Benefit-Cost Analysis of Variable Pricing Projects: QuickRide HOT Lanes Mark Burris, P.E., M.ASCE1; and Edward Sullivan, P.E., M.ASCE2 Abstract: Researchers identified a potential methodology for obtaining the incremental societal costs and benefits from a variable pricing project and applied that methodology to the QuickRide high occupancy/toll 共HOT兲 lanes in Texas. This is one of the longest running variable pricing projects in the United States and, as such, it provided useful historical data and trends upon which to estimate future benefits and costs. This analysis found that the incremental societal benefits of QuickRide exceeded incremental societal costs for the time period considered. A companion paper that used the same methodology to examine the benefits and costs of the SR-91 Express Lanes found similar results. However, the differences between the benefits and costs were dramatically different for the two projects, indicative of the relative size of the two projects and the number of travelers impacted. On SR-91, tens of thousands of travelers were impacted on a daily basis where QuickRide’s impact was limited to approximately 400 travelers per day. Interestingly, the benefit-cost ratios of the two projects were similar, both between 1.5 and 1.7. DOI: 10.1061/共ASCE兲0733-947X共2006兲132:3共183兲 CE Database subject headings: Benefit cost ratios; Pricing; Tolls; High occupancy vehicles; Texas; Traffic management.
Introduction Facing increasing difficulties in funding transportation infrastructure improvements to handle ever increasing traffic congestion 共Schrank and Lomax 2002兲, agencies in charge of highway transportation are increasingly considering the application of value pricing to their roadways. Value pricing includes 共FHwA 2002兲: 1. Managing demand through tolls that vary by time of day, day of week, or season; 2. Using flat or variable tolls to efficiently preserve congestionfree use of both new and previously existing spare capacity, such as in high occupancy/toll 共HOT兲 lanes; 3. Cordon and area pricing with flat or variable tolls; 4. Priced queue-jumps and other congestion-bypass facilities; 5. Converting traditionally fixed costs 共insurance, registration, etc.兲 to variable costs; 6. Cash out of subsidized employee parking; and 7. Car sharing. The first four methods listed above have the potential to use tolls to both help finance infrastructure improvements and manage congestion—dealing with two of the critical dilemmas facing transportation agencies. Some pricing methods increase options available to travelers in the form of both premium and discount 1 Assistant Professor, Dept. of Civil Engineering, Texas A&M Univ., 3136 TAMU, College Station, TX 77843. E-mail:
[email protected] 2 Professor, Dept. of Civil and Environmental Engineering, California Polytechnic State Univ., San Luis Obispo, CA 93407. E-mail:
[email protected] Note. Discussion open until August 1, 2006. Separate discussions must be submitted for individual papers. To extend the closing date by one month, a written request must be filed with the ASCE Managing Editor. The manuscript for this paper was submitted for review and possible publication on November 12, 2004; approved on May 4, 2005. This paper is part of the Journal of Transportation Engineering, Vol. 132, No. 3, March 1, 2006. ©ASCE, ISSN 0733-947X/2006/3-183–190/ $25.00.
services such as access to a faster lane 共methods 2 and 4兲, and discounts for adjusting time of travel 共methods 1 and 3兲. Variable pricing is a feature of many value pricing projects, where tolls vary by time of day. These pricing strategies reduce traffic demand for the most congested times and locations, improving traffic flow. In theory, the use of variable pricing in the form of marginal cost pricing has been shown to maximize net societal benefits 共Walters 1968; Hau 1992; Small and Gomez-Ibanez 1997; Mohring 1999兲. These benefits occur because the total cost of travel 共marginal social cost兲 exceeds the average private cost paid by drivers during congested periods 共see Fig. 1兲. This disparity leads to an inefficiently high demand 共q1 in Fig. 1兲 and a welfare loss 共the hatched area of Fig. 1兲. The optimal traffic flow 共q2兲
Fig. 1. Optimal pricing of transportation
JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MARCH 2006 / 183
occurs at the point where the demand for travel 共Curve D-D in Fig. 1兲, which measures the marginal benefit to users of each additional trip, intersects the marginal social cost curve. At the optimum flow, q2, any increase or decrease in traffic would decrease the net societal benefits derived from the use of the road. In theory, the optimal flow of traffic can be achieved if drivers are charged a toll of . Planners, engineers, and economists have developed theoretical models to predict the impact of congestion pricing 共both in its pure marginal social cost pricing form and using variations of marginal social cost pricing兲. Applying these models to actual traffic conditions, researchers predicted the resulting societal costs and benefits 共Hyman and Mayhew 2002; May et al. 2002; Santos and Rojey 2003兲. In almost every case, societal benefits exceeded societal costs—although the impacts were often regressive to neutral with respect to the relative benefits enjoyed by different income groups. To improve the equity of the pricing program some writers examined various toll revenue allocation schemes 共Small 1992; Litman 1999; Parry and Bento 1999兲. Despite these findings, many transportation agencies remain reluctant to proceed with variable pricing projects. Part of this reluctance may stem from the lack of actual empirical data on the benefits and costs of operational variable pricing projects. This lack of actual data is not surprising since only 19 variable pricing projects have been implemented 共Burris and Pendyala 2002兲, and many of those have been implemented too recently to provide the data required for a benefit-cost analysis 共BCA兲. This research was undertaken to provide an empirical reference point by performing a BCA on one of the US’s longest running projects: the QuickRide Program 共HOT lanes兲, in Houston 共see Fig. 2兲. 共A companion paper used the same methodology to perform a BCA for the SR-91 Express Lanes.兲 In analyzing this project, researchers used 4 years of historical data to derive historical costs and benefits and to project future costs and benefits. Based on these data, net societal benefits exceeded costs for the period considered, although the difference was small. The next section of the paper includes an overview of the BCA methods used. In the succeeding section the benefits and costs of the variable pricing project are documented. The last section of the paper includes a summary and conclusion.
Framework for the Benefit Cost Analysis In this research, the net incremental societal benefits and costs of a variable pricing project were quantified. This included the incremental benefits and costs resulting from the implementation of the variable pricing project, but did not include any transfers of wealth among different groups. In such an analyses it is critical to carefully define the attributes of the “base case” scenario. This scenario did not exist, but rather it is what would have likely occurred if the variable pricing project had not been undertaken. As such, it provides the reference point against which the incremental societal costs and benefits of the variable pricing project are measured. Our base case assumed that the HOT lanes in Houston continued to operate as HOV lanes, without ever incorporating the option of toll paying travelers 共see the following section for details specific to the individual project兲. The variable pricing project examined here, along with most pricing projects around the world, rely on electronic toll collection 共ETC兲 for vehicle identification and toll payments. ETC equipment has a limited life span, and therefore benefit-cost
analyses of ETC projects 共along with other intelligent transportation system projects兲 typically examine benefits and costs for a 10-year time frame 共USDOT 2002兲. This research also used a 10-year time frame for cost and benefit streams. Although this may be seen as conservative, it was appropriate since one of the HOT lanes was slated to undergo major reconstruction and expansion after this time period. All benefits and costs were converted to year 2002 dollars using discount and inflation rates published by the federal government 共real⫽3.1%, nominal⫽5.1%, inflation⫽1.9%兲 共OMB 2002兲. Benefits In general, the benefits travelers derive from making trips is measured by the area formed between the demand curve 共D-D in Fig. 1兲 and the horizontal line indicating the cost of travel at the given demand 共line p1 in Fig. 1兲. If a change in transportation infrastructure or travel options causes the area between these two curves to increase then the change results in a net benefit for those travelers. As discussed previously, society as a whole benefits to the maximum extent possible when the cost of the trip equals the marginal social cost—eliminating the welfare loss caused primarily by the additional time penalty each new vehicle imposes on existing drivers under congested conditions. Unfortunately, the shapes of these curves are generally unknown and calculating a change in benefits often amounts to estimating the change in travel costs for travelers and aggregating these changes. It is common to estimate user cost changes due to changes in travel time savings, vehicle operation and ownership costs, and safety improvements 共USDOT 2000; AASHTO 2003兲. However, cost savings associated with safety were not examined here because the accident rates for the travel options were similar and no benefits were assumed from a safety aspect. One additional item examined here was costs associated with emissions, due to the increased interest in such impacts. Travel Time Savings The benefits of this variable pricing project, and the majority of transportation improvements, are dominated by travel time savings. Although it is relatively straightforward to estimate the amount of travel time saved due to a transportation improvement, it is far more difficult to determine the value of that travel time saved. A great deal of research has focused on determining appropriate values of time 共VOT兲 共Calfee and Winston 1998; Small et al. 1999; Bates et al. 2001; Hensher 2001兲. Much of this research has found that the values of time for commute trips range from 20% to over 50% of the drivers’ wage rates 共USDOT 2003兲. These values have been shown to increase with increased traffic congestion 共Small et al. 1999兲, and there was considerable traffic congestion in the free lanes beside the variable pricing project examined. The VOT was further refined in this research based on the premise that, other things being equal, people who chose to pay tolls to avoid congestion must have had values of time at least equal to the toll divided by the amount of time saved. Similarly, people who chose not to pay tolls to avoid congestion would usually be expected to have had values of time less than the toll divided by the time that could have been saved. However, this simple premise was complicated by eligibility issues, such as the requirement to have a transponder in order to use these toll facilities.
184 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MARCH 2006
Table 1. Unit Values of Emission Reductions Delucchi
Pollutant
Low 共$/kg兲
a
High 共$/kg兲
Small and Kazimi and Levinson b 共$/kg兲
1.59 23.34 NOx VOC 0.13 1.45 CO 0.01 0.10 PM10 a Delucchi 共2000兲. b Values obtained by Small and Kazimi 共2002兲.
Values used in this analysis 共$/kg兲
1.33 1.71 0.0063
1.59 1.45 0.01 12.17
共1995兲 as modified by Levinson
the default value used in the Federal Highway Administration 共FHwA兲 STEAM model 共FHwA 2003兲. Costs were inflated according to an average health care index 共+3.5% per year, USHCFA 2001兲 to calculate values for the entire evaluation period.
Fig. 2. Houston’s HOT lanes 共figure created by John Hobbs, used with permission兲
Vehicle Operating and Ownership Costs Another potential benefit derived from roadway improvements is the reduction in vehicle operating and ownership costs. In this application, since the distance traveled for each option was identical, it was assumed that there would be no significant difference in ownership costs and the bulk of operational cost savings would be derived from any reduction in fuel consumption. Based on the changes in speeds caused by the variable pricing project, the changes in fuel consumption were estimated using Fig. 3 共West et al. 1999兲. The total change in fuel consumption was converted to a dollar value based on the price of fuel at the pump 共approximately $1.50 in 2000兲. Taxes were subtracted from the pump price because taxes are transfers of wealth and should not be included in calculating incremental societal costs and benefits. The net benefit from reducing fuel consumption was valued at approximately $1.12 per gallon 共year 2000 dollars兲. Emissions The final benefit category considered was the change in vehicle emissions. The changes in emissions were found using Mobile 5b in Houston and the California EMFAC model for SR-91 in the companion paper 共CARB 1996; Barth et al. 2001; CARB 2003兲. Mobile is the mobile source emissions estimation program used in the majority of the United States while EMFAC is specifically calibrated for the California vehicle fleet. Once the changes in the quantities of pollutants 关nitrous oxides 共NOx兲, volatile organic compounds 共VOC兲, carbon monoxide 共CO兲, and 共in California兲 particulate matter 共PM10兲兴 were determined, the values of these changes were estimated using default values shown in Table 1. To convert emissions to monetary values, the costs associated with most pollutants were obtained from research done by Delucchi 共2000兲 and Small and Kazimi 共1995兲 共see Table 1兲. Both of these sources based their costs of emissions on the cost of health care to treat diseases related to the emissions of motor vehicles. The values used in this analysis were on the conservative end of the estimates. In the case of PM10 emissions 共in California兲, the unit cost of $12.17/kg was adopted, which is
Costs Incremental costs were more straightforward to estimate. All costs 共excluding transfers兲 that the agency incurred due to the project were included. Start up costs were amortized over the 10 years evaluation period and added to any operation and maintenance costs of the project over and above costs that would have occurred under the base case scenario.
Benefit-Cost Analysis of Houston Quickride For the analysis period, the Houston QuickRide project consisted of high occupancy/toll 共HOT兲 lanes on the Katy Freeway 共I-10兲 and Northwest Freeway 共US 290兲 in Houston 共see Fig. 2兲. Each highway had one barrier-separated reversible high occupancy vehicle 共HOV兲 lane that was open on weekdays 5–11 a.m. inbound and 2–8 p.m. outbound. Normally open to HOV 2+ vehicles, the lanes were restricted to HOV 3+ vehicles during 6:45–8:00 a.m. on both freeways and 5:00–6:00 p.m. on Katy Freeway. However, HOV 2+ drivers who registered for the QuickRide program could use the lanes during these hours if they paid a $2 toll per trip. In this way the toll paid by HOV2 travelers on the HOV lane varied by time of day, with the peak period costing $2 and the off-peak having no toll. The HOV lane on the Katy Freeway was constructed in 1984. The lane was originally restricted to buses and vanpools but the lane operator quickly relaxed these restrictions to accept any HOV 2+ vehicle. By 1988, the need to alleviate congestion led to the additional occupancy restrictions during the morning peak hour. The afternoon peak period followed shortly after. The same restrictions were later placed on the HOV lane on the Northwest Freeway during the morning peak period. This was how the HOV lanes operated for several years prior to QuickRide beginning in mid-January 1998 on the Katy Freeway and November 2000 on the Northwest Freeway. Prior to QuickRide, the increased occupancy restriction eliminated HOV lane congestion, but resulted in significant excess capacity 共Burris and Hannay 2003兲. This was chosen as the base case against which the costs and benefits of QuickRide were compared. In this base case, which assumed no QuickRide, HOV2 vehicles could not legally use the HOV lane during the HOV 3+ restricted periods. The analysis covered the time period
JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MARCH 2006 / 185
Table 2. Travel Alternatives to QuickRide
Travel alternative
Table 3. 2001 Travel Time Savings Estimated percentage of trips that used this travel method prior to QuickRidea
Travel by bus on the HOV or GPLs 8 HOV2 during off-peak on HOV lane 7 HOV2 during peak on GPLs 26 HOV3+ on HOV lane 2 SOV on GPLs 54 Other 3 a These estimates were obtained from a 1998 survey of QuickRide enrollees.
from mid-January 1998 to mid-January 2008, corresponding to 10 years of QuickRide usage on the Katy Freeway and just over 7 years on the Northwest Freeway. QuickRide Benefits The number of QuickRide trips was determined using a database that contained all recorded QuickRide trips. Future QuickRide usage was estimated to be the average of volumes for the years 2002 and 2003. This was likely a conservative estimate as it assumed a 0% growth in the number of QuickRide trips. For each of these trips it was assumed that the benefits derived from using QuickRide exceeded the benefits derived from the travelers’ alternative methods of travel. The most probable travel alternatives are listed in Table 2. Note that the data in Table 2 can be considered only rough estimates since they were obtained in 1998 and did not include US 290 travelers. Each of these travel options provided a reduced utility for travelers with respect to using QuickRide. It is extremely difficult to both 共1兲 quantify the number of current QuickRide trips that would fall into each of the categories in Table 2 and 共2兲 to quantify the additional benefits of a QuickRide trip versus each alternative. Therefore, in this analysis, the benefits of QuickRide were derived by comparing a QuickRide trip to a single occupancy vehicle 共SOV兲 trip at the same time on the general purpose lanes 共GPL兲. Thus the additional benefits derived from using QuickRide compared to SOV travel were assumed to be similar to the benefits of QuickRide compared to other travel options. For example, this assumes benefits derived by HOV2 travelers who
兺 segments 兺
冤
Freeway
Time
Katy a.m.
6:45–7:00 7:00–7:15 7:15–7:30 7:30–7:45 7:45–8:00 Weighted average 共a.m.兲
11.1 19.5 23.6 23.5 10.2
29.76 27.25 24.48 23.37 24.79 25.50
53.98 59.81 60.21 60.11 59.48 59.22
11.6 15.4 18.6 20.1 18.1 17.3
Katy p.m.
5:00–5:15 5:15–5:30 5:30–5:45 5:45–6:00 Weighted average 共p.m.兲
7.0 14.2 12.2 6.7
28.35 26.13 26.97 28.61 27.19
57.19 58.34 57.63 58.70 57.98
13.7 16.2 15.2 13.8 15.0
Northwest a.m. 6:45–7:00 7:00–7:15 7:15–7:30 7:30–7:45 7:45–8:00 Weighted average 共a.m.兲
2.8 8.0 14.0 16.2 7.2
34.36 31.89 28.72 27.44 30.09 29.35
53.01 57.91 58.85 59.52 59.82 58.72
6.3 8.6 10.9 12.0 10.1 10.5
used to travel during the off-peak period but could now drive at their preferred time of travel in the HOT lanes using QuickRide were similar to the benefits derived by an SOV peak-period traveler switching to HOV2 and using the HOT lane as a QuickRide traveler. Value of Travel Time Savings To determine the travel time savings gained by using the HOT lanes, vehicle travel times on both the GPLs and the HOT lanes were analyzed. The dataset used included several million speed entries for individual vehicles based on their actual travel times between several automatic vehicle identification 共AVI兲 readers placed along the HOT and general purpose lanes 共GPLs兲 of the Houston freeways 共data available at http://traffic.tamu.edu兲. Using these data, the average travel speeds on both the HOT and GP lanes were found for each 15-min period, on each highway, for each year, using Eq. 共1兲
兺 observations users
兺 n=1
work days
average speed =
Eq. 共1兲 was used to determine the average space mean speed 共Banks 2002兲 weighted by the number of observations for a given segment and the length of that segment. To determine the travel time savings, these average speeds plus the average distance traveled on the HOT lanes were required. Survey data from QuickRide users, taken in 1998, was used to find that the average
GPL HOT lane average average Time Vehicles speed speed savings per day 共mph兲 共mph兲 共min/veh兲
1 ⫻ # observations speedn
兺
冥
⫻ lengthsegment
lengths
segments
# work days in year
共1兲
distance traveled on the Katy HOT lane by QuickRide users was 11 mi 共2.3 mi less than the length of the facility兲. For the Northwest Freeway, data on entry and exit of all vehicles on the HOT lane was used since there were no data on QuickRide users alone. Assuming QuickRide trips were of similar length as all HOT lane trips, the average distance traveled was 10.6 mi 共3.1 mi
186 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MARCH 2006
Fig. 3. Fuel consumed versus travel speed 共data from West et al. 1999兲
less than the length of the facility兲. Using these data the average travel time savings of travel in the HOT lanes versus the GPLs was determined. Next, the average carpool formation time of 4.33 min 共from the survey of drivers兲 was subtracted to account for the time lost in forming carpools. These average travel time savings for each year 共see Table 3 for representative values for a single year兲 were entered into Eq. 共2兲 to determine the total value of travel time savings. 10
VTTS =
兺 y=1
冋
9
VoTy ⫻
共TSyi ⫻ Vehyi兲 ⫻ QRdaysy 兺 i=1
册
共2兲
where VTTS⫽value of travel time savings 共$兲 for all QuickRide users 共in 2002 dollars兲; 1–10; these indicate the 10 years of QuickRide usage examined 共the first three on US 290 had 0 VTTS兲; VoTy⫽value of time in year y⫽$15.56 per person-hour, $31.13 per vehicle-hour in 2002 and adjusted by the discount rate 共3.1%兲 for other years. This was based on 35% of the wage rate of QuickRide users 共from the survey兲 and $0 for children
共who comprise approximately 21% of QuickRide carpool passengers兲; i = 1 – 9; these indicate the nine, 15-min time segments 共as shown in Table 3兲 for the Katy Freeway or i⫽1–5 for the five, 15-min time segments on the Northwest Freeway; TS⫽travel time saved by traveling in the HOT lane compared to the GPLs 共h兲; Veh⫽average number of vehicles per day that used QuickRide 共vehicles/day兲; and QRdays⫽number of QuickRide days per year, typically just over 250 per year. The total value of travel time savings over the entire period 共converted to year 2002 dollars兲 was $2.36 million. Fuel Usage The fuel saved by QuickRide users versus GPL users was estimated using travel speeds from the GPLs and HOT lanes. The localized travel speeds were converted to fuel use 共gal/mil兲 using Fig. 3. The results for the HOT lanes were likely to be more accurate than for the GPLs because vehicle speeds on the HOT lanes contained much less variability 共see Fig. 4兲 and therefore
Fig. 4. Speed distributions on the HOT and GP lanes JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MARCH 2006 / 187
Table 4. 2003 Fuel Consumption 共FC兲 Savings on Katy Freeway
Table 5. Weighted Average Emission Savings 共2002兲
FCHOT Daily Vehicles FCGPL 共per day兲 共gal/veh mi兲 共gal/veh mi兲 fuel savings 共gal兲
Time 6:45–7:00 a.m. 7:00–7:15 a.m. 7:15–7:30 a.m. 7:30–7:45 a.m. 7:45–8:00 a.m. 5:00–5:15 p.m. 5:15–5:30 p.m. 5:30–5:45 p.m. 5:45–6:00 p.m. Total
8.59 0.024 0.022 15.68 0.025 0.021 21.22 0.026 0.020 22.71 0.027 0.021 15.81 0.027 0.021 9.28 0.032 0.021 17.43 0.037 0.021 15.97 0.037 0.021 11.35 0.034 0.020 138.04 ⫻254 days⫻ $1.13/ gal=
0.3 0.7 1.4 1.6 1.1 1.1 3.1 2.8 1.7 13.8 $3,980
Katy a.m. Katy p.m. US 290 a.m.
10
兺 y=1
再
9
CoFy ⫻
⫻ QRdaysy
冎
VOC 共g/veh兲
CO 共g/veh兲
NOx 共g/veh兲
5.3 6.1 3.6
35.2 67.0 7.9
−6.8 −5.0 −6.9
2.
the AVI readings of travel times between two AVI locations 共1–5 mi apart depending on location兲 more accurately represented the actual vehicle speeds. Conversely, vehicle speeds on the GPLs during these times were highly variable. In this case, an average speed of 15mph over a freeway segment was likely indicative of stop and go traffic where speeds actually varied from near 0 to 30 mph. Since fuel usage increases rapidly at speeds below 25 mph the true fuel usage on the GPLs was likely greater than what was found using average speeds. A more accurate estimate would use second-bysecond speed profiles, but such data were unavailable. Therefore a conservative estimate of fuel savings was derived using Eq. 共3兲 FS =
Time period
关共FCML − FCHOT兲yi ⫻ Vehyi ⫻ D兴 兺 i=1 共3兲
where FS⫽value of fuel savings 共$兲 for all QuickRide users 共in 2002 dollars兲; y⫽1–10; these indicate the 10 years of QuickRide usage examined; CoFy⫽cost of fuel in year y⫽$1.116/gallon in 2000 共$1.50/gallon at the pump $0.384/gallon in taxes兲 and adjusted by the discount rate of other years; i⫽1–9; these indicate the nine, 15-min time segments 共as shown in Table 4兲 for the Katy Freeway or i⫽1–5 for five, 15-min time segments for the Northwest Freeway; FC⫽fuel consumed on the GPLs or the HOT lane 共gallons per mile兲; Veh⫽average number of vehicles per day that used QuickRide 共vehicles/day兲; D⫽average distance traveled per vehicle, 11 mi on Katy, 10.6 mi on US 290; and QRdays⫽number of QuickRide days per year, typically just over 250 per year. These calculations resulted in a total fuel savings of $33,700. An example of the results from 1 year is shown in Table 4. Emissions The final benefit examined was the change in emissions from vehicles using QuickRide. Once again, the vehicle speeds were used to calculate the changes in emissions from a vehicle traveling on the GPLs versus one traveling on the HOT lane. Note that this method should have also provided a conservative estimate, for two reasons: 1. As with fuel consumption, vehicle emissions increase significantly in stop and go traffic. The fact that speed data contained only average speeds over 1–5 mi segments reduced the ability to accurately account for stop and go conditions on the GPLs; and
If QuickRide did not exist, over half of the QuickRide travelers would have chosen to travel in SOVs. Therefore each QuickRide vehicle corresponded to more than one vehicle removed from the GPLs. However, we conservatively assumed a one-to-one vehicle replacement in this calculation. Calculation of the changes in emissions was done in much the same manner as fuel consumption, as shown in Eq. 共4兲 and Table 5 3
PR =
10
兺 兺 p=1 y=1
再
9
CoPpy ⫻
⫻ QRdaysy
冎
关共PEML − PEHOT兲yi ⫻ Vehyi ⫻ D兴 兺 i=1 共4兲
where PR⫽value of pollution reduction 共in 2002 dollars兲 due to all QuickRide users; p⫽1–3; these indicate the three pollutant species 共VOC, NOx, SO2兲 examined; y⫽1–10; these indicate the 10 years of QuickRide usage examined; CoPy⫽cost of pollutant 共see Table 1兲 共$/kg兲 in year y 共adjusted for other years using the discount rate of 3.1%兲; i⫽1–9; these indicate the nine, 15-min time segments 共as shown in Table 4兲 for the Katy Freeway or i⫽1–5 for five, 15-min time segments for the Northwest Freeway; PE⫽pollutant emitted on the GPLs or the HOT lane as a function of speed 共based on Mobile 5a兲 共kg per mile兲. Veh⫽average number of vehicles per day that used QuickRide 共vehicles/day兲; D⫽average distance traveled per vehicle, 11 mi on Katy, 10.6 mi on US 290; and QRdays⫽number of QuickRide days per year, typically just over 250 per year. An example of this is shown in Table 6. The total benefit from reduced vehicle emissions on both freeways was found to be −$1,100. The shift from the GPLs to the HOT lane reduced the overall volume of pollutants. However, it was the pollutant with Table 6. Values of Emissions Savings: Katy Freeway Total emissions savings
Year 1998 1999 2000 2001 2002 2003 2004a 2005a 2006a 2007a 2008a Total a Projected.
188 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MARCH 2006
QR 共days兲
VOC 共$兲
CO 共$兲
NOx 共$兲
Total 共$兲
238 253 254 252 253 254 253 253 253 253 15
164 192 181 224 264 411 389 395 400 406 24
2 0 0 5 15 26 24 25 25 26 2
−316 −416 −387 −405 −321 −367 −353 −358 −363 −369 −22
−150 −224 −205 −175 −42 70 60 61 62 63 4 −475
Table 7. Summary of Costs and Benefits for QuickRide 共2002 Dollars兲 Benefit category
Value
Travel time savings Fuel savings Emissions savings
$2,360,000 $33,700 $−1,100
Costs Start up costs Operations and maintenance
$477,900 $1,010,700
Benefit cost ratio
1.61
the smallest health impact, and therefore lowest economic impact, CO, that was reduced the most. The most costly pollutant, NOx, increased due to the HOT lane vehicles traveling at speeds near 60 mph. This caused the negative total economic value of the change in emissions. As stated earlier, this was a conservative estimate of the reduction in emissions. Due to this fact, the actual result may have been closer to neutral.
Acknowledgments
QuickRide Costs The costs of the QuickRide project included agency start up costs plus annual operation and maintenance costs. The $2 toll and $2.50 monthly enrollment fee were not included as they were simply transfers from the driver to the QuickRide agency. Additionally, the travel speed and level of service on the HOT lane were not degraded so previous users of the HOT lane did not experience additional travel costs due to the presence of QuickRide participants. Conversely, the limited number of QuickRide enrollees leaving the GPLs did not significantly impact travel on those lanes and therefore no benefits were calculated for GPL users. The Texas Department of Transportation and Houston METRO supplied cost figures for QuickRide from 1997 to 2002. These cost figures included all costs incurred due to the QuickRide program over and above the HOV lane operations 共the base case兲. Based on these figures, researchers attempted to separate start up costs and annual costs and arrived at the following cost estimates: Average monthly operation and maintenance costs: Katy start up costs: US 290 start up costs: Total costs:
thousands of travelers were impacted on a daily basis where QuickRide’s impact was limited to approximately 400 travelers per day. Interestingly, the benefit-cost ratios of the two projects were similar, both between 1.5 and 1.7. Although these ratios were similar for these two projects, other variable pricing projects are likely to yield different results. The analysis also found that the majority of benefits were derived from travel time savings. The VOT was the key factor in determining total benefits. Although a great amount of research has been done to estimate VOT, the bulk of it has been through stated preference surveys, often analyzing different mode choices. More research is needed on VOT derived from operational toll projects, particularly ones where the toll varies by time of day. This will allow researchers to better quantify both the VTTS and the benefits derived by travelers who gain the option of traveling congestion-free at their preferred time of day. This data may be derived from existing variable pricing projects and, in turn, help quantify the benefits of proposed future projects.
$8,263 共in 2002 dollars兲 $362,389 共in 1997 dollars兲 $52,482 共in 2000 dollars兲 $1,488,600 共in 2002 dollars兲
QuickRide Summary The total benefits of QuickRide over the 10 year period exceeded costs by 61% 共see Table 7兲.
Conclusions This analysis of QuickRide, one of the earliest operational variable pricing projects in the U.S., found that the incremental societal benefits exceeded costs. Similar results were found in a companion paper that investigated the corresponding benefits and costs for the SR-91 Express Lanes. The magnitudes of the differences in benefits and costs were dramatically different for the two projects, indicative of the relative size of the two projects and the number of travelers impacted. On SR-91, tens of
The writers would like to thank the FHwA’s Value Pricing Program for support of the projects analyzed in this research. Additionally, Houston METRO and TxDOT partnered with the FHwA in supporting these projects and the efforts made by all of those agencies are gratefully acknowledged. Additional thanks to Doug Lee and two anonymous reviewers for examining an earlier version of this paper. Any errors and omissions are entirely the fault of the writers.
References American Association of State Highway and Transportation Officials 共AASHTO兲. 共2003兲. “User benefit analysis for highways.” NCHRP Project 02-23, AASHTO, Washington, D.C. Banks, J. H. 共2002兲. Introduction to transportation engineering, 2nd Ed., McGraw-Hill, New York. Barth, M., Malcolm, C., Younglove, T., and Hill, N. 共2001兲. “Recent validation efforts for a comprehensive modal emissions model.” Transportation Research Record 1750, Transportation Research Board, Washington, D.C., 13–23. Bates, J., Polak, J., Jones, P., and Cook, A. 共2001兲. “The valuation of reliability for personal travel.” Transport research Part E 37, Pergamon, New York, 191–229. Burris, M., and Hannay, R. 共2003兲. “Equity analysis of the Houston, Texas, QuickRide project.” Transportation Research Record 1859, Transportation Research Board, National Research Council, Washington, D.C., 87–92. Burris, M. W., and Pendyala, R. M. 共2002兲. “Discrete choice models of traveler participation in differential time of day pricing programs.” Transp. Policy, 9共4兲, 66–77. Calfee, J., and Winston, C. 共1998兲. “The value of automobile travel time: Implications for congestion policy.” J. Public Econ., 69, 83–102. California Air Resources Board 共CARB兲. 共1996兲. Methodology for estimating emissions from on-road motor vehicles. Volume II: EMFAC7G, Technical Support Division, Mobile Source Emission Inventory Branch, Mobile. California Air Resources Board 共CARB兲. 共2003兲. EMFAC 2001/EMFAC 2002: Calculating emission inventories for vehicles in California— User’s guide, available on-line at: 具http://www.arb.ca.gov/msei/ on-road/previous_version.htm典 共Sept. 10, 2004兲. Delucchi, M. 共2000兲. “Environmental externalities of motor-vehicle use in the U.S.” J. Transp. Econ. Policy, 34共2兲, 135–168.
JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MARCH 2006 / 189
Federal Highway Administration 共FHwA兲. 共2002兲. Summary of funded value pricing pilot projects, 具http://www.hhh.umn.edu/centers/slp/ projects/conpric/index.htm典 共Dec. 10, 2002兲. Hau, T. 共1992兲. “Economic fundamentals of road pricing: A diagrammatic analysis.” Policy Research Working Papers WPS 1070, The World Bank. Hensher, D. 共2001兲. “Measurement of the valuation of travel time savings.” J. Transp. Econ. Policy, 35共1兲, 71–97. Hyman, G., and Mayhew, L. 共2002兲. “Optimizing the benefits of urban road user charging.” Transp. Policy, 9, 189–207. Levinson, D. 共2002兲. Financing transportation networks, Elgar, Northampton, Mass. Litman, T. 共1999兲. Using road pricing revenue: Economic efficiency and equity considerations, Victoria Transport Policy Institute, 具www.vtpi.org典 共Dec. 3, 2002兲. May, A., Milne, D., Shepherd, S., and Sumalee, A. 共2002兲. “Specification of optimal cordon pricing locations and charges.” Transportation Research Record 1812, Transportation Research Board, Washington, D.C., 60–68. Mohring, H. 共1999兲. “Congestion.” Essays in transportation economics and policy, Brookings Institution, Washington, D.C., 181–221. Office of Management and Budget 共OMB兲. 共2002兲. Circular No. A-94: Discount rates for cost-effectiveness, lease purchase, and related 具http://www.whitehouse.gov/omb/circulars/a094/a94_ analyses, appx-c.html典 共Dec. 12, 2002兲. Parry, I., and Bento, A. 共1999兲. Revenue recycling and the welfare effects of road pricing, Policy Research Working Paper, The World Bank Development Research Group, The World Bank. Santos, G., and Rojey, L. 共2003兲. “Distributional impacts of road pricing: The truth behind the myth.” 82nd Annual Meeting Transportation Research Board, Transportation Research Board, Washington, D.C. Schrank, D., and Lomax, T. 共2002兲. The 2002 urban mobility report,
Texas Transportation Institute, Texas A&M Univ., College Station, Tex. Small, K. 共1992兲. “Using the revenues from congestion pricing.” Transportation, 19共4兲, 351–389. Small, K., and Gomez-Ibanez, J. 共1997兲. “Road pricing for congestion management: The transition from theory to policy.” Transport economics: Selected readings, Harwood Academic, The Netherlands, 373–404. Small, K., and Kazimi, C. 共1995兲. “On the costs of air pollution from motor vehicles.” J. Transp. Econ. Policy, 共1兲, 7–32. Small, K., Noland, R., Chu, X., and Lewis, D. 共1999兲. “Valuation of travel-time savings and predictability in congested conditions for highway user-cost estimation.” Rep. 431, National Cooperative Highway Research Program, Transportation Research Board, Washington, D.C. United States Department of Transportation 共USDOT兲. 共2000兲. Highway economic requirements system: HERS user’s guide v2.36, Federal Highway Administration, Washington, D.C. United States Department of Transportation 共USDOT兲. 共2002兲. ITS unit cost database, 具http://www.benefitcost.its.gov/its/benecost.nsf/ ByLink/CostHome典 共June 25, 2002兲. United States Department of Transportation 共USDOT兲. 共2003兲. Revised departmental guidance on the valuation of travel time in economic analysis, memo, USDOT, February 11, 2003, 具http:// ostpxweb.dot.gov/典 共Dec. 10, 2003兲. United States Health Care Financing Administration 共USHCFA兲. 共2001兲. Index levels of medical prices: 1997-2001, 具http://www.hcfa.gov/stats/ indicatr/tables/t8.htm典 共June 15, 2002兲. Walters, A. 共1968兲. The economics of road user charges, International Bank for Reconstruction and Development, Johns Hopkins University Press, Baltimore.
190 / JOURNAL OF TRANSPORTATION ENGINEERING © ASCE / MARCH 2006