This paper examines international wage determination using a new data set of wages for the year 1998 from a sample of medium to large companies in 58 countries. The data were collected to provide simple and comparable information about wage levels in a large cross-section of diverse countries. Good wage data exist from official sources for several high-income countries. However, these data are sometimes too detailed and usually not immediately comparable, due to the lack of synchronization in reporting and the lack of common standards. For many poorer countries the problem is simply lack of data.
The data will be discussed more extensively below, but briefly, it was collected in early 1999 through identical surveys of managers in large firms in 58 countries. Each executive was asked to report monthly take-home-pay in their company for five occupations: Janitors, Drivers, Secretaries with five years experience, mid-level managers and top managers. Companies were also asked to report their sector of operation, number of employees, multinational affiliation and a number of other questions pertaining to the nature of their business and labor market practices in their company and main country of operation. The strategy was to ask relatively simple questions to obtain a high response rate, with full awareness that there would be some cost in terms of precision. The advantage of this data is that it covers a large number of countries, and asks the same questions about wages to executives within each country at the same point in time. The average sample size is 56 firms per country.
The composition of the sample was controlled for ex-ante through sample design and ex-post through regression analysis. The ex-ante part is that the sample across economic sectors in each country was designed to be proportional to the distribution of the non-agricultural labor force. The ex-post part is that we control for firm sector, type, size etc. in regressions before calculating the mean wages used in further cross-sectional regressions. As the reader can no doubt imagine with such an exercise, we cannot guarantee complete accuracy in the reported wages and salaries. Nevertheless the data do withstand a number of quality checks.
This paper is based on two kinds of regressions. In the first kind of regression wages by occupation, company, sector and country are regressed on a vector of company characteristics, sector-specific dummy variables and country specific dummy variables. This kind of regression exploits within country variation to learn about the characteristics of companies that are associated with wages. These regressions examine the association of wages with the multinational status of the company, public/government nature of the company, size of the company, recent performance of the company, the nature of the competitive pressures facing the company, and the economic sector of the company. These results differ from previous studies in that we have a world-wide sample, but we confirm several results found in previous individual country studies. In the second kind of regression mean wages by occupation are regressed on a list of country-specific variables. Our measure of the mean wage is not the unconditional mean but rather the estimated coefficients on the country dummy variables from the first kind of regression. Since these country fixed effects are taken from regressions with size and industry dummy variables, they represent estimates of the mean wage by country after controlling for the composition of company size and industry in the sample.
This second kind of regression exploits between-country variation conditional on average local GDP per worker to learn about the determinants of the average level of wages by occupation in different countries. These regressions test for the impact of product market competition, foreign language attainment, labor market regulations, wage setting institutions, and democratic rights on average wages. The section also tests whether foreign language attainment by executives serves to de-link their pay from the standard of the domestic labor market.
Download
PDF Ebook International Wage Determination and Globalization
