Efficient Public Sector Downsizing

Most downsizing operations exhibit high financial returns, but their economic returns crucially depend on their design. After briefly comparing public sector employment across countries, this paper analyzes the optimal design of downsizing operations from a microeconomic perspective. It first discusses how to identify redundant workers when individual productivity is observable, as is often the case in state-owned enterprises. It shows that comparisons of productivity and labor costs are misleading, because over-staffing is only one among several distortions. It therefore proposes to use a shadow cost of labor, much the same as in standard investment projects. The paper then moves to the identification of redundancies when individual productivity cannot be observed, as in government administration. It shows that voluntary separations in exchange for severance pay lead to an adverse selection problem, whereby the best workers leave the public sector and the worst ones stay. It hence discusses other self-selection mechanisms, which potentially create an incentive for the best workers to stay rather than quit. Next, the paper deals with compensation issues. It shows that the rules-of-thumb underlying severance pay offers tend to over-compensate the workers who take them. It therefore analyzes how labor market data can be used to predict the loss displaced workers will experience and tailor compensation to their individual characteristics. Finally, the paper addresses other design issues, including the appropriate sequence of downsizing and privatization, the consequences of early retirement programs and the usefulness of training programs and other active labor market policies. The main findings are organized under the form of a decision tree.