A Note on the Foreign Exchange Market Efficiency Hypothesis: Does Small Sample Bias affect Inference?

This study examines whether small sample bias affects the standard inference about the foreign exchange market efficiency hypothesis. Our findings indicate that the bias is large enough to result in rejection of the efficient market hypothesis even when it is true. We use bootstrapping to adjust for the bias and find that the hypothesis cannot be rejected for the Swiss franc and French franc. We also find that the bias plays a significant role in the inference that expectation error causes inefficiency in the foreign exchange markets. After bias adjustment, the rational expectation hypothesis holds even at one month-horizon.