Better Jails


My AFR oped today is on prison reform. You can’t put acknowledgements in an opinion piece, but the piece owes a substantial debt to Justin Wolfers, who first suggested the idea of smarter prison contracts about 7 years ago, when we were strolling the streets of San Francisco together. I’m also very grateful to a senior criminologist who took the time to discuss the contents, but who probably doesn’t want to be named here. ANU PhD student Daniel Suryadarma did a great job tracking down all the relevant statistics for me.

Full text over the fold.

Prison Reform Hard Labour, Australian Financial Review, 17 November 2009

It’s a set-piece of every Hollywood movie featuring an ex-con. A clean-shaven man walks out the prison gate carrying his possessions in a box under his arm, and puts his hand up to shade his eyes from the unfamiliar glare of the sun.

In 2009, over 20,000 Australians got a ‘get out of jail’ card, and walked through the gates of one of the 120 or so prisons dotted across the country.

But the sobering fact is that during the next two years, nearly half of them will have been sent back. Some will get to play that prison release scene again and again.

Prisons do reduce crime, but mainly because of what criminologists call ‘the incapacitation effect’ (when you’re doing time in Long Bay, it’s harder to hotwire a car). There may also be some deterrence effect, but this is small by comparison. And there is little evidence of a rehabilitation effect.

Why are prisons less a portal to a new life than a revolving door? Part of the problem lies in the fact that politicians and public servants have few incentives to create better rehabilitation programs. A policymaker who creates a cutting-edge program to improve inmates’ sense of self-worth risks being ridiculed by the shock jocks. A politician who advocates work release programs must worry about the tabloid headlines if some inmates abscond.

To encourage innovation, we should start publicly reporting the outcomes that matter most. Rather than merely telling the public how many people are held in each jail, governments should publish prison-level data on recidivism rates and employment rates.

Just as with schools and hospitals, ‘prison league tables’ would need to take account of each institution’s caseload. Effectively, such comparisons should aim to answer the question: given the mix of inmates they handled, which prisons did the best job? Carefully constructed prison league tables should give taxpayers and policymakers a strong sense of where innovation is occurring in the system, and help smarten up the criminal justice debate. (A similar approach could also be taken in the case of community corrections programs, which handle about twice as many people as jails.)

As well as focusing on the important outcomes, Australian states should rethink the contracts they write with private providers. At present, about 16% of inmates are held in a private jail. Unfortunately, the contracts for private jails bear a remarkable similarity to sheep agistment contracts.

Providers are penalised if inmates harm themselves or others, and rewarded if they do the paperwork correctly. Yet the contracts say nothing about life after release. A private prison operator receives the same remuneration regardless of whether released inmates lead healthy and productive lives, or become serial killers.

A smarter way to run private jails would be to contract for the outcomes that matter most. For example, why not pay bonus payments for every prisoner who holds down a job after release, and does not reoffend? Given the right incentives, private prisons might be able to actually teach the public sector a few lessons on how to run a great rehabilitation program.

According to the Productivity Commission’s annual Report on Government Services, states and territories have formally agreed to ‘Provide program interventions to reduce the risk of re-offending’. Yet the reality is that less than one-third of prisoners are engaged in any formal education program, and that many of the jobs done by inmates do not provide the skills required by the regular labour market.

Given that each prisoner costs the taxpayer about the same as a five-star hotel room, reducing recidivism rates has a direct benefit for all of us.

But criminal justice – and particularly juvenile justice – is also the pointy end of social policy. Ex-prisoners are disproportionately high users of other government services, such as counselling and income support. So improving corrective services helps in other social areas as well. This particularly affects Indigenous Australians, who comprise one-fortieth of the general population, but one-quarter of the prison population.

As a nation where a convict ancestor is a badge of pride, Australians know better than anyone that life can have a second act. The challenge today is to encourage our corrective services to correct, not just punish. Can we make jail work?

Andrew Leigh is an economist in the Research School of Social Sciences at the Australian National University.

A footnote for the data wonks: so far as I can tell, there are no publicly available estimates on prisoner releases from jail. However, we do know the annual number of new entrants, and if the total isn’t rising, the inflows must equal outflows. In 2007 and 2008, “Sentenced Prisoner Receptions into full-time custody” were 27,000 and 29,000 respectively; while the prison population was 27,000 in both years. Although this implies an exit rate closer to 30,000 per year, I use “over 20,000” because of the possibility that some people enter/exit more than once in a given year.

(xposted at

Update: Michael Pascoe in the Age likes the idea (headline: “Shock! Economist Has Really Good Idea”)

12 Responses to "Better Jails"
  1. While we’re at it, governments should also attempt to increase the deterrence effect by taking paid space to disseminate information about the penalties meted out by the courts for serious crimes. The media are very selective in what they report.

  2. I doubt whether further publicising penalties will have much of an effect. Few people who break the law expect that they will get caught, and in many cases their optimism is justified.  If a penalty were really draconian – say, life imprisonment for using a hand-held mobile phone while driving – it might have a deterrent effect. A much stronger deterrent normally is any measure that increases the likelihood of getting caught and convicted.

  3. In contraindication to MikeM’s statement, there is an article in a recent edition of the Economist that documents a higher deterrence effect on repeat crime if a (smaller) penalty is applied immediately following the crime, rather than after a long period involving a trial with longer sentences. An immediate threat of punishment carries greater weight.

    I guess this applies to potential recidivists who are caught in the criminal act.

    Is the fact that only 1/3 of prisoners are involved in formal education due to a a lack of supply or a lack of demand or a mix of both? Exceptions abound, but convicted criminals aren’t generally open to educational possibilities. I don’t see cause and effect demonstrated. Are prisons ineffective at rehabilitation because there is no opportunity for prisoner improvement or because of the characteristcs of the prison population itself?

  4. Restructing the private prison contracts around the outcomes we actually want – it could be as simple as recidivism rates – is a damn good idea.  In fact I’d go so far as to say it’s a no-brainer.

  5. If prisons are rewarded on the basis of non-recidivism – say two years without a conviction following release – then will the delayed reward be enough incentive to spend the capital today?
    Sounds like a similar incentive problem for making sure the CEO doesn’t game the salary outcome on the basis of achieving short-term objectives over long-term corporate need. Of course, CEOs are not repeat offenders surely 😉

  6. I think the debate needs to be more alive to the subtleties of the situation. What follows are some observations based on 10 years or so interation with criminals, criminologists and social psychologists.

    Only a very small proportion of people who commit crimes go to jail (most crimes are not reported to police; there is not enough admissable evidence to take a high proportion of those that are reported to court, and then a good many are acquitted).  one back of the envelope calculation I made (stats are not good in this area) suggest that less than 5% of crimes are formally punished in any way.

    Magistrates and judges are aware of the poor rehabilitation record of jails, and mostly try to avoid sending people there without really good cause (one is to trigger the incapacitation effect).

    There is a good deal of evidence that crime polarises around two ends of the spectrum – most single offences are done by the young and reckless, who mostly grow out of offending by 25 or so, and so are best deterred or corrected by immediate small-scale action (telling their mums often works).

    The other end is that a large proportion of instances of crime are committed by people for whom crime is a way of life, and jail is just one of the things you cope with (eg around 90% of burglaries are done by 10% of offenders). It has the plus side that you get to make new acquaintances and learn a few more tricks of the trade – a bit like a boring version of an academic convention.

    These people are really hard to change – the best track records come from transportation to some place you actually have to do honest work to survive, or putting them in the army. Neither is cheap. A new contract with a jail will do little unless you can follow up beyond jail with some drastic intervention.

    There are some promising alternatives – eg interventions to keep teenagers from getting on the track to becoming career criminals in the first place. Again, not cheap. 

    Not saying don’t try it – just don’t expect too much of it.

  7. The point of the contract would be to give the private prison providers a financial incentive in finding the alternatives and methods that work.

  8. “Governments should publish prison-level data on recidivism rates and employment rates.” I couldn’t agree more Andrew and have been arguing this for many years – which leads me to a radical  approach to parole decision.s
    Start with the observation that we do not lock up our entire population to reduce the crime rate. There is a chance that any person will commit a crime, but the baseline risk is not sufficient for pre-emptive incarceration. The flip side is that once you commit a crime you reveal yourself as high risk and so you should be incarcerated. I think it follows that, unless evidence can be produced that your chance of committing another crime has fallen to baseline, you should not be released – EVER.  (Let’s narrow the discussion to violent crime to avoid the valid point that prison may be more costly than letting petty burglars continue to burgle.)
    Getting such evidence to assess recidivist risk is not easy. One approach would be to employ criminologists and psychologists to make the call (perhaps subject to committee veto). Prison league tables would be part of the assessment but also individual level assessment (obviously). The economists amongst us would then need to design a monetary payoff table to reward or punish the assessor for their correct and incorrect assessments.
    OK. It is never going to happen I know. But I do resent the fact that habitual criminals are released into the community, simple because they have served their time. Punishment is a necessary but not sufficient part of a fair system.

  9. Plans like this should also be wary of unintended consequences.  When profit is associated with a particular outcome, there is motivation not just to achieve the outcome but also to define the outcome, and to influence the measurement of the outcome.
    In the case of criminal recidivism the relevant measurements are made by the police and the justice system – any risk of encouraging the application of commercial influence to these organs of society should be given sober consideration.

%d bloggers like this: