Browsing by Author "Mitrunen, Matti"

Job polarization, loss of middle skilled employment relative to low and high skilled, has been visible in many western countries in the past decades (Goos et al. 2009), but its causes are not completely understood. Most of the research blames uneven technological change, which allows machinery to replace routine middle skilled work (Autor et al. 2003, Autor & Dorn 2012), but cheap foreign workforce could also contribute to the job polarization trend by replacing middle skilled manufacturing labour. This relationship between job polarization and offshoring has been explored with indices, but not with actual trade data. The contribution of this thesis is to show the Finnish job polarization trend in detail, and use firm-level imports data to examine its relation to offshoring. I have in my disposal Structure of Earnings Statistics, Finnish Longitudinal Employer-Employee Data and firm-level imports data from 2000 to 2008. These data allow me to illustrate job polarization trend in Finland, as well as the trend within firms. I divide Finnish labour in low, middle and high skilled occupations, and create a firm-level variable for how much each of these groups is employed in a firm. I also construct a variable measuring firm’s offshoring, which is done following Feenstra & Hanson (1999). I estimate basic Ordinary Least Squares (OLS) regressions to figure out how much exactly the offshoring of a firm affects the skill structure of a firm. I control for technological aspects and add firm and time specific effects. I perform this exercise also at the industry level, because it is more likely that the structure of industry is changing due to increasing offshoring. To assure causality I develop my own industry-level instrumental variable (following Autor et al. 2011), and run Two-Stage Least Squares (2SLS) regressions. I find that there has been job polarization in Finland, where the middle skilled labour has lost approximately 7 percent points of its share between 2000 and 2008. Furthermore, I find that the losing occupations perform excessively routine tasks. In this study I find that offshoring does not explain job polarization very well at the firm level, when added firm and time specific effects. My instrumental variable approach states that offshoring does cause statistically significant demise in the share of the middle skilled at the industry level, but these results are subject to some doubt, since the instrumental variable might be weak. I conclude that offshoring might have an effect on job polarization, but this effect should not be overestimated. Like the previous literature, I come to the conclusion that routine-intensity of work might be better explanation for diminishing middle skilled labour than offshoring.