nach oben

Erschienen in:

2015 | OriginalPaper | Buchkapitel

43. Rationality and Heterogeneity of Survey Forecasts of the Yen-Dollar Exchange Rate: A Reexamination

verfasst von : Richard Cohen, Carl S. Bonham, Shigeyuki Abe

Erschienen in: Handbook of Financial Econometrics and Statistics

Verlag: Springer New York

Einloggen

Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.

search-config

KI-gestützte Suche

Aus

Abstract

This chapter examines the rationality and diversity of industry-level forecasts of the yen-dollar exchange rate collected by the Japan Center for International Finance. In several ways we update and extend the seminal work by Ito (1990, American Economic Review 80, 434–449). We compare three specifications for testing rationality: the “conventional” bivariate regression, the univariate regression of a forecast error on a constant and other information set variables, and an error correction model (ECM). We find that the bivariate specification, while producing consistent estimates, suffers from two defects: first, the conventional restrictions are sufficient but not necessary for unbiasedness; second, the test has low power. However, before we can apply the univariate specification, we must conduct pretests for the stationarity of the forecast error. We find a unit root in the 6-month horizon forecast error for all groups, thereby rejecting unbiasedness and weak efficiency at the pretest stage. For the other two horizons, we find much evidence in favor of unbiasedness but not weak efficiency. Our ECM rejects unbiasedness for all forecasters at all horizons. We conjecture that these results, too, occur because the restrictions test sufficiency, not necessity.

We extend the analysis of industry-level forecasts to a SUR-type structure using an innovative GMM technique (Bonham and Cohen 2001, Journal of Business & Economic Statistics, 19, 278–291) that allows for forecaster cross-correlation due to the existence of common shocks and/or herd effects. Our GMM tests of micro-homogeneity uniformly reject the hypothesis that forecasters exhibit similar rationality characteristics.

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

über 102.000 Bücher
über 537 Zeitschriften

aus folgenden Fachgebieten:

Automobil + Motoren
Bauwesen + Immobilien
Business IT + Informatik
Elektrotechnik + Elektronik
Energie + Nachhaltigkeit
Finance + Banking
Management + Führung
Marketing + Vertrieb
Maschinenbau + Werkstoffe
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Springer Professional "Wirtschaft"

Online-Abonnement

Mit Springer Professional "Wirtschaft" erhalten Sie Zugriff auf:

über 67.000 Bücher
über 340 Zeitschriften

aus folgenden Fachgebieten:

Bauwesen + Immobilien
Business IT + Informatik
Finance + Banking
Management + Führung
Marketing + Vertrieb
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Vorheriges Kapitel Earnings Quality and Board Structure: Evidence from South East Asia

Nächstes Kapitel Stochastic Volatility Structures and Intraday Asset Price Dynamics

Nur mit Berechtigung zugänglich

If in addition the residuals from the cointegrating regression are white noise, this supports a type of weak efficiency.

Pretesting the forecast error for stationarity is a common practice in testing the RNMEH, but the only study we know of that applies this practice to survey forecasts of exchange rates is Osterberg (2000), and he does not test for a zero intercept in the cointegrating regression.

Market microstructure theories assume that there is a minimum amount of forecaster (as well as cross-sectional forecast) diversity. Also, theories of exchange rate determination that depend upon the interaction between chartists (or noise traders) and fundamentalists by definition require a certain structure of forecaster heterogeneity.

It is important to note that the result from one type of rationality test does not have implications for the results from any other types of rationality tests. In this chapter we test for unbiasedness and weak efficiency, leaving the more stringent tests of efficiency with respect to publicly available information for future analysis.

A large theoretical literature relaxes Muth’s assumption that all information relevant for forming a rational forecast is publicly available. Instead, this literature examines how heterogeneous individual expectations are mapped into an aggregate market expectation, and whether the latter leads to market efficiency. (See, e.g., Figlewski 1978, 1982, 1984; Kirman 1992; Haltiwanger and Waldman 1989.) Our paper focuses on individual rationality but allows for the possibility of synergism by incorporating not only heteroscedasticity and autocorrelation consistent standard errors in individual rationality tests but also cross-forecaster correlation in tests of micro-homogeneity. The extreme informational requirement of the REH led Pesaran and Weale (2006) to propose a weaker form of the REH that is based on the (weighted) average expectation using only publicly available (i.e., common) information.

The extent to which private information influences forecasts is more controversial in the foreign exchange market than in the equity or bond markets. While Chionis and MacDonald (1997) maintain that there is little or no private information in the foreign exchange market, Lyons (2002) argues that order flow explains much of the variation in prices. To the extent that one agrees with the market microstructure emphasis on the importance of the private information embodied in dealer order flow, the Figlewski-Wachtel critique remains valid in the returns regression.

Elliott and Ito (1999) show that, although a random walk forecast frequently outperforms the JCIF survey forecasts using an MSE criterion, survey forecasts generally outperform the random walk, based on an excess profits criterion. This supports the contention that JCIF forecasters are properly motivated to produce their best forecasts.

To mitigate the confidentiality problem in this case, the survey typically withholds individual forecasts until the realization is known or (as with the JCIF) masks the individual forecast by only reporting some aggregate forecast (at the industry and total level) to the public.

Laster et al. (1999) called this practice “rational bias.” Prominent references in this growing literature include Lamont (2002), Ehrbeck and Waldmann (1996), and Batchelor and Dua (1990a, b, 1992). Because we have access only to forecasts at the industry average level, we cannot test the strategic incentive hypotheses.

See Elliott and Ito (1999), Boothe and Glassman (1987), LeBaron (2000), Leitch and Tanner (1991), Lai (1990), Goldberg and Frydman (1996), and Pilbeam (1995). This type of loss function may appear to be relevant only for relatively liquid assets such as foreign exchange, but not for macroeconomic flows. However, the directional goal is also used in models to predict business cycle turning points. Also, trends in financial engineering may lead to the creation of derivative contracts in macroeconomic variables, e.g., CPI futures.

The efficiency aspect of rationality is sometimes tested by including additional variables in the forecaster’s information set, with corresponding hypotheses of zero coefficients on these variables. See, e.g., Keane and Runkle (1990) for a more recent study using the level specification and Bonham and Cohen (1995) for a critique of Keane and Runkle’s integration accounting.

As we report in Sect. 43.5, this lack of power is at least consistent with the failure to reject micro-homogeneity at all three horizons.

Note that, for illustrative purposes only, we compute the expectational variable as the four-group average percentage change in the forecast. However, recall that, despite the failure to reject micro-homogeneity at any horizon, the Figlewski-Wachtel critique implies that these parameter estimates are inconsistent in the presence of private information. (See the last paragraph in this subsection.)

However, in the general case of biased and/or inefficient forecasts, Mincer and Zarnowitz (1969, p. 11) also viewed the bivariate regression ‘as a method of correcting the forecasts … to improve [their] accuracy … Theil (1966, p. 33) called it the “optimal linear correction.”’ That is, the correction would involve (1) subtracting α _i,h and then (2) multiplying by 1/β _i,h. Graphically, this is a translation of the regression line followed by a rotation, until the regression line coincides with the 45° line.

Other researchers (e.g., Bryant 1995) have found similar vertical scatters for regressions where the independent variable, e.g., the forward premium/discount f _t,h − s _t, the “exchange risk premium” f _t,h − s _t+h, or the difference between domestic and foreign interest rates (i − i ^*), exhibits little variation.

It is also possible to estimate the cointegrating parameters and jointly test whether they are zero and one. A variety of methods, such as those due to Saikkonen (1991) or Phillips and Hansen (1990), exist that allow for inference in cointegrated bivariate regressions.

As expected, exporters failed to reject at the 10 % level in all three tests.

The direction of the bias for exporters is negative; that is, they systematically underestimate the value of the yen, relative to the dollar. Ito (1990) found the same tendency using only the first two years of survey data (1985–1987). He characterized this depreciation bias as a type of “wishful thinking” on the part of exporters.

Ito (1994) conducted a similar analysis for the aggregate of all forecasters, but without an explicit test for structural breaks.

This is consistent with the finding of nonstationary forecast errors for all groups at the 6-month horizon.

Zacharatos and Sutcliffe (2002) note that the inclusion of the contemporaneous spot forecast (in their paper, the forward rate) as a regressor assumes that the latter is weakly exogenous; that is, deviations from unbiasedness are corrected only by movements in the realized spot rate. These authors prefer a bivariate ECM specification, in which the change in the future spot rate and the change in the contemporaneous forecast are functions of an error correction term and lags of the dependent variables. However, Zivot (2000) points out that if the spot rate and forecast are contemporaneously correlated, then our single-equation specification does not make any assumptions about the weak exogeneity of the forecast.

Our empirical specification of the ECM also includes an intercept. This will help us to determine whether there are structural breaks in the ECM.

Since we include an intercept, we also test the restriction that the intercept equals zero – both individually and as part of the joint unbiasedness hypothesis.

The only exception is for exporters at the 1-month horizon.

The standard errors in the univariate regression are about the same as those for the ECM. (By definition, of course, the R ² s for the univariate regression equal zero.)

Since we estimate the restricted ECM with an intercept, unbiasedness also requires the intercept to be equal to zero.

Since the intercept in Eq. 43.10 is not significant in any regression, the simple hypothesis that α _i,h equals one also fares the same as the simple unbiasedness tests.

For purposes of comparison with both the bivariate joint and simple unbiasedness restrictions, we have used the ECM results using the robust standard errors. In all cases testing the ECM restrictions using F-statistics based on whitened residuals produces rejections of all restrictions, simple and joint, except a zero intercept. Hakkio and Rush (1989) found similarly strong rejections of Eq. 43.9, where the forecast was the forward rate.

Notice that the first two sets of weak efficiency variables include the mean forecast, rather than the individual group forecast. Our intention is to allow a given group to incorporate information from other groups’ forecasts via the prior mean forecast. This requires an extra lag in the information set variables, relative to a contemporaneously available variable such as the realized exchange rate depreciation.

This is a general test, not only because it allows for an alternative hypothesis of higher-order serial correlation of specified order but also because it allows for serial correlation to be generated by AR, MA, or ARMA processes.

We use the F-statistic because the χ ² test statistics tend to over-reject, while the F-tests have more appropriate significance levels (see Kiviet 1987).

Elliott and Ito (1999) used single-equation estimation that incorporated a White correction for heteroscedasticity and a Newey-West correction for serial correlation. (See the discussion below of Ito’s tests of forecaster heterogeneity.)

Unlike Breusch and Pagan’s (1980) LM test for cross-sectional dependence, Pesaran’s (2004) CD test is robust to multiple breaks in slope coefficients and error variances, as long as the unconditional means of the variables are stationary and the residuals are symmetrically distributed.

There are three instances of statistically significant negative test statistics for lags greater than h-1, none for lags less than or equal to h-1. Thus, some industries produce relatively high forecast errors several periods after others produce relative low forecast errors, and this information is not fully incorporated in some current forecasts.

The nonrejection of micro-homogeneity in bivariate regressions does not, however, mean that one can avoid aggregation bias by using the mean forecast. Even if the bivariate regressions were correctly interpreted as joint tests of unbiasedness and weak efficiency with respect to the current forecast, and even if the regressions had sufficient power to reject a false null, the micro-homogeneity tests would be subject to additional econometric problems. According to the Figlewski-Wachtel (1983) critique, successfully passing a pretest for micro-homogeneity does not ensure that estimated coefficients from such consensus regressions will be consistent. See Sect. 43.2.1.

Recall that our group results are not entirely comparable to Ito’s (1990), since our data set, unlike his, combines insurance companies and trading companies into one group and life insurance companies and import-oriented companies into another group.

Chionis and MacDonald (1997) performed an Ito-type test on individual expectations data from Consensus Forecasts of London.

Elliott and Ito (1999), who have access to forecasts for the 42 individual firms in the survey, find that, for virtually the same sample period as ours, the null hypothesis of a zero deviation from the mean forecast is rejected at the 5 % level by 17 firms at the 1-month horizon, 13 firms for the 3-month horizon, and 12 firms for the 6-month horizon. These authors do not report results by industry group.

We put less weight on the results of the weaker tests for micro-homogeneity in the bivariate regressions.

He also included regressors for adaptive expectations and the forward premium.

Banerjee, A., Dolado, J. J., Galbraith, J. W., & Hendry, D. F. (1993). Co-integration, error correction and the econometric analysis of non-stationary data. Oxford: Oxford University Press.CrossRef

Batchelor, R., & Dua, P. (1990a). Forecaster ideology, forecasting technique, and the accuracy of economic forecasts. International Journal of Forecasting, 6, 3–10.CrossRef

Batchelor, R., & Dua, P. (1990b). Product differentiation in the economic forecasting industry. International Journal of Forecasting, 6, 311–316.CrossRef

Batchelor, R., & Dua, P. (1992). Conservatism and consensus-seeking among economic forecasters. Journal of Forecasting, 11, 169–181.CrossRef

Batchelor, R., & Peel, D. A. (1998). Rationality testing under asymmetric loss. Economics Letters, 61, 49–54.CrossRef

Bonham, C. S., & Cohen, R. H. (1995). Testing the rationality of price forecasts: Comment. American Economic Review, 85(1), 284–289.

Bonham, C. S., & Cohen, R. H. (2001). To aggregate, pool, or neither: Testing the rational-expectations hypothesis using survey data. Journal of Business & Economic Statistics, 19, 278–291.CrossRef

Boothe, P., & Glassman, D. (1987). Comparing exchange rate forecasting models. International Journal of Forecasting, 3, 65–79.CrossRef

Breusch, T. S. (1978). Testing for autocorrelation in dynamic linear models. Australian Economic Papers, 17, 334–355.CrossRef

Breusch, T. S., & Pagan, A. R. (1980). The Lagrange multiplier test and its application to model specifications in econometrics. Review of Economic Studies, 47, 239–253.CrossRef

Bryant, R. C. (1995). The “exchange risk premium,” uncovered interest parity, and the treatment of exchange rates in multicountry macroeconomic models. Technical report, The Brookings Institution.

Cavaglia, S. M. F. G., Verschoor, W. F. C., & Wolff, C. C. P. (1994). On the biasedness of forward foreign exchange rates: Irrationality or risk premia? Journal of Business, 67, 321–343.CrossRef

Chionis, D., & MacDonald, R. (1997). Some tests of market microstructure hypotheses in the foreign exchange market. Journal of Multinational Financial Management, 7, 203–229.CrossRef

Clements, M. P., & Hendry, D. F. (1998). Forecasting economic time series. Cambridge: Cambridge University Press.CrossRef

Cohen, R. H., & Bonham, C. S. (2006). Specifying the forecast generating process for exchange rate survey forecasts. Working paper, Department of Economics, University of Hawaii Manoa.

Davies, A., & Lahiri, K. (1995). A new framework for testing rationality and measuring aggregate shocks using panel data. Journal of Econometrics, 68, 205–227.CrossRef

Dickey, D. A., & Fuller, W. A. (1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74, 427–431.

Ehrbeck, T., & Waldmann, R. (1996). Why are professional forecasters biased? Agency versus behavioral explanations. The Quarterly Journal of Economics, 111, 21–40.CrossRef

Elliott, G., & Ito, T. (1999). Heterogeneous expectations and tests of efficiency in the yen/dollar forward exchange rate market. Journal of Monetary Economics, 43, 435–456.CrossRef

Engle, R., & Granger, C. W. S. (1987). Cointegration and error correction representation, estimation and testing. Econometrica, 55, 251–276.CrossRef

Figlewski, S. (1978). Market “efficiency” in a market with heterogeneous information. Journal of Political Economy, 86, 581–597.CrossRef

Figlewski, S. (1982). Information diversity and market behavior. Journal of Finance, 37, 87–102.CrossRef

Figlewski, S. (1984). Information diversity and market behavior: A reply. Journal of Finance, 39, 299–302.CrossRef

Figlewski, S., & Wachtel, P. (1983). Rational expectations, informational efficiency, and tests using survey data: A reply. Review of Economics and Statistics, 65, 529–531.CrossRef

Frankel, J. A., & Froot, K. A. (1987). Using survey data to test standard propositions regarding exchange rate expectations. American Economic Review, 77, 133–153.

Frankel, J. A., & Froot, K. A. (1990). Chartists, fundamentalists, and trading in the foreign exchange market. American Economic Review, 80, 181–185.

Froot, K. A., & Frankel, J. A. (1989). Forward discount bias: Is it an exchange risk premium? Quarterly Journal of Economics, 104, 139–161.CrossRef

Gavin, W. T. (2003) FOMC forecasts: Is all the information in the central tendency? Economic Review, Federal Reserve Bank of St. Louis, May/June, 27–46.

Godfrey, L. G. (1978). Testing for higher order serial correlation in regression equations when the regressors include lagged dependent variables. Econometrica, 46, 1303–1310.CrossRef

Goldberg, M. D., & Frydman, R. (1996). Imperfect knowledge and behaviour in the foreign exchange market. Economic Journal, 106, 869–893.CrossRef

Granger, C. W. J. (1991). Developments in the study of cointegrated economic variables. In R. F. Engle & C. W. J. Granger (Eds.), Long run economic relationships: Readings in Cointegration (pp. 65–80). Oxford: Oxford University Press.

Hakkio, C. S., & Rush, M. (1989). Market efficiency and cointegration: An application to the sterling deutschemark exchange markets. Journal of International Money and Finance, 8, 75–88.CrossRef

Haltiwanger, J. C., & Waldman, M. (1989). Rational expectations in the aggregate. Economic Inquiry, 27, 619–636.CrossRef

Hamilton, J. (1994). Time series analysis. Princeton: Princeton University Press.

Hansen, L. P., & Hodrick, R. J. (1980). Forward exchange rates as optimal predictors of future spot rates: An econometric analysis. Journal of Political Economy, 88, 829–853.CrossRef

Holden, K., & Peel, D. A. (1990). On testing for unbiasedness and efficiency of forecasts. Manchester School, 58, 120–127.CrossRef

Ito, T. (1990). Foreign exchange rate expectations: Micro survey data. American Economic Review, 80, 434–449.

Ito, T. (1994). Short-run and long-run expectations of the yen/dollar exchange rate. Journal of the Japanese and International Economies, 8, 119–143.CrossRef

Keane, M. P., & Runkle, D. E. (1990). Testing the rationality of price forecasts: New evidence from panel data. American Economic Review, 80, 714–735.

Kirman, A. P. (1992). Whom or what does the representative individual represent? Journal of Economic Perspectives, 6, 117–136.CrossRef

Kiviet, J. F. (1987). Testing linear econometric models. Ph.D. Thesis, University of Amsterdam.

Lai, K. (1990). An evaluation of survey exchange rate forecasts. Economics Letters, 32, 61–65.CrossRef

Lamont, O. (2002). Macroeconomic forecast and microeconomic forecasters. Journal of Economic Behavior and Organization, 48, 265–280.CrossRef

Laster, D., Bennett, P., & Geoum, I. S. (1999). Rational bias in macroeconomic forecasts. Quarterly Journal of Economics, 293–318.

LeBaron, B. (2000) Technical trading profitability in foreign exchange markets in the 1990’s. Technical report, Brandeis U.

Leitch, G., & Tanner, J. E. (1991). Economic forecast evaluation: Profits versus the conventional error measures. The American Economic Review, 81, 580–590.

Liu, P., & Maddala, G. S. (1992). Rationality of survey data and tests for market efficiency in the foreign exchange markets. Journal of International Money and Finance, 11, 366–381.CrossRef

Lyons, R. K. (2002). Foreign exchange: Macro puzzles, micro tools. Economic Review, Federal Reserve Bank of San Francisco, 51–69.

MacDonald, R. (1992). Exchange rate survey data: A disaggregated G-7 perspective. The Manchester School, LX(Supplement), 47–62.

Maynard, A., & Phillips, P. C. B. (2001). Rethinking and old empirical puzzle: Econometric evidence on the forward discount anomaly. Journal of Applied Econometrics, 16, 671–708.CrossRef

Mincer, J., & Zarnowitz, V. (1969). The evaluation of economic forecasts. In Economic forecasts and expectations. New York: National Bureau of Economic Research.

Muth, J. (1961). Rational expectations and the theory of price movements. Econometrica, 29, 315–335.CrossRef

Newey, W., & West, K. (1987). A simple positive semi-definite heteroscedasticity and autocorrelation consistent covariance matrix. Econometrica, 55, 703–708.CrossRef

Osterberg, W. P. (2000). New results on the rationality of survey measures of exchange-rate expectations. Federal Reserve Bank of Cleveland Economic Review, 36, 14–21.

Pesaran, H. M. (2004). General diagnostic tests for cross section dependence in panels. CESifo working paper series 1229, CE-Sifo Group Munich.

Pesaran, H. M., & Weale, M. (2006). Survey expectations. In G. E. C. W. J. Granger & A. Timmermann (Eds.), Handbook of economic forecasting. Amsterdam: North-Holland.

Phillips, P., & Hansen, B. E. (1990). Statistical inference in instrumental variables regression with I(1) processes. Review of Economic Studies, 57, 99–125.CrossRef

Phillips, P. C. B., & Maynard, A. (2001). Rethinking an old empirical puzzle: Econometric evidence on the forward discount anomaly. Journal of Applied Econometrics, 16, 671–708.CrossRef

Pilbeam, K. (1995). Exchange rate models and exchange rate expectations: an empirical investigation. Applied Economics, 27, 1009–1015.CrossRef

Saikkonen, P. (1991). Asymptotically efficient estimation of cointegration regressions. Econometric Theory, 7, 1–21.CrossRef

Stockman, A. C. (1987). Economic theory and exchange rate forecasts. International Journal of Forecasting, 3, 3–15.CrossRef

Theil, H. (1966). Applied economic forecasting. Amsterdam: North-Holland.

Zacharatos, N., & Sutcliffe, C. (2002). Is the forward rate for the Greek Drachma unbiased? A VECM analysis with both overlapping and non-overlapping data. Journal of Financial Management and Analysis, 15, 27–37.

Zarnowitz, V. (1985). Rational expectations and macroeconomic forecasts. Journal of Business & Economic Statistics, 3, 293–311.

Zellner, A (1962). An efficient method of estimating seemingly unrelated regressions and tests for aggregation bias. Journal of the American Statistical Association, June, 348–368.

Zellner, A. (1986). Biased predictors, rationality and the evaluation of forecasts. Economics Letters, 21, 45–48.CrossRef

Zivot, E. (2000). Cointegration and forward and spot exchange rate regressions. Journal of International Money and Finance, 19, 785–812.CrossRef

Titel: Rationality and Heterogeneity of Survey Forecasts of the Yen-Dollar Exchange Rate: A Reexamination
verfasst von: Richard Cohen
Carl S. Bonham
Shigeyuki Abe
Verlag: Springer New York
Buch: Handbook of Financial Econometrics and Statistics
Print ISBN: 978-1-4614-7749-5

Electronic ISBN: 978-1-4614-7750-1

Copyright-Jahr: 2015
DOI: https://doi.org/10.1007/978-1-4614-7750-1_43