Efficiency and compensation of public officials

This paper recommends a system of rewards for public officials which is linearly related to the performance of the official. The main advantage of such a system of compensation is that it serves as a regular ‘self checking’ of the performance of an official and it reduces the need for a process of direct supervision. This is an important characteristic, since direct supervision by an ‘uninformed government’ is quite difficult, at least on a regular basis. On the other hand, judgement by result seems to be much easier to carry out. Furthermore, we suggest that the system of linear compensation could be improved with an ‘entrance fee’ which is determined in an auction among qualified candidates competing for the right to work for the Government. We have also shown that the dismissal practice on the basis of realized results will lead to an inappropriate choice of policy and inefficiency. We suggest, therefore, that this practice (of dismissal) should be limited only to a screening mechanism, where official can be dismissed for bad decision but not for bad ‘luck’. Bad decision is a choice of a non optimal policy by an incompetent official (i.e A i notA k ). While bad ‘luck’ is an occurrence of an undesired outcome (X ij ) eventhough the optimal policy A k was selected. This view of the role of dismissal suggests a need for a more detailed examination of the decision process before a dismissal takes place. As suggested by G. Tullock one can reduce the errors of dismissal by using a detailed survey of the actual choices facing the official for a sample of cases. Such a sample will include the non-random cases which otherwise require dismissal, as well as a random sample of other decisions. Another advantage of the random sample is that it may reveal some cases in which a ‘good outcome’ is a result of poor decision and good luck. In such cases an official may be dismissed for his poor decision despite his realized good performance. This procedure improves the automatic mechanism of screening suggested in our model. Finally, we have shown, that an incentive wage scheme can be used to insure the optimal level effort by an official on the job. Our compensation scheme may reduce the number of dismissals and it will lead to an increase in the use of wage reduction as well as wage increase as a function of observed performance. Under this proposal the official is more directly involved with his work and acts as a partner to the Government in the particular aspect of the job. Copyright Martinus Nijhoff Publishers bv 1980