On behalf of the committee, I welcome the public, the press and our witnesses. I introduce Mr Kenneth MacKenzie, our principal witness, the accounting officer, secretary and head of the Scottish Executive Development Department. I welcome Mr MacKenzie and ask him to introduce his colleague.
Thank you, convener. Allow me to introduce John Howison, who is the department's chief road engineer and was principally engaged in the A74(M) project.
You are both very welcome. The approach that the committee will take in this meeting will be that individual members, in turn, will ask a series of questions on a particular theme of their choice. Afterwards, there will be an open session in which members of the committee will be able to pursue follow-up questions. Questions from members may be put directly to both Mr MacKenzie and Mr Howison, who have confirmed that they are happy with that approach.
Yes, I think that the report is fair and balanced, presenting an accurate account of the position. It recognises that the department's handling of the project was generally competent, that the price that we got from Autolink Concessionaires is likely to remain value for money, and that the PFI approach was justified in this case. We have noted several comments on various aspects of the report and I am sure that you will want to ask us about those. We will state our position clearly. We have already accepted or adopted four of the recommendations made in the report and there are two that we are prepared to consider.
Thank you. I want to start with a general question. There are lessons to be learned from the way in which the project was conducted and my colleagues will no doubt press you on those matters that could have been handled more effectively. Will you set out those aspects of the projects that were handled well and what benefits there have been to the public as a result?
I ask the committee to bear in mind that the A74 route is the main thoroughfare between Scotland and England. By 1995—apart from the 28 km section in the middle, from Paddy's Rickle Bridge to Cleuchbrae, which remained a dual carriageway—the road had been upgraded to motorway standard from the outskirts of Glasgow to just beyond the English border, following the pledges of the previous Administration. At that time, our cost-benefit analysis suggested that the upgrading of that section would produce net benefits. The main attraction was in bringing the main motorway from Scotland to England—Scotland's principal and most important trading link—up to the same standard throughout.
That question allowed you to tell us what went well. I also want to ask the opposite question. What went wrong and what lessons have you learned? Can you tell us, with hindsight, what did not work?
I cannot give a final answer on that, because we have been operating the completed road only since April. I do not know whether members of the committee have had the opportunity to drive on it, but it is a dream to drive on. It is safer than before and it is quicker, so there are economic benefits for users and businesses that depend on it. To that extent, it is a success. I cannot give you a definitive answer until the 30 years are up and we see whether it is operated and maintained to the standard that we specified and that the financial structure on which it is based is delivering value for money.
We have to act now rather than wait for 30 years.
The only thing that would cause me not to do it in the same way would be if I had any doubts about whether it was value for money. As an accounting officer reporting to you I am not primarily concerned with how the project is delivered, how the money is raised and how the work is done, provided that the solution that the department adopts gives best value for the money that we are spending. It should not necessarily be the cheapest solution; it should be the best affordable solution at that time. In a future case, I would ask the same questions, particularly how the project looked against the public sector comparator.
That issue will, I hope, be explored later.
I will ask Kenneth the question, but he can pass it on if that is appropriate.
The first major repair of the road will occur between years 17 and 24. On whether it should be two lanes or three lanes, I remind you that the report says that the specification
Were ways of avoiding congestion other than building three lanes, such as traffic calming, examined in detail?
The motorway is the principal trade route to England. Maintaining it to a very high standard has been a top Government priority for more than 20 years. It would have seemed odd, therefore, to try to reduce the speed of traffic using this main artery, to increase its capacity without building the extra width.
Could we compare opting for three lanes with a young couple setting up home and going for a five-bedroom house instead of a two-bedroom semi? We are going for a much bigger and more expensive road.
One might make a comparison with a starter home, parts of which can be improved: when the couple want extra room for the baby, they will start knocking the house down and creating dust and disturbance.
The decision has been taken and we have the road now. Paragraph 1.11 on page 11 confirms that a two-lane motorway would have been cheaper to construct. It also says that, under normal conditions, traffic forecasts do not justify the addition of a third lane. How much cheaper would it have been to construct a two-lane motorway?
I do not have an exact figure. It depends on whether the idea is to build the motorway to two-lane standards, or to build it to two-lane standards in such as way that it can be economically upgraded to three-lane standards. The second approach saves little because the earthworks and bridges have to be constructed to accommodate three lanes from the beginning. Although a two-lane motorway would be cheaper, the benefits to the road-user would be very much less. This is the primary trade route to the south and it carries a very high percentage of heavy goods vehicles, so a third lane is very useful in preventing fast-moving vehicles being slowed down by banks of lorries using the two internal lanes.
I will follow up on that. Mr MacKenzie, you said that we were perhaps liable to focus on the asset rather than the service. You explained how the decision was made within the department—that it would clear the path if the statutory requirements were done and the environmental statement was produced and so on. Why did you not consider consulting the bidders about alternative ways of providing that service before you had done that?
I think that there was a fair amount of consultation with potential bidders. We invited expressions of interest, and six such expressions were received at the beginning. Those were whittled down to four, and ultimately to two. At that informal stage, before any commitments are made, there is a process by which potential bidders can express an interest in exactly what the project will be like, and we can bounce ideas off them. However, it is an informal process, not a workshop.
Mr Howison, do you want to add to that?
We must bear in mind that, originally, this job was going to be procured conventionally. Therefore, procedures were put in place for a conventional contract well in advance of the decision to go along the DBFO route. Inevitably, for conventional procurement, the tendency is to build the optimum economic standard of road at the beginning, rather than to convert it later on. With the statutory procedures before us, the question is this: do we want to alter or amend those statutory procedures, to allow a variation? In this case, we felt that by the time the question of the two-lane or three-lane option was raised, we would have been so well into the competition that we might have opened ourselves to challenge by bidders if we had changed the nature of the competition in such a major way.
The Official Report and I hate acronyms. I would appreciate an explanation of DBFO.
Design, build, finance and operate.
In paragraph 1.10, on page 11 of the NAO's report, it is clear that the economic case for the project was critically dependent on the traffic forecasts that were made at that time, which were updated in 1995 and 1996. That is also clear from the answers that you have given. How do the traffic volumes now compare with the original forecasts on which the decisions were based?
Quite a lot of work was done on traffic forecasting, as it is critical to the cost of the scheme. We have not had much time, since the road was opened, to pursue the forecasts, but traffic volumes are broadly in line with what we projected.
So there has been no change to the basis for the economic case?
I want to add something about justifying the standard. Two approaches are taken in justifying the standard. The first is to consider the capacity of the road. That gives a broad idea of the sort of carriageways that might cope with the projected amount of traffic. The second approach is to conduct a cost-benefit analysis, to work out which is the most economically advantageous option. In this particular case, although the traffic could be carried either by a dual two-lane or a dual three-lane carriageway, at the moment, the economic assessment pointed us clearly towards the dual three-lane carriageway.
I have a question on a slightly tangential matter. Some of my colleagues will want to pursue the question of shadow tolling with you, but the implication of that is the possibility of real tolls being introduced at some point. The Executive is considering what kinds of road charging might help to tackle problems of congestion and so on. Have you carried out any calculation of how the introduction of real tolling might affect the traffic volume, and thereby the economic case in this instance?
According to the partnership agreement, road charging will be considered where it seems sensible to do so. No decision has been taken on the introduction of road charging on this motorway or anywhere else. The Executive is still considering the responses that it has received to the consultation document "Tackling Congestion". That is where we stand, formally.
Do you want to add anything to that, Mr Howison?
The terms of the contract compensate the contractor only for diversionary effects of traffic reduction on this road, not for absolute effects caused by tolling. In other words, if tolls did not have a diversionary effect when applied across the network, the road would be cheaper to the department.
The NAO report notes the fact that the Executive required bidders to include an option to upgrade to motorway standard the adjacent section of the A74 over the border between Gretna and Carlisle. That was to be done at a predetermined price until January 1999. Apparently, the costs were £48 million for the construction and £7 million for the maintenance over the contract period.
I accept that the tendering costs for the English section have necessarily had to be borne as part of the shadow tolls. However, it is not possible to quantify exactly how much is attributable to that section. The company has not revealed those figures and we are not entitled to demand them. We have no reason to believe that the sum would be significant.
That was a very comprehensive answer, thank you. Can I confirm, then, that you cannot quantify the increased costs of the Scottish road as a result of the option being included?
No, because we do not know precisely what the incremental addition to the costs of design and tendering for Autolink were, unless it tells us. We have no entitlement to demand that information.
Has the question been asked at all? I know that you have no entitlement to make demands, but has the question been asked within the department?
We have asked the question, but we have not received an answer.
I know the feeling. With the benefit of hindsight and the fact that the option was not exercised, do you think that it might have been wise to ask the bidders to provide a two-stage bid—one for the Scottish section and an add-on bid for the English section? I am interested in how you would justify the extra costs that the Executive has incurred, bearing in mind that if the Highways Agency goes ahead, a new commercial arrangement will need to be entered into.
The original proposition was that the option would have to be exercised by the end of December 1998, which would have meant that it started before the completion—in April 1999—of the Scottish section. The construction phase could have been an end-on, if not entirely continuous, operation. There was no question of leaving it for some years and then returning to it, but that is precisely the situation that now arises as a result of the Highways Agency not taking up the option.
I do not want to go back over old ground, but I refer you to paragraph 1.15 on page 13 of the report. I want to address not so much the specification of a three-lane road as the flexibility bidders have to provide a solution. The NAO makes the point that the department did not allow bidders the flexibility to investigate alternative solutions. Have you any comments on that?
Apart from the specification that the road should be a dual three-lane motorway, bidders had ample flexibility to suggest innovations in the way in which the service could be delivered. Figure 3 on page 15 of the NAO report gives an example of the kind of innovation that they introduced. Within the constraints of the envelope of land that had been subject to statutory procedures and the stipulation that it should be three lanes rather than two, there was a good deal of scope for design innovation. That is shown by the way in which the surface and the vertical and horizontal alignments were chosen, and the way in which some bridges, embankments and cuttings were designed. We think, and I think that they would endorse this, that there was ample scope for innovative design—some of which must have contributed to the earlier completion date.
I would like to return to some of the innovations later, but I have finished with that section.
You said, Mr MacKenzie, that you were awaiting an answer from your English colleagues. How long have you been waiting?
We are not waiting for an answer now. The option under our contract expired on 31 December last year. Although we are party to the steering committee that is looking at alternative solutions, we are no longer bound by that arrangement. This is now entirely a matter for the Highways Agency.
I am interested to know how long you waited for an answer when you needed one?
The organisation of which we asked the question that has not been answered was Autolink. We asked it out of curiosity—with the meeting with you in mind—if it could say how much was attributable to the design and tendering of the English section. It has not answered us and we are in no position to press it. It is the private consortium that has not responded.
I would like to concentrate for a moment on that question. "Out of curiosity" was the phrase that you used about the costs of tendering and so on.
That was taken into consideration, but we could not, at that point, identify what the costs were likely to be. It was the medium-sized risk that we took. There was a sharing of risk. The English section not going ahead was part of the risk that the department undoubtedly took.
So we have no way of knowing the costs of its not going ahead? It is clear that the sum of money may not be vast—nothing like the cost of construction—but are you satisfied that, if such a situation arises in future and there is no scheme for when projects do not go ahead, you will have the appropriate information and will be reimbursed?
I can see the desirability of that. I think that, as a result of this experience, we would be even more careful in entering into such an arrangement.
You talk picturesquely about the Cumberland gap, but it is not a green field between the two sections of road, nor a muddy track. What are the significant differences between the so-called English section—the Cumberland gap—and the motorway standard? As I recall, there is some similarity between the two.
The phrase "Cumberland gap" is not of my devising; it has been used in the Carlisle local press. The road is similar to the gap in the M8—on the A8 between Newhouse and Baillieston—where there is a discernible difference in the quality of the road and there are far more access points. The A74 between Guardsmill and Carlisle is an all-purpose road: a tractor can come on to it, which motorists would not be expecting if they had just come off 130 km of motorway. This is a question of the kind of traffic, the number of access points, the number of lanes and the fact that motorists, some of whom may have driven all the way from Birmingham on continuous motorway, suddenly find themselves on that stretch of the A74 and may not be prepared. It is a question of safety.
If there is a significant difference, would that be expected to show up in accident rates? How can the Highways Agency give the scheme a low priority if a cost-benefit analysis shows that, because of accident rates, upgrading would be highly desirable? Is there a high accident rate on that stretch of the A74? What figure came out of the cost-benefit analysis? Was it a plus or a minus?
The cost-benefit analysis examines the statistical likelihood of accidents. The section of road is relatively short, so the actual accident record would not necessarily be statistically reliable. The cost-benefit analysis will take into account the average accident rate for an all-purpose road and will compare that with the average rate for a motorway. I do not have the actual accidents figure for you.
I have some concerns about that. The document seems to suggest that it would be neat—and no more than neat—to complete the motorway section, whereas we are trying to assess whether public money should be expended on something that was necessary. Is one of the lessons to be learned from the project that, in similar situations, there should be more intensive investigation as to whether there would be a real cost benefit or just an apparent gain from upgrading the road?
We have to remember that this is not a Scottish road. We were not responsible for making the justification for it; our role was simply to act as an agent for the Highways Agency of the Department of Transport. How the analysis should have been undertaken, therefore, was largely a matter for that agency.
But the Scottish Executive is still bearing costs for this, so it is not quite as simple as saying that it is a matter for the English authorities, is it?
Ministers gave a high priority to completing the motorway section, to the extent that when John MacGregor announced the go-ahead for DBFO projects in England he included this scheme. We had previous experience of cross-boundary co-operation with the Highways Agency and its predecessors at the A1, where we undertook a joint contract to improve the road, and at the A74, where we had already constructed a length of the English motorway and had been repaid. With hindsight, of course, one would come to a different conclusion, but at the time we felt that there was a high degree of commitment from the Highways Agency. We had no reason to believe that it would not pursue the option, provided that the statutory procedures could be completed.
Can we turn to the chosen method of payment? Paragraphs 1.22 to 1.24 on pages 16 and 17 of the NAO report show that shadow tolls create a risk for the private sector that can result in higher prices. The private sector is no better placed than the public sector to be able to predict traffic flows. Paragraph 2.12 records the difficulties that bidders encountered in meeting the requirements that were set out, particularly on the point that Euan mentioned on the English section. Why do you think that using shadow tolls was a sensible way in which to reimburse the contractor? What other methods, if any, were considered?
We acted in line with Government policy for roads DBFO projects at that time. Eight English DBFO schemes had been awarded on the same basis, so we adopted that pattern. This project was seen as part of a clutch of road schemes that the then Government was anxious to fund in this way. We should bear in mind the fact that, when we let this contract, we had not had the benefit of the NAO's report on the first four DBFOs; that report was not published until January 1998. The Public Accounts Committee, following its hearing on that NAO report, published its report in June of that year. The scene moved on after the time at which we had to reach a decision on this scheme, which was seen as part of the first stage towards a customer-financed road operating industry and, as I said, was part of that clutch.
When you say "this approach", do you mean the alternative approach or the approach that you adopted?
The approach that we adopted—shadow tolling.
So, even with the benefit of hindsight and despite all the other options that have been highlighted, you believe that shadow tolling was the best option.
Even with the additional costs that shadow tolling may have involved, this project is likely to remain value for money.
I appreciate that it may still be deemed to be value for money. However, given the alternatives that were highlighted in the previous audit report and the consideration of the scheme in London, would you still say that we should have gone for shadow tolling? Do you think that we should now consider the other options?
Obviously, the orthodoxy has moved on and methods have been refined. The criticism here is that placing the volume risk with the contractor was not entirely fair, in terms of the contractor's degree of control and ability to manage the risk. The principle is always that risk should be allocated to where it is best managed, and the volume of traffic is not entirely within the management control of the operator. It may be possible to refine risk sharing and to devise a better and more satisfactory method, although we should bear in mind the fact that the operator signed up to this—it was not forced in any way. However, the new approaches may be more satisfactory, and we will certainly be prepared to consider them.
Paragraph 3.18 and figure 17 on page 44 of the report show the sculpting factors that were agreed between Autolink and the Executive. Those have the effect of skewing the shadow toll payments towards periods of high expenditure for Autolink, such as debt repayment and major maintenance works. Taking into account the sculpting factors, how much traffic and revenue risk is Autolink bearing in this project?
Behind the question of value for money, which must always be paramount, is the issue of the categorisation of the project and its accountancy terms. To be classified as off the balance sheet, the project requires the beneficial ownership of the road to lie with the operating company. That is decided not by the physical ownership, but by the rewards that the company obtains over the period of the contract. It also depends on the risks that the company takes. Inevitably, to the extent that the transfer of risk will be expensive, the company would endeavour to reduce the risk that it takes on board to the minimum that is required to satisfy the accountancy treatment. The answer to your question is that Autolink will have taken on board what it considers to be the minimum risk in relation to the accountancy treatment, having regard to producing as low a tender as possible.
I can accept that—that is obviously what all companies want to do. What are the real financial risks that Autolink have taken on in terms of variations in traffic flow?
Figure 16 on page 43 shows how, in sculpting the bands and the different tolls per vehicle-kilometre, Autolink has had to calculate what it would expect to get back over the 30 years. The slopes are its calculation of what risks it thinks it could bear. The reason why the lower bands are more expensive to the department than the higher bands is precisely the unfolding of the degree of risk that the company is prepared to carry.
We now move to questions on the choice between private finance and conventional procurement.
I want to focus first on your public sector comparator with the private finance initiative project appraisal. The net benefit that you have calculated is £17 million. What is the error term in that? What is the sensitivity analysis in terms of the scope for it to be incorrect in your calculations?
The report says that the department's forecast of the public sector comparator was reasonable. It was checked out for NAO by PriceWaterhouseCoopers and Halcrow, which confirmed that it was generally well founded—I think that that was the phrase used. The report acknowledges that there can be no absolute certainty about the calculation of a public sector comparator, especially in an instance such as this where we have a new process with the inclusion of operation and maintenance agreements is novel to such a transaction.
Let me put it this way. The report also says that the estimated overstatement in the comparator was about £9 million. By my calculation, the benefit of £17 million is subject to an error of more than 50 per cent, which makes the benefit calculation very inaccurate. Is that correct?
The £9 million is well within the compass of the £17 million.
Yes, but it is more than 50 per cent out, which is pretty poor. Do you agree with that?
I do not see that as poor in comparison with other public sector comparators and the extent to which one might have room within the two figures.
Okay. That is in terms of actuals rather than proportions.
The figures are in the report.
Yes, but is there any reason why that should not be the case across other projects in which your department is involved?
All the department's projects are subject to examination by the NAO. I do not see any difficulty if, as in this case, we agree with the NAO that the tables recording the figures should appear in a published report.
I want to go back a stage. The £17 million benefit, or the £9 million benefit—whichever it is—requires operational savings of between, roughly, £42 million to £50 million, compared with the finance cost, which is estimated to be £33 million higher. Can you specify the saving efficiencies in the private sector comparator compared with the public sector that gave you those figures?
I am sorry, I did not quite understand the question.
If, as the report states, the finance costs are £33 million higher but the net benefit, in your calculation, is £17 million, the efficiency savings from the operation of the scheme, compared to the public sector option, should amount to £17 million plus £33 million—that is, £50 million. Where are these operational savings identified in the comparison between the public and private sectors? What are the main sources of the private sector efficiencies?
We can do no more than talk about the generalities of where those efficiencies might have arisen, as we do not have a detailed analysis of Autolink's bid, which created those efficiencies.
Thank you for that useful response.
There was a positive net present value down to, I think, 5 per cent, where the gap was £8 million; however, at 4 per cent, it was negative.
So, it is around about the 5 per cent mark. Is that calculation easy to obtain for most projects? Is it straightforward, as it comes from the cash flows? Do you go for that sensitivity in every case?
We chose to do so in this case. At the time that the project was entered into, the test discount rate for roads projects was 8 per cent, as rail was still in public ownership. Investment in roads and rail had to compared on a level playing field. In February 1997, the Treasury told us that 6 per cent would be more appropriate in future, in line with the general guidance on public sector investment projects. Because of rail privatisation, the condition was no longer applicable. In future, we would want to test—as we did here—to a range of discounts, as the NAO report on the Highways Agency's first four projects enjoined us to do.
Thank you—that was useful.
Good afternoon, gentlemen. Since this arrangement, I understand that the Executive has not provided other roads, using the PFI approach or otherwise. Can you give us some idea of the costs involved and the implications for your budget? In other words, for the next 30 years, what proportion of the annual moneys available for new roads is the scheme going to take? Why have you not used this approach to make a lot of money available to tackle road congestion and so on?
It will not surprise you to hear that the roads programme in Scotland has been diminishing since 1994-95. We estimate that the payments on this scheme will be £23.6 million this year and about £24.2 million next year. That represents our payments to Autolink for the shadow tolls in respect of this project. That part of our budget is spoken for by this commitment.
What proportion of the total budget for this type of work is that?
The budget for motorways and trunk roads in the current year is £163 million, so you are comparing—
But is that £163 million all for this type of thing? As I understand it, that figure is broken down into different programmes. Can you score the £23 million or £24 million against the part of the roads budget that is allocated for this type of work?
The motorways and trunk roads budget pays for the construction of new roads and the maintenance of existing ones. That is the £163 million in the current year, of which £23 million is taken to pay for this scheme.
But is it not true to say that the £163 million is broken down and that some of it is already committed elsewhere?
Yes.
Is it true to say that that £24 million—a substantial part of the budget—was available and uncommitted?
That amount was budgeted for. We are in the current financial year, so the £163 million is already committed on maintenance and the completion of the roads that are under construction.
What I am trying to get at, Mr MacKenzie, is a breakdown of the £163 million. It is not just an overall sum—it is allocated to separate headings. Can you give us an idea of what those headings would be and how this £24 million per annum might score against those headings?
By comparison with a road that we are building by conventional procurement, for example?
Where does that amount fit into your overall budget?
Our overall budget is £163 million; £24 million of that is for this road. Are you asking what the rest consists of?
Your overall budget is £163 million, which is not solely for the provision of new roads.
No, it is for maintenance of existing roads—
Exactly. That is the point that I am making. I would like to know how the £163 million is divided up. Some of it will be for the maintenance of existing roads. What proportion of the new roads budget does the £24 million represent?
Can you give an answer now?
I will have to come back to you on that. You want to know how much is spoken for by maintenance and how much by new construction, however that is financed.
We want to know where the £24 million fits in to the overall budget, because it is a major commitment of money. We would appreciate it if you could give us the figures.
Why have you not used PFI schemes subsequently?
We have not had any new projects subsequently. It was intended to use that approach for the M8, but the project was cancelled last year as a result of the comprehensive spending review.
We are concerned with value for money. Is any consideration given to affordability, with regard to the revenue that might be available? The advantages of PFI are that you do not need to find the capital and the cost is spread over 30 years, but that has a significant effect on the available revenue. Has any consideration been given in your budget to affordability, in terms of factoring in future projects and value for money? Is that part of your value-for-money considerations?
The determination of value for money is a calculation that is done internally, as demonstrated in the report, by comparing what it costs to build a road by PFI, discounted against the cost of building it by conventional means.
It may be better if I did not use value for money in its technical sense.
Perhaps you mean affordability.
If you have a limited amount of revenue, as is clear from looking at your overall budget, a disproportionate amount of it is spoken for for the next 30 years, unless there is a significant change in the roads budget.
Precisely. I have no idea what the roads budget will be in 30 years' time, but I do know that by virtue of this arrangement a proportion of it is already tied up. That judgment had to be made when we entered into this project.
So affordability did not come in to the considerations as to whether PFI was an appropriate way to finance the budget?
No, I am not saying that. We would not have entered into the contract had we not been able to afford the £23 million this year.
I do not find that answer to be satisfactory.
What we could not have afforded this year was the £214 million that it would have cost to build the road immediately.
Are you saying that the capital costs were £214 million?
That would be the equivalent cost of building the road by public sector procurement. The cost would not all have fallen in one year.
To be fair, you have been rather selective in the figures that you have produced for us. You talked about what the revenue costs of the shadow tolling would be in years one and two, when they are very low. You are now aggregating 30-year costs and allocating them in year one. Some of those costs are maintenance costs and not capital costs.
We have to budget for both.
I am not denying that you have to budget for both when you are doing it to have a valid comparator, but I do not think that you are offering us that.
Perhaps I may step in here. The preconception is that all of the cost of this project is new construction. In fact, about 45 per cent of the total money involved in this project will be for maintenance.
I accept that a significant proportion of the money is for maintenance. You have already explained the significance of the road—it is the main route, and no one disputes that—but I am concerned that we have tied up so much money for 30 years. If we are constrained by a fixed budget, which the Scottish Executive is, there will be no flexibility to do other things. I do not see evidence anywhere in this report of a comparison of affordability with a fixed budget.
The report, rightly, covers value for money. Affordability was considered when the decision to take the road project forward was made. At that time, it was decided that three such schemes should be taken forward. One of the factors was to ensure that the composite cost of the three schemes would not be too great a proportion of the money available. As Mr MacKenzie has said, though, after the contract was awarded, the amount of money available for trunk roads was reduced. The issue is not that the money represents a high proportion of the overall money that is available, but that allowance has been made for the scheme even though the money that is available has been reduced.
I would like to add that the answer that I gave to Mr Wilson's question was not my opinion but a quotation from paragraph 3.39 of the NAO's report. The NAO said that its review
I think that that condemns the procedure in any case. If we cannot have robust calculations, what is the point in having a comparator? That question is especially relevant if the comparator is a way of making judgments at the margin and if the comparator has a degree of variation that is perhaps 5, 10 or 15 per cent.
The document simply says that the levels of precision are not very high, but the levels are robust.
I would like to consider the part of the report that is concerned with the procurement process. It seems that this was an innovative process, particularly in the way that two bids were short-listed for a second round of bidding. Was that to ensure that there was a level playing field and that unconditional bids were received? I am interested in what led to that decision.
The report says that we managed what the NAO regards as a competitive process and that we fielded an effective project team. It quotes the bidders as saying that they found that the negotiations were conducted in a reasonable way and that communications and confidentiality were maintained.
That is helpful, Mr MacKenzie. I thank you. That leads on to consideration of the additional costs. My colleague, Mr Macdonald, will want to question you more specifically on costs to the department.
If, in a future project, we found that we were unable to attract high-quality bids without offering to reimburse costs, that is clearly something that we would have to think hard about.
At the moment, you are open-minded about that? You would leave a decision until whenever the situation arose again?
When we got into negotiations, I would want to know what the bidders were prepared to do.
Thank you very much.
I have a question on the additional cost to the department. Pursuing any PFI contract involves extra costs and advice on dealing with the complexities. However, the figures—particularly in figure 11 on page 31—show that your outturn costs were more than twice the budget that was set by the department. Of those extra costs, how much would you have saved if you had chosen a conventional design and build project, instead of a design, build, finance and operate project?
We would not have produced a fairly sophisticated public sector comparator, for a start. That would have been unnecessary. We would still have required the additional work that we found was necessary on the traffic projections—to know how much load this road was going to take—although we would not have required that for the same purpose of shadow tolling. We would still have required whatever the extra cost turned out to be, in those circumstances, for the English section.
The public sector comparator was the one thing that you identified as incurring an additional cost. What share of the total outturn do you estimate that would account for?
You will see that, in the middle block of figure 11, opposite Scott Wilson, our technical and engineering advisers, the cost of developing a public sector comparator—which was not budgeted for—contributed to the difference between the £711,000 budget and the £1,883,000 outturn. However, I do not have a note of the precise amount that is attributable purely to the public sector comparator.
The implication is that most of the extra outturn would have been attributable to the public sector comparator? Is that correct?
No, because that amount includes all the items that you see in figure 11: the additional work that arose from the second round of bidding, which I mentioned to Miss Goldie; the addition of the English section, although we might have contemplated that under conventional procurement; the verification of the statutory orders; and the overrun on the production of tender documents. Of course, there would still have been a tendering cost with conventional procurement.
So, quite a small share of the additional outturn would have been saved in those circumstances? What about in-house costs such as the legal involvement in your team, and so on? How much of that is attributable to the fact that this was a privately financed project?
None of the West Merchant Bank costs would have arisen if there had been conventional procurement. Nor would any of the additional legal secondee costs have arisen if there had been conventional procurement. The Percy Thomas Partnership costs would still have arisen. As Mr MacKenzie said, it is a very complex engineering contract and it would have had a very high value, even as a design and build contract. We therefore think that a substantial amount of Scott Wilson's costs would still have been involved in tendering for a design and build contract, but we cannot tell you what the proportion of those costs would have been.
Adding together the merchant bank costs, the legal costs and a substantial part of the £1.9 million payment to Scott Wilson, I guess that we are talking about between a third and a half of the total outturn costs. Is it fair to estimate the figure to be in the region of between £1 million and £2 million?
It could be of that order.
How did you deal with those extra costs in appraising the private sector option? Did you add those to the winning bidder's price when assessing whether the contract represented value for money?
When we consider whether a proposition represents value for money, we always disregard the sunk costs so that, by the time the final bids are in, much of that expenditure would already have been spent one way or another. Therefore, that did not come into the assessment of the bids when they were finally made.
So you have not done the sum for the extra internal costs, nor have you included the comparisons that you made?
That is right.
We have had quite a long session and I would like to cover all aspects of the situation, so I ask Brian Adam and Andrew Wilson to finish off this section.
Following on from what Annabel Goldie and Lewis Macdonald were asking, I note that you originally had six people seeking the contracts and that four of them made it through the pre-qualifying stage. I also note that the bidder that was ultimately successful did not make it through the pre-qualifying stage. Do you have any concerns about the qualifying conditions, bearing in mind that, in terms of value for money, Autolink was the only bidder to put in a bid that was better than the public sector comparator?
Although we had six applicants originally, one of them—the Wimpey Morrison Babtie group—fell out because of Wimpey's merger with Tarmac Construction. That provided the opportunity for Autolink to strengthen its team significantly with the additional engineering expertise of Babtie. It also brought in Barr, a major Scottish contractor with significant local knowledge. We believe that our decision, based on the analysis of the six contenders, was correct. It was the strengthening of the team within Autolink that allowed it to do so well in the end.
Let us move on to the use of private finance, which is mentioned in paragraphs 3.4 and 3.5 of the document, and to whether, through this project, we can gain experience to strengthen public sector provision. What specific skills do you believe banks and investors have brought to the project that could not have been provided by the Executive?
They can bring the fruits of long experience. The discipline that is imposed on a contractor by the people from whom he has borrowed money, the monitoring that is applied and the finance mechanisms that exist between him and them, appear to drive that contractor to meet delivery times, to contain costs and to meet specifications in a way that has not traditionally been experienced by the public sector in its dealings with contractors.
That is not to say that the public sector does not make improvements. I have an analysis here of 42 schemes, with a total value of £540 million. The first batch of 17 schemes consisted of contracts that were awarded in the old-fashioned way. We call it measure and value—in other words for example, the contractor is paid for each cubic metre of cement that it supplies. Those contracts, which were widely used in the industry before 1990, were notoriously expensive. The average increase of the outturn over the tender was 33 per cent.
Presumably, that greater level of scrutiny has a cost element, or do you think that it is simply the result of greater experience of managing money? Have you any other suggestions as to what may motivate investors and banks to exercise stricter disciplines on a project than Government departments?
I think that it is the result of their experience of knowing what level of risk they can pass on to investors, and of balancing the optimum level of risk with the amount of scrutiny they need to devote to the project to give investors confidence.
In that case, could you and Mr MacKenzie say what factors were causing the cost overruns in the first series of projects to which you referred, which you were later able to squeeze out?
The old-fashioned contract forms did not produce a particularly effective incentive for contractors to restrain their costs when they discovered difficulties that they claimed they were unable to foresee at the time of tendering.
Can you give us an example from recent years, prior to this project, of how much of a problem cost overruns and delays have been for major road projects in Scotland?
The figure of 33 per cent that I cited is an average. Cases ranged up to more than 60 per cent cost overrun, so it was a very serious problem. Not only was the overall cost higher than it needed to be, because there was no incentive to reduce the price, but the variability of the costs was difficult to manage.
We need to press on.
I have the answer to Mr Adam's question about the proportion of the budget. Would it be convenient to give that now, or would you rather that I supplied it in writing?
Can you give us the answer now, so that it can be put on record?
The figure of £163 million breaks down into £93 million of construction expenditure and £70 million of maintenance—current expenditure. Within that £70 million, £38 million is for routine and winter maintenance and £32 million is for other current expenditure, including the shadow toll payments. The current expenditure does not come out of the £93 million for the construction and improvement of motorways and trunk roads.
Is it fair to say that £24 million comes out of the £32 million?
Yes.
We are always impressed by instant detailed answers. We will take a question from Andrew Wilson, to finish off this part of the meeting.
Let us try to square the circle of the value of having a private lender. Of the five contracted consortia and companies, one is the Innisfree PFI Fund.
If I may say so, that is not altogether so. If working outwith busy times costs in terms of lane penalties or lane occupation charges, one will do the work when the cost is least—for example at night, when occupation charges and penalties are low. However, if one is not under that constraint, one will do the work in the daytime, when one does not have to pay staff overtime for night work. That means occupying the lanes and inconveniencing motorists, however, which brings an economic disbenefit at peak motorway usage times.
If the decision is of benefit to the public sector, however, you, as a procurer, could simply tell the maintenance team that it must operate in the evenings. There is nothing to stop you doing that; it is not a great skill that must be bought from Innisfree PFI Fund. It is a simple aspect of project management that you could do. I do not see the connection between private finance and operational savings.
The connection is that—however it was done—the project came in £17 million below the public sector comparator, which had no financing cost, despite the fact that the private financing cost £33 million. That is not a fact that can simply be ignored.
We now move on to risk management.
One of the attractive features of PFI is the ability to pass out risk and for risk not to be retained solely by the public sector. What evidence has emerged so far from this particular project to demonstrate that risk transfer is resulting in value for money, or is it too early at this stage to assess?
Figure 14 shows that we have achieved the intended allocation of risk to those parties best placed to manage it, but we are, effectively, only in year two of 30. Having looked at other PFI projects, the NAO concludes in its report that the manner in which we placed the risk was consistent with obtaining value for money, not only due to the quality of the performance standards specification that is written into the contract, but due to the discipline of the compensation arrangements—what we will receive if the performance standards are not met—and of the financial investors, as I mentioned in response to Mr Adam.
More specifically, are you aware of any risks that may crystallise, which could increase the cost of the contract for either party?
One risk that has crystallised is that the English option has not been exercised, which was the department's risk. It was reckoned to be of medium likelihood that that would occur, but as we explained earlier, it has. There is always a possibility of legislation from you, the European Community, or whoever, which might have a bearing on the arrangements—I confess that I am in the realm of speculation. That risk is shared.
What is the nature of the compensation provisions in the contract? Will they be subject to further negotiation if any of the risks is manifested?
The contract would be applied to provide the remedy for a particular risk. Negotiation would not be about the principle, but would simply be about whether compensation figures were agreed.
The contract would simply be interpreted as at the date of the emergence of the risk in the risk claim and the underlying principle of the contract would not be renegotiated.
That is right.
A provision has been built in to compensate Autolink, for example, on the assumption that we might enter the world of real tolls. How is it possible to make an informed decision on quantification of such compensation, when it is so unpredictable?
Autolink is not compensated for a general drop in traffic if tolls are introduced. It is compensated only for the effect of a differential in the spread of traffic across the network. The answer is that we maintain traffic counts on all the roads in the network, so it would be easy to see whether the proportion of traffic using, for example, the A1 rather than the M74 increased or decreased.
So there is a formula to determine how the compensation provision would be applied?
That is right.
Are you satisfied that the contract, which cannot be renegotiated, has not been over-generous to Autolink?
Yes, as far I can be, looking into the future.
I call a very patient Paul Martin to continue on risk management.
The road was delivered some six months earlier than Autolink's scheduled date, which the Executive had felt at the time of bidding was ambitious. In paragraph 3.9, the NAO estimates that that might have increased the cost by more than £7 million. Given that the road has now been completed successfully, what is the current estimate of the increase in the cost of tolls as a result of early delivery?
The current estimate is still about £7 million—£7.6 million. It is marginally variable because of what, in the household analogy, is called snagging. We will not sign the certificate for the project until the final snagging has been carried out. That has an effect on how soon we start paying the full shadow tolls. We have been informed that all the work required to certify the road was completed by last Friday. Mr Howison hopes to go down to Beattock tomorrow to sign for the project, so your question is timely.
Can you be specific about the impact on this year's current roads programme, as £7 million is a fair sum? The public will want to know where you have found those resources and how you have offset the costs. Can you be specific about where we have lost the £7 million?
The £7 million does not all fall this year. We expect to pay £23.6 million this year. Payment has been advanced by a few months, but it has not been at the expense of other projects that we might have started this year.
Can you be specific about where that £7 million, which we no longer have in our control, has been offset?
Bear it in mind that, in any given year, the amount with which one starts each maintenance project is flexible. If we thought that we were running out of money, we would rephase the timetable of maintenance schemes, to reduce the claim on the total budget.
Before we move on from that, I want to clarify whether you have overrun by more than £7 million. Is that not the case?
We have not overrun, because we have not started to spend it yet.
However, the project has overrun by £7 million and we have lost £7 million.
We have not lost it, because we have gained considerable benefits. Road accidents are much less likely on that stretch of road because construction disruption has now been removed. Furthermore, the removal of the 50 mph speed restriction along the road has allowed traffic to move more speedily, which results in economic benefits.
I move on to paragraphs 1.10 to 1.12, which tell us that, considering the forecast traffic volumes, under normal operating conditions, there is no justification for upgrading the road to three lanes. However, the contract was structured in such a way that the winning bidder—Autolink—had an incentive to complete the three-lane motorway as soon as possible, to increase the volume of traffic and the income from tolls. That is shown in figure 17. Why was the contract not structured in such a way as to eliminate the potential risk of Autolink securing additional payment for early completion? Can you elaborate on that?
Much of the gain in time has been as a result of the innovations that we have been discussing. Autolink's technical engineering achievements have assisted in that. Furthermore, the precise timing of such a project is a function of good luck, the weather, the technical capability of the employees and the volume of equipment that is dedicated to the project. Finishing early has been achieved at a cost to the contractor—it is not just net gain. The contractor had to spend more to finish sooner.
The other question that I would like to ask relates to the additional costs that would have been incurred under traditional procurement, which are not comparable with those of the contractor. Is it fair to say that those costs would have been incurred under a publicly funded project, so the marginal benefit of PFI is very slim indeed? We need more details on that.
Which marginal cost are you comparing?
The difference between a PFI project and a public project.
That is the £17 million?
Yes. Can you give us a comparable cost? The additional costs would not have been incurred under a traditional procurement process.
Do you mean the additional costs of earlier completion?
Yes.
Conventional contracts often include built-in bonus payments for contractors. In this, the additional costs would depend on whether we had included such bonuses, had we gone for a conventional contract.
I want to clarify the benefits of finishing early. Clearly, there is the extra cost of £7 million spread over 30 years. Then there is the economic benefit of between £5 million and £9 million. That is a recurring benefit—or is it a one-year benefit?—but it is also a notional sum rather than an actual sum. While there is a comparison with an actual sum spread over 30 years, the benefit is in truth a notional sum—is that correct?
All cost-benefit analysis is on that basis.
Right. Thank you.
Are there any other questions for the open session?
You were asked earlier about the change in the traffic forecasts between 1995-96 and the present, and whether there were any discernible changes that might invalidate those forecasts. Your response was no. Can we conclude then, that in your judgment there has been no effect on traffic volumes from the increase in the fuel tax escalator?
I do not think that we could pull that out from such a short period of monitoring. A whole range of factors will affect the traffic on this particular road. Fuel charges will be one of them. Broadly speaking, traffic levels are in line with the predictions that we made.
One factor that has emerged is that the Executive's policy on lower car usage will have an impact in time. Alternatives to shadow tolling will have an additional attraction if they can be made to work and if the policy is not to be inconsistent with the Executive's integrated transport policy.
The discounting arrangement gives a figure of £217 million in terms of the costs. What are the actual costs? You have told us that the costs in the first two years would be £23.6 million and £24.2 million. What do you project as the total cost over 30 years?
In pound notes, by 2027 the cost will be £728 million.
It will be £728 million.
If the Government's policies to reduce car use and transfer freight to rail succeed over the next 30 years, will that save us money on the project?
That is one of the risks that Autolink runs—that vehicle kilometres will or will not exceed the point at which payments are capped.
That would reduce the £728 million, for example.
By the very nature of these inquiries, they tend to be negative. I will focus on two positive aspects. Paragraph 1.18 on page 15 highlights the innovative procedures that Autolink used to renegotiate parts of the alignment of the road and the accommodation underpasses. Those procedures could be picked up by public bodies. A little lateral thinking would probably result in some savings for the public purse.
In design and build contracts, we have moved to help contractors alter the specification in that way, to help provide savings.
Did I understand correctly when you said that the Executive had no input into the materials specified in the contract, the reason being that maintenance would fall to Autolink and that materials were not relevant to the Executive? Was there no reference to materials at all? I did not quite understand the point.
I was trying to say that, in this case, we could be much more relaxed about allowing materials that might be unproven in Britain. An example, given in the report, is the use of stone-mastic asphalt as a surfacing. At the time, we had exclusively used hot-rolled asphalt for such work.
If there is little input into the choice of materials and a problem develops, the costs might fall to Autolink, but what about the disruption to the general public and the loss of economic efficiency?
If Autolink has to undertake remedial works on the road, it will have to pay the lane occupation costs that are built into the contract.
I will indulge myself by asking a quick question. The first 78 rpm record that I bought was Lonnie Donegan singing "Cumberland Gap". However, the Cumberland Gap remains a residual problem. I would not like that area to become a notorious accident spot, which is a possibility, given what is happening elsewhere. Can you indicate when that residual problem will be solved?
I am afraid that that is not within our gift—it is entirely a matter for the Highways Agency. Nevertheless, the Scottish Executive ministers have influence in such matters and will doubtless make their concerns known.
I hope that they will so do and that they will read the Official Report.
Are you referring to the test discount rate?
Yes.
We understand that it is proving to be appropriate in the view of the Treasury for public sector investment appraisals. Until we see guidance to the contrary, we expect to apply that rate.
Thank you. No doubt, we will consider those matters further.
Meeting suspended.
On resuming—
The committee will now discuss the final two agenda items, which include the next steps in this inquiry, the preparation of a draft report and future business. I believe that it would be more appropriate to hold those discussions, which will include contributions from our advisers, in private, so I propose to consider the final agenda items in private. Is that agreed?
Thank you. I am afraid that the committee will now move into private session. I thank the press for attending today.
Meeting continued in private.