We use data from the period 1978-92 to perform earnings-equation estimations that control for a wide range of worker-productivity characteristics. Overall, we find that controlling explicitly for permanent location and occupation effects consistently yields a positive unemployment elasticity, while excluding these local fixed effects produces a negative relationship. These full-sample results suggest that, while there may be a longer-run wage-curve relationship in Taiwan, the transitory elements of wages and unemployment are positively correlated. Hence, the compensating-differentials framework is a better description of shorter-term pay determination in Taiwan. However, modest wage-curve effects are present for particular groups such as women and less-educated workers, even when we control for permanent region and occupation effects. We conclude that the wage-curve mechanism may indeed operate in certain settings, although it fails to describe Taiwan's short-run relationship between wages and unemployment in the overall labor market. Our results are in contrast with Blanchflower and Oswald's finding of a consistently negative relationship between wages and industry unemployment in Korea during the 1970s and 1980s. A potential explanation for this discrepancy is that Korea's tradition of lifetime employment during that period left firms with less maneuvering room in the labor market, forcing them to use real wage cuts to respond to contractionary conditions. Taiwan's firms, less bound to this tradition of lifetime employment, appear to have used layoffs, rather than real wage cuts, as the adjustment mechanism.
All Science Journal Classification (ASJC) codes
- Economics and Econometrics