Other Writing

Models tell us more than hindsight

According to my esteemed colleague Gideon Rachman, economists should be swept off their thrones by historians. Economists have had far too strong a stranglehold on the levers of power, he claims. They think they are scientists. They think they can foretell the future. They are wrong: “pseudo-scientists”, “peddling brash certainties”. Historians such as Gideon and Professor Niall Ferguson, hitherto relegated to backwaters such as the FT’s op-ed page, should at last be paid a bit of attention.

In pondering how to respond, I suffered an acute shortage of brash certainty. Gideon is quite right about the importance of history. When it comes to economics, however, the chief source of brash certainties appears to be Gideon, who wouldn’t know an economic model if it paraded down a catwalk at him.

I know as little about history as Gideon knows about economics, but no doubt he is right to suggest that an important role of the historian is to emphasise the knotty particularity of time and place, and the difficulty of producing sweeping scientific laws that accurately describe a complex social world. Economists, sociologists, psychologists and anthropologists should appreciate this just as keenly. The best do. Many do not and, sadly, they are over-represented in the media. Perhaps this is the reason Gideon misunderstands the task and the methods of economics.

He argues that while economic edifices are always collapsing, “buildings constructed according to the laws of physics seem to stand”. This is an odd statement. Buildings constructed according to the laws of physics have a habit of falling down. Henry Petroski, engineer and author of Success through Failure, observes that structural engineers tend to learn by constructing ever more ambitious structures. When one of them falls down or wobbles, engineers figure out what was wrong with their models. Sometimes the results are tragic: when the innovative Malpasset dam cracked thanks to inadequate geological modelling, nearly 400 people died. Sometimes they are delicious: the award-winning Kemper Arena collapsed, with no loss of life, just 24 hours after hosting the American Institute of Architects Convention. From his riverside eyrie, I think Gideon can just see the famous wobbly bridge across the Thames. Is this really a damning indictment of the laws of physics?

But, of course, our grasp of the laws of physics is not to blame. The trouble is the difficulty of modelling them in a world with snowdrifts, clay seams and error-prone contractors. In short, buildings, like economic institutions, stand up not because of our grasp of the laws governing them, but because they have survived a process of trial and error in a complex world.

Economic institutions are more complicated and unique than any building. No wonder that progress is so difficult. But Gideon is too quick to dismiss “models and equations”. I agree that macroeconomic models have proved fairly useless. I also agree that economists, like historians, sociologists, political scientists and newspaper columnists, make terrible forecasters. But few academic economists bother to try, and forecasting models represent a small slice of the mathematics deployed by economists.

What, for instance, of the famous contention by the economist Steven Levitt and his co-author, John Donohue, that legalised abortion in the US reduced the crime rate about 18 years later? This is a hypothesis about history, but one that no historian is well-qualified to judge. Instead, the hypothesis has been tested statistically with some ingenuity; the statistical models themselves have been contested, pulled apart, found wanting in some respects, double-checked using alternative data and tested against the experience in other countries. The debate continues. Is this process “science”? I am not sure. But it certainly isn’t idle banter.

No doubt economists can learn from historians, but the search for economic regularities should not be abandoned. It is not limited to traditional economic approaches. We know much more about economics thanks to the work of Robert Axtell (computer scientist), Cesar Hidalgo (physicist), Duncan Watts (sociologist), Esther Duflo (an economist who runs the kind of controlled experiments which, according to Gideon, don’t happen in economics) and Daniel Kahneman (psychologist). All of them use those pesky “models and equations”. Are mainstream economists receptive enough to such invaders? The best ones are. The majority are not, but that is a fact about academia, not economics.

At least Kahneman has been rewarded for his efforts with a Nobel memorial prize. (Or as Gideon would put it, a “fake Nobel”. The Nobel Memorial Prize in Economics was established in 1968, long after the dynamite magnate’s death, unlike the rigorously scientific Nobel Peace Prize awarded to Henry Kissinger, or the rigorously scientific Nobel Prize for Literature denied Leo Tolstoy.)

My own supervisor, Paul Klemperer, is presumably the kind of economist Gideon despises: a game theorist who tries to understand the world through mathematical models. How quaint and arrogant. But my clearest recollection of Paul’s teaching is a series of classroom demonstrations calculated to undermine predictions of game theory and to open his students’ minds to the likelihood that the models lacked something important.

Paul used those models to help design an auction that raised £22.5bn for the British government, and then helped the Bank of England design an auction that would help them to inject liquidity into banking. Paul can’t forecast the future, but the auctions have – like many buildings – stood up well so far.

Historians deal in hindsight. It is a wonderful thing. But it is not the only thing. I wonder whether Gideon, intoxicated with a heady brew of Niall Ferguson and Herodotus, has forgotten that.

Gideon Rachman responds:

In his spirited and learned defence of his beloved economists, Tim Harford makes a mistake that is a characteristic of the profession that we both really belong to – journalism. That mistake is to rest too much of his argument on a couple of anecdotes. The story about the collapse of the Kemper Arena is indeed delightful. The implication Tim seems to draw is that the laws of physics and engineering are no more reliable than the “laws” of economics. They are all just hypotheses, which are gradually improved by trial-and-error in the real world. But it is my perhaps lazy impression, that the hard sciences have established a body of settled, scientifically-testable knowledge that economics simply cannot lay claim to. Is Tim really saying that ain’t so?

Tim also admits that macroeconomic models have proved “useless” at forecasting but suggests that this is a minor matter, since few academic economists “even bother to try”. But the starting point for my piece was Joseph Stiglitz’s suggestion that “the failure of much of the economics profession to see the crisis coming should be a cause of great concern.” And Professor Stiglitz is a winner of the Bank of Sweden prize for economics – sometimes referred to as the Nobel prize.

Also published at ft.com.

11th of September, 2010Other Writing • Comments off