There is a wide agreement among researchers that increase in wage inequality across the OECD countries was caused by introduction of skill-biased production methods, which generated higher demand for skilled than for unskilled workers. However, does a skill-bias of production methods originate from the skill-biased nature of new global technological paradigm (i.e. Information Technology) or from the fact that following the increase in a number of college graduates firms have chosen to exploit the new technological paradigm in a way that favoured skilled workers?

To study this question, I first observe that while the source of the latter cause is global, the source of the former rests in labour market at the country-level. Therefore to calibrate the model I isolate global from local sources of college wage premium using both cross-section and cross-time dimension of the data. The exercise implies that endogenous technology choice at the local level can explain 30% of the increase of college premium in the OECD countries. In econometric analysis I find that countries which experienced higher growth in a number of college graduates than other countries witnessed also higher growth of college wage premium a decade later. A number of arguments (and a setup of the regression) suggests that endogenous technology choice at country level is the most plausible explanation for this finding. One implication is that any policy that affects supply of skilled workforce will have an impact on skill-bias of equilibrium technology and wage inequality dynamics. At theoretical level, I set a microfoundation for the model by showing how research in the R&D sector might generate a tradeoff between skill- and unskill-biased technologies.