|
Many empirical studies have demonstrated that seniority and tenure
have substantially greater effects on earnings in Japan than in the UK
or US, which may reflect a greater emphasis on training and thus account
in part for Japan's remarkable growth performance since World War II. In
Discussion Paper No. 974, Research Associate Giorgio Brunello and
Kenn Ariga note that earnings may alternatively rise with tenure
simply because good matches last longer. If this is the case,
conventional estimates of earnings profiles based on cross-section data
with unobserved heterogeneity of worker, job and/or match quality may be
biased. If the size of this bias varies substantially across countries,
comparative evidence on tenure effects may be misleading. Unobserved
worker-, job- and match-specific information may be inferred, however,
from firms' allocation of workers to jobs and training to workers both
internally and in the market-place. If `better' workers are assigned to
jobs requiring more sophisticated skills and greater responsibility,
their `rank' will provide information on both their own quality and the
quality of their match, which can be used to construct an unbiased
estimator of tenure effects. |