Rioters convicted after the London riots of 2011 were given much harsher sentences than were previously standard for such crimes. According to research by Laura Jaitman, Brian Bell and Stephen Machin, published in the May 2014 issue of the Economic Journal, this led to a significant drop in riot crimes across the whole of London (not just the riot areas) in the six months after the riots.

Their study concludes that the deterrence effect of tougher sentencing was prolonged by the perception among prospective criminals that the change in the justice system was permanent and not transitory. The intense media coverage of the riot sentencing seems likely to have reinforced the belief of more severe sentencing still taking place.

Can tougher sentencing deter crime? To answer this important question, the new study analyses the aftermath of what happened in London in August 2011, some of the worst riots in the city''s history. The riots took place in highly localised geographical areas, with crime going up by a huge 57% in the affected neighbourhoods. After the riots, large numbers of rioters were convicted and given much harsher sentences than for people convicted of the same kinds of crimes before the riots.

The study finds a significant drop in riot crimes across London in the six months after the riots, consistent with a deterrence effect from the tougher sentencing. More evidence of general deterrence comes from the observation that crime also fell in the post-riot aftermath in areas where rioting did not take place.

Whether tougher sentencing deters crime is a fundamental policy question. But persuasive empirical evidence on the question is hard to find. This is, at least in part, because tougher sentencing does not usually arise randomly but occurs in response to crime increases that develop over time.

The 2011 London riots occurred in ''hot spots'' that were not experiencing differential rises in crime relative to non-riot areas prior to the riots happening. The subsequent harsh sentencing that ensued was not anticipated at the time (especially given the speed with which the criminal justice system responded) and thus offers a credible setting to look for deterrence effects on crime after the riots.

The August 2011 riots were unparalleled in terms of speed, scale and geographical spread. After a fatal shooting by police officers on 4 August, a peaceful protest turned into violent disorder, which in the following days escalated and spread through England (to a number of the larger cities including Liverpool, Manchester and Birmingham).

It is estimated that between 13,000 and 15,000 people were involved in the riots, among them criminals, opportunists and spectators. During the five days of disorder, more than 5,000 crimes were committed. Apart from five fatalities, the offences were mainly burglary, criminal damage and violence against persons.

These riot crimes took place in highly localised geographical areas – in only 1.8% of London sub-wards (the smallest administrative unit), where crimes went up by a huge 57% (from 19 crimes per 1,000 population to 22.7 crimes per 1,000 population) – and did not rise at all in the other London areas.

In the last days of the riots, the police changed the usual dispersal tactics employed in cases of disorder, making immediate arrests and the criminal justice worked intensively to deliver fast and tough sentences. By September 2012 (a year after the riots), 4,600 people were arrested and 2,250 appeared before court for incidents related to the riots.

Rioters who were taken to court were almost three times more likely to be placed in immediate custody than offenders with comparable demographic characteristics who had committed the same offences the previous year. In addition, convicted rioters were sentenced to significantly longer sentences (approximately two months more on average than the 13 months received on average by similar offenders the previous year).

The severities of the sanctions for rioters were widely covered in the media and the perception of the population was that the sanctions imposed were tougher than expected.

Did the tougher sentencing have a deterrent effect on crime? The results show a significant drop in crime in both riot and non-riot areas of London in the six months after the riots. On average, there was a 0.25 points decrease in the overall crime rate in non-riot areas after six months (3% from its pre-riot average level).

The researchers observe a decline in crime even in London areas located far from the riot incidents (more than 3km away or in the outer 5km of London excluding riot locations) and in police force areas of England and Wales that were unaffected by the riots.

Moreover, they find that non-riot crimes actually went in the opposite direction and increased. This suggests a rational response from criminals who seem to have substituted away from the types of crimes that received tougher sentences to those that did not.

This is consistent with the operation of a deterrence effect from tougher sentencing rather than only being the effect of incapacitation of criminals. The researchers find little evidence that spatial crime displacement to other areas or extra police presence on the streets of London in the wake of the riots accounts for these patterns of change.

''Crime Deterrence: Evidence from the London 2011 Riots'' by Laura Jaitman, Brian Bell and Stephen Machin is published in the May 2014 issue of the Economic Journal. Laura Jaitman and Stephen Machin are at University College London. Brian Bell is at the University of Oxford.