Applying the Mincer Wage Equation to Japanese Data

Author Name KAWAGUCHI Daiji  (Faculty Fellow, RIETI)
Creation Date/NO. March 2011 11-J-026
Research Project Empirical Analysis of Japan's Labor Market: Policy Responses to Fertility Decline and Population Aging
Download / Links


This paper points to several caveats in an estimation of the Mincer wage equation using Japanese data. An estimation of the Mincer wage equation using microdata of the Basic Survey of Wage Structure 2005-2008 reveals following six points as caveats.

  1. Wage profile is discontinuous at age 60 because of mandatory retirement.
  2. Wage rate, which is the dependent variable of the equation, should be log transformed.
  3. Educational attainments should be included as discrete dummy variables rather than a single index as the years of education.
  4. Log wage-potential experience profiles differ across educational backgrounds.
  5. Log wage-potential experience profiles are well approximated as quadratic functions.
  6. Log wage variance within a group defined by educational background and potential experience increases as the years of potential experience increases. Thus, the error term of the Mincer wage equation is heteroskedastic.
The analysis also finds that log wage-potential experience profiles of Japan are steeper than the US counterparts based on the Current Population Survey January Supplement.