As governments around the world combatted COVID-19, the US witnessed unprecedented levels of funding transfers from the federal government to state and local municipalities. This column examines how the influx of federal aid affected the population’s health. Leveraging the fact that states with excess representation in Congress received more assistance than underrepresented states, the authors show that federal subsidies significantly improved individual health outcomes. At the same time, aid to state and local governments evinced fewer economic benefits, suggesting that federal subsidies are more constructively channelled toward critical health care inputs.
As the COVID-19 pandemic recedes into recent history, data are increasingly available for assessing the effectiveness of the public sector’s response. In contrast with more conventional economic crises, the pandemic’s dramatic blend of health and economic costs is among its most notable features. At the pandemic’s outset, Cutler and Summers (2020) estimated that its cost over two years would amount to roughly $16 trillion in the US, or roughly three quarters of 2020 GDP, with similarly sized costs from declines in economic activity and loss of life. The magnitudes of both forecast and realised costs varied substantially across countries (Yeyati and Fillipani 2021), raising the question of whether government interventions played a meaningful role in mediating the pandemic’s toll.
Governments around the world devoted substantial resources to combatting the COVID-19 pandemic (Deb et al. 2022). In the US, the federal government authorised roughly $6 trillion in expenditure through an array of programmes that reached state and local governments, businesses, and households. While a substantial body of research has studied the effects of these aid packages on economic outcomes, relatively few studies have assessed the extent to which federal relief mitigated the pandemic’s adverse effects on health. In a recent paper (Clemens and Mahajan 2025), we take up this question with an emphasis on the US federal government’s $1 trillion in aid to state and local governments.
A challenge in assessing the causal effect of federal aid on health is that aid may have been targeted, in part, to states with greatest need. If need drives aid, then a naïve comparison of health outcomes in states that received more versus less aid will tend to deliver an unduly pessimistic assessment of aid’s effects. Our solution to this problem makes use of the fact that federal aid was disproportionately generous to states that, due primarily to the structure of the US Senate, have more per capita representation in the US Congress. Past work (Clemens and Veuger 2021) has shown that per capita representation strongly predicted historically large variations in federal aid to state and local governments; an additional representative per million state residents predicts an additional $1,000 in aid per resident.
Crucially for our estimation strategy, per capita representation is uncorrelated with many proxies for states’ pandemic driven needs. We further show that per capita representation is uncorrelated with pre-pandemic mortality levels, pre-pandemic mortality trends, the age distribution of states’ populations, and the baseline prevalence of chronic conditions. These findings support our use of per capita congressional representation as an instrument for estimating the causal effect of federal aid to state and local governments on health outcomes.
Figure 1 shows the reduced form relationship between our instrument and cumulative COVID-19 mortality from 2020 through 2022. The top panel shows the simple bivariate relationship between these variables; the best fit line indicates that an additional representative per million state residents predicts 22 fewer COVID-19 deaths per 100,000 state residents. For the bottom panel, we residualise mortality and representation with respect to population density and a measure of political preferences, both potentially important sources of variation in COVID-19 related health outcomes. Controlling for these factors substantially tightens the relationship between representation and mortality while having a modest impact on the slope; an additional representative predicts 26 fewer deaths per 100,000 state residents, with a standard error of five.
Figure 1 2020–2022 cumulative mortality and congressional representation
Figure 2 presents event study estimates of the relationship between per capita representation and both COVID-19 mortality (top panel) and all-cause mortality (bottom panel) from 2016 through 2022. The bottom panel illustrates that our instrument is uncorrelated with changes in all-cause mortality until the pandemic itself. Our estimated impacts on mortality unfolded steadily between April 2020 and February 2022, stabilising beyond that point. Notably, a substantial share of the gains emerged in the months immediately following the introduction of COVID-19 vaccines, suggesting a possible role for federal aid in hastening or expanding states’ vaccination campaigns. The bottom panel shows, further, that reductions in COVID-19 mortality were not offset by changes in mortality from other causes; if anything, moderate reductions in mortality from other causes complemented the COVID-19 mortality improvements associated with more generous allocations of federal funds. By the end of 2022, we estimate that states with more representatives per capita experienced 26 fewer deaths from COVID-19, and 38 fewer all-cause deaths per 100,000 state residents. Translated into dollar terms, our estimates imply $3.8 million allocated per COVID-19 related death averted, and $2.6 million allocated per death averted from all causes.
Figure 2 Effects of an additional representative on monthly COVID-19 and all-cause mortality
Through what mechanisms might federal aid have driven reductions in mortality? We find clear evidence that greater representation – and its associated federal aid – predict substantially higher rates of COVID-19 testing and moderately higher rates of vaccination. A back of the envelope calculation suggests that heightened rates of vaccination may account for nearly half of the mortality reduction we find in states that received more generous aid allocations. Consistent with early detection, which could improve treatment outcomes and reduce COVID-19’s spread, heightened testing is associated with lower positive test rates. By contrast, we find no evidence that federal aid predicts investments in hospital capacity or larger numbers of enrolees in states’ Medicaid programmes, suggesting that these mechanisms played little if any role.
Figure 3 The reduced form effect of an additional representative on COVID-19 vaccination, testing, hospital capacity, and Medicaid enrolment
Finally, we assess the economic value of the gains in life expectancy associated with federal aid. This analysis has two key inputs beyond our estimated mortality declines. First, we use demographic life tables to infer the average number of years gained from the mortality reductions we estimate. Because mortality reductions were concentrated among the elderly, we apply appropriately differentiated estimates for elderly, middle-aged, and young people; we also conduct sensitivity analyses to allow for the possibility that deaths were averted primarily among individuals with lower remaining life expectancies than among others in their cohort. We next apply estimates of the value of a statistical life year that are consistent with the range considered by US regulatory agencies. The central estimate from this analysis is that the last $331 billion in federal aid, which corresponds with the in-sample variation in federal aid across states, resulted in benefits worth $591 billion through life years saved.
Our findings contribute to the body of work on the effectiveness of public policy in mitigating the pandemic’s severity. Outside of analyses of Operation WARP Speed, previous work on key elements of the US federal response focused primarily on economic outcomes. Past work has shown, for example, that Economic Impact Payments may have spurred consumption in the pandemic’s early months (Chetty et al. 2024), but they primarily buoyed household balance sheets (Parker et al. 2022). Evidence on the effects of expanded unemployment benefits is mixed; some research finds that expanded benefits may have aided consumption smoothing without discouraging job entry (Ganong et al. 2023), while other work finds meaningful negative effects on employment (Holzer et al. 2023). Other research has found that the Paycheck Protection Program preserved employment and business continuity, though at substantial cost per job-year saved (Hubbard and Strain 2020, Autor et al. 2022, Splinter et al. 2025).
Previous work on the effects of federal aid to state and local governments, which is the focus of our study, has considered a range of economic and political outcomes. An earlier paper found evidence, complementary to ours, that federal aid had substantial effects on rates of COVID-19 testing (Clemens et al. 2023). Other work finds that this aid improved electoral outcomes for incumbent governors and senators (Clemens et al. 2024). Estimates of federal aid’s effects on employment and aggregate output, however, imply modest if any impact (Clemens at al. 2022b). From a cost-benefit perspective, this prior research finds no evidence that federal aid to state and local governments had substantial economic benefits (Clemens at al. 2022a). Our recent paper thus brings an additional layer of nuance to this line of research; cost-benefit analyses of federal fiscal aid may hinge on its health impacts.
What might these findings imply for the design of federal responses to future pandemics? Taken together, the relevant research finds the effects of federal aid for state and local governments to be substantial with respect to health outcomes and modest with respect to economic outcomes. The effects we estimate for health outcomes are sufficiently large to yield a positive cost-benefit analysis of this line of spending. At the same time, the additional vaccination and testing that appear to drive these gains can account for less than ten cents of every dollar spent. In light of federal aid’s modest economic effects, a key question is whether the federal government could have produced similar health gains at more modest expense by constraining its aid to the financing of critical health care inputs.
Source: cepr.org
