Developing and Testing a Dynamic Theory of Motor Carrier Safety Jason W. Miller 1 , Susan L. Golicic 2 , and Brian S. Fugate 3 1 Michigan State University 2 Colorado State University 3 University of Arkansas D eveloping an understanding of the longitudinal relationships between different measures of motor carrier safety is important to advance theory and practice regarding this significant supply chain management and public policy issue. In this article, we combine core principles from several theoretical traditions to propose a dynamic theory of motor carrier safety that specifies the longitudinal relationships between three core measures of safety publically reported by the Federal Motor Carrier Safety Administration: unsafe driving, hour-of-service compliance, and vehicle maintenance. We test this theory using four years of longitudinal data on motor carrier safety for a random sample of large, for-hire motor carriers. Results from fitting a vector multivariate autoregressive moving average time-series model are largely consistent with the theory we propose. We describe the implications of our research for supply chain management theory and practice, summarize limitations, and suggest directions for future research. Keywords: transportation; motor carrier; safety; longitudinal structural equation modeling; time series; panel data INTRODUCTION Studying dynamic relationships is important in the continually evolving logistics and supply chain management (L&SCM) disci- pline. For instance, consider the topic of motor carrier safety. Motor carrier managers and their drivers, either willfully or through negligence, take many actions that influence the propen- sity of drivers to be involved in accidents, which affect not only shippers’ and consignees’ supply chain operations, but also soci- ety at large. These actions include operating trucks in a danger- ous manner (e.g., texting while driving), violating hour-of- service (HOS) rules (e.g., driving when excessively fatigued), or failing to properly maintain vehicles (e.g., operating a truck with defective brakes or worn tires; The Volpe Center 2013). Examin- ing how these actions unfold over a multiyear time period as a dynamic system provides both theoretical and practical insight and extends previous work that has studied how characteristics of carrier’s drivers (Monaco and Williams 2000) and operations (Mejza et al. 2003) affect carrier safety. Despite the value of such longitudinal investigations (McArdle and Nesselroade 2014), this type of analysis is rare in L&SCM (cf., Autry and Golicic 2010) and, to the best of our knowledge, has never been conducted with the goal of understanding how various actions linked to motor carrier accidents unfold as a dynamic system. This article addresses the question, “what are the longitudinal relationships between three core measures of motor carrier safety: unsafe driving, HOS compliance, and vehicle maintenance?” We focus on these three measures given that carriers with poor per- formance in these three areas have been found to have higher accident rates in independent studies conducted by the Federal Motor Carrier Safety Administration (FMCSA; The Volpe Center 2014) and the American Trucking Research Institute (ATRI; 2012). In answering this question, we propose a dynamic theory of motor carrier safety that details the underlying processes (Whetten 1989; Sutton and Staw 1995) for why unsafe driving, HOS compliance, and vehicle maintenance should be longitudi- nally linked. Beyond offering explanations for why these mea- sures should be interrelated, our theory specifies the magnitude of the autoregressive relationships for each safety measure. The logic for this dynamic theory is drawn from multiple sources including strain theory (Agnew 1992, 2014), behavioral theory of the firm (Cyert and March 1992; Gavetti et al. 2012), organi- zation turnover theory (Heavey et al. 2013; Park and Shaw 2013), and theories from the medical and safety science literature concerning the effect of fatigue on driving (Lal and Craig 2001; Ting et al. 2008). This synthesis of theory responds to calls for theoretical pluralism when studying complex L&SCM phenom- ena that cut across multiple disciplines (Sanders and Wagner 2011) to develop middle-range theories (Merton 1968) specific to the L&SCM domain (Holmstr€ om et al. 2009). We subject our theory to testing using four years of longitudinal data obtained from the FMCSA on carrier unsafe driving, HOS compliance, and vehicle maintenance using a multivariate time-series model estimated in the covariance structural equation modeling (SEM) framework (du Toit and Browne 2007). Our results are largely consistent with our dynamic theory and indicate a complex series of interrelationships between these three safety measures. For example, we find carriers that are less compliant with HOS rules at time t are predicted to have higher levels of reckless driving at time t + 1, but carriers that have higher levels of reckless driving at time t are predicted to have lower levels of HOS violations at time t + 1. We evaluate the robustness of our findings by incor- porating key covariates previously linked to motor carrier safety. We further examine the generalizability of our model through a cross-validation analysis (Cudeck and Browne 1983; Browne 2000) performed on an independent sample of carriers. This article makes several contributions to the extant literature. First, this research extends prior work that has examined factors that predict motor carrier safety, and in particular accidents (e.g., Cantor et al. 2010), by investigating the longitudinal relation- ships between safety measures that are proximal to accidents (ATRI 2012; The Volpe Center 2014). Studying how these Corresponding author: Jason W. Miller, Department of Supply Chain Management, Michi- gan State University, 370 North Business Complex, 632 Bogue Street, East Lansing, MI 48824, USA; E-mail: mill2831@msu.edu Journal of Business Logistics, 2017, 1–19 doi: 10.1111/jbl.12149 © Council of Supply Chain Management Professionals