It has been suggested that spillover effects between politically autonomous local governments are common, and that they present a welfare problem for the community as a whole. A local government might under-invest in roads, for example, because they are used by non-residents, or in schools because school-leavers move out of the area. Other researchers, and many consumers, have found fault in the mechanisms of public choice or have claimed that electorates have negligible effective control over the local bureaucracy as far as output is concerned. There might well be a serious local welfare problem even without spillovers. This paper offers an empirically operational methodology for the welfare economic evaluation of local public spending. The setting is the system of local government in Sydney, Australia, although the theoretical approach can be readily adapted to other settings. The questions addressed are: (a) Is there evidence of positive spillovers, and do they imply welf arerelevant externalities when prices, quantities and locations have adjusted? (b) Is an increase in local spending desirable for local residents? (c) On balance, what can be said about the effect on aggregate welfare of a local fiscal expansion? The key to the theoretical approach and its testable implications is the idea that the residential mobility of households between local governments reveals preferences over local public goods. This can be traced to Tiebout (1956), although the present model is appropriate to a considerably more general setting in which restrictions on mobility, spillover effects and local government inefficiencies are allowed. The central problem for the present analysis is to infer consumer preferences over local fiscal policies from observed differences in the rates of population growth between local governments. As far as spillovers are concerned it is useful to make the following distinction from the outset. Direct externalities are said to be present when, for given consumptions of non-government goods and given residential locations, an individual in one local political jurisdiction is not indifferent between alternative expenditures by another government. Mobility externalities result when the welfare of a local resident is affected