Prices for new infrastructure assets steadily increased over the past 10 years, on the back of intense competition for scarce new assets, fuelled by the strong debt and equity markets. In the competition for new toll road assets, bidding consortiums used higher gearing and aggressive traffic forecasts as value-levers to satisfy investor hurdle rates for this expensive asset class. Debt markets began to move closer to these aggressive assumptions, and as such have been exposed to much more of the potential downside of traffic risk than they have traditionally been comfortable with.

This article discusses the results of this approach, and outlines the options available for future greenfield projects in the sector. It also suggests that for toll roads, equity should perhaps shoulder the risk of initial patronage, since projects are at the mercy of the economic and political fluctuations that are possible during the lengthy period between bid and commercial operation and errors on underlying patronage forecasts can be as much as 50% either side of target.

The casualties

The traditional toll road operator's business model is now in question, in light of the difficulties experienced by the last four Australian greenfield toll road projects; including Cross City Tunnel (CCT), Lane Cove Tunnel (LCT), WestLink (M7) and now ConnectEast (CEU) have all missed initial traffic projections. The combination of overly optimistic traffic projections and aggressive gearing has been the key problem with these projects and, embarrassingly enough, left both the CCT and LCT in receivership within just 18 months of opening.

The failure of these projects has been very public and has created significant political issues for State governments, which want to undertake an ambitious pipeline of PPP projects across a range of social assets. The situation has also left many investors, understandably enough, looking for better risk-reward opportunities elsewhere. This attitude is borne out by the poor performance of two recently-listed greenfield toll road projects RiverCity Motorway (RCY) and BrisConnections. Both companies are now trading at greater than 50% discounts to their IPO share prices.

The gearing of these projects has typically been justified using project finance debt sizing methodologies. Project financing relies upon ratios including interest coverage, debt service coverage, loan and project life coverage ratios. These ratios are effective for the regearing of stabilised brownfield toll road assets where the debt metrics can be produced with reference to historical performance and are expected to improve over time. Additional debt capital can be drawn, tightening debt metrics against defensive cash flows.

However this approach seems ineffective when considering the tenor, magnitude and growth issues that constitute opening traffic risk. Project financing metrics are used in bid models in order to maximise the capital structure and in some instances the upfront payment to government. Bidding consortiums attempt to size appropriate debt levels based on minimum debt ratios that rely upon the opening traffic expectations.

But these opening traffic expectations are reliant on estimates of (i) regional GDP growth (ii) fuel prices and (iii) commuter behaviour over the period between the bid date and commencement of operations. On a greenfield project this can be a forecast as much as 5 years into the future. Accordingly it is not surprising that on a competitive bid these traffic forecasts are prone to error. The difference between a highly and moderately geared project could mean the difference between a project surviving its expected ramp-up period and a project being declared insolvent.

Changes to gearing are necessary

Using the process outlined above greenfield toll road projects have achieved gearing as high as 75%. These high debt levels were acceptable to equity investors as the additional debt was accretive to their internal rates of return (IRRs). However, in our view, the debt levels over the past decade became excessive. With investors rediscovering concerns around highly leveraged vehicles, combined with the recent failures of greenfield toll road projects, we expect weighted average cost of capital (WACC) calculations to be adjusted to force lower gearing levels on projects in the future.

Corporate finance theory teaches us that an increased proportion of debt lowers the WACC for a project. Increased gearing levels will approach a minima or, tipping point after which WACC will increase rapidly with increased debt levels. Since recent toll road projects have been geared to a range of between 65-75% we would propose the market has identified this range as the optimal gearing level for these projects. With recent failures this range is set to fall to a more sensible range of 50-60% on greenfield project, which would allow equity to accept all opening traffic risk. As an asset matures and traffic ramps up, project financing techniques may then be employed to increase these gearing levels without changing the perceived risk within the business. However with the failure of greenfield toll road projects, combined with higher costs of equity, as well as an aversion to gearing we expect this range to fall significantly.

Financial returns from toll road projects are principally driven by the underlying traffic growth profile of the asset. Aggressive investment banks have used debt structures to manufacture early artificial returns from immature assets and to boost equity IRRs. Macquarie Bank and Babcock & Brown, for instance, have proven experts at projecting spectacular returns from this kind of deal structuring.

The lessons from the fallen

The Cross City Tunnel (CCT) and Lane Cove Tunnel (LCT) projects are instructive when considering the issues undermining the existing projects as well as what needs to be considered in new projects in order to avoid a repeat of these scenarios.

The Cross City Tunnel cost around A$1.02 billion ($900 million to build and finance and with featured senior debt of around A$680 million, meaning the project was geared at around 58%-60%. The Lane Cove Tunnel had a purchase price of A$1.685 billion and was geared at 68% (with roughly A$1.15 billion of debt). LCT's owner and operator is Connector Motorways. Opened in March 2007, the tunnel and gateway are operated by Transfield Services.

Problem # 1: Patronage Estimation

There was a fundamental miscalculation in the daily patronage experienced of the CCT. The winning bid was based on around 90,000 vehicles per day; the experienced reality is around a third of this figure. This significant difference resulted in the CCT going into receivership and being subsequently sold for around A$700 million. This situation can be partly attributed to the aggressive sealed, highest-bidder-wins, auction process run by the New South Wales State government.

Successful patronage projects depend on forecasting a dimensionally complex probability of the movement of varying demographic, geographic and socioeconomic clusters. Layered on top of this we burden the system with an economic cost (the toll) and a benefit (the time saving). This framework then needs to be compared to a new route instead of a cost-free and familiar alternative. The complexity can be said to captured using sophisticated traffic network models. These models can even predict retrospectively historic traffic movements and hence argued to be calibrated.

A project finance banker can interrogate patronage modelling in the light of economic scenarios such as increased or decreased inflation, inaccurate time-value estimates and more tangible ones such as junctions working less efficiently than planned or alternative routes not being closed as promised. These scenarios can interact with the financial model for debt structuring, sizing and credit analysis. What is very difficult to capture, no matter what market research study has been performed, is the behaviour of a free-thinking, independent, and in this case strong-willed Australian population.

The significant increase in fuel prices, increasing CPI indices, and borrowing rates that are already straining the wallet of the average road user, could be blamed for low patronage of a toll road, but even after the significant passage of time, although acceptance is increasing the tunnel is barely at levels expected in bank downside cases.

Problem # 2: Capital Structure

With the ability to structure toll road vehicles as triple-stapled unit trusts, investors are able to realise returns without the usual requirement of available post tax profits. This in itself drives higher returns than those available in other asset classes.

BrisConnections offers a telling example of the dangers of marrying high levels of leverage to the fickleness of the retail infrastructure equity market. BrisConnections, in Brisbane, links the city's airport to its central business district, and is Australia's largest toll road project to date. It comprises the construction of two road tunnels and a new airport connection – the 6.5km Airport Link underground toll road, the 3km Windsor-to-Kedron section of the Northern Bus way, and a 750m fly-over above the airport roundabout.

The A$3.15 billion 10-year term syndicated term loan supporting BrisConnections was reported as being difficult to syndicate. Total project cost is A$4.8 billion (including land costs) with the state of Queensland putting up A$1.5 billion. The combined projects will cost over A$4 billion to design and build, A$3.4 billion for the Airport link, A$444 million for the Bus way, and A$272 million for the fly-over. The enterprise value, once completed, is expected to be as high as A$5.5 billion. At the end of June BrisConnections launched a $1.2 billion initial public offering, which was underwritten by shareholder and financial adviser Macquarie Bank. Investors that participated on in the first instalment of $1 for stapled units dumped them on 31st July. The sales represented 11% of those issued.

Summary

The problems facing the sector today are largely a product of the financial environment in which the projects were initiated, often with leadership from investment banks, which dominated consortiums bidding on projects in Australia. Over the past decade these banks recognised the abundance of new capital for projects on the back of the credit bubble. In this environment of strong debt and equity markets and intense competition for new assets bidding consortiums pushed bid assumptions too far. WestLink widely considered a success has survived due to stronger than expected average trip distances, a lucky escape that delivered the additional revenues required on the back of lower than expected forecast trip numbers.

There are many factors that put strain on the success of a toll road project. Overly aggressive patronage forecasts have been proved a weak link time and time again. Aggressive financial structuring can put undue stress on the operational cashflows especially during the often uncertain ramp-up phase. The only viable solution is to raise equity in order to repay the upcoming debt obligations. This option is being now being pursued by companies not only in the toll road sector but also by companies involved in the wider infrastructure sector.

The major toll road player Transurban has recently adopted this approach, and Macquarie Infrastructure Group, which in the last two years has earned around A$2.4 billion in profit from its whole portfolio, has proposed the sale of a 50% stake in Sydney's WestLink M7 and a A$600 million share buyback that it hopes will restore investor confidence in its business model and in doing so inject life back into its depressed share price. Transurban is widely considered as the major contender to buy half of the orbital road because of its existing pre-emptive rights.

Nick Crawley is managing director of Navigator Project Finance, a specialist project finance modelling consultancy. www.navigatorPF.com.