The floating exchange rate in the post-Bretton Woods era appears to be disconnected from its underlying macroeconomic determinants most of the time. Empirical work has often failed to present evidence of stable relationship between nominal exchange rate movements and fundamental variables suggested by the exchange rate determination models. "Indeed, the explanatory power of these models is essentially zero," as Evans and Lyons (2002, p.170) assert. In addition, a plethora of empirical studies find that the nominal exchange rate is excessively volatile relative to the underlying macroeconomic variables during the recent floating period. Flood and Rose (1999), for example, show that there are no macro-fundamentals capable of explaining the dramatic volatility of the exchange rate.
In this paper, I consider modeling the effects of the macroeconomic determinants on the nominal exchange rate to be channeled through the transition probabilities in a Markovian process. Many researchers have sought ways to model the possible nonlinearities in the relationship between the exchange rate and macro-fundamentals. Little work, nevertheless, has ever studied the transitional effects of the macroeconomic determinants on the exchange rate. In effect, allowing fundamentals to affect the transition probabilities in the Markovian process is intuitively attractive: the market responds to the updated news in the macro variables deviation of the exchange rate from its fundamental value and in turn alters the belief in the chance of the process staying in certain regime next period.
In this paper, I consider modeling the effects of the macroeconomic determinants on the nominal exchange rate to be channeled through the transition probabilities in a Markovian process. Many researchers have sought ways to model the possible nonlinearities in the relationship between the exchange rate and macro-fundamentals. Little work, nevertheless, has ever studied the transitional effects of the macroeconomic determinants on the exchange rate. In effect, allowing fundamentals to affect the transition probabilities in the Markovian process is intuitively attractive: the market responds to the updated news in the macro variables deviation of the exchange rate from its fundamental value and in turn alters the belief in the chance of the process staying in certain regime next period.
On the other hand, the Markov-switching model popularized by Hamilton (1988, 1989) has proved especially successful in modeling time series with changes of regime. Nevertheless, the Hamilton's Markov-switching model takes little consideration of the movements in the variance. For example, Pagan and Schwert (1990) show that Markov switching specification is not satisfactory in explaining the monthly U.S. stock-return volatility from 1834 to 1925. In this regard, an extension combining the traditional Markov-switching model with ARCH specification turns out to be a natural motivation.
Download
The Exchange Rate and Macroeconomic Determinants: Time-Varying Transitional Dynamics
