They Pay Farmers Not to Grow Crops, Don’t They?
Despite the odd title, this is an optimistic report. Before describing a breakthrough in how to reduce urban traffic congestion, let me summarize the problems with current strategies.
We are all travelers. We go to work, school or college, shop, out to eat, visit medical facilities, friends and relatives, and take part in sporting or recreational events of all kinds. Since most travel involves roads for part of or our entire trip, traffic congestion interferes with each of these activities.
Traffic congestion is listed as one of the nation’s most serious transportation problems. The Texas Transportation Institute’s Urban Mobility report estimates the cost of extra travel time and fuel at $121 billion in 2011, with corresponding large costs due to environmental emissions.
Given the importance of reliability and predictability in today’s economy and the role that access to labor/jobs plays in economic productivity, the overall cost to the economy is likely much greater. The nation’s economic history shows significant periods of growth around times when “excess” capacity served to stimulate economic growth – the Interstate Highway System is the classic transportation example – but the Internet and wireless communication offer other examples of the economic and social power of network economics.
Before addressing a new approach to reducing peak congestion, let’s first review our current list of tools and why they have failed.
- The traditional approach is to simply “live with it.” That is we simply adapt where we live and work and when we travel in order to control the annoyance and economic losses from congestion.
- The classic answer of “add new lanes” has been replaced with “we can not build our way out of congestion.” The lack of funds, along with environmental hurdles, makes this difficult (other than a few locations where toll roads make sense). According to one State DOT Secretary: “The best we can do is to slow the pace with which congestion gets worse.”
- Improve operations. This is a practical approach, with a host of tools ranging from coordinated traffic signals to peak-hour use of shoulders to high occupancy toll (HOT) lanes. Delcan and Ken Button from George Mason University recently completed a study for the Federal Highway Administration that estimates the impact on traffic congestion of a number of these techniques. Each has a positive impact but the focus is on squeezing additional capacity from the current system.
- In time automated vehicles and connected vehicles will offer an opportunity to reduce headways thus creating the potential to add significant effective capacity to roadways. Despite claims of early deployment (Google possibly by 2017; Nissan and GM by 2020), many questions remain and it will be 2030 or later before a significant portion of the traditional fleet has turned over.
- That leads to congestion pricing as a way to encourage less important trips to change when they travel or to shift modes. This technique has a long and noble history: William Vickery received the 1996 Noble Prize in economics in part for his work on congestion pricing. There are many variations on how congestion pricing might be implemented. Successful systems operate in Singapore (the grandfather), Stockholm, and London among others. Despite considerable promotion, relevant technology, and even an attempt at some federal subsidies under the Bush administration, nothing has been implemented in the United States. HOT lanes with variable prices for parts of specific roads are more popular.
Congestion pricing involves a profound change in how we use transportation. All roads become toll roads, sometimes with quite high rates – more than $15 dollars (ten pounds) per weekday to travel into Central London. Rates in Stockholm are lower, a bit more than $3 during the peak hours. Implementation can be expensive. Political support is crucial and despite success in Singapore and a number of cities in Europe, congestion pricing has yet to make progress in North America.
The fundamental principle behind transportation pricing is that not all trips are equally important. A higher price during peak periods will encourage some travelers to shift travel to off peak periods, shift to another mode (transit, bike, walk, or carpool), or not travel at all (telecommute). While economists like this idea, most travelers do not appreciate being forced to pay something for a previously free service.
What if, instead of making everyone pay more, we paid a small number of people NOT to travel during peak periods? These folks would be quite happy. Regular commuters would be pleased as well since overall traffic congestion would shrink, although probably not eliminated.
But this is a crazy idea! Who ever heard of paying someone NOT to do something? The U.S. farm program pays subsidies to farmers not to grow crops in environmentally sensitive areas and makes payments to farmers based on what they have grown historically, even though they may no longer grow that crop. Of more relevance, several transportation programs exist that do pay people not to travel at peak periods or not to use a single occupancy car.
Example 1: Rotterdam
Between October 2009 and December 2012, travelers along the A15 roadway near the harbor were offered financial incentives to avoid the road during the 6:00 to 9:00 AM peak and 4:00 to 7:00 PM peak while construction was underway. The payments were quite large: 6.5 Euros for avoiding both peak periods and as much as 5 Euros for the AM peak alone. Participants received a smartphone to record their plans and after one year or 100 peak period avoidances, they could keep the phone. The project reduced peak hour travel by 7 percent or about 800 morning trips. The firm in charge of the program was paid based on the number of trips that were diverted. Participants were recruited via the Internet and were required to remain active in order to stay in the program. In addition to monitoring the smartphone location, license plate surveys were used to confirm off peak use. Just over half of the potential peak hour trips were diverted, with most travelers simply changing the time of day that they traveled; a few changed their route, others used transit or bikes, and a small fraction worked from home. The Rotterdam project is now complete and a similar system in underway in Utrecht, the Netherlands.
Example 2: Washington DOT Trip Reduction Performance Program
For about 10 years the Washington State Department of Transportation (WSDOT) has operated a Trip Reduction Performance Program (TRPP) as part of its Commuter Trip Reduction program, which was established by the state legislature rather than WS DOT. Every two years WSDOT issues an RFP to public and private firms with more than 100 employees. Grants are based on the number of peak-hour single occupancy trips that each agency will reduce and the requested payment per trip. Trips can be diverted to transit, carpools, or vanpools. Successful bidders also provide their own funds. Overall, the program has exceeded its targets. The annual public subsidy paid per trip reduced range from $235 in 2005 to $375 for 2009 – less than $1 per trip! Unlike Rotterdam, funds are not paid directly to individual travelers, but rather used to subsidize transit, carpools or vanpools. Telecommuting has not yet been a focus of the program. In addition, the state has tax credit programs to encourage use of carpools and vanpools. Overall, the WSDOT program says it has reduced single occupancy trips by 4.8 percent and VMT for work trips by 5.6 percent. WSDOT estimates that this has reduced congestion by more than 10 percent. For the 2007-2009 period, the legislature appropriated $2.5 million for this program but the program has since been eliminated due to State budget problems.
Example 3: Stanford University
The CAPRI (Congestion and Parking Relief Incentives) system at Stanford University provides employees and students with an incentive to bike or walk and to avoid driving between 8:00 and 9:00 AM and 5:00 and 6:00 PM. This system was developed by Professor Balaji Prabhakar and funded with a $3 million grant from the USDOT. Participants receive either cash of ten cents per trip avoided or a chance to win up to $50. The ten cents per trip (at most $1 a week) is significantly lower than the costs for the Rotterdam experiment and the WSDOT program.
Example 4: Singapore peak-hour train
The INSINC program is similar in concept to the Stanford CAPRI system but aimed at encouraging commuters to avoid the 7:30 to 8:30 AM peak travel time on commuter trains in Singapore. Travelers receive points based on shifting to off peak trains and the points can be used in a game that then makes payments in cash. It does seem ironic to use a pricing scheme to help resolve overcrowding on transit lines in the city that invented congestion pricing for cars as one way to encourage commuters to shift to transit.
The idea of paying people not to drive a single occupancy car during the rush hour has some appeal. It is a targeted approach. Rather than trying to add capacity for everyone or to charge everyone more for their normal trip, funds can be focused on those who are willing and able to shift when they travel or how they travel. Costs should be significantly lower than for the other approaches, and the politics of implementation much easier. The ability to shift even five percent of peak hour traffic can generate a significant drop in congestion. One problem is the need to ensure peak-hour trips actually are avoided. Technology helps, but any system should involve incentives to minimize cheating. Financial incentives to the program manager as used in Rotterdam make sense, but it makes sense to calculate the fraction of cheating that could occur without harming the end result.
The examples mentioned here vary widely. The Rotterdam experiment shows how technology can be used to recruit and then confirm behavior. While this system is performance-based and run by private operators, the financial incentive of as much as $6.50 (5 Euros) per AM trip seems quite high. The Stanford and Singapore rail examples offer much lower costs per trip, perhaps the chance to win more appeals to participants. The WSDOT program covers multiple urban centers in the state and thus is less direct in terms of geography and relies on matching funds from public and private partners.
How might such a direct payment system work in a U.S. environment? What can be learned from the technology used in Rotterdam? How much of a payment is needed? Are there opportunities to reduce costs even further with the gaming system used at Stanford or with the indirect subsidies used by WSDOT? How many trip reductions are needed to make a noticeable improvement in congestion? Finally, are there any jurisdictions in the U.S. that are willing to test this concept?
Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the official policy or position of The Eno Center for Transportation.