We analyze the empirical power and specification of test statistics designed to detect abnormal bond returns in corporate event studies, using monthly and daily data. We find that test statistics based on frequently used methods of calculating abnormal monthly bond returns are biased. Most...