<p>The analysis of the expenditure decisions of English local authorities has assumed great importance as central government has sought to exercise increasing control over the activities of local government. In particular, in a variety of contexts, central government has sought to estimate from empirical observation what a local authority 'naught' to spend. Unfortunately, such an undertaking is becoming increasingly complex, as the influence of previous government policy itself assumes greater importance in local authority expenditure decisions. For example, central government grant allocations to local authorities are based on a simple statistical analysis of previous spending patterns. These grant allocations will to some extent influence current spending. The expenditure responses in turn are likely to affect future grant allocations, and so the cycle continues. Such circularity formed an important component of criticisms of current local government finance arrangements by the Audit Commission (1993) and the House of Commons Select Committee on the Environment (1994). This paper seeks to underline the difficulties by demonstrating the statistical methods that are required to model spending patterns amongst non-metropolitan districts satisfactorily. The structure of the paper is as follows. </p>