Hansard

Members should have a copy of my briefing statement and it might help them to follow the text, but I welcome the opportunity to introduce what is quite a long and detailed report and I hope that my briefing will help members to come to terms with some of the report's key conclusions and findings.I will start by outlining the background to the report in the context of my previous reports and I will explain the scope of the new report and my understanding of how it relates to the inquiry by Lord Fraser of Carmyllie. I will then describe the main events of the past four years and I will offer general comments about the project as a whole. I will go on to outline my views on the reasons for the late delivery of the project and for the increase in costs. I will also comment on some of the key features of project management and control and I will conclude by mentioning some of the lessons that I think can be learned from the Holyrood project.I start with the context of my report. My first report to the Parliament on the matter was made in September 2000, partly in response to a request from Andrew Welsh MSP, who was the convener of the Audit Committee at the time, and partly in response to the general concern in the Parliament and among the wider public about the state of the project in spring 2000. The report was unusual for two reasons. First, it was unusual for an auditor to examine a project that had not been completed—in summer 2000, the bulk of the spending on the project had yet to be committed. Secondly, my statutory remit ran back only to the day on which the Scottish Parliament came into being. I was grateful for the co-operation of the Scottish Executive in allowing me to inform myself about certain issues and events that occurred before devolution, but I did that only to the extent that was necessary for a proper understanding of the situation in spring 2000. Therefore, my 2000 report was not a full audit examination of the pre-devolution management of the project.After my report in 2000 was published, the Audit Committee expected me to keep the project under review and to make further reports, should the need arise. I reported for a second time in December 2002. That report concentrated on various contract management issues following the termination of a major contract with Flour City Architectural Metals (UK) Ltd and made recommendations for urgent action by the project management team to tighten control of the contracts.In June last year, I advised the Audit Committee that I intended to examine and report again on the project and I subsequently indicated that I would make my report in summer 2004. Also in June last year, the First Minister, after discussions with the Presiding Officer and me, announced that there would be an independent inquiry by Lord Fraser of Carmyllie. It was agreed that Lord Fraser would build on my existing findings and take account of my intention to examine the economy, efficiency and effectiveness with which resources have been used in the project.The report that has been laid before the Parliament today takes the story forward from my report of September 2000. I have not revisited the matters that I covered in 2000, nor have I altered my view on the conclusions and recommendations that I made in 2000. The most recent audit examination relates to how the project has been managed and controlled over the past four years. It describes the delays and escalating costs and it contains my key findings on the reasons for the delays and cost increases. In other words, my report concerns the management and control processes applied to the Holyrood project. I concentrate on the performance of what has been termed collectively the "project management". The project management consists of the clerk and chief executive of the Parliament, who is the principal accountable officer, and the project team, which is led by the project director. The legal client for the Holyrood project is the Scottish Parliamentary Corporate Body, which has powers under the Scotland Act 1998 to provide the new Parliament building. The corporate body formally delegated the function of completing the Holyrood building project to the clerk.Project management is responsible for managing and delivering the project. Project management is advised and guided by the Holyrood progress group, which was set up by the corporate body to look after its interests as the client on behalf of the Scottish Parliament. The corporate body stated in 2000 that it expected the clerk to act on the advice of the progress group and he has done so.I emphasise that, through the clerk, project management is accountable for the delivery of the project. The Holyrood progress group includes MSPs and independent experts. The progress group provides advice to project management and monitors progress on behalf of the corporate body. It has responsibilities for advice and monitoring, but it is not accountable for the delivery of the project. For that reason, I have not looked in detail at how the progress group worked. I imagine that Lord Fraser, with his wider remit, is likely to comment on how the progress group functioned on the basis of the extensive evidence that he has taken from members of the group.There is one final and important contextual point to make. Despite the slippage and extra costs that have affected the project, a great deal has been achieved in the course of the building operations. Several architectural observers have already reviewed the building favourably. Arguably, it satisfies the requirement for a high-quality, landmark building that reflects the aspirations of Scotland as a nation. We should bear that in mind when we assess the slippage and cost increases that have affected the project. I turn now to a brief summary of what has happened to the project over the past four years. Part 2 of my report includes a detailed analysis. I will draw out a few of the key events because that might assist members' understanding of the main findings. At the time of my September 2000 report, construction was at a very early stage. The frame of the MSP building was going up, but the site of the main assembly building was literally a hole in the ground. However, I concluded that there was a sound basis for advancing the project provided that my recommendations were followed through. In particular, an important milestone had passed when the approval of the stage D design occurred in June 2000. My report highlighted previous major changes in the requirements for the area and layout of parts of the building, with the client unable to freeze either its requirements or the design. Approval of the stage D design in June 2000 seemed to end that uncertainty and promised a firm anchorage for the project. The key findings and recommendations of my 2000 report are summarised in exhibit 1 on page 12 and in exhibit 37 on page 63 of my current report. In November 2000, three important things happened: a new project director started work; a cost plan was prepared; and a risk assessment was completed. In the early months of 2001, the cost pressures were building up and initiatives were taken to find cost savings. Progress on construction continued—for example, construction work started on Queensberry House and the major contract for the assembly building frame was approved. However, by June 2001, significant difficulties had come to the surface again. The corporate body concluded that the previous target of £195 million was not achievable. The Parliament approved a motion that removed the overall budget constraint of £195 million. Also in June 2001, the progress group was concerned about poor co-ordination, a misunderstanding of responsibilities for cost control and other aspects of the project. At that point, there was also a change of project director. During the rest of 2001 leading into 2002, the difficulties of project monitoring and co-ordination continued and extra project management capacity was brought in. The target completion date of December 2002 looked increasingly unrealistic and, in March 2002, the completion date was targeted for April 2003. Also in March 2002, the Finance Committee was advised that the project costs might reach £265 million. By September 2002, the figure reported to the Finance Committee was £295 million, which is equivalent to £309 million when one includes the landscaping costs that were excluded from the reports to the Finance Committee. In the autumn of 2002, a new programme was issued that moved the completion date to June 2003, although that programme was heavily qualified. In December 2002, the corporate body discussed options that included stopping the project, instructing a further independent review and putting a cap on the costs. The corporate body reported to the Finance Committee that the estimated cost excluding landscaping was now £311 million. Problems continued during 2003, with various measures suggested or introduced to improve project management and the effectiveness of the design team. In June 2003, the corporate body reported to the Finance Committee project costs, excluding landscaping, of £359 million. In August 2003, another programme was issued. It gave a target date for completion of the debating chamber by April 2004. There was intense construction activity during those months. In April 2003, there were 1,100 people working on the site. During the rest of the year, there continued to be problems in resolving design issues, managing congestion on the site and managing the interdependencies between contracts. By December 2003, construction activity reached a peak with 1,500 people on site every day. Also by December 2003, the MSP building and Queensberry House were substantially completed.In February 2004, the final version of the programme was issued and it included a target for completion of the debating chamber by August this year. Also in February, the corporate body reported to the Finance Committee a cost estimate of £431 million, and that estimate still applies today.I turn now to the reasons for the late delivery of the project. As I mentioned earlier, I intend to concentrate on what has happened since the events reported in my first report of September 2000. In December 2000, the completion date was set for two years later—December 2002. That was based on what seemed then to be a settled design and a firm budget. As I have described, there were successive delays until the current completion target of August this year. I want to make one very important point about the role of the corporate body as client. Media reporting over the past two or three years might have left people with the impression that the corporate body was changing its requirements, but that is not true. The client did not alter significantly its user requirements after it approved the stage D design in June 2000. Therefore, changes by the client in its specification did not contribute to the delays and neither did they contribute significantly to the costs. The construction costs of changes initiated by the client amounted to £600,000. That is just 0.2 per cent of the current construction cost estimate. My report has a chapter about the reasons for the late delivery. I have concluded that the main causes of the 20-month delay to the project since September 2000 were the production of detailed design variations and the late supply of information during the construction process. I will return to that shortly.Other factors also contributed to the slippage. One factor was the choice of construction management as the method of procuring the project. I described the features of construction management in my 2000 report and I summarise those features again in the current report.I suggest that, in general, construction management is unsuited to most building projects in the public sector; however, it was not necessarily the wrong choice for Holyrood. Different forms of contracting are intended to transfer risk to those who are best able to manage it. Under construction management, the design is incomplete and uncertain when construction starts, so the risk stays with the client. The client must manage design development and have a project team that involves professionals who are experienced in that construction method.Unfortunately, in the Holyrood case, project management did not implement construction management fully in accordance with usual practice. Experience and expertise in construction management were not present in the Holyrood project's early stages, so the client and project management did not fully appreciate the risks and challenges. Another major factor to be recognised was simply the challenge of construction. Major difficulties were associated with constructing a very complex and unusual building on a densely developed site against very tight deadlines. Anyone who has visited the site—as I have—and observed the extraordinary architecture can appreciate that. My report contains several images that show the design's complexity. For example, the chamber roof, which is shown on page 14, was an extremely challenging assignment in engineering design and construction terms. A third factor that explains the slippage was that absolutely no room for slippage was provided in the original timetable, which was very compressed and challenging to deliver. As a reflection of all those factors, the architects and some trade contractors found it difficult to deliver some critical elements of the work on time.As I have mentioned, I consider that the main causes of the 20-month delay to the project were detailed design variations and the late supply of information. Therefore, if the client did not change its requirement in any significant way, what happened? The corporate body estimates that some 10,000 change orders have been issued over the course of the project. It would have been completely impractical to examine all those changes as part of the audit. However, in exhibit 20 on page 41 of my report, members will see the cost consequences of changes in each of the 58 contracts that comprise the main construction works. When tenders were approved for those contracts, the estimated cost was £129 million. The current estimate of the final cost of those contracts is more than £220 million. That is an increase of 71 per cent.The stage D design, which was approved in June 2000, was the design team's firm proposal in response to what the corporate body as the client had specified, but it was by no means a fully developed design. The shape and area of all the main building elements were fixed but, in June 2000, much of the necessary detail was not present. The design changed as the architect translated the stage D design into the more detailed design that was necessary for procurement and construction. The client's needs stayed the same, but many changes were introduced to improve the design's buildability, performance and aesthetics. The foyer roof provides a good example of what happened subsequent to June 2000. In exhibit 21 on page 42, members will see computer-generated images from 2000 and 2001 and a photograph of the roof when it was under construction in 2004. Not only has the design changed radically, but the estimated cost has risen from £1.8 million to £7.3 million.In short, the project involved a very high level of design development. In my view, the design's complexity was not fully appreciated early enough in the project. Thereafter, I am not confident that project management controlled the design development process effectively.I emphasise that I did not seek to form any opinion on whether any individual contractor has been at fault. My emphasis has been entirely on the performance of the project management—the officials who have been responsible for the project. Project management is responsible for managing its consultants and contractors and for assessing their performance.Under the construction management approach for the Holyrood project, much of the design development happened at the same time as—not before—construction. Combined with the time pressures and the emphasis on quality, that made the design development process much more difficult to control than normal.At the root of the project slippage was the production and release of design information later than at the time agreed in the construction manager's programme, which the construction manager prepared in consultation with the design team and contractors. There was tacit agreement by those parties to the assumptions in the programme, but it was revisited and revised several times.The construction manager conducted eight programme reviews between December 2000 and February 2004. The client consistently maintained a drive for completion by the earliest possible date. However, programmes were heavily qualified. New target dates for completion that were agreed with the design team and the contractors were subsequently not achieved.Exhibit 23 on page 44 gives examples of programme qualifications. Exhibit 7 on page 19 summarises the six main programmes that were issued between 2000 and 2004 and the repeated slippage in construction in all areas of the site. The Holyrood progress group challenged the construction manager and the architects about the slippage, but no decisive improvement was achieved. New targets were set, but the fundamental problems of non-performance were not overcome. The audit examination considered progress on 20 of the largest contracts. The average delay in awarding those contracts was 37 weeks. On page 47, members will see an exhibit that describes the delays on eight large works contracts and they will also find another exhibit that shows the number of design variations after contract award for five critical works contracts. As an example, the assembly building's concrete frame was originally programmed for November 2000 to December 2001. However, the period of variations to that contract ran from January 2001 right through to March 2003, and some 1,800 draft change orders were issued.Some further delays resulted because the design team had to approve design elements that were provided by trade contractors. Another factor was the very tight completion date that the client required. As a consequence, some work was out of sequence and unproductive, which added to the delays.The challenges of delivering the programme were clearly enormous, but I am not clear that project management did enough to address the underlying causes of the problems. With each revision to the programme, all parties agreed that the new targets were achievable. However, the repeated programme slippage must raise questions about all parties' performance, which no one seems to have addressed effectively.I note in my report that although project management raised significant questions about some aspects of some of its consultants' work, it has not systematically assessed their performance. In the face of repeated slippage, project management should have managed performance by measuring what was being achieved against contractual obligations and enforcing those obligations strictly.I turn to the reasons for the increased costs of the Holyrood building project and start by repeating a point that I made earlier: the corporate body, as the client, did not significantly alter its requirements after the middle of 2000, but although the size and layout of the building did not change, the cost of realising the design escalated enormously. If estimated landscape costs are included, the actual total, with the £195 million target that the Parliament set in 2000, was £209 million. Since that time, project costs have more doubled to £431 million.There was a cost plan prepared in November 2000, and that plan underpinned the £195 million target. However, the question arises whether the basis of the £195 million target was adequate. The construction manager stated that there was insufficient design information to provide reliable cost estimates. In my view, much of the information in the November 2000 cost plan can be regarded only as an indication of a cost target—it was not a reliable prediction of the cost.A further weakness that contributes to my view is that, even before the cost plan was agreed, it was apparent that the impact of inflation would consume the entire £11 million contingency that lay within the £195 million budget. As a consequence, and although all parties accepted that risks were very likely to occur, no allowance at all was made for other risks. A risk workshop, which was held in November 2000, quantified additional costs for all risks including inflation at some £61 million. Although all those risks had to be eliminated to sustain the £195 million budget, it was not made clear how that was to be done.In the event, the estimated final cost of most of the trade contracts now greatly exceeds the original cost-plan allowance. Of the 58 trade contracts that I mentioned previously, 41 have an estimated final cost that is 21 per cent or more above the original cost-plan allowance. Five contracts—at a cost of £28 million—increased to between three and eight times their original tender value. In other words, all the project risks that were identified in November 2000 crystallised subsequently and the cost of those risks proved to be significantly higher than was first expected.Some of the increased costs have been due to the extended construction period; in any project, time slippage can mean that additional costs arise from prolongation, delay and disruption. Prolongation costs are the additional costs of doing the same work over a longer period. On page 54, two examples—the assembly windows package and the specialist glazing—show how time delay can result in extra costs.Much of the risk that was identified in November 2000 came from design decisions that were yet to be made. I have explained the high level of design development that affected the project. Because of the large amount of novel and complex features, realisation of the design of many packages pushed costs well above what had been estimated in the stage D design and the November 2000 cost plan. I return to an example that I mentioned earlier. Members will find set out on page 55 of my report a description of the impact on the cost of the foyer roof and the glazing contract. The stage D cost plan included £1.5 million for that part of the building work, whereas the current estimated cost is about £7.4 million.Another example that I touch on in the report is the impact of blast-protection measures. My report recognises that an important feature of the Holyrood project was the need to develop the design to satisfy security and blast considerations. Overall, I have concluded that the main reasons for cost increases since 2000 are design development and delay in the construction process. My reasoning for that conclusion is set out on pages 56 and 57 of the report and summarised in exhibit 32. In summary, construction cost increases fall into three main areas: inflation at £19 million; prolongation, disruption and delay at £73 million; and design development at £68 million. Another £4 million represents the project management's estimate of the extra costs that were incurred as a result of the demise of Flour City in 2001.My work included a review of 20 of the main construction contracts, which represent more than half the estimated construction costs. I show that works packages were often let at times when there was uncertainty about what work would be involved, which affected the project in several important ways. First, it was often difficult to achieve good competition and to deliver value for money. Secondly, where contracts were awarded with uncertain scope and design, project management was not well placed to resist extra time-related costs from contractors. Thirdly, and importantly, there was the ripple effect as a result of which uncertainty in one contract created difficulties in other contracts.The project has about 58 separate main construction contracts. For construction to work efficiently, the common boundaries between such trade contracts and how they integrate with each other must be carefully defined. However, a drive for early completion often resulted in contracts being placed before all the design work was completed. With the benefit of hindsight, I believe that that did not provide value for money. If work had been fully designed before the contracts were placed, the boundaries between contracts would have been much easier to define and integrate.Another consequence of letting contracts before the design was fully developed was the risk that true competitiveness would not be achieved. If the scope and design of the work had remained unchanged, it would have been easier to ensure that contractors complied with the programme obligations, but the changes meant that contractors were reluctant to accept responsibility for delay.All building projects are subject to change and design development. The Holyrood project is a large, demanding and complex project and the degree of uncertainty in many of its contracts was very high. On page 60 of my report, I identify 11 contracts—from a sample of 20 that were reviewed—in which evidence suggests that the design was insufficiently detailed at the tender stage. I also give examples of the uncertainty that was involved in each case.I turn now to part 5 of my report, which examines issues relating to project management and control. Any construction project must achieve a balance between time, quality and cost. For example, a tighter deadline for a project can be achieved if quality standards are lowered or if more money is made available. In 1998, the client required that the building be completed by summer 2001. Although time was clearly a priority, quality was equally important throughout the project. Construction management was seen as being the only method of delivering high quality within very tight deadlines.The client also set a budget at the outset. However, having carried out my examination, the importance of cost is not clear when it is assessed alongside time and quality. Although construction management is not a method that works well on projects that have fixed cost constraints, it was nevertheless the method that was chosen for the project. In 2000, Parliament set a fixed budget of £195 million, but by 2001 that budget was considered unachievable. Thereafter, although the client reported successive cost increases to Parliament, a new budget limit was never established. Any project should have an approved budget limit—if it looks likely that a budget will be exceeded, project management must take action to reduce costs or obtain extra funding. A decision on which course of action to take should be based on a reassessment of value for money at the new project cost level.A very important point was reached in June 2001. Parliament approved a lengthy motion, which—in my interpretation—removed the previous cost constraint of £195 million. The motion has been included on page 25 of my report as exhibit 11. During the debate, it was said that, although a firm figure for the final outturn cost could not be set at that time, there was no question of giving out a blank cheque. However, I question the wisdom of project management not seeking to establish a new budget or approved cost ceiling after the June 2001 debate. In my opinion, the absence of an approved budget contributed to weaknesses in cost reporting and financial control. I have described those weaknesses on pages 71 and 72 of my report.Although I recommended in my 2000 report that project management should report project costs on a monthly basis, it was not until July 2003 that regular reporting started to the Finance Committee of the total estimated costs of the project. Until July 2003, the reports were provided only quarterly or less often. None of the reports before July 2003 provided information about landscaping costs, and only four of the eight reports that were provided between June 2001 and June 2003 reported all the other main cost items. There was no other regular reporting of total project costs until in July 2003 the corporate body started routine monthly reporting to the Finance Committee on total project costs.I make an important distinction between budgets and forecasts in my report. A budget is an approved sum that is allocated for a project, which can be viewed as a cash limit—only the authority that approves a budget can vary it once it is set. A budget is not the same as a forecast. A forecast may vary throughout the life of a project as circumstances unfold.Every fortnight for four years, the cost consultant reported estimated construction costs to the client, but those projections were not a budget. However, project management seemed to regard the costs reports as setting a construction budget. Contractors' invoices were not checked against an approved budget; they were checked against the total estimated costs for the package that were reported by the cost consultant. The danger of confusing forecasts with budget limits is that forecasts will become self-fulfilling if effective action is not taken to contain costs within an approved budget. Exhibit 47 on page 73 shows how cost estimates by the cost consultant marched in step with the rising cost figures that were reported to the Finance Committee between 2001 and 2004.I will deal briefly with consultants' fees. The corporate body limited its exposure to increases in consultants' fees in 2003, but in my opinion project management could have taken more action at an earlier stage. For example, there was from 2000 an opportunity under the construction manager's contract for project management to convert part of its fee to a fixed lump sum, but project management did not secure that until August 2003. Therefore, it is possible that the client may have paid more on fees than was necessary over the life of the project. The absence of financial incentives to help to bring the project in on time and within budget has, in my opinion, been even more significant.My final main point relates to the overall leadership and control of the client organisation. In my view, the organisation of the Holyrood project did not provide the necessary clear direction and leadership. The responsibility for management and leadership was divided among several parties and there was no single point of leadership and control. Normally, leadership and control should reside with the project director and the client should give the project director responsibility for making the project happen within specified boundaries and performance targets. In the Holyrood project, leadership and control of the project was not clearly established in such a way.I invite members to examine the two diagrams on pages 66 and 67. On page 67, exhibit 41 shows the Holyrood project organisation from June 2000 onwards. That diagram also appeared in my earlier report and was agreed with project management at the time. The Holyrood progress group was established then; in 2000, I concluded that that organisation was appropriate. That is still my view. Members will see that the organisation has all the key features of the Treasury's recommended model, which is shown in exhibit 40. However, the diagrams do not show where the major risks lie. In a conventional contract that is run on the Treasury model, construction risk should be placed with the main contractor, in the central box on the bottom line. The main contractor would contract with, and pass some risk to, the suppliers of goods and services. Neither the project manager nor anyone in the higher boxes, which represent the client interests, carries any significant construction risk in the Treasury model.In the Holyrood project, things were quite different because the construction management method was used. Construction risk moved right to the top of the tree and has rested with the client, advised and supported by project management, the construction manager and the cost consultant. In my view, that is fundamental to understanding the problems that occurred in ensuring clear leadership and control on the part of the client organisation. The client could not, if you like, stand back from the contractor who was carrying the main risk and let the contractor sort out the delivery problems. There was no main contractor in Holyrood. Therefore, the client carried most of the risk and the project's management, who were advised by the Holyrood progress group on behalf of the client, had to manage that risk.A few moments ago, I said that the single point of accountability for, and control of, a project should normally reside with the project director. The need for absolute clarity about who was taking decisions, and the need for that person to have a clear budget limit to work with, was even more important for Holyrood than for a normal construction project because the client was carrying the construction risk. That was one of the main reasons why I emphasised in my 2000 report the importance of having a senior and very experienced construction professional in at least one of the three senior posts of project owner, project director or project manager.The project owner, who is the clerk and chief executive of the Parliament, is not a construction professional, and neither is the current project director, who was appointed in June 2001, before the main intense period of construction activity started. The senior project manager/project manager is a construction professional, but is less senior in the hierarchy and is not a single authoritative point of command. In my latest report, I have said that the project director has in practice acted as the senior project administrator who co-ordinates the interests of the various parties and the high-level communication and reporting. That is not a pejorative comment or a criticism of her; she did what she was qualified to do and, by all accounts, did it very well.The Holyrood progress group is an advisory and monitoring group, which was created by the corporate body and has no executive power. Therefore, it is not accountable for delivery of the project. That accountability lies with the clerk and chief executive, the project director and the project team. The Holyrood progress group was not accountable for delivery but, in practice, it quickly evolved into an executive decision-making body. I have given examples of its decision-making role in exhibit 42 on page 68. I do not criticise the group for that. There were professionally qualified and experienced people on the group as advisers and it is right that their expertise was used to the full. It is also perfectly understandable that, because the big financial risks were being carried by the corporate body—as the client—rather than by a contractor, the Holyrood progress group did everything that it could to support the project team in managing the risks. However, my conclusion is that there was not a clear single point of leadership and control in this very complex and challenging project. That has been a weakness in the system; it is not a criticism of the many individuals who were doing their best to manage the project.On page 8 of my report, I have listed the lessons from the Holyrood project that I believe are relevant for managing big public sector projects. Briefly, I say that the form of contracting must be chosen with care, with a sound understanding of where the risks lie and how they will be managed; that the construction management method is unsuited for most public sector building projects; that the gateway review process that allows scrutiny of a project at different stages should be followed; that there should be performance payment incentives for contractors; that there should be a single point of leadership and control; that performance should be measured as the project unrolls; and that there should be adequate time for planning before the project starts.I will summarise my conclusions on the management of the Holyrood project. I remind the committee that I did not seek to form opinions about the performance of the organisations that are involved in the project; it is for the client and project management to consider and address the performance of their contractors.Despite the slippage and extra cost that have affected the project, a great deal has been achieved. The building is now almost ready and in my opinion it is likely to satisfy the requirement for a high-quality landmark building that reflects the aspirations of Scotland as a nation. We are reaching the point at which everyone in Scotland will be able to make the aesthetic judgment that must be part of the final assessment of the project.A number of factors have caused slippages in the project, but the main causes of delay since September 2000 have been the production of very large numbers of design variations, and late supply of information during the construction process. Delays to the project have been a major reason for the increases in costs that I quantify in my report. The other main cause of cost increases has been that the cost of realising the design that was approved by the corporate body in 2000 increased enormously. I also quantify that in my report.There were enormous challenges in the project, but I am not clear that project management did enough to address the underlying causes of the problems of slippage and increased costs. In my opinion, the absence of an approved budget contributed to weaknesses in cost reporting and financial control. The construction management procurement method left most of the construction risks with the client. It was, therefore, essential to have clear direction and leadership to manage the performance of contractors. However, responsibility for direction and leadership was divided between several parties and there was no single point of leadership and control. That is a weakness in the system and not a criticism of the many individuals who were doing their best to manage this challenging and complex project.That brings me to the end of my comments on the context of my report and the opinions that I have formed as a result of the audit examination. Before I conclude, I wish to inform the committee about the status of my report with regard to clearance of matters of fact that it contains. Normally, I give accountable officers three weeks to comment on the facts in my draft reports. In the case of the Holyrood report, that period expired on 14 June. The arrangement is a convention that I adopt—there is no statutory or other requirement for me to do so. Prior to 14 June, there were discussions and exchanges of information between officials of the corporate body and the Audit Scotland team. All the concerns and representations that were made to us were considered and, where appropriate, amendments were made to the draft report.On 16 June, the principal accountable officer wrote to me indicating that he was not able to agree my report in the time that was available and that that would not happen until 16 July at the earliest, which would have delayed publication of my report until some time in August. As the committee is aware, I agreed to make my report available in time for Lord Fraser to use it in his report. However, if I had moved my publication date back to August, it would have been impossible for Lord Fraser to report in September.The principal accountable officer has indicated his disagreement with my interpretation of and conclusions on the management of the project. He is entitled to do so, because the interpretation and conclusions are my own, but it is not my practice to offer only conclusions that have the agreement of an accountable officer. The essence of my role as Auditor General is to offer an independent opinion based on my audit examination.However, it is unusual for me to present a report in respect of which the accountable officer has not indicated which matters of fact are of continuing concern. That was the position in which I found myself until late yesterday. In the hours since I finalised this briefing paper, the accountable officer has indicated the matters that are of concern to him; I will be happy to continue to work with him and his team to narrow any areas of difference and I will provide the committee with a note of the outcome of those discussions. However, in the short time that has passed since first I saw the accountable officer's submission, I have not found anything that causes me to change the principal conclusions of my report.In the meantime, I am personally satisfied that the audit examination has been rigorous and that my conclusions are based soundly on that analysis. It would be surprising if there were not, in a project that has been as challenging and complex as this, some differences of emphasis and interpretation. However, the audit examination has been thorough and I have, therefore, fulfilled the commitment that I gave to the First Minister and the Presiding Officer to assist Lord Fraser's inquiry to the best of my ability. I have done that by making my report to Parliament today.