In the highly partisan climate of politics today, a refreshing conversation is happening across party and sector lines. Results for America recently launched the “Moneyball for Government” campaign to encourage the US government to allocate funding where evidence shows it can have the most impact on improving the lives of families, children, and communities. Results for America issues “scorecards” that grade federal agencies based on the extent to which they encourage groups they fund to measure results—and whether the agency uses that data to determine what to fund. The purpose of the scorecard is to encourage competition among government agencies to incorporate evidence of results into funding decisions.

Research by America Achieves, according to Managing Director Michele Jolin, shows that remarkably little federal funding flows to evidence-based programs—an estimated $593 million out of a trillion-dollar federal budget. While I am a champion of this new agenda, because I believe we must use government funds to achieve the greatest possible impact, we must remain mindful of the potential pitfalls that could stifle innovation and progress if we do not implement it well.

I recently had a chance to represent REDF, which invests capital and expertise in nonprofit businesses that create jobs for people who face barriers to work, at the Results for America Capitol Hill briefing for Social Innovation Fund (SIF) grantees. There, I helped make the case that the government should direct mainstream public funds toward sustaining successful evidence-based programs launched by the SIF and other innovation funds. I and other speakers highlighted the current state of play, and shared both opportunities for and potential challenges of an evidence-based approach.

Gene Sperling, two-time director of the National Economic Council, shared two examples that make a nuanced case for the power of data to drive results:

  • First, President Obama and his economic advisors convinced Fortune 500 CEOs that their companies should stop discriminating against people who were unemployed for long periods of time, citing studies of individuals with the same demographics and work experience. Those who were unemployed for a year or more had to send out an average of 15 times more resumes before finding a job than those who were unemployed for a month.
  • Second, only 8 percent of low-income, high-achieving high school students applied to institutions that matched their abilities. Research tested and found that it was possible to significantly increase this percentage by sending customized packages to these young people, along with texts encouraging them to complete enrollment—at a cost of just $6 per person. Funding from universities, philanthropy, and business is now flowing to this approach.

Yet pursuing more evidence-based programs is a complex exercise. According to Sperling and other commentators, their experiences and observations have led them to conclude that the guiding principles must include:

  • No double standards—we do not turn away from lung cancer treatment and research because we have not identified a cure. The same should be true of programs for low-income people.
  • While governments should have the guts to terminate programs that do not work, Sperling, who worked in both the Clinton and Obama Administrations, urged public officials not to rely on a single study to “flip the light switch,” as sometimes new research contradicts earlier studies. Researchers agree that multiple studies in different locales under different circumstances are required to verify the results of social programs.
  • Adopt elements of programs that work, even if programs as a whole were not completely successful.
  • Set clear desired outcomes to determine what to measure. Ron Klain, former Chief of Staff to Vice Presidents Gore and Biden, shared that the American Recovery and Reinvestment Act resulted in many roads laid more cheaply than planned but employment objectives fell short. In this case, the outcome of cost effective infrastructure development undermined the overriding objective of the ARRA, which was to create jobs.
  • Be careful not to punish “early adopters” by prematurely cutting off funding of those that assess their programs but develop evidence that the program does not achieve the anticipated results, or else others will not step forward.
  • Finally, given extreme partisanship, Jim Nussle, a former Republican Congressman and director of the Office of Management and Budget said that working on a local level might be a better place to start.

REDF’s founder George R. Roberts built results measurement into REDF’s DNA so that the organization would understand whether and how its work creates value, and could adjust as necessary to create as much impact as possible. I am a big believer in the transformative power of the “what works” agenda, as long as we are not using inadequate or poor-quality data manipulatively, or to drive hasty decisions. We must remain clear on our objectives, offer sufficient resources to measure with rigor, and enable practitioners on the ground to use data to improve and change course as necessary.

Sperling aptly quoted Franklin D. Roosevelt: “It is common sense to take a method and try it: If it fails, admit it frankly and try another. But above all, try something!”