HomeAbstract of 10th-icrbmfPanel-Corrected Standard Errors in Executive Compensation Research: Implications for Estimation and Inference

Panel-Corrected Standard Errors in Executive Compensation Research: Implications for Estimation and Inference

  • Jun 24, 2026
  • Posted by:
  • Category: Abstract of 10th-icrbmf
No Comments


Abstract Book of the 10th International Conference on Research in Business, Management and Finance

Year: 2026

[PDF]

Panel-Corrected Standard Errors in Executive Compensation Research: Implications for Estimation and Inference

Dr. Ramesh Dangol, Peter D Woodlock; Yogesh Uppal

ABSTRACT:

Executive compensation research has traditionally relied on pooled ordinary least squares (OLS), fixed effects (FE), and feasible generalized least squares (FGLS) estimators to examine the determinants of CEO pay. This reliance reflects the empirical structure of executive compensation data, in which key explanatory variables such as firm size, industry norms, executive tenure, and strategic orientation are time-invariant or change slowly. While pooled OLS preserves this economically meaningful variation central to executive compensation theories, it implicitly assumes that errors are independently distributed across firms and over time. This assumption is unlikely to be held in executive compensation settings, where firms are subject to industry changes and peer-group pay-setting practices, thereby generating cross-sectional dependence in the error structure. On the other hand, Fixed-effects models attenuate or eliminate the effects of such variables, leading researchers to incorrectly conclude that time-invariant and slowly changing explanatory variables have little influence on compensation. FGLS estimators, by contrast, retain these variables but tend to underestimate standard errors and inflate test statistics. Departing from traditional estimation practices, we show that panel-corrected standard errors (PCSE) provide more reliable inference by explicitly accounting for heteroskedasticity and cross-sectional dependence while retaining slow-moving determinants of pay. Using time-series cross-sectional data on executive compensation, we demonstrate that inferences about the effects of firm size, industry, CEO tenure, and strategy are sensitive to estimation choices. These findings suggest that methodological choices, rather than differences in theory or data, may explain inconsistent conclusions in the executive compensation literature.

Keywords: Executive Compensation; Estimation Methods; Panel-Corrected Standard Errors (PCSE)