Cost overruns in large-scale transport infrastructure projects: A theoretical and empirical exploration for the Netherlands and worldwide

In the Netherlands, the costs of railroad infrastructure construction work out 11 per cent higher than forecast; the percentage for roads is 19 and 22 for bridges and tunnels. While this puts the Netherlands in a favourable light compared to other countries, the actual cost overruns are probably far higher, believes Chantal Cantarelli, who obtained her doctorate on this subject from TU Delft on Monday, 28 November. International research into cost overruns among 258 large-scale transport infrastructure projects has shown that roads generally turn out to be 20 per cent more expensive, bridges and tunnels 34 per cent and railroads 45 per cent. 'But what about the Netherlands?', TU Delft PhD student Chantal Cantarelli asked herself. From her study of the situation in the Netherlands, it appeared that the costs were higher than predicted in more than half the projects involving roads, bridges, tunnels and railroads. Cost overruns occur less frequently than abroad and they are also smaller in extent. Railroads usually turn out 11 per cent more expensive, roads 19 per cent and bridges and tunnels 22 per cent. Worldwide, however, the cost overruns involved in constructing railroads are higher than in the case of roads. Another notable fact is that the greatest cost increases take place in the phase prior to construction and not during construction itself. However, it would be a mistake to conclude that cost overruns aren't that much of a problem in the Netherlands, warns Cantarelli. Overruns are, she points out, often based on a comparison between the costs upon final delivery and the costs estimated at the time the formal decision was taken. But politicians have often made an informal decision at a much earlier stage. This phenomenon is known as lock-in. The estimates at the time of the unofficial decision are often much lower. Cantarelli examined this effect in relation to two Dutch projects, the Betuweroute (a 160 km long double-track railway from the port of Rotterdam to the German border) and the HSL-Zuid (a 125 km long high speed rail line between the Netherlands and the Belgian border). It transpired that politicians had in fact committed themselves to the project four years prior to the moment the formal decision was taken. If one calculates the cost overruns on the basis of the underestimated costs at the time of the informal decision, the costs for the Betuweroute and HSL-Zuid work out as three to four times higher. Cost overruns are therefore presumably far higher than is actually known. According to Cantarelli, it is a matter for concern that cost estimates have not improved in the last twenty years. As estimation methods have certainly improved, there have to be other reasons for systematic overruns. Cantarelli concludes that to improve cost forecasts, another cost methodology should be applied in addition. The method Reference Class Forecasting is particularly suitable for this purpose. On the basis of this methodology, estimates are calculated by comparing the actual costs and cost overruns of similar realised projects. This helps to avoid overly optimistic estimates. But strategic behaviour on the part of parties, in the form of deliberately estimating lower costs, may also play a role. The rules of play should change in order to make it harder for interested parties to exhibit such behaviour. In this regard, it is also important that responsibility for future cost overruns should be borne by those drawing up the estimates.