The Proper Scope of Government: Theory and an Application to Prisons Hart, O. Shleifer, A. Vishny, R. QJE 1997 For details of the model see handwritten diagramatic notes.

Issue with original model: whole discussion is motivated by incompleteness. But in renegotiation stage we have a specification of the quality improvement that is presumed to be enforceable. Even though we allow for cost cutting to harm quality surely it is the case that the greater the intended quality improvement the greater the potential damage (and advantage to private contractor) from cost cutting. That is b(e) (cost to benefits of effort on cost cutting) is not independent of beta(i) (benefits of innovation in quality).

As discussion makes clear the major cost improvements by private firms are not of the 'efficiency' kind at all but simply involve hiring poorer quality staff. See subsequent anecdotal evidence that in part substantiates this

In model monitoring ability and contract specification for private manager is same as for public manager (except for firing issue). This is unrealistic. An 'inside' manager (e.g. public manager) will often be easier to monitor and will have incentives more closely aligned with government (the converse is possible through bureaucratic cronyism). In fact, in one sense, this is implicit in the setup in that the private manager magically has no monitoring/contract problems while the government does (the private firm itself has to employ managers and draw up contracts). Private contractors also have incentive to post facto undermine enforcement of contract provisions. (See evidence on corruption: . Also