• D24;
  • E22;
  • L32;
  • L33;
  • privatisation;
  • public ownership;
  • refuse collection;
  • production costs


Many comparisons of public and private firms use a public/private ownership dummy variable to capture cost differences. If, however, public and private firms use different production technologies, the dummy-variable approach is misspecified. Data from public and private firms should not be pooled. Secondly, selectivity bias may arise, making it more difficult to identify cost differentials that actually exist. Thirdly, if data should be pooled, the resulting empirical model may be logically inconsistent. This paper compares public and private firms using the refuse collection costs of 170 firms in 115 Swedish municipalities. First, public production costs were 6 per cent lower than private production costs. Secondly, cost differences did not affect producer choice. It is crucial to adjust for selectivity. Data for private and public firms should not be pooled. The dummy-variable model is misspecified.