³ÉÈË´óƬ

In many countries, the central government transfers resources to local governments. This raises the question of whether local expenditures respond asymmetrically to changes in intergovernmental grants. The existing literature conjectures that in a regression explaining expenditures as a function of grants and other variables, a difference in the estimated coefficients for increases and decreases in grants would reflect an asymmetry. However, it finds virtually no evidence of this. In this paper, I argue that this approach is flawed. I formulate a model in which a vested interest engages in costly lobbying to resist expenditures cuts. Lobbying is triggered when grants fall below a threshold that depends on other conditions that influence expenditures. This threshold is equal to previous grants (as the standard approach assumes) only by coincidence. The model suggests the possibility of multiple steady states. Due to this multiplicity, a temporary change in grants can have a permanent effect on expenditures by shifting the system from one steady state to another. I estimate the model using data on administrative expenditures by Mexican municipalities during the years 1990-1999. When the standard approach is applied to the same data, no asymmetry is found. In contrast, the estimates of my model imply substantial asymmetries. Resistance by the vested interest increases expenditures by 6.2% on average. When I correct for the potential endogeneity of grants, this figure increases to 19.6%. Restrictions that lead to the standard model are rejected at high levels of confidence.