Compare, for example, the evolution of modern medicine with the typical approach taken in government. In the 20th century, randomized controlled trials revolutionized medicine, replacing guesswork with a rigorous, scientific method to determine what works and what does not. But in government and philanthropy, all too often decisions about how to allocate scarce resources have continued to be been informed by hunch and anecdote. Replacing hunches with facts has dramatic consequences for the efficacy of government programs, particularly those that deliver services to assist the poorest in our society.
But how do you go about building a system of evidence-based policymaking? In an era of tight budgets and hard choices, here are five concrete steps that state and local policymakers can take to ensure that taxpayer money is being used in the most effective ways possible:
1. Add requirements and support for rigorous evaluation into existing funding streams. When authorizing pilot programs, lawmakers should encourage agencies to roll them out in a way that allows agencies and scholars to compare the effect of the programs on those who receive services from them against a statistically identical population that receives only pre-existing services. Policymakers should seek out technical assistance for agencies to become better producers and consumers of evidence and to create a culture of evaluation across the jurisdiction's executive branch.
2. When allocating scarce resources to oversubscribed programs, consider determining eligibility by lottery rather than first-come-first-served. This is a way for agencies to evaluate the impact of a program on recipients and make the case for further funding if it is demonstrably effective. Lotteries can also be the fairest way to select individuals off a wait list. This approach was used to great effect by the state of Oregon when it expanded Medicaid to previously ineligible applicants by lottery, allowing scholars to observe precisely what effects expanded Medicaid access had on beneficiaries' financial circumstances and physical and mental health.
3. Require that agencies link administrative datasets. Government programs often affect people's lives in ways that are not confined to bureaucratic silos. A housing program can have a profound effect on health, an education program can affect students' job prospects, or a program for substance use disorder treatment can influence the likelihood that patients will get into trouble with the law. But these outcomes can be hard to detect if data is confined to silos within agencies that do not work together. Policymakers should insist that agencies think collaboratively about how to link data so that policymakers and researchers can observe the true effects of programs. South Carolina, for example, has established an integrated data system to make it easier for government and independent evaluators to study the impact of initiatives such as the Nurse-Family Partnership program.
4. Institutionalize best practices by creating independent evaluation offices. These are tasked with identifying opportunities for randomized evaluations and other rigorous research, linking administrative datasets across state agencies to facilitate these evaluations, and applying existing evidence to improve the efficacy of state programs. For example, the Washington state legislature established the Washington State Institute for Public Policy to carry out policy-relevant research and to use this evidence to advise legislators. In a similar model with a focus on behavioral science, Philadelphia and Washington, D.C., have created behavioral "nudge" units, along the lines of the White House Social and Behavioral Science Team, to apply research from behavioral science to improve the efficacy of their programs.
5. Take a page from Congress' book by establishing state-level commissions on evidence-based policymaking. These would be charged with carrying out systematic reviews of existing data and evaluation infrastructure and finding better ways to institutionalize the generation and use of evidence in government policy.
By implementing these kinds of suggestions, state and local government policymakers can build the infrastructure to design and fund programs that work -- a win-win for policymakers of all political persuasions and for the citizens they serve.
