User:Bkwillwm/Forward premium puzzle

The forward premium puzzle concerns the apparent inability of the forward exchange rate to predict actual, future spot rates. Fama (1984) found that current spot rates are a better predictor of future spot rates that forward rates. The apparent "excess returns" possible from can be seen as a violation of uncovered interest rate parity or rational expectations, since exchange traders may appear systematically irrational. Fama's results also showed that the variance of the risk premium was higher than those of expected spot rates.

Econometrically, the puzzle has been documented in regressions of the forward rate on the corresponding spot rate. Estimates on the Beta (coefficient on the forward rate) would be expected to be near 1, given a constant risk premium, or near zero if the spot rate followed a random walk. Evidence suggests that the estimates of beta are negative, contradicting both these possibilities.

One explanation for this phenomena has been the "peso problem" discussed in Krasker (1980). Krasker notes that that was a consistent excess return to holding the peso during much of the 20th century. This was explained by the fact that central bank pegged the peso, but there was always a chance that a devaluation could occur, so there was a consistent risk premium that took this into account in the short run. The futures rate consistently showed a chance of a devaluation when none occurred, so there appeared to be an excess return. However, the chance of a devaluation was real and rational investors would be expected to take this in to account. Thus the excess returns are consistent with rational expectations. See Taleb distribution.

Survey data by Froot and Frankel (1989) suggest that exchange traders make systematic prediction errors.