Contracting out refuse collection

In this paper we seek an explanation for the reservations of local authorities towards contracting out. Although empirical evidence suggests that contracting out results in a significant cost decrease, a majority of Dutch municipalities provides for waste collection services themselves. Based on theoretical insights we model the choice between private, public, in-house, and out-house refuse collection. The models are estimated using a database comprising nearly all Dutch municipalities. We find evidence that the number of inhabitants, the transfer by central government, and interest group arguments are important explanations. Interestingly, ideology seems to play a minor role. Compared to earlier studies we estimate more general models. Although the same qualitative results are found for parametric and semiparametric models, we find strong statistical evidence that a parametric specification is far too inflexible. Differences between the parametric and the semiparametric marginal effects are substantial. Thus, more attention is needed for the implications of model specification. Copyright Springer-Verlag Berlin Heidelberg 2003Key words: Refuse collection, institutional choice, ideology, interest groups, semiparametric estimation,