Abstract
Recent work uses the change in equity value surrounding an announcement of a change in financial policy as an explanatory variable in regressions that examine whether changes in financial policy convey information about firm performance. We explore the methodological and econometric issues of this approach and show that using the change in equity value as an explanatory variable can severely bias ordinary least squares estimates. We demonstrate the effect of this bias on standard statistical tests and conclude that the bias has a significant impact on the power of these tests. We propose an estimator to partially correct the biases.
Original language | English (US) |
---|---|
Pages (from-to) | 209-216 |
Number of pages | 8 |
Journal | Journal of Business and Economic Statistics |
Volume | 8 |
Issue number | 2 |
DOIs | |
State | Published - Apr 1990 |
Keywords
- Corporate finance
- Errors in variables
- Event studies
ASJC Scopus subject areas
- Statistics and Probability
- Social Sciences (miscellaneous)
- Economics and Econometrics
- Statistics, Probability and Uncertainty