A New Perspective on Returns to Scale for Truckload Motor Carriers

• Date: 01 September 2020
• Published date: 01 September 2020
• Author: Jason W. Miller,William A. Muir
• DOI: http://doi.org/10.1111/jbl.12234

A New Perspective on Returns to Scale for Truckload Motor

Carriers

Jason W. Miller

1

and William A. Muir

2

1

Michigan State University

2

Naval Postgraduate School

Understanding how motor carriers’size affects their productivity (e.g., miles per power unit) is of fundamental importance to carrier man-

agers, shippers, and investors, because the nature of this relationship should influence carriers’strategies with regard to growth. In the

truckload (TL) sector, the prevailing assumption is that TL carriers face constant returns to scale such that productivity differs little between

large and small carriers. While empirical findings from several studies conducted since deregulation are consistent with this assumption, we con-

tend that the true relationship between carrier size and productivity is more nuanced and is contingent on carriers’level of technical efficiency.

Specifically, we develop and test middle range theory that predicts increasing returns to scale for carriers with low technical efficiency, constant

returns to scale for carriers with average technical efficiency, and decreasing returns to scale for carriers with high technical efficiency. We test

our theory by estimating production functions using quantile regression for data collected from the U.S. Department of Transportation for 1,068

TL carriers. Results from our analyses corroborate our predictions. Our findings hold valuable implications for the logistics literature as well as

TL carrier management, shippers, and other industry stakeholders.

Keywords: motor carrier; productivity; quantile regression; transportation

INTRODUCTION

Questions regarding the relationship between full truckload (TL)

motor carrier size and productivity (e.g., miles per power unit)

are of intense interest to logistics scholars and practitioners. Car-

riers evaluating whether to add capacity want to understand the

likely impact on productivity (Stephenson and Stank 1994) and,

ultimately, firm competiveness and survival (Fawcett et al.

2016). Investors speculate on how proposed mergers, divesti-

tures, and acquisitions that alter TL carrier size may influence

future financial performance, since firm productivity correlates

with profitability (Foster et al. 2008; Braguinsky et al. 2015) and

survival (Syverson 2011). This is made all the more important

by the fact that carriers’ability to increase size in the short run

is constrained by their ability to recruit and train new drivers as

well as purchase equipment (Bowman 2016). Furthermore,

reducing size is equally as difficult because carriers must trim

their workforce and, if they are vertically integrated, find buyers

for used trucks (Bowman 2016). Consequently, decisions that

affect carriers’size have long-term implications. It is for these

reasons that research into TL motor carrier size and productivity

remains of great interest to academics and is of relevance and

value to practitioners.

The literature examining productivity in the TL sector—the

largest sector of the motor carrier industry based on revenue and

number of firms (Corsi 2005)—has traditionally assumed con-

stant returns to scale (McMullen and Stanley 1988; Grimm et al.

1989; McMullen and Tanaka 1995; McMullen 2005). In this

view, larger size is assumed to offer TL carriers no discernible

benefit with regard to the productivity of their resources (e.g.,

drivers and trucks). However, several arguments exist as to why

larger TL carriers should be more productive than their smaller

counterparts. These include larger TL carriers (1) having more

ability to leverage economies of density (Powell 1996), (2) being

able to serve large-volume shippers with stable demand (Keeler

1989), (3) having the scale to justify investments in specialized

technology or personnel (Galbraith 1977; Viscelli 2016), and (4)

having reduced exposure to stochastic fluctuations in demand

(Boyer 1993). Additional doubt has been cast by a string of car-

rier mergers and acquisitions in the TL sector, such as the mer-

ger of Swift and Knight (Gensler 2017). Wilson Logistics

acquired Haney Truck Line in 2017 (Bearth 2017). Heartland

Express acquired Interstate Distributor Co. in 2017, with the

CEO of Heartland Express noting one reason for the decisions

was that “Heartland will gain significant additional traffic density

in the West, and our stronger eastern network will improve ser-

vice for IDC’s customers in the East”[emphasis added] (Ashe

2017). Given these anomalies between theory, empirics, and

industry practice, it is important to revisit prior assumptions that

bear on the issue (Laudan 1977).

Accordingly, in this manuscript, we present a challenge to the

prevailing assumption that TL carriers universally face constant

returns to scale with regard to productivity. Instead, our study

presents a more nuanced investigation into the size–productivity

relationship: At a given size, a TL carrier may face constant,

increasing, or decreasing returns to scale depending on how effi-

ciently it is able to transform the services of productive resources

into outputs relative to peers (i.e., its technical efficiency) (Coelli

et al. 2005). To do so, we evaluate whether returns to scale with

regard to productivity are conditional on carriers’technical effi-

ciency using quantile regression (Koenker and Hallock 2001).

This approach aligns our research with calls for scholars to eval-

uate whether production functions are contingent on firms’levels

of technical efficiency (Bernini et al. 2004; Behr 2010). We

Corresponding author:

Jason W. Miller, Department of Supply Chain Management, Eli

Broad College of Business, Michigan State University, 632 Bogue

Street, East Lansing, MI N370, USA; E-mail: mill2831@msu.edu

Disclaimer: The views presented are those of the authors and do not

necessarily represent the views of the U.S. Department of Defense

or its components.

Journal of Business Logistics, 2020, 41(3): 236–258 doi: 10.1111/jbl.12234

© 2020 Council of Supply Chain Management Professionals

devise middle range theory by incorporating contextual features

of TL load assignment (Powell et al. 1988), arguments as to why

returns to scale may exist in the TL sector (Powell 1996; Corsi

2005), and information processing theory (IPT) (Galbraith 1977;

Tushman and Nadler 1978). In doing so, we theorize that

increasing returns to scale exist for technically inefficient carriers,

constant returns to scale exist for carriers with average technical

efficiency, and decreasing returns to scale exist for technically

efficient carriers.

To test our theory, we collect archival data from the Depart-

ment of Transportation concerning the productivity of over 1,000

truckload motor carriers. We test our moderation hypothesis

using quantile regression (Koenker and Bassett 1978; Koenker

and Hallock 2001) to estimate how the elasticity of carriers’

vehicle miles traveled (output) relative to carrier size (measured

as power units as well as power units and drivers)—which cap-

tures productivity—varies across models estimated at different

quantiles (Behr 2010). This approach for testing moderation

hypotheses, discussed by Li (2015), complements the commonly

seen approach of including product terms in regression models

(Goldsby et al. 2013) and, critical to this inquiry, allows us to

answer a research question that could not be investigated using

other approaches. The results corroborate our theory and have

important managerial implications.

This research makes several contributions. First, we draw on

IPT to offer a new explanation for returns to scale with regard to

productivity in the TL sector. In doing so, we enrich the litera-

ture by offering a new explanation—a theoretical contribution

per Makadok et al. (2018)—that reconciles inconsistencies

between theory and prior empirics and, furthermore, has greater

consilience (Thagard 1978) than prior accounts in that our theory

can explain a wider array of findings. Simultaneously, we enrich

IPT by demonstrating how this theory’s central elements (e.g.,

assumptions, mechanisms), with appropriate contextualization,

can accommodate a wider array of empirical findings. As noted

by M€

aki (2001), this represents one of the key ways scholars

make theoretical contributions when working with general theo-

ries like IPT. Third, to the best of our knowledge, this is the first

manuscript in Journal of Business Logistics to utilize quantile

regression. As noted by Li (2015), quantile regression is a pow-

erful tool that expands the scope of research questions that can

be econometrically examined, which can contribute to the devel-

opment and testing of new theory with important practical impli-

cations.

The remainder of this manuscript is structured into five sec-

tions. The next section reviews the relevant literature. This is fol-

lowed by the theory and hypothesis development. We then

explain our research design and summarize our measures. The

penultimate section summarizes our econometric approach, pre-

sents the results, and describes robustness tests. The final section

explains theoretical contributions, details managerial implications,

notes limitations, and suggests directions for future research.

LITERATURE REVIEW

We begin by delineating the distinction between productivity and

technical efficiency. Productivity is an absolute measure captur-

ing the ratio of outputs to inputs in a production setting.

Equivalently, productivity can be discussed as the elasticity of an

output with respect to one or more inputs (Coelli et al. 2005). If

output elasticity exceeds one (i.e., a one percent increase in pro-

ductive inputs results in a greater-than-one percent increase in

output), there exist increasing returns to scale; conversely, if out-

put elasticity falls below one, there are decreasing returns to

scale. Constant returns to scale exist when this elasticity does not

differ from one. In contrast, technical efficiency is a relative mea-

sure and refers to a decision-making unit’s (DMU) ability to gen-

erate output with a given set of inputs relative to the maximum

output that another DMU might generate given that set of inputs

(Greene 2008). Thus, technical efficiency relates, by way of dis-

tance, the level of output for a firm to some level on a produc-

tion possibilities frontier, where firms existing on the frontier are

considered to be perfectly efficient and firms below the frontier

are considered to be technically inefficient by some degree. Tra-

ditionally, the concepts of “returns to scale”and “scale econo-

mies”have referred the shape of the production frontier (Schmidt

and Lovell 1979); however, as we will later discuss, transporta-

tion researchers have overwhelmingly estimated elasticities at the

conditional mean, which makes a strong assumption that the

average production function is a neutral shift of the frontier.

Lastly, as noted by Muir et al. (2019), when there are increasing

returns to scale with regard to productivity, it is possible for one

DMU (Firm A) that is larger in size to be more productive than

another DMU (Firm B) that is smaller in size but the latter DMU

(Firm B) may be more technically efficient, illustrating that these

concepts are distinct.

With these definitions in mind, we turn to the literature. A

large body of research exists concerning the nature of returns to

scale and, by duality, economies of scale within the motor carrier

industry. This research stream can be segmented within many

facets, including sampling frame (e.g., TL or LTL carriers),

research methodology (e.g., production function, cost function),

and regulatory regime—most notably, whether studies were con-

ducted prior to deregulation by the Motor Carrier Act of 1980

(Roberts 1956; Emery 1965; Ladenson and Stoga 1974; Koenker

1977; Rakowski 1978; Harmatuck 1981; Sugrue et al. 1982;

Friedlaender and Chaing 1983) or after regulatory reform

(McMullen and Stanley 1988; Grimm et al. 1989; Ying 1990;

Xu et al. 1994; McMullen and Tanaka 1995; McMullen and Lee

1999). To improve manageability of our review, we limit our

focus to studies that (1) were conducted after deregulation and

(2) examine directly or via duality the relationship between car-

rier size and productivity. This latter criterion results in us

excluding several studies that investigate technological change

and/or temporal shifts in industry-level productivity (e.g.,

McMullen 2004; Corsi 2005; Scheraga 2011); such investiga-

tions do not address our research question of interest. Given the

large number of studies resulting from our literature search, we

summarize relevant study characteristics in Table 1.

We wish to call attention to four findings from Table 1. First,

scholars have devoted more energy toward examining carrier

productivity and efficiency in the LTL sector (Harmatuck 1991,

1992; Allen and Liu 1995; Nebesky et al. 1995; Giordano 1997,

2008; McMullen and Lee 1999). This is understandable given

the worry that the higher fixed-cost nature of LTL operations

could give rise to economies of scale, resulting in industry con-

solidation (Allen and Liu 1995). However, LTL-specific studies

237