How do states spend federal aid windfalls?
Core argument: For every $1,000 per capita of committed federal pandemic aid, state intergovernmental grant revenue rose $645 over 2020–2022, from which the authors infer roughly 64.5 cents per dollar of the nearly $1 trillion in aid reached state governments by the end of fiscal 2022.
In 2020 and 2021, federal lawmakers enacted nearly $1 trillion in financial aid for state and local governments. In general, funds were not intended to be used to offset tax reductions or as contributions to pension funds. For every $1000 of incremental committed federal aid per capita, $645 appears as increased intergovernmental grant revenue in states over the years 2020-2022, from which we infer that approximately $0.645 of every dollar of committed federal aid to state and local governments was disbursed to state governments by the end of fiscal year 2022. Second, total revenues of states increased by approximately as much, indicating that states did not offset the federal revenue windfall with other revenue sources. Third, total expenditures of states increased by $770 per capita for every $1000 of federal aid committed, statistically indistinguishable from the revenue estimate. Fourth, the spending categories in which expenditures increase most strongly and significantly are General Government Expenditures ($379 per $1000 of incremental federal aid, or 38 cents per dollar) and state and local pension contributions ($72 per $1000 of federal aid, or 7 cents per dollar). General government expenditures reflect spending on the non-healthcare, non-educational functions of government, and include expenditures on administrative services and various departments. This category accounted for nearly half of the $770 spending increase we document while accounting for less than 10% of total expenditures. We do not find any statistically significant evidence that aid increased expenditures in any one of the functional categories of healthcare, education, or infrastructure, nor are there statistically significant increases in these aggregated functional categories, although we cannot reject the hypothesis that they also increased.



