All public sector organisations in the UK have witnessed changes in funding arrangements during the 1980s as part of the Government's drive to make them more accountable to the tax-payer. The development of performance indicators is seen as an essential step to ensure that such organisations provide value for money. This paper examines the possibility of constructing measures of the performance of UK universities. A methodology is developed in the framework of production theory and uses multiple regression techniques to estimate the relationship between the outputs and inputs of universities. Around 80% of the inter-university variation in four output measures can be explained by corresponding variations in several input measures. This highlights the need to take into account the inputs available to a university when comparing its output performance with that achieved by other institutions. The problems of interpreting an array of performance indicators are also clearly demonstrated.