DISCUSSION
L'achèvement du Projet tient du miracle compte tenu des différences culturelles linguistiques, morales et sociales. La réussite résulte probablement dans l'adhésion d'une majorité à un objectif commun (TML Manager responding to UCL survey, October 1993).
As a starting point for analysis, it will be useful to place the performance of the channel tunnel project in context. The benchmarks provided by the RAND Corporation survey (Merrow 1988) of megaprojects (>$500m @ 1984 prices), which cover the period from the commencement of detail design through to handover, can be compared to the peformance of the channel tunnel project, as shown in table 4. If the facts that the programme data for the Channel Tunnel also include scheme design and that this project was at the top end of the RAND Corporation's sample in terms of scale and complexity, are taken into account, then the performance of Eurotunnel and TML can only be described as good. The Channel Tunnel was not a failure in conventional project management terms.
Performance Criterion
Megaprojects Average
Channel Tunnel
Budget Increase
88%
69%
Programme Overrun
17%
14.2%
Conformance Quality
53% performance not up to expectations
performance as expected
Operational Profitability
72% not profitable
operationally profitable, but overwhelmed by financing costs
Table 4 Project Management Benchmarks
However, the project simply is not financially viable, and the world's savers, via rights issues and additional loans from the syndicate banks, have been obliged to pour something like £6bn* more than they intended into the project. The financial restructuring, incomplete at the time of writing, is very much an attempt to make the best of a very difficult situation. The following discussion will attempt to identify how a relatively successful project got into such a situation, so that lessons can be learned for other major projects in Europe and elsewhere.
The symbiotic relationship between TML and Eurotunnel is fundamental to this story. The attempt to clearly delineate the responsibilities of concessionaire and principal contractor did not work. Jonathan Aitken, a local British MP, described the quarrels between the two as a "lover's tiff" (cited Kakoullis 1991 p 44). The more appropriate metaphor is parent and child. TML's members acting as promoters gave birth to Eurotunnel and nurtured it early on; Eurotunnel then had to find its own way in the world and therefore publicly proclaimed its independence from a hurt parent The problem remained that the child was tainted in the eyes of the world by its parent. For some, such as the deposed Directeur John Reeve, "things started to go wrong when we pulled out of ownership of the Channel Tunnel; we should have stayed there as owner operator - just like Trafalgar House has done on the Dartford river Crossing" (The Sunday Times 8/10/89). This view surfaced again when rumours that the French members of TML wanted to take over Eurotunnel spread during the middle of 1993 (New Builder 14/5/93). These points are manifestations of a profound ambiguity in the role of the members of TML - they promoted the project as a concession contract, but were not prepared, or able, to take the risks of being concessionaire. This ambiguity soured relations on the project because Eurotunnel was obliged to try to appear to be playing tough with TML in order to secure the confidence of bankers and investors, which led at times to public displays of almost pure theatre. In the opinion of Jack Lemley, the Chief Executive of TML, the rows themselves cost as much as £1b* and caused the job to be 9 months late (Anderson and Roskrow 1994 p199). This is almost certainly an overestimate, but it is clear that they did not help the project.
These problems were reinforced by the inappropriate choice of contract type for the terminals and fixed equipment, which was driven by the insistence of the financiers for risk transfer (Byrd 1994; Cohendet 1993; Stannard 1990), rather by an appraisal of the appropriate balance of risks between client and contractor. It is no accident that most of the acrimony revolved around the lump sum contract, for this was the one which transferred risk largely to TML. At the same time, Eurotunnel could not fulfill the basic condition for the use of such a contract type - no changes in technical requirements. This inability was inherent in the structure of the project. Firstly, the time scale required for the completion of the project on a privately financed basis compressed the programme so that, in effect, it was a fast-track project. Secondly, critical aspects of the design were subject to a public regulatory authority, the IGC. As others have noted (Merrow 1988; Stinchcombe and Heimer 1985), regulatory authorities are not project orientated, and are not happy making decisions in rhythm with project milestones. Thirdly, Eurotunnel was a novice client building a unique facility. Major changes in technical requirements were inevitable on a project of this type, and under such conditions it was totally inappropriate to attempt to transfer the bulk of the risk to the contractor.
The general air of mistrust between TML and Eurotunnel generated by these factors meant that there was a strong tendency towards bureaucracy - "puisqu'on se méfie de ses partenaires et puis qu'il faut justifier ses propres compétances, on mulitplie les contrôles tatillons, on fait appel à des sociétés de consultants coûteuses en faisant grimper par là même les coûtes de transaction et de contrôle" (Cohendet 1993 p 20). The large number of different actors involved - contractor, concessionaire, financiers, concessor - all preferred to retain their own independent advisors. One commentator summarised the situation as thus: "there are so many advisors and sub-advisors involved in the project that I hate to think of the amount of paperwork being produced. One thing is for sure, none of them are digging a tunnel" (The Sunday Times 8/10/89). The generation of this massive weight of transaction costs on the project is a product of the complex set of simultaneous interactions between the actors compounded by deep mistrust between the two principal ones. In turn, the financiers, at least in the early stages, were suspicious of the nature of their relationship, which therefore generated further antagonism between Eurotunnel and TML.
Given the structure of the project, changes in technical requirements were inevitable, yet the mechanisms for negotiating the time and cost implications of those changes appear to have been totally inadequate. Three third party actors were involved here - the MdO, the Disputes Panel, and the Arbitration Panel. None appeared to have adequate legitimacy to resolve the intractable disputes between the parties. In the end, the issues were resolved by the two principal parties sitting down and thrashing out the issues round the table, but it was only after the middle of 1993 that real progress was being made along these lines. In the meantime, the uncertainties involved compounded the project management problem. For instance, it was clear from May 1991 that the June 1993 deadline could not be met due to delays on the procurement items and the M&E works, yet agreement on its revision took another two years. In the meantime, TML managers were obliged to work to deadlines which they knew to be unrealisable (Anderson 1992).
TML itself faced a number of problems although, perhaps surprisingly, these did not appear to stem, for the most part, from its binational or ten member consortium nature. This may be due to the large-scale recruitment of staff especially for the project, and a testament to the wisdom of Henderson's comment which opens this paper. However, the appointment of an American to the strengthened Chief Executive role in 1989 does suggest underlying national tensions in the move away from a twin-headed organisation. Apart from the tensions associated with this change, TML's problems seemed to stem largely from the pressures generated by a highly uncertain and hostile environment. The high turnover amongst senior management must be a cause for concern. Although any project organisation evolves through a number of phases related to the progress of the works, the main shifts in organisation structure within TML appear to have only been possible with the departure of senior managers and their replacement by outsiders. It seems unlikely that this was due to problems of resistance to change, at least after 1989; more probably, it was due to the extraordinary pressures experienced in the job.
These problems were compounded by a fundamental weakness in the structure of the consortium. The 10 members were essentially civil engineering contractors, and they "viewed the project very much as a civil engineering one" (Byrd 1994 p 24). One third of the contract sum was, however, for power engineering and rolling stock items, and it was the rolling stock which experienced the most ferocious cost inflation. While TML did recruit senior executives from rail operators such as SNCF, this is not the same as having corporate involvement to back up the skills. The inclusion of a power and railway engineering corporation in the consortium, such as GEC Alsthom, would almost certainly improved cost forecasting and control for these items. It is notable that most other consortia and joint ventures for rail concession contracts do include such corporations.
This weakness in the structure of TML is symptomatic of the failure to properly scope the project. The fundamental assumption of all the parties at project definition stage around 1985 was that this was a civil engineering project. This is reflected in the rousing conclusion to the UK government White Paper which argued that "the fixed link is a challenging and exciting project. It will be the largest civil engineering project for many years and the largest in Europe ever undertaken by the private sector" (cmnd 9735 1986 £ 64 my emphasis). It is to this failure of scoping that the failure to plan adequately for the commissioning process must be attributed. As one informant put it, what is being built is not just a tunnel, but "an integrated transport system" (interview 12/2/93), but the implications of the systems nature of the problem were not fully appreciated by senior management until too late.
These problems might have been recognised earlier had the engineering of the fixed equipment and termini not been starved of cash during 1987. It is clear that not enough effort was put into engineering prior to commencement on site. The combination of pressure from the banks, the civil engineering culture of the project, and uncertainty due to the tortuous British legal processes led to a lack of full engineering effort in the early stages. This meant that crucial aspects of the design had to be changed relatively late, in addition to the problems associated with the IGC.
Pressure from the banks pushed the project into fast track mode - it went on site after only around 15% of the programme had elapsed. This contrasts starkly with the Severn Bridge (see Working Paper 12), which did not go on site until one third of a 6 year programme had elapsed, on a project that, while technologically challenging, did not feature the systems complexity of the Channel Tunnel.
Organisationally, the decision to let the construction contract to a single integrated consortium meant that any particular contractual dispute affected the project as a whole. The usual strategy in the face of complexity is to break the overall problem down into smaller problems. This was the strategy adopted on Storebælt (see Working Paper 14), where the bridge and tunnel project was let as four separate packages, and the rail transport systems were taken outside the construction contract. While this project has faced enormous technological problems, particularly in the east tunnel, client/contractor relations have never become as bad as on the channel tunnel, and it has been possible to manage each problem separately. Both of these solutions of allowing more time for design and breaking the project down into a number of separate packages has been adopted to an even greater degree on the channel tunnel rail link (Hirst 1996).
It is often argued that the external environment poses some of the largest risks for projects of this type (e.g. Merrow 1988). Certainly, the relatively rapid growth of the latter part of the eighties caused problems - price inflation was serious, particularly for the procurement items. However, this same boom also allowed the refinancing of the project in 1990 and 1994 on the basis of upwardly revised revenue projections based on increased traffic forecasts (Stannard 1990). In the end, these revisions were robust enough to withstand a 69% cost increase. What they could not withstand was a 12 month overrun in construction time; and, in particular, an 18 month delay in the provision of a full revenue-generating service. The main dynamic of risk on the project was internally generated by the structure of the project - the decision to fast track due to the source of finance; the choice of an inappropriate contract type due to the source of finance; the family feuds do to the necessity for Eurotunnel to prove its independence to the sources of finance. It is difficult to escape the conclusion that the attempts to reduce uncertainty in order to enhance financing capability are, of themselves, the main sources of the uncertainty, and hence cost and time overruns, on the project. The project was lucky because these errors were absorbed by unforeseen favourable changes in the external environment
The Channel Tunnel project is one of the largest and most complex integrated construction projects ever undertaken. It demanded an extraordinary range of technical expertise, deployed on a massive scale. It was inevitable, then, that things would go wrong - in retrospect, it is remarkable how many things have gone right