Adam Ozimek on the Power of Econometrics and Data

Episode Summary

Executive Summary: Russ Roberts and Adam Ozimek debate how much economics can really “change minds.” They focus on empirical evidence in trade, minimum wage, fiscal stimulus, and market behavior, arguing that economics does learn from data but also that worldviews, anecdotes, and theory strongly shape interpretation. Both agree economics progresses, but slowly and imperfectly.

Main Topics: Can evidence change economists’ minds? (Priority: 5/5): Ozimek argues that economists do revise beliefs when strong empirical work accumulates, but change is uneven because priors and worldview still filter what people accept. Trade with China and adjustment costs (Priority: 5/5): They discuss Autor, Dorn, and Hanson’s research showing larger-than-expected, long-lasting local labor-market harms from import competition, which softened Ozimek’s earlier confidence in labor-market flexibility. Minimum wage empirical debate (Priority: 5/5): The conversation examines how newer studies finding weak employment effects have shifted some economists’ views, while Ozimek thinks broader evidence still points toward job loss and other hidden costs. Fiscal stimulus and the Great Recession (Priority: 4/5): Roberts and Ozimek compare stimulus debates to other policy disputes, noting that the 2008 package bundled many interventions, making clean causal evaluation difficult, but that some components like Cash for Clunkers look clearly weak. Market efficiency, ideology, and worldview (Priority: 4/5): Roberts reflects on becoming more skeptical that markets work as smoothly as he once believed; Ozimek argues economists should be more skeptical of their lenses and less dismissive of empirical work. Limits of economics versus other sciences (Priority: 3/5): They contrast economics with medicine and life sciences, concluding that even fields with better experimental methods face bias and slow paradigm change, so economics is not uniquely flawed though it is harder.

Key Arguments: Ozimek’s central claim is that economists do change their minds when empirical evidence is strong, and his own shift on trade exposure from China is a prime example. The China trade literature showed that adjustment costs were larger and more persistent than expected, especially in heavily exposed local labor markets, with workers dropping out of the labor force and moving to disability programs. Roberts argues that trade’s benefits are still real, but that much support for free trade comes from theory, casual observation, and broad historical reasoning rather than definitive econometric proof. On the minimum wage, Ozimek says the recent literature is persuasive enough to have moved many economists, but he expects future evidence to restore a stronger consensus that higher minimum wages reduce employment. Roberts pushes back that accepting no employment effect should imply broader beliefs about firm behavior, incentives, and worker adjustment, not just a narrow conclusion about jobs. Both agree macroeconomics is much harder than microeconomics because time-series data are noisy, policy bundles are complex, and causal identification is weaker. They treat Cash for Clunkers as an example of a stimulus policy that likely only shifted demand forward briefly rather than creating lasting gains. Roberts emphasizes that economists often use empirical results to dress up deeper ideological or worldview commitments, while Ozimek argues that empirical studies still meaningfully discipline those commitments.

Data Points: Episode date: January 26, 2016 - Introductory metadata for the EconTalk episode. Minimum wage workforce share: around 5% or less - Roberts notes the subgroup most affected by minimum-wage policy is a small share of the labor force. Stimulus package size: $780 billion - Roberts references the broad public shorthand for the American Recovery and Reinvestment Act. Expected Medicare spending reduction: about one-third less - Ozimek cites David Cutler’s work suggesting end-of-life care costs could fall if guidelines were followed. Time span of demand shift in Cash for Clunkers: one to four quarters - Ozimek says the program mostly shifted car purchases forward in time. Survey cited: University of Chicago economist survey - Used to support the claim of a strong consensus that free trade benefits the average person.

Pivotal Quotes: "doing nothing is more costly than we thought, which lowers the cost of doing something" — Adam Ozimek: Ozimek explains why recognizing higher adjustment costs from trade exposure changes policy thinking. "we have these sort of vague ideas that come from our personal experience, they're not made up... and then we dress it up with these studies" — Russ Roberts: Roberts summarizes his critique that empirical studies often validate preexisting worldviews rather than replace them. "I think we need to be more skeptical of our ideological priors and try to use empirical evidence better because it really can be persuasive" — Adam Ozimek: Ozimek argues that evidence should discipline theory and policy beliefs rather than be dismissed.

Implications: Listeners are left with a nuanced view: economics does learn from data, but not cleanly or automatically. Policy debates should account for adjustment costs, hidden margins, and the limits of causal inference, while remaining open to revision as evidence accumulates.