We now move on to item 2, on the Scottish Prison Service. I invite the Auditor General to brief the committee on the report that was recently published by Audit Scotland.
In November 2003, the Scottish Prison Service signed a contract, on behalf of Scottish ministers, with Reliance Secure Task Management Ltd—commonly referred to as Reliance—for the provision of prisoner escort and court custody services throughout Scotland. The contract is worth about £126 million over seven years and covers the transport of prisoners between prisons and police stations, for example, and court, and the safe custody of prisoners while at court. In 2003-04, there were some 140,000 prisoner escorts throughout Scotland.
Throughout the report, it appears that the information that was available from the police and the SPS about what happened in the past was not reliable. I am therefore concerned that some of the anticipated savings might well not be achieved, but there seems to be no tracking of that. Do you have further information on that point?
You asked first about the information that was available. The SPS has advised us that one of the subsidiary reasons for considering the form of contract that was used was the fact that in the past very poor information about performance was available. It is not unfair to say that, generally, the agencies that carried out the service in the past did not retain and manage good information about how the service was delivered. As a consequence, the SPS reached the general conclusion that it would not be practical to attempt to gather such information so long after the event, which meant that limited information could be provided to potential tenderers about what they would be taking on. I referred to that in my opening comments.
As the Auditor General said, the main difference between the service that was provided in-house and the service that Reliance performs is that the former was based on input whereas the latter is based on outputs. There is a shortage of information about input and even if such information were available it would be difficult to compare it with the output performance against which the new contract is measured—that is not to diminish the disadvantage of not having the information but to recognise that it would be difficult to compare one type of information with the other.
That is why I am concerned. How can we say that savings are being made when we do not know how many hours used to be input by the SPS, the police and Kilmarnock prison, for example? Where does the figure come from?
The figure of £20 million represents savings against the calculation of what might be called the public sector comparator. It is a calculation of what the public sector cost of continuing to pay for the service in-house would have been, as against what the contract is costing. It is a calculation, but with the disadvantage that it does not have all the information behind it about what was being spent.
So the figure may not be as robust as we would like.
I am sure that the Scottish Prison Service would say that it was not an absolutely accurate, guaranteed figure but the best estimate, using the public sector comparator, of what could be saved.
The concept of a contingency may for some people imply a catastrophic failure. However, if the contract is not a success, it will not necessarily close suddenly—it could wind down and leave time for contingency measures to be put in place. That is not to say that a catastrophic failure is impossible, but it is the least likely outcome.
The Auditor General said that it would be "inappropriate" for the committee to be concerned about such a failure. Why did you use that term?
The contract has been properly formed and at present is being rolled out with a longer timescale. I assure the committee that I see no immediate prospect that the contract will not continue to plan.
What is the plan? You say that the contract is properly formed, but it strikes me that it was based on poor or limited information. The contract is supposed to save £20 million. The use of penalty clauses comes after the event—we should not need them because we should get the saving that is predicted. However, the estimate for the saving seems to be poorly based. Is it not the case that the SPS estimate of a saving of £20 million was largely speculative, especially given that the costing was not based on specific activity data?
Before the Auditor General answers that question, I ask him to answer Margaret Jamieson's third question, which was about a newspaper report.
I am sorry, but you have the advantage on me on that issue—I may not have been around when the press report came out. Will Margaret Jamieson help me with that?
The report was by the home affairs correspondent for Scotland on Sunday, Kate Foster, and appeared in that newspaper on 3 October 2004. It states:
Thank you for helping me with that. I am sure that the committee appreciates that my role as the Auditor General is to report to Parliament on matters relating to accounts for which accountable officers are responsible. I do not have powers or duties to report on contractors or what they might say.
Very good. We will now deal with Andrew Welsh's question.
As Arwel Roberts and I said, the Prison Service calculated its best estimate of the potential saving using the public sector comparator, which is an appropriate action for all departments or agencies in considering such matters. I would not wish to hold bodies to a requirement that there be an exact calculation of savings that will eventually be achieved. I imagine that the Prison Service takes useful knowledge from the fact that the Reliance tender was the lowest of the three tenders and that there was good competition for what is a fairly specialist job. The Prison Service accepted the lowest tender and is satisfied with the technical specification. It is not my role to second-guess the decisions that the Prison Service took, but I am satisfied that it undertook a proper and robust process and that, under the circumstances, it probably let the contract well.
The contract could be technically correct, but it may not have been sensible. Is it the case that Reliance was supplied with specific activity data and advice on staffing numbers only after the contract was awarded?
Reliance was provided with that information when the tenders were considered. Before the contract was let, Reliance and other potential contractors were given information. Subsequently, after the contract was let, Reliance was given more information.
It is the word "subsequently" that bothers me. How significant was the subsequent information? When the contract was signed, how much did Reliance know about what it had to do?
I refer you to paragraph 2.11 of our report, which states:
How significant was the subsequent information? What was the contract based on when it was signed? Was the basis of the awarding of the contract adequate, or was there important subsequent information that would have affected negotiations?
The Prison Service reacted to Reliance's request for additional information. Reliance clearly did not feel that the information in the invitation to tender was sufficient, so that information was subsequently supplemented.
In tendering processes for public contracts, it is not uncommon for potential tenderers to ask for additional information. Paragraphs 2.12 and 2.13 of our report, which attempt to explain what happened as precisely as possible, state:
How realistic was the £20 million projected saving?
I have attempted to answer that question; any more specific questions would have to be put to the accountable officer in person. However, I assure the committee that the Scottish Prison Service used its best endeavours to compare bids with the public sector comparator, based on the best information that it had. That is a significant assurance.
On management and monitoring, you say that the SPS
We will find it difficult to provide a comprehensive and reliable response to that question because, as I described, the audit is unusual in that it was undertaken partly in response to a request from the minister to provide independent assurance and the contract is still bedding in. Moreover, much of the information for which you ask is not fully available to us and the situation changes weekly as work proceeds to bed the contract in. However, I ask the team whether it can help you with some information on that.
I refer the committee to paragraph 3.8, which is on page 15 of the report. It indicates that the SPS has finalised service-level agreements with individual prison establishments, the Scottish Court Service, the Crown Office and Procurator Fiscal Service and four local authorities. Service-level agreements have still to be finalised with the remaining local authorities and the eight Scottish police forces, but there is a working group that is seeking to finalise those agreements as soon as possible.
It strikes me that the situation is still in the melting pot rather than being set and in operation. Have any of the penalty clauses been implemented yet?
We understand that they have.
In paragraph 6 of the report's summary, you say:
On the point about value for money being equivalent to the £20 million saving, neither the SPS nor Audit Scotland would expect best value to be equivalent to cost reduction. The SPS is hopeful that it will get a better quality of service, as well as a more efficient one, measured against the public sector comparator.
I can only repeat the point that because the contract with Reliance is based on a definition of outputs, to make a comparison with historical inputs might not give an accurate picture of the value for money to which Robin Harper refers. I stress that the £20 million figure is based on an estimate, due to the difficulty of obtaining information. The concept of the public sector comparator is similar to the one that is used when a contract is put together under the private finance initiative. Under that system, estimates are made of the likely public cost against the private cost. Similarly, with the prisoner escort contract, an estimate of the cost of delivering a particular output was made and compared with the cost that Reliance offered. That is what influences the cost element of value for money, separately from quality.
I take the Auditor General's point about best value involving quality. Is he saying that, in our monitoring of the audit over the years to come, we should concentrate entirely on outputs?
Yes.
Some aspects of the contract are commercially confidential. To what extent might that change in the future as a result of the operation of the freedom of information legislation and to what extent would more availability of commercial information be in the public interest? Would it help to drive down costs as competitors become better able to compare costs?
It is fair to predict that the freedom of information legislation will make it more difficult for any party to use commercial confidentiality as a reason for withholding information that would otherwise be put into the public domain. That would have to be considered case by case. In the instance about which we are talking, the contract provides for Reliance to have significant control over the release of what it considers to be commercial information.
One thing that still bothers me is that Reliance's bid was the cheapest and that the present tendering system tends to lead to the cheapest bid being accepted. Reliance's bid was accepted even though it was not technically the best, but it had enough in it to satisfy the technical requirements and was the cheapest. However, there were no activity data, and it bothers me that those involved seem to have talked about the detailed activity in an activity-based process after the contract had been signed. Is that normal practice and would you recommend it?
I apologise, because I am not giving a good explanation of what happened in the contracting-out process. The SPS specified the contract appropriately and included sufficient information for the tenderers to prepare their bids. In the course of preparing their bids, the tenderers came back and asked for various pieces of additional information, which were provided. That in turn led to a good competition under the circumstances. The lowest bid was accepted, but only after an extensive technical evaluation, which included a range of factors that I have detailed in the report. Therefore, there was a careful assessment of whether the contractor with the lowest bid was capable of delivering the required standard of service, and the Prison Service satisfied itself that that was the case.
Thank you for that clarification.
With that, we close agenda item 2. Members will be able to consider, under a later agenda item, what action, if any, the committee wishes to take in regard to the discussion that we have just had.