Changes in the composition of labour supply : implications for wages and unemployment
This thesis aims to show that usually neglected compositional changes of the labour force over the last 30 years (more young workers, more women, more workers with low attachment to the labour market, less experienced workers, more educated workers), can explain important stylized facts. First, about unemployment: the countries with high female unemployment are also countries with high young unemployment (the correlation across OECD countries between the two rates is positive and quite high at about 0.84). This can be explained by a shift towards more competition among young / women in a secondary labour market due to increased labour supply of the two groups. About 4 points of total unemployment are associated with observed changes in labour supply, if endogeneity of participation variables with respect to unemployment is properly accounted for. Second, about wage inequality: the most important trend in wage inequality in the US is the rising return to experience, or equivalently the deterioration of the relative position of younger workers. The second fact can be explained by a historical change in the skill composition of the labour force - workers though more educated are yet less experienced. This substitution of skills can also explain a significant part of the unexplained rise in the return to education, depending on the substitutability between education and experience in human capital: more inexperienced workers in the labour force generate a relative scarcity of human capital that increases the demand for education. In the theoretical part of the thesis, two trends of the OECD labour markets are explored within matching models. First, it is shown that more short-term employment can be explained by higher active population growth and lower productivity growth. Second, stronger urban unemployment gradients and higher aggregate unemployment can be shown to reinforce each other when location choices within agglomerations are endogenous.