Long-run Behavior of Earnings Response Coefficient on A Multi-Country Analysis

Previous studies reveal that accounting information are value relevant to stock price, and that earnings response coefficient (ERC) value increases as the disclosure window is increased. To understand the long-run behaviour of ERC, this study will attempt to estimate the size of ERC using different test windows. It will examine what selected factors (e.g., capital structure, risk, firm size) affect ERC. The difference in model performance will also be estimated using both individual companies setting and portfolio setting. Meanwhile, the size of ERC will be measured in different settings. The study will also explore why the share’s book value is not equal to its market value as predicted by theory, estimating the correlation between market value change and selected factors. In addition, the study will estimate multi-country differences in ERC size among the OECD countries and Malaysia. Countries from both smaller and larger economies will be included and will be categorized into two groups accordingly. Our study will add new findings on ERC of several major economies which have not been studied yet. We will apply newer research processes to refine the measurement of ERC and also explore the correlation between ERC and selected factors not yet used in prior research, which includes identifying key determinants of market value change using panel regression, checking significant differences in country-specific ERCs, and testing the size of the difference between market value and book value under different test windows. To mitigate the problem of “errors in variables”, our research data will be grouped into three portfolios: individual companies in portfolios, sectoral portfolios, and country portfolios. Regression analysis, including panel regression, will be used in the study. In addition, the event study method will be used to examine the impact of unexpected earnings announcement (event) on stock price