BROOKINGS AEI STUDY:Time in traffic worth only a fifth of pay on average

August 15, 1998
By Peter Samuel

BROOKINGS AEI STUDY:Time in traffic worth only a fifth of pay on average

Originally published in issue 30 of Tollroads Newsletter, which came out in Aug 1998.


Subjects:value of time saved willingness to pay toll discount

Agencies:Brookings AEI Winston

Sources:Clifford Winston Calfee

A new study of commuters by two economic researchers concludes that on average they will only pay about 7c/minute, or a fifth of their hourly pay-rates, to save congestion time on their commute. Some previous studies of the value of time saved have suggested people will pay up to half their pay-rate or 15 to 20c/min in time saving tolls. Uncongested driving time is worth even less to people, only about 2.5c/min, the study says.

The report, quite a downer for tolls, is titled “The value of automobile travel time: implications for congestion policy” and is authored by Clifford Winston of the Brookings Inst and John Calfee of the Amer Enterprise Inst. and published in the JOURNAL OF PUBLIC ECONOMICS (#69, July 1998 p83-102.) They say that previous more optimistic estimates of value-of-time-saved have been derived from mode choice and route choice modeling. These are flawed, they write, because they pick up superfluous valuations of time in a car versus time in a bus or train, and are likely to be comparing the preferences of different people.

The authors think that the reason why real world commuters have such a low average value for time saved is that commuters with a high value of travel time have made residential and workplace location adjustments that result in a short commute and one with little congestion.

Those who hate congestion will organize their lives so that they won’t have to put up with much of it, they suggest. Such adjustment to congestion is also cited as a reason why congestion seems to have stabilized in the last few years. The basis for these findings is a revealed preference survey in which they had a sample of people rank their preference for various combinations of driving commutes involving different levels of congestion and toll cost.

X-lanes don’t deal with average

It seems on our reading to be a well thought out and well conducted study that deserves to be studied by everyone in the forecasting business. But it will have more relevance to complete tolling of a highway than the tolling of a lane or lanes, which is the form of congestion pricing being most actively pursued in this country. On 91-X for example motorists are paying at least twice the Winston-Calfee estimate — 10c to 20c/min saved and on I-15 they have been paying up to 50c/min saved.

Winston-Calfee calculate average value-of-time saved, which is relevant to full highway tolling. But their average value data completely miss the point about toll express lanes which offer every patron on every trip the option of paying the toll and going express, or taking his chances in the free lanes. Toll express lanes are after the business at the top of the distribution of trips by value of time saved, not average time value trips. A great many of their users are occasional patrons, people who normally ride the free lanes but make conscious decisions to use the express lanes on occasional trips when the value of time saved to them is especially high, so the average value of time saved, as calculated by Calfee-Winston is (Can’t resist the jibe) academic.

To spell it out because apparently these guys didn’t cotton on — the value of time saved of the 270k daily patrons on CA-91 might indeed be an average of 7c/min, but 91-X in the middle doesn’t care about that. It is catering to about 10% of those 270k trips where there is a time savings valued at 10c to 20c/min. On I-15X in San Diego it appears that people will pay a lot more, though it is a much smaller proportion still of the total users of the road, because the I-15 X-lanes allow HOV2s in for free and therefore have far fewer spare ‘slots’ to sell.

The authors do not appear to have thought through what is needed for the assessment of the cream-skimming toll express lanes. And these paper exercises of asking people what they might buy are a second best now that there are motorists out there on 91-X and I-15X buying various time savings packages every day. Go west guys and get real data!

Also, a couple of subsidiary findings. Winston & Calfee find no evidence that motorists are concerned with how toll revenues are spent and show little interest in equity issues. And they don’t care whether the toll lanes are run by a public or private agency. In addition motorists are not prepared to pay much to avoid driving alongside trucks. One thing they don’t deal with is motorists’ persistent exaggeration of how much time they are spending in traffic. Time seems to go more slowly in stop-&-go conditions. Repeatedly it is found in other surveys that motorists in traffic will think their trip is, say, 15 or 20 minutes longer than the trip in the express lanes when test floater cars measure the difference as only 4 or 5 minutes. If people cannot accurately judge time savings, and systematically exaggerate it, maybe this seriously detracts from the validity of revealed preference survey data? (Contact Clifford Winston Any distrustful skeptic or pedant who finds this version inadequate can get the original Click on Tables of Contents & Abstracts, select the July 1998 issue, download winston1723.pdf and read it with the awesome Acrobat.)

Leave a comment: