The dismally obedient science

Empirical inquiry has its place in economics, but economists must not let the pursuit of causality displace the pursuit of explanation

Economics has spent the past three decades celebrating what it calls a credibility revolution. Armed with natural experiments, randomized controlled trials and increasingly sophisticated statistical techniques, economists claim to have become less ideological and more scientific. The discipline now prides itself on answering questions with facts and evidence, rather than theory and assumptions.

At the heart of this transformation lies a paradox. As economics has gotten better at identifying facts, it has gotten worse at understanding the forces that produce them. Today’s economists can tell us with remarkable precision whether a particular policy affects employment, school attendance, firm productivity or household income, but they are unlikely to ask why labor markets, education systems, firms or inequality take the form they do in the first place.

This is not a new problem. More than a half-century ago, Stanford University economist Paul Baran warned against the “fallacy of misplaced concreteness” — that is, the tendency to mistake observable facts for objective reality — in the social sciences. Unlike physical laws, social facts are not bestowed by nature. Wage levels, poverty rates, wealth distributions, educational outcomes and employment patterns are products of a range of human-generated forces and factors, such as institutions, power relations and social conflict. Any social science that treats them simply as data risks conflating outcomes and causes.

Illustration: Mountain People

Some of the discipline’s most celebrated achievements demonstrate that contemporary economics is increasingly making precisely this mistake. Studies of social mobility, particularly the work of Harvard University public economics professor Raj Chetty and his collaborators, have shown how neighborhoods, schools and local social environments shape life chances and intergenerational mobility.

Research on minimum wages, beginning with economists David Card and Alan Krueger’s landmark natural experiment comparing fast-food restaurants in New Jersey and Pennsylvania, has generated influential evidence showing that increases in minimum pay do not necessarily reduce employment. Card, together with econometrician Guido Imbens and economist Joshua Angrist, later won a Nobel Prize for work that had “revolutionized empirical research” and transformed economists’ ability to answer causal questions.

Massachusetts Institute of Technology economics professors Abhijit Banerjee and Esther Duflo, together with the University of Chicago economics professor Michael Kremer, received a Nobel Prize for their “experimental approach to alleviating global poverty,” which relied on randomized controlled trials to identify effective interventions. From educational attainment to immunization rates, the findings of such experiments could, many economists believe, transform development policy.

All these achievements are valuable, but, in each case, economists are reporting outcomes, which is different from identifying underlying causes. Determining how a minimum-wage increase affects employment is not the same as showing why wages remain low. Explaining that neighborhoods shape mobility is not the same as deciphering why segregation and spatial inequality persist. Identifying which intervention improves vaccine uptake is not the same as understanding the social and economic underpinnings of poverty.

This approach — studying adjustments within the system (what happens when a variable changes), rather than the system itself (how the variables came into existence) — is increasingly prevalent.

One recent meta-analysis of more than 44,000 economics working papers showed that the share of claims documented via causal-inference methods (as opposed to general claims) soared from 4 percent in 1990 to nearly 28 percent by 2020, and that causal-inference methods are increasingly rewarded by top journals.

As economics becomes increasingly capable of producing credible knowledge about segments of reality, it risks becoming less capable of understanding the whole picture. Some might argue that this is the price of scientific rigor. Large questions about capitalism, power, class or social structures are difficult to test empirically. Better to answer narrow questions reliably than broad questions poorly. However, this argument, despite its merits, fails to recognize that economics research does not merely provide answers; it also shapes what counts as a legitimate question.

Treating the social order as a static condition has intellectual and political consequences. A social science that does not investigate a society’s underlying structure treats the institutions, power dynamics and inequalities that generate economic problems as somehow natural and even immutable. The status quo is upheld not through explicit ideological defense, but through analytical omission. A discipline committed to objectivity becomes a vehicle for apologetics.

Baran’s warning remains as relevant today as when he first issued it. The greatest danger facing economics is not ideology masquerading as science. It is the belief that science can progress by focusing only on observable facts, while failing to examine the structures that shape them. Identifying causal effects has its place, but if facts are allowed to become detached from the social forces that underlie them, what looks like objective knowledge might form a partial and even misleading story.

The remedy is not to abandon empirical inquiry. Rather, it is to ensure that the pursuit of causality does not displace the pursuit of explanation. Beyond identifying causal effects within existing arrangements, economists must investigate how those arrangements are produced, reproduced and contested. Only then can empirical analysis become genuinely critical, rather than serving as a sophisticated defense of the status quo.

As Baran puts it: “refusing to accept as a datum or to treat as immune from analysis any single part of the whole [is an] inescapable necessity.”

The purpose of social science is not to measure the world as it is, but to explain why it is that way and how it might be otherwise. When economics forgets this distinction, it risks becoming an obedient science. The great irony of contemporary economics is that it has never been more rigorous — or less ambitious.

Carolina Alves is associate professor in economics at the Institute for Innovation and Public Purpose at University College London and a fellow at Girton College at the University of Cambridge.