Skill-biased technical change has occupied empirical economists for much of the 90s. However, the empirical literature has not progressed much beyond observing a positive correlation between technology indicators and demand shifts. Two hypotheses on the root causes of skill biases in technical change, the acceleration effect and the market size effect, have been suggested in the literature. In this paper both are studied in a unified theoretical framework to derive the sufficient and necessary conditions for both hypotheses. Confronting them with the evidence the paper concludes that it favors the acceleration hypothesis but further empirical work needs to be done.