Matt Sinclair: What does the empirical evidence tell us about the causes of riots?

By Matthew Sinclair of The TaxPayers' Alliance.

There have been a lot of calls to understand the causes of the riots. Unfortunately most of the attempts thus far haven't been grounded in what empirical evidence we have. Mary Riddell for example, says that we shouldn't blame unemployment, cuts or race but poverty and inequality. She talks about the rising wealth of the richest and how "successive British governments have colluded in incubating the poverty, the inequality and the inhumanity now exacerbated by financial turmoil."  Riddell's argument is passionate and powerful rhetoric, but the best study we have on what causes riots suggests almost exactly the opposite.

In 1996, after the Los Angeles riots in 1992, Denise DiPasquale and Ed Glaeser produced a landmark study of the causes of riots, The Los Angeles Riot and the Economics of Urban Unrest. Of course one paper isn't the end the debate, and there is other research, but their extremely high quality study has to be the starting point in considering what has gone wrong.

The evidence suggests that relative poverty doesn't cause riots. The authors report that: "Our cross-national and the 1960s evidence provided little support for the popular notion that poverty is a major determinant of which cities riot."  And contrary to the idea insufficient public spending is the issue: "The size of government is also somewhat positively correlated with rioting, perhaps because community level gains from rioting are higher when there is a greater amount of government expenditures to divide."

There are some qualities that make London inherently susceptible to riots. Urbanisation and ethnic heterogeneity are both connected to rioting.  Given that London will continue to be a large and ethnically diverse city, we will always need to be careful about the potential for these kinds of disturbances. We can't do a lot about that.

There are two other factors that we can do something about.

First, economic conditions. While relative poverty doesn't appear to be important, unemployment does. Idle hands really are the devil's plaything. That means the worst thing to do is attempt to subsidise relative poverty out of existence with generous benefits, while not reforming welfare and increasing non-wage labour costs. That combination ruins the incentive to work and creates the conditions in which riots are more likely. The results were clear even before the recession. In July 2008, the OECD reported that in 2007, the unemployment rate for 16 to 24 year olds was 14.4 per cent in Britain, up from 11.0 per cent in 2002. In the wider OECD that rate went down from 14.6 per cent to 13.4 per cent.

We need to look at a combination of benefit and tax reform in order to reduce the wedge between the amount paid by employers and the resulting opportunities for workers. Higher national income also helps, which is one more reason why supply-side reforms would be a good idea.

The second factor is the probability and extent of punishment:

"Our results support a basic neoclassical view that probability and size of punishment drive some portion of rioting behaviour.  The dictatorship results from the cross-national data, the police expenditures results from the cross-city data, and the south dummy all suggest than an increase in the probability of arrest lessens the probability and size of riots."

If people don't think they will be caught, or don't think they will be seriously punished, then they are much more likely to to riot and loot. It is pretty clear from the media coverage that lots of the looters were enjoying themselves, and there needed to be consequences to put them off. That is why reports local authorities would kick rioters out of council houses were a helpful deterrent, for example.

I'm not going to get into the debate over how we best catch criminals and enforce the law in the kind of circumstances the police have faced so bravely in recent nights. Various commentators have made powerful arguments that more or less CCTV would improve things. Others have said we need to spend more on law enforcement - or we need more prison places - but they'll need to set out what else they would cut to maintain the fiscal adjustment, and we should look to reduce waste in police forces.

The explanations for why our fellow citizens are doing this that best fit the data are pretty simple: Rioters don't have anything better to do and think they can get away with it. Low opportunity costs and low risks. In lieu of a better explanation, or better data, those are the problems to address.