The importance of risk analysis
As mentioned in Chapter 1, in terms of risk, too many feasibility studies and appraisals of megaprojects assume projects to exist in a predictable Newtonian world of cause and effect where things go according to plan. In reality, the world of megaproject planning and implementation is a highly stochastic one where things happen only with a certain probabil- ity and rarely turn out as originally intended. The failure to reﬂect the probabilistic nature of project planning, implementation and operation is a central cause of the poor track record for megaproject performance documented above.1
It is fairly common in feasibility studies and appraisals of major trans- port and other infrastructure projects to make a mechanical sensitivity analysis examining the effect on project viability of hypothetical changes
in, for instance, construction costs, interest rates and revenues.2 The typical range for such sensitivity analysis is from +10 per cent to +20 per cent. It is, on the other hand, unfortunately rare that risk analysis is
made by identifying alternative future states of costs, revenues and effects and a probability distribution estimated for the likelihood that these states would actually occur. This information is required in order to estimate the expected values of costs, revenues and effects, or, in other words, the most likely development, including the associated variances. Approaching risk analysis in this way is essential in order to curb what has been called ‘appraisal optimism’ and to give decision makers a more realistic view of the likely outcome of projects, instead of the incomplete and misleading view on which decisions are often based today.3 Risk analysis is also the basis for risk management, that is the identiﬁcation of strategies to reduce risks, including how to allocate them to the parties involved and which risks to transfer to professional risk management institutions, namely in- surance companies. In this chapter we describe how risk was treated in the Channel, Great Belt, Øresund and other projects. In addition,
review the implications of risk for costs and ﬁnancing. Finally, we spell out the lessons to be learnt regarding risk.4
Channel tunnel, Great Belt, Øresund and other projects
When Eurotunnel went public as a company in 1987, investors were told that the project was relatively straightforward. Regarding risk, the prospectus read:
Whilst the undertaking of a tunnelling project of this nature necessarily involves certain construction risks, the techniques to be used are well proven… The Directors, having consulted the Maˆitre d’Oeuvre, believe that 10%… would be a reasonable allowance for the possible impact of unforeseen circumstances on construction costs.5
Two hundred banks communicated the ﬁgures for cost and risk to in- vestors, including a large number of small investors. As has been observed elsewhere, anyone persuaded in this way to buy shares in Eurotunnel in the belief that the cost estimate was the mean of possible outcomes was, in effect, misled.6 The cost estimate of the prospectus turned out to be a best possible outcome based on the unlikely assumption that everything would go according to plan with no delays, no changes in performance speciﬁcations, no management problems, no problems with contractual arrangements or new technologies or geology, no major conﬂicts, no po- litical promises not kept, and so on. The assumptions were, in other words, those of an ideal world. The real risks for the Chunnel venture were several times higher than those communicated to potential investors, as evidenced by the fact that the real costs of the project were higher by a factor of two compared with forecasts.
Similarly, before ratifying the Great Belt project, the members of the Danish Parliament were informed by the Minister of Transport regarding risk that:
I do not consider the uncertainty of the overall construction cost estimate for a ﬁxed link across Great Belt to be larger than for other large bridge or tunnel projects carried out in this country.7
However, the cost of even the largest bridge and tunnel projects carried out in Denmark before the Great Belt link was less than a tenth of the original Great Belt budget, and none of these projects included a bored tunnel.8 So for reasons of size and innovation alone, the risks associated with the Great Belt link were more important than for any other trans- port infrastructure project in modern Danish history. In addition, the geological and technological risks – and hence the risks associated with
construction costs – were higher. Nevertheless, no ﬁnancial or economic risk analysis was made for the Great Belt project.9 As in the case of the Channel tunnel, the cost estimate for the Great Belt link turned out to be optimistic and was closer to an unlikely best possible outcome than to a most likely one. With a cost overrun of 110 per cent for the Great Belt rail tunnel, risks caught up with the rail link, which proved non-viable and was rescued only by cross-subsidisation from the road link (see Chapter 4). Total construction costs were 54 per cent higher than forecast costs. But, unlike the Chunnel, it is taxpayers’ money that has been placed at risk at Great Belt and not investors’ money.
For the Øresund link, partial risk analyses were carried out. In one such analysis, a group of government ofﬁcials assessed that given histor- ical experience a 50 per cent cost overrun ‘cannot be seen as an unreal- istic maximum estimate’ for the link.10 In addition, the group found the project ﬁnancially non-viable even without this overrun, as did three other appraisals carried out by ofﬁcials in the months before the project was presented to the Danish Parliament. Yet, when the Minister of Transport presented the project for ratiﬁcation in 1991, none of this information was mentioned. Neither the proposed law nor the accompanying comments contained any information about risks of non-viability.
When, more than two years later, it became publicly known that such information existed and had been withheld from Parliament, the result was a sharp critique of the Minister of Transport by the Auditor- General of Denmark.11 The Auditor-General found that the assumptions on which the estimates of project viability were based ‘represent the economic assumptions that have to be made for the project to be self- ﬁnancing’, that is the assumptions that would make the project seem viable on paper, and not the assumptions that would have reﬂected the most likely development.12 For later budgets, worked out by the state- owned enterprises responsible for construction of the Danish access links and the ﬁxed link proper, the Auditor-General similarly found that it was ‘less evident’ that the budgets were within claimed levels of uncer- tainty.13 Moreover, the Auditor-General found that the risks regarding the future development of interest rates and ﬁnancing costs had been ignored and that it ‘would have been natural’ to mention this in the project budgets.14
On this basis, the Auditor-General resolved to monitor the Øresund project and to carry out audits in order to establish whether the assumed basis for the project would actually be realised, including whether the project would, in fact, be self-ﬁnancing as required by the Danish political agreement behind the project and by the agreement between Denmark and Sweden to build the link.15
The treatment of risk in the Channel tunnel, Great Belt and Øresund projects has clearly been inadequate. Even so, the organisational and institutional set-up of the Great Belt and Øresund links, as loan- and user-ﬁnanced state-owned enterprises, may be improvements over earlier transport infrastructure projects in Denmark in terms of risk identiﬁca- tion and allocation, because costs, revenues and viability have become more visible than in earlier projects, as have environmental impacts. Still, there is substantial scope for improvement for future projects as shown above and as will become clear from what follows. For other major projects – transport and non-transport alike – the conclusions are similar: the risks involved are high and are typically treated in a deﬁcient, sometimes even deceptive, manner in feasibility studies and project appraisals, if treated at all. In a World Bank study of ninety-two projects, only a handful was found to contain ‘thoughtful’ risk analy- sis showing ‘good practice’.16 Appraisals of World Bank projects are typi- cally more complete and more rigorous than appraisals of other projects. Nevertheless, it is important to note that there are good examples as well. We have already mentioned in Chapters 2–4 the construction of the Paris South-East and Paris Atlantique high-speed rail lines.17 Also, the technologically high-risk Apollo aerospace programme is considered a classic success story of megaproject planning and implementation. The cost overrun on this us$21 billion project was only 5 per cent. Few know, however, that the original budget estimate included us$8 billion of con- tingencies.18 By allowing for risk with foresight, the programme avoided ending up in the type of large cost overrun that destabilises many major projects during implementation. The Apollo approach, with its realistic view of risks, costs and contingencies, should be adopted in more major
A typology of risks
The main sources of ﬁnancial risk in major transport infrastructure projects are:
- construction cost overruns induced by, for instance, government, client, management, contractor or accident;
- increased ﬁnancing costs, caused by changes in interest and exchange rates and by delays; and
- lower than expected revenues, produced by changes in trafﬁc volumes and in payments per unit of trafﬁc.
Although less signiﬁcant, ﬁnancial risks are also related to costs of oper- ations, maintenance and management. From an economic point of view the main risks are cost overruns, delays and lower realised demand than that assumed during appraisal.
From an analytical point of view, it is expedient to identify the fol- lowing types of risk of relevance to both a ﬁnancial and an economic perspective:19
. Project-speciﬁc risks;
. Market risks;
. Sector-policy (including force majeure) risks;
. Capital-market risks.
The two ﬁrst types of risk include those that are conventionally associated with a project, and that have been in focus so far. As for project-speciﬁc risks, the conventional assumption is that their effect can be eliminated – at least in part – by risk pooling or risk spreading; see further below. This does not normally apply to market risks that are explained by more funda- mental events that affect economic activities in a similar way, for example the overall economic development in a country. Sector-policy risks arise from the fact that the outcome of a project is dependent on speciﬁc sector policies; for transport projects, for instance, complementary investments in access links, taxation of transport or other regulation of road transport or of the environment. Some of these risks can be identiﬁed and can also be eliminated by providing a stable regulatory environment and by proper contracting. These types of risk are not necessarily eliminated by having projects operated by state-owned companies. Ultimately, such companies serve the general public, and if changes in regulatory policies imply that a project cannot be used as originally envisaged, private parties, that is taxpayers, will be affected in a negative way, as would be the case if the project were to be operated by a private entity.
Capital-market risks are created through borrowing, in particular in the international market, in order to ﬁnance projects. Such risks mainly con- sist of two elements, interest rate risks and currency risks. The capital market is today able to provide ﬁnancing on conditions that allow borrow- ers to protect themselves against currency and interest rate adjustments in the short and medium term. But, of course, such insurance comes at a cost.
The cost of risk
The condition that a project is associated with risks gives rise to an economic cost. People are normally risk-adverse and are prepared to pay something – an insurance – in order to reduce or totally elimi- nate risks. The cost of risk is an economic concept and reﬂects the
maximum amount that an individual is willing to pay to eliminate a particular type of risk, so that the future of a particular type of event would become risk free for that individual.20 In practice, different lev- els of riskiness associated with different types of investment are reﬂected in different minimum rates of returns, which are required in order to persuade individuals to commit their money.21 The lowest return is nor- mally required on government bonds; here the return is about 3–4 per cent in real terms, as these bonds are considered to be virtually risk free. Private-sector debentures are normally associated with a somewhat higher return, say 5–6 per cent in real terms. The return on equity, namely ownership capital with risk, starts at about this level, but may be- come much higher depending on the perceived riskiness of the enterprise concerned.22
The risk costs associated with investments in infrastructure can be expected to be high. There are two basic reasons for this. The ﬁrst is the fact that an investment in a major infrastructure project is basically a sunk cost, that is it cannot be retrieved. Once, for example, a bridge has been built it cannot really be used for anything else, so if the decision to build the bridge turns out later to have been a poor one, it cannot be repaired. The second reason is that the beneﬁts of investments are often highly correlated with economic growth. If economic growth is high, then the project will fare well; conversely if growth is poor, the project will perform poorly. As noted, overall economic performance affects the market risk of a project, particularly when seen in an economic perspective. From an economic viewpoint it is difﬁcult to offer insurance against market risks within a country, as it affects everybody, that is the risks cannot be spread and pooled when seen in a national perspective.
That infrastructure investments are viewed as risky is also brought out by the experience of private-sector involvement in infrastructure projects through a concession. The available data suggest that ﬁnancing for con- cession companies will only be forthcoming provided: (i) the equity makes up about 20 per cent to 30 per cent of the total ﬁnancing requirements; and (ii) that those who invest in equity can be expected to receive a return of between 15 and 25 per cent in real terms and after tax.23 Assuming that the remainder of the capital, the long-term debt, can be mobilised at a real cost of 6 per cent, the implication is that the project will have to achieve a ﬁnancial rate of return of about 7.5 per cent to 12 per cent in real terms (disregarding taxation effects).
The difference between the return required for this type of investment, say 9 per cent, and the return on a virtually risk-free investment, say government bonds at 4–5 per cent, can be viewed as a measure of the cost of risk associated with the project. It should be emphasised that this
type of cost is incurred not only when an infrastructure project is de- veloped through a private concession company. It is also present when the project is developed by a state-owned enterprise and the ﬁnancing is secured through a sovereign guarantee. In the world of risk there is no such thing as a free lunch. The risk and its costs exist under any cir- cumstances, even if promoters of projects backed by sovereign guarantees often impart the impression to politicians and the general public that this is not the case, as happened for the Øresund and Great Belt links. Here the promoters seemed to hold this belief themselves. One high-ranking Danish Ministry of Transport ofﬁcial expressed it to us in the follow- ing way in his comments on an earlier draft of the manuscript for this book:
In my view you should mention [in the book] that the costs the private sector would demand to have covered for taking on the full risk [of projects] would make the projects more costly. Thus the advantage of ﬁnancing the Great Belt and Øresund projects by a sovereign guarantee is that this lowered ﬁnancing costs.24
Not only do we hold that sovereign guarantees hide ﬁnancing costs instead of lowering them. We also hold that there are reasons to believe that the risk costs associated with ﬁnancing could increase if the project is underwritten by a sovereign guarantee. The reason is that the guarantee will transfer most of the risks to taxpayers, that is risk bearers who can be assumed to be less able, on average, to protect themselves against risks than those persons who act in the capital market, hence increasing overall risk costs.
In addition, there is good reason to mention a point made by the World Bank that the money saved by lower interest rates on loans backed by sovereign guarantees may be offset by inefﬁciencies arising from relaxed discipline as a result of such guarantees. Lenders backed by sovereign guarantees have no, or much less, incentive for supervising projects than do commercial banks without such guarantees, resulting in relaxed pressure on project performance. According to the Bank it may take several percentage points of interest advantage to offset even moderate inefﬁciencies in terms of cost overruns and delays stemming from the inefﬁcient supervision of projects.25 For the Great Belt link the direct interest advantage of a sovereign guarantee on loans is approximately 1 to 2 percentage points, depending on the size and riskiness of the non- guaranteed projects that one compares it with.26 For the Øresund link the interest advantage is of a similar size. If one were to compare interest on guaranteed loans against return on equity, the advantage would be higher.
The total risk costs associated with infrastructure projects can normally be expected to be substantial. Take, for instance, the Øresund link, which cost some d 17 billion (excluding access routes), and is expected to yield a ﬁnancial return of about 5 per cent. Assume, for example, that the actual requirement is an additional 5 per cent, that is in total 10 per cent in order to ensure that the costs of risk also will be compensated for. Then the cost of risk is equal to about one third of the total investment, that is about d 6 billion. Whilst this example is not an exact calculation of the cost of risk, it illustrates how important it is to take risk into account in project planning and appraisal.
Strategies of risk assessment
The most consequential problem regarding risk analysis in megaproject feasibility studies and decision making is not the absence or inadequacy of risk analysis in itself, but the neglect of relevant downside probabilities in the calculation of project viability.27 To a wide extent, risk is simply dis- regarded in feasibility studies and appraisals by assuming what the World Bank calls the EGAP-principle, ‘Everything Goes According to Plan’.28 We explained above how the unrealistic EGAP-principle was used for the Channel, Great Belt and Øresund links and what the problems involved were.
A cure recommended here, when undertaking a feasibility study, is to substitute what we call the ‘MLD-principle’ for the EGAP-principle, MLD standing for ‘Most Likely Development’. The cure should be seen as part of a wider strategy for public-sector involvement in developing megaprojects, which is taken further in Chapters 10 and 11. By follow- ing the MLD-principle, the roles of feasibility study and appraisal are redeﬁned from the optimistic and unrealistic everything-goes-according- to-plan estimation of project viability to the realistic and experience-based assessment of the most likely development of projects. Carrying out MLD appraisals, the focus is on identifying the most likely risks and the most risky parts in a given project in order to reduce these risks and, if possible, drop those parts. What the World Bank calls ‘switching values’ would be calculated for key variables, including environmental variables, under- stood as the level of the variable at which the project turns from viable to non-viable. The likelihood of switching values actually occurring would be estimated. In addition to this, what the Auditor-General of Sweden calls ‘threshold levels’ would be established for costs, revenues, environ- mental impacts and viability, namely levels that, if crossed, redeﬁne the project as a new project that must be appraised and approved anew. The result of such measures would be more robust projects.
A further technique that should be made use of in feasibility studies sponsored by public-sector organisations is the analysis of worst-case sce- narios, a method frequently employed in the private sector. The basic idea is simple: identify negative conditions from the point of view of the project and analyse the implications for the project’s viability and ﬁnancing. This approach is helpful for determining the robustness of the project, but also for identifying supplementary actions required in order to mitigate risks and ensure success. Worst-case scenarios may also be useful for identi- fying projects that should be dropped altogether since the risks and their implications appear to be all too signiﬁcant.
In addition to identifying risks carefully and making them visible, a main instrument for reducing the costs of risk is to prepare a risk management plan as part of a feasibility study. The purpose of such a plan is to identify how various risks are to be managed and by whom. In the public sector, the establishment of a credible risk management plan should be a part of the documentation required before any decision is taken on whether to go ahead with a project or not (see Chapter 11).29 Figure 7.1 shows the main elements involved in risk management.
The main challenge to the preparation of a risk management plan is to actually fully identify the scope for risk management, and to communicate that it is much wider than what is normally appreciated. To a large extent this lack of appreciation of the role and scope of proper risk analysis and management is due to the history of contracting for infrastructure facili- ties. Because contracting in the infrastructure sector has to a large extent been on behalf of the public sector, the contracting format has become dominated by public-sector thinking. A key aspect of public-sector be- haviour is thus that it is typically rule-based and not performance-based. In contracting, this is reﬂected in the fact that contracts typically are based on technical speciﬁcations being given to the contractor; the contractor’s task is to build according to these speciﬁcations and not necessarily to achieve a certain level of performance (for so-called build contracts, see Chapter 9). However, this also means that the incentive and scope for developing new techniques in order to reduce costs, to reduce certain types of risk, and so on are limited. Consequently, present contracting techniques to a considerable extent eliminate the possibility of managing risks. A main challenge to risk management will therefore be to change the present contracting format, as is further discussed in Chapter 10.
There are several basic approaches to be considered as part of a risk management plan, some of which are partially overlapping. One approach
Figure 7.1 The risk management process
Source: Council of Standards Australia and Council of Standards New Zealand, Risk Management, AS/NZS 4360:1995 (Homebush: Standards Australia and Wellington: Standards New Zealand, 1995), p. 11.
involves eliminating risk altogether. This applies, for example, to certain sector-policy risks, for which it, given the circumstances, may be possible for the entity responsible for undertaking a speciﬁc type of infrastructure project to enter into an agreement with the central or local government to
ensure that certain policy actions are not taken, or at least not taken with- out compensation. A second approach involves buying risk management services. This may be the approach used in order to deal with capital- market risks and it might also be available for what some consider as force-majeure risks.
A third approach involves allocating risks to parties who have an in- centive to reduce the negative impacts of risk, either by reducing the likelihood of the event or to reduce the negative impact itself, if the event were to occur. Again this applies to such events that are normally consid- ered to be of a force-majeure nature, although they may not necessarily be a major event. An example may be the occurrence of unexpected geologi- cal phenomena creating delays in the construction of a bridge or a tunnel. By allocating geological risks to contractors from the outset the result is likely to be a more thorough analysis of such risks before contracts are ﬁnalised and a faster and more effective containment of negative impacts should unexpected geological phenomena occur during construction. If, conversely, the owner accepts geological risks, or if the placement of such risks is unclear, experience shows this to be a sure road to delays and cost overrun, especially for projects with substantial underground work. Another example may be the occurrence of protest actions during the ini- tial construction phase of an infrastructure facility, where the likelihood of the actions actually taking place can be signiﬁcantly reduced through speciﬁc measures, for instance by replacing the conventional, closed for- mat for megaproject development with a more open and transparent one and by paying adequate compensation to parties negatively affected by projects, as done in Boston’s Big Dig project.
The nature of project-speciﬁc risks is such that their costs can be elim- inated by appropriate pooling or risk spreading. There are several institu- tional arrangements to handle this. As has been demonstrated by Arrow and Lind, one approach to achieving effective elimination through risk spreading is to allocate project-speciﬁc risks to the public in general.30 A way to achieve this would be to operate the project as part of the public sector, or by securing the ﬁnancing for the project by way of govern- ment guarantees. But also the private sector has instruments available to achieve risk spreading. One of the reasons for establishing speciﬁc project companies for undertaking large infrastructure projects is thus to enable widespread participation by the capital market in the project, thereby allowing individual investors to pool their investments, and al- lowing the speciﬁc risks of the project considered to be spread between many investors, and thereby permitting the aggregate cost of this risk to be reduced.
The most difﬁcult risks to manage are, as mentioned, market risks. One reason is that such risks are quite different when seen in an economic perspective from when seen from a ﬁnancial point of view. As a rule, the economic cost of a market risk cannot be managed; the main issue to be considered is who should bear the cost of this risk, which is an important income distributional question and may also have institutional implications (see further Chapter 9).
Lessons regarding risk
As documented in this and the previous chapters, the risks associated with major infrastructure projects are substantial. Key factors contributing to risk are the facts that the investment will be irreversible and the viability highly dependent on general economic development. Given the magni- tude of the uncertainties involved, feasibility studies of major projects without risk analysis are less than useful since such studies will often de- ceive decision makers and the general public regarding the outcomes of projects. Risks cannot be eliminated from major projects, but they can be acknowledged and their impacts reduced through careful identiﬁcation and by allocation of risks to those best suited to manage them.31
In most democracies the civil service has an obligation, deﬁned by law, to provide the Cabinet and Parliament with ‘all relevant information’ pertaining to their decision taking and law making. Clearly, risks of cost overruns of 50–100 per cent on multibillion-dollar projects, together with large uncertainties regarding revenues and environmental impacts, are ‘relevant information’. Thus such information must be brought to the attention of politicians and the general public.
The following conclusions can be drawn regarding risk:
- Public and private investors, parliaments, media and the general pub- lic are routinely inadequately informed and misled regarding the risks involved in megaprojects, cases in point being the Channel tunnel, Great Belt and Øresund projects;
- A full risk analysis based on the MLD-principle (Most Likely Devel- opment) should be carried out as part of feasibility study and apprai- sal – undertaken by public-sector organisations – for any megaproject. In addition, a risk management plan should be prepared. Such risk
analysis and management would identify the most risky parts of a project. The objective is to reduce risks and to change or drop the most risky parts of the project. Finally, the aim is also to allocate risks appropriately to the involved parties;
- Risk analysis should also comprise worst-case scenarios, in order to illustrate what happens if worst comes to worst. The experience with ﬂooding and ﬁre in the Great Belt rail tunnel illustrates the pertinence of this point as do the cost overruns and ﬁre in the case of the Channel tunnel;
- Feasibility studies and risk analyses for future projects should be carried out together with considerations regarding the possible in- stitutional, organisational and ﬁnancial set-ups for the project. The set-ups will substantially inﬂuence risks and costs, just as risks and costs may inﬂuence the set-ups. Institutional change may be a pre- requisite for risk reduction, as discussed further in Chapters 10 and 11;
- Public ﬁnancing or ﬁnancing with a sovereign guarantee and no risk capital, as known from Great Belt and Øresund, does not reduce risk or risk costs. It only transfers risk from lenders to taxpayers, and so is likely to increase the total risks and costs of a project.
Get expert help for Dealing with risk Assignment and many more. 100% safe, Plag free, 24X7 support, Order Online Now!No Fields Found.