On the evolution of wage inequality in Acemoglu's model of directed technical change

Abstract

In Acemoglu's model of directed technical change, the skill-premium increases in consequence of an increase in the relative supply of skilled labour. In this article, I argue that other measures of wage inequality such as the Gini-coefficient do not necessarily rise as well. The Gini-coefficient depends positively on the skill-premium, but the effect of an increase in the relative supply of skilled labour is ambiguous. A simulation of Acemoglu's model shows that the growth in the relative supply of skilled labour has led to increased wage inequality in the past, but will lead to decreasing wage inequality in the future.

Acknowledegment

I am grateful to Melanie Lührmann for valuable comments and suggestions.

Notes

¹ See, e.g. Gottschalk and Smeeding (1997), Berman et al. (1998), Katz, and Autor (1999), and Acemoglu (2002b).

² See Acemoglu (1998, 2002a, b) and Kiley (1999).

³ In the extreme case where all workers have become skilled, the wage premium may be gigantic, but it's irrelevant.

⁴ The requirement for this result is that the elasticity of substitution between the production factors is larger than 1. In the case of unskilled vs. skilled labour, this is supported by the empirical evidence.

⁵ For ease of presentation, I abstain from endogenous growth. The spirit of the model of directed technical change is not affected by this simplification.

⁶ In this model, A_l and A_h are just two factor inputs that are chosen such that marginal productivity equals marginal cost.

⁷ See also Table 1, column (0).

⁸ I tried 5, 6 and 7-year lags. The results are similar. Results with 7-year lags are reported.

⁹ The values proposed by Acemoglu imply w_h /w_l  = c ⋅ (H/L)^0.006. For the simulation, I chose c = 1.7 from the baseline estimation. Other values for c yield very similar results.

¹⁰ For the specifications with lags, I calculated the steady state relationship between w_h/w_l and H/L in order to display the Gini-coefficient in the two-dimensional Gini-H/L-space.

¹¹ This argument has also been used to promote subsidies to education.

¹² Education subsidies would then seem counterproductive for fighting wage inequality: ‘… subsidies to education lead to an increased tendency to acquire education, and also to a larger education premium due to the directed technology effect.’ (Acemoglu (1998) p. 1078).