Working Paper: NBER ID: w11280
Authors: Mitchell A. Petersen
Abstract: In both corporate finance and asset pricing empirical work, researchers are often confronted with panel data. In these data sets, the residuals may be correlated across firms and across time, and OLS standard errors can be biased. Historically, the two literatures have used different solutions to this problem. Corporate finance has relied on Rogers standard errors, while asset pricing has used the Fama-MacBeth procedure to estimate standard errors. This paper will examine the different methods used in the literature and explain when the different methods yield the same (and correct) standard errors and when they diverge. The intent is to provide intuition as to why the different approaches sometimes give different answers and give researchers guidance for their use.
Keywords: Standard Errors; Panel Data; Finance; Fama-MacBeth; Clustered Standard Errors
JEL Codes: G1; G3; C1
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| OLS standard errors are biased downward when residuals are correlated across observations (C20) | incorrect inferences about the significance of coefficients (C20) |
| Fama-MacBeth standard errors are biased in the presence of firm effects (C51) | underestimation of standard errors (C20) |
| clustered standard errors account for the residual dependence created by firm effects (C23) | unbiased standard errors (C51) |
| the choice of standard error estimation method significantly affects the validity of statistical inferences in finance research (C51) | accuracy of standard error estimates (C20) |