When Public Spending on Social Policies is Fiscally Responsible
Sakshi Parihar 
• Bookmarks: 202
Governments enact federal programs with the hope that they will improve the well-being of their citizens. But how do we know which programs have the highest long-term returns on investment? It is difficult to compare policies from different domains because of a lack of common factors. In a recent paper, Harvard economists Nathaniel Hendren and Ben Sprung-Keyser attempted to solve this problem by evaluating the returns of 133 policy changes made in the United States over the last 50 years using a single metric: the marginal value of public funds. Using this measure, they found that policies affecting the health and education of children in low-income families had the highest long-term returns on investment.
Simply, the MVPF divides the amount of welfare accrued by beneficiaries (measured by their willingness to pay) by the cost to the government to enact the policy or program. When the government recoups its investment through additional future taxes collected from beneficiaries and reduced transfers, the policy is said to have an MVPF of one. If spending on the policy causes individuals to work less and future tax revenue falls, then the policy has a negative MVPF. If the policy causes large, positive ripple effects and revenue generated is in excess of initial investments (with accounting for the changing value of money over time), then the policy has an infinite MVPF with a net cost of zero. Under this framework, the researchers compared the per dollar benefits of a variety of social programs relating to social insurance, education, taxes, and public transfers.
The researchers found that when it comes to low-income children, federal programs have high returns on investment. Programs like Medicaid expansion, childhood education spending, and expenditures on college earned the government an average $1.78 for every dollar spent. Human capital investments did not diminish as children aged: high MVPFs on education spending were found for both young children and teenagers. A particularly striking example was the Perry Pre-School Project — a preschool education program targeting children from disadvantaged backgrounds — which had an MVPF of 44. On the other hand, programs targeting young adults, such as college subsidies and job training programs, had relatively lower MVPFs.
The MVPFs on adult programs were usually significantly lower than programs for low-income children, and sometimes had negative values. The exception was the reduction in the top marginal tax rate, which had an MVPF ranging from 1.16 to infinity as it had spillover effects on children. There was a lot of variability observed within adults. For example, tax reform in 1993 provided $1.85 per dollar spent to individuals in the top marginal income tax bracket and only $1.12 to lower-income individuals claiming the Earned Income Tax Credit. For both children and adults, the authors’ results were sharpened by adjusting for publication bias that favors studies finding positive effects for children and negative effects for adults.
While the MVPF is a useful indicator to evaluate policies across domains on equal footing, it should not be misconstrued as a policy recommendation in itself. For example, even policies with high MVPFs could have negative spillovers, and the costs have to be weighed against the benefits. For example, food stamps issued under the Supplemental Nutrition Assistance Program reduced willingness to work among adults but increased the lifespan of children. Further, Medicaid increases crowded out private insurance coverage but reduced chronic illness amongst kids.
The same policy also impacts different members of the family in different ways. For example, the Moving to Opportunity experiment that provided vouchers to families to move to lower poverty neighborhoods generated negative MVPFs for adults, null or slightly negative MVPFs for teenagers, and infinite MVPFs for children below the age of 12. The MVPF can tell us how policies need to be adjusted to better target different subsets of individuals.
The authors also note that efficiency may not always be the desired policy goal. For example, transfers to disabled children and their families have very low MVPFs. Does this mean that such investments should not be made? Just because a tax reduction in the highest tax bracket has a larger MVPF than cash transfers to disabled children, it doesn’t necessarily mean governments should adopt tax reductions instead of cash transfers if doing so breeds inequality.
Short-term thinking is the greatest enemy of a good government. The government should plan beyond the immediate return on investment and have the patience to invest for the long term. Increases in future earnings and additional tax revenue can be a good gauge of a policy’s success. The MVPF offers a useful metric to evaluate different policies intended for the same set of beneficiaries and choose the policy with maximum social impact given budget constraints. The MVPF can boost the toolkit of governments across the world to effectively use evidence to analyze policy.
Hendren, Nathaniel and Ben Sprung-Keyser. 2020. “A Unified Welfare Analysis of Government Policies”. Quarterly Journal of Economics 135, no. 3: 1209-1318. https://doi.org/10.1093/qje/qjaa006.
