What is the influence of the structural features of local systems of representation in local policy-making? In particular, what is the effect of the size and structure of city councils in the patterns of local government spending? This research addresses the link between local government expenditures and political representation defined as the ratio of representatives to constituents. The goal here is to explore a possible trade-off between the accuracy in representing constituency preferences and local legislative monitoring and decision-making costs. An increase in the density of representation is likely to increase the accuracy with which constituent preferences are represented, but this may be offset by increased bargaining costs in the city council and by higher costs in monitoring a larger number of representatives. Since Portuguese city councils are equally divided between parish presidents (Presidentes de Junta) representing local subunits and directly elected members, one may argue that the size and structure of the council affects the costs and the quality of representation as well as the pattern of local policy choices. Three arguments concur to this: first, parish representatives may be interested in promoting parish interests much in the same way as district representatives in local elections in the United States; second, directly elected members take a broader approach, similar to the one assumed by representatives elected at-large in US municipalities; finally, the sheer size of city councils (in many cases above 100 representatives) improves the density of representation at the expense of increased monitoring and decision-making costs. The main hypothesis argues that an increase in the number of parish representatives and city council members is associated with significant increases in local government expenditures in distributive policies targeting geographical interests. Since city councilors have the final vote in budget approval, we also expect an increase in total expenditures as a result of logrolling and vote trading among councilors. We test the hypotheses derived from extant research and discuss the implications of our findings for the literature on local institutions and political representation.