Value Pricing and Freight Traffic by alicejenny


									64                                                                                                           s Transportation Research Record 1707
                                                                                                                                Paper No. 00 -1112

Value Pricing and Freight Traffic
Issues and Industry Constraints in Shifting from
Peak to Off-Peak Movements

Pierre Vilain and Paul Wolfrom

The Port Authority of New York and New Jersey (PANYNJ), in associa-               facilities between the hours of 6 and 10 a.m. underpriced if the toll is
tion with Louis Berger & Associates, Inc., completed a study to determine         invariant? If value pricing strategies were initiated, would a signifi-
the potential for reducing peak-period commercial traffic at the interstate       cant proportion of commercial traffic now using the crossings be
crossings between New Jersey and New York City. One of the policy                 encouraged to switch to the off-peak? (Whereas the value-pricing lit-
options examined, and the focus here, was the feasibility of encouraging          erature usually emphasizes congestion surcharges, PANYNJ carried
a reduction in peak-period congestion through the use of congestion- or           out the study under the assumption that value pricing could take sev-
value-pricing incentives. (PANYNJ recognizes the need for a coordinated           eral forms: a peak-period surcharge, an off-peak discount, or even a
approach to reducing congestion involving as many of the regional trans-          mixture of both. These were usually the types of value-pricing options
portation authorities as possible. Also, it is clear that congestion reduction    discussed with trucking industry representatives during interviews.)
strategies must focus on all vehicles and not simply commercial traffic.             In answering these questions, theoretical models of the freight
The purpose of this research was to better understand the constraints fac-        industry are ambiguous. Typically, such models stress that a funda-
ing that segment of the market. It was not intended to suggest pricing poli-      mental aspect of commercial traffic is its nature as a derived demand,
cies focusing exclusively on commercial traffic.) Summarized are some of          namely derived from the demand for the good being shipped. Further,
the findings of a large number of interviews carried out with trucking             incorporating the assumptions of microeconomics regarding the
firms, in particular key personnel of the firms in positions of responsibil-        behavior of producing firms, one can derive a demand curve for
ity or authority with respect to scheduling of deliveries. The interviews,
                                                                                  freight transport that is itself downward sloping with respect to the
each fairly detailed and in-depth, elicited significant and valuable infor-
                                                                                  shipping rate (5). Accordingly, imposing a peak-period congestion
mation to help understand what could be the response of commercial
                                                                                  surcharge could be expected to reduce the demand for peak-period
traffic to value-pricing initiatives. Another part of the analysis also is dis-
                                                                                  freight deliveries, possibly increasing the demand for a substitute such
cussed, involving calculating the total value of tolls at the interstate cross-
                                                                                  as off-peak deliveries.
ings as a proportion of the generalized cost of travel (GCT) facing trucks.
                                                                                     However, extending the analysis to include inventory costs, as well
The analysis was carried out to assess how much the GCT would be
                                                                                  as uncertainty with respect to the actual timing of demand facing
affected by value-pricing incentives.
                                                                                  firms, complicates matters. As will be discussed, firms have a strong
                                                                                  incentive to minimize these logistic costs (which we define as includ-
The rationale for value pricing is by now well known and described                ing inventory costs, as well as costs associated with the reliability of
in numerous publications, including Vickrey, Keeler and Small, Hau,               shipping deliveries). This could imply that, for certain firms, the value
and Kain (1–4). Briefly, the theory states that the private cost of using          of deliveries made at an optimal time with minimum degrees of cer-
a road network may be significantly below the “social” marginal                    tainty is high. In other words, the derived demand curve for certain
social cost, as users typically will not be considering either their              types of deliveries may be quite inelastic to the shipping rate, imply-
effect on overall congestion on the road or other factors, such as the            ing that congestion surcharges could have limited impact on this type
cost of road maintenance. Under these conditions, the use of the road             of commercial traffic. For many firms operating under the constraints
is underpriced, and the eventual number of users will be higher than              of inventory cost minimization, it could well be that the cost of a
would be the case in a social optimum. One possible remedy is to                  freight delivery made after the start of the business day outweighs an
impose a toll that would add to the private cost facing users (typically          imposed congestion surcharge levied on the peak-period delivery.
time in travel and vehicle operating costs). Ideally this toll could                 The interviews carried out with members of the regional freight
approximate the difference between the private cost and the social                community confirmed the proposition that a significant amount of
marginal cost.                                                                    deliveries made during the peak period were indeed deliveries made
   A large focus of value pricing is to address the imbalances                    to firms operating under strong inventory cost constraints. But other
between congestion levels at peak versus off-peak periods. These                  factors also are significant in explaining why so many deliveries con-
imbalances imply that a unique toll charged for the use of a facility             tinue to be made during the peak period of the working day. These
could be underpricing the use of the facility during peak periods. To             factors are described after detailing the characteristics of PANYNJ’s
what degree can this rationale be applied to commercial traffic using             interstate crossing facilities.
the interstate crossing of the Port Authority of New York and New
Jersey (PANYNJ) during the morning peak period? Is the use of the
                                                                                  PANYNJ INTERSTATE FACILITIES
P. Vilain, Louis Berger & Associates, Inc., 797 South Second Street, Philadel-
phia, PA 19147. P. Wolfrom, Port Authority of New York and New Jersey, Office     PANYNJ’s interstate transportation facilities include the following
of Policy & Planning, One World Trade Center, 61N, New York, NY 10048.            six crossings between New Jersey and New York City:
Vilain and Wolfrom                                                                                                  Paper No. 00 -1112        65

  •   George Washington Bridge,                                           Weekday truck traffic at the George Washington Bridge exhibits a
  •   Lincoln Tunnel,                                                     pronounced peak in truck traffic between 4 and 10 a.m., with peak-
  •   Holland Tunnel,                                                     hour traffic of more than 1,100 trucks between 6 and 7 a.m., the high-
  •   Goethals Bridge,                                                    est hourly truck volume among the Port Authority’s six crossings.
  •   Outerbridge Crossing, and                                           Weekday truck traffic at the Lincoln Tunnel also displays a morning
  •   Bayonne Bridge                                                      peak pattern, with peak hourly volume reaching 380 between 5 and
                                                                          6 a.m. Truck traffic at the Goethals Bridge displays an even flatter
   Figure 1 provides a map of the 17 counties that comprise the           morning peak pattern, with its peak truck volume reached between
PANYNJ 17-County Metropolitan Region. Figure 2 outlines the               8 and 9 a.m. As with the George Washington Bridge, trucks at
major truck corridors in the region, and Figure 3 shows the crossings     Goethals Bridge consist mainly of large trucks, with their share
in relation to the major regional port facilities. These maps show how    accounting for 50 to 80 percent, depending on the time of day.
crucial the interstate transportation facilities are to the flow of com-      This pattern is evident at all six crossings, with each facility dis-
mercial traffic within the region, providing the primary road link        playing a peak period during the morning hours, though the timing
between New York City and Long Island on the one hand and New             and spread of the peak may vary. It is interesting to note that except
Jersey and other points to the west on the other. Further, a significant   for the George Washington Bridge and the Lincoln Tunnel, the peak
volume of traffic bound for New England also uses various interstate      hour for trucks typically is 1 h earlier than for all traffic. Overall,
crossings, particularly the George Washington Bridge.                     the data imply that a significant proportion of truck traffic already
   Table 1 outlines the average weekday eastbound traffic volumes         tries to avoid the periods of peak congestion by making use of the
for each of the crossings (or southbound in the case of the Bayonne       peak-period “shoulders.” Despite this use of prepeak periods, the
Bridge). Since PANYNJ currently imposes tolls only in the east-           heaviest truck traffic still occurs between 5 and 8 a.m., contributing
bound direction, our analysis is focused on eastbound traffic pat-        significantly to congestion at the crossings.
terns and the possibility of influencing the morning eastbound
peak through value pricing.
   As shown in Table 1, truck traffic is a significant proportion of      TOLLS AT INTERSTATE CROSSINGS AS
total vehicle traffic at the crossings. Also included in the table is a   PROPORTION OF GENERALIZED COST OF TRAVEL
more accurate measure of the contribution of commercial traffic
to congestion. This is obtained by converting the truck population        Trucks face a current toll in the eastbound direction of $4 per axle,
into personal car equivalents (PCEs), in which large trucks typi-         which equals $16 for the typical large truck. How important is this
cally are each given values of 3 car equivalents and small trucks         toll as a percentage of the generalized cost of travel (GCT) facing
2 car equivalents. It is calculated that in terms of PCE, the share       trucks making a peak-period interstate crossing? The question is of
of total traffic accounted for by trucks is over 23 percent for the       interest, since a congestion surcharge that only marginally affects
George Washington Bridge, 10 percent at the Lincoln Tunnel, and           the GCT may not significantly change the behavior of these trucks,
nearly 25 percent at the Goethals Bridge.                                 while a surcharge that affects the total GCT to a considerable degree
   In terms of patterns by time of day, Figures 4 through 6 show the      should be more likely to do so.
hourly patterns for this weekday eastbound traffic, by vehicle type          In order to carry out the analysis rigorously, the Port Authority’s
for three representative facilities. Note that the data again are dis-    Interstate Network Analysis (INA) model was used. The INA model
aggregated into small trucks and large trucks of more than two axles.     is a regional road network assignment model that covers the entire

      FIGURE 1   PANYNJ 17-County Metropolitan Region.
FIGURE 2   Major truck corridors in the Port Authority of New York and New Jersey metropolitan region.
FIGURE 3   Port Authority interstate crossing facilities and regional port facilities.

                TABLE 1     Daily Eastbound Traffic at PANYNJ Interstate Crossings, Typical 1998 Weekday
68     Paper No. 00 -1112                                                                                  Transportation Research Record 1707

                                                                            pocket toll expenses. For example, the estimate for the GCT for trucks
                                                                            traveling between Bergen County, New Jersey, and lower Manhattan
                                                                            during the peak period is $56, while the GCT for the longer trip
                                                                            between Bergen County and Nassau County on Long Island is over
                                                                            $92. Not surprisingly, the highest GCTs are for the relatively long
                                                                            trips between Orange County, in the Hudson River Valley, and the
                                                                            various destinations in New York City and Long Island, or those
                                                                            between Middlesex County, New Jersey, and Nassau County.
FIGURE 4 Weekday hourly percentage of daily traffic, George
Washington Bridge eastbound.                                                   The relative importance of the interstate crossing tolls as a propor-
                                                                            tion of GCT then is calculated for the same origin-destination pairs,
                                                                            as outlined in Table 3. As expected, the $16 toll is a fairly high pro-
                                                                            portion of short trips, such as those between Bergen County and lower
                                                                            Manhattan, but a smaller proportion of longer trips. For example,
                                                                            for trucks traveling between Middlesex County and Nassau County
                                                                            on Long Island, the $16 toll is equal to 10 percent of the GCT. In
                                                                            this case, even a doubling of the toll would only raise the total trip
                                                                            cost by 10 percent. However, doubling the toll for trips between
                                                                            Bergen County and lower Manhattan has a more pronounced effect,
                                                                            increasing total trip cost by 29 percent.
                                                                               A tentative conclusion of the analysis is that, while significant for
FIGURE 5 Weekday hourly percentage of daily traffic, Lincoln
Tunnel eastbound.                                                           certain trips, it is not clear that a peak-period congestion surcharge
                                                                            would entail such a large increase in the total trip cost. It also is worth
                                                                            considering how politically feasible a doubling of peak-period tolls
                                                                            would be. If not, smaller congestion surcharges would obviously
PANYNJ region as well as several counties in Connecticut and the            have even smaller effects on behavior. Further, if other types of costs
Hudson Valley region of New York State. (The INA model contains             were accounted for, the relative importance of the congestion sur-
several hundred origin-destination pairs, with estimates of the peak-       charge might be even more reduced. For example, what costs would
period and off-peak-period traffic volumes on each of these pairs. The      be incurred by trucking firms if they were to shift operations to empha-
model uses detailed information about the road network to calculate         size off-peak movements? Further, is the GCT even the appropriate
the likely route, travel time, and travel distance for traffic on each of   measure of shipping costs facing firms?
the origin-destination pairs.) With the use of the INA model, an esti-
mate was derived of the total peak-period travel time for trucks mak-       LOGISTICS COSTS: BEYOND GCT AS INDICATOR
ing trips between each origin-destination pair contained in the model.      OF SHIPPING COSTS
This estimate also was combined with an estimate of the total distance
traveled on each origin-destination pair. After reasonable estimates        An important point regarding the estimate of GCT is that these
were assigned for the value of time of trucks and vehicle operating cost    estimates include only factor costs involved in making the various
per mile, estimates of the peak-period GCT for all origin-destination       eastbound trips, namely, crew-time costs, vehicle operating costs,
pairs were derived. Since the INA model also accounts for tolls, the        and tolls. As such, they reflect the costs incurred by trucking firms
estimates included tolls charged at the interstate crossings.               in bringing deliveries to their clients; and given the competitive
   Table 2 outlines estimates of the GCT for selected origin-               environment in the industry, these costs, along with the appropri-
destination pairs that use the interstate crossings. For ease of pre-       ate markup for profit, will approximate the basic shipping rate the
sentation, the very precise (and numerous) origin-destination pairs         clients will pay.
have been aggregated to county level, except for the case of Man-              While the estimated GCT may be a good indication of the basic
hattan, which is separated into several zones. In these calculations, the   shipping rate faced by the producing firms that are users of freight-
INA model produced estimates of the least-cost route available given        shipping services, the shipping rate is only part of the consideration
the volume of traffic on the entire network for each of the origin-         facing producing firms in their decisions regarding shipping. Since the
destination pairs. Trips along these least-cost routes then were given      demand for shipping services is itself derived from the demand for the
a cost in terms of travel time, vehicle operating costs, and out-of-        goods or services of the manufacturing, service, or retail firms (3),

                                 FIGURE 6 Weekday hourly percentage of daily traffic, Goethals Bridge
Vilain and Wolfrom                                                                                                      Paper No. 00 -1112         69

                  TABLE 2 Average Generalized Cost of Travel for Trucks During the Morning Peak Period for Selected
                  Origin-Destination Pairs

other considerations become extremely important. For example, as             costs are further reduced by having deliveries of goods made at the
noted by Allen et al. (6), the reliability of the delivery time of ship-     start of the business day.
ments can be significantly more important than the shipping rate for             The fact that the logistics costs are a more appropriate measure of
the producing firm. The importance of reliability is due to various fac-      true shipping costs is demonstrated in studies estimating the observed
tors, including its impact on the amounts of safety inventory required       value truckers themselves place on time savings in transit. Since
or the risks of disruptions to the production process of the producing       truckers need to respect the wishes of their clients at least in part, one
firms. Reliability also is crucial in reducing the risk of perishability of   would assume that the value they would place on time savings would
goods or even affecting the producing firm’s reputation due to delay          reflect in part the value their clients place on reliability. This “sub-
in its own deliveries.                                                       jective” value of time that truckers express, either through surveys or
   The total logistics costs for shipments therefore should include          through their actual behavior, therefore should be higher than the
the shipping rate charged by the trucking firm as well as the inven-          simple factor costs (wage rates, vehicle operating costs, and tolls)
tory costs for goods while in transit and the “excess” inventory that        that comprise the GCT. As reported by De Jong in a review of sev-
is necessary to guard against unreliability in deliveries (6 ). As an        eral studies, the subjective value that truckers place on their own time
example of the last item, consider a manufacturing firm that must             usually is significantly above the GCT (7). This presumably reflects
always have several costly components in inventory, as running out           the fact that these truckers are giving a value to their time that incor-
would severely disrupt production. If deliveries were made with cer-         porates, at least partly, the more extensive logistics costs that matter
tainty at a given date and time, the inventory costs of holding these        most to their clients.
extra components could be reduced significantly. Similarly, con-                 A major finding from the interviews with trucking operations
sider the example of the supermarket that must always keep a buffer          was that logistics costs usually are far more important to producing
inventory of 378 L (100 gal) of milk to avoid running out and los-           firms in the region than mere shipping rates. This is due to the fact
ing customers. Again, deliveries made with certainty at a particular         that the potential disruption to production, or loss of retail sales,
time can allow the supermarket to hold a smaller buffer inventory,           from having a delay in delivery time has attendant costs that far out-
reducing these inventory costs. Note that in both cases, inventory           weigh the magnitude of the shipping rate. This explains a point that

                  TABLE 3     Tolls as a Proportion of Generalized Cost of Travel for Trucks During the Morning Peak Period
70     Paper No. 00 -1112                                                                               Transportation Research Record 1707

was mentioned repeatedly by trucking firms during the interviews:          (The interviews did involve an exercise in which interviewees were
customers (in other words, the producing firms that have goods            presented with a series of congestion-reduction policies and asked to
shipped) usually insist on deliveries in quite narrow time windows.       estimate their response to the strategies. The strategies proposed
The reasons for this is that morning deliveries are crucial to firms       included various pricing scenarios. However, the exercise produced
concerned with reducing inventory costs as well as maintaining a          only a limited sample of quantitative data and therefore will not be
smooth flow of production while minimizing the amount of wasted            presented in this paper.)
or idle labor. Not surprisingly, the impetus to reduce inventory costs
is encouraged further by the high cost of real estate in the region.          • As suggested by the data in Figures 4 to 6, a significant pro-
                                                                          portion of commercial traffic tries to avoid the most congested part
                                                                          of the peak period. Trucks making an off-peak crossing do so
SELECTING FIRMS TO BE INTERVIEWED                                         because they can. For example, trucks making a longer trip between
                                                                          southern states and New England can time their trip so as to avoid
The interview approach used contrasts with the typical stated-            the peak period if they use the PANYNJ facilities. Alternatively,
preference study, in which a large number of fairly short interviews      truckers making pickups or deliveries to 24-hour facilities also ben-
would be administered to drivers, presumably at the point of mak-         efit from increased flexibility in their time of crossing. Trucking
ing the interstate crossing. Here, a selected group of 50 trucking        firms are more aware than anyone of the increased cost of conges-
firms was interviewed, with the interview conducted with various par-     tion that can be expected with a peak-period crossing, and whenever
ties intimately knowledgeable about current shipping practices as well    possible they will try to avoid the peak. Various factors conspire to
as the constraints imposed by the demands of their clients.               force truckers into the peak, however. Some are examined here.
   The selected firms comprising the sample were chosen to be rep-             • The single biggest constraint in avoiding peak-period interstate
resentative of users of the interstate crossings during peak periods.     crossings are customers. This constraint applies to all segments of
An estimate of some of the characteristics of peak-period commer-         commercial traffic to varying degrees, whether for-hire or private car-
cial users of the interstate crossings was derived from a windshield      riers. Most customers require a particular level of service, includ-
survey at three of the six crossings. These three crossings—the           ing deliveries during working hours. This is particularly the case
George Washington Bridge, the Lincoln Tunnel, and the Goethals            for trucking firms making express overnight deliveries, as their main
Bridge—each serve a different segment of the population using the         reason for being is precisely to make deliveries early in the working
crossings. The George Washington Bridge serves what is commonly           day. In general, customers want their deliveries in the morning and
known as the Northern Corridor into Manhattan, the Lincoln Tun-           pickups in the afternoon. This makes sense in the context of inventory
nel serves the Midtown Corridor, and the Goethals Bridge serves the       cost minimization, in which finished goods are shipped out rapidly,
Southern Corridor into Staten Island.                                     and goods are delivered for use in production or sale at the start of the
   The truck population was broken down into for-hire carriers            working day.
[including less-than-truckload (LTL), truckload (TL), and drayage]            It should be noted that “just-in-time” inventory management is a
and private carriers. The segmentation was based on visual identifi-       significant factor cited by many of the firms interviewed, not only
cation of certain characteristics (carrier name, size of truck, origin,   with respect to manufacturing firms but also various types of retail.
commodity shipped), and was to a degree arbitrary and possibly            It is not surprising that inventory management and cost minimization
imprecise. However, with these caveats in mind, it was estimated that,    would be so prevalent in a region where real estate costs are high.
depending on the facility, between 22 and 47 percent of the trucks            There may be an increase in after-hours operations for warehouses
using the crossings are LTLs, and most of the remainder are TLs (with     and some grocery facilities, but such delivery flexibility still charac-
drayage a fairly small proportion of the volume). Differentiating         terizes a limited share of the customer base of most truckers. For
between LTL and TL trips was based on a variety of factors, includ-       example, most large grocery stores require deliveries in the morning,
ing name recognition of large carriers. Deduction also was used to dif-   typically in an 8-to-10-a.m. window, which in turn allows for shelv-
ferentiate. For example, a two-axle panel truck with local plates doing   ing by the larger daytime staff. Smaller grocery stores, including the
business as “XYZ Express Lines” would be assumed to be an LTL             ubiquitous deli, are much more flexible in delivery times. Given their
trip, whereas a five-axle truck with nonlocal plates usually would be      longer hours, these stores have staff on hand to accept deliveries at
assumed to be a TL trip.                                                  times when large stores refuse deliveries. However, these stores still
   With respect to the origin of the trucks, visual inspection of the     are relatively limited shares of the customer base of the typical trucker
vehicle license plates revealed that, depending on the facility, be-      making retail deliveries, and the typical delivery route between New
tween 73 and 91 percent of the trucks were local (in other words, from    Jersey and New York usually must include deliveries to large stores
either New Jersey or, to a lesser extent, New York). These observed       in the 8-to-10 a.m. window.
characteristics of commercial traffic during the peak period were the         There are, of course, items that are less time sensitive, and delivery
basis for the selection of interview firms. For example, a large major-    of these goods may provide more flexibility. For example, newsprint
ity of firms interviewed were based in New Jersey, and the break-          requires a period to adapt to the temperature of the printing environ-
down between for-hire and private carriers was representative of          ment. Clearly, a delivery of this commodity does not need to coincide
peak-period commercial users.                                             with the start of the business day. But a trucking operation involved
                                                                          in delivering newsprint, interviewed during the study, is perhaps typ-
                                                                          ical: it already is taking full advantage of the flexibility provided by
FINDINGS OF INTERVIEWS                                                    the commodity, making off-peak movements whenever possible.
                                                                              • Important restrictions facing truckers also included pier oper-
A summary of interview findings is outlined here, grouped by theme.        ating hours, neighborhood curfews, and union-negotiated hours of
Note that no attempt is made to attach a numerical value or statisti-     operation. The region’s maritime terminals typically operate between
cal significance to any of the responses elicited during the interviews.   6 a.m. and 5 p.m., which generally moves truckers into peak-period
Vilain and Wolfrom                                                                                                     Paper No. 00 -1112         71

crossings. For example, New Jersey truckers making pickups at the          highly competitive nature of the trucking industry, it is not surpris-
Howland Hook Marine Terminal in New York City would almost                 ing that an operating cost increase would be passed on to customers.
certainly make an eastbound interstate crossing no earlier than 5 a.m.     However, it is not clear that the rate increase would be passed on to
Crossing any earlier would simply mean that drivers would be idle,         only peak-period deliveries. The fact that an interstate crossing is
waiting at the terminal until it opens. Neighborhood curfews were          typically followed by several deliveries would mean that the increased
mentioned by some firms as posing restrictions to evening deliveries        rate would probably be uniform regardless of the time of the actual
or pickups, principally to some retail stores in residential neighbor-     delivery, reducing the effectiveness of a congestion surcharge. The
hoods. In terms of labor rules, some of the firms interviewed oper-         reason for this is simple: If a trucker were to make a peak-period
ated under union contracts, which sometime stipulate the hours of          crossing and pay a congestion surcharge, which subsequent delivery
operation. In the case of one trucking firm, a union contract stipulated    should absorb the surcharge? Would the 10 a.m. delivery be charged
hours of operation starting at 8 a.m., which would guarantee traveling     more than the noon delivery? Would this be feasible, particularly if
during some of the more congested parts of the peak.                       the timing of the particular delivery really reflected the scheduling
    • The more time sensitive the goods delivered, the less likely the     preferences of the trucker? For example, firms located in Manhattan
impact of a congestion surcharge. In general, firms with highly time-       near the interstate crossings typically receive deliveries earlier in the
sensitive deliveries will be unlikely to switch their delivery times.      working day, as they are the logical first deliveries for a truck com-
This is particularly true of express delivery services guaranteeing        ing into New York City from New Jersey. Following the imposition
overnight deliveries. As also reported in Golob and Regan, this seg-       of a congestion surcharge, should these firms then pay a higher deliv-
ment of the market may well favor some forms of congestion sur-            ery rate than those located in Queens or Brooklyn, which, because of
charges if they lead to reductions in the peak-period congestion that      their location, may be scheduled for later deliveries?
they cannot avoid (8). For this segment, the reduction in peak-period         Given the difficulties in assigning a peak-period surcharge to spe-
congestion (and presumably, variability in travel time) would be a         cific deliveries, it appears likely that trucking firms would continue
valuable commodity worth the surcharge.                                    to charge a uniform rate, albeit a higher one, to all customers. If the
    An interview with an express shipper was particularly enlighten-       delivery charge to individual firms would fail to reflect the conges-
ing. The regional operations manager of the company was quite ada-         tion surcharge at the crossing, part of the purpose of the surcharge
mant in stating that the company would be happy to pay for various         (namely, encouraging firms to opt for later or earlier deliveries)
congestion mitigation strategies (notably, hypothetical “truck only”       would not be accomplished.
lanes at various crossings). To them, the value of reduced uncertainty        • Rates currently charged by trucking firms do reflect the different
in travel times is high, particularly as their business involves quite     costs attributable to deliveries at different locations but not different
strict delivery windows. For this reason, off-peak discounts (includ-      delivery times. For example, a trucking firm based in New Jersey was
ing those currently offered by the New Jersey Turnpike) are of no use.     found to charge $200 to $225 for deliveries between Port Newark and
The service they offer simply does not allow rescheduling deliveries       Carteret, New Jersey, as opposed to $375 to $425 for a delivery to
to avoid the peak period.                                                  Maspeth, Queens. Although such rate differentiation based on loca-
    • It is not clear that there are sufficient incentives for many firms   tion is the norm, no rate differentiation based on time of delivery
to accept evening, night, or early-morning deliveries. The imperative      was used by any of the 55 firms interviewed, even though an 8 a.m.
to minimize inventory costs, particularly of perishable goods, makes       delivery is arguably more expensive in time and resources than a
early-morning deliveries at the start of the business day optimal for      2 p.m. delivery. As mentioned previously, a major reason for this is
many, if not most, businesses. Off-hour deliveries, on the other hand,     undoubtedly that trucks typically make more than one delivery
typically entail the added costs of extra workers, possibly at higher      along a route. Given this fact, it becomes difficult to pinpoint which
hourly rates. Would any congestion surcharge ever be enough to             delivery should be charged more, particularly as the actual delivery
motivate this type of reorganization? For example, it is interesting       time may be due to the trucker optimizing the efficiency of his or her
to note that there are essentially no off-hour deliveries made to firms     delivery schedule rather than customer preferences for a particular
in the garment district in midtown Manhattan, even though these            delivery window.
firms are located in a very congested part of New York City. In the           • The lack of alternatives to the interstate crossing facilities implies
words of one manager at a trucking firm specializing in deliveries to       that peak-period surcharges would not lead to significant deviation
the garment sector, “Nighttime deliveries will never happen there.”        toward other routes. Trucking operations are keenly aware of tolls
Although one can argue with so categorical a conclusion, it is still       and, where practical, will try to avoid them to minimize their costs.
clear that garment firms do not find it in their interest to reorient      If they can select a route that avoids tolls while not adding on more
operation in order to accommodate off-hour deliveries.                     costs in terms of time and operating costs, they will choose it. How-
    • Even if most customers were willing to accept deliveries after       ever, as can be seen from Figures 2 and 3, there are relatively few
10 a.m., it is not clear that truckers would choose to postpone their      options for avoiding the tolls at the interstate crossing facilities, with
time of crossing. Truckers making an interstate crossing may have a        the Tappan Zee Bridge the main alternative eastbound crossing. This
route that will take them to various customers, ending the working         implies that although trucking operations do look to avoid toll charges
day by 4 or 5 p.m. In order to optimize the revenue from the route, it     if it is efficient to do so, it is unlikely that limited peak-period toll
makes little sense to wait for the peak period to end before making        surcharges, for example, would lead to significant traffic deviation.
the crossing. In the words of an operations manager for a firm mak-            • In general, the interviews suggested that the response to conges-
ing retail deliveries, “Nobody leaves after 6 a.m.—it’s just not worth     tion surcharges would be relatively modest. When faced with hypo-
it to them.” Similarly, trucks leaving at 3 or 4 a.m., while avoiding      thetical peak-period congestion surcharges of up to 100 percent, most
significant congestion, would find themselves in New York City with          operations managers indicated their response would be fairly mini-
no one open to accept deliveries.                                          mal. Notable exceptions included trucking firms that deliver fairly
    • Many (but not all) of the trucking firms indicated that a conges-     time-insensitive items to wholesalers and smaller grocery stores. One
tion surcharge would be passed on as higher delivery rates. Given the      such firm indicated that a 50 percent surcharge (which would raise the
72     Paper No. 00 -1112                                                                                 Transportation Research Record 1707

toll to $6 per axle) would lead to a shift of 30 percent of peak-period       A tentative conclusion of the analysis is that “realistic” value-
crossings to the off-peak, while a 100 percent surcharge would lead        pricing scenarios may result in only modest changes in behavior by
to a 50 percent shift. This response was the most extreme of all the       commercial traffic. The costs of shifting to off-peak are apparently
firms interviewed.                                                          too high in most cases. Further, it is not clear that a congestion sur-
                                                                           charge would not simply be passed on to customers as a uniform rate
                                                                           increase, regardless of the delivery time. In this case, one of the
CONCLUSIONS                                                                implicit objectives of a congestion surcharge, making the customer
                                                                           assume the marginal social cost of the peak-period delivery, would
An analysis of available data reveals that commercial traffic using        not be achieved.
the PANYNJ interstate crossings does try to spread its peak period            The present research suggests a relatively inelastic peak-period
to a degree. However, even despite these efforts, the hours of heavi-      demand for commercial users of the interstate crossings. Although
est use of the crossings are similar to those of other vehicle types,      this implies modest changes in peak-period volumes under the sug-
meaning that commercial traffic is a major contributor to peak-period      gested levels of value pricing, this does not necessarily argue against
congestion.                                                                such policies. The general impetus for value pricing is to try to ensure
   Would commercial traffic respond to value pricing? An analysis          that users pay a charge that approximates the marginal cost of their
of the generalized cost of travel facing trucks using the crossings        trip. Whether traffic volumes are significantly altered in the pro-
revealed that current tolls account for between 10 and 29 percent of       cess is not necessarily the only yardstick for imposing this type of
the GCT, depending on the origins and destinations. Assuming a             pricing.
doubling of the tolls were politically feasible, this would increase
the total GCT by 10 to 29 percent. Given the highly competitive
nature of the trucking industry, it is assumed that these costs would      REFERENCES
be mostly passed on to producing firms. However, the factor costs
that make up the GCT are only one aspect of the logistics costs fac-       1. Vickrey, W. Congestion Theory and Transport Investment. American
ing producing firms, for which the costs of switching to off-peak              Economic Review Papers and Proceedings, Vol. 59, No. 2, 1969,
deliveries may far outweigh the higher peak-period tolls.                     pp. 251–260.
   Therefore, it is not surprising that the single most important con-     2. Keeler, T., and K. Small. Optimal Peak-Load Pricing, Investment and
                                                                              Service Levels on Urban Expressways. Journal of Political Economy,
straint facing truckers is delivery windows imposed by customers.             Vol. 85, 1977, pp. 1–25.
These delivery windows tend to force trucking operations toward            3. Hau, T. Economic Fundamentals of Congestion Pricing: A Diagrammatic
peak-period interstate crossings. Delivery windows are most bind-             Analysis. Policy Research Working Papers WPS-1070. The World Bank,
ing in the case of express overnight delivery; however, some trucking         Washington, D.C., 1992.
                                                                           4. Kain, J. Impacts of Congestion Pricing on Transit and Carpool Demand and
firms reported that extremely rigid delivery times are sometimes
                                                                              Supply. In Special Report 242: Curbing Gridlock: Peak-Period Fees to
imposed by some manufacturers and retailers as well. Other factors            Reduce Traffic Congestion, TRB, National Research Council, Washington,
that constrain the ability of trucking operations to shift to off-peak        D.C., 1994, pp. 502–553.
deliveries include the operating hours of piers, curfews, and, in some     5. Allen, W. B. The Demand for Freight Transportation: A Micro Approach.
instances, union regulations.                                                 Transportation Research, Vol. 2, 1977, pp. 9–14.
                                                                           6. Allen, W. B., M. M. Mahmoud, and D. McNeil. The Importance of Time
   Another constraint in avoiding peak-period interstate crossings is         in Transit and Reliability of Transit Time for Shippers, Receivers and
due to regional geography. Typically, a trucking firm making an east-          Carriers. Transportation Research B, Vol. 198, No. 5, 1985, pp. 447–456.
bound crossing will do so early enough to be able to schedule a prof-      7. De Jong, G. Freight and Coach Value of Time Studies. Presented at Pub-
itable number of pickups and deliveries in New York City and Long             lic Transportation Resources Council Seminar on the Value of Time,
Island. This will always force the crossing into the peak period, as          London, 1996.
                                                                           8. Golob, T. F., and A. C. Regan. Freight Industry Attitudes Towards Poli-
waiting until 11 a.m. to cross simply will mean postponing the begin-         cies to Reduce Congestion. Presented at 78th Annual Meeting of the
ning of deliveries. Given the constraints imposed by customers, this          Transportation Research Board, Washington, D.C., 1999.
really translates into losing half the working day. Likewise, crossings
made before 5 a.m. translate into drivers being idle, even if deliveries   Publication of this paper sponsored by Committee on Urban Freight Trans-
begin in Manhattan as early as 7 a.m.                                      portation and Committee on Taxation and Finance.

To top