Charles Manski Wins Frontiers of Knowledge Award for Teaching Economics to Doubt Itself

Most economists, Charles Manski will tell you, are in the business of making things sound more definite than they are. Their models produce single numbers - a policy will raise employment by 3.2 percent, or cut poverty by a specific margin - figures that project confidence even when the underlying data cannot support it. Manski has spent five decades pushing back against this tendency, and the BBVA Foundation has now recognized that effort with its eighteenth Frontiers of Knowledge Award in Economics, Finance and Management.

The Northwestern University professor's central contribution is a concept he calls partial identification. Instead of forcing an analysis toward a single point estimate, partial identification acknowledges when the data can only narrow outcomes to a range - and treats that range as the honest answer. The width of the range itself carries information: a narrow band means the evidence is strong; a wide band means it is not, and policymakers should know that before acting.

The Homeless Study That Reframed the Question

The intellectual origin of this work traces back to the late 1980s, when Manski was directing the Institute for Research on Poverty at the University of Wisconsin-Madison. A colleague was conducting a longitudinal study tracking homeless individuals over time. The problem was that roughly 40 percent of the original sample could not be located a year later - precisely because they lacked fixed addresses.

Standard statistical practice at the time assumed these missing subjects were essentially identical to those who could be found. Manski rejected that assumption on logical grounds: people who had vanished from contact were likely experiencing worse conditions than those who remained reachable. Treating the missing data as random would systematically distort the conclusions.

That insight pushed him toward a different framework. Rather than plugging the data gap with an assumption and proceeding to a point estimate, he asked: given what we actually know and what we genuinely do not, what is the full range of conclusions consistent with the evidence? The result was partial identification - now a standard concept in econometrics, though it was met with resistance when first proposed.

His 2007 book Identification for Prediction and Decision (Harvard University Press) laid out the formal methodology. The core trade-off he identified has become known as the law of decreasing credibility: the more assumptions you make to tighten your confidence interval, the less believable your conclusions become. Strong assumptions can produce clean answers; they cannot produce credible ones.

Surveying What People Actually Expect

A second major strand of Manski's work addresses how economists measure expectations - what firms and households anticipate about their own futures. This matters because economic behavior is driven not just by current conditions but by what people expect to happen. If households expect their incomes to fall, they cut spending now; if businesses expect demand to grow, they invest.

Before Manski's interventions, economists typically modeled expectations through theoretical assumptions rather than by asking people directly. He argued for a more empirical approach: use surveys to ask people the probabilities they assign to various outcomes, such as the chance their home value increases or the likelihood they face unemployment in the coming year.

The idea was greeted with skepticism. Economic traditionalists doubted that survey respondents could meaningfully report probabilities. Manski and collaborators collected the data anyway and showed it was informative. Today, central banks including the Bank of Spain, the Bank of Italy, and the Federal Reserve Bank of New York conduct regular expectations surveys that draw directly on his framework. What was once considered naive is now routine monetary policy infrastructure.

The Reflection Problem in Education

Manski's doctoral thesis - an analysis of how high school seniors decide whether to attend college - grew into a long engagement with education policy. His 1983 book College Choice in America, co-authored with David Wise of Harvard's Kennedy School, examined how economic assistance programs affect enrollment decisions.

The policy question was concrete: the U.S. federal government launched a grant program in the early 1970s designed to increase college attendance among low-income students. Manski used counterfactual methods - constructing alternative scenarios and comparing them to observed outcomes - to estimate how many additional students would enroll under different policy designs. The approach required honest acknowledgment of what the data could and could not establish.

Later, he formalized what is now called the reflection problem: the difficulty of separating how much a student's performance reflects peer influence from how much it reflects their own ability, when both are operating simultaneously. If a student does well in a high-performing class, is that because talented classmates raised their performance, or because they were already talented? The two effects are mathematically entangled in cross-sectional data. Manski showed this is not a problem that clever analysis can simply overcome - it is a fundamental limit of what the data structure can reveal.

Applying Uncertainty Methods to Clinical Medicine

His 2019 book Patient Care under Uncertainty (Princeton University Press) extends the same logic to clinical medicine. Manski argues that physicians face structurally similar problems to policymakers: they must make decisions for individual patients using evidence that was generated in populations that may not resemble the patient in front of them. Clinical trial results from a limited sample are routinely generalized to the broad population, a step that introduces uncertainty which medical practice rarely acknowledges explicitly.

The book examines several failure modes: statistical errors that arise when selecting one therapy over alternatives, trial results that do not replicate in real-world practice, and drug efficacy claims that dissolve under scrutiny of post-market evidence. Manski's prescription is the same as in economics: quantify the uncertainty rather than suppress it, and use that quantification to make better decisions under ambiguity.

He notes that the methodological gap is not unique to medicine. Doctors have strong incentives - professional, institutional, and psychological - to project confidence. Patients expect definitive recommendations. The combination produces what Manski calls incredible certitude: the appearance of precision that the evidence does not support. The same phrase applies equally to economic forecasting, policy evaluation, and clinical judgment.

What the Award Committee Said

The Frontiers of Knowledge committee, chaired by Nobel laureate Eric Maskin of Harvard, described Manski as a "foundational figure" whose work "has profoundly influenced empirical research across education, health policy, labor markets, industrial policy, and social programs." Committee member Sir Richard Blundell of University College London credited him with teaching economists to examine the assumptions underneath their analyses before trusting the results.

Manuel Arellano of Spain's Center for Monetary and Financial Studies described Manski as a "major innovator in methods of empirical measurement" who has also made fundamental contributions to quantifying the range of possible outcomes under conditions of genuine uncertainty. A total of 82 nominations across 70 candidates were received for this edition. Manski was nominated by Thierry Magnac of the University of Toulouse and Richard J. Smith, Emeritus Professor at Cambridge.

Now 77 and still teaching at Northwestern, Manski is converting the methods he developed over five decades into a practical software tool: a web application that would allow non-specialists - clinicians, policymakers, administrators - to apply uncertainty quantification to real decisions without needing mathematical training. Whether that project shifts behavior broadly remains to be seen. The academic record, at least, has now been formally honored.