Labour Economics
Turnover costs

In partial equilibrium models of the effects of `employment protection' measures, with wages and business conditions given, rising turnover costs reduce employment fluctuations, but their implications for its average level are ambiguous, since obstacles to firing also discourage hiring. In Discussion Paper No. 601, Research Fellow Giuseppe Bertola develops a dynamic model of labour demand in which turnover costs are a component of labour costs and hence inversely related to employment. Firms' revenues depend on the labour they use and a set of exogenous, cyclical variables. For a given wage rate, firms incur positive costs in hiring or firing, which should be equal to their effects on future discounted cash flows. Firms first decide whether to act or passively to accept a steady reduction in employment; they then choose the level of control to exercise at each point in time.
Bertola finds that the direction of turnover costs' effect on average employment depends on the relative steepness of labour's marginal product when firing and hiring; labour attrition over a typical hiring-firing cycle; and the relative size of hiring and firing costs. The optimal labour demand path for a discounted objective function is partially myopic, and firing costs tend to increase myopic firms' average labour demand.
Bertola notes that such a positive effect of enhanced job security on employment may explain employed and unemployed workers' support for permanent restrictions on firing, which would otherwise protect incumbents at the expense of average employment. His analysis also suggests that long intervals may be required between hiring and firing for the latter's costs to raise labour demand, which may then reduce the average labour demand of firms whose business is highly seasonal or volatile or which have little market power.
This theoretical model also has interesting implications for empirical work. It suggests that turnover costs affect employment dynamics much more strongly than its average level; while job-security provisions should have relatively small effects on average labour demand over the cycle. These results are also relevant to general equilibrium analysis and, in particular, to the joint determination of wages and turnover costs. Making both employment and wages endogenous would require detailed modelling of both labour and product markets, in which case imperfect information and market incompleteness would significantly influence the real effects of job-security provisions. The finding that job-security provisions have small and possibly positive effects on average employment may also be relevant to studies of optimal contracts in an explicitly dynamic framework.

Labour Turnover Costs and Average Labour Demand
Giuseppe Bertola

Discussion Paper No. 601, December 1991 (HR)