Short-Term Spending and Long-Term Dynamic Effects

Posted by
Phill Swagel
November 1, 2023

Federal policies related to health care, education, infrastructure, the environment, and research and development often involve short-term expenditures that give rise to economic and budgetary effects far in the future. CBO’s conventional cost estimates project the budgetary effects of proposed legislation over a 10-year period and incorporate the assumption that nominal gross domestic product (GDP) remains unchanged. In dynamic analyses, CBO studies how proposals that would significantly change federal spending and tax policies would affect the overall economy, as well as how such effects would feed back into the federal budget (CBO 2023a).

The agency has been building analytic capacity to consider longer time frames in a dynamic framework for policies that would have long-term effects. That capacity could be used to supplement the analysis in conventional cost estimates and provide additional information about effects that are more than 10 years in the future and that alter nominal GDP.

In a working paper released today, CBO considers the long-term dynamic effects of spending on health care for children when that spending is projected to affect their earnings as adults (CBO 2023b). That paper complements CBO’s other analyses of policies that would incur upfront costs and potentially generate long-term budgetary effects. For example, the agency has assessed two illustrative scenarios that would boost federal funding for a mix of physical infrastructure projects (CBO 2021a). More recently, CBO analyzed the economic and budgetary effects of policies that would impose caps on discretionary budget authority (CBO 2023c). In addition, the agency has been studying the channels through which policies that would address the effects of climate change or help entities adapt to those changes would affect the federal budget (CBO 2021b).

Long-Term Fiscal Effects of Medicaid Spending on Children

CBO’s paper on the effects of Medicaid coverage for children describes a general framework for estimating the long-term fiscal effects of spending on children that affects their earnings as adults—in this case, using Medicaid as an example. Recent studies have estimated the link between enrolling in Medicaid as a child and earnings as an adult (Brown, Kowalski, and Lurie 2020; Goodman-Bacon 2021; Miller and Wherry 2019). Those studies allow CBO to estimate the long-term fiscal effects of that spending.

In the childhood Medicaid paper, CBO considers two illustrative policies: one that would increase children’s enrollment in Medicaid over the next decade, and another that would decrease it. Those illustrative policies would affect slightly different populations, so the results are slightly asymmetric. In that analysis, federal Medicaid spending is projected to be approximately $1,700 per child for each year of enrollment.

Increasing children’s enrollment in Medicaid would reduce the future federal deficit by between roughly $800 and $3,400 per child per year of enrollment, CBO estimates. Reducing enrollment would increase the federal deficit by about $900 to $3,500 per child per year of enrollment.

Those long-term fiscal effects are shown as present values, which express the flow of future income in terms of a single number. The ranges of numbers, in turn, depend on the discount rate used to translate future cash flows into current dollars. The larger numbers reflect the use of Treasury discount rates, which are based on the government’s cost of borrowing and are more useful for understanding the expected effect of policies on federal debt. The smaller numbers reflect the use of fair-value discount rates, which account for market risk and can help policymakers compare policies that involve different risks.

Any policy’s overall effect on the federal budget depends on its method of financing. The main analysis in the childhood Medicaid paper incorporates the assumption that changes made in other areas of the budget would, on net, leave the federal deficit unchanged. The results would differ if an increase in Medicaid spending was instead financed by additional borrowing. In that case, the federal budgetary cost of increasing children’s enrollment in Medicaid would stem from two factors: the greater public borrowing to cover the added spending on Medicaid, and the crowding out of the funding available for private-sector investment as a result of that additional borrowing. In CBO’s estimation, roughly one-quarter to one-half of the costs of that added borrowing would be offset by the long-term fiscal effects of Medicaid spending.

Directions for Future Research

CBO’s estimates of policies’ long-term effects are uncertain, and more evidence could help the agency refine its analyses. Uncertainty can compound over time, so accurately estimating a model’s parameters would be particularly useful in a long-term context.

In CBO’s analysis of Medicaid coverage for children, one area of uncertainty is the extent to which greater coverage would affect their future earnings. Findings from future research could refine estimates of the size of that effect and help CBO determine how to project that effect as trends in health insurance availability and coverage change.

In addition, future research could examine other ways that Medicaid coverage might influence the federal budget—including, for example, through its effects on associated health care costs, expenditures for postsecondary education (such as Pell grants), costs to the justice system, and affected children’s other family members (through a change in the mental or physical health of their parents or a change in the number of hours those parents work). Research about the effects of other programs that benefit children also would enable CBO to use the same framework to estimate the long-term fiscal effects of policy changes in those programs.

More broadly, refinement of parameters that factor into the modeling of policies with upfront costs and long-term effects would be useful to CBO. For those types of analyses, choosing which discount rate to use to estimate the present value of long-term benefits is particularly important. The agency has written extensively about the use of discount rates to measure fiscal effects, especially for credit programs (CBO 2021c). CBO would benefit from more discussion about how to measure the long-term fiscal effects of other programs on a present-value basis.

In general, a better understanding of the channels through which long-term effects occur and how those channels interact would be helpful. In the context of Medicaid spending on children, one approach would be to combine estimates from different strands of literature, such as the effects of Medicaid on school completion and the effects of school completion on earnings. A drawback of that approach is that it can increase uncertainty in estimated effects. Moreover, different channels may have overlapping effects on earnings, creating ambiguity when aggregating outcomes across channels (Athey and others 2019). Research that integrated evidence on multichannel short-term effects with long-term aggregate effects could advance that type of work.

Phillip L. Swagel is CBO’s Director. This blog post includes contributions from CBO staff members Elizabeth Ash, William Carrington (a consultant to CBO), Berna Demiralp, Michael Falkenheim, Rebecca Heller, Grace Hwang, Joseph Kile, Junghoon Lee, and Xiaotong Niu.