The Guardian, July 2003

There are a lot of chief constables who would happily strangle George Dixon. It's not so much that the old BBC copper with his folksy winking ways makes any real officer look inadequate, nor even that he had the infuriating advantage of scriptwriters to deliver his perfect results. The real problem is that Dixon is cemented into the public imagination - and he's not very good at his job.

Consider the case of the car park in the country. Ramblers used to turn up there, park their cars, head off on a hike and then return to find their car windows smashed and their belongings stolen. Dixon's answer would have been to get out his handcuffs, hide behind a bush until the thieves turned up, nick them and wobble off home on his bicycle. In the real world, another thief would have been back breaking into the cars before he reached Dock Green and the offenders he arrested would have been bailed to join in again the next day.

For decades, British police would have been happy enough with that failure: they had no option. But in the last 20 years, researchers in the United States and Britain have been urging them to think in a radically different way - to stop relying on the conventional tools of arrest and conviction and to start looking beneath the surface to manipulate the causes of crime, to stop being dragged around by events in favour of stepping back and trying to change them. They call it problem-oriented policing. More and more chief constables can see that when it works, it is a great deal more effective than the esteemed Dixon.

In the case of the car park, instead of reaching for their handcuffs, the local police stopped and analysed the crime. They realised that the real problem was simply that, once the ramblers hit the hills, there was nothing and nobody to stop the thieves doing exactly as they pleased, so, instead of making pointless arrests, they decided to solve the problem at its root - by building picnic tables at the side of the car park. To make sure that picnickers turned up, they arranged for a licenced vendor to sell drink and food up there. Natural surveillance. They cut the thieving by 48% in a year.

In the same way, there was a city force who had a problem with a take-away pizza parlour. Every Friday and Saturday night, groups of drunken people would gather outside, obstructing the pavement, disturbing the locals, occasionally fighting. The old-fashioned answer would have been to send a uniformed officer to stand there and arrest anybody who broke the law - inefficient and ultimately ineffective, because after a week or two, the officer would have had to be deployed elsewhere. Instead, they analysed the problem and discovered that the crowd was building up because the pizza parlour could not cope with the sudden rush of customers after the pubs closed, so they installed hot lines from all the local pubs to the pizza maker, who took their orders in advance and had their pizzas ready when they turned up: no more crowd on the pavement, no more trouble.

This problem-oriented approach is the heart of a far wider effort to bury George Dixon: to say that arrests and convictions have their uses but they also have their limits; to break out of the boundaries of the criminal justice system by pushing resources and energy into working with other agencies and delivering something other than punishment; to attack the infinitely complex roots of crime with infinitely flexible tactics. It's not so much criminal justice as crime reduction.

It is the simple equivalent of doctors admitting that they can't cure cholera with medicine alone: somebody has to find the source of the diseased water and clean it. It means reversing John Major's simple edict on crime that "society should understand a little less and condemn a little more" and working at three levels: to look at the deep reasons why some adolescents devote their lives to crime while a sibling in the same family will not, to deal with family, schooling, housing, health; to see the opportunities for crime as problems which can be solved (like the rural car park and the pizza parlour); to make maximum use of the 'target hardening' techniques of traditional crime prevention. It means giving George Dixon a brain as well as truncheon.

This is the story of the boldest departure in British criminal justice strategy in the last hundred years, the opening of a second front in the war against crime, drawing on the thinking of radical criminologists and pressure groups who were previously locked out of power. It has unfolded almost invisibly, largely unreported and sometimes misunderstood by news media, effectively camouflaged (by accident or design) by the government's relentless release of hard-line criminal policy statements. Most important, at this particular moment, this is a story which looks like it may have a miserable ending. Behind the scenes, this ground-breaking campaign has turned into the policing equivalent of the Flanders fields - misguided generals, destructive commands, confusion in communication, disorder in the ranks, brave young ideas broken and bruised on the ground. But the fight is not yet over. The great danger is that the ministers who launched the campaign may now desert it in its hour of need.

It all started in the spring of 1997. Labour had won the election, and Jack Straw moved into a Home Office which was unlike any other government department in two key respects. First, it had not been starved of resources and so, even though Gordon Brown imposed existing Tory spending plans across Whitehall, the Home Office still had money to spend. Second, other departments were trapped by Tory policies which were destructive but deemed to be politically untouchable (the sale of council houses, the privatisation of railways, Kenneth Baker's education reforms), but the Home Office had nothing. It was a policy wasteland. For nearly 30 years, the masters of the criminal justice system had resorted to a belief which they summarised in two simple words: 'Nothing works'. The last Tory Home Secretary, Michael Howard, had exploited the vacuum to embark on a cynical exercise in political populism, with other two other simple words: 'Prison works'. It didn't. Jack Straw inherited a police service which was detecting fewer crimes, courts which were clogged with delays, fines which were not being collected, community sentences which were not being served and overcrowded prisons which were failing to change the behaviour of prisoners. But he had money to spend and the political opening to do so.

Better still, he had the Prime Minister on side. Under the influence of 'radical realist' criminologists, Tony Blair and the Labour electoral machine had finally seen through the left's blind spot on law and order, understanding that crime was not some kind of romantic rebellion by an alienated working class attacking their middle-class tormentors: it was a miserable plague which was overwhemingly likely to be inflicted on the working class for whom they claimed to speak. For nearly 20 years under the Tories, an epidemic of poverty had swept through the old public housing estates, infecting them with burglary and mugging and harrassment, and it was no longer good enough to have a criminal justice policy which consisted of leaning back over the After Eights and sneering at the cops. Five years earlier, as shadow Home Secretary, Blair had come across the germ of a strategy.

He fastened on to a report by a management consultant, James Morgan, who had been hired by the then minister, John Patten, to look for a solution to rising crime. Morgan had come back with a radical plan to make it a legal requirement for all local authorities to form partnerships with their police in order to attack the causes of crime. Patten took one look at it, took fright at the idea that a Tory government would pass any power at all to local authorities and shelved it. In opposition, Blair met informally with chief constables and local government leaders who encouraged him to adopt the Morgan Report as Labour policy; he goaded the Tories over their failure to act on it. Now, Jack Straw took the Morgan Report down from the shelf.

At the same time, Straw told the Home Office's research department to trawl the archives for evidence of effective tactics. This exercise finally swept away the defeatism of Nothing Works and produced a report, Reducing Offending, which drew heavily on American experiments to pour a stream of fresh thinking into the Home Office. "Nothing works" gave way to "What works".

In effect, the researchers took the rule-book of law enforcement and tore it up. With tightly-argued evidence, they exposed the weakness of traditional policing technique (police patrols "do not have a marked effect on crime levels"; laying more charges against offenders "does not have any noticeable effect"; reducing crime by attacking drug "does not happen") and of traditional punishment (prison was no more effective than fines or community sentences and was so expensive that it would cost £380 million to cut total crime by only 0.6%.) Then they showed the alternatives: deep social programmes like the American pre-school scheme which worked with problem children so that only 7% of them went on to be arrested by the age of 27, compared to 35% of their unaided neighours; using design to cut the opportunities for offenders, like the Dutch programme which cut burglary by 70% by insisting that all houses be secured against crime, the banks who had redesigned their credit cards and cut fraud in half, the local authorities who had redesigned their wheelie bins to stop burglars using them as climbing aids. They recorded how the British car industry had finally started building security into their cars after the Home Office in 1992 published a Car Theft Index which diverted buyers from the riskiest vehicles.

One of these researchers, Professor Ken Pease, of Huddersfield University, later captured the theory of crime reduction in a single sentence: "It is probably not going too far to say that the best strategy for crime control is now clearly a combination of proven techniques for the reduction of individuals' tendency to commit crime through intervention in childhood, and the manipulation of environments to make that more difficult."

By the autumn of 1998, both these initiatives had taken off. The thrust of the Morgan Report had been converted into a legal requirement to set up 376 Crime and Disorder Reduction Partnerships (CDRPs), spread across England and Wales, in which police were compelled to meet with their local social workers, teachers, housing officials and health workers to deliver crime reduction. They were to recruit the wishes of local communities to set their priorities, using their local knowledge and contacts to come up with the infinitely complex tactics to fit each area's own particular needs. And the researchers' radical findings had been developed into the outline of a Crime Reduction Programme. University specialists were hired to direct it and to check its progress. It was backed by the Home Secretary, by Downing Street and by the Treasury. For its first three years, from April 1999, the programme was given £250 million, which soon rose to £400 million with a huge investment in CCTV. Those who were involved speak of a buzz of excitement inside the Home Office - like a birth.

The baby was no sooner born, than it crawled into trouble. Ministers took a decision which, at the time, was controversial and which, in retrospect, was startlingly destructive. They ruled that the CDRPs should have none of the money from the new crime reduction programme. Not a cent. James Morgan had wanted each of them to be given a small budget, just enough to hire one or two staff, an office and a phone, to keep the wheel turning. Ministers, however, over-ruled Morgan's plan and insisted that police and local authorities must fund the new partnerships from their mainstream budgets. The practical result on the ground was that, without staff or cash, the CDRPs had no fuel.

That immediately exposed the CDRPs to a second weakness, which was to cause even more problems: most of the partners were not interested in partnership. The whole initiative was being launched on to a public sector which was exhausted by years of Tory spending cuts, none of which had been reversed. Demoralised and distracted by survival, numerous local authorities could not see what crime had to do with them. In some cases, too, the police were suspicious that this was all about invading their patch. In other cases, the police tried to kick-start the scheme, only to find that this alienated some of their partners, who saw their own patch being invaded.

If the CDRPs themselves had no fuel to run on, and if many local agencies were not willing to invest any effort or funds of their own, the crucial task of analysing local crime simply did not happen. In the autumn of 1998, they were told to provide local crime audits by April 1999, but the deadline arrived, with the alarming news that 117 of the 376 CDRPs had failed to deliver any audit at all. Among those who did, according to a Home Office review, many simply supplied an executive summary without any real data and, even those who tried to collect data found that their partners were not willing or able to provide it: only 38% of the audits had data on school truancy, only 27% of them had health data on drug treatment, only 27% had social services data on offenders.

This was a serious failure, right at the root of the evidence-based, problem-solving philosophy of the programme. Having snapped off the root of the new plant, ministers now dumped manure on it. From April 1999 they started to release the £400 million from the crime reduction programme: the intention was apparently benign, but the result was destructive. They broke the money down into a series of perfectly sensible priorities - burglary, CCTV, prostitution, drug workers, school management etc - but these did not necessarily overlap with local plans, so, for example, a rural CDRP which was most concerned with traffic speeding through its villages would get nothing unless it could find some sex workers to help. Second, they did not offer the money specifically to the CDRPs, with the result that individual local agencies reached out for the cash, bypassing the new partnerships. Finally, they handed out cash only in response to bids, and the most fragile CDRPs lacked the time and energy to submit them effectively. Many CDRPs continued without any funds at all. Others eventually received cash through the police to tackle problems which had never been their priority.

Less than a year had passed, and the whole structure of the CDRPs had already been weakened at the base by three problems which were to continue to haunt it: the clumsy decisions of ministers; the lethargy and weakness of local partners; the heavy hand of Whitehall control. Now, the top of the pyramid also became weakened as, behind the scenes in Whitehall, a kind of craziness took over the other flank of the new strategy, the £400 million crime reduction programme.

The engine house for the programme was the Home Office's Police Research Group which had been given £25 million to feed research to the police and to evaluate their work. Since 1983, the group had been part of the Home Office's police directorate, where it could work directly with forces around the country. Just as the new programme was launched, it was moved, against the wishes of many of its staff, and became part of the Home Office Research and Statistics Unit. It lost its link to the real world. Worse, the group then became caught up in feuding between civil servants in different departments of the Home Office who had an interest in their work, variously ignoring it or attempting to control it. Some junior civil servants were reduced to drifting into other departments to complain about the antics of their bosses. One official simply shut himself in his room to escape the tension. Disunited and dysfunctional, the Home Office then became vulnerable to the demands of the Treasury.

Some of the researchers wanted to go out onto the ground with the police, to help them run their projects; but the Treasury said they must stand back and make objective evaluations without interfering. The researchers wanted to spend the money slowly, partly because they had so much of it that they were running out of academics to spend it on, but, more importantly, because it was the core of the new approach to work slowly, gathering data, analysing it, searching for the precise strategy that would work. At one point, the Treasury agreed to drop their demand for 'annuality' - that each tranche of cash must be spent before the end of the financial year - but rapidly they reneged on the agreement.

One source who was closely involved describes the administrative equivalent of a three-way bar-room brawl: "There were the ministers who had basically said 'Get on with it', which was fine. There were senior Home Office civil servants who were just desperate to spend the money before the Treasury started clawing it back. And there were the researchers saying 'Hold on a minute, this is wrong.' The civil servants were saying 'Spend it. Just get the money out of here.'"

Arguably the most important single initiative was a £30 million programme to run problem-oriented projects with the police. Like the rural car park and the pizza parlour, this can be simple: Oakham Castle in Leicestershire, for example, suffered a long spate of broken windows caused by vandals chucking rubble at the building. For several years, the police tried and failed to control the vandals by occasionally hiding in the building and running out to arrest them - "both expensive and ineffective" as a Home Office report later concluded. Then they adopted the problem-oriented approach, thought laterally and instead of going after the vandals, they went after the rubble. It was swept away, and the windows were safe again. In other cases, it can be hugely complex.

The American criminologist George Kelling was recruited to advise on the epidemic of crime in the New York subway in the late 1980s. The police were trying to deal with the problem by harrassing homeless people who slept in the subway; crime kept rising. Kelling spent a year simply analysing the problem, asking who was in the subway, why, where did they live, what did they do. Armed with this data, he discovered that all of the police officers and pundits who blamed the robbery on homeless people were wrong. The crime was being committed by a relatively small number of people who had homes and who treated the subway as a private playground. Once he had that under his belt, he could devise a strategy which targetted law enforcement on the core of robbers, helped the homeless people through social services, alerted the whole community through a publicity campaign and finally cut subway crime by 80%. Kelling now reports that there are days when the whole subway system is entirely free of crime.

In Britain, Lord Justice Taylor's 1989 report into the Hillsborough disaster is another striking example of wiping away crime with a prolonged and detailed analysis of its roots: his advice that all football spectators should be given seats, instead of wandering loose on terraces, by-passed conventional law enforcement and transformed crime and disorder in stadia.

But the very complexity of the approach was a danger. HM Inspector of Constabulary had just echoed the warnings of a sequence of earlier research that the police generally had not mastered the problem-oriented approach. HM Inspector found ("Beating Crime", 1998) that in 32% of the problem-oriented projects he examined, there was no evidence that the supposed problem even existed; that only 27% of the projects showed evidence of proper analysis; and that only 7% claimed to have succeeded. The researchers warned their civil servants that the new initiatives would succeed only if the police were given two vital supports: time and training. But the civil servants said that was impossible: they had to spend the £30 million before the Treasury's deadline. The result was that police were invited to bid for the money and given only one month to come up with projects. And, in order to get rapid results, they were given no time or funds for training. The researchers watched in dismay as their most important project was launched into dangerously choppy water.

At the same time, they launched another key part of the programme - the Reducing Burglary Initiative, funded by £24million. The plan was to use the problem-oriented approach to develop a package of measures tailored to local needs in a strategy that had been startlingly successful in earlier pilots: in Kirkholt where they cut burglary by 75% over three years; and later in Huddersfield where they cut it by 30%. The police bid for the money in two rounds: 60 special projects which were to develop innovative tactics; a further 186 which would use a more traditional approach. The researchers wanted to divide all the money between the 60 ground-breaking schemes, but senior civil servants over-ruled them and divided the cash equally between all 246. Less than £100,000 each: it was not enough, the researchers warned, to guarantee success from any of them. And why were all of them being given the same money regardless of their scope?

By early 2000, some researchers were so despondent that they considered by-passing the civil servants and going direct to ministers to warn them of what was happening. But they shrunk back from mutiny. In the meantime, the CDRPs had hit a new wave of troubles. Ministers who had already damaged them by denying them fuel now grabbed for the steering wheel and deliberately pushed them off course. Under pressure from Downing Street and the Treasury, who wanted quick, measurable and electorally attractive results, they used the 1999 Local Government Act to require all CDRPs from April 2000 to set new five-year targets for vehicle crime, domestic burglary and robberies. There was no longer any pretence of being guided by the voice of local communities.

In the original CDRP audits, 87% had put domestic violence at the top of their list, but this was nowhere in the new key targets; only 8% of the partnerships wanted to focus on robbery, yet now it was one of the three main national objectives. Those CDRPs who had come up with the same priorities as the government had offered very different aims. The Home Office itself recorded that CDRPs chose targets for reducing vehicle crime by anything between 2% and 25%; Whitehall now ordered all of them to make a 30% cut. In the same way, the local schemes had set targets to cut burglary by between 1% and 16%; Whitehall ordered a cut of 25% from all of them. Beyond that, local needs changed, but Whitehall's new targets were set in stone for five years. And this whole move was produced, in the case of vehicle crime, by an unscripted public remark by the Prime Minister; and, with burglary and robbery, by internal negotiations between the Home Office and the Treasury. There was no attempt at consulting the CDRPs or even the police, many of whom to this day remain privately indignant at the distortion of their work.

Cash was now pouring downwards. Ministers belatedly heeded James Morgan's advice to fund the CDRPs, but now the funding became part of the machinery of Whitehall control. All ten regional governments in England and Wales were told to hire a regional crime director whose job was not simply to get the partnerships working but to ensure that each of them hit the Whitehall targets, which meant a regime of business plans and audits and data-collection - "shedloads of paperwork" as one senior police officer put it - and more and more staff to handle more and more paper in the regional crime directors' offices. Simultaneously, as this series has already described, the same regime of micromanagent was being imposed on the police and the Drug Action Teams, their working lives silting up with bureaucracy.

Beneath the surface, the impact of this central drive began to erode the foundations of the new approach, forcing police back onto the conventional tactics of arrest and conviction in order to deliver the rapid and measurable results which Whitehall was demanding. George Dixon was making his return.

And while central control was tightening, local arrangements were coming unstuck as every outpost of the public sector was re-organised. One partner would reach out for another only to see it lurching away like a drunkard dancng. When the 376 CDRPs were created in 1998, they were told to link up with the 370 police Basic Command Units (the new jargon for a police division). Within 18 months, the police had shuffled their internal boundaries so that there were only 318 BCUs: some of them included several CDRPs; some of them shared a single CDRP with several other BCUs; only 120 now matched the area of their local CDRP. In the meantime, the network of less than 100 Drug Action Teams was expanded so that by the autumn of 2001, there were 149, only a few of which matched BCU and/or CDRP bundaries. (In Wales, the five DAATs were no sooner up and running than they were abolished and replaced by 22 Community Safety Partnerships). In the background, there were 54 probation areas, none of which matched the boundaries of any BCUs or CDRPs or DATs; and in April 2001, they were rearranged into 42 areas. A year later, the structure of eight regional NHS offices was radically altered into a new national network of 304 primary care trusts. The education authorities which had been organised at county level were ordered to devolve funding and powers to individual schools. And annually, the government issued new targets, new guidelines, new funding streams, new law.

There is a perfect glimpse of partnership life on the ground in a recent report by the Inspector of Probation, Rod Morgan, who looked at what happened when the government told the probation service to run the new Drug Treatment and Testing Orders, which allowed courts to order drug-users to accept treatment. The order came through in June 2000. Probation set off in October without any infrastructure to run the scheme. By that time, they were half way into reorganising themselves for April 2001 into their 42 new areas, eight of which were amalgamations needing completely new budgets and plans. By December, the DATs, who were their key partners in the exercise, were also being reorganised - new teams, new boundaries, more new plans. In April 2001, probation were given £36 million to run the scheme. Weeks later, half of it was taken away from them and given to the new National Treatment Agency which was to run the DATs. More new plans - but not until the NTA agreed who was going to spend what, and they didn't start work until the autumn. That dispute was still running in early 2002 when the entire national health service, which was delivering the key treatment for the drug users, was completely reorganised - more new boundaries and plans.

By now, some of the DATs had decided that the DTTOs were too expensive and so they refused to hand over the money that was finally being passed down by the National Treatment Agency. And some of the new primary care trusts, struggling for funds, refused to hand over valuable beds to help law-enforcement when they needed them to hit their own targets. To deal with that, probation had to agree to let the primary care trusts assess the drug users, in addition to probation's own assessment, with the result that only 20% of the assessments were completed within the three weeks required by government. The system was up and running, but only 24% of offenders were actually undergoing the twice-weekly urine tests which lay at its heart - and almost half of them were failing their tests more than 90% of the time. Many probation officers registered this failure but decided it was best to allow the offender to continue with treatment rather than being shunted back to the dead end of imprisonment. Among those who were taken to court for their orders to be revoked, some escaped because there was nobody to impose the revocation.

Around the country, police officers were trying to link up with probation, only to find that probation were too short of funds to do their work; drug action teams referred addicts to treatment, knowing that there was almost no effective treatment out there; social workers sent excluded children for education, but there were almost no out-of-school facilities; courts jailed homeless beggars because there was nowhere else to house them. Even when agencies had the funds and the plans to work together, they could be prised apart by clashing targets from the centre - like the CDRP plan in Bristol to send drug users on DTTOs to live in a probation hostel. The hostel had space, local probation liked the idea, but they had a national performance target which required them to keep their hostels at least 90% full and, since chaotic drug users were the least reliable residents, they had to block the plan.

In the autumn of 2000, the Home Office received a clear signal that all was not well on the ground when HM Inspector of Constabulary and the Audit Commission jointly reported that the new CDRPs were struggling. They urged government to stop over-riding local plans, to measure police against 'community sentiment', to set crime targets for health and education. And, in particular, they urged them to extend the Crime and Disorder Act so that it required partnership work not only from local authorities but from all central government departments as well. The advice was rejected.

As time passed, the researchers began to see the results of the crime-reduction initiatives which they had launched into this confusion. Their prized £30 million scheme to promote problem-oriented policing had crashed into the same barriers which had obstructed earlier efforts. Without the time or training to do the job properly, schemes generally had failed to gather the right data, failed to analyse it properly and failed to device effective solutions. This project was central to the whole crime reduction programme. In a report published earlier this year, two researchers hired by the Home Office, Karen Bullock and Nick Tilley, recorded: "The conclusions are not encouraging."

They noted the destructive impact of the Treasury's timetable: "A much longer initial period would have been needed to make progress.... In so far as the funding regime itself provoked premature closure over problem-definitions and measures to address them, ironically it may, >>> Of course, have undermined the problem-orientation it was intended to promote." And they noted the equally destructive impact of ministerial priorities. British police were now employing more analysts for problem-solving, they wrote: "What they currently do, though, appears to be more oriented to a detection and enforcement agenda and to satisfying performance indicator requirements, than it is to an agenda concerned with dealing with police-relevant community problems."

The former head of the Police Research Group at the Home Office, Gloria Laycock, co-wrote an analysis of the failure, pointing to the capitulation of the police to Home Office pressure to carry on arresting: "There are far too many cases where persistent problems are ignored and where, even when the police do think about solving a problem, they fall back on the standard approach of arrest and prosecution." The bold new programme had wandered far from home.

The £24 million Reducing Burglary Initiative had also suffered from the same cramping of the problem-oriented approach and from the Treasury's attempt to stop researchers intervening to help the police. Three consortia of university specialists were hired to follow the 60 most innovative schemes, to measure changes in burglary rates and to check the funding against the £2,300 average social cost of a burglary. The results were far from successful.

One of the consortia, led by Professor Tim Hope of Keele University, reported that only one of its 20 schemes had secured a significant cut in burglary at a cost-effective price; five others had scored a less significant cut at less than £2,300 per crime. The rest had failed: seven of them had had a more or less neutral effect on burglary at an unacceptably high price, with one of them costing eleven times the £2,300 for each offence it cut; and, worst of all, seven other schemes had stimulated a clear rise in burglary, sending it spiralling upwards in one case by 39%, usually at vast cost. Prof Hope told us: "We found all the implementation problems on the ground that need time and thought and effort to get right. Most projects were on a hiding to nothing from the start."

Most of the schemes had made a particular effort to protect 'repeat victims' - the 4% of the population who suffer 44% of all crime. But Prof Hope's researchers found that, out of their 20 sample schemes, none had reduced the risk of these victims, and six had actually increased it. And this turned out to be a crucial clue to what was going wrong. Just as conventional law enforcement often failed because it assumed that offenders would behave rationally and be deterred, so now crime reduction discovered that the victims of crime would not necessarily behave rationally either.

Most of the burglary schemes had used 'target hardening' as the core of their strategy, offering free packages of extra security to any household which wanted it. The problem was that the most vulnerable households were also the least likely to accept help: the victims were alienated from officialdom and suspicous of the approach; the houses were owned by absentee landlords who could not care less about protecting their tenants; the victims lived chaotic lives in short-term housing and had moved before the offer was made. In some cases, when the police target-hardened the most co-operative houses in the area, the effect was to divert all the burglary onto the most vulnerable, with a dramatic increase in the suffering of the repeat victims.

This clue, in turn, led back to the structural weakness which had been inflicted on the CDRPs who had been created to run schemes like this with the police. The researchers found clear evidence that the most successful burglary projects were planned from the bottom up, using community knowledge and contacts; when they went wrong, they rapidly understood why and adapted. The least succesful ones were imposed by the police alone who sometimes saw the emerging failure, knew no way to adapt the plan and reacted by simply deciding to spend all the money before it was clawed back.

For several years, the Police Foundation think-tank have been trying to ring alarm bells warning that constables can be inherently conservative and even cynical and tend not to do very much about initiatives that are handed down from on high without any kind of consultation with them. A scheme works its way from Whitehall down through the ranks until finally it arrives on the desk of a duty sergeant along with a whole lot of other paper and becomes, for the constables, just the latest fad that they are expected to deal with. It is an exact echo of the problem of crime victims who do not co-operate with burglary initiatives which are passed down from distant desks. From the Home Office's point of view, abandoning the reality of consultation, it was the opposite of rolling a snowball down a hill and watching it gather weight, more like throwing a vase off a rooftop and watching people on the ground pick up bits and wonder where that came from.

Supporters of the burglary initiative were dismayed by the results but not defeated. The research had done precisely what it was supposed to do: by mid 2002 they had identified the problem and so now they could deal with it by strengthening the partnerships and using their local knowledge to adapt tactics on the ground. But now something else went wrong: the Home Office lost its nerve. They simply sat on the research, denying the police and the CDRPs the chance to learn the lessons. A year passed, and last month (June 2003), a brief summary was published amidst such controversy about its claims of success that Prof Hope's consortia demanded that its name be removed from it on the grounds that it was "highly misleading and selective".

At a conference on criminology last month, Prof Hope complained that the Home Office summary simply failed to say anything about whether the burglary schemes had been cost effective and undervalued the fact that burglary had been falling across the country, crediting the schemes with a success to which they were not necessarily entitled. The Home Office summary, he said, "produces results that at best contradict our own and at worst seriously distort the actual pattern of project outcomes as far as we can judge. These tend to point in a much more positive and favourable direction than our own." This looked, he said, like "the regrettably familiar government game of policy-led evidence". "What works" had started to give way to "Pretend it works."

In the background, the Home Office was rapidly discarding other elements of the new programme to beat property crime. In 1999, they commissioned a group of specialist insurance executives, police officers, fire officers and academics to advise them on imaginative new approaches to burglary, arson and criminal damage. They were known as P-crat, the Property Crime Reduction Action Team. In May 2001, they produced an initial report which was full of energetic ideas.

They wanted building regulations to include a requirement for all new homes to be secured against crime; a £1 computer chip installed in all new electronic goods so that they could be rendered inoperable if they were stolen; a national database of stolen valuables to make it harder to fence them; a new library of good practice in crime prevention; an expansion of the CDRPs to include more local agencies, including fire officers. They had already negotiated preliminary figures from insurance companies who were willing to slash premiums for home owners who took up their ideas and they had set up five working parties to carry forward their plans.

Their work was welcomed by the then minister, Charles Clarke, who said: "I look forward to the delivery of their full report and to doing all I can to help them to achieve reductions in property crime." But there never was a full report. P-crat was quietly disbanded. None of the ideas has materialised in the Home Office, although their arson ideas have been taken over by John Prescott's Arson Control Forum. One of those closely involved with P-crat told us: "There was an election; there was a new Home Secretary. The whole thing lost momentum."

At the same time, the Home Office had funded work by a Department of Trade inquiry, chaired by Lord Sharman, into technology and crime. They too produced innovative ideas, based on the advice of 60 experts and 26 seminars, to protect computer systems from hacking, to allow the police to record crime and to appeal for help through the internet, to build crime prevention into the life cycle of every product, and to tie central government into the kind of partnership work which was now required of the local CDRPs. They undertook to make sure that all their ideas were acted on. "Crime reduction is not the priority for everyone that it should be, and we need to change that," Lord Sharman wrote in his foreword to the report. Thus far, the government has set up funding for more research, but none of the other ideas has materialised. Sharman warned, in particular, that mobile phone companies should design their products so that they could be blocked if they were stolen: nothing was done until the theft of mobile phones soared.

Part of the difficulty seems to have been a resistance from mainstream business. One of P-crat's problems was that they pushed hard for a compulsory scheme for new houses to be 'secured by design' with high quality doors and window frames and estate lay-outs which expose thieves to the natural surveillance of neighbours. As a voluntary scheme, this scored startling success in the early 1990s, with some projects reporting that burglary fell to zero. But the big construction companies soon discovered that buyers were not willing to pay extra for the security and stopped co-operating. P-crat wanted to compel them, but one insider suggests that the steady fall in burglary since 1993 has weakened the government's will to take on the construction industry.

Some of the big car park operators have resisted a long-standing programme to win 'secure car park status' by installing CCTV and other crime-prevention, even though car parks are the site of 22% of all vehicle crime and the secure status cuts crime by an average of 70%. The government aimed to have 2000 car parks in the scheme by the end of 2000: more than two years later, they still have only 1,210. Like the hoteliers, car park operators profit from consumer law which omits them from the definition of 'goods and services' which would otherwise make them liable for thefts on their property. Similarly, at various points, the Home Office has looked at amending the 1956 Hotel Proprietors Act to stop hotels denying liability for the theft of their guests' possessions. In France, for example hotels are legally responsible and, therefore, are far more security conscious. But British hoteliers have dug in their heels.

Four years after its official launch, the crime reduction programme now is buried in a rubble of disappointment. But it is not dead. There are signs of some partnerships finally beginning to work. In Bristol, for example, where the CDRP was particularly weak in its early years, there has been a revival, brought about from an unlikely source. Tony Blair's 2002 Street Crime Initiative on mugging, which was a classic example of centralised control reinforced with a rigid performance regime, appeared to be a further blow to the CDRPs by imposing yet another law-enforcement requirement on the police, but ironically the sheer political pressure to come up with a result forced local agencies to work together. The chief constable of Avon and Somerset, Steve Pilkington, chaired a special leadership group to run the initiative: he pulled together the same agencies who were failing to work together in the CDRP and gave them a chance to succeed. That has now had a knock-on effect on the CDRP, which finally is showing signs of life.

And some of the schemes run by the crime reduction programme appear to have done better than their initiatives on burglary and problem-oriented policing. The Youth Justice Board has established Young Offender Teams and Youth Offender Panels in every area to divert juveniles from prison and into constructive schemes. Parenting Orders and Anti Social Behaviour Orders - both of which received a bad press when they were announced - have proved to be less repressive and more constructive than was feared. The huge Sure Start programme has generated projects around the country to work with pre-school children, on the same model as the successful American schemes, but the long-term nature of the work means the evaluation will not be finished until 2007.

And whereas P-crat was dumped, the equivalent action team for vehicle crime, V-crat, has done better, with a little help from the Department of Transport. Their key idea, of having immobilisers fitted to all cars so that they can not be driven away by thieves, fell foul of technical problems but they persuaded government to tackle the wrecking of stolen cars by scrap-metal firms and to make life more difficult for car thieves by improving the data flow to the police from the DVLA, car insurance companies and MOT records, with the aim of clamping down not only on drivers without correct paperwork but on the apparently large pool of cars which exist nowhere in official records and are routinely used for crime.

But the drum beat from the Home Office is not so encouraging. It recently posted a one-line message on its website, noting that the crime reduction programme which was originally conceived by researchers as a ten-year project, had closed in March last year after only three years. The Home Office say it has been replaced by a new funding stream, the Safer Communities Initiative, but its £20 million last year was a trickle compared to the original programme's £400 million. The officials who ran it have been dispersed throughout the Home Office. Some of the local schemes are still running, but there is no more central funding or training for problem-oriented policing or radical new initiatives on burglary. Indeed, the Treasury, which was blamed for many of the problems with the burglary projects, used the weak results to claw back £26 million which had originally been earmarked for a second phase of 250 more schemes. Funding for other schems including design against crime drained away last year. Crucially, there is no more central guidance from researchers; and no more central funding for the kind of evidence-based innovation which lay at the heart of the original programme.

The CDRPs themselves now benefit from direct funding and will receive some £84 million a year for the next three years - but this is a substantial cut compared to last year, because £25 million of their funding has been diverted to police BCUs. Although some local police commanders are now committed supporters of the CDRPs and will spend the money with them, others are being encouraged by the Home Office to spend it on DNA banks and car number-plate recognition systems and new finger-print databases, all the new technology of old law enforcement.

As the Home Secretary ratchets up his rhetoric for more detections and longer sentences, there are clear signs that the balance of policy has tipped back towards law enforcement. The crime reduction programme aimed to cut the prison population by 10%; but it has risen as the government pushes a conventional agenda. It has just set up new criminal justice boards across the country with the specific task of achieving its central performance targets.

The whole radical new strategy is now balanced between the power of its underlying thinking and the weakness of its implementation. The law still requires that police and local authorities work in partnership. There is still money flowing through the CDRPs. More than ever before, there are police officers and council officials and health workers and teachers and social workers who can see the value of solving problems instead of merely trying to punish their results. They can see that if police make the roads safer, hospitals will have fewer casualties to treat; that if schools work well with their truants, police will have fewer burglaries to handle; that if any of them can break through on drug treatment, all of them will be better off.

In a speech in Hackney last September, Tony Blair declared: "It is crucial to address poverty, social exclusion, unemployment, lack of education, the sense of hopelessness that is so often the breeding ground for crime." The message from the battered body of the crime reduction programme is that he can still do that, but his Home Office ministers too often have looked for quick fixes when the real solutions are long term; demanded measurable results when the real progress is often invisible; forced great simplicities on a problem which is notoriously complex; imposed rigid centralism where local flexibility is needed; and, most of all, failed to stand up for their project, failed to educate the public in the truth about the weakness of conventional law enforcement, failed to explain the virtues of their crime reduction programme and the new partnerships, and reverted to the easy populist rhetoric which made George Dixon a star.