Skill Biased Technical Change and Misallocation: A Unified Framework

Due to strict reliance on competitive labor markets, standard approaches which measure skill biased technical change (SBTC) conflate labor market distortions which prevent firms from choosing the efficient ratio between skilled and unskilled labor and 'true' SBTC. This contrasts with recent evidence on decoupling between wages and productivity. To overcome this limitation, we present a unified framework to estimate SBTC which accounts for factor accumulation (FA) effects, and quantifies the discrepancy (i.e., relative misallocation) between the wage ratio (skilled to unskilled) and the marginal rate of technical substitution (MRTS). The suggested methodology takes advantage of recent developments in nonparametric estimation methods (i.e., discrete smoothing) that allow us to estimate the marginal productivity of inputs at the country-sector level directly from data. Over the 1995- 2005 period, we find a 3% yearly growth rate for the MRTS between skilled and unskilled labor and show such change to be almost entirely driven by SBTC, rather than FA, on average. In most cases, SBTC and MRTS growth do not come with increasing skill premia; this fosters the decoupling and results in relative misallocation patterns (across countries and sectors) which are quite heterogeneous, for which we report a 6% overall increase. Finally, we show evidence that relative misallocation increased less in country-sectors in which it was larger at the beginning of the period and grew more in country-sectors characterized by - (a) higher skill-intensity; (b) lower bargaining power of skilled over unskilled workers; and (c) lower FA effects.