For those of us who teach federalism, we often make the case for having decentralized local governments on the basis of the idea that preferences for public goods and services differ across place. A populace in North Dakota might want to maintain an outdoor ice rink, while another in Florida a public swimming pool. The potential gains to decentralization depend then on the degree of preference divergence and their spatial correlation. How much of that potential is realized partially depends on the quality of local government. A lot of research is directed at identifying local government failure, particularly in the form of tests for “fiscal illusion.” In these models, scholars look for variables that should not affect some public expenditure or public revenue outcome, like the salience of the tax rates, confusing property tax rates with levies, or the presence of renters. Very often, but not always, these variables matter and are taken as evidence that voters are systematically fooled by local politicians in an exploitable way. This increases the case for state provision and/or oversight of local government services.

I (along with co-authors Vikki Perez and Kosali Simon) have a new working paper that tries to do the opposite of this previous research in that we look for evidence that local governments respond to exogenous shocks in a way that plausibly represents local preferences. We compare local governments’ own expenditures on hospitals from 2006 to 2015. During this period, some states expanded Medicaid generosity under the incentives of the Affordable Care Act. If low income people gain access to coverage through Medicaid expansion, it becomes cheaper for local governments to support hospital services that are more likely to be reimbursed through a 3rd party payer. A local government could respond to this change by spending more on these services, or it could use the cost savings to spend on other goods, including private consumption via lower taxes. We find evidence that local governments in areas that voted for Obama in the 2012 election responded to the ACA with greater levels of spending on hospitals, while those in areas voting Romney tended towards reducing property taxes. We take this as evidence that local governments do a good job of representing local preferences. Here is the abstract:

Local governments participate in the health care system as both a patient care provider and third-party payer. In doing so, local governments represent the majority of the public health service economy, outspending state governments about $3-to-$2 on hospitals. Using data from the Census of Governments from 2006 to 2015, we employ a difference-in-difference framework to study the effect of the state decision to expand Medicaid on their local governments’ fiscal decisions. We find that local governments in areas that voted for Obama in the 2012 presidential election responded to their state’s Medicaid expansion by increasing expenditures on hospitals, whereas those that voted for Romney reduced hospital expenditures and used the savings to reduce property taxes. This finding is not driven by an urban/rural distinction. We then extend this analysis to determine (1) the extent to which observed changes in local government financing are affected by changes in hospital revenue; and (2) assess the effect of these changes in local government funding on the finances of surrounding hospitals. We rule out the possibility that changes in local government spending are a response to hospital competition and find evidence that hospitals in Medicaid expansion states observed increased profits, following the expansion. This paper provides evidence on the responsiveness of local governments to the preferences of their voters under incentives introduced by the Affordable Care Act (ACA).

This is not published work, so comments and suggestions by email on how to improve the paper are welcome and appreciated.