Consistent Estimation of Conditional Conservatism

In this paper, we demonstrate analytically that Basu model estimates are affected by two biases, except if very restrictive conditions are met: an aggregation bias produced by the absence of the book-to-market ratio in Basu model, and (2) an aggregation problem produced by the use of total market returns instead of separating the good news and bad news of the period. To solve this problem, we propose an alternative econometric model that is robust to both biases. The empirical results obtained using archival data demonstrate the advantages of robustness of our alternative econometric model compared to Basu model: it controls the omitted variable bias, is robust to the aggregation effect, its estimates are less influenced by the extreme values of market returns, and it presents a higher explanatory power