CAREER: Market Imperfections and the Welfare Measurement of Government Intervention

Project Outcomes Statement

A primary outcome for this project was the development and implementation of new tools for measuring the welfare impact of government policies. There is a growing literature estimating the causal effects of many government policy changes, as highlighted by the so-called "credibility revolution". But, subsequent welfare analyses of these policies is often ad-hoc and done using different methods across different domains. For example, papers evaluating a tax policy might measure its marginal deadweight loss, those evaluating an education policy might conduct a form of cost-benefit analysis; those studying a healthcare policy might construct a cost per life saved; or those evaluating a labor market policy might measure the cost per job saved. The use of different measures makes it difficult to compare the welfare impacts of policies across different domains (e.g. spending more on education versus healthcare).

This project developed a unified approach, known as the Marginal Value of Public Funds (MVPF), to measuring the welfare impact of government policy changes. The MVPF is the ratio of the benefits provided to policy beneficiaries relative to the net cost of the policy to the government. The key to calculating the MVPF is to accurately measure the net cost to the government by not only including the upfront cost of the policy but also measuring the impact of the policy on future taxpayer behavior. For example, if an education policy increases earnings in adulthood, this also raises tax revenue and lowers the effective cost to the government over the long run.

We measured the MVPF for 133 historical policy changes over the past half-century in the United States, focusing on policies in social insurance, education and job training, taxes and cash transfers, and in-kind transfers. The figure below illustrates the results by plotting the estimated MVPFs as a function of the age of the beneficiaries on the horizontal axis. Historically, direct investments in children -- especially low-income children's health and education -- have historically had the highest MVPFs, on average exceeding 5. Many such policies have paid for themselves as governments recouped the cost of their initial expenditures through additional taxes collected and reduced transfers. We find large MVPFs for education and health policies amongst children of all ages, rather than observing diminishing marginal returns throughout childhood. For adults, the MVPFs are smaller (generally between 0.5 and 2), but can be larger if a policy targeted to adults has spillovers onto children.

This project led to the launch of Policy Impacts (www.policyimpacts.org) that is now continuing this line of work. Most notably, the Policy Impacts Library hosts a comprehensive list of MVPF estimates from existing literature, enabling researchers and policymakers to view up to date information that translates the high quality causal effects estimated in existing literature into their implied MVPFs that enable comparisons of policies both across and within different policy areas.