Response to Garside’s Police, Prisons and Crime Rates

nick cowen, 12 July 2012

The most interesting and strident response to our recent crime analysis so far has come from Richard Garside. He claims that our report is ‘a rather flaky argument’ that is ‘built on hopelessly shaky foundations’. While he picks up on some important points regarding how to interpret our results, I don’t think they amount to a fundamental refutation of the paper. I have two points responding to his criticisms and one point showing that our results actually agree with some of Garside’s own concerns.

Is Recorded Crime a useless indicator?

Garside’s most developed claim is that using recorded crime data ‘to assess levels of crime in an area is akin to assessing the health problems of a neighbourhood through a Friday night visit to the local A & E department.’ He argues that our results are distorted by this selection process and that, in the end, all we are measuring is relationships between different bureaucratic processes, not the actual experience of crime in Britain.

Recorded crime data is collected by people reporting crimes to the police in their local area. Recorded crime counts will thus depend on recording conventions, including how crime is defined and labelled, what rules the police have on recording crime and whether those rules are followed consistently. Recorded crime can be treated as a relative measure for comparing times and places rather than as an absolute measure. But that won’t do much good in the end if there is no real relationship between recorded crime and overall crime prevalence either between districts or over time.

Fortunately, there is a way of checking the validity of some recorded crime data by comparing it to trends in the British Crime Survey. The BCS asks random members of the public if they have been a victim of crime in the last year (it is conducted independently from police reports). We can look at a subset of one of our crime types, domestic burglary (source: p. 75 of this edition of Crime in England and Wales). Domestic burglary is recorded as such by police and is also estimated by the BCS:

It’s not quite a case of ‘did they ever fit together’ but there is a family resemblance. In fact, it appears that around half of BCS estimated domestic burglaries is reported to the police. These two trends were generated using entirely independent methods but tell a similar story. The BCS, narrow in its remit as it is, uses a valid method for detecting likely trends. So it is plausible to suggest that recorded crime, at least with respect to domestic burglary, is tracking a real social trend too. This means that variables that measure criminal justice activity (like detection and sentencing) that are observed to affect recorded crime rates are also likely to be affecting the real underlying experience of crime.

The case might be harder to make for some of the other crime types, not least because the BCS and police records often itemise crimes subtly differently. The BCS, because it is based on a random sample of a few thousand people, does not always produce stable estimates of relatively occasional crimes. So we haven’t assumed recorded crime is a perfect indicator, only that there is a prima facie case for saying that there is a relationship between recorded crime and real crime for the crime types we are covering.

Do we assume that people are rational calculators?

Garside’s other main claim of substance is more theoretically motivated, that the paper relies on ‘a naive view of human nature, peculiar to economists, that individuals are largely rational calculators who calmly weigh up the benefits and disbenefits of any given action’. This is not strictly the case. All the paper claims is that people respond to incentives. Why they respond to incentives is an open question that the paper does not address. The methodology of economic analysis does not rely on rational calculation as the actual mechanism for action. It only claims that the assumption of rationality often yields predictable results. As Chris Dillow sometimes points out, people often behave as if they were rational, even they aren’t actually thinking rationally.

When Garside tries to ridicule this approach by saying it assumes that ‘were it not for the threat of imprisonment I would happily mug my fellow passengers on the train home’, he is piling up a straw man. For the majority of people, the fact that stealing other people’s property is morally wrong is reason enough not to commit the crime. However, when looking at offenders and potential offenders, we are dealing with people who have a tendency to make morally bad decisions. For them, a greater prospect of punishment can discourage them from offending.

Offenders don’t need to examine crime statistics like we do to see what their chances are of being caught. They might just notice other criminals, a fence who handles their stolen goods, for example, being caught and punished and conclude that criminal acts are a bit more risky than before. Some will offend less frequently while others might slowly shift into more civilised patterns of behaviour.

Where I agree with Garside

In the past, Richard Garside has argued against a policy focus on ‘persistent offenders’ because they are essentially defined by the focus of currently recorded crime, rather than crime more broadly conceptualised. Arguably, these persistent offenders would look like less of a priority if we could more accurately record things like domestic violence and abuse, cybercrime and high-level corruption:

offences the BCS or police recorded crime either do not cover at all, or count inadequately, include benefit fraud, whitecollar and corporate crime and environmental crime.

I think the policy focus on street crime is somewhat justified by public concern. On a subjective level, we probably feel more worried about a street mugging, or a house burglary, than being subject to an elaborate business fraud. This is not necessarily a rational perspective but it is one that policymakers may respond to in a democracy. People may feel more violated by crimes that harm our person and our personal possessions compared to crimes that simply disadvantage us financially.

However, our results are consistent with Garside’s critique in one respect, in so far as they indicate that fraud (which is more of a ‘white collar’ crime) is consistently reactive to longer sentences, with longer sentences associated with fewer recorded frauds. This might suggest that we are some way away from punishing fraud at an optimal level of severity, although we could only know for sure by looking at different, more expansive, measures of fraud. In other words, perhaps we are too soft on fraudsters compared to other types of property criminal at the moment. So Garside may have made an important point in the past that our analysis confirms. The criminal justice system (and British society) could be made fairer and safer were we able to catch, convict and punish more white-collar criminals. I think everyone can think of a few white-collar sectors that might benefit from a greater criminal justice focus!