Ministers wanted to pay us all to learn. Alas, their scheme simply created a cowboy heaven

The minister for adult skills ought to be a buxom blonde in a saucy 1970s film entitled Confessions of a Labour Aide. The actual holder of that title, John Healey, is not remotely pneumatic. But he is making politics a little livelier at the moment by his "it's a fair cop" responses to questions from the Commons education select committee on the untimely death of individual learning accounts (ILAs).

With uncharacteristic frankness for this most anal of governments, he has explained how one of new Labour's better ideas foundered on the inadequacies of execution. It is worth rehearsing, as a microcosm of wider problems in what the PM calls "delivery".

The idea was simple and appealing. Anyone would be able to open a "learning account". Both the government and the individual would inject funds into it. And these funds would be used for the purchase of training or vocational education. In this way, millions would gain the habit of periodically assessing and improving their capabilities.

Hyped by Brown, Blunkett, Harman et al long before the 1997 general election as a "stakeholder" policy, the accounts were an important plank of the new Labour edifice. After several reworkings, they were finally introduced in 2000. Blunkett, as Education Secretary, went for a two-year scheme that provided up to £200 of training subsidy to each applicant. The greatest incentives were to acquire IT skills.

It looked a stonking success, and was lauded as such by Blunkett and his successor, Estelle Morris, till they noticed the stench of something going bad. By last October, after just 18 months, 2.5 million people had opened accounts, more than double the one million target for March of this year. Well over one million had registered for courses with 8,000 training providers.

The corollary was that the budget was overspent by more than a third - £62.9m, according to figures Healey has supplied to the Commons education committee - and that the overspend threatened to go higher. As always with this kind of initiative, there are also questions about how many people received government money for training they would have bought in any case. And there is evidence that far too little was done to promote the scheme to the most disadvantaged groups, those with the fewest skills.

But the real scandal is that the scale of fraud and profiteering was huge and unsustainable. The further I go into this, the more I wonder how many of the 2.5 million account-holders are real.

The system required account-holders to sign up for courses with training companies, or "learning providers". The providers then received the government's share of the fees. A letter from Healey to the select committee, which goes into arrangements for assessing claims for payment from the providers, is unnerving. It concerns claims received before 21 November, or two days before the scheme was summarily closed down: "Before Christmas we were able to pay 91 per cent of provider claims, covering almost 1,400 providers, with a value of £3.5m. However, 136 providers' claims, with a total value of £11.3m, were not paid . . . These are being held while complaints and concerns about the learning delivered are investigated."

In other words, the Department for Education and Skills questioned the validity of £11.3m out of £14.8m claimed, or 76 per cent. These figures may apply only to a smallish batch; but, because more than £200m in total has been disbursed, they are worrying to say the least.

Healey goes on to say that investigators are examining 82 "learning providers" and the police are probing 13. There have already been 39 arrests in cases involving two registered providers. The programme was terminated, Healey says, because of the "very serious risk that public monies could be obtained improperly".

Much of the coverage of the debacle has concentrated on the role of the private sector company Capita, which created and managed the ILA computer systems. It is possible that Capita's procedures for safeguarding valuable data were not tight enough - though it denies this.

But I think the scheme's deficiencies were more fundamental. Training companies could receive government payments simply by submitting the names and account numbers of individuals they claimed to have taught. Minimal authorisation was required from the account-holder. This system might have worked if the companies themselves were subject to rigorous vetting. But the government wanted a light regulatory touch to keep administrative costs down. So more or less anyone could set themselves up as a "learning provider". It was cowboy heaven.

The opportunities ranged from knocking on doors to get whole roads of people signing up for fictitious training, through to over-charging for low-quality "educational" CD-Roms. What eventually led to panic was the discovery of a computer disk that offered for sale the names and details of 1,000 account-holders.

The lessons? One is that there is no point trying to implement an important policy on the cheap. Certain minimal standards of control and regulation are essential. Another, I think, is that the scheme was insufficiently "stakeholderish": I suspect there would have been less fraud if individuals had been given more control over their accounts. When people feel they own something, they tend not to hand it over to the first salesman who knocks on the door offering career transformation through spurious digitised training.

Robert Peston is editorial director of Quest ; www.csquest.com; e-mail rpeston@csquest.com