Skip to main content

Table 6 Variance decomposition: estimated firm effects and predicted wage from worker characteristics

From: The role of the firm in worker wage dispersion: an analysis of the Ghanaian manufacturing sector

  Surge in inequality Decline in inequality
  1992 to <1998 1998 to 2003
  (1) (2)
All firms   
Earnings dispersion: Var(lnw) 1.460 1.318
Individual characteristics: Var(κ) 0.235 0.331
Firm fixed effect: Var(δ) 0.595 0.385
Cov(κ,δ) 0.175 0.164
Sorting: worker–firm ρ δ 0.745 0.495
Residuals: Var(u) 0.498 0.385
Firms with private domestic ownership   
Earnings dispersion: Var(lnw) 1.551 1.369
Individual characteristics: Var(κ) 0.308 0.400
Firm fixed effect: Var(δ) 0.387 0.232
Cov(κ,δ) 0.124 0.081
Sorting: worker–firm ρ δ 0.403 0.202
Residuals: Var(u) 0.580 0.535
Firms with any foreign ownership   
Earnings dispersion: Var(lnw) 0.803 0.825
Individual characteristics: Var(κ) 0.139 0.214
Firm fixed effect: Var(δ) 0.134 0.441
Cov(κ,δ) 0.036 −0.075
Sorting: worker–firm ρ δ 0.259 −0.350
Residuals: Var(u) 0.808 1.413
  1. Data from CSAE-RPED on the Ghanaian manufacturing sector from 1992 to 2003. Regressions are extensions of Mincerian equations and include the following independent variables: age, age squared, education, education squared, experience and its squared, and firm fixed effects. Dependent variable is log real hourly earnings, κ is predicted wage from observed worker characteristics, δ is firm effect on earnings, and ρ δ is the correlation between the worker characteristics and firm effects