What this covers

Russ Roberts and Noah Smith debate whether economics qualifies as a science, disagreeing sharply on what standard to apply. Roberts contends that economics fails the test because even well-executed empirical studies rarely settle disputes decisively—economists simply find methodological grounds to dismiss findings that contradict their prior beliefs. Smith argues the bar for "science" is lower than Roberts imagines: physics too features smart people in genuine disagreement, and economics accumulates knowledge through the accumulated weight of evidence rather than singular thunderbolt studies. The conversation moves between specific case studies (the stimulus, minimum wage, education lotteries, deworming programs) and foundational questions about how causal identification works in messy real-world settings.

The substance centers on the natural experiment as economics' best empirical tool—and its serious limits. Smith explains that natural experiments beat mere correlation but fall short of controlled lab experiments, and macroeconomics faces a peculiar bind: its phenomena are global and entangled, so clean comparisons almost never exist, leaving priors and ideology to fill the void. He traces belief revision in economists like Robert Lucas and Steve Williamson, showing that scholars do update when evidence accumulates. The discussion covers the dreaded third thing—the omitted variable that can undermine any causal claim—and why disagreements often hinge on whether that concern is plausible or merely pedantic. Roberts pushes back with historical examples: the postwar period saw Keynesians' gloomy predictions fail, yet their convictions remained unmoved. He and Smith diverge most visibly on the minimum wage's real-world effects and whether schooling signifies skill or simply selects for it, illustrating how empirical economics can produce honest disagreement despite abundant data.

Sharpest takeaway

Roberts argues economics fails as a science because its empirical methods—even natural experiments—rarely produce results decisive enough to change minds, while Smith counters that economics is as scientific as physics, with knowledge accruing through accumulated evidence rather than single decisive studies.

  • Macroeconomics has almost no clean natural experiments because policy changes are non-exogenous and effects are global, so priors and ideology dominate debates
  • Even hard sciences like physics feature smart people disagreeing, so the bar for being a science is lower than imagined
  • Knowledge in economics progresses by accumulated weight of consistent studies, not by 'thunderbolt' studies that decisively change minds

The argument · threads4 threads · 27 claims
0.85

Economics cannot settle disputes empirically; natural experiments generate journal articles but rarely convince people to change ideological commitments.

3 pointscentrality 5/5
  • Roberts cannot think of any economic study so decisively executed that opponents had to throw up their hands and change their views, which suggests economics has serious problems claiming to be a science—and the recent $800 billion stimulus is a prime example, since brilliant credentialed economists with Nobel Prizes still cannot agree whether it succeeded.

    “Can you think of a study that was so decisively performed... that people on the other side of the debate had to throw up their hands and say, 'Well, I guess I was wrong.' Because I can't think of one”

  • Roberts's strong claim is that natural experiments are really good for generating publications in economics journals but not good for convincing people; rather than being merely not-quite-as-good as real experiments, they are only good for producing journal articles, since most studies don't convince anyone—people simply find what's wrong with them and dismiss them based on prior ideological or methodological commitments.

    “Natural experiments are really good for getting publications in economics journals. They are not so good for convincing people”

  • The end of WWII was a natural experiment across time: Keynesians like Paul Samuelson predicted that a sharp drop in military spending and a flood of returning workers would cause a depression worse than the Great Depression, but military spending fell dramatically with no cushioning policy and no disastrous downturn followed, yet this failure did not shake the Keynesian convictions of the day.

    “the end of WWII brought a massive decrease in government spending, an enormous increase in the number of people suddenly looking for work. And there was basically--nothing happened”

0.71

Minimum wage harms low-skill workers by pricing them out of jobs and opportunity despite aggregate prosperity for educated families.

3 pointscentrality 4/5
  • While the economy works fine for skilled people like Roberts's children, less-skilled workers are handicapped by a minimum wage that makes them too expensive to hire relative to their skills, cutting off their access to job experience and the engine of prosperity, which is a key reason to worry about expanding government even when aggregate life is good.

    “they are handicapped by the fact that their labor market opportunities are cut off by, I believe, a minimum wage that makes them too expensive to hire relative to their skills”

  • Influenced by Bryan Caplan, Roberts is more open to the possibility that schooling is mainly signaling—that the higher wages associated with education result from uncontrolled person-specific variables (like natural ability) rather than from the education itself—though this is not a core position that his worldview hinges on.

    “I'm more open to the possibility that school is mainly signaling; that the higher wages associated with schooling are really a result of uncontrolled variables that are person-specific rather than related to the education itself”

  • Life is pretty good in America despite complaints that it's all going to hell, and it's especially good for educated, advantaged families whose children have a strong shot at material success even in an economy where government is 40% of GDP.

    “Life is pretty good in America, say, where you hear people complaining that it's all going to hell in a hand basket”

0.65

Economic policy lacks dependable empirical relationships that can serve as reliable basis for decision-making.

1 pointcentrality 5/5
  • Economics is not much of a science in the everyday sense because it does not produce reliable empirical relationships that can be used as the dependable basis for economic policy.

    “It doesn't produce reliable empirical relationships that can be used as the basis for economic policy”

0.61

Economic studies often fail to replicate or scale reliably across contexts, from macro predictions to development interventions.

2 pointscentrality 4/5
  • Monetarists predicted that the Fed's balance sheet expansion from $800 billion to $4 trillion would cause massive inflation, but it did not materialize; while they argue it simply 'hasn't happened yet,' they should consider that their understanding of the relationship between the Fed's balance sheet, velocity, and price levels may be imperfect.

    “the balance sheet of the Fed went from $800 billion to $4 trillion and every economist on "my side" said that's going to cause massive inflation. And nothing happened”

  • The widely promoted claim that deworming African and poor children improves educational and school performance was based on a natural experiment that later proved unreliable—it didn't scale, worked in some places but not others, and its educational benefits are very much in question—illustrating that reliable results are hard to find in complex systems even outside macro.

    “it turned out maybe it wasn't so reliable; it didn't scale; it didn't here, it didn't work there... the benefits in terms of improved educational and school performance are very much in question”