Joshua Angrist

Joshua Angristis an Israeli-American economist who, with the Dutch-American economist Guido Imbens, was awarded one-half of the 2021 Nobel Prize for Economics (the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel) for his “methodological contributions to the analysis of causal relationships” in labour markets. The other half of the prize was awarded to the Canadian-American economist David Card “for his empirical contributions to labour economics.” The work of the three economists showed how certain “natural experiments,” or real-world social developments arising from policy changes or chance events, because of their resemblance to controlled or randomized experiments in medicine and the physical sciences, could be used to clarify causal relationships in the analysis of labour markets, such as the relationship between employment rates and the minimum wage and the relationship between level of education and income. The laureates’ approach to natural experiments provided a solid empirical ground on which to address important questions of social and economic policy and, more broadly, “revolutionised empirical research” in the social sciences, in the words of the Economic Sciences Prize Committee.

Angrist received a bachelor’s degree in economics from Oberlin College in Ohio in 1982 and master’s and doctoral degrees in economics from Princeton University in Princeton, New Jersey, in 1987 and 1989, respectively. He taught economics at Harvard University, the Hebrew University of Jerusalem, and the Massachusetts Institute of Technology (MIT) before being appointed professor of economics (1998–2008) and later Ford Professor of Economics (2008– ) at MIT.

A long-standing challenge to empirical research in economics has been that of clearly identifying the economic or social effects of changes in economic policy and the economic or social causes of changes in economic conditions. Such causal relationships are difficult to establish because the nature of the phenomena under study makes it generally impossible for researchers to create control groups—that is, groups sharing the same relevant features as a corresponding experimental group, except that the latter is subjected to a specific change, or “intervention,” which can then be identified as the cause of any resulting change or effect in that group. To test the hypothesis that additional higher education results in higher incomes, for example, researchers conducting a standard experiment would need to randomly assign large numbers of individuals to control and experimental groups and then ensure that members of the latter group received additional higher education and that members of the former did not. In reality, of course, researchers cannot perform such an experiment, because they cannot control how much education other people receive.

Although causal relationships in economics and other social sciences generally cannot be identified through standard experiments, the work of Card, Angrist, and Imbens has demonstrated that many such questions can be addressed on the basis of natural experiments. Angrist and Imbens’s important contributions were to explore the strengths and limitations of natural experiments and to develop a method for drawing valid causal conclusions from them. In an influential paper published in the mid-1990s, “Identification and Estimation of Local Average Treatment Effects,” they considered the general problem of identifying a causal relationship between correlated interventions and effects in situations where effects vary between subjects and researchers have no control (or incomplete control) over which subjects undergo the intervention and which do not. (One source of uncertainty in such situations is that researchers would be unaware of the subjects’ possible motives for undergoing or avoiding the intervention—assuming they have a choice—which could act as additional or alternative causes of a given effect and thus make it difficult to identify the intervention itself as a single cause.) Angrist and Imbens’s solution enabled them to calculate an average causal effect for a given intervention, which they called a “local average treatment effect,” or LATE, despite these complicating factors. The framework they developed has enhanced scholarly understanding of the operation of labour markets and has greatly extended the insights available to empirical researchers in other social sciences.