Exploring reasons why cross-section analysis underestimates the union premium in Korea

Abstract

This study evaluates the cartel effects of Korean labour unions from the perspectives of job separation and loss of wage premium. Unlike the previous cross-sectional analysis showing that the wage premium of Korean labour unions is small, a high level of wage premium was observed in unionized companies. This was more prominent among long tenured union workers and among union workers in monopoly firms than it was amongst workers in other comparable groups. This demonstrates that those who work in unionized companies and enjoy a high level of employment security concentrate on building up firm-specific skills, and not on building-up alternative skills as preparation for potential job turnover.

Notes

¹ Until now, the wage effect of Korean labour unions has been estimated using cross-sectional data. For example, Cho and Yoo (1997) estimated the Equation 1 using annual cross-sectional data from 1987 to 1994. According to this, labour union wage premium was not observed both in male and female workers, and in some cases, the premium was negative (−). Wage premium started to go up in 1987, reaching a peak in 1989, but even at the peak it was only 4.4% for male employees and 2.2% for female employees. Afterwards, the wage premium started to go down to–1.8% in 1993, and–2.3% in 1994 for male workers. In conclusion, the wage premium of Korean labour unions has never gone over the level of 5%.

² References for the American wage premium are Lewis (1986), Jakubson (1991), Freeman and Medoff (1984), and Karier (1985). According to the Jakubson method (1991, Table 2, p. 979) that controlled the workers’ heterogeneity, the union premium goes down to 7.4%. The 5% wage premium in the Korean case may go down slightly if it is re-analysed based on this method.