THE SULLIVAN REPORT: 91X validates toll express lane principle

91-Express (91X) is “working well” and the idea “merits consideration” elsewhere. Those are the main conclusions of the official independent evaluation of the celebrated first toll express lanes project done for Caltrans with Fed Hwy Admin support. The $965k 4-year study led by Edward C. Sullivan, engineering prof at Calpoly San Luis Obispo, gathered and analysed a vast amount of before and after traffic count data from pavement loops, performed public opinion polling, origin & destination work, filmed weaving patterns and other driver behavior, did extensive interviews and observation, and ran test runs for travel times. The report, written in low key prose, paints a picture of successful operation, major transp benefits and broad public acceptance.

Concludes Sullivan: “The express lanes constructed on California State Route 91 have demonstrated that providing new highway travel options, in this case, premium service for a premium price, can win public acceptance and produce significant travel changes. In many urban highway corridors there are limited options available to persons whose trip urgency or personal situation justify paying extra to better satisfy their travel requirements. Providing additional choices through pricing is working well in its first major American implementation in California. What we have learned from studying the impacts of the SR 91 value priced express lanes suggests that this innovative approach merits consideration for further experimentation elsewhere.”

It is too early to judge the financial viability of the project, because while it has reached cash-flow break-even, it faces a difficult-to-predict loss of business to the partially competitive Eastern toll road (CA241) which opens within weeks. (See TRnl#30 Aug 98 p9)

12 to 13min time saving

The operating 17km (10mi) X-lanes facility in mid-97 (the last survey date) was providing average time savings of 12 to 13mins in the normal peak hour, and that implies a value of time saved for 91-X users of $13 to $14/hour, Sullivan says. The time savings were obtained by clocking test cars running in the X-lanes and the regular lanes.

Once again Sullivan documents the way motorists have an exaggerated notion of X-lane time savings, though the exaggeration is not as sharp as reported on I-15X in San Diego. 11% of 91X-lane users say they save over 30mins, 24% say they save over 21 to 30mins. 48% get it about right saying they save 11 to 20mins, while 18% are on the low side saying less than 10mins. Of course the variability of time savings from day to day makes such calculation necessarily rough. Perhaps people dwell on the over-average and most satisfying time savings. They are buying greater certainty about trip time — greater reliability — as well as a better average time. Time savings cannot be the only benefit however, Sullivan says, since the X-lanes continue to be utilized and tolls paid by some when there are negligible differences in driving speeds.

Average time savings are cited by 47% as their only reason for being in the X-lanes, the report says. 6% stress superior predictability as another main reason for using the X-lanes, 25% cite greater safety, 16% cite easier and more comfortable driving (trucks are not allowed) as draw cards for the X-lanes in addition to time-savings. 7% of those using the X-lanes said “Yes” to the point that they get a vicarious satisfaction from or enjoy feeling priveleged and passing folks in congestion.

About 7% of traffic on 91 chooses the X-lanes midday weekdays when there are negligible time-savings and the toll is $1.00, which suggests a considerable number of people willing to pay for the non-time savings aspects.

Used very selectively

Most striking is the finding that a large majority of patrons use the X-lanes selectively, despite a major program involving discounts by the operator Calif Private Transp Company (CPTC) to recruit routine frequent users. Only 23% of those on 91X use it every weekday (for 1, 2 or more trips), 29% use it 2 to 4 times per week (about a third of the their trips if they are commuters), and 16% use it about once a week. A third of those on 91X at any time use it less than once a week. Nearly half of those less than once-a-weekers (15% of the total) use it less than once a month. CPTC’s business plan assumed a much higher number of regular users and many fewer of the infrequents.

Over 100k transponders are on issue by CPTC but nearby TCA’s larger inventory of transponders are regularly used on the 91X lanes as well, so there are perhaps 180k transponders on issue that get used on 91X or 5 to 6 times the number of daily transactions.

The most common reason given by users for choosing to use the X-lanes is that the congestion in the free lanes is “unusually bad.” The prospect of being late for work or for an appointment are other important factors that generate use of the toll X-lanes, according to Sullivan’s surveys.

About half the travellers on 91 say they have used the X-lanes and about half say they never use the X-lanes. Some of these use interchange ramps off the free lanes not served by the X-lanes. The others say simply that the time savings of the X-lanes are not worth the toll price. A small proportion would be ineligible to use 91X (big trucks).

Rich use more but...

Higher income motorists use the toll facility more often, but all income levels use it. Those with household incomes under $40k use it for about 10% of trips, whereas those $40k to $60k use it for about 20% of trips, the same percentage as those over $60k income. Though there is clearly a skew toward higher income use, Sullivan says, the substantial use by lower and middle income groups shows it is “incorrect” to claim the benefits go mainly to the rich, as critics are wont to charge. Women make greater use of the X-lanes than men, 42% of women being frequent users vs 28% of men.

These data seem to consistently reinforce the point that a very diverse group of people make up the X-lanes users, with diverse reasons for using them. The clever “Lexus lanes” label of a few critics is a simplistic ideological caricature that, while it contains a small element of truth, is considerably more wrong than it is right. A minority of those using it no doubt don’t give a damn about the cost, but the sheer variability of usage suggests the large majority of patrons do some mental calculation about whether the higher quality service of the X-lane is worth the toll which is posed on the variable message sign on the approaches. Most people are really making choices. It is not the “two class” system of anti-road choice demagoguery, a label which might be justified if there was one coherent group of motorists that used the facility all the time and another that never used it.

“Induced traffic”

Total traffic on 91 increased from 199k veh/day Dec 95 before the opening of the X-lanes to 227k in Dec 96. That’s an increase of 28k veh/day, 14.1% up. Sullivan estimates that prior trends this decade would account for 5.4k of the increase (2.8%) leaving 22.6k or 11.4% as the extra traffic attracted onto 91 by the extra capacity and associated better level of service that 91X brought. Sullivan traces 6k veh/day as a diversion from nearby competitive surface arterials (La Palma & Santa Ana Canyon avenues), leaving new traffic in the corridor at 16.7k veh/day, 8.4% of the pre-X lanes volume — new traffic or latent demand.

Transit in the corridor does not appear to have been affected one way or another by the X-lanes, Sullivan concludes. It consists of express bus (300 to 500 riders) which seems most affected by service frequency (which changed) and a Metrolink commuter rail service (5 trains a day) which opened in Oct 95 just 10wks before 91X. The rail too is very small, about 1,000 persons total daily ridership. Park & Ride and van usage didn’t change much overall.

Vehicle occupancy trends on 91 overall are conflicting. HOV3 which got free rides in the X-lanes in the first two years (they by now pay half toll) increased markedly. PM peak hour HOV3s went from 430 in 1995, to 705 in 96, to 780 in 97, an 81% increase over 2 years. By a slightly different measure after 6 months HOV3s increased their share of vehicles in the total traffic from 3.7% to 4.4%. The more numerous HOV2s went up only 10% (2.8k to 3.1k) and dropped from 20.3% to 17.4% of total vehicles. SOVs in the rush hour went up 31% (from 10.5k to 13.8k) just slightly increasing their presence in the traffic (from 76% to 78%).

Overall there is a slight decline in average vehicle occupancy, Sullivan calculates — a conclusion contrary to the impression generated earlier in reports on 91X by a focus on the strong rise in HOV3s alone. A graph 1994 to 1997 suggests the slight downward trend in average occupancy was in place well before the X-lanes so the X-lanes themselves seem to have had little independent net effect.

Snub-nosed bullet graph

The Sullivan report comments that traffic in the free and toll lanes on 91 seems to have reached an “equilibrium” in the rush hour in which the 2x2 lane toll X-facility regularly carries up to 35% of total 91 traffic in the overall 12-lane 230k veh/day roadway. Over the full 24 hours weekdays it carries only a bit over 30k veh/day, just over one-eighth the total traffic, and at weekends 17k veh/day a similar proportion.

91X’s varied toll rates keep traffic in the express lanes free flowing during the rush hours, carrying approx the same 1.5k to 2k veh/lane/hr as the regular lanes. The great difference however is that 91X runs that volume at 65mph whereas 91-regular lanes alongside run about the same per-lane volumes at average speeds of 30mph to 40mph.

At the peak of the pm rush while average speeds in the 2x4-free lanes were 15mph before the opening of the X-lanes, 6 months after the opening of the X-lanes speeds in the free lanes were nearly 40mph — due mainly to the removal of a lot of the rush hour traffic to the X-lanes. After another year (Jun 97) and the full increase in overall traffic on the 12-combined lanes of over 14% (just about exactly what the X-lanes are carrying on a daily basis) speeds in the free lanes were down to about 32mph in the worst of the pm peak. The toll lanes are taking the peak hour traffic away and the free lanes traffic is less peaked, or more spread.

Before the X-lanes the 2x4 regular lanes were running 25k veh/lane/day, but the addition of 2x2 X-lanes increased overall capacity 33%, which with the 14% increase in traffic, reduced daily vehicles/lane to 19k. The report doesn’t say this but Ed Sullivan points out in discussion that in constructing the 91-X Caltrans had the CPTC contractors add some auxiliary lanes to the unrestricted lanes outside the termini of the X-lanes, which in themselves may have helped reduce bottlenecks, so before-&-after analysis is not a clean or completely clearcut measure of what the X-lanes alone wrought.

“Apparently an equilibrium has been established between the free lanes and the (toll) express lanes, “ writes Sullivan in an important proposition.

That magic equilibrium

The rush hour equilibrium is especially fascinating since it would appear — the report does not go much beyond the general statement — that the X-lanes are carrying about the same per-lane volumes at free flow 65mph as the general lanes are carrying at around 32mph (see TRnl snub-nosed bullet graph sketch). Perhaps we have the managers of the X-lanes raising tolls to cut back volumes in the X-lanes preventing their facility going from the top of the nose onto the slippery tip and sliding back down under. Maybe we have a LOS B/C on the top slope of the tip and LOS E/F on the bottom slope?

Here’s a hypothesis crying out for expert validation, examination and analysis. A more fine grained examination of the X-lanes vs regular lanes data than Sullivan has been able to do in his report could provide great benefits to planners and managers of possible future toll express lanes.

If indeed the same rush hour volume of traffic/lane can be carried in comfortable free flow at twice the average speed of unrestricted traffic by using a variable price to meter entry volumes and to prevent traffic flows from going over the edge of the snub-nosed bullet to the stop&go, creep/fly/brake/creep-again travel speeds down under on the graph (see sketch), then the variable price regime could be (1) an extremely potent tool for traffic managers, (2) a possible way of reducing average air emissions, and (3) such a source of benefit to motorists that it should be capable of generating a large revenue stream. Now, that last long sentence was deliberately hung on a big “If” because the analysis to fully support it has not been done in this report beyond Sullivan’s important statement that an “equilibrium” including rush hour traffic streams of 65mph in X-lanes and 32mph in unrestricted lanes has apparently been established with similar volumes per lane. The characteristics of that equilibrium need to be fleshed by further research.

We do know the promising fact that in the rush hour on 91 up to a third of the traffic will pay a toll of nearly 20c/km (32c/mi) to travel free-flow in X-lanes at double the speed of the uncongested lanes. That’s a strong willingness to pay from a third of the motorists in a major traffic stream.

The second rich lode for research results will be laid down shortly when the Eastern toll road opens and siphons a proportion of traffic off 91 completely. The whole new situation 91-pre Eastern versus 91-post Eastern, and CPTC’s management of the change on 91X seems likely to be a rich new source for information about X-lane performance.

Public opinion

Sullivan’s report includes extensive data on public attitudes toward toll express lanes. 80% of commuters on 91 supported increased capacity however it was financed, with over 60% of the public supporting toll-financing. Support for varying the toll rates with congestion was only 45% before implementation but since opening of 91X has risen to the 60 to 70% range. There is now close to 50% support for the more flexible dynamic pricing of tolls (I-15 style) as compared to the bar-chart style stepped hourly schedules in use by CPTC on 91X, the Sullivan report finds. And there is somewhere over 85% patron support for photo-enforcement of toll violations.

Comment

This X-lane stuff is apparently a political winner in southern Calif — because it works for motorists. Whether and in what circumstances it works for investors remains unclear. The new Eastern pike may add so much extra capacity (6-lanes) that the new equilibrium 91X vs 91-free will be at non-viable toll rates. For congestion pricing, you must have congestion!

But we know now that variable tolls are potent and quite workable tools for adjusting traffic. And that those who doubted it would work, or be politically acceptable, were plain wrong. Jerry Pfeffer who set up CPTC and designed 91X, and Greg Hulsizer who carried it on, and their counterparts at Caltrans Carl Wilson and Joe el Harake have pioneered something very big for roads and potentially great for motorists.

[The 133 page first volume of the report titled “Evaluating the Impacts of the SR 91 Variable Toll Express Lane Facility, Final Report” dated May 1998 became available mid-Sept. Principal author is Ed Sullivan, prof civil eng at Calpoly San Luis Obispo tel 805 756 2947 esulliva@calpoly.edu A 2nd larger volume contains the blank questionaires used in the many surveys. Sullivan was assisted by about 50 students, graduates, lecturers, and a few subcontractors and consultants. Paper copies of the reports are available from Mike Gray Caltrans 916 654 6133 Mike_Gray@

dot.ca.gov It will be downloadable from a Calpoly website shortly possibly http://airship.ardfa.calpoly.edu/~jwhanson/sr91.html. The data in an Oracle database and on floppies in Excel format is for sale at cost from Sullivan at esulliva@calpoly.edu Major problem encountered in the study was the erratic performance of Caltrans pavement loops. Sullivan’s team wrote special software to subject all the loopy data to various logical tests, then fixed minor anomalies by interpolation or discarded data as bad.]