Toll Road Forecasts Come Under Scrutiny in Virginia

An important discussion is taking place in Northern Viriginia over the veracity of traffic and revenue forecasts for the Dulles Tollroad. The forecasts are critical for this corridor because higher toll rates have been justified based on their ability to finance a the extension of the Washington, DC Metro to the Dulles Airport (and beyond).

According to (Feb 24, 2012), the forecasting company, CDM Smith (formerly Wilbur Smith Associates), has made over 200 forecasts for 100 projects. In a response to criticism that the CDM Smith forecasts suffered from “optimism bias,” the company’s CEO writes (in a letter available at

“The previous traffic and revenue studies undertaken for the DTR [Dulles Toll Road] highlight our continued success and reliability in effectively forecasting its traffic and revenue potential. Comparison of CDM Smith’s 15yr forecast for the DTR in 1989 with the actual performance indicates that actual 2003 transactions were 98.2% of forecast (the 2004 toll increase was not assumed in 1989).

“The 2009 study forecasts also compare favorably to the actual revenues following the 2010 and
2011 toll adjustments:

– 2009 Study 2 yr forecast: Actual 2010 revenue was 100.7% of forecast

– 2009 Study 3 yr forecast: Actual 2011 revenue is expected to be 97.5% of forecast

“The 2005 DTR Study cannot be tested against actual performance as none of the toll scenarios in that study were implemented.”

This is very important issue, and we’ve written about this elsewhere (see here). Given a few high-profile bankruptcies (e.g., the South Bay Expressway in San Diego, Southern Connector in South Carolina), and the rising important of public-private partnerships in financing these projects, the accuracy and reliability of these forecasts are critical, and I personally have worried that the forecasting track record hasn’t received enough public scrutiny.

The article at TollroadsNews does a good job of exploring the complexity of these forecasts. In many ways, it’s more art than science. Nevertheless, decisions about long-term infrastructure need to be bounded, and forecasts are an essential part of evaluating risk and uncertainty. But they also have to be subjected to independent scrutiny and the full range of uncertainties in these forecasts need to be part of the public decisionmaking process.

Thanks to TallroadsNews for helping to broker this public discussion.