Identifying and Correcting Publication Selection Bias in the Efficiency-Wage Literature: Heckman Meta-Regression

Publication selection bias represents one of the most serious challenges to the integrity of empirical economics. We develop Heckman regression methods to solve this potentially persistent problem and apply these meta-regression methods to seventy five empirical estimates from the efficiency-wage literature. Although many researchers find mixed or ambiguous support for the efficiency wage hypothesis (EWH), our meta-analyses give unambiguous confirmation of the EWH. After correcting for publication selection bias, we estimate the wage elasticity of output to be 0.32, much smaller than what the neoclassical version of the efficiency wage hypothesis demands. This wage elasticity also depends significantly upon whether the researchers’ model accounts for the simultaneity of wages and productivity and whether their empirical model includes capital. In both cases, the ‘correct’ specification increases the magnitude of the wage elasticity of production, thereby further corroborating the EWH. JEL Classification Codes: J3, C2, E24.