Details

Autor(en) / Beteiligte

• Choi, Heenam

Titel

Risk premia or forecast errors: Explaining excess returns in foreign exchange markets

Ort / Verlag

ProQuest Dissertations & Theses

Erscheinungsjahr

1997

Quelle

ProQuest Dissertations & Theses A&I

Beschreibungen/Notizen

• There is a systematic relationship between the rate of appreciation or depreciation of the US dollar and the direction of bias in the forward rates. During periods in which the dollar appreciated, forward rates significantly underestimated the future value of the dollar $\rm(f\sb{t}>s\sb{t+1})$ and in periods in which the dollar depreciated, the forward rates significantly overestimated the future value of the dollar $\rm(f\sb{t}<s\sb{t+1}).$ Therefore, the forward foreign exchange rate is a biased and inefficient predictor of the future spot rate; returns from speculation in spot and forward currency markets exist and are volatile. The purpose of this dissertation is to tackle this excess return puzzle. The first chapter investigates the variances of excess returns using a vector of excess returns to uncover unobservable factors which make the agents expectations differ from the actual rate. Comovements among excess returns show that there exist common elements that can be captured by underlying, unobservable variables. Using a factor model, this chapter decomposes the volatility of excess returns into unobservable components associated with world, domestic, and foreign factors which are closely related with news regarding the fundamentals of the exchange rates of all countries. The results obtained give some implication about the source of variation in excess returns. In particular, they suggest the strong role of US economic policy in affecting the volatility of excess returns even though the role becomes less important than before. These results suggest that if we are going to analyze the volatility of excess returns or make a forecast on future spot rates of the major currencies, we should first consider domestic factors related to US economic policies such as monetary and fiscal policy. The second chapter examines the failure of efficient market hypothesis due to expectational errors. By including both monetary and fiscal policy variables as observed market fundamentals to decide the current exchange rate regime in market's information set in addition to exchange rate changes, I find that this model explains well the learning behavior of economic agents when a regime change takes place and explains up to 97 percent of the forecast errors in forward exchange rate markets. Although the empirical results presented in this chapter find the role of market fundamentals in the formation of agents' expectations to be important, other factors such as the trade deficit or the monetary and fiscal policy of foreign countries may also help to explain forecast errors. Moreover, to completely understand the existence of forecast errors, it would be desirable to integrate the explanations provided by the learning effect and the peso problem. Finally, the purpose of the third chapter is to test whether risk premium expressed as time variation in conditional variances of the exogenous processes helps in reproducing the time series properties of actual excess returns. Since the impact of changes in uncertainty in exogenous variables may undergo regime changes, this research combines the Markov regime switching model and the ARCH model to describe time series data. Estimation results demonstrate that the conditional variances fail to explain the movements of excess returns. It suggests that the existence of risk premium is open to question. Given these results, answers to the excess returns puzzle appear more likely to arise from on the role of the forecast errors. (Abstract shortened by UMI.)