How to Translate Research Findings into Policy

Recommendations for successfully communicating with policymakers
October 20th 2022

This speech was originally presented at the NBER Summer Institute panel on Challenges and Opportunities in Translating Innovation Research into Policy in July, 2022.


Many people, like myself, become economists not only because we’re intellectually fascinated by the subject and its tools, but also because we believe good economic analysis can make the world a better place. But when economists try to turn their work into practical advice for policymakers, we’re often stymied and frustrated. Fortunately, speaking so that policymakers will listen isn’t an art form: there are specific things you can do to improve your odds of success.

I’ve spent most of my career at institutions that receive research findings and apply them to make policy decisions.  I gave five recommendations to a group of economists attending a summer conference at the National Bureau of Economic Research, and I replicate (in slightly amended form) those recommendations below.

1. Most policymakers care a lot about who wins and who loses from a policy change, and care much less about abstract notions of efficiency

Because most policymakers are not trained in economics, they don’t understand what efficiency means in economists’ sense of the word. However, even policymakers who do understand our view of efficiency recognize that compensating side payments rarely happen. To them, the idea that “everyone could be made better off” is a nice thought, but not operationally very useful.

The implication of policymakers’ focus on distributional effects is that you should try to translate your research findings into the effects of policy changes on income for different groups of people. It won’t always be possible to do in any meaningful way, but when it is possible, do it. Also explain why incomes change, because policymakers and their audiences will not take a result on faith without hearing a plausible explanation.

Some useful heuristics: Qualitative descriptions of income changes are much better than no descriptions at all. Rough numerical approximations are much better than qualitative descriptions. Median incomes are generally more interesting than mean incomes. Identifying any types of households that lose income from a policy change is important — as well as explaining what else they might gain besides income, and how they might be at least partially buffered from a policy’s effect.

2. Focus on key government agencies or non-governmental groups that influence policymakers, and think about those agencies’ or groups’ interests

I’m not against direct lobbying of elected officials and their personal staff. That kind of lobbying can work, because elected officials tend to be people-oriented and responsive to direct interaction. But more often, in my experience, elected officials get their information through dedicated agencies or outside organizations, and you can significantly increase your impact if you draw those agencies and organizations to your perspectives. How do you do that?

To anyone in the marketing business, “putting yourself in your customers’ shoes” does not come as surprising advice. The key here is to understand the agencies’ or groups’ own mandates and approaches to work. Most agencies need to accomplish specific goals, not just make the world better in a broad sense. You want to address those goals. Most agencies have specific ways of working that developed as responses to external forces. Even if those ways of working seem wrong to you, taking those external forces seriously may be a prerequisite for moving the agency you have in mind.

An obvious example here is CBO’s default approach of estimating budgetary effects over ten years only, without allowing total work effort or other aspects of GDP to change. That default approach can lead to misleading estimates, and I worked hard with my colleagues at CBO to move beyond that default; I hope that, over time, CBO can keep moving. But that default approach is not an accident — it’s the result of certain forces — and attempts to change CBO’s approach need to take those forces seriously.

3. Communicate in non-technical ways that are not standard in the academy

Stay away from jargon and technical detail as much as possible. Don’t dumb down your message, because you’ll generally be talking to smart people. But these people are usually not trained as economists, and even the ones who have that training do not have the time or incentive to keep up with the latest research. That’s why they need you. Don’t throw around the latest shorthand labels, or make your starting point the last paper in the chain before yours, as you might in a seminar. Good ideas in economics can almost always be explained in plain English if you take the time to think them through.

Also, communicate in formats that are shorter than papers and less about text. Your audience has limited time and maybe short attention spans, especially because many influential staff members are young. When I was at CBO, a survey suggested that Capitol Hill staffers went to our website most often on their phones while sitting in meetings — so although we often thought of our main products as printed reports or estimates on 8½ x11 pages, our readers were looking at our work on little screens. That changes the optimal presentation. 

Roughly a decade ago, we also started to produce infographics and chartbooks along with standard reports. One of our first infographics was about deforestation and carbon emissions, and the author of the report initially worried about turning her sophisticated analysis “into pictures.” But later, someone watching C-SPAN spotted the Senator who requested the analysis walking around on the Senate floor showing other Senators the infographic. That’s success in communications.

It is important, but quite difficult, to be honest about uncertainty. In my CBO experience, presenting confidence intervals when possible made me and my colleagues feel better, but didn’t do much for our audience; they either found the extra dimensionality of the information confusing, or found a way to distort the extra information for their advantage.

4. When translating research findings into policy, be patient and persistent

Policymakers and their staff are not sitting around on free afternoons waiting for economists to turn up and tell them how they should be doing their jobs differently. They are very busy — facing pressing deadlines for passing bills, issuing regulations, signing contracts, responding to legal challenges, and much more. They are also dealing with multiple constituencies — voters, donors, interest groups, other governments, supposed experts, true experts, and more. 

By putting yourself in their shoes and communicating in non-technical ways, you can use their time effectively and gain their trust. Don’t expect them to turn on a dime because you arrive one day with a new and better answer to a policy question. You need to build their trust in you and your judgment, give them time to check what you say against their other sources of information, and figure out how they can make the best use of what you tell them.

Of course, even if you can persuade a public official or staffer of your view, those converts need to deal with many other officials and staffers to get something done. The key implication for economists is that we need to lay the groundwork of understanding as broadly as we can so that public officials are ready when the political stars align. An example from my days at CBO: The Affordable Care Act was possible only because two critical elements came together — the political alignment and focus in 2009, and the research, modeling, and teaching done by health economists over the preceding two decades.

5. Play the long game by encouraging students to take policy jobs in government or government-adjacent organizations

Our country desperately needs more people who understand social science to be working directly with policymakers. Graduate and undergraduate students can find rewarding opportunities in the federal government, in state and local governments, and in organizations like the Center for Global Development, the Federation of American Scientists, and many others. Some of these opportunities are in so-called “permanent” jobs; others are in jobs that are explicitly time-limited.

Of course, I don’t recommend that a smart student take just any job in a government or a policy organization. Students should explore carefully and try to get a good feel for a place before accepting a position, but there are lots of great opportunities.

Moreover, you could take a job in a government or policy organization, on a long-term or short-term basis. For example, when I was CBO director, we were extremely lucky to persuade Debbie Lucas to take leave from MIT to build up CBO’s capacity for financial analysis, and we were extremely lucky to persuade Jeff Kling to move to CBO and oversee a wide array of economic analysis.

None of these recommendations are silver bullets. Policy is tricky. Policymakers’ attention is hard to keep, and much of a given political debate is outside of your control. But by paying close attention to how policymakers operate and what they need, you can increase your odds of communicating to them successfully.