Dynamic Benchmarking of Mass Transit Systems in the United States Using Data Envelopment Analysis and the Malmquist Productivity Index

• Author: Young‐Hyo Ahn,Hokey Min
• DOI: http://doi.org/10.1111/jbl.12148
• Date: 01 March 2017
• Published date: 01 March 2017

Dynamic Benchmarking of Mass Transit Systems in the United

States Using Data Envelopment Analysis and the Malmquist

Productivity Index

Hokey Min

1

and Young-Hyo Ahn

2

1

Bowling Green State University

2

Incheon National University

For more than five decades, the federal, state, and local governments have subsidized mass transit systems through sales, gasoline, and

property taxes with an expectation that it would improve mobility to low-income citizens, reduce carbon footprints and traffic congestion,

and facilitate regional economic growth. However, in times of financial crisis and chronic government budget deficits, the inefficient use of a

mass transit system can increase public outcry over the wasteful spending of government funds and taxpayers’monies. To find ways to utilize

mass transit systems more efficiently across the United States, this paper aimed to identify the benchmark transit practices that every mass

transit system can emulate and then continuously improve its performance. To achieve these goals, this paper analyzed the multiple years of

past performances of 262 mass transit agencies in the United States using data envelopment analysis and the Malmquist productivity index and

then provided practical guidelines for enhancing mass transit efficiency.

Keywords: mass transit systems; dynamic benchmarking; data envelopment analysis; longitudinal study

INTRODUCTION

The American Public Transportation Association (2014a)

recently reported that mass transit use in the United States rose

to 10.7 billion trips in 2013—the highest ridership use in

57 years. The continuous growth of the transit ridership is attrib-

uted to rising gasoline price, recent economic slowdown, and

demographic shifts triggered by the Baby Boomers’return to

urban areas and Millennials’desire for more affordable travel

options (American Public Transportation Association 2015).

Growing demand for mass transit often necessitates the expan-

sion/diversification of service offerings, the improvement of

transportation infrastructure, the replacement of old vehicles with

new ones, and the additional hiring of mass transit employees

including drivers and maintenance crews. As such, mass transit

authorities call for increased funding. As the main sources of

transit funding primarily come from government subsidies

through increased sales, gasoline, and property taxes, the general

public who should bear the financial burden of supporting mass

transit systems have increasingly scrutinized the investment in

mass transit services which generally refer to large-scale public

transportation services on a fixed route in shared vehicles in a

city or a metropolitan area. These vehicles include cable cars,

electric streetcars, trolley coaches, gasoline and diesel-powered

buses, underground and above-ground rail rapid transit, ferries,

and some commuter trains. In the United States, mass transit has,

for the most part, meant some kind of local bus or passenger rail

service (Schrag 2000).

In an era of tight budgetary constraints and government down-

sizing, however, the mass transit authority cannot afford to spend

wastefully or to make risky future investments that cannot be

paid back fully. In other words, a thorough analysis of current

mass transit efficiency and some type of managerial improve-

ments based on such analysis are essential for sustaining the

vitality of mass transit systems.

Generally, important benefits of mass transit services may

include (1) enhanced travel choices with public transportation

alternatives; (2) improved mobility (especially for the handi-

capped and low-income citizens); (3) enhanced living environ-

ments with less traffic congestion and reduced CO

2

emission; (4)

a greater opportunity for advancing transportation technologies

such as the use of biofuel for transit vehicles; (5) increased traf-

fic safety with less accidental risks as compared to private trans-

portation; and (6) stimulus for local economic development (Min

et al. 2015). In particular, capital investment in public infrastruc-

ture such as mass transit systems is often linked with regional

economic development. For instance, based on the review of aca-

demic literature on economic benefits of public infrastructure,

Bhatta and Drennan (2003) observed that such investment tended

to yield long-term economic benefits such as higher residential

property value, higher real wages for local workers, greater job

opportunities, and reduced travel time. Similar conclusions are

drawn from the more recent studies of Africa’s transportation

infrastructure (e.g., Boopen 2006) and China’s transportation

infrastructure (e.g., Zhou et al. 2007). Overall, investment in

mass transit can yield 50,731 jobs per $1 billion invested and

sustained higher investment in mass transit can create a total eco-

nomic value of $3.7 billion per $1 billion invested annually

(American Public Transportation 2014b); especially, the eco-

nomic impact of a mass transit system on America’s poor-

income families is known to be greater because the mass transit

system is an important means of accessing an affordable trans-

portation option (Moulding 2005).

On the other hand, a mass transit system can increase financial

burden for municipal, state, and federal governments. According

to the American Public Transportation Association (2013), the

Corresponding author:

Hokey Min, Department of Management, BAA 3008C, College of

Business Administration, Bowling Green State University, Bowling

Green, OH 43403, USA; E-mail: hmin@bgsu.edu

Journal of Business Logistics, 2017, 38(1): 55–73 doi: 10.1111/jbl.12148

© Council of Supply Chain Management Professionals

U.S. mass transit system consumed $56 billion for operation,

maintenance, and capital investment in 2010. Controlling mass

transit operating costs as well as meeting service demand

remains the greatest challenge for mass transit authorities, private

transit service providers, and public policy makers (Cervero

2004; Savage 2004; Polzin and Chu 2005).

Considering the significant impact of mass transit systems on

public well-being, economic development, and government

finances, a growing number of local and state government offi-

cials have attempted to find ways to improve mass transit ser-

vices, while better utilizing resources (e.g., drivers, dispatchers,

maintenance crews, vehicles, equipment, depots) required for

mass transit services under tight budget constraints. These

attempts include the assessment of the recent performances of

mass transit systems across the United States (including Puerto

Rico) in terms of their operating and financial efficiency (e.g., a

greater amount of revenue sources, a greater access to mass tran-

sit services, better utilization of assets and financial resources

including tax dollars). Because mass transit operating efficiency

may hinge on the community setting (i.e., the density of housing

development, urban sprawl) and municipal size, a majority of the

published literature regarding public services has focused on the

discussions of appropriate municipal size and its potential impact

on the efficiency of public services such as mass transit services

(Kain 1967; Real Estate Research Corporation 1974; Ladd 1992,

1994; Rosen 1992; Carruthers and Ulfarsson 2003; Moore et al.

2005; Garcia-Sanchez 2006; O’Sullivan 2007). For example,

some of these prior studies attempted to verify the theory that in

densely populated urban areas, the distances transit vehicles must

travel were short, but heavy traffic could cause delays and subse-

quently could undermine transit efficiency.

In contrast with the large urban metropolitan setting, sparsely

populated suburban areas pose challenges for offering adequate

mass transit services because dispersed populations require much

greater vehicle service hours than in dense urban areas. Also,

limited financial resources, communication gaps, and a lack of

skilled drivers in suburban or satellite city areas may compound

the problem of delivering mass transit services to their residents

(O’Sullivan 2007; Min and Lambert 2010). Thus, the small satel-

lite city setting can adversely influence the efficiency of mass

transit services.

PRIOR LITERATURE

Despite a significance of mass transit systems to our daily lives

and regional economic development, the published literature

evaluating the efficiency of mass transit systems has been scant.

However, some pioneering attempts have been made to assess

or improve the efficiency of mass transit services from opera-

tional and financial perspectives. Examples of such attempts

include Ball et al. (1983) who proposed a match-based heuris-

tics to schedule vehicles and drivers simultaneously to improve

the cost efficiency of the Baltimore Metropolitan Transit

authority. Extending the work of Ball et al. (1983), Haase et al.

(2001) developed both the exact algorithm built upon the

branch-and-bound method and the heuristic version of a set-

partitioning algorithm to solve the complex problem of scheduling

mass transit vehicles and their crews simultaneously. Their

study improved the solution accuracy as compared to Ball

et al.’s study (1983). Likewise, a vast majority of the existing

literature focused on the development of analytical tools/meth-

ods intended for the better utilization of transit vehicles, drivers,

and/or other resources (including maintenance crews and capital

resources).

Narrowing down the scope of the mass transit system to a

paratransit system, Bower (1991) investigated the impact of an

automated paratransit routing and scheduling system called

COMSIS on the operating cost and service quality of paratransit

services. As expected, COMSIS turned out to be useful for

reducing scheduling errors, reducing the cost of generating

schedules, and identifying traffic patterns. Thus, Bower (1991)

concluded that COMSIS improved the overall efficiency of para-

transit service quality. Similarly, Chira-Chavala and Venter

(1997) analyzed the impact of automated vehicles and passenger

scheduling methods on the operating costs of paratransit systems.

They found that such methods saved unit paratransit transporta-

tion cost by 13%. Further extending the earlier works of Chira-

Chavala and Venter (1997), Pagano et al. (2002) assessed the

impact of the computer-assisted scheduling and dispatching

(CASD) systems on the service quality of paratransit systems in

central Illinois. They found that CASD systems allowed passen-

gers to experience less riding time and greater on-time services

at both pickups and drop-offs and subsequently enhanced their

overall satisfaction with paratransit services. On the other hand,

the use of CASD to promote higher vehicle productivity led to

slightly longer ride times, while callers to the system had to wait

longer for responses. Overall, they concluded that the quality of

service, which was considered one of the transit efficiency indi-

cators (e.g., Vuchic 2005), was positively affected by the imple-

mentation of the CASD system.

Rather than dealing with the mass transit routing and schedul-

ing issues, other earlier studies focused on the assessment of the

efficiency and effectiveness of mass transit services from a finan-

cial or administrative perspective. For instance, Jackson (1982)

compared the real costs of service provided by subsidized mass

transit (especially paratransit) operations to those of private-

sector run operations in the New England region. He discovered

that cost figures per passenger trip by nonprofit and publicly

owned mass transit services were seriously underestimated and

did not truly reflect the actual costs or the cost efficiency of mass

transit services provided. The study by Nolan et al. (2001) was

one of the first to propose a data envelopment analysis (DEA) to

measure the comparative operating efficiency of the 25 selected

mass transit systems in the United States. They also identified

various factors influencing mass transit efficiency using the Tobit

regression analysis. Their study found that average fleet age

adversely affected transit efficiency and federal subsidies under-

mined transit efficiency, whereas local-based subsidies had a

positive impact on transit efficiency.

Following suit, Fu et al. (2007) evaluated efficiency levels of

individual paratransit systems in Canada with the specific objec-

tive of identifying the most efficient paratransit systems and the

sources of their efficiency using DEA. Through identification of

the most efficient systems along with the key influencing factors,

they developed new paratransit service policies and operational

strategies for enhanced resource utilization and quality of ser-

vices. Their study is one of the few that have attempted to

56 H. Min and Y.-H. Ahn