TORONTO:407 video/ET costs revealed


TORONTO:407 video/ET costs revealed

Originally published in issue 40 of Tollroads Newsletter, which came out in Jun 1999.

Page:10

Subjects:video imaging costs on 407

Facilities:407-ETR

Agencies:407 OTCC 407 ETR Conbcession Company

Locations:Toronto Canada

Toronto’s 407-ETR toll system, though having abolished the labor of on-site toll collection remains quite expensive to operate, according to accounts released in connection with a bond sale by the new owners. But the 70c premium payable by non-transponder patrons seems, on our calculations, to roughly cover the high costs of the advanced license plate video processing. And the open road design made possible with automated tolling helps attract traffic that would not be available to a more conventional toll road.

With 230k tolls/weekday and annual revenue heading toward $100m this year, 407 is already one of north America’s major toll roads. (By comparison the Dallas North Toll Road is at $75m, Orlando system $105m, the Sam Houston is $140m, the Oklahoma Turnpike system $124m, Pennsylvania Turnpike $186m, Florida’s Turnpike $280m, the TCA’s three toll roads in southern CA are at $46m)

407 incurred $19m in toll collection costs in 1998 on revenues of $82m. (All US$s) Costs projected for this year and next do not seem proportionately much different. It is early days for 407 of course (tolling began Oct 97) but toll costs over 20% of revenue are on the high side compared to most toll roads. (For example the NJ Turnpike spends about $55m or 19% of its $336m revenue on toll collection costs.) Tolling on 407 is done by Advanced Toll Management Corp a joint venture of Hughes/now Raytheon and Bell Sigma/now CIG on undisclosed terms in a 7-year contract with that goes to 3/31/04. Those are the same companies that built the toll system and they retain the intellectual property rights to it.

The 407 toll system cost $55m to construct and perhaps half its cost is attributable to the need to transmit and process digital images of the rear of vehicles of non-transponder patrons. A 407 image consists of 5 megabytes of data vs 280 bytes for a transponder toll message, 18,000 times more data to move and process and store. Many of the toll costs are common to keeping toll accounts – the costs of customer service, preparing, mailing and tracking accounts. But clearly the great bulk of the 407 fiber optic network and central office computer processing expense is attributable to the demands of video. About 80% of video images are handled by the automatic character recognition algorithms, we were told recently. 15% have to go to operators for what is called Video Exceptions Processing (VEP) where staff sitting at up to 16 workstations use filters, and zoom tools to enhance the images on their screens. About 5% of the images are discarded as unreadable.

If it is assumed that two-thirds of the total toll cost of toll operations is attributable to the video tolling, that is just about recovered in the toll premiums charged for traveling the road without a transponder.

The road is currently running about 73m toll transactions/year of which 70% are by transponder, 30% are by video. Revenue from the video premium of 70c is therefore 73m x 0.3 x 0.7 = $15m.

Conclusion: Enough people are apparently prepared to pay that substantial video premium and to in effect underwrite the costs of the expensive video tolling in order to use the toll road spontaneously. In addition to the premium revenue they pay about $25m in regular per-km tolls. In that light the fancy video tolling system seems to be well justified.

That conclusion is ‘one in the eye’ of American skeptics who have said that the expense of automatic license plate imaging systems can only justified for enforcement purposes and cannot profitably be used for regular tolling.

If only used transponders...

Ed Regan of Wilbur Smith who was in charge of traffic and revenue studies for the 407 says that they calculated during the design phase of the system that if the road had been built like 91-Express, requiring all users to have transponders, then it would be operating with about 20% less traffic than it has. The top fifth of the road’s business therefore is the result of allowing spontaneous use by non-transponder equipped vehicles. That seems consistent with the observed 30% of users who prefer not to get transponders despite the per trip premium of the video tolling.

Another unusual feature of 407 as a toll road is its extraordinary connectedness to the area road network. The completed central section has 28 interchanges (ICs) in 69km (42mi), a close spacing of ICs that sometimes requires ‘basketweave’ grade separation of an entry ramp of one IC from the exit ramp of the next IC. A toll road with on-site toll collection would probably have only about 8 to 10 ICs in that length and therefore would lose some (10% to 15%?) of the trips 407 picks up with connections every 2.5km (1.6mi) on average. By 2002 when extensions east (16km) and west (24km) are complete the total 108km (66mi) highway will have 42 interchanges for an average of one IC every 2.6km.

According to PUBLIC WORKS FINANCE Cintra, the majority owner based in Spain, projects toll revenues of $205m in 02 for an operating surplus of $180m. PWF says SNC-Lavalin has estimated the toll road will return it 12% on equity longterm.

The company reports in its bond prospectus that 407 is presently running with 6% of tolls as “unbillables.” This figure is fairly much in the middle range of toll losses on open systems. (Gated systems like New York’s MTA B&T or closed ticket systems like the NJ Turnpike have almost zero losses by evasion, but systems with coin machines often experience losses in the double digit %.) The 6% loss on 407 includes not only the unreadable license plates, but plates out of jurisdiction for billing (Manitoba, US outside of NY, OH, MI and PA), people who have left no change of address and whose account letters are returned by the postal service, and so-called half-trips – these are trips lacking matching entry and exit registrations, whether by transponder or video. 407 levies tolls by trip and presently requires both entry and exit to be registered and successfully matched in order to compute a toll.

The new owners propose to levy a minimum toll for all such ‘half-trips’ generated by flaws in the toll system. They say they will “attempt to reduce the number of unbillable trips through technology improvements, more reliable transponder transactions and increased enforcement actions.”

It was recently revealed that the 407 system had some serious early difficulties, not reported previously. The most serious was the vehicle classification system. The equipment – laser profilers from SEO of Orlando FL – worked as specified and accurately measured vehicle volume. But the vehicle class was defined around registered gross vehicle weight (5 tons) and the tollster had faulty lookup tables for the registered vehicle mass to volume equivalents. Dozens of calls poured in each day to the customer service center about the system classifying cube vans as commercial vehicles of over 5 tons (hence needing a transponder under the rules) when they were in fact registered at less than 5 tons.

Specifications called for the key gantry antenna-vehicle transponder communications (DSRC) link to work at 99.995% accuracy. One consultant says that this was close to achieved on opening, but that only around 99% of transponder trips produced a full matching pair of transponder transactions. The one percent where there was a bad transponder read on either one or both portions of a trip generated an automatic 70c video premium charge, and over 500 people per day were being erroneously charged.

Customer service overwhelmed

The customer service center was designed with 25 workstations, a number that proved completely inadequate. After the first invoices went out there were over 3,000 customer service calls per day for many weeks. An extra bank of telephone work stations had to be established, and the toll system had to be modified to eliminate the automatic $18 fee for near-5ton vehicles without transponders that the volumetric laser profiling system could not properly classify, and also the 70c video charges for transponder owners when only one transponder read registered in the system.

The customer service center got few complaints about the tolls, almost no problems from the different toll rates at different times of day, and many compliments about the quality of the road and how smoothly it got patrons where they wanted to go. Major complaint was about the $1.40 monthly account fee, which is levied for everyone who uses the road – to cover account establishment and billing costs. Many people who used the road just once in the month were indignant because the one trip triggered the monthly account fee and their single trip was therefore more expensive by that amount than they had expected.

Warranties from the toll system supplier apparently expire late this summer, about 2 years after the system was first deemed fully operational, so the new owners have not had much time to take advantage of expertise about the Raytheon system. Some have expressed surprise at the new owners’ apparent lack of interest in the technical knowledge of consultants who worked with Hughes to bring the system online in 1997 and who developed the work-arounds of the early days – people who know about the system’s strengths and weaknesses from the inside.

On the operational side the province’s agency OTCC never apparently finalized a contract with Canadian Highways Management Corp (CHMC) which has conducted the non-tolling aspects of operations and maintenance of the toll road on a month by month cost-plus deal since it opened. The new owners are completely free to renegotiate that arrangement, or terminate CHMC at any time, and put in their own staff or subcontractor.

In line with the provincial government’s strange secretiveness about 407 operations, the details of the 407 sale agreement and conditions are still not fully public. But we do know this is no laissez-faire regime.

No laissez-faire

The following provincial government controls exist:

• a toll rate ‘threshold’ starts at C11c/km vs C10c being charged now and increases 1.5% then 2%/yr to 30% total, plus inflation/yr

• the govt will establish traffic flow ‘thresholds’ based on peak-hour flows when extensions open end-01, then thresholds are increased 1 to 3%/yr depending on previous year’s traffic to max of 1500 veh/lane/hr

• within the year there is no restriction on toll rate increases if traffic is over the established traffic threshold (allowing a kind of congestion pricing short term) but if the traffic is below the threshold there are traffic shortfall penalties

• if traffic isn’t above the traffic threshold and if toll rates go above the toll threshold then are excess toll rate penalties

• penalties are imposed also for increases in admin fees above C$60/yr, video surcharge going over the threshold of C$1, $1.50 up to $3, increases in the multiples for larger vehicles, if charges off-peak are set at more than peak charges (!?!)

• there are requirements to widen the road within two years after traffic flow exceeds specified lane per hour levels

The province seemed concerned to have the leverage to prevent the 407 operator following a high toll rate/low traffic volume strategy that would dump traffic on nearby free roads maintained by the province or local jurisdictions. Its controls seem designed not to freeze the toll rates and policies of the new owner, but to prevent big and controversial changes in toll rates and structure. Gradual adjustments are possible. The regime seems at first glance very bureaucratic but few complete prohibitions are in place, just penalties for insufficient traffic attracted or for excessive tolls without sufficient traffic.

The conditions of sale provide no guarantees of non-competition from expanded free roads and the new owners can be forced to provide right of way for transit within the 407 corridor (not that transit is ever likely to be competitive with the road.)

The new owners seem to think they have sufficient flexibility. We spoke to Ken Walker the chief financial officer and secretary of the new company. He told us that the group was proud to own an excellent toll road which they hopes to improve on further. He was very positive about the toll technology and the opportunity to manage the road: “We are very proud of the road. We think it (the technology) gives us a lot of flexibility about how we manage it. We think more can be done to take advantage of the technology.”

407 has many firsts. It is one of first toll roads to have differential pricing with the three different toll rate regimes (7c/km in rush hours, 4c daytime offpeak and 2.5c night time.) As Ed Regan of Wilbur Smith says this has enabled 407 to increase its attractiveness to patrons by pitching its toll rates closer to their value of time saved.

The new owners are also very positive about differential pricing saying in their bond prospectus: “Variable tolling will maximize revenue by allowing the operator of the highway to charge based on usage, by time of day (currently peak, off-peak and night-time) by vehicle type and by section of highway” (p13)

Unlike the US where vast quantities of before and after data are collected and analyzed about new systems, the Canadians have sponsored almost zero professional research into the operations and impact of this important new toll road. They don’t seem to know the importance of what they have achieved. (Contact Ken Walker 407 ETR Int Company 416 326 9384 kwalker@407etr.com)