Edward Leigh, chairman of the House of Commons Public Accounts Committee, is a difficult man to please. But even he praised executives at HM Revenue and Customs (HMRC) for successfully transferring a huge IT contract from one supplier to another.
Leigh said last year that the facts showed that, "It is possible to run a successful competition even when the incumbent seems firmly ensconced."
He was referring to the Aspire (Acquiring Strategic Partners for the Inland Revenue) contract, in which services supplier Capgemini took over the running of the Inland Revenue's IT department from EDS in July 2004.
It had been thought impossible to dislodge EDS, which had run a bewildering complexity of more than 110 major tax systems for 10 years.
The EDS transfer to Capgemini had its difficulties. Some key staff at EDS, for example, did not move to the new supplier and had to be contracted to work temporarily for Capgemini after the start of the contract. And there were failures in the IT services which caused distress to the Revenue's corporate users, some of whom contacted Computer Weekly.
Still, more than 2,000 IT staff transferred from EDS to Capgemini without the work of the department falling apart.
What nobody knew then, however, was that the total cost of the Aspire contract would soar from Capgemini's initial bid of £2.9bn to an estimated £8bn over 10 years.
By the time the projected costs had increased last year to £7bn, Leigh and the National Audit Office became concerned. A statement issued on 19 July, 2006 by Leigh said, "I am concerned about the delays to critical projects and the steep rise in the overall cost of the contract.
"These have come about because the department has changed its requirements and because there has been a higher than forecast demand for IT services. The department must get a firmer grip on these projects and on managing the contract."
But since this statement the projected cost of the contract has risen by a further £1bn to £8bn, and no detailed explanation of the increased spend has been provided by HMRC or the National Audit Office.
The Revenue has said that Capgemini is performing well and the department is in control of its costs. But MP Richard Bacon, a member of the Public Accounts Committee, said that the increased cost of Aspire indicated that the department may not be in full control.
This concern is underlined by evidence from the National Audit Office that the department may not have the measures in place to exert the amount of control that is needed. Indeed, IT costs at the Revenue seem to have risen unstoppably.
The Inland Revenue's main IT contract between 1994 and 2004, before the merger with Customs and Excise, was with EDS. It began in 1994 at a cost of £1bn and was £2.5bn by early 2004.
By the time Capgemini had taken over a revised contract in the summer of 2004, the cost of running the department's systems over the next 10 years from 2004 to 2014 was estimated to be £3.5bn.
Then, in 2005, the Inland Revenue and Customs and Excise were merged. As a result, Fujitsu's £1bn contract with Customs and Excise was merged into the Revenue's £3.5bn Aspire deal with Capgemini. That made the Aspire contract worth about £4.5bn.
This figure has now risen by an estimated £3.5bn to £8bn, and the National Audit Office is not sure where the extra money is going to come from. Its report on the Aspire contract said there were questions about "how the department will fund the additional spending on IT under the new contract". And that was at a time when the audit office expected the total contract to be worth £7bn.
It has also emerged that Capgemini's profit margins on the hugely expanded contract are around 10% - about the same percentage as the profit margins on the lower figure of £4.5bn - which hardly sounds like tough negotiating on the part of HMRC.
So for all the extra money Capgemini is receiving, is it bringing something to HMRC that EDS did not?
To date, the National Audit Office has found nothing seriously awry with Capgemini's performance, though there had been "some delays and cost increases on business-critical projects, which have on the whole been caused by the department changing its requirements".
It has also taken about 18 months for the performance of Capgemini to bed down.
"The department had originally intended to produce an annual value for money report in November 2005," said HMRC. "But it now intends to produce a financial scorecard and a customer service assessment reporting mechanism in late 2006 to assess the supplier's overall performance."
Computer Weekly asked HMRC for a copy of that overall performance assessment. Last week the department refused, saying that the report was in commercial confidence. It did, however, say it was happy with Capgemini's performance.
As things stand, then, the outside world is left to take it on trust that there is solid, independent evidence to support the department's claim.And this is arguably the biggest weakness with Aspire - and with some other large government IT-based contracts, such as those related to the ID cards scheme. Every penny of the money spent on IT contracts may be worthwhile. But nobody outside the department has the evidence on which to judge.
HMRC is able to spend an extra £3.5bn on an IT services contract without having to account in detail to anyone outside the department over how the money is going to be spent. If the department has provided these figures to the National Audit Office, they have yet to be published.
There was, in fact, an expectation that costs would fall when the Inland Revenue and Customs and Excise merged in 2005 - that was the point of the merger. Indeed, the National Audit Office reported in 2006 that HMRC wanted to reduce IT costs to less than 20% of its total budget.
However, the reality is that £8bn may not be the final figure.
The EDS outsourcing contract rose by two-and-half-times the original £1bn. If Aspire were also to increase in cost by two and a half times, it would cost about £10bn - or four times Capgemini's original bid price for Aspire.
The trouble is that, to HMRC, all accountability means is providing the Public Accounts Committee with a list of headings under which the extra money is being spent.
And even the headings are vague: operational services charges, service credits, business application development and enhancements integration, desktop applications, rate-based services. No sub- headings are provided.
The department is willing to say that there is a "continuing high demand for IT services", but at no point does it explain how the money is being spent.
Some might say it is encouraging that HMRC is investing heavily in IT to address the shortcomings highlighted by the Revenue's CIO Steve Lamey in his speech in 2005. Lamey said then that some of the technology at the Inland Revenue would not look out of place in a museum of IT.
If well spent, the extra money could cut the overall costs at the Revenue and could improve the reliability and breadth of the services to the public.
But there is no evidence of this happening, and this evidence is what is needed. In the US, politicians would require a line-by-line breakdown of an increased spend of $7bn. In the UK, the public are not entitled to know how the department spends public money.
So we do not know whether the extra £3.5bn has been thrown into fighting crisis after crisis - and perhaps not even then successfully - or if it has been spent on the effective delivery of IT services.
As things stand, the National Audit Office audits the books of HMRC, but a very British convention still requires the factual content of the audit report to be agreed with the department before publication.
Much of the wording of the auditor's reports is, therefore, polite and far from explicit. For example, it is only in a footnote in a 61-page report on Aspire that the National Audit Office reveals that the annual profit margin on running the National Insurance Recording System 2 - now a subset of the Aspire contract - has ranged previously from 25.9% to 39.7%.
This compares with average profit margins on Private Finance Initiative schemes of about 10% to 13.5%.
HMRC may say it is becoming more open and accountable, but the evidence is not there yet. And that is a major concern when £8bn of public money is at stake.
Comment on this article: email@example.com