TRADE REFORM: Wage Inequality and Globalization 5

To control for institutional features of labor markets, we include plant shares of social-security contributions, non-wage payments, and profit-sharing with workers in total labor costs.

Table 5 presents results for the relative-wage regressions. Relative wages are defined as the log ratio of white collar to blue-collar average annual average annual wages. To address the possibility of multicollinearity, we begin by regressing in the first column relative wages on the two trade policy variables without any other controls. We then redo the analysis in first differences in the second column. Unfortunately, due to very little time series variation in many of the other independent variable, we only redo (1) in first differences. In the third column, we add other measures of openness. In the fourth column we add technology and labor market institution variables. Finally, in the fifth column we add additional measures of technological change. so

In columns (1) and (2), the results are mixed: high industry tariffs are associated with greater wage inequality, while high quotas are associated with the opposite. The second column reports the same regression in first differences: changes in relative wages were regressed on changes in tariffs and quotas. The results are again inconclusive: there is no statistically significant correlation between changes in relative wages and changes in trade policy.

In the next three columns, we redo the analysis, controlling for other measures of openness, technology, capital intensity, size, and labor market institutions. Consider first the results for export activity, foreign ownership, and trade protection. Plants that participate in foreign product or capital markets appear to pay relatively high wages to skilled labor. Relative wages are positively and significantly correlated with the industry share of exports in sales, indicating that the skilled-unskilled wage gap is higher in high-export industries.

Relative wages are also positively correlated with plant-level foreign investment, but not with sector-level FDI. These results suggest that foreign investment locates in sectors with more income inequality, but that foreign firms themselves pay a higher premium to skilled workers. Consistent with the results in columns (1) and (2), relative wages are not significantly correlated with tariff rates or import-license coverage rates.

Among the technology variables, only the share of royalty payments in plant sales is positively and significantly correlated with relative wages. This result suggests that the skilled-unskilled wage gap is higher in plants that upgrade their technology through licensing arrangements. No other technology variables are statistically significant. Other plant characteristics are also important in explaining variation in the skilled-unskilled wage gap.

Tags: , ,