The joint-effects of conservatism and market inefficiency on the returns-earnings relation

I investigate the implication of the joint effects of accounting conservatism, growth, and market inefficiency for earnings response coefficients (ERCs). Penman and Zhang (2002) earn positive future returns using a strategy that goes long (short) in companies with conservative accounting and a recent increase (decrease) in operating assets. Given that the market appears to misinterpret conservatism, I expect biased coefficients in the returns-earnings relation due to the dependent variable reacting to accounting information with a delay. To detect potential bias in coefficients, I apply a procedure from Aboody, Hughes and Liu (2002) that adjusts current returns (or stock price) for future risk-adjusted returns so that coefficients reflect contemporaneous and delayed market reactions to current accounting variables. Significant differences between coefficients from the adjusted and unadjusted regressions imply bias from market inefficiency. For my analysis, I expect larger differences between coefficients for samples exhibiting high conservatism and high growth compared to samples exhibiting high conservatism and low growth. While I find some evidence of coefficient bias due to conservatism and growth significant enough to affect inferences, I do not find significant differences in ERCs across conservatism samples that I expect based on the evidence in Penman and Zhang (2002). Thus, the results in this study do not support the notion that the relation between conservatism and future returns is due to the market's over- or under-reaction to earnings. My results are robust to sensitivity analyses using alternate measures of conservatism combined with growth and sample selection.