Forward-Rate Bias, Imperfect Knowledge, and Risk: Evidence from Developed and Developing Countries

This paper examines the stability of the basic regression underpinning the forward-discount anomaly for 20 developed and 32 developing countries. It finds that the correlation between the change of the spot exchange rate and forward discount is piece-wise linear in every country, involving stretches of time in which forward-rate bias is negative and other stretches of time in which it is positive. The results also point out to a new empirical finding that the average magnitude of the positive and negative biases that are found in each developed country tends to be larger than those for developing countries. The paper shows that a new risk premium model based on Imperfect Knowledge Economics can account for this pattern across developed and developing countries, thereby undercutting the widespread view in the literature that developed-country currency markets are characterized by a greater degree of irrationality than those for developing countries.