Previous articleNext article FreeSummaries of ArticlesSummaries of ArticlesPDFPDF PLUSFull Text Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinked InRedditEmailQR Code SectionsMoreThe Responsiveness of Medicaid Spending to the Federal SubsidyM. Kate Bundorf and Daniel P. KesslerAlthough economic theory suggests that the federal government can influence state spending through subsidies to state programs, no recent work has quantified the magnitude of this effect for Medicaid. Medicaid, the largest health insurance program for low-income people in the United States, is financed jointly by the federal and states’ governments according to a statutory formula that effectively subsidizes states’ spending on the program. Because Medicaid is a substantial share of state and federal budgets, identifying the consequences of Medicaid’s financing system on its spending is an important unresolved economic and policy question.We find that states’ spending per enrollee on Medicaid is responsive to the magnitude of the federal subsidy — the Federal Medical Assistance Percentage, or FMAP. We evaluate the impact of an increase in the FMAP (the “enhanced FMAP”) created by the Affordable Care Act (ACA) on the spending of enrollees who were Medicaid eligible before the passage of the ACA (“nonexpansion enrollees”). We find that the enhanced FMAP increased Medicaid spending per nonexpansion enrollee, holding other factors constant.Specifically, we estimate the elasticity of Medicaid spending per nonexpansion enrollee with respect to the after-FMAP price of Medicaid spending faced by the states — that is, the effect on a state’s spending of a 1 percent decrease in the portion of spending that the state bears, after taking account of the increased federal subsidy. According to our model, a 1 percent decrease in the after-FMAP price leads to an increase in spending of 0.46–0.58 percent.Our results are consistent with a long literature in public finance documenting the expenditure-increasing effects of open-ended matching grants. In fact, the confidence intervals around our estimates of the elasticity of Medicaid spending with respect to the after-FMAP price include the elasticity of spending on other federally subsidized state programs with respect to the after-subsidy price estimated in previous research.Our study has an important limitation: the validity of our estimates depends on our assumption that the ACA affected nonexpansion enrollee spending only through its effect on the FMAP, conditional on variables we include in our model. We cannot rule out the possibility that aspects of the ACA other than the increased FMAP — such as the expansion in enrollment — increased spending on nonexpansion enrollees. In addition, we cannot rule out the possibility that our estimates will not generalize to future changes in the FMAP that could occur in different economic environments.The Implications of Uncertain Economic Paths for Revenue ProjectionsLeonard Burman, Benjamin Page, and David WeinerMembers of Congress rely on the budget forecasts of the Congressional Budget Office (CBO) when planning changes to tax and spending policy. Forecasts of the economy and the deficit are inherently uncertain. CBO has analyzed its budget projection errors and determined that revenue forecast error is the largest source of error in projected deficits with uncertainty about the future path of the economy explaining much of the error in revenue projections. A better understanding of the magnitude of the errors in revenue forecasts and any potential bias in those forecasts could help policy makers put proposed changes to spending and revenues in proper context.Previous research on the degree of uncertainty in revenue forecasts has relied on the historical forecasts compared with actual revenues as a measure of forecast error. However, legislation enacted after forecasts have been made complicates the retrospective analysis of the underlying uncertainty, and such analysis may not produce the best estimate of the future uncertainty of revenue projections. We take an alternative approach to measuring the uncertainty. We first produce multiple projections for the future path of the economy based on historical economic growth. We then take advantage of cloud-based computing resources to simulate numerous future revenue paths for the two largest sources of federal revenues: individual income taxes and payroll taxes.In addition to measuring the variation in income and payroll taxes for each scenario, we also test whether budget forecasts that rely on a single economic forecast result in a biased projection of revenues. While the economic path reflects the average expected future path for the economy, revenues based on that single path may not reflect the average of revenues under multiple possible paths. In particular, the structure of the individual income tax should lead to a pessimistic revenue forecast from a single economic path as compared with the average of estimates for multiple paths for the economy.Like previous researchers, we find that uncertainty about the economy leads to considerable uncertainty about future revenues, and that the uncertainty grows with the length of the forecast. We find that forecasts of income and payroll tax revenues based on a single economic path are overly optimistic, but the degree of bias is small. When measured at the furthest forecast horizon, the bias was about 1 percent for the individual income tax and less than 0.3 percent for the payroll tax. We developed a simple formula to estimate the bias based on actual variation in gross domestic product (GDP) and the responsiveness of the tax sources to changes in GDP. The estimate of the bias from the formula was very similar to that from the simulations.The Revenue Productivity of India’s Subnational VATAstha Sen and Sally WallaceOver the last 60-plus years, a national value-added tax (VAT) has been embraced as an economically efficient revenue source that has the potential to raise substantial levels of revenue. Most of the VAT adoptions have been at the central level of government. At the subnational level, the VAT is uncommon, and its impacts are not well known. A subnational VAT has been implemented in just three countries — Canada, Brazil, and India. Canada has one of the most credible tax administrations in the world, so the performance of its subnational VAT is unlikely to generalize to developing and less developed countries, which are often hampered by weak tax administrations and low levels of revenue. Identification of robust sources of revenue is important for subnational governments to provide public goods and services that fall under their jurisdiction. India’s subnational VAT reform in the mid-2000s provides an opportunity to evaluate the revenue productivity of an important policy reform.India’s subnational VAT reform aimed to simplify a complex sales tax system and increase states’ revenue, and it was implemented over the period of 2003–2008. We use the staggered implementation of this important reform to gain insights into the revenue performance of a subnational VAT in India. An understanding of the impact of India’s VAT reform could be useful to other developing countries grappling with limited tax base and instruments. Using state-level data for 29 states, we empirically estimate the impact of the implementation of the state sales tax revenue, controlling for other factors that could affect revenue productivity. We observe that the VAT, on average, increases the state government’s sales tax revenue by 13 percent in the more developed (nonspecial) states of India. Moreover, our empirical evidence shows that the impact of the VAT reform grows over time in the more developed states of India, implying a stronger impact of the tax policy reform in the long run.Real-time Forecasts of State and Local Government Budgets with an Application to the COVID-19 PandemicEric Ghysels, Fotis Grigoris, and Nazire ÖzkanState and local policy makers are responsible for the provision of a variety of goods and services, ranging from education and social welfare programs to public safety, transport, and the retirement benefits for government employees. In 2019 alone, the cost of providing these and other services almost exceeded $4 trillion. While state and local governments are responsible for providing their constituents with a vast quantity and variety of amenities and services, their spending is constrained by the fact that most subnational governments in the United States face balanced budget requirements. As such, local policy makers seeking to avoid unexpected tax increases and spending cuts need to ensure that annual expenditures remain below anticipated revenues.With these constraints in mind, more accurate and timely forecasts of fiscal outcomes (i.e., revenues and expenditures) would undoubtedly help policy makers in at least three ways. First, policy makers could better anticipate how evolving economic conditions will influence their state’s budget. Second, policy makers could prepare for the effects of deteriorating economic conditions to ensure their state’s balanced budget requirement is not violated in times of a fiscal crisis. Finally, policy makers could avoid the need for more drastic actions when fiscal shocks materialize. Unfortunately, however, most traditional fiscal forecasting models fail to provide policy makers with timely and accurate forecasts of their government’s fiscal outcomes, as these models rely exclusively on low-frequency (e.g., annual) data. Consequently, these traditional models can only provide policy makers with updated forecasts of their government’s fiscal health once a year.In this paper, we propose an easy-to-implement and data-driven forecasting methodology that allows policy makers to obtain accurate and timely forecasts of their government’s revenues and expenditures. The model we propose — the autoregressive distributed lag mixed-data sampling, or ADL-MIDAS, model — leverages mixed-frequency (e.g., monthly, quarterly, and annual) data to estimate each state’s revenues and expenditures in a timelier manner than estimates produced by traditional forecasting models. In fact, we show that the ADL-MIDAS model produces forecasts of state-level fiscal outcomes that are up to 38 percent more accurate than the same forecasts from traditional fiscal forecasting models. While the forecast accuracy of the ADL-MIDAS model varies somewhat across states, we show that the ADL-MIDAS model delivers more accurate forecasts in states with a greater reliance on corporate income and property taxes.Finally, we show how policy makers can use the model we propose to conduct scenario analyses in the face of unprecedented economic conditions, such as those instigated by the COVID-19 pandemic. We use the economic effects of the coronavirus pandemic as our laboratory to illustrate this process, and we note how the ADL-MIDAS model allows policy makers to update their forecasts of their government’s revenues and expenditures within a given fiscal year, a feat that is not possible with traditional low-frequency fiscal forecasting models. Previous articleNext article DetailsFiguresReferencesCited by National Tax Journal Volume 75, Number 4December 2022 Published for: The National Tax Association Article DOIhttps://doi.org/10.1086/723651 Views: 148Total views on this site © 2022 National Tax Association. All rights reserved.PDF download Crossref reports no articles citing this article.