1

1

2

The Use of Passenger Transit Infrastructure for Goods Movement: 3

A Bay Area Economic Feasibility Study 4

5 by 6 7

K. Sivakumaran* 8

Graduate Student, UC Berkeley 9 416G McLaughlin Hall, UC Berkeley, Berkeley, CA 94704 10

Tel : (304) 517-0961 Fax: (510) 665-3537 11 Email : ksivakum@berkeley.edu 12

13

X. Y. Lu 14 California PATH, ITS, UC Berkeley 15

Richmond Field Station, Bldg 452, 1357 S. 46th Street, Richmond, CA 94804-4648 16

Tel: (510) 665-3644, Fax: (510) 665-3537 17 Email: xylu@path.berkeley.edu 18

19 M. Hanson 20

Division of Research and Innovation, Caltrans 21

1127 O Street, Sacramento, CA 94273 22 Tel: (916) 654-8171 23

Email: matt_hanson@dot.ca.gov 24

25

R. Lu 26 BART Division of Operations Planning 27

San Francisco Bay Area Rapid Transit District 28 Mail Drop: 1KB-6, 1000 Broadway, Oakland, CA 94612 29

Email: Rlu@bart.gov 30

31

F. Zaman, and R. Zhou 32 Federal Express , Regional - Sort Operation, Planning and Engineering 33

1 Sally Ride Way, Oakland, CA 94621-0002 34

Email: faisal.zaman@fedex.com 35 36

For presentation and publication 37

3rd

National Urban Freight Conference 38 Long Beach, CA 39

October 21-23, 2009 40

Word Count: 41

5, 224 words + 3 figures + 9 tables = 8,224 42

* Corresponding author

2

1

ABSTRACT 2

Truck traffic due to goods movement can increase traffic congestion, harmful emissions, 3

maintenance costs, and accidents all increase. One potential alternative to truck transport of 4

goods is mixed passenger-goods movement on urban rail networks, a concept motivated by 5

several potential benefits to involved parties, including additional revenue to rail operators, more 6

reliable travel time for goods, reductions in harmful trucking externalities, and more efficient 7

usage of existing facilities. 8

As a case study, the use of the Bay Area Rapid Transit (BART) system, for FedEx 9

Express cargo is examined in the following paper. FedEx Express was particularly selected due 10

to the location of the company’s major western regional hub at Oakland International Airport 11

(OAK). Four alternatives - A1, A2, B1, and B2 - are compared to the status quo of truck-only 12

transport. Alternatives A1 and A2 consider only minor capital investment, but differentiate in 13

that A2 utilizes electric vehicles rather than long-haul trucks for local transshipments. 14

Alternatives B1 and B2 assume far greater capital investment, including a jointly operated 15

BART/FedEx facility at OAK, but again differentiate in that B2 utilizes electric vehicles rather 16

for local transshipments. Truck VMT, FedEx operating costs, BART operating costs, and CO2 17

emissions are determined for the status quo and each alternative. Analysis shows not only that 18

significant truck VMT savings can be accrued from mixed-goods service, but that such service 19

can both be profitable for passenger rail systems and cost-effective for air cargo carriers. This 20

concept could potentially be pursued by transit systems in other metropolitan areas such as Los 21

Angeles, Washington D. C., New York and Chicago. 22

23

3

1

1 INTRODUCTION 2

As demand for air freight movement increases, so does truck traffic on California’s roads. As 3

truck transport grows, so do the problems attached with it, including traffic congestion, harmful 4

emissions, maintenance costs, and accidents. There exists an imperative to explore alternatives 5

to truck transport for freight movement, specifically the potential for existing passenger rail 6

networks to operate as a mode of freight transport. This concept is motivated by several 7

potential benefits to all parties. A passenger rail network offering service to both passengers and 8

air freight carriers, such as FedEx Express, can provide itself with a) a more efficient use of its 9

existing capacity, particularly during the off-peak periods, b) an additional source of revenue, 10

and c) potential source of cross-subsidization, where revenue from freight revenue can be 11

reinvested into improved passenger movement. The freight carrier, by shifting goods from truck 12

transport to rail transport, can potentially benefit from a) more reliable travel times for goods 13

movement, particularly during periods of non-recurrent congestion, b) fewer truck accidents, and 14

c) cheaper transport costs. Society, as a whole, may benefit from reduced truck traffic through 15

reductions in a) harmful emissions and GHGs, b) truck-induced delay, c) road maintenance costs, 16

and d) other negative externalities, such as noise and truck-related accidents. 17

With this concept in mind, this paper offers a case study of the potential for the use of 18

transit systems such as BART (the Bay Area Rapid Transit), in the movement of air freight. The 19

economic feasibility of such a system is presented from a purely operational perspective. A later 20

report will include the required capital costs, but for the purposes of this paper, capital 21

investments are considered external costs incurred by the government. The paper is structured as 22

follows. Section 2 describes the critical characteristics of the system under analysis. Section 3 23

outlines the assumptions and methodology of the economic feasibility model. Section 4 discusses 24

the quantitative results of our economic analysis, and Section 5 discusses how this type of 25

service might be further investigated. 26

27

2 SYSTEM CHARACTERISTICS 28

This section explores the characteristics of the system under analysis, specifically that of the 29

major passenger rail operator in the Bay Area, BART, as well as the area’s major air freight 30

4

carrier, FedEx Express. These two parties present unique capabilities and restrictions which 1

must be considered for accurate cost-benefit analysis. 2

3

2.1 Characteristics of the BART System 4

BART is the Bay Area’s primary passenger rail transit system which connects more than 26 5

cities across four counties covering over 104 miles of track connecting 43 stations (see FIGURE 6

1). BART provides five lines of service which generally operate at headways of approximately 7

15 minutes. However, due to overlaying, the frequencies between some stations are very high 8

during peak periods. Due to placement of interchanges at key transfer stations, BART trains can 9

travel between any two points seamlessly although daily passenger movement service only fixed 10

for certain lines. On average, BART utilizes roughly 30% of its mainline capacity, which 11

provides some motivation for exploring increased system utilization. 12

Unlike many other urban passenger rail systems, BART utilizes a nonstandard gauge of 5 13

ft 6 in; the standard gauge most commonly found in the U.S. is 4 ft 8.5 in. Consequently, the 14

system relies on custom-made rolling stock, which totals 669 cars. The dimensions for the 15

vehicles currently used by BART are 70 ft long, 10.5 ft wide and 7 ft high. Door dimensions are 16

approximately 4.5 ft X 6.5 ft (width X height). Due to each BART station’s length (~710 ft), 17

BART typically does not operate trains longer than 10 cars. Each BART car has a load capacity 18

of 30,000 lbs. BART trains are also able to operate in both directions without the need to turn 19

around. During a typical morning peak-hour commute, 541 cars are in service, while the 20

remainder are on “active duty”, and either being used to build spare trains, in repair, or involved 21

with other maintenance work. In lieu of the increased stress being placed on BART’s current 22

vehicles, the transit agency is looking to purchase 700 new cars in an effort to upgrade its entire 23

fleet by 2024. Any retired BART cars could potentially be retrofitted and used for freight 24

movement. Future planned BART extensions include connections to the Oakland International 25

Airport (OAK) as well as a track extension of the Fremont line southward to San Jose. These 26

future projects have the potential to strengthen the viability of mixed-goods service by BART. 27

28

2.2 Characteristics of the Air Freight Carrier: FedEx 29

Air freight carriers, such as FedEx, DHL, and UPS all have a strong need for frequent, 30

reliable, and cost-effective chain of transportation. With the recent withdrawal of DHL from the 31

5

U.S. air and ground carrier markets, a significant portion of carrier market share will likely be 1

subsumed by FedEx and UPS, which will further stress their existing transport chains. 2

Furthermore, the overall market volume is expected to increase with the growing number of 3

imports from rapidly emerging regions such as the Pacific Rim. 4

5

6

FIGURE 1 BART System Map 7 8

FedEx Express has been actively participating in an economic feasibility study due to its 9

significant presence in the Bay Area. The FedEx Western Regional Hub at OAK acts as a 10

sorting and distribution center to seven states on the West Coast with distribution/collection 11

centers scattered throughout the Bay Area. FedEx Express also maintains flights to and from the 12

San Francisco International Airport (SFO). As traffic congestion continues to increase along the 13

Bay Area’s urban corridors, FedEx has been exploring service alternatives which provide greater 14

reliability. 15

All container types, other than the USPS container, are transported throughout the FedEx 16

transport chain using ball-bearings; that is, items are not lifted unto pallets in order to be loaded 17

onto trucks. Rather, containers are simply slid across platforms and aboard trucks via the casters 18

mounted throughout loading platforms. 19

FedEx utilizes two truck types, CTV4 and CTV5, which differ in truck length 20

(unpublished data). The CTV4 is approximately 45 feet long, and the CTV5 is approximately 53 21

feet long. Both trucks are 10.5 ft wide. As mentioned earlier, the trucks also use ball-bearings 22

on their beds to allow for seamless loading and unloading of FedEx containers. It is also 23

6

important to note that FedEx containers operate within a “closed loop” for security purposes – 1

that is, once a container is initially filled and closed before distribution, it remains closed until it 2

reaches its final local sorting center. 3

FedEx Express demand is time dependent. FedEx packages originating from locations 4

outside the Bay Area and destined for the Bay Area are distributed in the morning between 3:00 5

AM and 5:00 AM. FedEx packages originating from the Bay Area and destined for locations 6

outside the Bay Area are collected between 5:00 PM and 9:00 PM. 7

8

3 MODEL ASSUMPTIONS AND METHODOLOGY 9

This section outlines more definitively some assumptions as the foundation for the next 10

economic analysis. 11

3.1 Case Study Alternatives 12

This analysis examines four alternatives for BART mixed-goods service over the period 13

between 2021 and 2040. These alternatives will be compared against the status quo (SQ), where 14

only trucks will be used for transport of FedEx products. Alternatives A1 and A2 will consider 15

the scenario where little capital investment is required – only existing BART yards and 16

maintenance areas will be used as FedEx transfer terminals. For Alternatives B1 and B2, far 17

greater capital investment is assumed, specifically a jointly operated BART/FedEx facility at the 18

FedEx Oakland Regional Hub. This would connect OAK to the existing BART network and 19

eliminate truck transshipments along the same link. Alternatives B1 and B2 also include 20

sufficient capital investment to allow for retrofitting of existing BART stations – specifically the 21

Glen Park and South San Francisco stations – so that these stations can adequately facilitate 22

FedEx goods movement. However, A1 and B1 make use of FedEx long-haul trucks for all goods 23

movement, while A2 and B2 utilize smaller, electric delivery trucks for local transshipments. 24

Note that for all alternatives, only operating cost expenditures are considered (with all 25

costs expressed in 2008 $). All alternatives actually require capital expenditures towards 26

loading/unloading platforms for containers, stripping and retrofitting of BART cars, and other 27

station modifications to ensure segregation between passengers and FedEx containers. A future 28

report is expected to outline these necessary capital expenditures. Note also that A1 and A2 29

require expansion of the FedEx truck fleet, because transshipments are required at both ends of a 30

travel link. 31

7

3.2 Exclusive rights by FedEx Express to BART Freight Movement Services 1

Here, economic analysis only considers the transport of FedEx products. It may be 2

beneficial to also consider the products transported by other carriers, such as UPS; doing so may 3

increase the financial viability from BART’s perspective, as the increased product demand 4

source might lead to the critical demand threshold being passed. However, this would require 5

further inspection into the container types and handling requirements of other carriers. While 6

economic analysis at this stage focuses only on FedEx product transport, demand values have 7

been scaled to incorporate possible demand sources from other carriers. 8

3.3 Locations 9

There are six key FedEx distribution centers in the Bay Area which are located near 10

existing/future BART facilities: Oakland, Concord, Hayward, Dublin, downtown San Francisco, 11

south San Francisco, and Milpitas. The Milpitas station is assumed to be ready for service by 12

2021. One exception is node 4b in A1 and A2 (FIGURE 2); while there is no BART station at 13

this location, there is a small BART maintenance yard known as the Oakland Annex Shop, 14

where there is a spur track and adequate room for truck loading/offloading. BART stations such 15

as Colma, Union City, and Concord are of particular interest due to their proximity to BART 16

yards. These nearby yards could be used for assembling freight consist. 17

18

19

20

21

22

23

24

25

26

27

28

FIGURE 2 BART network under consideration for goods movement in A1 and A2 29

30

2a: HWD

(Bay Fair)

4b: Oakland

Annex Shop 4a: OAK

5a: CCR

3a: LVK

3b: Dublin/

Pleasanton

2b: Union City

(YARD)

1b: Milpitas 1a: RHV

5b: Concord

(YARD)

(YARD)

7a: SQL

6a: SFO

(Oakland City

Center/12th St.)

6b/7b: Colma

(YARD)

8

1

2

3

4

5

6

7

8

9

10 11

FIGURE 3 BART network under consideration for goods movement in B1 and B2 12 13

See Figures 2 and 3 for a simple graphical depiction of the network for each alternative. 14

BART stations are represented by white circles, FedEx centers by black circles, and BART 15

transfer points by white squares. Node 4ab acts jointly as a BART station and FedEx center in 16

B1 and B2. Solid lines represent BART rail lines, and dashed lines represent truck 17

transshipments. 18

19

3.4 Demand 20

An origin-destination (OD) demand matrix has been developed through communications 21

with FedEx, but scaled so that true demand is not disclosed (for proprietary reasons). This 22

matrix, shown in Table 1, roughly conveys the daily demand, in [lbs] across the assumed 23

locations for the year 2009. A demand growth rate term 6% per year is also incorporated in 24

order to capture the expected future increases in goods movement. This growth rate value was 25

determined according to the Metropolitan Transportation Commission’s (MTC) current and 26

future estimates of Bay Area air cargo movements (2004). 27

All demand follows the time windows mentioned in Section 2, where all distribution 28

takes place during the AM period, and all collection takes place during the PM period. Because 29

items may be particularly time sensitive, both the AM and PM periods have been split into two 30

sub-periods, and demand has been assumed to be split equally amongst these sub-periods. 31

(Bay Fair)

6a: SFO

(Oakland City

Center/12th St.)

6b: Glen Park 7b: S. San

Francisco

5b: Concord

(Yard)

(YARD)

5a: CCR

7a: SQL

6b.7b: Colma

(YARD)

4ab:

OAK

2b: Union City

(YARD)

1b: Milpitas

2a: HWD

3a: LVK

3b: Dublin/

Pleasanton

1a: RHV

(Coliseum/

Oakland Airport)

9

1

TABLE 1 O-D Demand Matrix (from OAK and to OAK) in [lbs] 2 3

1a:

RHV

2a:

HWD

3a:

LVK

5a:

CCR

6a:

SFO

7a:

SQL

4a: From OAK 18,300 13,800 12,900 13,200 25,500 15,900

4a: To OAK 60,000 219,00 99,000 51,000 76,500 36,000

4 3.5 Container Type 5

Note that all FedEx container types, other than the USPS containers, would not be able to 6

both a) fit through existing BART car doors and b) be transported via BART, because they 7

require caster decks for movement. Therefore, to avoid retrofitting BART cars to include caster 8

decks and wider doors, which may be prohibitively expensive, it is reasonable to only consider 9

transport of USPS containers at this stage; using an existing container type would limit the 10

startup costs required from FedEx. A load density of 5.5 lbs/ft3 is assumed, which is roughly 11

equivalent to the average load density of most existing FedEx container types. 12

13

3.6 Dedicated Freight Cars 14

It is assumed that BART uses retrofitted bi-directionally operable “C” cars for freight 15

movement; some cars should be available for mixed-goods movement after the BART begins its 16

fleet overhaul in 2018. A BART “C” car could readily be stripped of all seating and fitted with 17

locking mechanisms for wheeled FedEx containers. 18

19

3.7 Vehicle Capacity 20

For SQ, A1, and B1, all FedEx trucks are assumed to be of type CTV5. Truck capacity is 21

constrained both by container shape and weight; when considering those constraints, CTV5 22

trucks have a capacity limit of 16 USPS containers. A constant fuel economy of 10 mpg is 23

assumed for the analysis period. 24

For A2 and B2, it is assumed that the local transshipments between FedEx distribution 25

centers and BART stations are made by zero-emission FedEx delivery vehicles (EVs), which are 26

already in service in the UK. These vehicles are further assumed to have the same capacity as 27

the commonly used FedEx W700 vehicle, which can carry 4 USPS containers. 28

However, A2 will maintain the use of CTV5 trucks for the transshipment between OAK 29

and Oakland Annex. 30

10

For all alternatives, it is assumed that BART uses retrofitted bi-directionally operable “C” 1

cars freight movement; these cars should be available after the agency’s fleet overhaul in 2018. 2

A BART “C” car could readily be stripped of all seating and fitted with locking mechanisms for 3

wheeled FedEx containers. A single BART car, when considering the weight and shape of 4

USPS containers as well as the dimensions of car doorways, can hold 24 USPS containers.1 5

6

3.8 Empty Container Returns 7

Empty container returns must also be considered, since the demand O-D table is likely to 8

be unbalanced (TABLE 1). The number of empty containers which must be transported to/from 9

a given location is simply the difference between the location’s outgoing and incoming demand. 10

11

3.9 Freight Car Storage Areas and Empty Train Travels 12

For A1 and A2, it is assumed that there is no storage space for BART cars at the Oakland 13

Annex Shop, so all freight consists must be assembled at one of BART’s four yards before 14

leaving for a particular route. During AM operations, since all movement is from OAK, the 15

Union City BART yard is assumed to supply freight consists to the Oakland Annex Shop station. 16

After unloading all goods at a given destination, the empty train will then depart for the nearest 17

yard. During PM operations, each origin will have trains supplied from whichever yard is 18

closest. After unloading, each empty train will return to the Union City yard. However, a new 19

train must then be assembled which returns empty containers to a given location (trains may be 20

either originating from or destined for OAK). Empty container returns simultaneously ensure 21

that the same number of freight cars will be in each yard at the beginning of each day. 22

For B1 and B2, it is assumed that there is adequate BART car storage at the jointly-23

operated FedEx/BART facility at node 4ab. Thus, this facility will assume the role played by the 24

Union City yard in A1 and A2. All other procedures from A1 and A2 are adopted for B1 and B2. 25

26

1 Note that if trucks and BART cars were fully packed with containers, they would be able to

carry 24 and 36 containers, respectively. However, some room must be within vehicles for locking

mechanism, as well as access to any container. Finally, while consolidation is the aim, vehicles may have

to leave their origin before being completely filled due to delivery time window limits or problematic

traffic conditions. Thus, reduced capacity constraints are used.

11

3.10 Link Travel Times 1

In order to determine the operating costs incurred for BART cars, the travel times across 2

the system network must be considered. Since it is infeasible for a single train to transport all 3

goods throughout the network due to FedEx delivery time window, particular routes have been 4

assigned for BART freight trains, as seen in Table 2. Route travel times were approximated 5

from BART schedules (Bay Area Rapid Transit, 2009). However, those provided times include 6

the dwell times across intermittent stops; thus, an assumed 40 second dwell time per stop has 7

been subtracted from the provided BART trip times. Furthermore, the aforementioned empty 8

train travel times (from/to the yards corresponding to the beginning and end of a particular route) 9

should be added to all route travel times in order to calculate the true cost of service. The 10

resulting total service times, as well as the route travel times, are given in Table 2. 11

12 TABLE 2 BART Travel Times for A1, B1, A2, and B2 13

14

Route

Route Travel Time

[hrs]

Total Service Time

[hrs]

A1

an

d A

2 1b 2b 4b 0.66 1.33

3b 4b 0.51 1.34

5b 4b 0.52 0.92

7b/6b 4b 0.56 0.96

B1

an

d B

2 1b 2b 4ab 0.66 0.93

3b 4ab 0.46 0.90

5b 4ab 0.65 0.65

7b 6b 4ab 0.72 0.77

15 16

It is also necessary to determine the link travel distances and times across the road 17

network for truck transport of FedEx goods. Travel distances for all links are determined 18

according to the locations of BART stations and FedEx distribution centers. These distances are 19

listed in Table 3. However, while it can safely be assumed that transshipment travel times 20

between local FedEx distribution centers and BART stations will remain constant, the line-haul 21

travel times in the Status Quo might be expected to change over time due to rising freeway 22

demand over time (and subsequently, reduced travel speeds). Thus, analysis makes use of the 23

predicted inter-regional travel speeds given by the Bay Area’s Metropolitan Transportation 24

Commission (2008a). 25

26

12

3.11 Handling 1

A handling time of 0.025 hrs (1.5 minutes) per container is assumed. This handling time 2

is defined as the time to move one container from a BART vehicle across a station platform and 3

secure the container onto a FedEx truck (and vice-versa). With respect to SQ, A1 and A2 will 4

require two additional handling movements (one movement at all local distribution centers and 5

one movement at node 4b), and both B1 and B2 will require one additional handling movement 6

(at all local distribution centers). Total handling cost is estimated by simply multiplying the 7

assumed handling time per container by the number of containers handled, the number of 8

handling movements, and the handler costs. Regardless of how responsibility is divided for 9

loading/offloading items at BART stations, BART must still incur some labor cost as train 10

operators must stand by until trains are loaded. Similarly, FedEx truck operators will at least be 11

on standby while trucks are loaded. Therefore, this total handling time cost is assumed to be 12

incurred by both FedEx and BART. 13

TABLE 3 Truck Travel Distances for Various Alternatives 14 15

Link

Truck Travel Distance

[mi]

ST

AT

US

QU

O 4a 1a 33.1

4a2a 15.5

4a 3a 20.9

4a 5a 31.6

4a 6a 22.4

4a 7a 28.6

A1

an

d A

2

1b 1a 3

2b 2a 4

3b 3a 3.6

5b 5a 4

6b/7b 6a 8.6

6b/7b 7a 4.9

4a 4b 8.8

B1

an

d B

2

1b 1a 3

2b 2a 4

3b 3a 2.6

5b 5a 4

6b 6a 3.8

7b 7a 3.4

16

13

3.12 Truck VMT Estimation 1

The number of trucks required between any two points can be approximated as nij =D ij

K, 2

where Dij = demand originating from location i and destined for location j, in [containers], and 3

K = truck capacity, in [containers]. With these capacity considerations, the demand matrix 4

provided earlier can be readily converted into a “Truck Trips” matrix. From the number of truck 5

trips, the annual truck VMT, external social costs, and total truck operating costs can be 6

calculated for each alternative. All trucks are assumed to travel round-trip. 7

8

3.13 External/Social Costs 9

There are four key external social costs that arise from heavy-duty truck VMT: 1) 10

congestion/delay costs, 2) emissions costs, 3) accident costs, and 4) infrastructure maintenance 11

costs. 12

The delay savings which arise from reduced truck VMT are difficult to estimate, given 13

that the savings depend on the freeway’s total demand. While it should be noted that there may 14

be some delay reduction from the proposed mixed-goods service, no quantification of these 15

savings are included here. 16

To quantify reductions in harmful GHG’s, a CO2 emissions rate of 22.2 lbs/gallon diesel 17

fuel is used (Office of Transportation and Air Quality 2008). A constant fuel economy of 10 18

mpg has been assumed over the analysis period. 19

By reducing truck VMT, the number of freeway accidents will likely be reduced. To 20

quantify this cost, a truck accident rate is used. This rate is derived from the 2006 statistics for 21

the United States (Federal Motor Carrier Safety Administration, 2009a). The average cost of all 22

heavy truck accidents in 2005 was $91,112 (Federal Motor Carrier Safety Administration, 23

2007b). The annual accident cost is given by the product of two factors together with the annual 24

truck VMT. Half of this cost is allocated to FedEx annual operating cost, and half to society. 25

The road damage unit cost due to urban truck travel is taken as $0.074 per truck-mile 26

(Parry, 2008, p. 651-668). 27

28

3.14 Operating and Labor Costs 29

The 2008 per-mile truck operating costs were given as $3.28 in an unpublished report by 30

R. Rai (whose previous work is continued in this study) but because this per-mile cost included 31

14

driver costs, a lower operating cost of $2.70/mile is used for CTV5 vehicles so that the per-mile 1

trucking cost only includes fuel, maintenance, and ownership/depreciation costs. The baseline 2

per-mile operating cost for FedEx EVs was taken to be $2.40/mile. Note that both these 3

operating cost values are 2010 baseline values; we assume that these operating costs increase 4

each year with rising fuel costs and inflation. Diesel fuel costs are assumed to follow the same 5

rising trend as gasoline costs, which from recent forecasts are expected to rise to roughly 6

$7.50/gal by 2035 (Metropolitan Transportation Commission, 2008b). For EVs, we assume the 7

hourly operating cost to rise at $0.03 per year. 8

The per-hour cost includes FedEx driver wages, which are taken as $22/hr (PayScale, 9

2009), and both health insurance and worker’s compensation, leading to an hourly cost of $30 10

(Bureau of Labor Statistics, 2009). Hourly handling labor cost is assumed to be the same as 11

FedEx driver costs ($31/hr). Other components of driving cost include $8 bridge tolls, a 12

$0.25/mile congestion charge, and a 20% carbon tax on all VMT. All three of these charges are 13

expected to be implemented by 2035 in the Bay Area (Metropolitan Transportation 14

Commission, 2008a). The congestion charge is only applied to the freeway travel component of 15

each truck trip, and the bridge toll cost only to those trucks which utilize links 4a6a and 16

4a7a (only these two links require travel across the Bay Bridge). 17

Difficulties arise in approximating BART operating cost due to data limitations; through 18

communications with BART’s Financial Planning Office, an operating cost of $250/per car-hr, 19

was obtained. This value is simply the total annual operating cost by the total annual car-hours. 20

However, the operating cost for trains, which possess economies of scale, is required. Therefore, 21

rather than directly apply the given operating cost per car-hr to a train with 𝑛 cars, we assume the 22

average train length to be six cars, which corresponds to an hourly operating cost of $1,500. 23

However, because freight trains will not require the same level of amenities – i.e. air 24

conditioning – as passenger trains, a reduced operating cost value of $1,250/hr is used for a six-25

car train. Because the relationship between train length 𝑛 and cost should show economies of 26

scale, it is assumed that the hourly operating cost can be expressed as the concave function 27

1000𝑛1/8, where 2 ≤ 𝑛 ≤ 10. Using this function, the hourly freight train operating cost ranges 28

from $1,090 for a 2-car train to $1,330 for a 10-car train. 29

Total BART train hours are determined similar to total truck VMT. The number of 30

BART cars required is simply given by nk = Dk

KBART, where Dk = demand for route , in 31

15

[containers], and KBART = BART car capacity, in [containers]. From the number of cars required, 1

one can determine the number of trains required by dividing the required number of cars by the 2

10-car limit; the appropriate operating cost value (according to the function above) is then 3

assigned to the required trains and their respective lengths. A single train’s daily operating cost 4

is given by the product of its hourly operating cost and its total time in service. The annual 5

operating cost is found by summing the operating cost over all trains and days for the year. 6

7

3.15 Security 8

All FedEx goods, once containerized at their origin distribution center, are not opened 9

until reaching their destination distribution center. Furthermore, all FedEx goods leaving the 10

Oakland Regional Hub must pass airport security. Thus, it is reasonable to expect that FedEx 11

containers which meet stringent airport security requirements will be more than capable of 12

meeting any security requirements imposed by BART. 13

14

4 RESULTS 15

A summary of the different alternatives is tabulated in Table 4. Analysis results for all 16

alternatives are tabulated in Tables 5 through 9. The level of subsidy required will simply be the 17

difference between the total cost of a given alternative and the total cost of the status quo. Any 18

years which show a negative level of required subsidy indicate that no subsidies are required; 19

these years instead show the potential for BART profits and FedEx savings. Note that several of 20

the columns in Tables 6 through 9 show savings in truck VMT and truck-related externalities 21

rather than the raw values. 22

Alternatives A1 and A2, which entail minimal capital investment, require some subsidy 23

throughout the timeline. A1 requires between roughly $4M and $6M each year, while A2 24

requires roughly between $12M and $25M annually. However, tremendous savings in truck 25

VMT can be achieved; the cumulative truck VMT savings over the analysis period amounts to 26

nearly 42 million VMT for A1 and more than 60 million VMT for A2. These savings translate 27

to CO2 emission savings of more than 45,000 tons for A1 and more than 70,000 tons for A2. 28

For B1 and B2, which both require more significant capital investment but also eliminate 29

one of the transshipments required in A1 and A2, even greater savings in truck VMT are 30

achieved. For B1, the cumulative VMT savings amount to more than 80 million VMT, which 31

16

translates to more than 92,000 tons of CO2 emission savings. For B2, the cumulative VMT 1

savings amount to nearly 105 million VMT, which translates to nearly 120,000 tons of CO2 2

emission savings. 3

Perhaps most interestingly, no subsidy is ever required for B1. Thus, if container demand 4

is sufficient, BART mixed-goods service can both be profitable for BART and beneficial for 5

FedEx from solely a fiscal perspective. The exact levels of profit for BART and savings for 6

FedEx will simply depend on a mutually agreed-upon price for transported containers. 7

Assuming a discount rate of 5%, the accumulated amount of total cost savings compared to the 8

status quo amounts to roughly $51M. These savings can be channeled towards recovering the 9

initial capital investment, as well as towards improvements in passenger service, which can 10

further incentivize transit ridership. Note that a clear trend is established in the results: the 11

higher the demand level, the more profitable mixed-goods operation becomes. 12

However, note that there exist other social benefits not included here, such as reduced 13

congestion, noise, particulate matter reduction, and more economic land use. Government 14

agencies must weigh all of these benefits against any required subsidies or start-up costs in order 15

to determine whether or not mixed-goods service is worthy of pursuit. 16

17

18

19 20

17

TABLE 4 Summary of Case-Study Alternatives 1 2

1 2

A

CTV5 Trucks for local transshipments;

Existing BART yards and maintenance

areas for access point;

Dedicated freight train

Electric trucks for local transshipments;

Existing BART yards, stations and

maintenance areas for access point;

Dedicated freight train

B

CTV5 Trucks for local transshipments;

BART connection between OAK and

Colliseum Station;

Retrofitting of existing BART stations

for goods movement;

Dedicated freight train

Electric trucks for local transshipments;

BART connection between OAK and

Colliseum Station

Retrofitting of existing BART stations for

goods movement;

Dedicated freight train

18

TABLE 5 Analysis Results for the Status Quo

Year

CTV5 Truck

VMT FedEx Costs BART Costs Total Cost

Minimum

Subsidy

Required (+)

# CTV5

Accidents

Accident

$

Maintenance

$ CO2 (lbs)

2021 2,907,770 $13,181,300 $0 $13,181,300 $0 5 $240,450 $215,170 6,455,200

2022 3,112,990 $14,165,200 $0 $14,165,200 $0 6 $257,420 $230,360 6,910,800

2023 3,245,130 $14,805,800 $0 $14,805,800 $0 6 $268,350 $240,140 7,204,200

2024 3,444,130 $15,763,300 $0 $15,763,300 $0 6 $284,800 $254,870 7,646,000

2025 3,604,270 $16,559,500 $0 $16,559,500 $0 6 $298,040 $266,720 8,001,500

2026 3,873,810 $17,844,600 $0 $17,844,600 $0 7 $320,330 $286,660 8,599,900

2027 4,091,370 $18,902,800 $0 $18,902,800 $0 7 $338,320 $302,760 9,082,800

2028 4,345,100 $20,145,600 $0 $20,145,600 $0 8 $359,300 $321,540 9,646,100

2029 4,525,330 $21,041,500 $0 $21,041,500 $0 8 $374,210 $334,870 10,046,200

2030 4,798,540 $22,387,100 $0 $22,387,100 $0 8 $396,800 $355,090 10,652,800

2031 5,112,190 $23,936,700 $0 $23,936,700 $0 9 $422,740 $378,300 11,349,100

2032 5,309,050 $24,931,200 $0 $24,931,200 $0 9 $439,010 $392,870 11,786,100

2033 5,725,770 $26,969,100 $0 $26,969,100 $0 10 $473,470 $423,710 12,711,200

2034 6,040,700 $28,544,200 $0 $28,544,200 $0 10 $499,520 $447,010 13,410,400

2035 6,427,790 $30,481,300 $0 $30,481,300 $0 11 $531,530 $475,660 14,269,700

2036 6,666,310 $31,711,800 $0 $31,711,800 $0 12 $551,250 $493,310 14,799,200

2037 7,144,280 $34,105,200 $0 $34,105,200 $0 12 $590,770 $528,680 15,860,300

2038 7,590,080 $36,349,500 $0 $36,349,500 $0 13 $627,640 $561,670 16,850,000

2039 8,014,650 $38,502,000 $0 $38,502,000 $0 14 $662,750 $593,080 17,792,500

2040 8,471,460 $40,839,000 $0 $40,839,000 $0 15 $700,520 $626,890 18,806,600

19

TABLE 6 Analysis Results for Alternative A1

Year

CTV5 Truck

VMT Savings FedEx Costs BART Costs Total Cost

Minimum

Subsidy

Required (+)

# Truck

Accidents

Saved

Accident

$ Saved

Maintenance

$ Saved

CO2

Saved

(lbs)

2021 1,193,810 $9,475,200 $10,136,000 $19,611,200 $6,429,900 2 $98,720 $88,340 2,650,200

2022 1,280,000 $10,138,700 $9,722,200 $19,860,900 $5,695,700 2 $105,850 $94,720 2,841,600

2023 1,324,570 $10,668,300 $10,259,700 $20,928,000 $6,122,200 2 $109,540 $98,020 2,940,600

2024 1,405,860 $11,340,500 $10,993,700 $22,334,200 $6,570,900 2 $116,250 $104,040 3,121,000

2025 1,456,460 $11,988,200 $10,983,200 $22,971,400 $6,411,900 2 $120,430 $107,780 3,233,400

2026 1,582,120 $12,804,600 $11,491,400 $24,296,000 $6,451,400 3 $130,830 $117,070 3,512,300

2027 1,658,520 $13,615,800 $11,195,800 $24,811,600 $5,908,800 2 $137,140 $122,730 3,681,900

2028 1,773,020 $14,435,000 $11,701,200 $26,136,200 $5,990,600 3 $146,610 $131,210 3,936,100

2029 1,813,350 $15,262,500 $11,832,400 $27,094,900 $6,053,400 3 $149,950 $134,180 4,025,600

2030 1,922,650 $16,218,800 $12,531,300 $28,750,100 $6,363,000 3 $158,990 $142,270 4,268,300

2031 2,058,260 $17,256,700 $12,764,600 $30,021,300 $6,084,600 3 $170,210 $152,310 4,569,400

2032 2,100,890 $18,186,900 $13,086,500 $31,273,400 $6,342,200 3 $173,720 $155,470 4,664,000

2033 2,318,150 $19,359,100 $13,444,800 $32,803,900 $5,834,800 4 $191,690 $171,550 5,146,300

2034 2,416,840 $20,615,600 $13,395,000 $34,010,600 $5,466,400 4 $199,860 $178,840 5,365,400

2035 2,591,210 $21,877,400 $13,061,000 $34,938,400 $4,457,100 4 $214,280 $191,750 5,752,500

2036 2,625,530 $23,114,400 $13,539,000 $36,653,400 $4,941,600 5 $217,110 $194,290 5,828,700

2037 2,845,080 $24,629,400 $13,498,400 $38,127,800 $4,022,600 4 $235,260 $210,540 6,316,100

2038 3,035,420 $26,149,600 $14,765,700 $40,915,300 $4,565,800 5 $251,010 $224,630 6,738,700

2039 3,192,140 $27,746,500 $15,297,500 $43,044,000 $4,542,000 6 $263,970 $236,210 7,086,500

2040 3,361,060 $29,467,100 $15,130,400 $44,597,500 $3,758,500 6 $277,930 $248,720 7,461,500

20

TABLE 7 Analysis Results for Alternative A2

Year

CTV5 Truck

VMT

Savings FedEx Costs BART Costs Total Cost

Minimum

Subsidy

Required (+)

# Truck

Accidents

Saved

Accident

$ Saved

Maintenance

$ Saved

CO2 Saved

(lbs)

2021 1,790,260 $14,821,200 $10,136,000 $24,957,200 $11,775,900 3 $148,040 $132,470 3,974,300

2022 1,923,670 $15,821,700 $9,722,200 $25,543,900 $11,378,700 4 $159,070 $142,350 4,270,500

2023 1,992,980 $16,875,000 $10,259,700 $27,134,700 $12,328,900 4 $164,810 $147,480 4,424,400

2024 2,115,680 $17,940,800 $10,993,700 $28,934,500 $13,171,200 4 $174,950 $156,560 4,696,800

2025 2,199,530 $19,112,600 $10,983,200 $30,095,800 $13,536,300 4 $181,880 $162,770 4,883,000

2026 2,379,310 $20,440,900 $11,491,400 $31,932,300 $14,087,700 5 $196,750 $176,070 5,282,100

2027 2,507,110 $21,788,100 $11,195,800 $32,983,900 $14,081,100 5 $207,310 $185,520 5,565,700

2028 2,671,080 $23,214,400 $11,701,200 $34,915,600 $14,770,000 5 $220,870 $197,660 5,929,800

2029 2,748,080 $24,731,900 $11,832,400 $36,564,300 $15,522,800 5 $227,250 $203,350 6,100,700

2030 2,913,580 $26,390,500 $12,531,300 $38,921,800 $16,534,700 5 $240,930 $215,600 6,468,200

2031 3,115,030 $28,126,500 $12,764,600 $40,891,100 $16,954,400 6 $257,590 $230,510 6,915,400

2032 3,204,180 $29,865,700 $13,086,500 $42,952,200 $18,021,000 6 $264,950 $237,110 7,113,300

2033 3,499,720 $31,869,000 $13,444,800 $45,313,800 $18,344,700 7 $289,390 $258,980 7,769,400

2034 3,666,550 $34,052,500 $13,395,000 $47,447,500 $18,903,300 7 $303,200 $271,320 8,139,800

2035 3,919,000 $36,290,100 $13,061,000 $49,351,100 $18,869,800 7 $324,070 $290,010 8,700,200

2036 4,004,930 $38,680,000 $13,539,000 $52,219,000 $20,507,200 8 $331,180 $296,370 8,890,900

2037 4,325,820 $41,240,500 $13,498,400 $54,738,900 $20,633,700 8 $357,710 $320,110 9,603,300

2038 4,605,560 $43,976,100 $14,765,700 $58,741,800 $22,392,300 9 $380,840 $340,820 10,224,400

2039 4,850,610 $46,848,200 $15,297,500 $62,145,700 $23,643,700 10 $401,110 $358,940 10,768,300

2040 5,114,440 $49,972,000 $15,130,400 $65,102,400 $24,263,400 10 $422,920 $378,470 11,354,000

21

TABLE 8 Analysis Results for Alternative B1

Year

CTV5 Truck

VMT

Savings FedEx Costs BART Costs Total Cost

Minimum

Subsidy

Required (+)

# Truck

Accidents

Saved

Accident

$ Saved

Maintenance

$ Saved

CO2 Saved

(lbs)

2021 2,311,320 $3,540,200 $7,893,300 $11,433,500 -$1,747,800 4 $196,310 $171,030 5,131,100

2022 2,469,320 $3,811,700 $7,618,900 $11,430,600 -$2,734,600 4 $209,790 $182,730 5,481,900

2023 2,576,720 $3,985,400 $8,026,200 $12,011,600 -$2,794,200 4 $218,890 $190,680 5,720,300

2024 2,734,310 $4,236,800 $8,623,700 $12,860,500 -$2,902,800 4 $232,270 $202,340 6,070,200

2025 2,861,200 $4,456,900 $8,658,100 $13,115,000 -$3,444,500 4 $243,050 $211,730 6,351,900

2026 3,076,630 $4,778,600 $9,030,900 $13,809,500 -$4,035,100 5 $261,340 $227,670 6,830,200

2027 3,242,780 $5,089,100 $8,864,900 $13,954,000 -$4,948,800 5 $275,520 $239,960 7,198,900

2028 3,447,040 $5,400,600 $9,272,800 $14,673,400 -$5,472,200 6 $292,840 $255,080 7,652,400

2029 3,590,600 $5,653,100 $9,388,200 $15,041,300 -$6,000,200 6 $305,040 $265,700 7,971,100

2030 3,807,610 $6,004,500 $9,923,400 $15,927,900 -$6,459,200 6 $323,470 $281,760 8,452,900

2031 4,055,420 $6,408,800 $10,146,800 $16,555,600 -$7,381,100 7 $344,540 $300,100 9,003,100

2032 4,205,770 $6,721,700 $10,449,600 $17,171,300 -$7,759,900 7 $357,370 $311,230 9,336,800

2033 4,544,200 $7,202,800 $10,759,600 $17,962,400 -$9,006,700 8 $386,030 $336,270 10,088,100

2034 4,791,000 $7,633,500 $10,785,000 $18,418,500 -$10,125,700 7 $407,040 $354,530 10,636,100

2035 5,100,000 $8,123,200 $10,566,700 $18,689,900 -$11,791,400 8 $433,270 $377,400 11,322,000

2036 5,286,910 $8,492,500 $10,984,400 $19,476,900 -$12,234,900 9 $449,170 $391,230 11,736,900

2037 5,663,550 $9,108,200 $11,049,600 $20,157,800 -$13,947,400 9 $481,200 $419,110 12,573,100

2038 6,019,940 $9,677,100 $12,043,100 $21,720,200 -$14,629,300 10 $511,450 $445,480 13,364,300

2039 6,356,180 $10,246,700 $12,521,000 $22,767,700 -$15,734,300 11 $540,020 $470,350 14,110,700

2040 6,718,080 $10,860,800 $12,487,600 $23,348,400 -$17,490,600 12 $570,770 $497,140 14,914,100

22

TABLE 9 Analysis Results for Alternative B2

Year

CTV5 Truck

VMT Savings FedEx Costs BART Costs Total Cost

Minimum

Subsidy

Required (+)

# Truck

Accidents

Saved

Accident

$ Saved

Maintenance

$ Saved

CO2 Saved

(lbs)

2021 2,907,770 $8,644,000 $7,893,300 $16,537,300 $3,356,000 5 $240,450 $215,170 6,455,200

2022 3,112,990 $9,233,600 $7,618,900 $16,852,500 $2,687,300 6 $257,420 $230,360 6,910,800

2023 3,245,130 $9,920,000 $8,026,200 $17,946,200 $3,140,400 6 $268,350 $240,140 7,204,200

2024 3,444,130 $10,549,900 $8,623,700 $19,173,600 $3,410,300 6 $284,800 $254,870 7,646,000

2025 3,604,270 $11,278,600 $8,658,100 $19,936,700 $3,377,200 6 $298,040 $266,720 8,001,500

2026 3,873,810 $12,090,200 $9,030,900 $21,121,100 $3,276,500 7 $320,330 $286,660 8,599,900

2027 4,091,370 $12,915,400 $8,864,900 $21,780,300 $2,877,500 7 $338,320 $302,760 9,082,800

2028 4,345,100 $13,816,300 $9,272,800 $23,089,100 $2,943,500 8 $359,300 $321,540 9,646,100

2029 4,525,330 $14,741,900 $9,388,200 $24,130,100 $3,088,600 8 $374,210 $334,870 10,046,200

2030 4,798,540 $15,773,400 $9,923,400 $25,696,800 $3,309,700 8 $396,800 $355,090 10,652,800

2031 5,112,190 $16,848,700 $10,146,800 $26,995,500 $3,058,800 9 $422,740 $378,300 11,349,100

2032 5,309,050 $17,951,800 $10,449,600 $28,401,400 $3,470,200 9 $439,010 $392,870 11,786,100

2033 5,725,770 $19,232,200 $10,759,600 $29,991,800 $3,022,700 10 $473,470 $423,710 12,711,200

2034 6,040,700 $20,562,900 $10,785,000 $31,347,900 $2,803,700 10 $499,520 $447,010 13,410,400

2035 6,427,790 $21,995,200 $10,566,700 $32,561,900 $2,080,600 11 $531,530 $475,660 14,269,700

2036 6,666,310 $23,497,600 $10,984,400 $34,482,000 $2,770,200 12 $551,250 $493,310 14,799,200

2037 7,144,280 $25,119,400 $11,049,600 $36,169,000 $2,063,800 12 $590,770 $528,680 15,860,300

2038 7,590,080 $26,865,000 $12,043,100 $38,908,100 $2,558,600 13 $627,640 $561,670 16,850,000

2039 8,014,650 $28,674,200 $12,521,000 $41,195,200 $2,693,200 14 $662,750 $593,080 17,792,500

2040 8,471,460 $30,653,800 $12,487,600 $43,141,400 $2,302,400 15 $700,520 $626,890 18,806,600

23

5 CONCLUDING REMARKS 1

The presented economic feasibility study shows significant promise for exploring the 2

possibility of mixed-goods service on passenger-rail systems. Two of the proposed service 3

alternatives show a trend suggesting that the higher the container demand, the lower the level of 4

subsidy required. Furthermore, with adequate capital investment, BART can actually derive 5

profit from mixed-goods service while the air freight carrier (FedEx Express, in this case study) 6

can derive savings. The fact that economic feasibility is so closely linked to demand should 7

motivate researchers to explore ways to tap into other demand sources so that mixed-goods 8

service can be both profitable and environmentally sustainable. An additional benefit for freight 9

carriers not quantified here was the higher level of reliability provided by rail transport compared 10

to truck transport. Including this benefit may further skew the results towards mixed-goods 11

service. 12

There are several future extensions to this study which will be pursued. First, the 13

presented results are based on parameter values which may require revision; the true operating 14

costs of FedEx vehicles and BART trains should be used in order for appropriate economic 15

analysis. One critical consideration is the capital cost required for such service. Once a valid 16

estimate of initial capital and other “start-up” costs is included, all future benefits can be 17

discounted so that a true economic analysis can be conducted. Future studies should also 18

examine in greater detail some of the logistical issues inherent in mixed-goods service, including 19

non-interference with existing passenger transport operations and routine maintenance activities. 20

Another critical element that can contribute to the success of mixed-goods service is the 21

inclusion of other cargo movers as BART “customers”. Doing so can ensure that demand levels 22

are high enough to promote profits for BART and savings for FedEx customers. 23

These results should be of particular interest to other urban areas such as such as Los 24

Angles, Washington D. C., New York and Chicago across the U.S. where passenger rail systems 25

exist in close proximity to major air cargo terminals. Some of these systems may possess 26

particularly favorable characteristics towards mixed-goods movement, such as intermodal 27

transfer stations, containers which can interface between multiple modes, and standard gauge 28

rails. 29

30

31

24

32

ACKNOWLEDGEMENT 33

This project is sponsored by UCTC (University of California Transportation Center) and 34

the State of California Business, Transportation and Housing Agency, Department of 35

Transportation. The contents of this report reflect the views of the authors who are responsible 36

for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect 37

the official views or policies of the State of California. This paper does not constitute a standard, 38

specification, or regulation. All the ideas presented here are for research purposes. Neither 39

BART nor FedEx carries any commitment or liability for any content of this paper. 40

Support from Tom Messer, Marcus Evans and Michele Fell, in Caltrans Goods 41

Movement are gratefully acknowledged. The demand data used in all calculations has been 42

scaled, and is not true FedEx demand data. 43

44

25

REFERENCES 45

BAY AREA RAPID TRANSIT (2009). BART Trip Planner. www.bart.gov/. Accessed 46

April 19, 2009. 47

48

BUREAU OF LABOR STATISTICS (2009). Employer Costs for Employee Compensation 49

Summary. www.bls.gov/news.release/ecec.nr0.htm/. Accessed July 29, 2009. 50

51

FEDERAL MOTOR CARRIER SAFETY ADMINISTRATION (2007a). Commercial 52

Motor Vehicle Facts. http://www.fmcsa.dot.gov/facts-research/facts-figures/analysis-53

statistics/cmvfacts.htm/. Accessed July 27, 2009. 54

55

FEDERAL MOTOR CARRIER SAFETY ADMINISTRATION (2009b). Motor Carrier 56

Safety Progress Report. http://www.fmcsa.dot.gov/facts-research/art-safety-progress-57

report.htm /. Accessed July 27, 2009. 58

59

METROPOLITAN TRANSPORTATION COMMISSION (2008a). Change In Motion: 60

Transportation 2035 Plan for the Bay Area. 2008. 61

62

METROPOLITAN TRANSPORTATION COMMISSION (2004). Regional Goods 63

Movement Study for the San Francisco Bay Area: Final Summary Report. 64

65

METROPOLITAN TRANSPORTATION COMMISSION (2008b). Travel Forecasts Data 66

Summary: Transportation 2035 Plan for the Bay Area. 67

68

OFFICE OF TRANSPORTATION AND AIR QUALITY (2008). Calculating Greenhouse 69

Gas Emissions: Key Facts and Figures. United States Environmental Protection Agency. 70

71

PARRY, IAN (2008). How should heavy-duty trucks be taxed? Journal of Urban 72

Economics. Vol. 63, Issue 2, pp. 651-668. 73

74

PAYSCALE (2009). Median Hourly Rate by Job, Employer: FedEx/FedEx Express 75

Corporation. 76

www.payscale.com/research/US/Employer=FedEx_/_Federal_Express_Corporation/Hourly_77

Rate>. Accessed July 29, 2009. 78