Many businesses struggle to forecast revenue and costs over the short term, yet, for many infrastructure transactions, there is a need to forecast these over the long term, sometimes for more than 50 years. How is this possible and what are the consequences of getting it wrong?
A key to the reliability of long-term forecasting is the extent to which there is an ability to fix both revenue and costs, including finance costs, over the long term. For example, many concession-type contracts will fix the payments to the concessionaires if they achieve the required functionality and/or operational performance, or if revenue from some economic infrastructure can be fixed over the long-term. The concessionaire may be able to negotiate long-term subcon-tracts-for example, for asset operation and maintenance over the concession period.
Basing contracts on fixed costs (especially operational costs with a high fixed element) can attract a premium because the operator is being asked to forecast his performance and costs over the long term but has no one to pass those costs on to in the event those costs differ from the original forecast. There are ways around this, however; the most common is to build in periodic reviews of costs and adjust the revenue to reflect any changes revealed in these reviews. The real risk with this approach, and one that is often overlooked, is that the long-term counterparty will survive the test of time. This should be a concern whether it is the public sector, subcontractors, or even the financial hedge provider making the payments.
In 2008 significant parts of the world's financial system come close to collapse, and it is not unheard of for public-sector parties to default on payments. For example, in 2001 the government of the State of Maharashtra, India, had to bail out its subsidiary, the Maharashtra State Electricity Board (MSEB), when MSEB failed to make payments to the Dabhol Power Company under a power purchase agreement. It is worth noting, however, that this non-payment was only part of a complex set of issues with the Dabhol power project.1
The difficulties that often attract the most controversy are transactions with demand risk; one example is toll roads that rely on long-term forecasting of both traffic and toll levels. The only thing that is certain is that these forecasts will be wrong. In bull markets, there is a danger that bidding can be biased by optimism, typically by overestimating traffic forecasts and the way those will grow over the concession period. It is common for traffic growth to be linked to GDP growth and for forecasts to assume year-on-year constant growth, but there is growing evidence, particularly in developed economies, that the linkage between traffic growth and GDP is being lost as people change their travel habits. It is necessary to be realistic about the level of tolls that users will be willing to pay and the alternative routes they may have. For example, when the Cross City Tunnel project in Sydney, Australia, was being bid, the capital costs increased significantly. Consequently, the forecast toll level was increased to reflect this. Very soon after the contract was awarded, it became clear that the tolls were too high and that, as a result, drivers were not using the tunnel. As a result, within months of opening, the project company was insolvent (see Case Study 2: The Cross City Tunnel).
The manner in which the bidding process is structured and the criteria by which the contract is awarded can also encourage over-optimism; for example, if the decision to award the contract rests on the size of the upfront payment to the public authority, unrealistic assumptions about the size of that upfront payment can result. The flip side of over-optimism is overly conservative forecasts that are exceeded significantly by the operator, resulting in materially greater returns than expected. In some circumstances, such overshooting the expected goal can be an issue for the public sector (see also the discussion in Chapter 1.5 about public perception).