Some argue that technological advances have so increased skill requirements that the relatively uneducated may find themselves not only unemployed but `unemployable' if they lack the skills required by the modern labour market. Economic theory suggests, however, that downward real wage rigidity and increased technological sophistication should raise the demand for and wages of skilled labour, not generate unemployment among the unskilled.

In Discussion Paper No. 689, Research Fellow Gilles Saint-Paul develops a matching model with rigid real wages in which firing is costly or impossible. High- and low-skill workers are in distinct labour markets; the numbers of matches per period depend on numbers of unemployed and vacancies posted in each market. The total number of jobs remains fixed, but each may be held by either worker type, and low-skill workers are always less productive than high-skill workers. Firms' decisions to hire low- or high-skill workers generate arbitrage between the values of vacancies in the two markets; they earn more from high-skill than from low-skill workers, but posting vacancies in the skilled market tightens it relative to the unskilled market. Vacancies therefore fill in both markets until firms are indifferent between worker types.
Saint-Paul finds that increases in the relative productivity of the skilled raise unskilled unemployment, reduce skilled unemployment, and unambiguously increase aggregate unemployment if firing costs are high. A rise in the proportion of skilled workers raises unemployment rates in both markets (so aggregate unemployment may rise or fall). These effects weaken as firing costs fall, which suggests that the paradoxical rise in the unemployment rate of the less numerous unskilled results from the irreversibility of the hiring decision.

Saint-Paul concludes that technological change generates unemployment by increasing either the relative productivity or the supply of skilled workers. As the unskilled earn less, firms are less willing to risk being stuck with them, so they become `unemployable'. These results suggest that the effects of training programmes often advocated as means of reducing unemployment, which is typically higher in low-skill categories depend on their implementation and may be quite perverse. If such programmes train only a proportion of the unskilled to become skilled workers, aggregate unemployment is likely to increase. Raising the productivity of the entire unskilled labour force will probably reduce aggregate unemployment, however, which suggests that improvements in schooling are more likely to reduce unemployment than more limited training programmes.

Are the Unemployed Unemployable?
Gilles Saint-Paul

Discussion Paper No. 689, August 1992 (IM)