How does a firms technological change affect wages? The question is fundamental for our understanding of the sources of both individual wage growth and wage dispersion. Yet different theories provide very different predictions. In a competitive frictionless model workersmwages respond to aggregate labour market conditions and personal characteristics, such as human capital. The situation of the firm or its sector of operation does not play a role unless for compensating differentials related to (dis)amenities in the workplace. Similarly, in Azariadis (1975) risk&neutral firms insure risk&averse workers against wage fluctuations related to productivity and demand shocks. In contrast, rent&sharing models suggest that individual wages respond to firmsmprofitability, and consequentially fluctuate with firm movements in demand and productivity. This study uses matched employer employee data with unique firm side information to assess the role of firm level productivity for the fluctuations and dispersion of wages.
Differences in wages paid between firms are a large and growing part of the overall wage dispersion in many countries (Lazear and Shaw (2009)) including Sweden, from where we draw our data. From the extensive literature building on Abowd, Kramarz, and Margolis (1999), it is by now an established fact that some firms pay higher wages than others, even to identical workers. Consistent with this finding, a large literature has established the importance of rent sharing. However, few hard facts exist regarding the deep determinants of these persistent differences (although e.g. Mortensen (2003) argues that productivity differences should be crucial), and even less is known about how changes in a firms productivity due to technological advances affect the wages of its workers.
In a study related to ours, Guiso, Schivardi, and Pistaferri (2005) use Italian data to show that firmms pure transitory shocks in sales are not transmitted into workerms wages , while permanent shocks are less than fully transferred. Thus, in this sense firms provide insurance to workers. A more important feedback of innovations in sales into wages is found for Hungary (Kátay (2008)). An alternative interpretation of the results is that firms share rents with their workers (an argument also made by Cardoso and Portela (2009)), suggesting that local wage bargaining is important for actual wages. While innovations in firms sales are related to both productivity, demand shocks, and factor price shocks our effort is on isolating the effects of productivity.
Our analysis is therefore also related to a recent literature that estimates the impact of human capital on productivity differences across firms. Fox and Smeets (2010), using matched employer employee data for Denmark similar to ours, find that a wide range of human capital measures are not able to explain much of the observed productivity dispersion across firms. Irarrazabal, Moxnes, and Ulltveit-Moe (2009) use similar methods on Norwegian data to examine a related question: to what extent the productivity premium of exporter firms is associated with a better workforce. Their findings suggest that at most 30% of the productivity differentials between exporters and non exporters can be explained by worker sorting into exporter firms. Hence, as Lentz and Mortensen (2010) in their review paper conclude "both production studies and estimates of structure models of wage dynamics suggest that firm (rather than worker) heterogeneity is the more important explanation of productivity differences".
Download
Wage Adjustment and Productivity: Evidence from Matched Employer-Employee Data
