Effects of Motor Carriers’ Growth or Contraction on Safety: A Multiyear Panel Analysis

• DOI: http://doi.org/10.1111/jbl.12178
• Author: Matthew A. Schwieterman,Yemisi A. Bolumole,Jason W. Miller
• Date: 01 June 2018
• Published date: 01 June 2018

Effects of Motor Carriers’Growth or Contraction on Safety:

A Multiyear Panel Analysis

Jason W. Miller, Matthew A. Schwieterman, and Yemisi A. Bolumole

Michigan State University

Motor carrier safety remains a highly relevant issue for supply chain managers and scholars because carriers’safety affects supply chains

as well as the welfare of the motoring public. This article enriches understanding regarding this topic by investigating how motor carri-

ers’growth or contraction since the start of the Compliance, Safety, and Accountability (CSA) program in 2010 affects their safety perfor-

mance. Drawing on core principles from theories regarding internal adjustment costs from economics and nonscale free capabilities from

management, we explain why carriers’growth or contraction should differentially affect various safety metrics tracked by the CSA program. To

test our theory, we assemble a multiyear panel data set for over 1,000 of the largest for-hire motor carriers operating in the United States by

melding together several different governmental data sources. We fit a series of multivariate seemingly unrelated regression models to test our

hypothesized effects. Our results corroborate our theorized predictions and are robust to alternative model specifications. We conclude by detail-

ing how this work contributes to extant theory, summarizing managerial and policy implications, highlighting limitations, and suggesting direc-

tions for further pursuit.

Keywords: motor carrier; safety; seemingly unrelated regression; firm growth; panel data

INTRODUCTION

Firms’safety performance is a topic of great importance to logis-

tics and supply chain management (L&SCM) scholars given that

firms’safety affects multiple stakeholders including employees,

customers, stockholders, and the general public (Brown 1996;

Cantor 2008). One setting where safety is of paramount impor-

tance is the motor carrier industry (Viscelli 2016). Although

fatality rates from accidents involving large commercial trucks

have declined since the 1970s, motor carrier operations still have

a tremendous impact on societal welfare. Recent statistics from

the Department of Transportation (DOT) for 2014 reveal there

were 3,424 fatal accidents involving large commercial trucks that

claimed 3,903 lives (Federal Motor Carrier Safety Administration

[FMCSA] 2016). Moreover, the fatal accident rates have risen

from a low of 1.29 per 100 million vehicle miles in 2010 to 1.40

in 2014 (FMCSA 2016). As such, although motor carrier safety

has been studied for many decades, it remains a highly relevant

topic for L&SCM scholars today (Miller 2017a).

Researchers have devoted much attention toward identifying car-

rier-level characteristics thatpredict carrier-level safety (Cantor et al.

2009, 2017; Britto et al. 2010). To date, the vast majority of these

studies focus on understanding the effects of staticpredictors, which

capture carriers’standing on a characteristic at a given point intime.

Examples of static predictors of carrier-level safety include driver

turnover rates (Corsi and Fanara 1988; Shaw et al. 2005), carrier

size (Cantor et al. 2016; Miller 2017b), age (Corsi and Fanara

1989), and financial performance (Corsi et al. 1988; Bruning 1989;

Britto et al. 2010). Comparatively, scholars have expended limited

energy toward understanding the effects of dynamic predictors,

which capture carriers’change on a characteristic over a period of

time (cf, Miller and Saldanha 2016). As carriers must continually

evolve to remain competitive by altering characteristics such as

financial strategy (Corsi and Scheraga 1989; Scheraga et al. 1994),

degree of less-than-truckload (LTL) versus truckload (TL) focus

(Feitler et al. 1997; Scheraga 2011), tractor-to-trailer ratio (Jin et al.

2017), and balance of owner-operators versus employee drivers

(Corsi and Grimm 1987; Nickerson and Silverman 2003), devising

and testing theory regarding dynamic predictors of carrier-level

safety providesan avenue for extending extant theory.

This article seeks to further knowledge regarding dynamic predic-

tors of carrier safety by examining how changes in carriers’size

over a multiyear period affect carrier-level safety. We focus on

changes in carriers’size—growth or contraction—given that the

decisions managers make to increase or decrease fleet size represent

a key facet of carriers’strategic positioning (Feitler et al. 1997,

1998; Scheraga 2011). Focusing on changes in size allows us to

draw on theory in economics concerning internal adjustment costs

(Hamermesh and Pfann 1996). This, coupled with management the-

ories concerning nonscale free capabilities (Penrose 2009; Levinthal

and Wu 2010; Knudsen et al. 2014) and organizational decline

(Cameron et al. 1987), allows us to devise middle-range theory

(Merton 1968; Stank et al. 2017) that postulates asymmetric effects

of carriers’growth or contraction on different dimensions of safety

tracked by the FMCSA as part of the Compliance, Safety, and

Accountability (CSA) program. In particular, we explain why carri-

ers’growth will have a more detrimental effect on their driving

safety than their vehicle maintenance. Conversely, we explain why

contraction will have a more detrimental effect on vehicle mainte-

nance than drivingsafety. To test our theory, we assemble a six-year

panel data set from December 2010 till December 2016 for 1,127

large, for-hire motorcarriers operating in the United States. We con-

duct our econometric analysis via simultaneous estimation of a sys-

tem of equations to compare the magnitude of regression

coefficients across safety dimensions (Greene 2012). We further

conduct a series of robustness tests to evaluate the stability of our

results to alternative modeling assumptions. Results from these anal-

yses corroborate o ur theoretical predictions.

Corresponding author:

Jason W. Miller, Eli Broad College of Business, Michigan State

University, 632 Bogue Street N370, East Lansing, MI 48824, USA;

E-mail: mill2831@msu.edu

Journal of Business Logistics, 2018, 39(2): 138–156 doi: 10.1111/jbl.12178

© Council of Supply Chain Management Professionals

The remainder of this article is structured in five sections. The

first section summarizes the pertinent literature and explains how

this research expands upon these works. The second section

devises the underlying logic for our hypothesized predictions.

The third section explains the data collection protocol, variables,

and data handling procedures. The penultimate section explains

the econometric modeling approach, presents results, and details

various robustness tests. The final section explains theoretical

contributions, describes managerial and policy implications, notes

limitations, and suggests directions for future work.

LITERATURE REVIEW

We organize our literature review around three themes to provide

a basis for framing how this research contributes to the literature.

These themes encompass studies that (1) examine factors affect-

ing carrier-level safety, (2) analyze changes in motor carriers’

size, and (3) investigate consequences of firms’growth or con-

traction outside the motor carrier industry.

Carrier-level safety

Given the current state of research on motor carrier safety, we

focus our attention on studies that have examined carrier-level

factors that predict why some carriers are safer than others. Lim-

iting our focus to carrier-level predictors of safety follows Mill-

er’s (2017a) categorization of these predictors into studies that

examine driver-, haul-, and carrier-level characteristics. More-

over, as changes in carriers’size occur at the carrier level, limit-

ing our attention in this manner is justifiable.

As noted in the prior section, studies examining the relations

between carrier-level variables and carriers’safety have, by and

large, emphasized the role of static predictors. These include car-

riers’use of electronic logbooks (Cantor et al. 2009), carrier age

(Corsi and Fanara 1989; Cantor et al. 2017), carrier size (Cantor

et al. 2016), financial performance (Corsi et al. 1988; Bruning

1989; Beard 1992; Naveh and Marcus 2007; Britto et al. 2010),

unionization (Corsi et al. 2012), and driver turnover (Corsi and

Fanara 1988; Shaw et al. 2005; Miller et al. 2017c).

In contrast, far less research has examined the effect of

dynamic predictors on carrier safety. Two exceptions are Hunter

and Mangum (1995) and Miller and Saldanha (2016). Hunter and

Mangum (1995) examined how carriers’year-over-year change in

revenue per mile and net income percent affected their accident

rates for two cross-sectional samples of carriers in 1976 and

1986. Miller and Saldanha (2016) studied how carriers’average

change in earnings before interest, taxes, depreciation, and amor-

tization over sales for 2009–12 explained between-carrier differ-

ences in safety for 2010–13. This study builds upon these two

articles by investigating the consequences of another dynamic

predictor—changes in carriers’size—on carriers’safety perfor-

mance.

Changes in carriers’size

Scholars have undertaken investigations to understand changes in

carriers’size, albeit for different purposes. Several studies have

treated changes in carriers’size as the outcome variable of

interest. Rakowski (1988, 1994) and Kling (1988, 1990) docu-

ment how the largest LTL carriers experienced dramatic

increases in market share following the passage of the Motor

Carrier Act in 1980. They attributed these findings to large LTL

carriers being able to leverage economies of size. Feitler et al.

(1998) corroborate these findings by noting that LTL carriers

grew from an average size of 561 employees in 1976 to 1,922

employees by 1993. Pettus (2001) studies 59 LTL carriers to

examine how carriers’pattern of actions following deregulation

affected their growth in sales, employees, and total assets. Gior-

dano (2008) utilizes data from 392 TL carriers from 1981, 1991,

and 2001 to test Gibrat’s law of proportionate growth (Sutton

1997) using carriers’reported ton-miles as a measure of carriers’

size and finds evidence inconsistent with predictions implied by

this theory.

1

Giordano (2014) conducts similar analyses using

data on 244 LTL carriers over this same horizon and again finds

evidence inconsistent with Gibrat’s law. Feitler et al. (1997) treat

changes in carriers’revenue as one dimension of a broader con-

struct representing strategic change and examine antecedents of

strategic change for LTL carriers over an 18-year horizon span-

ning 1976–93.

A smaller number of studies have treated changes in carriers’

size as a predictor variable. Feitler et al. (1998), again treating

changes in carriers’size as one dimension of strategic change,

test whether LTL carriers’strategic change is related to their

financial performance. Scheraga (2011), using data from 69

general freight carriers for 1999–2003, finds that carriers that

pursued a strategy focused on growth saw a decline in operat-

ing efficiency but an increase in their technical innovation.

2

Fawcett et al. (2016) includes carrier growth, measured as quar-

ter-over-quarter change in sales, in their panel models for 41

publicly traded motor carriers from 1998 to 2014. They find

that while growth is positively related to return on invested

capital, it is negatively related to return on assets and operating

margin.

3

This article builds on these two sets of studies by

extending understanding regarding the performance conse-

quences of changes in carriers’size to a new domain. More-

over, this research uses a measurement approach for size

changes that (1) focuses on carriers’owned and leased physical

assets and (2) considers size changes over a multiyear horizon

to reduce concerns regarding the random noise associated with

studying size changes over short time windows (Weinzimmer

et al. 1998).

1

Gibrat’s law postulates that there is no correlation between firms’size at

time t

0

and firms’subsequent growth rates over a long time horizon (i.e., t

0

to t

n

where nis many years in the future). If this law holds, we would expect

to see (1) no correlation between firms’size at t

0

and their subsequent growth

rates where growth rates are couched on percent terms, (2) homogenous

growth rates across different size classes at time t

0

, (3) the distribution of

firm size to be lognormal, and (4) the variance of firm size should increase

over time (Giordano 2014).

2

To clarify the meaning of this terminology, Scheraga (2011) constructs a

Malmquist Productivity Index for these firms. In this type of index, efficiency

is measured as distance from an efficiency frontier, whereas technical innova-

tion represents a shift in the efficiency frontier.

3

Given their use of firm fixed-effects, these results are more properly

interpreted as changes in a carrier’s growth relative to its average growth rate

(Certo et al. 2017).

Growth and Motor Carrier Safety 139