Friday Notes, October 7, 2022
Dear Friends -
On Tuesday, Dave Evans asked on Twitter: “If you were going to give one piece of advice to an early career researcher on how to do research that influences policy, what would it be?” My answer: “Spend time in government.”
And that got me thinking about a few of the many things I learned in my time in and around the U.S. Federal government about doing research that influences policy. Namely:
Fixing. The most influential forms of policy research are not the ones that describe or explain some big social or economic problem. They are the ones that provide a solution to a known and salient problem that affects a defined constituency. Now, that does not in any way negate the value of theory, or evermore in-depth descriptive or analytic work. It’s hugely important, for example, to have a sense of the magnitude of a labor shortage, and the myriad reasons for a spike in people dropping out of the labor market. Those types of studies, though, just feel like a piling on of bad news to the beleaguered bureaucrat or people-pleasing politician. If you want to see a straight line between analytic work and policy change, focus on fixes. This means developing policy innovations, like incentives for retraining; or evaluating existing policies and offering evidence-rich recommendations for improvements.
Anticipating. Government policy is a blunt instrument affecting the futures of lots of people and business with a single clause, or even a single word. This means that researchers who aspire to have their work influence policy have an obligation to think through what would actually happen in the real world if their findings and ideas were taken up. In general, policymakers are not skilled at asking critical questions about whether a particular study result is generalizable; or if there are potential unintended negative consequences — beyond the fiscal impact — of adopting a particular policy path. Researchers themselves have to discipline their enthusiasm for a “fix” by thinking through and analyzing longer-run and second-order effects. For example, what would we guess — or estimate from historical data — would be the effect of student debt forgiveness on college tuition and fees five years hence? Are there additions to the near-term “fixes” that would mitigate future problems? This, too, is part of the mandate of the policy researcher.
Timing. Almost all opportunities for policy research to come alive in the form of legislation, regulation, or other actions are tied to events: the arrival of a new Administration; the occasion of a big international meeting where the President or another representative has to offer up a “deliverable”; the State of the Union address; an election. The preparation time is always absurdly short. This is not because it’s impossible to predict when the events will occur; they are all planned years or decades in advance. It’s because everyone at a decision-making level in government is putting out fires (often dumpster fires) all of the time, and almost anything that isn’t burning gets little attention, and even less brain space. This means that there are predictable moments when researchers, or those who are communicating about research, have the chance to contribute — to fill the vacuum with ideas. During those windows of opportunity, a conversation over coffee with someone near the West Wing can lead to your policy proposal, based on your research, making its way into a speech, a bill, and maybe even a budget. All the rest of the time, it’s probably just another cup of coffee.
Connecting. Along with finding the right moment to present an idea based on research, a savvy policy researcher also needs to find a frame, or connection to a pressing concern, that helps a decision maker see the merits. There is a crude way to do this and a more subtle one. The crude way is to attach an idea to the political boogeyman du jour, even if the connection is remote: “applying my research findings will help us win the war against [China’s unfair trade practices] [White nationalism] [inflation] [epidemics] [the mental health crisis] [climate change].” For a taste of that, just observe the radical swings in the argument about how much to spend on foreign aid, and for what. These jump from humanitarian rationales to anti-terrorism to global health security to anti-China to anti-immigrant, depending on some larger (and often ugly) narrative. The subtle way is to think about the incentives of people whose could advance the application of a new idea, and frame accordingly. For example, lots of people in high-level positions in government, whether elected or appointed, like to think of themselves as “champions” of one thing or another, and as heroes to specific constituencies. Find the person whose status will be elevated by your idea, and make a friend.
Serving. In academia and other parts of the research ecosystem, your name is really, really important. First authorship. Citations. Even names of statistical tests and distributions (here’s looking at you, Kaplan, Meier, and Poisson). Extremely high technical quality and precision are also really, really important, almost no matter how long it takes to achieve them. But when you offer up ideas based on your research findings into the policy arena, you need to be prepared to let attribution and credit go — and you need to respond quickly when questions come up, even if that means working with less deliberation than you’re used to. An effective policy researcher needs to be prepared to pivot from “I have an idea” to “I’m ready to be part of the sausage-making,” all without claiming credit.
So much more than fit in a tweet!
A random assortment of highly recommended videos (two short, one long but worth every minute):
Have a good weekend,
-Ruth