What this covers
Russ Roberts interviews Steven Levitt about the evolution of economics since Freakonomics and the tension between academic incentives and real-world relevance. Levitt reflects on how the 2005 book became a cultural phenomenon, what it got right and wrong, and how his thinking has shifted over two decades. The conversation centers on a single core complaint: academic economics is structured to reward answering small questions with perfect rigor rather than tackling large questions with partial understanding, which has hollowed out the profession's ability to matter beyond the seminar room.
Levitt traces this problem through several domains. Academia systematically underinvests in broad thinking—explaining why China succeeded, for instance—because there is no academic payoff to it. Natural experiments and causal-inference methods, once scarce and valuable, have become commodified recipes, so economists have migrated to more complicated structural models, making the field less policy-relevant. He contrasts this with the communication skill that made Milton Friedman effective: packaging ideas simply enough that ordinary people could explain them onward. On Incentives Matter, Levitt argues that getting prices and incentives right is often far more powerful than behavioral nudges, though he credits Richard Thaler with showing that changing defaults can yield real results. He discusses replication failures, perverse incentive systems in medicine and real estate, and GPS monitoring as a policy lever that could simultaneously reduce both violent crime and incarceration. The conversation also touches on climate change, education reform, and the gulf between what economists know and what they can explain to others—a barrier Levitt attributes to the Curse of Knowledge rather than complexity.
Levitt argues that economics is most valuable when applied as a practical, data-driven toolkit aimed at real-world impact, but that academic incentives have pushed the profession toward answering small questions perfectly rather than big questions usefully, making it less relevant.
- Academic economics rewards answering little questions 100% rather than big questions halfway
- Getting incentives and prices right is often more powerful than behavioral 'tricks'
- Real-world impact requires moving ideas to proof-of-concept, not just publishing them
Incentives broadly construed—financial, moral, social—drive human behavior; defaults and design often work better than education or persuasion.
- The word 'incentives' should be understood broadly to cover three categories—financial, moral, and social incentives—which together essentially cover the universe of things that lead attentive people to do what they do.
“those three things--financial, moral, and social incentives--essentially cover more or less the universe of things that lead people who are paying attention, to do the things they do.”
- Because people are busy and inattentive rather than stupid, it is often easier to get desired outcomes by quietly changing defaults or design ('tricking' people) than by educating them or changing incentives—an insight Levitt credits to Thaler and now places at the top of his toolkit, while noting it raises ethical concerns.
“a lot of times, it's just easier to trick people into doing what you want them to do than to actually either, like, educate them, use information, or to change incentives.”
- When the University changed incentives so the Economics Department was paid on the margin for additional students taking courses, Levitt and List responded by creating a popular non-majors economics course that drew about 500 students—demonstrating that incentive structures shape even academic behavior.
“they changed the incentives at the University so that, to the Department, there was a benefit from having more students take our courses. We started getting paid on the margin.”
- People are too stuck in the status quo and fear change because its costs are largely up front and many of us are hyperbolic discounters; Levitt found via a randomized coin-flip experiment that if you've been unable to pull the trigger on a major life change for a month, the right answer is almost surely to make the change.
“if you really are stuck... and month at a time and you just can't pull the trigger, almost for sure the right answer is, you should pull the trigger.”
Academic economics rewards answering small questions completely rather than big questions halfway, producing systematic underinvestment in broad thinking.
- Academia does not reward answering big questions halfway; it rewards answering little questions 100% of the way, which produces a systematic under-investment in broad thinking (like why China succeeded) because there is no academic return to it.
“academics does not reward answering big questions halfway. Academics rewards answering little questions 100% of the way.”
- When natural-experiment and causal-inference techniques were new 30 years ago, doing them was a scarce, hard talent that carried academic rewards; now there is a cookie-cutter recipe for extracting causal results from a law change, so the skill is no longer scarce and academics have moved on to complicated structural models, making economics less policy-relevant.
“Now that talent is not scarce. Now there's a cookie cutter recipe for taking a data, looking at a law change or some other natural experiment, and getting a result.”
- Levitt observed repeated 'defining moments' in which University of Chicago economists, brilliant at applying Chicago economics in their papers, made real-life departmental decisions (classroom allocation, tenure timing, allocating trivial sums) completely at odds with the rational, incentive-based reasoning they espouse professionally.
“everything I believe about Chicago economics is turned on its head in the actual practice of life by Chicago economists--who are great in their papers at acting like Chicago economists, but awful in real life.”
Behavioral economics produces large lab effects but modest real-world results; getting incentives and prices right is far more powerful.
- Behavioral Economics produces large behavioral effects in laboratory experiments, but in the real world most nudges (with the exception of defaults like automatic retirement enrollment) yield only modest results, while getting incentives and prices right is often far more powerful.
“In laboratory experiments, you often can generate really big impacts on behavior... In the real world... most of the stuff that people have tried to exploit... haven't yielded very much in terms of empirical results.”
- Asking people to 'do the right thing' on climate change when the benefits accrue to others never really works; therefore any real solution to climate change will likely be a technological one—a position Levitt says the world has increasingly moved toward over the past decade.
“as economists, we all know that just asking people to do the right thing, when the benefits accrue to other people, it never really works. And, that, if we're going to have a solution, it's probably going to be a technological solution.”
- Much chemotherapy is relatively ineffective with enormous side effects and high expense, yet a perverse incentive system in which doctors receive part of the revenue from selling the treatments sustains its use; if patients understood this system better they would be less open to it.
“the crazy incentive system whereby doctors get part of the revenue that is generated from the sale of those.”
Problems persist because they are genuinely hard or because unpopular solutions require institutions willing to not be liked.
- Big social problems remain unsolved either because they are genuinely too hard, or because the answers are not that hard but are so unpopular that no one with common sense will pursue them—so a center freed from caring about being liked can carve out a different space by tackling those unpopular-but-effective solutions.
“The problems that haven't been solved, the big social problems, haven't been solved either because they're too hard or because they're not that hard, but the answers are really unpopular--so nobody with any common sense would actually go out and try to do it.”
- Ideas alone are not enough to change policy; what is missing is the entrepreneurial intermediate stage of taking an idea to a working proof of concept, which is what RISC aims to do—demonstrating in Cook County that monitoring works, saves money, and is replicable via open-source software so other jurisdictions can adopt it turnkey.
“ideas are not enough to win the day.”
- Placing GPS monitoring on people under criminal-justice jurisdiction creates a strong deterrent—monitored people commit virtually no crime because locations can be cross-referenced with crime databases—so offering voluntary early release in exchange for monitoring could, by back-of-envelope estimates, reduce violent crime by 25% while simultaneously reducing the prison population by 25%, benefiting government, society, and offenders.
“if you put people on these bracelets that have GPS, they do virtually no crime.”
Success requires tolerating off-path steps, pursuing passions with willingness to quit, and treating others well because kindness simplifies life.
- Contrary to the fear that stepping off the expected path is disastrous, life is long rather than short, and observing many wildly successful people suggests the off-path steps are often key to their eventual success—so one should be more tolerant of not being in a hurry.
“life is long, not short. I was always in a hurry... it's often the steps off the path, I think, that have been the key to their eventual success”
- It is good to pursue what you love (like being a rock star) knowing you will probably fail, but the real self-harm comes from never stopping even after the writing has long been on the wall—so the key is to combine pursuing passions with the willingness to quit when it's no longer working.
“Where you really hurt yourself is when you start being a rock star and then you never stop trying to be a rock star, even though the writing was on the wall a long time before that.”
- Adopting a rule of trying to be nice, giving people the benefit of the doubt, and offering compliments or help when one doesn't have to is valuable not only for others but because it makes one's own life easier—being a jerk is hard work that takes a lot of time and effort.
“being a jerk is hard work. It takes a lot of time and effort.”
- Levitt found interviewing surprisingly harder than being interviewed, carrying a real sense of obligation and weight, and discovered that the deep preparation for interviews puts him into a flow state—an intense immersion he rarely experiences in his other work.
“the preparation for my interviews is very much a flow state for me. I become extremely focused in a way that I haven't”
- For a relatively rare medical condition, an individual can study hard for a day and come to know more about that specific malady than their general-practitioner doctor, illustrating how the Internet has dramatically increased the availability and usefulness of information to individuals.
“If you are diagnosed with some relatively rare thing, you can know more about that malady than your general practitioner-doctor, in about a day.”
- Levitt sees himself as not a conventional economist but as data-driven and interested in society and culture, using economic tools (especially incentives and externalities) to study questions at the edges of what counts as economics.
“I've just never been that interested in economic systems, as much as I have been in using economic tools to study questions that are further afield”
- The educational system fails on two opposite dimensions by an archaic curriculum: it omits practical mental-health tools (trauma, mindfulness, conflict resolution) and fails to teach genuine data literacy, both of which would better prepare students to function in the real world than what is currently taught.
“I think we make a real mistake in the educational system by not giving mental health topics a place in our curriculum.”
Economics has poor replication standards, tolerates suspected fraud without penalty, and economists struggle to explain their own research clearly.
- Unlike psychology where suspected fraud carries huge penalties, economics tolerates suspected fabrication because the field's problem space is so vast relative to the number of economists that only a handful care about any one result; Levitt and List failed to replicate two outrageous experimental papers and suspected fraud, but editors forced them to remove fraud language and the author paid no price.
“the editors made us take out anything about fraud or malfeasance. Unlike in psychology, where there have been huge prices paid by people who it's been suggested to have been fraudulent.”
- Replication in economics is narrowly defined as whether posted code reproduces the published tables, but starting from raw data most—if not all—papers contain enormous numbers of mistakes; most such errors push results toward zero rather than away from it, and full replication is unlikely to happen for most findings.
“my hunch is that most, if not all papers, have enormous numbers of mistakes in them.”
- Economists are articulate about many topics but uniquely bad at explaining their own research, because they know it too well and resist any simplification as omitting essential nuance—analogous to the realization that newspaper articles about a subject you know are full of errors, then wondering if that's true of everything they write.
“The thing that economists are worst at talking about is their own actual research.”
Freakonomics succeeded through luck and good timing with strong storytelling, while errors came from credible-seeming sources rather than data.
- The success of Freakonomics was attributable mostly to luck and good timing combined with Dubner's journalistic writing ability and the storytelling-friendly nature of Levitt's research, rather than to any reproducible formula.
“Mostly luck, honestly. I think we were in the right time--the right place at the right time.”
- The two things in Freakonomics that turned out to be plain wrong—the Stetson Kennedy KKK infiltration story and the OrangeJello/LemonJello naming anecdote—were both based on credible-seeming sources rather than Levitt's own data analysis, while his research-based findings have stood the test of time.
“the things that I regret now that are wrong are the things that were well-researched journalism, as opposed to actual academic research.”
- After SuperFreakonomics discussed climate change, a group invoking the authority of science (the Union of Concerned Scientists) ran what Levitt characterizes as a propaganda exercise to discredit them, even though the book cited top scientists' work in Nature and Science, illustrating how the reputation of science can be strategically weaponized regardless of truth.
“the strategic use of the reputation of science to destroy things that the people who are arguing want destroyed, regardless of truth, regardless of right and wrong, was used incredibly effectively.”