Toll collector productivity

350 vehicles/hour is traditionally a standard average workload for US toll collectors. Working flatout in the rush hour an attendant can take tolls from 600. Since an expressway lane has been thought to have a capacity of about 1,800 to 2,000 v/h it has been a rule of thumb of tolling that at mainline barrier plazas you must have 3 tolling lanes for each traffic lane (1800/600) and so a 4-lane toll road has minimum 12-toll-lane plazas, 6-lane toll roads 18-toll-lane plazas, and so forth for two-way tolling (or half those numbers in case of one-directional tolling as on the Hudson River NY crossings.) Traffic flows faster and at lesser headways in the 1990s than when most of these standards were derived — in the 1960s and 1970s. Better brakes and tires, perhaps, or Atari-style computer games, encouraged the current generation of drivers to beat along faster and closer to the guy in front. In lanes with cars predominant, such as the inner lanes of wide 8 and 10-lane expressways (as in Los Angeles) or where there are separate cars-only carriageways (as on the inner 6-lanes of the NJ Turnpike), or where entering traffic is smoothly introduced with ramp meters (Minnesota) up to 2,500 vehicles/lane have become common.

The greater throughput on the mainline means more pressure on existing plazas. Either they have to be widened and provided more toll lanes (at an extreme to a 4/1 ratio of toll lanes to travel lanes in place of the old 3/1 ratio.) Or you have to upgrade your productivity. Electronic tolling obviously is one way, though it is not easy to main the proper balance between dedicated e-toll lanes and mixed lanes. In countries with a functioning coinage system, automatic coin machines (ACM) are an option (they'll usually double throughput), but in the US, with the 'quarter' the largest circulating coin, too few patrons will carry the number of coins needed to pay tolls at ACMs. Some tollsters mint their own 'coins' called tokens. But for the forseeable future there's going to be a need for attendants at many toll plazas and a need to have them work as productively as possible. (In the U.S. some of them cost over $20/hr — a gross multiple of that made by this humble hack!)

Henry J. Stanton, chief of the New York State Bridge Authority (NYSBA) which operates five toll bridges over the Hudson River (70km to 200km north of New York City) was challenged by his state comptroller's office a while back to defend his existing practice of adding a collector whenever another 350/veh/hr were added to throughput. He found few studies of toll collector productivity. That got him into an interesting investigation of his coillectors' productivity. The authority's toll monitoring computers were already recording to a hundredth of a second, Stanton was delighted to discover, the time of treadle registers, classification key-ins, and cash register hits.

Wide variability: Stanton found huge variability in toll collector performance: "Toll collectors can be extremely erratic in the way they process transactions. Some ring up vehicles as they roll in. Others wait till the vehicle stops. Still others wait until they have cash in hand. Some (not the good ones!) do it differently from vehicle to vehicle."

The second busiest bridge, the Mid-Hudson at Poughkeepsie had the most productive toll collectors — average 600 v/h in its mixed traffic lanes in rush hour. At the busiest on I-84's Newburgh-Beacon bridge mixed lane throughput in rush hours averaged only 322 v/h. Part of the discrepancy is accounted for by a higher proportion of commuters' at the Mid-Hudson (61% v. 40% at I-84) since regular users are more adept at paying efficiently and pay by simply handing over a ticket. Also I-84 has more big trucks (13% vs 5%).

Across the 5 bridges commuters paying with a ticket took 4.8sec average, cars and other light vehicles paying 75c toll took 10.6sec and others, mainly large trucks took 18.5sec. The NYSBA has now set performance standards for its collectors: 5sec, 11sec, 20sec. And Stanton points out that if the toll was raised to $1, the second class of transaction time might be halved to about the same 5sec as ticket payment.

The authority has implemented major economies following the study:

• replacement of five fulltimer shifts/week with five 4-hour part-timer shifts at the 3 smaller bridges

• reduced staffing at Newburgh-Beacon I-84, and use of only 6 lanes in place of 7 or 8 in rush hours — boosting average productivity 18%

• with increased ACM lanes 5 fulltime collector shifts were eliminated on I-84

• other staff reduction by attrition and a retirement incentive

Overall toll collection costs have been cut 10% with the potential for another 10% once full staff cuts are achieved. (Contact H. Stanton tel 914 691 7245)