Tim Harford The Undercover Economist

Articles published in July, 2019

How economics can raise its game

How can economics become a more insightful discipline? Should it aim to be more like physics, with its precision and predictive power? Or should economists emulate anthropologists or historians, immersing themselves in the details of the particular and the unquantifiable?

There’s a case to be made either way. Some critics argue that economics is missing better physics: it got stuck in the 19th century with fusty old ideas like marginal analysis and equilibrium, and missed out on cool ideas like chaos theory and phase transitions that promise to shed insights on economic complexity or sudden crises. (See, for example, Philip Ball’s excellent book, Critical Mass.)

Others say that economics needs to put the mathematics down and back slowly away. As Immanuel Kant put it, “out of the crooked timber of humanity, no straight thing was ever made”, so economists should be less fond of putting rulers against everything.

The fact that both views have the ring of plausibility suggests that this is a tougher challenge than it might appear from the sidelines. Now a new paper addresses the question from the heart of academic economics: Nobel laureate George Akerlof, writing in the Journal of Economic Literature.

Prof Akerlof, now at Georgetown, argues that the academic discipline of economics rewards “hard” rather than “soft” research with publication in the top journals, and therefore with promotion and status. We know “hard” when we see it: numbers are harder than words, quantities harder than qualities. Causation is harder than correlation. Physics is “hard” and sociology “soft”.

There is much to be said for hard science. The downside is that certain questions cannot be answered — or perhaps even asked — in precise, mathematical, causal terms. They are still important, and if economics insists on “hard” methods it will overlook them.

Another Nobel laureate, the late Gary Becker, reflected on his college studies: “I began to lose interest in economics . . . because it did not seem to deal with important social problems. I contemplated transferring to sociology, but found that subject too difficult.”

What is fascinating about that remark is that Becker earned his prize by applying the hard-ish tools of economics to areas that seemed to belong to softer disciplines: addiction, discrimination, marriage and education. How well he succeeded remains a matter of debate. Personally I believe he contributed a lot.

Yet plenty of sociology remains outside the reach of the economists’ tools: it is important but too soft — or in Becker’s words, “too difficult”. More awkwardly for the economics profession, some key economic questions also seem more likely to yield to soft than hard approaches: what are the obstacles to social mobility? Where does innovation come from? Can we strengthen the institutions that matter for prosperity?

Beyond any particular problem there is also the challenge of combining insights from highly specialised subfields. Raghuram Rajan, when he was chief economist of the IMF, came closest to predicting the 2008 financial crisis. He later observed that economists had written insightfully on all the key issues but had lacked someone capable of putting all the pieces together.

Is this also a hard/soft problem? Prof Akerlof thinks so. He argues that the bias towards hard analysis also produces a bias towards specialised silo thinking, and that being a generalist is something of a soft skill.

I’m not sure that’s right. Certain mathematical tools are both highly portable and distinctly hard-edged. But it is surely true that the kind of synthesis that would have identified the looming crisis would have been too discursive to be published in a top economic journal.

So I am not completely persuaded that economics is, on average, “harder” than it should be. But I am in complete agreement with the recommendation that economics needs to be more tolerant of different methods, whether the latest ultra-hard physics or the softer explorations of anthropology or even a business-school case study. After all, the economy encompasses a lot of different things; why should it yield only to a particular set of analytical tools?

Economics has certainly profited from the insights of those outside the field, such as psychologist Daniel Kahneman, and the late Elinor Ostrom, a political scientist whose ambition to study economics was thwarted because, as a girl in the 1940s, she’d been steered away from mathematics. Both have Nobel memorial prizes in economics.

The insights can flow the other way, too. Charles Darwin’s theory of evolution was influenced by the economist Robert Malthus’s Essay on the Principle of Population. When working on the kinetic theory of gases, physicist James Clerk Maxwell drew inspiration from social scientists and their habit of thinking in statistical approximations about large populations.

It is easier to make such a recommendation than to embrace it. The further the leading economics journals stray from their core expertise, the more difficult they will find it to distinguish good work from bad. But at the margin, such moves offer a lot of promise.


Written for and first published in the Financial Times on 28 June 2019.

My book “Fifty Things That Made the Modern Economy” (UK) / “Fifty Inventions That Shaped The Modern Economy” (US) is out now in paperback – feel free to order online or through your local bookshop.

Free email updates

(You can unsubscribe at any time)

How the Brexit debate was flushed down the drain

On a scale of one to seven, how well do you understand how a flush lavatory works?

This was a question asked by two Yale psychologists, Leonid Rozenblit and Frank Keil, almost two decades ago. Before I explain why, here’s a follow-up exercise: write down your lavatory explanation in as much detail as you can. You may wish to draw a diagram, or explain it to a friend. Or not.

You may then reflect that you knew a little less than you realised. That was the experience of many of the study’s subjects — and not just for lavatories (why does all the water disappear down the U-bend?) but also for zips, quartz watches, helicopters, speedometers, cylinder locks, piano keys and sewing machines. People felt they understood the mechanisms that surrounded them, but their confidence was severely dented by the simple act of giving them pencil and paper and saying: “Show me.”

The same exercise can be performed with politics. In 2013, Steven Sloman and Philip Fernbach, authors of The Knowledge Illusion, were members of a research team that did just that, inviting people resident in the US to rate their understanding of American policy proposals such as introducing unilateral sanctions on Iran, a cap-and-trade system for carbon emissions and a national flat tax. They also asked people to rate their approval of each policy, which would have been unnecessary for lavatories and zips. (Lavatories are useful, zips self-evidently malevolent.)

Professors Sloman and Fernbach and their colleagues found that — just as with locks and speedometers — people tended to overrate their knowledge at first, and then discover some humility when asked to be more specific.

Perhaps British voters could use a dose of the same medicine when it comes to our understanding of Brexit. Leave or Remain, many of us came late to the realisation that there was a difference between the single market and the customs union. I am still not sure most people can explain what that difference is. Many people have strong views about Prime Minister Theresa May’s withdrawal agreement; rather fewer could give a convincing account of how it differs from the political declaration that accompanies it.

When Mrs May began her premiership with the statement that “Brexit means Brexit”, it dawned on most of us that the details of the whole project might need a little more work. But she wasn’t the only one who was vague.

I’d love to see the contenders for the Conservative party leadership quizzed a little less about their cocaine habits and instead forced to sit down and write a detailed explanation of what a no-deal Brexit actually is. While we wait, perhaps the same exercise could be given to the 160,000 Conservative party members who are about to select the country’s next prime minister.

How long, for example, will HM Revenue & Customs wave through imports without inspections? Will the French reciprocate? What are the implications of “trading under World Trade Organization rules” for the UK’s banking and insurance industries? How large are those industries?

How many other developed countries are content to rely solely on WTO arrangements in their trade with the EU? Is the WTO capable of enforcing the rules anyway, given the current crisis in its appellate body? How likely is the EU to grant permission to British farmers to sell meat, milk or cheese? Would any of these decisions be different if the UK refused to pay the “divorce bill” it had negotiated?

I don’t think it is especially shameful that we ordinary voters are incurious about the ins and outs of Brexit, any more than we should be obliged to understand the workings of a quartz watch. An ability to read the time is generally sufficient. But I am stunned by just how little we seem to demand of our political leaders.

We want tailors to understand sewing machines, locksmiths to understand locks and plumbers to know that a lavatory is basically a siphon. But our standards for politicians seem far lower. The next prime minister is likely to be a person who believes that if we demanded it with enough gusto, sewage would remove itself from our homes in some scatological remix of Mary Poppins — and that anyone who tells you otherwise is clearly a shill for Big Porcelain.

We should expect more of anyone who wants to lead the country. And since our politicians have grown so fond of punting the hard questions back to us, perhaps we should also demand more of ourselves.

Profs Sloman and Fernbach found that asking people to explain the workings of the policies they so fervently supported or opposed had a humbling effect. When people realised that they knew less than they had once believed, they quite reasonably wound their necks in as a result. It seems strange to die in a ditch for something we can’t clearly explain, even to ourselves.

Next time you find yourself in some heated political debate, perhaps you should suggest that both sides pause to explain the policy in question. You may find you understand less — and agree more — than you realised.

Written for and first published in the Financial Times on 21 June 2019.

My book “Fifty Things That Made the Modern Economy” (UK) / “Fifty Inventions That Shaped The Modern Economy” (US) is out now in paperback – feel free to order online or through your local bookshop.

Free email updates

(You can unsubscribe at any time)

How the US is weaponising the world economy

Back in 2002, serious people were worried about the possibility of a nuclear exchange between India and Pakistan. Millions might have died — and the prospect seemed real enough that both the US and the UK advised their citizens to flee the region. How, then, was the crisis defused?

Thomas Friedman, author of The World Is Flat, is fond of telling the story that US businesses (in particular Dell) told their Indian suppliers (in particular Wipro) to calm things down or get cut out of the loop. And things did indeed calm down, so perhaps it was the concerns over Dell’s supply chain that prevented catastrophe. Perhaps.

Mr Friedman duly coined the phrase “the Dell Theory Of Conflict Prevention”: no two countries will go to war if they are part of the same global supply chain. He was never entirely serious about that, but the question now arises: did he have it backwards? Rather than a line of defence against hostile action, might global supply chains be a line of attack?

One example is the recent executive order banning US companies from working with Huawei, effectively denying the Chinese telecoms company the use of Qualcomm’s chips and Google’s Android operating system. Another was Mr Trump’s crude — and fleeting — threat to slap tariffs on Mexico if it didn’t satisfy him on immigration policy.

It is tempting to view such actions as uniquely Trumpian. Would any other president threaten sanctions against one of its largest trading partners, via Twitter, with 10 days’ notice, at the very moment they were presenting the new Trump-championed trade agreement to the Mexican senate?

Yet while a different president might act with more subtlety, the US seems unlikely to abandon the aggressive tweaking of the nerves and sinews under the skin of the world economy. Henry Farrell and Abraham Newman — political scientists at George Washington and Georgetown Universities, respectively — have popularised the term “weaponised interdependence”, the title of a forthcoming article in the journal International Security.

Messrs Farrell and Newman point out that supply chains and digital networks can be used both as a “panopticon” to see everything that happens and as a “chokepoint”, denying access to some vital service. Both approaches require a certain bureaucratic apparatus — something that would be hard to disassemble once in operation. There is more going on here than the whim of “Tariff Man”.

Consider Swift, the international financial messaging system. Although Swift does not directly handle transfers of money between banks, it provides the secure service that makes those transfers possible. Swift is a private company based in Brussels, yet last summer it found itself on the receiving end of US demands to cut off Iranian banks. The EU, in turn, demanded that it did not comply. Forced to pick a side in this tug of war, late last year it picked the US.

That is an indication of just how much power the US can wield if it is determined to do so. And the temptation is strong: it seems far safer to attack Iranian interests through stern letters to a messaging service in Brussels than with a carrier strike group.

That very temptation, of course, risks over-reach. The US is not the first global superpower to ponder the use of financial and communication networks as a weapon of war. In the early 20th century, modern economies were increasingly underpinned by complex financing. Britain viewed the central role of the City of London in the world’s banking, telegraph and marine insurance system as potentially decisive when coupled with the power of the Royal Navy. Should war break out with Germany, these networks could be used to sustain the UK economy while crushing that of Germany.

The idea seemed plausible, but needless to say, these plans for economic shock and awe failed both to head off the first world war and to limit its duration and brutality.

Two questions arise: would the US be wise to use its economic leverage more sparingly? And should other nations be building alternative networks beyond the hegemon’s gaze and grip?

It is too easy to say that the US should restrain itself in its own enlightened self-interest. The logic of network effects is self-reinforcing. Having established a central position in finance through Wall Street and the mighty dollar, and in digital networks thanks to Silicon Valley and the Pentagon’s role in funding the early internet, the US advantage may endure many abuses.

Still, the more the US seeks to coerce others through its privileged position in banking and the internet, the greater the incentive to develop alternatives that cut the US out of the loop — for example, a Chinese-built operating system for smartphones, or Instex, the special purpose vehicle launched by France, Germany and the UK to allow companies to do business with Iran beyond the reach of US punishment.

Wolfgang Munchau described Instex as “a dysfunctional insurance vehicle for small carpet traders”; it turns out that these alternatives are not cheap or easy to develop. In ordinary circumstances, few would even bother to try. But if the US presses too painfully on the global economy’s nervous system, its rivals and even its allies will look for relief.


Written for and first published in the Financial Times on 14 June 2019.

My book “Fifty Things That Made the Modern Economy” (UK) / “Fifty Inventions That Shaped The Modern Economy” (US) is out now in paperback – feel free to order online or through your local bookshop.

Free email updates

(You can unsubscribe at any time)

What to do when blessings come well-disguised

Keith Jarrett’s 1975 concert in Cologne should have been a musical catastrophe. Owing to a string of mix-ups and bad luck, he was faced with the choice of attempting his widely admired improvisations on a beaten-up old piano with sticky keys and a harsh upper register — or walking out altogether. He was all for walking out, but felt sorry for the concert promoter and agreed to play the unplayable piano against his better judgment. The result was not a catastrophe but a masterpiece, and a bestselling one at that: The Köln Concert album.

I’m fond of that story, of the way that an obstacle can unleash a creative response — partly by concentrating the mind and partly by forcing the artist to explore fresh approaches. It’s not the only such tale. The great jazz guitarist Django Reinhardt developed his distinctive style after suffering severe burns that left him with only two fully functioning fingers on his fretting hand.

A quotidian parallel is the Tube strike of 2014, which partially shut down the London Underground, prodding several thousand commuters into finding better routes into work.

The three stories share another common element: nobody at the time would have predicted that anything good would come of the problem. Keith Jarrett could have sought out a bad piano; young Django could have tied two fingers together; commuters could have tried a new route any day. But they didn’t, until there was no choice. The blessings came well disguised.

David Epstein’s new book, Range, has brought other instances to my attention. Range is mostly about the benefits of being a generalist rather than a specialist, but a recurring theme is of teaching or training techniques that seem to fall flat but pay off in the long run.

For instance, in the US Air Force Academy, students are initially assigned at random to academic instructors, then later randomly reassigned to follow-on courses — a nice natural experiment. Scott Carrell and James West looked at the data and found snatches of unplayable piano in it: students assigned the “best” instructors for introductory courses, as measured by short-term pupil performance, earned the highest student evaluations but went on to produce the worst results in the long run.

In a nutshell, teachers who “teach to the test”, or otherwise provide simple problem-solving procedures, immediately improve grades and their students thank them for it. But in the long-term, the students suffer from not having been forced to think more deeply.

This isn’t just about pedagogical techniques. One study that stuck with me when I was researching my own book, Messy, was by Katherine Phillips and other psychologists. The researchers set small groups working on a problem before introducing a new team member to help, sometimes a stranger and sometimes a familiar face. The groups forced to work with the stranger were much more likely to solve the problem, but also enjoyed the experience less and sharply underrated their success. Groups of friends did much worse, but had fun and were under the illusion that they had done a good job.

Again the striking thing is not just that the obstacle turned out to be helpful, but that nobody thought so at the time. Only with hindsight did people realise.

Three more examples. Sports scientists now suggest that top endurance athletes should keep most of their training at low intensity, resisting the temptation to train too hard. Traffic planners are familiar with Braess’ Paradox: closing a road can and often does improve traffic flow, even if the number of journeys stays the same. The closure prevents drivers from choosing routes that make sense for them but cause congestion for others.

The inverse is even more common: a pleasing intervention that does no good. For instance, preliminary research suggests that using virtual reality to promote empathy — for example, for refugees, homeless people or those with disabilities — can promote a surge of emotional support in the short run without delivering any long-run insight.

We shouldn’t leap to the conclusion that every blessing comes well disguised, and that every problem is an opportunity. Sometimes blessings are perfectly apparent, and we should remember to count them. Sometimes a problem is simply a problem. But the prevalence of these counterintuitive results is a reminder that we know less about the world — and about our reaction to it — than we like to think.

I draw three lessons. The first is the need to gather solid evidence — for example, by running randomised trials of teaching techniques — and to ensure that the evidence looks at a good range of outcomes over a long enough time to be meaningful. Common sense can lead us astray, especially in situations where short-term pain leads to long-term gain.

The second lesson is to experiment with more variety in our own lives — whether choosing a holiday or a commuting route. It shouldn’t take industrial action to prompt us to try an alternative route to work.

The final lesson is the simplest. When life confronts us with an unplayable piano, perhaps we should sit down and try to play it.

Written for and first published in the Financial Times on 7 June 2019.

My book “Fifty Things That Made the Modern Economy” (UK) / “Fifty Inventions That Shaped The Modern Economy” (US) is out now in paperback – feel free to order online or through your local bookshop.

Free email updates

(You can unsubscribe at any time)


  • 1 Twitter
  • 3 RSS
  • 5 Podcasts
  • 6 Facebook


  • Fifty Inventions That Shaped the Modern Economy
  • Messy
  • The Undercover Economist Strikes Back
  • Adapt
  • Dear Undercover Economist
  • The Logic of Life
  • The Undercover Economist

Free Email Updates

Enter your email address to receive notifications of new articles by email (you can unsubscribe at any time).

Join 188,297 other subscribers.