The Effect of an Income Shock on Subnational Debt: Micro Evidence from Mexico

Abstract

This paper examines how the borrowing decisions of local governments, often financially constrained, respond to a shock that affects the distribution of revenue from the central government. The shock stems from the discrete updating of population census data that is plausibly uncorrelated with short-term financing needs. For a one-standard-deviation increase in the population shock, I find that federal transfers to Mexican municipalities increase by 3% over the first two post-census years. Using supervisory loan-level data, I show that the probability of municipalities being indebted declines by 0.2 percentage points over the same period. This response is driven by governments with relatively more own-source revenue and, hence, less dependent on transfers, which lenders perceive as more creditworthy. The shock has no differential effect when debt is granted by a public lender, not even during electoral years or when the mayor is political aligned with the central government. Most of the additional revenue goes to finance short-term, current expenditures, especially by the more transfer-dependent governments. These findings reveal a small capacity to smooth shocks in credit markets, restricted to few governments with a diversified revenue base that borrow from private intermediaries.