The purpose of this paper is to estimate the union wage effect of Korea before and after 1987 using occupational wage survey data. When we estimate the union wage effect using firm size dummy variable as most earlier studies did, results are likely to depend on the results of the most large sample group. The estimated results by firm size showed that the union wage effect varies from firm size to firm size. This is caused by different estimated unionization rates by firm size. To obtain, therefore, the proper union wage effect, it should be estimated controlling for firm size.
