Exchange rates and monetary fundamentals: what do we learn from long-horizon regressions?

Abstract

The use of a new bootstrap method for small‐sample inference in long‐horizon regressions is illustrated by analysing the long‐horizon predictability of four major exchange rates, and the findings are reconciled with those of an earlier study by Mark (1995). While there is some evidence of exchange rate predictability, contrary to earlier studies, no evidence is found of higher predictability at longer horizons. Additional evidence is presented that the linear VEC model framework underlying the empirical study is likely to be misspecified, and that the methodology for constructing bootstrap p‐values for long‐horizon regression tests may be fundamentally flawed. Copyright © 1999 John Wiley & Sons, Ltd.

First‐ and Second‐Moment Exchange Rate Exposure: Evidence from U.S. Stock Returns

“…As mentioned earlier, Chow, Lee, and Solt (1997a) claim that using longer horizon data increases the chance of detecting significant exposure. To explore this possibility, we estimate the models using monthly nonoverlapping data because of the potential biases resulting from the use of overlapping data (see, for example, Andrews and Monahan, 1992;Boudoukh and Richardson, 1994;Kilian, 1999;Campbell, 2001). The results (not reported) show that at the monthly frequency there are fewer instances of time-varying volatility in exchange rates and stock returns.…”

Exchange Rates and Fundamentals: Footloose or Evolving Relationship?

“…Sweeney (2006) also provides evidence that challenges the conventional wisdom that industrial-country floating exchange rates contain unit roots and that, in out-of-sample forecasts, mean-reversion models beat random walks on average, in some forecast periods significantly. However, others remain skeptical (e.g., Kilian 1999;Berkowitz and Giorgianni 2001;Faust, Rogers, and Wright 2003;Engel and West 2005), so that evidence that exchange rate forecasts obtained using fundamentals models are better than forecasts from a random walk remains elusive (e.g., Cheung, Chinn, and Garcia Pascual 2003;Sarno 2005).…”

Testing Long‐horizon Predictive Ability With High Persistence, and the Meese–rogoff Puzzle*

“…The result of this article apply to situations in which the researcher is interested in predictions at long horizons in the presence of small sample sizes. This situation is quite common in finance and international macroeconomics (e.g., Mark, 1995;Kilian, 1999;Berkowitz and Giorgianni, 2001). Diebold and Kilian (2000) found evidence that unit root pretests are likely to improve the forecasting accuracy relative to forecasts from models in levels.…”