Benchmarking road safety of U.S. states: A DEA-based Malmquist productivity index approach
Highlights
► In this paper, U.S. highway safety efficiency is studied considering 50 states. ► A DEA based Malmquist index model is developed. ► The relative efficiencies U.S. states in terms of road safety are analyzed. ► Results indicate that, more than half U.S. states are still not efficient in utilizing societal and economical resources towards the goal of zero crash.
Introduction
In the United States, federal and state governments utilize individual policy-driven procedures and strategies to reduce crash frequency and societal cost. However, the number of crashes has not fallen below 30,000 in the last 50 years, even though a substantial effort toward improving highway safety through the reduction of transportation-related fatalities has been undertaken by agencies and government organizations. The vast majority of the funding is available through the authorization of the federal surface transportation program; currently SAFETEA-LU (the Safe, Accountable, Flexible and Efficient Transportation Equity Act: A Legacy for Users). In return, State Highway Safety Offices administer a variety of highway safety grant programs designated in the federal transportation authorization. An example of the programs and the federal funding available includes, but is not limited to, the following: State Highway Safety Grant Programs, $234.8 million; Occupant Protection Incentive, $25 million; Child Safety and Child Booster Seat Incentive, $7 million.
Contributing to the significant number of crashes are rising travel demand, annual vehicle miles of travel, inexperienced or young drivers along the road network and older drivers with slower reaction times. Several methods have been employed to assess highway safety and assist policy makers in their decision making processes. Data envelopment analysis (DEA) is one of the methods that has recently been used to analyze highway safety efficiency (Hermans et al., 2008). DEA is a very robust tool which compares similar elements based upon various input(s) and produces a desired output(s). The robustness of DEA comes from its wide applicability to various problems. DEA has been widely applied to several problems from various fields such as economics (Zhang et al., 2008, Filippetti and Peyrache, 2011), sustainability (Lee and Farzipoor Saen, 2012), finance (Kirikal and Tehnikaülikool, 2005), healthcare (Feng and Yuan-Biao, 2010), construction (Xue et al., 2008), transportation (Ozbek et al., 2009, Cooper et al., 2011), etc.
With regard to highway safety, Odeck (2001) analyzed the efficiency of the Norwegian road network using two competing methods: a data envelopment analysis and a deterministic frontier analysis. In another study, the efficiency of targeted operational achievements of the Norwegian Public Roads Administration was investigated with DEA and Malmquist Indices (Odeck, 2006). A bootstrapping method was also utilized to determine confidence intervals for efficiency scores and test hypotheses regarding productivity growth. To provide a better indication of highway safety in different regions or countries, a comparison was made with respect to safety performance indicators (SPIs). The properties of SPIs were studied in terms of their relationship to the outcome of highway crash fatalities (Tingvall et al., 2010). The assumption of linearity between SPIs and a final outcome was partly rejected in these studies. In addition, the availability of data largely limited the scope of the comparison.
While evaluating highway safety, there are several indicators, including, but not limited to, infrastructure characteristics, vehicle miles traveled, alcohol involvement and safety belt usage, which need to be considered. To enable an overall comparison of different countries or regions, a single highway safety index is required (Hermans et al., 2009). To be able to have single highway safety index, the creation of a combination of relevant highway safety indicators is then required. Prior to defining the index variable, a weighting strategy must be employed to appropriately reflect the relative importance of such safety indicators. Even though several studies have been conducted, there is neither a widely accepted nor a reasonable method for the assignment of weights (Tatari and Kurmapu, 2011). For these reasons, the DEA methodology provides an acceptable alternative solution for highway safety efficiency determination. DEA provides a theoretical framework, does not require a weighting procedure, and rates the highest among five multi-criteria evaluation methods (Hermans et al., 2008). In a latter work, Hermans et al. (2009) compared roadway safety of 25 European countries. Several SPIs were considered such as alcohol and drug use, vehicle protective systems, vehicle characteristics, infrastructure quality, and trauma management. The comparison was completed by using a non-bias weighting procedure incorporated into a linear programming optimization-based efficiency analysis model, so called “single-period output-oriented DEA”.
Even though, road safety performance assessment is a crucial concept and significant research has been conducted for European countries from a holistic perspective, to the best knowledge of authors U.S. states road safety assessment has not been addressed in literature. The overarching goal of this study is to develop an analytical tool that can be used to analyze and compare the road safety performance of the U.S. states. To assess and benchmark U.S. states road safety, a well-known linear programming-based method, data envelopment analysis (DEA), is utilized. The rest of the paper is as follows. In Section 2, basic explanation about data envelopment analysis, productivity and Malmquist index concepts are provided. In Section 3, research aims are explained. Study methods are described in Section 5. Section 6 provides insights about the experimentation and results. Concluding remarks and future direction of research are mentioned in Section 7.
Section snippets
Data envelopment analysis (DEA), productivity and Malmquist index
Data envelopment analysis is a linear programming and production theory-based mathematical approach developed by Charnes et al. (1978). In a typical DEA model, the objective is to compare similar element types based on predetermined inputs and outputs. Charnes et al. explained that the comparison should be made based on “decision making efficiency”. Therefore, a decision making unit (DMU) is considered the element subject to comparison. Schools, manufacturing plants, countries, hospitals,
Research aims
Traffic fatalities have been a major concern for both federal and state governments represented by several reports published by both government and private organizations. The number of traffic fatalities in the nation has reached to its lowest level since 1949 (Traffic Safety Annual Assessment Highlights, 2010) as shown in Fig. 1. Even though this declining trend can be interpreted as a positive sign toward highway safety and in support of recent efforts, there are other characteristics of the
Identification of safety performance indicators
According to the European Transportation Safety Council's definition, a Safety Performance Indicator (SPI) is “a measurement that is causally related to accidents or injuries, used in addition to a count of accidents or injuries in order to indicate a safety performance or understand the process that leads to accidents”. In a recent work, Hermans et al. (2009) considered alcohol and drugs, speed, protective systems, vehicle, infrastructure and trauma management as SPIs and studied their
DEA model
The inputs considered in the model included highway safety expenditure, the number of registered vehicles, the number of registered drivers, total vehicle miles traveled, total roadway length, overall road condition and safety belt usage rate. The output was the number of fatal crashes. To summarize, seven inputs and four outputs were considered. In a typical DEA model, the minimum number of DMUs required is the maximum of sum and product rules, which are shown in Eq. (13), where ninput is the
Results
The results of the proposed model were calculated by solving the linear programming models and utilizing the Malmquist productivity index formulation. The results are shown in Table 5. The values are means of the period between 2002 and 2008. There are three components of MPI in FGNZ&FLGR model, namely: PEC, SEC and TC. As previously stated, PEC (Pure Efficiency Change) is the efficiency change of the DMU's under variable returns to scale. On the other hand, the multiplication of PEC and SEC
Conclusions and future work
In this paper, the nations’ highway safety efficiency growth was analyzed. Recent annual highway safety reports indicate that there is a declining trend in the fatality rates in the nation. Every year, millions of dollars are invested on maintenance, construction, education, research and various other areas to reduce fatality rates. While the Federal Government provides guidance to the states, each state creates individualized strategic plans, and implements projects in the aim of reducing
Gokhan Egilmez is a PhD graduate of Industrial and Systems Engineering at Ohio University, USA. He is also enrolled in MS program of Transportation Engineering and working with Dr. McAvoy. He obtained his Industrial and Systems Engineering (ISE) master's degree in 2009 at Ohio University. Prior to ISE master's education, he obtained his BS in Industrial Engineering at Istanbul Technical University, Turkey in 2007. He is currently working as postdoctoral research associate at University of
References (30)
- et al.
The patterns of technological capabilities of countries: a dual approach using composite indicators and data envelopment analysis
World Development
(2011) - et al.
Modeling undesirable factors in efficiency evaluation: comment
European Journal of Operational Research
(2004) - et al.
Undesirable factors in efficiency measurement
Applied Mathematics and Computation
(2005) - et al.
Benchmarking road safety: lessons to learn from a data envelopment analysis
Accident; Analysis and Prevention
(2009) - et al.
Combining road safety information in a performance index
Accident; Analysis and Prevention
(2008) - et al.
Measuring corporate sustainability management: a data envelopment analysis approach
International Journal of Production Economics
(2012) Identifying traffic safety best practice: an application of DEA and Malmquist indices
Omega
(2006)- et al.
Modeling undesirable factors in efficiency evaluation
European Journal of Operational Research
(2002) - et al.
A cone-ratio DEA approach for AMT justification
IJPE
(2000) - et al.
The properties of safety performance indicators in target setting, projections and safety design of the road transport system
Accident; Analysis and Prevention
(2010)
Eco-efficiency analysis of industrial system in China: a data envelopment analysis approach
Ecological Economics
Super-efficiency and DEA sensitivity analysis
European Journal of Operational Research
Some models for estimating technical and scale inefficiencies in data envelopment analysis
Management Science
The economic theory of index numbers and the measurement of input, output, and productivity
Econometrica
Measuring the efficiency of decision-making units
European Journal of Operational Research
Cited by (93)
An integrated decision model with reliability to support transport safety system analysis
2023, Reliability Engineering and System SafetyMeasurement of road safety situation by CRITIC-TODIM-NMF: A lesson system of legislation and regulation for the United states
2023, Measurement: Journal of the International Measurement ConfederationSafety performance measurement in collectivized oil companies in China: Contribution of leading indicators to lagging indicators
2023, Journal of Loss Prevention in the Process IndustriesLinking investment decisions-based on firm performance and open innovation practices in Vietnam's wire and cable market using data envelopment analysis models
2023, Journal of Open Innovation: Technology, Market, and Complexity
Gokhan Egilmez is a PhD graduate of Industrial and Systems Engineering at Ohio University, USA. He is also enrolled in MS program of Transportation Engineering and working with Dr. McAvoy. He obtained his Industrial and Systems Engineering (ISE) master's degree in 2009 at Ohio University. Prior to ISE master's education, he obtained his BS in Industrial Engineering at Istanbul Technical University, Turkey in 2007. He is currently working as postdoctoral research associate at University of Central Florida with a focus on sustainable systems analysis and transportation safety.
Dr. Deborah McAvoy (MS Adviser) is an associate professor in the Department of Civil Engineering at Ohio University. She is a registered professional engineer in the states of Michigan and Ohio and certified Professional Traffic Operations Engineer (PTOE) through the Institute of Transportation Engineers (ITE). Dr. McAvoy's principal area of research interest is in highway safety, traffic engineering, traffic signal system optimization and progression, roadway design and human factors engineering. She has worked on projects for the Michigan Department of Transportation (MDOT), the Federal Highway Administration (FHWA) and the Michigan State Police Office of Highway Safety Planning (OHSP). As a part of these studies, she has conducted technical analyses for signal timing, congestion, crash, gap studies and has substantial experience in corridor progression and work zone safety analyses.
- 1
Tel.: +1 740 593 1468; fax: +1 740 593 0625.