Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
In this paper we propose a jackknife method to determine individual and time effects in linear panel data models. We first show that when both the serial and cross-sectional correlations among the idiosyncratic error terms are weak, our jackknife method can pick up the correct model with probability approaching one (w.p.a.1). In the presence of moderate or strong degree of serial correlation, we modify our jackknife criterion function and show that the modified jackknife method can also select the correct model w.p.a.1. We conduct Monte Carlo simulations to show that our new methods perform remarkably well in finite samples. We apply our methods to study i the crime rates in North Carolina, ii the determinants of saving rates across countries, and iii the relationship between guns and crime rates in the U.S.