The Individual Failings of Economics
In recent decades, economics has gone from defining itself as a set of questions to defining itself as a set of methods, all based on individuals making decisions. By doing so, it has undermined its own ability to make progress.
CAMBRIDGE – Economics could advance enormously if it relaxed one of its most precious assumptions: methodological individualism, or the idea that any explanation needs to be related to individuals making sensible decisions. This requirement puts the discipline at a huge disadvantage vis-à-vis the natural sciences, because it prevents progress in understanding the relationship between the micro and the macro.
Physics explains all behavior by assuming some fundamental laws at the (very) micro level. Quarks give rise to protons and neutrons, which, together with electrons, generate atoms, in turn giving rise to molecules and macro-molecules such as DNA, genes, and proteins. These produce cells, multicellular beings, and whole ecosystems that live on a planet that rotates around the sun. In theory, one should be able to explain all of this by going back to the fundamental laws of particle physics. In practice, this is not only impossible but also unnecessary, facilitating progress.
We know about all of these levels because scientists looked into them and described them in as much detail as possible, enabling other scientists to explain them in terms of lower-level determinants. Each layer can somehow be related to the layer below, all the way back to quarks and electrons.
Whereas going back one step is not easy, but often doable, going forward even one step is hard. We can work out the amino acid sequence of a protein from the gene that codes for it, but we still are unable to establish what three-dimensional shape the protein will take, which is fundamental to determining its function.
Making things even harder is a phenomenon known as emergence, whereby a next-level property does not exist at the previous step. Diamonds and graphene have very different properties, for example, but are chemically identical. Neurons give rise to consciousness, but only at the level of millions of networked neurons; we would never have guessed it by looking inside the neuron.
Contrast this with economics today. Methodological individualism requires that all phenomena ultimately be explained in terms of individuals making decisions they have sound reasons to make. Studying regularities in aggregate data – typical of macroeconomics before the 1970s – is thus uninteresting if these cannot be grounded in rational individual behavior.
Subscribe to Project Syndicate
Enjoy unlimited access to the ideas and opinions of the world's leading thinkers, including weekly long reads, book reviews, topical collections, and interviews; The Year Ahead annual print magazine; the complete PS archive; and more. All for less than $5 a month.
As the Nobel laureate economist Robert Lucas argues, governments could not trust these regularities to be stable if they based policies on them, because individuals would respond to those measures in ways that would undermine the regularities. The data might suggest a trade-off between inflation and unemployment, but if governments tried to “buy” less unemployment through a bit more inflation, people would change their inflation expectations in ways that would make the whole exercise futile.
The economics profession thus developed models with strong micro foundations, centered on individuals making rational decisions and responding to well-understood incentives. To make progress while abiding by these requirements, economists had to simplify or dumb down the layers of interaction between the individual and the aggregate outcomes they were trying to explain. One common way to do this is to assume that all individuals are identical, or that they are heterogeneous in predictable ways. But requiring all economic explanations to be based on individual behavior is like attempting to explain global warming with quantum physics.
Fortunately, this methodological approach is crumbling. At the micro level, behavioral economics has dented belief in the assumption of individual rationality. In a series of papers, Harvard’s Xavier Gabaix has shown that all the basic tenets of both micro and macroeconomics change a lot if we assume that there are limits to agents’ rationality. Likewise, Gabaix’s Harvard colleague Joseph Henrich argues that the way people make decisions is not universal, but rather depends on a society’s culture.
More to our point is the issue of going from decision-making individuals to the aggregate level. The extremely talented late Harvard economist Emmanuel Farhi, working with UCLA’s David Baqaee, showed that we need to consider the (unexplained) input-output structure of production in order to understand macroeconomic fluctuations: we cannot just derive it from individual representative agents.
Similarly, Harvard’s Pol Antràs (with co-authors) has recently been reconstructing the theory of international trade by assuming that the world is organized through global value chains instead of standard markets. This apparently minor assumption makes huge differences both in theory and in terms of trade-policy implications. We are barely starting to understand what it means in practice, because, up to now, we had not bothered to collect the requisite firm-to-firm data.
Seen from this perspective, the neoclassical theory of economic growth looks quaint. Its main contribution, in the words of the Nobel laureate economist Paul Romer, is to show how hard it is to ground long-term growth in theory. Alas, neoclassical theory has been next to useless for any practical purpose, mainly because it blatantly disregards the meso-structures that exist between individuals and aggregate economic outcomes.
Fortunately, some researchers have tried to uncover these meso-structures, using big data with network science and other techniques. For example, they have identified complex structures of skill complementarities and patterns of relatedness within and across industries, technology classes, and scientific areas.
These studies show that meso-structures matter for how cities and countries grow, and how technologies develop. Given the current orthodoxy, these papers have been unpublishable in economics journals, because they cannot show how these structures are linked to individuals making decisions under constraints. But they have been published in prestigious scientific journals such as Nature and Science, as well as in the Journal of Urban Economicsand Research Policy. As a result, other researchers can ask questions about how these meso-structures arise from, say, individual decision-making.
In recent decades, economics has gone from defining itself by the questions it asks to defining itself by the methods it uses. By restricting its approach to methodological individualism, it has undermined its own progress.