Purpose – This study investigates whether the future earnings predictability of the fundamental variables (AFS) is different from that of discretionary accruals. Given that AFS is the aggregate score index based on financially fundamental variables which predict the change in future earnings, AFS information is expected to be distinct from discretionary accruals which convey information on earnings persistence, but contain accruals estimation errors. Design/Methodology/Approach – This study uses AFS estimated using Penman and Zhang (2006) variables as the fundamental variable and discretionary accruals measurements according to Dechow et al. (1995) and Kothari et al. (2005). Frequency analysis and portfolio tests on firms listed in Korean stock exchanges (KSE) are conducted. Also, the hypothesis is verified by regression analysis with robust statistic test. Findings – The findings are as follows. First, AFS systematically mitigates the negative effect of discretionary accruals on future earnings increases. Additionally, AFS-based 12 month buy-hold returns are effective, but the discretionary accruals-based returns are not, in part. These results hold consistently on both the portfolio test and regression analyses. Research Implications – The results imply AFS contains information on future earnings increases independent of discretionary accruals. This study contributes to academic research by suggesting the future earnings predictability of accounting information based on financial statement analysis. Also, this study has practical implications in that two measures reflect the partially indifferent information of each other, and suggest an investment strategy exploiting ex-ante information.
Read full abstract