Performance measurement in the public sector: in theory and practice

This thesis examines attempts to improve the efficiency of local authority bureaucracies during the 1980's. A number of significant policy initiatives such as the establishment of the Audit Commission, the pressure for local authorities to implement systems of performance measurement and the introduction of performance related pay schemes reflect this quest for improved value for money. The thesis reviews existing economic theories of bureaucracy which show how inefficiency arises in the public sector. Two major hypotheses within this literature are identified; bureaus are inefficient because it is in the interests of bureaucrats either to produce too much output (allocative inefficiency) and/or to produce output at above minimum cost (X-inefficiency). The policy prescriptions arising from this theoretical framework suggest that strategies to reduce inefficiency must aim to change bureaucratic behaviour. One strategy seeks to induce bureaucrats to produce efficiently, whilst the other seeks to provide sponsors with the necessary information on costs to enable them to force bureaucrats to produce efficiently. Performance related pay schemes, which aim to change bureaucratic behaviour, are concerned with eliminating labour X-inefficiency. Our research suggests that the diversity of current schemes reflects a lack of consensus over the definition of indicators of employee performance. The general result of the introduction of performance related pay has been increased salaries for senior officers. Our assessment of the work of the Audit Commission in the area of value for money audits and our empirical research on the impact of performance measurement in local authorities indicates that a wealth of information has been generated in the form of performance indicators (PIs). However, the use of this information as a control device is limited as these indicators are clearly biased towards measuring X-inefficiency as distinct from allocative efficiency.