One of the studies referencing the Danish Aalborg University study “Inaccuracy in Traffic Forecasts” is the 2009 GAO report “High Speed Rail: Future development will depend on addressing financial and other challeges…“
I think it is fair to say that the GAO supports my conclusion that the cost/benefit of high-speed rail is entirely dependent on the particulars of the city-pair under discussion. In turn, the economics of that link depend not just on population densities, distances, intermediate stops, but upon how the link fits into the regional transit web. It is clear (to me) that there are very few US links that would qualify under any reasonable economic benchmarks. There are certainly no Australian links that make sense. The often-cited Shinkasen Tokyo – Osaka link has about 63 million people along its route. I have used that link — it is OK, but one reason it works is trains every 3-5 minutes work days.
The GAO does not consider at all the risk of obsolescence — which I personally think is very real. A huge capital investment becomes a huge white elephant.
The GAO report section on uncertainty touches on incentive design – the first hint I have seen towards leveraging market incentives to avoid bad projects and to optimize results from viable projects.
Uncertainty and Inaccuracy in Forecasts of Riders and Costs
Forecasts of riders and costs are two key components of evaluating the economic viability of high speed rail projects, and rider forecasts are the anchor for the array of public benefits that a new line might bring. However, as we have discussed, these forecasts are often optimistic, calling into possible question the credibility of information being used by decision makers to pursue high speed rail. Development of stronger policies, procedures, and tools could enhance the accuracy and credibility of the forecasts and contribute to better decision making. There are a variety of means that have been discussed in the transportation literature and could potentially be employed to strengthen the accuracy of forecasting.75 These means include the following:
• obligating state and local governments to share some of the risks of underestimated costs for those projects seeking federal financial support;
• obtaining forecasts and estimates from independent sources, such as a state auditor or a federal agency, rather than sources contracted to construct projects for a high speed rail project sponsor;
• subjecting forecasts to peer review with possible public disclosure of all relevant data and public hearings; and
• conducting horizontal comparisons of projects—that is, using data from different projects reported using a standardized accounting system to prepare probability distributions of the accuracy of project estimates of cost and demand—to evaluate new high speed rail projects.
Another potential means to improving the accuracy of these estimates is to align the incentives of public and private interests. For example, in Japan, for a new line to be built, the private operator must be able to make a reasonable profit over and above operating costs, maintenance costs, and lease payments made to the government for use of the track. The private operator then has an incentive to maximize riders, but also to minimize the lease payments, to increase its profit potential. Therefore, the private operator wants to be conservative regarding rider forecasting and wants the government to build the infrastructure in order to allow for the lowest cost operation and maintenance. The central government has an incentive to keep costs low in constructing the line and to extract the highest lease payment it can negotiate from the private operator. The private rail operator and the central government negotiate and agree upon a lease payment, which remains set over a 30-year period. These negotiations are based on forecasts of riders over the ensuing 30 years and the existing cost estimates. According to officials and academics in Japan, this structure has resulted in a discipline that has vastly improved the accuracy of rider forecasting and cost estimation. For one newly constructed line, actual riders were within 90 percent of forecasted riders, and the construction of the line was within budget and ontime.