- مبلغ: ۸۶,۰۰۰ تومان
- مبلغ: ۹۱,۰۰۰ تومان
In this paper, I focus on the time-varying and persistent exchange rate risk premiums in uncovered interest rate parity associated with changes in net foreign assets. The results of my analyses of the Dollar-Yen exchange rate provide evidence consistent with my risk premium formulation and the predictability of current account balances. I contend that the strong persistent effect causes nominal exchange rates to appear non-stationary in level. I also argue that the present value model of the level of exchange rates combined with the AR(1) approximation for interest rate differentials can reconcile a failure of uncovered interest rate parity.
4. Summary and concluding remarks
In this paper, I discussed the mechanism through which net foreign asset holdings affect exchange rates from the standpoint of asset pricing without explicitly using intertemporal budget constraints. Because risk premiums are the central issue in the current research on asset pricing, I have focused on the time-varying and persistent exchange rate risk premium related to uncovered interest rate parity. To do so, I have argued that spot exchange rate risk premiums vary through changes in net foreign asset holdings and, especially, increase with the accumulation of net foreign asset holdings. In the case of an infinitely long horizon, especially under the stationarity assumption, the uncovered interest rate parity equation is equivalent to a present value model of the level of exchange rates. My empirical results provide evidence consistent with my formulation of the time-varying and persistent risk premiums associated with changes in net foreign assets. The results also suggest that, in the longer-horizon case, the Japanese current account balance predicts the Dollar-Yen exchange rate level. While academics agree that nominal exchange rates are non-stationary in level generally, I argue that the strong persistent effect associated with changes in net foreign asset holdings causes exchange rates to appear non-stationary in level. The results of the regression replications with annual data show that the persistent effect shrinks in magnitude as the time interval becomes longer. These results indicate the stationarity in the level of nominal exchange rates.