A study of the dynamically priced High Occupancy Toll (HOT) lanes on Minnesota’s Interstates 394 and 35W show that when toll-lane prices go up, more drivers enter, contrary to what one would expect.

The MnPASS High Occupancy Toll (HOT) lanes installed on Minnesota’s I-394 and I-35W (in the Minneapolis/St. Paul area) have been of interest to road-pricing researchers. MnDOT converted the I-394 lanes from high-occupancy vehicle (HOV) to HOT in 2005; MnDOT added the I-35W lanes in 2009. They are the first form of road pricing in Minnesota. Based on observed traffic volumes, the MnPASS system changes toll prices every three minutes based on the assumption that increased prices will limit the number of vehicles entering the lanes, thereby preserving travel speeds.

Researchers Michael Janson and David Levinson analyzed prior toll and traffic-volume data and set up field experiments in which the researchers altered the relationship between traffic volumes and toll prices, unbeknownst to drivers. The researchers found that toll prices and lane use actually have a  positive relationship; when prices rise, more people decide to take the lanes.

The authors speculate that this phenomenon is the result of drivers interpreting the toll price as an indicator of upcoming congestion. Signs on the roadway provide the toll price, but not a projected time savings that one would enjoy by taking the lane or the level of congestion that one would encounter by continuing on the general-purpose lanes. Though the authors do not suggest this, these findings indicate that priced-lane projects should display both toll costs and benefits in the form of projected time savings.

Furthermore, researchers found that drivers were paying the equivalent of between $60 and $124 per hour to drive in the toll lanes. These values are far higher than transportation planners typically assume when estimating the value of time lost to congestion in commuting. The researchers propose that drivers pay these high per-hour prices because they are not only paying for actual travel-time savings, but also the travel-time-reliability that the pricing system is meant to guarantee.

LINKS

Abstract and Full Paper available from the web site of David Levinson’s research group.