Undercover Economist

When crime stops paying

To an economist, tougher sentencing in the wake of the 2011 riots offers a fascinating natural experiment

The third anniversary of the 2011 London riots is this week. They erupted so suddenly and spread so quickly across the capital and to other English cities that at the time the disintegration of British society seemed, if unlikely, at least conceivable. In the rear-view mirror, though, the riots are eclipsed by the London Olympics and much diminished by the passage of time.

For parochial reasons, the riots remain vivid to me. My son was born in Hackney just a few days before they started. As violence flared a couple of streets away to the south and to the north of us, my wife and son slept while I stood on the doorstep of our home and watched as a pair of helicopters droned directly overhead.

A year after the riots I wrote a column pointing out that they were essentially random events. They had a cause, of course. The spark was the shooting of Mark Duggan by the Metropolitan Police, and one source of fuel was the perception that police stop-and-search powers were being used crassly and with a racial bias. Yet similar grievances have emerged at other times and in other places without provoking mass civil unrest. Chance plays a major element in such stories.

The criminal justice system responded sharply to the riots. More than 1,000 suspected rioters were charged by the Metropolitan Police during the first week of trouble, and over the same time period more than 800 of them made a first appearance in court. By September 2012, 4,600 people had been arrested, out of about 13,000-15,000 people who are believed to have participated in the trouble in some way. Given the initial sense of impunity, that is a high rate of unwelcome police attention.

More striking was the way in which judges handed out sentences as though they were on steroids. Two people were sentenced to four years in prison each for Facebook postings inviting others to run amok in Cheshire, an unlikely location for a revolutionary uprising. Nobody showed up to “smash dwn in Northwich town” or “riot in Latchford”, so the sentences raised eyebrows. So did the 10-month sentence handed out to a teenager who carried two left-footed trainers out of a shop in Wolverhampton. She thought better of it and immediately dropped them – surely one of the most short-lived thefts in history. Sentences were, in general, more severe than normal. The thinking behind all this was that the true crime that needed to be punished was not theft or incitement but participation in a moment of grave civil peril.

Were these sentences an essential crisis response or a draconian overreaction? To an economist, they are something else: a fascinating natural experiment. With the news full of crushing punishments, it must have seemed plausible that the risks of committing a crime had soared. So did the threat of harsh punishments deter crime?

The usual statistical problem is that sentencing policy might influence crime rates but crime rates might equally influence sentencing policy. Cause and effect are hard to disentangle. In the case of the riots, however, the surge in crime that provoked the crackdown was sudden, unexpected, highly localised and brief. The sentencing response was drawn-out and stories of harsh sentences appeared in the national and London press for months.

. . .

As a result, a mugger or burglar in an area of London entirely unaffected by the riots might still feel conscious that the mood of the judiciary had changed. Three economists, Brian Bell, Laura Jaitman and Stephen Machin, used this sudden change in the judicial wind to measure the impact of tough sentences on crime. Across London, they found a significant drop in “riot crimes” – burglary, criminal damage and violence against the person – over the six months following the riots. Meanwhile, other crimes showed a tendency to increase, as though criminals were substituting away from the “expensive” crimes and towards the “cheaper” ones.

This shouldn’t be too much of a surprise. (I wrote an entire book, The Logic of Life, arguing that the most unlikely people in the most unlikely circumstances turn out to be greatly influenced by simple incentives.) But it’s a useful result because rigorous evidence on such matters is hard to find.

One of my favourite exceptions is an article by two economists, Jonathan Klick and Alex Tabarrok, who examined the impact of periodic terrorism alerts in Washington DC in the couple of years following the attacks of September 11 2001. Whenever alert levels were raised, police officers flooded sensitive locations, most of which (such as the White House and the Capitol) are on or near the National Mall.

Over the 16 months studied, the Mall and surrounding district experienced about 8,500 crimes, often theft from or of cars, not really al-Qaeda territory. Klick and Tabarrok argued that the occasional surges in police numbers were not caused by car thefts but did successfully deter them.

There may well be cheaper, more effective and more humane ways to reduce the crime rate.

But such studies have helped to build confidence that the world isn’t an entirely irrational place. Raise the costs of crime and criminals will respond.

Also published at ft.com.