic Value,” Journal of Financial Economics 89, 158–174. Della Corte, P., L. Sarno, and I. Tsiakas (2009). “An Economic Evaluation of Empirical Exchange Rate Models,” Review of Financial Studies 22, 3491–3530. Della Corte, P., L. Sarno, and I. Tsiakas (2011). “Spot and Forward Volatility in Foreign Exchange,” Journal of Financial Economics 100, 496–513. Della Corte, P., L. Sarno, and I. Tsiakas (2012). “Volatility and Correlation Timing in Active Currency Management,” in James, J., L. Sarno and I.W. Marsh (eds.) Handbook of Exchange Rates. London: Wiley (forthcoming). Diebold, F.X., and R.S. Mariano (1995). “Comparing Predictive Accuracy,”Journal of Business and Economic Statistics 13, 253–263. Engel, C. (1996). “The Forward Discount Anomaly and the Risk Premium: A Survey of Recent Evidence,” Journal of Empirical Finance 3, 123–192. Engel, C., N.C. Mark, and K.D. West (2007). “Exchange Rate Models are not as Bad as you Think,” NBER Macroeconomics Annual 2007, 381–441. Engel, C., and K.D. West (2005). “Exchange Rates and Fundamentals,”Journal of Political Economy 113, 485–517. Evans, M.D.D., and R.K. Lyons (2002). “Order Flow and Exchange Rate Dynamics,” Journal of Political Economy 110, 170–180. Fama, E.F. (1984). “Forward and Spot Exchange Rates,” Journal of Monetary Economics 14, 319– 338. Fleming, J., C. Kirby, and B. Ostdiek (2001). “The Economic Value of Volatility Timing,” Journal of Finance 56, 329–352. Goetzmann, W., J. Ingersoll, M. Spiegel, and I. Welch (2007). “Portfolio Performance Manipulation and Manipulation-Proof Performance Measures,” Review of Financial Studies 20, 1503–1546. 38 Gomez, V., and A. Maravall (2000). “Seasonal Adjustment and Signal Extraction in Economic Time Series,”in Pena D., G.C. Tiao, and T.R. Tsay (eds.), A Course in Time Series Analysis, New York: Wiley. Gourinchas, P.-O., and H. Rey (2007). “International Financial Adjustment,” Journal of Political Economy 115, 665–703. Goyal, A., and A. Saretto (2009). “Cross-Section of Option Returns and Volatility,” Journal of Financial Economics 94, 310–326. Groen, J.J.J. (2000). “The Monetary Exchange Rate Model as a Long-Run Phenomenon,” Journal of International Economics 52, 299–319. Hai, W., N.C. Mark, and Y. Wu (1997). “Understanding Spot and Forward Exchange Rate Regressions,” Journal of Applied Econometrics 12, 715–734. Han, Y. (2006). “Asset Allocation with a High Dimensional Latent Factor Stochastic Volatility Model,” Review of Financial Studies 19, 237–271. Hodrick, R.J. (1987). The Empirical Evidence on the E¢ ciency of Forward and Futures Foreign Exchange Markets, London: Harwood. Hodrick, R.J., and E.C. Prescott (1997). “Postwar U.S. Business Cycles: An Empirical Investigation,” Journal of Money, Credit and Banking 29, 1–16. Jondeau, E., and M. Rockinger (2008). “Time-Variability in Higher Moments Is Important for Asset Allocation,” Unpublished Working paper, University of Lausanne. Kilian, L. (1999). “Exchange Rates and Monetary Fundamentals: What do we Learn from LongHorizon Regressions?” Journal of Applied Econometrics 14, 491–510. Lothian, J.R., and M.P. Taylor (1996). “Real Exchange Rate Behavior: The Recent Float From the Perspective of the Past Two Centuries,” Journal of Political Economy 104, 488–509. Mark, N.C. (1995). “Exchange Rates and Fundamentals: Evidence on Long-Horizon Predictability,” American Economic Review 85, 201–218. Mark, N.C. (2009). “Changing Monetary Policy Rules, Learning, and Real Exchange Rate Dynamics,” Journal of Money, Credit and Banking 41, 1047–1070. Mark, N.C., and D. Sul. (2001). “Nominal Exchange Rates and Monetary Fundamentals: Evidence from a Small Post-Bretton Woods Panel,” Journal of International Economics 53, 29–52. Marquering, W., and M. Verbeek (2004). “The Economic Value of Predicting Stock Index Returns and Volatility,” Journal of Financial and Quantitative Analysis 39, 407–429. McCracken, M.W. (2007). “Asymptotics for Out of Sample Tests of Granger Causality,” Journal of Econometrics 140, 719–752. Meese, R.A., and K. Rogo¤ (1983). “Empirical Exchange Rate Models of the Seventies: Do They Fit Out of Sample?” Journal of International Economics 14, 3–24. Molodtsova, T., and D.H. Papell (2009). “Out-of-Sample Exchange Rate Predictability with Taylor Rule Fundamentals?” Journal of International Economics 77, 167–180. Neely, C.J., P.A. Weller, and J.M. Ulrich (2009). “The Adaptive Markets Hypothesis: Evidence from the Foreign Exchange Market,” Journal of Financial and Quantitative Analysis 44, 467–488. 39 Nelson, C.R., and M.J. Kim (1993). “Predictable Stock Returns: the Role of Small Sample Bias,” Journal of Finance 48, 641–661. Newey, W.K., and K.D. West (1987). “A Simple, Positive Semi-De…nite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix,” Econometrica 55, 703–708. Orphanides, A. (2001). “Monetary Policy Rules Based on Real-Time Data,” American Economic Review 91, 964–985. Orphanides, A., and S. van Norden (2002). “The Unreliability of Output Gap Estimates in Real Time,” Review of Economics and Statistics 84, 569–583. Rapach, D.E., J.K. Strauss, and G. Zhou (2010). “Out-of-Sample Equity Premium Prediction: Combination Forecasts and Links to the Real Economy.” Review of Financial Studies 23, 821–862. Rime, D., L. Sarno and E. Sojli (2010). “Exchange Rate Forecasting, Order Flow and Macroeconomic Information,” Journal of International Economics 80, 72–88. Rogo¤, K.S. (1996). “The Purchasing Power Parity Puzzle,” Journal of Economic Literature 34, 647–668. Rogo¤, K.S., and V. Stavrakeva (2008). “The Continuing Puzzle of Short Horizon Exchange Rate Forecasting,” NBER Working Paper No. 14071. Sarno, L. (2005). “Towards a Solution to the Puzzles in Exchange Rate Economics: Where Do We Stand?” Canadian Journal of Economics 38, 673–708. Sarno, L., and E. Sojli (2009). “The Feeble Link Between Exchange Rates and Fundamentals: Can We Blame the Discount Factor?,” Journal of Money, Credit and Banking 41, 437–442. Stambaugh, R.F. (1999). “Predictive Regressions,” Journal of Financial Economics 54, 375–421. Stock, J.H., and M.W. Watson (2003). “Forecasting Output and In‡ation: The Role of Asset Prices,” Journal of Economic Literature 41, 788–829. Stock, J.H., and M.W. Watson (2004). “Combination Forecasts of Output Growth in a Seven-Country Data Set,” Journal of Forecasting 23, 405–430. Taylor, A.M., and M.P. Taylor (2004). “The Purchasing Power Parity Debate,”Journal of Economic Perspectives 18, 135–158. Taylor, J.B. (1993). “Discretion versus Policy Rules in Practice,” Carnegie-Rochester Conference Series on Public Policy 39, 195–214. Timmermann, A. (2006). “Forecast Combinations,”in Elliott, G., C.W.J. Granger, and A. Timmermann (eds.), Handbook of Economic Forecasting, Amsterdam: Elsevier. Welch, I., and A. Goyal (2008). “A Comprehensive Look at The Empirical Performance of Equity Premium Prediction,” Review of Financial Studies 21, 1455–1508. West, K.D. (1988). “Dividend Innovations and Stock Price Volatility,” Econometrica 56, 37–61. West, K.D. (1996). “Asymptotic Inference about Predictive Ability,” Econometrica 64, 1067–1084. West, K.D. (2006). “Forecast Evaluation,” in: Elliott, G., C.W. Granger, and A. Timmermann, (eds.), Handbook of Economic Forecasting. Vol. 1, Chapter 3, 99–134. 40 West, K.D., H.J. Edison, and D. Cho (1993). “A Utility-Based Comparison of Some Models of Exchange Rate Volatility,” Journal of International Economics 35, 23–45. Wright, J.H. (2008). “Bayesian Model Averaging and Exchange Rate Forecasting,”Journal of Econometrics 146, 329–341. 41
© Copyright 2026 Paperzz