This paper presents a shift-share decomposition applied to industry-level data to identify contributions to the rise in the hourly wage cost in Belgium between 2000 and 2010. According to the results, individual wage increases account for the largest part of the economy- wide hourly wage cost increase (87%). The contribution of changes in the gender, age and skill structure of industry-level employment amounts to 16%, while changes in the industry composition of hours worked even have a negative effect on the hourly wage cost (-3%).
The 2013 report of the Expert Group on Competitiveness and Employment has highlighted major differences in hourly wage cost levels between industries for Belgium. Thus, changes in industry shares in total employment may have affected the economy-wide increase in the hourly wage cost. Moreover, trends in the hourly wage cost have also varied substantially between industries. This may have been influenced by industry-level differences in the gender, age and skill structure of the workforce. The aim of this article is to shed some light on both these issues using a traditional shift-share decomposition to identify contributions of three effects to the rise in the hourly wage cost in Belgium: changes in the industry composition of total hours worked (composition effect), changes in the structure of employment in terms of categories of workers (employment structure effect), and increases in the hourly wage cost of these individual categories (wage effect). The decomposition effects are calculated for the years 2000-2010 with industry-level data (A38) from the national accounts published in October 2014 according to the ESA 2010 and with EUKLEMS data for 18 categories of workers according to three criteria (gender, age and skill-level). Note that the Federal Planning Bureau Working Paper 11-14 investigated the issue of whether worker characteristics affect wage cost increases in Belgium with data for individual workers. Differences in the methodology and their influence on the results are discussed in the introduction section of this paper.
In the results of the decomposition, the wage effect largely dominates, accounting for 87% of the hourly wage cost rise. The composition effect is negative but small (-3%). This is mainly due to the fall in the share of manufacturing industries in total hours worked, which have a relatively high level of hourly wage cost, and the rise in the share of non-market service industries, which is the group of industries with the lowest hourly wage cost level, both in 2000 and 2010. Finally, the contribution of the employment structure amounts to 16% of the economy-wide rise in the hourly wage cost. This effect is closely related to the rise in the average skills and age of employees.
The shift-share decomposition has also been applied at the industry-level so as to determine to what extent changes in the gender, age and skill structure of the workforce have contributed to hourly wage cost rises in each industry. According to the results, the employment structure effect is below 50% for all industries. It is highest for the ‘Manufacturing of computer, electronic and optical products’ industry followed by four market service industries: ‘Telecommunications’, ‘Computer and information service activities’, ‘Publishing, video and broadcasting activities’ and ‘Advertising and other specialised activities’. In all of these industries, the share of workers with tertiary education increased fast over the period 2000-2010. Hence, the industry-level employment structure effect seems to be influenced in particular by changing skill levels.