Abstract:
It is often argued that present practices in evaluating employment effects of decreases in the cost of unskilled labor lead to under-estimations: an aggregation bias is ignored when microeconometric estimations are transposed. A model in which appear two elasticites of substitution, one in production, the other in consumption, clarifies the issue. When the first elasticity is smaller, the conjecture is usually validated. But the opposite case also exists.