Recent empirical studies have found that small listed firms yield higher average returns than large firms even when their riskiness is equal. The riskiness of small firms, however, has been improperly measured. Apparently, the error is due to auto‐correlation in portfolio returns caused by infrequent trading. Other anomalous predictors of riskadjusted returns, such as price/earnings ratios and dividend yields, may also derive some of their apparent power from this spurious source.