Abstract
On 21 April 1997, a White Paper was produced evaluating the costs of providing the newly-formed Scottish Assembly with a “home fit for the Scottish Parliament”. The intent was to have the building ready for the 2001-2002 session of Parliament, and the estimate was in the range £24.5 to £34 million.
In July 1998, the site of Holyrood Hill in Edinburgh was selected and the estimate revised to £55 million. Costs and delays rose uncontrollably and, in May 2003, with the cost approaching £300 million, and the delivery date still a year away, a formal enquiry under Lord Fraser was launched to understand the causes and responsibilities in the “Holyrood Project”. The report was delivered on budget and on time on 15 September 2004. The Holyrood building was finally ready for business on 7 September 2004 and was officially opened by Queen Elizabeth II on 9 October 2004. Total costs are estimated to be in excess of £450 million.
One underlying error was that everyone involved considered this endeavour to be a project – albeit a very complex one – whereas it needed to be considered from a wider, program management point of view and managed accordingly. This paper consolidates the details provided in the Fraser report as well as other sources into a consistent framework based on seven key control areas on which effective program governance should be based and with respect to which the governance effectiveness of any program can be assessed. The Holyrood team never effectively recognized the need for defining or even of understanding inception-to-delivery lifecycle management, dependency planning, baselining, change management or pro-active control of risks.
This massive undertaking, ignoring the cost and time overruns has, however, succeeded in delivering a stunning architectural achievement for the city of Edinburgh and to Scotland as a whole. This success could however have been achieved with less expenditure of time, cost and effort if correct governance procedures had been effectively applied.
A view of the Public Entrance of the Scottish Parliament building
Alert MPs Action This
The Rt Hon Lord Fraser of Carmyllie Q.C. was invited by the First Minister of the Scottish Parliament to undertake an inquiry into the cost over-runs and the delays in the construction of the Scottish Parliament building in order to produce “…. a full account of all of the key decisions and factors which have determined the costs and value of the Parliament…and to identifying the lessons to be learned for the procurement or construction of major buildings in the future.” (Fraser, 2004). This report took input from all available sources including two investigations by the Auditor General for Scotland (AG1, 2000; AG2, 2004). It should be noted that the need for effective management was recognized early: “Good governance of the Holyrood project requires the client, as the investment decision maker, to have complete, reliable and relevant information to inform stewardship and monitoring.” (AG1 23). It is however disappointing to discover that the Auditor General (in 2000) made eight very pertinent recommendations, but that no effort was made to treat his complaints and act on his recommendations. The goal of the current paper is not simply to rephrase the findings of the earlier reports: instead, it refocuses them within the framework of management best practice and draws lessons that should be generally applicable in similar situations and, as such, provides a sound theoretical basis on which the administrative and control structures for complex programs can be built.
Its Charms Let to Spain
Once the site of Holyrood had been decided, the choice was made to select the architect by international competition. It is not clear what the full set of decision criteria and the overall process were. However, the final decision was to retain the well-known Spanish architect Enric Miralles. A number of questions have been raised as to the wisdom of the selection and even the way in which the selection process was carried out, since this process did not seem to comply with all of the corresponding European requirements. This is addressed in more detail later in this paper.
Mortals, It Isn't Cheap
Exhibit 1: Evolution of cost forecasts over the life of the project (in £millions)
The initial White Paper on 24 July 1997 indicated costs in the range £10-40 million. Once the site of Holyrood and the architects had been selected, cost estimation could be made on a sounder basis. However, this did not do much to control the escalation, as shown in Exhibit 1. Similarly, the dates did not fare much better, as shown in Exhibit 2.
Exhibit 2: Evolution of completion date forecasts over the life of the project
How could the situation not only get out of hand, but also be allowed to carry on in such a way? The basic reason was that none of the basic rules of governance or sound project management were ever applied – compounded by an incompatibility between cultures: project management and public administration, creative Spanish architects and Scottish construction practices.
Patterns Limit Chaos
There are a number of ways in which control can be applied to the construction of a new building:
- You can sign a contract with an architectural company who will then carry out the work in an orderly manner in line with agreed budget and timelines.
- You can retain additional flexibility, control, and risk by appointing a management company to work for you in coordinating all of the contractors (“Construction Management”).
- You take total control and responsibility, and coordinate all of the contractors yourself.
The level of risk to the buyer is lowest in the first option and rises with each of the others, and, similarly, for the buyer's required level of involvement and technical competence. In all cases, however, the customer needs to be aware, willing, and able to control the progress of the major phases of the work. These phases (RIBA, 1973) are
shown in Exhibit 3.
Exhibit 3: Phases and main stages of a building project, with go/nogo reviews at the end of each phase
The management team for the project took three major decisions which worked together to compound their risks: they chose a foreign architect with no experience of working in Scotland or with Scottish contractors; they adopted the “Construction Management” approach (option 2 above) in order to exploit its flexibility in order to compress the timescales and they decided to compress the schedule by overlapping the phases (and therefore the stages) shown in Exhibit 3. An attempt to represent their “very fast track” sequence is given in Exhibit 4. What should be noted in that diagram is that the approval points (1 to 4 shown as diamonds) are no longer placed between activities. That is to say that no review or approval is needed before initiating the following activity. A clear example of this is that Operations Onsite (i.e. construction) was being carried out before Stage D was complete. Since Stage D corresponds to “Scheme Design”, at which all of the top-level plan is complete, it should serve as a coordinating and consolidating framework prior to the detail work and choice of subcontractors: “Two necessary elements of a robust Stage D are to be found in the agreement by all parties of a Cost Plan and in the existence of at least a preliminary structural design by the structural engineers. Neither had been achieved by June 2000.” (FR, 11.3). However the dependency that Stage D had to be complete before Stage K (Operations Onsite) could start was bypassed “in the interests of speed” in the Holyrood project. This is one reason why, to use an expression of Lord Fraser's, the wheels started to drop off!
Exhibit 4: Concurrent approach that was adopted, showing potential approval points as diamonds
The dotted lines indicate where the flow of activity bypassed all best practice forms of control
The way in which this concurrent working was compounded by the subsequent delays is shown in Exhibit 5 (AG1, 2.22): the diagram shows, for example, that there was a period of 15 months of construction prior to the completion of Stage D.
Exhibit 5: How concurrent working got out of hand – already in August 2000
The Auditor's 2004 report states (AG2, 5.58) that “for construction management to be successful, the common boundaries between the trade contracts and how they will integrate with each other must be carefully defined.” For the Scottish Parliament, the overall work included not only the main building but also the upgrade and modification of surrounding buildings, and this work should have been structured as a program, in which each major trade contract was defined and let as a well-specified project within the structure and constraints of the overall program and under the coordinating control of the construction manager. Only in this way would it have been possible for the work to be kept under control despite the “fast track” approach adopted; although the program was run in fact track mode, all of the details for each component project could still have been complete and kept under effective control.
Map To This is Central
In order to structure the analysis of the issues, a hierarchical “program governance model” (PGM) has been developed for this – as shown in Exhibit 6
Exhibit 6: The Program Governance Model
It is illuminating to map some findings from the Fraser Report (FR) and the prior reports of the Auditor General for Scotland (AG1, AG2) onto the PGM (the paragraph references from the Reports are included):
Communications control provided neither effective communication not control:
□ “Cost reporting by the project team upward to the project owner and the client was reactive and not regular and systematic in relation to the total costs of the project.” (AG1, 3.51)
□ “Surprisingly, Ministers were not informed of the resignation of the Project Manager until his departure was picked up by the media…” (FR, 8.43)
□ “When [the Quantity Surveyors and Cost Consultants] did submit estimates, they were either withheld from Ministers or altered”. (FR, 8.44)
Scope control.
The estimates and expectations were based on a set of assumptions which all changed
□ “…assumptions…the Old Royal High School would provide a suitable location.”. Once this assumption was rejected, the objective fairly swiftly changed from “a building suitable for the Scottish Parliament” to “a design worthy of the hopes and aspirations of the Scottish people.” (FR, 1.6) – without any debate or formal approval of such a radical departure from the initial vision.
Planning control
There never seems to have been an integrated view of all that was to be carried out
□ “There should have been a formal project execution plan…There should have been change control procedures based on a detailed cost plan agreed between all the parties at an early stage.…there is still no firm cost plan agreed between all the main parties responsible for the project.” [AG1, 21]
Cost estimating was at best very approximate:
□ “The political balance may well have been between a budget that was sufficiently realistic…yet sufficiently low to allow for its political acceptability.” (FR, 3.35)
Dependencies between activities were not taken into account:
□ “…[the need to plan] trade packages with a long lead time to be completed in advance of adjacent and dependent packages”. (AG2, 5.60) It should however be noted that the idea of using any form of critical path method was made even more difficult by the level of change throughout the program, since alterations hit CPMs
Risk management was at best a theoretical exercise with no expectation that it should lead to any corresponding actions:
□ No provision was made in order to reduce the impact of risks to an acceptable level: “When in response to MSPs' views the client required changes to the design of the chamber in June and July 1999 there was no allowance in the programme for the additional, disruptive design effort required.” (AG1, 3.25)
□ “Although risk workshops…did identify owners for specified risks, there was no monitoring or feedback on subsequent action.” (AG2, 5.35)
Integration Control
No understanding of the checks and balances of the building life cycle was shown. There are repeated comments in the Fraser Report of “…although Stage D was not yet complete…”. (also AG2, 3.12)
Change management was at best reactive in that no provision was made to predict, let alone challenge any requests for change:
□ “Construction cost estimates have increased in total, from £50 million to £108 million (116 per cent). Almost half the increase is attributable to a 47 per cent increase in the total area of the building, which is now some 31,000m2. At the outset officials carefully considered how much space would be needed but their original target allowance for “balance” areas (for circulation, stairs, lifts, void areas, plant rooms etc) proved to be unrealistically low.” (AG1, 12).
□ “The significant delay in finalising the design has extended the site preparation and initial works period, which will have added to costs. Also, the notional high efficiency office block design was replaced by a more complex concept, which has evolved into a larger building with a higher quality of finish.” (AG1, 2.12)
□ For the assembly building concrete frame, 1,800 changes were recorded – and the completion date slipped from August 2001 to April 2004 (AG2, Exhibit 25).
Lack of teamwork in change management:
□ The artist complains - “…the only manner envisaged for cost reduction…causing…a paradox of having a cheap building costing a huge amount of money.” (FR, 12.13 quoting the architect)
Execution control
Any chance of controlling construction costs was undermined by the approach to tendering:
□ “…packages went out to tender…even if the amount of design information was less than would normally be expected.…The trade contractor can…ensure that any risks on costs taken to win the contract can be absorbed or eliminated in the subsequent design development.” (AG2, 4.33)
Similarly, fees for services such as construction management were uncontrolled:
□ “Fees and costs payable to the design team members and the construction manager are related to the value of the construction work. There is a need for project management to agree with these parties a firm basis for total remuneration…” (AG1, 1.41).
The basic administrative and team management actions were neglected:
□ “…although there were significant changes in the project organization since October 2000, the plan has never been updated…” (AG2, 5.13).
The risks arising from overlapping design and construction became clear early on but no corresponding actions implemented:
□ “Some works packages had been let and work commenced on site. But because the design remained uncertain there was an increasing prospect that essential construction information from the design team would dry up. This created a risk of wasting expenditure on consequential claims from contractors for disruption and idle time.” (AG1, 1.18).
And, eventually, not even any pretence at control:
□ “After June 2001 project management did not work within an overall budget or approved cost ceiling for the whole project” (AG2 2004). It is clear that an overall consolidation of the costs with respect to a single program budget was missing.
AG1 3.37 understates the matter most succinctly: “Measured against the Treasury guidance there were areas where the [execution] processes applied to the Holyrood project were incomplete”
Oversight control
No one was encouraged – let alone designated – to feel responsible for overall management:
□ The project director's job description states “…single focal point for day-to-day management…within budget…”. “In practice…there has been less emphasis on leading and making decisions…” (AG2, 5.12)
□ “There was not an arrangement which required project management to provide full cost information to accountable officers or to the client on a regular and systematic basis.” (AG1, 22)
□ “The current senior project manager…was appointed in 2001 but has not had a formal job description since taking up the role.”(AG2, 5.12)
□ “The construction manager…did not have control over costs or quality…has not taken responsibility for cost management and reporting and control.” (AG2, 5.12).
□ “…Cost consultant…monitor costs…but cannot instruct other parties regarding costs.” (AG2, 5.12)
Stakeholder control
The ownership of the project passed from the Secretary of State for Scotland, to the “Corporate Body” once the elections for the Scottish Parliament were complete. This led to a change of many of the key team members; there is no indication of any planning or risk assessment associated with this major change nor of the incoming team considering that this was an ideal opportunity for a comprehensive review.
□ “At this stage there was a particular need for those accepting responsibility for the project to have reviewed it, with a degree of independence from the project team, in order to satisfy themselves about its status and health, but there was no such review” (AG1, 3.64)
There was a lack of consistency between the chosen life cycle and the contractual approach, which reinforced many of the inherent risks associated with each by removing virtually all means of monitoring and control from the project management team:
□ “the Scottish Office decided…to choose the construction management route, but the design and construction teams were appointed using traditional contract terms that set fees as a percentage of approved construction costs.” [AG1, 20]. As a feedback mechanism, this could only serve to increase costs.
The team responsible for the management of the project seemed to adopt the attitude that they had no power and were at best servants - and at worst victims - of the architect:
□ Introduction to the Report: “No one tells Enric to think about economy with any seriousness.” (FR)
□ “Between July 1998 and November 1999 the lead architect attended only six of the fifteen meetings which project management convened to review progress” (AG1, 2.28)
□ “Project managers are replaceable…architects are not.” (FR, 8.40) Enric Miralles died suddenly on 3 July 2000!
It should be noted that the two reports from the Auditor General (AG1, AG2) are exemplary: the information is clear and well-structured; the conclusions are precise and well-founded. It is disappointing, therefore, to find that the recommendations did not lead to any major improvement in the governance of the Holyrood project and will probably not be integrated into future projects; the reasons for this would merit an analysis in their own right.
O.T.T.: Time'n'Cash Spiral
As in most situations, it is the combination of circumstances that builds up to the final result. In this case, the seeds were planted early and were nurtured continuously:
□ The initial goal was to provide a building; the final intent was for a historic monument (“his important castle”).
□ Responsibility was so diffuse that each stakeholder was able to focus on their own interests rather than those of the project.
□ The various reports, such as the initial one by the Auditor General for Scotland, carried out during the project, were never acted on nor attempts made to treat the shortcoming that were identified.
□ There was no baselining: technical plans, budgets, time scales were a mixture of wishful thinking, political messages and optimistic forecasts – a mixture of budgets so variable and time charts so pliant that they could never be used for control nor for developing corrective actions.
□ In the interest of speed, many stages were run in parallel, by-passing all controls.
□ Speed was always considered the highest-priority factor and took precedence over any attempts to trim cash potentials.
□ Since the pre-design phase was never effectively completed, tendering was carried out on partial plans and the selected contractors were then required to create their own designs – with the client picking up all of the costs.
□ On-site planning failed to optimise the use of resources so that the various trades got in each others' way and the level of rework was considerable.
□ Even when plans were established, they tended to be modified frequently.
□ The role and authority of the construction manager was never clear and the financial incentives were contrary to those of the project.
It should be pointed out that all of this is compounded into a vicious circle by the lack of phase control and formal review points, since these safeguards are built into any effective life cycle in order to provide a cut-off at which the synergistic feedback between the various disruptive factors can be interrupted.
This Last Importance – Review of the Findings of the Fraser Report
The Fraser Report produced 10 recommendations. These are compatible with the program governance model (PGM) described above, and alignment of them with the model provides a more comprehensive basis for future program control than simply adopting them piecemeal:
- “Where competition is held for the selection of a designer, consultant or contractor…there should be…a consistency of approach…”. This is addressed by Procurement Control in the Execution component of the PGM. The Report recommends doing this in line with Best Practice – with which there can be no argument! For the Holyrood situation, however, there are still unresolved questions as to whether this was carried out in accordance with European rules of fair competition.
- Use of an “internationally renowned ‘signature’ architect…”. The Report recommends “confirm[ing] compatibility” of “cultures and practices”. It is however probably unrealistic to expect any highly-creative architect to have the same respect for commercial process as is required for effective governance. For this reason, it is important not only to analyse the mismatch (stakeholder analysis), but also to accommodate for any discrepancies by adapting and reinforcing all of the other governance components, such as Stakeholder Control and Oversight Control.
- “Construction Management as a procurement route…”. It is my view that, although this procurement route served to highlight and exacerbate the shortcomings in governance, it was not the root case of the over-runs. It has, however, been taken as an acceptable excuse by all parties since it conveniently absolves all of the participants from potential blame. The corresponding Fraser recommendation is disappointingly toothless: “…[project management should] set before the political leadership a full evaluation of the risks”: evaluation without resolution is not worth the effort expended. The PGM on the other hand prescribes a strictly-tracked life cycle that incorporates the reviews and sign-offs that effectively mitigate the principal risks – whether from the procurement structure of from other causes – as well as assigning responsibility for effective decisions and action. In this way, Construction Management can remain as a potential procurement route for public building projects so long as the civil servants and local government officials comply with the rules of best practice and are held accountable for this; if this is not the case, it is unlikely that any other form of procurement would serve to protect the endeavour – or at least the taxpayer – from their shortcomings.
- “Procurement rules”. The report states “no one should be put in charge…without a demonstrable appreciation of what is required under EU procurement rules”. There can be no argument with this statement, other than to ask why restrict the need for competent leadership simply to the area of procurement?
- Information to Ministers “should not be filtered by the Civil Service”. This is a subset of PGM Communications Control and begs the question as to the rights and duties of the civil service in the management of such a project.
- “…governance should be as clear as is now required in the public sector”. The Report does not, however, give any indication as to whether the same penalties for poor governance should also be applied!
- “…give SPCB wider powers of delegation…”. For delegation and sub-contracting, the rules of authority must be defined (Planning Control) and established (Stakeholder Control) globally early on in the project – if necessary, as recommended in the Report, by explicit legislation.
- “…considerations of security and safety…should not be regarded as ‘add-ons’…”. A number of other ‘add-ons’ also appeared, such as landscaping: PGM requires a full review of the main components of the scope in order to identify and correct any major oversights prior to the design phase. It is clear that the events of 11 September made additional demands in the area of building protection. Within the PGM, these would have been catered for by change control – an area in which the shortcomings caused immense damage to the Holyrood project and is not even addressed in the recommendations of the Fraser report, although the Auditor General for Scotland described the need very clearly in his report four years earlier.
- Access to information: “…procedures of the Scottish Parliament did not allow oral questioning of the Presiding Officer.” This was altered in 2004, but could have been identified earlier if PGM Communication Control was applied.
- Contract provisions: “…full contracts, guarantees and bonds should be secured at the outset…”. This was achieved in 2004 – somewhat late in the day for the project. This protection of the public purse should be ensured by effective Procurement Control and the approach validated prior to committing any funds to the endeavour.
The Fraser recommendations are precise and well-founded; however, there is one major omission: the need for clearly defined roles and responsibilities with specified authority and accountability, to be linked into the overall Program Governance Model. This would allow all of the specific lessons identified by the reviews of the Scottish Parliament to be consolidated and would provide a framework on which to build effective governance in the public and the private sectors.
The PGM provides effective control of time and cost while enhancing inspiration by liberating the creative minds within the project from endless debates on ill-defined situations. In all future undertakings of this type, an independent body should be set up with the responsibility and authority to ensure that the work is carried out from the start and for the entire duration in accordance with this model.
Control and vision are not incompatible. However the stakeholders must understand that the more enchanting the vision, the more likely it is to make the stakeholders forget the rules, and that stricter controls are therefore required to turn the vision into reality: a dream that gets out of control rapidly becomes a nightmare.
Acknowledgements
Many of the headings and some brief comments within this paper have been created as anagrams of the words “Scottish Parliament”, with the assistance of the software available from www.anagramgenius.com. For example, this program suggested alternative titles to the paper such as “Help, I can't sort it, Sam” or “This Capital Monster”.
Photograph of the building on first page: Image © Scottish Parliamentary Corporate Body - 2005.