The anthropologist Edward Evans-Pritchard described the fallacy of “if I were a horse”. Since we have not been, and never will be, horses, our speculations are unlikely to connect with reality. We should not use categories derived from our own experience to interpret behaviour in a different culture and environment.
Ronald Coase, the Nobel laureate in economics who will shortly celebrate his 101st birthday, described how his colleague, Ely Devons, extended the metaphor. “Let’s wonder,” Devons mused, what an economist would do if he wanted to study horses. He would go to his study and think “what would I do if I were a horse?”
If you think this is far-fetched, consider how one might study how participants in financial markets form and revise their expectations. You might start with the extensive literature on learning created by psychologists and educators. You would interview analysts and traders, hoping to understand what kind of information they used, how they processed that information, and how their views of the future evolved. These investigations would suggest hypotheses that could be tested against actual future behaviour.
This approach would not get a research grant, or even a job, in a good economics department. The economist who studied learning adopted the approach of “what would I do if I were a horse?” The economist asked: “How would I behave if I were a rational agent in a world populated by other rational agents, and I knew everything about the world except for the values of particular parameters?”
The economist asked: “What would I do if I knew that one of several possible models of the world was correct but did not know which one?”
These are certainly possible ways of looking at the problem. You might think they do not sound promising, although I might try to persuade you that similar thinking had offered some insights into problems with a very simple information structure, such as auctions. But experience has justified your scepticism.
You might be shocked but not surprised to learn that the use of these methods had provided a seriously misleading account of the events that gave rise to the 2007-08 financial crisis – attributing the credit boom not to crass errors in forecasting and valuation but to sophisticated design of products to minimise risk.
You might be shocked and surprised to learn that the economists who built these models regard their approach as the only valid economic way of looking at the world, and dismiss others as unscientific. And you might be surprised that the principal exponent of this approach – Thomas Sargent – has been to Stockholm this week to receive a Nobel Prize for his work.
The Swedish Academy was on firmer ground when it gave the prize to Mr Coase. The centenarian’s most important work, undertaken between the wars, was on the nature and role of large integrated corporations such as General Motors. Following fieldwork in these companies, Mr Coase developed a theory of the US carmaker by reference to the relative costs of hierarchical and market organisation. The notion that economic institutions are a product of the relative cost of alternative modes of transaction has illuminated many other economic phenomenon.
Mr Coase understood the inter-dependence of theory and observation, and the need to develop models in context. That is why the problem he saw that was solved one way at General Motors 80 years ago was tackled differently in Japan – and would be tackled differently again in the twenty-first century corporation.
Evans-Pritchard rejected desk-bound speculation in favour of the detailed fieldwork that is today the hallmark of anthropology. Gillian Tett, the FT’s US managing editor, with her PhD in anthropology, understood the origins of the 2007-08 crisis by watching the markets and interviewing the participants. Much though I admire Ms Tett, I do not think she should get a Nobel Prize. But nor should Mr Sargent.
Ms Tett’s approach is certainly not the only, or necessarily the best, approach. But it would be foolish, even absurd, to comment in the absence of such insights. Before I hired a horse doctor or horse trainer, I would always ask “have you ever seen a horse?”
