Wrong question. Wrong answer

Within that, the question of the role and purpose of prison has loomed large. Is prison there to deter potential criminals, or punish those who are caught? Should prisons reform and rehabilitate prisoners, or merely contain those who might otherwise endanger the public? Should prison be used sparingly, or enthusiastically?

For most of the twentieth century, the argument in favour of a sparing use of prison, with regimes focused on rehabilitation, attracted widespread support from across the political spectrum. And for good reason. On moral and humanitarian grounds there is widespread agreement that imprisonment has a profound impact on the inmate, to say nothing or his or her friends and family. For some this is reason enough to use prison sparingly or, in the favoured phrase, 'as a last resort'. On pragmatic grounds reformers will often point out the sheer ineffectiveness of imprisonment as a means of reform and rehabilitation. The courts reconvict more than half of all prisoners, and a much higher percentage of young offenders, of a new offence within two years of release. The true recidivism rate is almost certainly higher given that many reoffences must go undetected, unprosecuted and unconvicted. Many consider the basic fact of high reconvictions and reoffending the most compelling argument for a prison-lite policy. Moreover, though some victims of crime do gain a degree of comfort from seeing a convicted offender locked up, imprisonment is often far from the key demand made by victims, regardless of what some advocates of an expanded prison system would have us believe. Additionally, on financial grounds, the sheer cost of imprisonment raises important questions about the scale of social investment being made in a system that is so seemingly ineffective. It costs more to send someone to jail than to Yale, as the American saying would have it. In the UK a year in Feltham works out at roughly twice the price of a year at Eton.

By contrast, arguing in favour of an expansion of imprisonment has, until comparatively recently, been considered a rather disreputable occupation, somewhat akin to supporting vivisection, eugenics or nuclear power. An inquiry into alternatives to prison sentences by the House of Commons Home Affairs Committee in 1998 'emphatically reject[ed]' the suggestion from one set of witnesses that a prison population of 200,000 was needed to make a meaningful impact on crime rates. The Committee also argued that the 'huge rise in the prison population during the last five years is unsustainable. Unless halted--and in due course reversed-- it will end badly'.

Since the report's publication the prison population in England and Wales has grown by around 10,000. At the time of writing it has topped 76,000. The government now talks of 'stabilising' the prison population at 80,000, hardly an ambitious target for a supposedly progressive government keen also to claim that crime has fallen dramatically on its watch.

Politicians from all the main parties are undoubtedly more sanguine about these higher incarceration rates. The Conservatives went into the recent General Election pledging to increase prison places to some 100,000. In its manifesto Labour trumpeted the building of more than 16,000 prison places since 1997 as a sign of its commitment to tackling crime. The Liberal Democrats preferred to emphasise the importance of making prison work, largely ignoring the question of prison numbers.

The causes of this shift in the political mood music are complex, and arguably not that interesting given that the prison population has tended to increase regardless of which party has been in government. In 1945 the prison population stood at just under 15,000, less than it had been in 1900 and only slightly more than the average population during the 1930s. By the early 1960s it had doubled. It had doubled again by the mid-1990s. Current Home Office projections suggest that by the end of the decade the prison population could be close to 90,000. Some commentators consider this an underestimate.

One factor at work has been the perceived success of prison in bringing down crime rates. Between 1993 and 2003 the prison population in England and Wales rose by nearly 30,000, from some 44,500 to around 73,600. Crime measured by the British Crime Survey (BCS) fell from 18.5 million offences in 1993 to 11.7 million by 2003/04. Put another way, BCS crime fell by around a third in England and Wales at the same time as the prison population increased by two-thirds. The United States showed similar trends of falling official crime rates and burgeoning prison populations.

Such a correlation of rising imprisonment and falling crime makes a compelling case for the greater use of imprisonment according to the authors of Civitas' recent book, Crime and Civil Society. 'More than one factor was at work,' in England and Wales during the 1990s they acknowledge, 'but it is impossible to argue that incapacitating an additional 20,000 offenders on average per year had no effect at all on crime'. The argument in Crime and Civil Society is a detailed one, sometimes overly so. It also contains a number of overlapping arguments: about human nature and why people break, or abide by, the law; about the relationship between crime and imprisonment rates; about the effectiveness of prisons, community sentences and fines; about the role of rehabilitation and resettlement; about the family, schools, religious institutions and community organisations. The authors could have made their case in half the space. Frankly the book could have done with a good edit.

Strip the detail away however and the core argument of the book is a simple one. It can be stated in three propositions:

The crime rate in England and Wales is high in historic terms, and compared with other countries. By contrast, our imprisonment rate is low. There is a strong correlation between crime rates and imprisonment rates. When the risk of imprisonment has fallen the rate of crime has increased. Conversely, when the risk of imprisonment has increased, the rate of crime has fallen. There is therefore a good crime reduction argument for increasing the number of people we imprison. Prison deters people from committing crime and by containing convicted criminals it prevents crime. Given how many crimes are prevented by prison it represents good value for money.

Crime rates and imprisonment rates

Reformers often argue that the UK's imprisonment rate is high in comparison with other EU countries. In 2000 the UK imprisoned some 121 people per 100,000 of the population compared with an EU average of 87. The recent accession of some former Eastern Bloc countries - a number of whom have comparatively high incarceration rates - has muddied this picture somewhat, but according to Civitas the comparison is in any case misleading. The correct measure should be imprisonment as a proportion of crime recorded, not the population of the country. After all, a 'country with a high level of crime would expect to put more people in jail'. On this basis the authors argue that England and Wales is lenient in its use of prison:

'[I]f we compare the number of prisoners to the number of recorded crimes, the EU average was 17.7 and the figure for England and Wales was 12.7. In fact, 9 out of 15 EU countries had rates of imprisonment for every 1,000 crimes that were the same or higher. France was higher, Spain much higher and Germany the same.'

This is a seductive, though misleading argument.

Measuring the prison population as a proportion of the general population is in some ways a rather crude approach. It tells us little about other factors that might have an impact on prison numbers, such as different levels of crime, the efficiency of the police in catching criminals, the severity of sentencing, political priorities and so on. Thus the British criminologist Ken Pease argued some years ago that such a measure was 'useless' for assessing the relative punitiveness of different countries. It would be informative, he argued, only 'if people are imprisoned at random'.

However, is Civitas' preferred comparison with police recorded crime rates any more useful? Every year the police record a whole range of minor and petty offences that would not usually end up with an individual, if convicted, ending up in prison. So at the very least a more accurate analysis would involve comparing prison numbers with recorded offences likely, on conviction, to lead to a prison sentence. Around one in five police recorded offences in England and Wales in 2003/04 for instance, were for criminal damage. Yet most convictions for criminal damage do not result in a prison sentence, and nor should they.

Police recorded crime rates are also notoriously unreliable as a measure of the true level of crime in society. Drawing strong conclusions about crime levels from those the public could be bothered to report to the police, or the police chose to detect, is akin to drawing conclusions about the health of a neighbourhood from a visit to the local accident and emergency department. Once one attempts cross-national comparisons of recorded crime rates and prison numbers, the problems increase. Different jurisdictions operate to different criminal codes; police in different countries record crimes in different ways; different authorities can be responsible for recording crime in different countries. All of this means that cross national comparisons of police recorded crime and incarceration rates are highly problematic. Drawing conclusions about how lenient or punitive a country is on the basis of them is rather dubious.

More imprisonment means less crime

'I know what causes crime,' Michael Howard once remarked, 'criminals!'. Lock up more criminals and you reduce their opportunities to commit crime, so the argument goes. The tough stance also deters other would-be offenders from committing crime out of fear of punishment. As Howard famously told the 1993 Conservative Party Conference: 'Prison works. It ensures that we are protected from murderers, muggers and rapists - and it makes many who are tempted to commit crime think twice.' Prison, in other words, reduces crime because it contains criminals and deters potential criminals.

The US experience, where official crime rates fell while imprisonment rates grew enormously, has proved remarkably influential for advocates of this view. Unsurprisingly reformers keen to see a reduction of the use of prison in this country have tended to downplay the significance the US prison experience and discredit those who argue that the US has something to teach the UK. This is a shame because some recent US developments are starting to highlight the problems with the 'more prison equals less crime' argument.

California - which with some 160,000 inmates in 2002 had a prison population around double the entire UK - has started to use drug treatment in place of prison for tens of thousands of users and petty dealers who would otherwise have ended up in prison. The state with the highest incarceration rate, Louisiana, was one of the first to reduce the number of mandatory minimum sentences for a range of non-violent offences. As Michael Jacobson points out in Downsizing Prison, '[f]or a state considered by many to be the most conservative in the country in terms of punishment to be the first to repeal or revise... some of the harsh legislation passed in the prior decade and a half sent a powerful message... that serious sentencing reform was possible'.

Jacobson is far from the first US writer to critique America's love affair with prison, and challenge the assumption that high prison rates are behind the falls in crime. Robert Reiman's The Rich Get Richer and the Poor Get Prison has been a standard text for over twenty years. More recent works by writers such as Joel Dyer and Christian Parenti have catalogued and exposed the massive dysfunction and profound injustices caused by a system that uses the mass incarceration of poor, young, disproportionately black, urban males as a mainstream social policy intervention. Moreover, US advocates for the nation's high imprisonment rate are increasingly few and far between. John DiIulio, for instance, who was formerly a key ideologue for prison growth has more recently argued that the US should aim for zero prison growth in the short-term and reductions in the longer term.

What makes Downsizing Prisons more than just another book about prison reform is that the author was the New York City prisons commissioner under the hardline Mayor Rudy Giuliani. Jacobson is therefore difficult to write off as just another woolly liberal. His analysis of what went on in New York's jail and prison system while the city's official crime rate was plummeting offers a significant challenge to those who argue prison reduces crime.

In 1980 New York City's daily jail population stood at around 12,000. In little over a decade it doubled to around 24,000. Then a strange thing started to happen. During the years of Giuliani's tough on crime administration the City's jail population fell, and fell quite dramatically. By 2005 the daily jail population had fallen to 13,000, back almost to its 1980 levels. The numbers New York City annually sent to prison also declined, from around 20,000 in the early 1990s to 8,000 today. This dramatic decline coincided with New York's well-publicised crime drop. Between 1993 and 2003 known homicide dropped by 69 percent and recorded violent crime by 64 percent. 'Thus,' Jacobson notes, 'the largest and most well publicized crime decrease in the country was accomplished with not an increase but a decrease in the use of prison'.

New York City was not alone in experiencing a fall both in crime and imprisonment rates. San Diego likewise experienced falls in both prison usage and its crime rates. In contrast to New York, it also adopted a more community-oriented style of policing, unlike New York's macho approach.

Nationally, no single pattern emerges from the US experience. Idaho led the way in penal expansion, increasing inmates in its state prisons by 175 percent between 1992 and 2002. In the same period violent crime grew by 14 percent. Montana increased its imprisonment rate by 113 percent and experienced a whopping 128 increase in violent crime. Colorado increased its imprisonment rate by a similar amount, but saw a 21 percent fall in violent crime. At 3 percent Massachussetts increased its imprisonment rate the least of all the states. Its violent crime rate dropped by 33 percent.

Some caution over all these figures is necessary. Because Jacobson is interested in comparing city and state crime and imprisonment rates he has to rely on recorded crime figures rather than the US crime victimisation survey, which most criminologists agree offers a more accurate indication of the crimes it measures. Moreover, Jacobson's analysis does not prove that prison has no impact on official crime rates. Nor does he claim this. But at the very least Downsizing Prisons does show that New York did not imprison itself to lower official crime rates. More broadly, it challenges a simple-minded assumption that more prison means less crime.

Regardless of the US lessons, perhaps the UK did imprison its way to lower crime rates during the 1990s, as Civitas argues? Police recorded crime rates fell from 1993 until the late 1990s; the BCS has been on a long-term downward trend since 1995. During this same period imprisonment rose significantly.

Finding counter-examples is not hard. The numbers in prison doubled between 1951 and 1971 for instance. During the same period recorded crime trebled. Recorded crime trebled again between 1971 and 1991 while prison numbers remained largely unchanged. Did the overuse of prison 'cause' crime between 1951 and 1971, or was prison not being used enough? Were we impossibly liberal between 1971 and 1991 in the face or rising crime, or did we, rightly, decide that crime would rise, regardless of how many people we locked up?

It is easy to pick a particular run of figures from a particular period in history - whether recent or more distant - to make an argument. But using historical data in this way can give an argument a false certainty. Drawing conclusions from recent history is particularly problematic, not least of all because in general people of any generation like to think that they live in historically interesting and significant times. Evidence alone is unlikely to resolve the question of the relationship between imprisonment rates and crime rates precisely because the evidence does not speak for itself. And this is not just because official crime rates - be they police recorded crime or BCS - offer such an inadequate measure of real crime levels. Evidence needs to be interpreted and placed in context. Careful judgements need to be made which are as much ideological as they are strictly criminological.

The cost-effectiveness of prison

The nature of the ideological judgements being made by Civitas become clearer on examination of their cost-effectiveness argument over the use of prison.

There is widespread agreement that prison costs a fair amount of money. The baseline 2005-06 budget for the public prison service is around £2.6 billion. Given this, and the high reconviction rates, liberal reformers and fiscal conservatives alike argue that, as a former Home Secretary under Margaret Thatcher once remarked, prisons are an expensive way of making bad people worse.

For Civitas such arguments are entirely wrong-headed. 'The pattern of repeat offending is not the result of prison,' the authors, 'but rather the consequence of a failure to imprison habitual offenders for long enough to protect the public'. Moreover, rather than being an expensive luxury, prisons are something of a bargain because they contain repeat offenders who might otherwise commit high volumes of socially and economically costly crimes. The correct policy response is to lock up far more people, and for longer, the authors contend, if we are to have an impact on crime.

A fairly wobbly foundation supports Civitas' argument, so a brief analysis will suffice. Civitas first draws on the discredited argument that 100,000 persistent offenders are responsible for half of all crime. As only 20,000 of these are apparently in jail at any one time, Civitas concludes 'that the Government's first priority should be to incarcerate the remaining 80,000'.

An analysis by the Crime and Society Foundation pointed out last year that the 100,000 persistent offenders estimate is based on conviction records, not overall offending patterns. So it is plainly false to claim that such a group, if indeed it exists, is responsible for half of all crime. In reality, the group does not exist. A Home Office study from earlier this year estimated that there were more than half a million serious and prolific offenders in the general population, and even this figure is probably an underestimate. The fact is that crime and criminality is far more widespread and common than the reassuring myths of a small criminal hardcore would have us believe, and this in itself raises questions about how much impact imprisonment would have.

What of the cost that would be incurred if Civitas' recommendations were followed. The authors suggest that the significant prison building programme needed to incarcerate the additional 80,000 so-called persistent offenders could be managed at a rate of 5,000 places per year, implying a penal expansion programme lasting 16 years. The capital costs of this prison building programme they estimate at between £302.5 million and £525 million per 5,000 spaces. Annual running costs per 5,000 they estimate at £194 million.

Taking the more conservative capital cost estimate of £302.5 million first, the annual cost incurred during year one of the 16 year programme would be £302.5 million, rising to £496.5 million in year two once the £194 million running costs are factored in (I assume here that running kick in during year two of the prison building programme). By year 10 the annual outlay (capital costs plus cumulative running costs) stands at around £1.8 billion. The cost in year 16, the final year of the building programme, is £3.2 billion. Total expenditure over the entire 16 years period would be £28.1 billion.

Assuming the higher capital cost estimate of £525 million, the total expenditure at the end of 16 years will have been nearly £31.7 billion. Costs from year 17 would decrease with the end of the building programme, but annual running costs of just under £3 billion would be necessary to maintain the expanded part of the prisons estate. Further capital outlays on new construction as the ageing prison estate reaches the end of its useful life would also need to be factored in.

These are big sums of money, and certainly an underestimate of the real costs of such a programme. More prisoners would also require more probation officers to supervise them on release; more courts to process them; more police officers to catch them. But this would all be money well spent in Civitas' view because the much higher economic and social costs of the crimes committed by persistent offenders. Far from being an extravagant waste of money, prison is something of a bargain.

This is a dismal vision - of human nature, of society, and of our capacity to do anything meaningful at a public policy level to change things for the better. The ideological agenda at work is pretty transparent. Crime is presented as the problem to which prison is the solution, and this despite the authors' acknowledgement that various factors influence crime rates and that '[t]he part played by the criminal justice system is always of lesser importance than grassroots efforts to build and maintain a society of conscientious citizens'.

The social and economic costs of such a programme being adopted would also be significant. Budgets are not limitless. The decision to devote more resources to imprisonment and associated criminal justice interventions also means deciding not to spend that money in other ways. This is likely to include expenditure on those services necessary for a more socially just society in which crime and a broader range of harms were less prevalent. As Michael Jacobson points out, the use of mass incarceration in the United States has carried with it many costs, not least of all for other important government services:

'[T]he tremendous growth of incarceration over the last few decades relates directly to this country's present difficulties in adequately funding its education and health care systems. Money spent on spiralling corrections costs... came at the expense of other crucial governmental services'.

The focus on criminal justice interventions to reduce crime in any case distracts from the far more signficant social and economic drivers of crime and vulnerability. One likely impact of the increased gap between rich and poor since the late 1970s, for instance, has been the increased vulnerability to crime and harm among the poorest in society. Research for the Joseph Rowntree Foundation published in 2000 suggested that 7,500 people would not die each year were the government to implement a modest redistribution of wealth from rich to poor. An analysis shortly to be published by the Crime and Society Foundation indicates that a significant redistribution of risk of homicide has taken place since the early 1980s: the richest are less likely to be murdered now; the poorest far more likely.

These questions of crime and vulnerability, and their relationship to socio-economic status and government policy are not only ignored by Civitas' book, they are systematically excluded. Rather like the man with the hammer, who sees every problem as a nail, those who start by seeking a prisons solution to the problem of crime tend to ignore some of the most interesting, and important, questions.