Research in Transportation Business & Management 29 (2018) 127–135

Contents lists available at ScienceDirect

Research in Transportation Business & Management

journal homepage: www.elsevier.com/locate/rtbm

Toward more comprehensive evaluation of traffic risks and safety strategies T

Todd Litman
Victoria Transport Policy Institute, 1250 Rudlin Street, Victoria, BC V8V 3R7, Canada

A B S T R A C T

Despite large investments in traffic safety programs and technologies, motor vehicle accidents continue to impose high social costs. New strategies will be needed to
achieve ambitious traffic safety targets such as Vision Zero. Recent research improves our understanding of factors that affect traffic risks and ways to increase traffic
safety. Applying this knowledge requires a paradigm shift, a change in the way we define problems and evaluate potential solutions. The old paradigm assumed that
driving is generally safe and favored targeted safety programs that reduce special risks such as youth, senior, impaired and distracted driving. The new paradigm
recognizes that all vehicle travel imposes risks, and so, in addition to targeted programs also supports vehicle travel reduction strategies such as more multimodal
planning, efficient transport pricing, Smart Growth development policies and TDM programs. These strategies tend to provide large co-benefits, in addition to safety.
This article examines our emerging understanding of traffic risks and new ways to increase safety.

1. Introduction pricing, Smart Growth development policies, and TDM programs. Since
TDM strategies can provide large co-benefits, besides safety, the new
Despite large investments in traffic safety, motor vehicle crashes paradigm supports more comprehensive analysis that considers these
impose huge social costs. According to a major study, in 2010 U.S. impacts.
motor vehicle crash damage costs totalled $800–2700 per capita Table 1 compares the old and new traffic safety paradigms.
(Blincoe, Miller, Zaloshnja, & Lawrence, 2015). International studies This article explores these issues. It describes traffic casualty trends
show similar results (Wismans et al., 2017), with traffic crash costs and the need for a new safety paradigm, summarizes recent research on
estimated at 5% of GDP in lower- and middle-income countries (Welle traffic risk factors and new safety strategies, evaluates the degree that
et al., 2018). current safety programs consider these factors, and provides re-
Although traffic casualty rates declined significantly during the last commendations for implementing new strategies to achieve safety
century, this long-term trend has started to reverse. In recent years, goals. It should be of interest to anybody who wants to identify the most
crash rates increased indicating that current traffic safety strategies efficient and cost effective ways to improve traffic safety.
have fulfilled their potential. Additional safety strategies are needed to
achieve ambitious safety goals such as Road to Zero (NSC, 2017). This 2. Why a New Paradigm?
will require a paradigm shift, a change in the way traffic risks are
measured and potential safety strategies are evaluated (Hughes, 2017; This section examines the need for a new traffic safety strategies.
Litman, 2018). Fig. 1 shows distance-based and per capita traffic fatality rates.
In a word, the new paradigm recognizes exposure – the amount that These rates declined during most of the Twentieth Century, but pla-
vehicles travel – as a risk factor. Total crashes are the product of dis- teaued between 2010 and 2014 and recently increased. Fig. 2 shows
tance-based crash rates (such as collisions per 100,000 vehicle-miles) 2007 to 2016 U.S. traffic fatality and fuel price trends. When fuel prices
times travel distance (such as per capita vehicle-miles); a change in were high, traffic fatality rates declined, but when fuel prices declined
either tends to cause proportional changes in total crashes. The old between 2014 and 2016, per capita vehicle travel and traffic deaths
paradigm assumed that vehicle travel is generally very safe, and so increased. This and other research described later in this article illus-
ignored exposure as a risk factor. It argued that most crashes are caused trate how factors that affect per capita vehicle travel, and therefore risk
by special risk factors, such as youth, senior, impaired and distracted exposure, affect crash rates.
driving, and so favors targeted safety programs. The new paradigm International comparisons indicate that large safety gains are pos-
recognizes that all vehicle travel imposes risks and so recognizes the sible. The U.S. has the highest per capita traffic fatality rate among its
additional crashes caused by planning decisions that increase vehicle peers (Fig. 3). Geographic factors do not explain this: Australia and
travel, and the safety benefits of transportation demand management Canada have lower population densities, and Sweden, Norway and
(TDM) strategies such as more multi-modal planning, efficient transport Finland have more extreme weather, yet all have much lower traffic

E-mail address: [email protected].

https://doi.org/10.1016/j.rtbm.2019.01.003
Received 16 November 2017; Received in revised form 9 January 2019; Accepted 9 January 2019
Available online 29 January 2019
2210-5395/ Crown Copyright © 2019 Published by Elsevier Ltd. All rights reserved.
T. Litman Research in Transportation Business & Management 29 (2018) 127–135

Table 1
Comparing the old and new traffic safety paradigms.
Factor Old New

Goal Make vehicle travel safer Make transportation systems safer

Risk measurement Direct user risks, measured by distance (e.g., occupant deaths per 100,000 Total risks, including risks to other road users, measured by distance and
million vehicle-miles) per capita
Solutions considered Roadway and vehicle design improvements Walking, cycling and public transit improvements
Graduated licenses Road, parking, fuel and insurance pricing reforms
Senior driver testing More connected and complete roadways
Seatbelt and helmet requirements Smart Growth development policies
Anti-impaired and distracted driving campaigns Transportation demand management programs
Analysis scope Program costs and traffic safety benefits All economic, social and environmental impacts

The old and new traffic safety paradigms differ in many ways.

Fig. 1. Total annual U.S. VMT and traffic fatalities (FHWA, 2015, Table FI-201).
Traffic death declined after 1973, but increased after 1993 and subsequently tracked total vehicle travel, and so recently increased when low fuel prices stimulated
more vehicle travel.

Fig. 2. Recent traffic fatality and fuel price trends (FHWA and GasBuddy Data).
Traffic fatality rates declined while fuel prices where high but increased after 2014 when prices went down.

Fig. 3. Traffic death rates by country (OECD, 2015–2016).

The U.S. has the highest traffic fatality rate among peer countries, nearly twice those of Australia and Canada, and three times those of European countries.

128
T. Litman Research in Transportation Business & Management 29 (2018) 127–135

Fig. 4. Traffic death rate by U.S. urban regions (CDC, 2012).

Crash rates vary significantly between cities, reflecting differences in their transport and land use patterns.

death rates and faster crash rates declines than the U.S.
There are also large variations between geographically similar re-
gions within the U.S., as illustrated in Fig. 4. For example, Seattle, San
Diego and Portland have less than half the rates of Atlanta, Houston and
Sacramento despite similar vehicles, roadways and traffic safety pro-
grams. Evidence described in the next section of this article indicates
that these variations largely reflect transport and land use policies that
affect motor vehicle travel.
Many people hope that new technologies will soon eliminate traffic
risks. Advocates claim that within a few years autonomous vehicles will
be ubiquitous and inexpensive, and since human errors contribute to
90% of traffic crashes, they will eliminate 90% of crashes (Keeney,
2017; Kok, Zou, Gordon, & Mercer, 2017). However, more objective
experts predict that autonomous vehicles will take longer to develop
and cost more than advocates claim, and are likely to introduce new
risks, such as hardware and software failures, malicious hacking, dan-
Fig. 5. Vehicle mileage and traffic fatality rates in OECD Countries (OECD
gers from platooning, increased risk-taking, and increases in total ve-
Data).
hicle travel (Hsu, 2017). As a result, autonomous vehicles will probably Per capita traffic fatality rates tend to increase with vehicle travel among U.S.
reduce crashes much < 90%, and their net safety benefits will depend states, as indicated in Fig. 6.
on public policies that affect how they are programmed and used
(Kockelman et al., 2016; Koopman & Wagner, 2017). For example, to
maximize mobility they can be programmed to operate at higher
speeds, take greater risks in unexpected situations, and have dedicated
platooning lanes, but to maximize safety they should be programed to
drive slower and be more cautious in unexpected situations, resulting in
more frequent delays, and public polices, such as road pricing and high
occupant vehicle priority, can be used to reduce total vehicle travel and
therefore risk exposure (Kaohsiung EcoMobility Festival, 2017).
This suggests that even if autonomous vehicles become common and
affordable, and reduce distance-based crash rates, the new safety
paradigm will still be justified. Policies reduce total motor vehicle
travel, and therefore crash exposure will provide safety benefits even if
they apply to autonomous vehicles.

3. New Understanding of Traffic Risk

This section describes new research concerning how transport and

land use factors affect crash risks. Also see Ewing, Hamidi, and Grace
Fig. 6. Vehicle mileage versus traffic fatalities in U.S. states (FHWA 1993–2002
(2016); Litman and Fitzroy (2016); and Welle et al. (2018). data).
This figure shows traffic fatality and annual mileage rates for urban and rural
3.1. Total Vehicle Travel portions of U.S. states. Per capita traffic fatality rates tend to increase with per
capita vehicle travel, particularly in rural areas.
Although many demographic, geographic and economic factors af-
fect casualty rates, all else being equal, that is, for a given group or area, countries, which can be explained by it also having the highest per
traffic casualties tend to increase with vehicle travel. For example, capita annual mileage.
among higher-income countries, per capita crash rates tend to increase Similar patterns occur at other geographic scales. Fig. 7 shows that
with per capita vehicle travel, as illustrated in Fig. 5. As previously regional traffic fatality rates tend to increase with vehicle travel, and
mentioned, the U.S. has the highest traffic death rate among peer

129
T. Litman Research in Transportation Business & Management 29 (2018) 127–135

increase in transit mode share, from about 1.5% to 4%, which alone
cannot explain the large safety gains. This suggests that factors that
encourage transit travel, such as compact development, good walk-
ability, and reduced parking supply, also tend to reduce vehicle travel
and increase traffic safety.
Fig. 10 shows that statistical relationship between transit ridership
and traffic safety is particularly strong for youths, age 15–25, which
suggests that many young people want to reduce their driving and as-
sociated risk, but can only do so if they have adequate alternatives.
Because active modes (walking and bicycling) have high distance-
based casualty rates, some researchers to conclude that “a shift from
passenger vehicle travel (lower risk) to nonmotorized travel (higher
risk) could result in an overall increase in the numbers of people killed
in traffic” (Beck, Dellinger, & O'Neil, 2007). However, numerous studies
find that both active and total (all mode) crash casualties tend to de-
cline as walking and bicycling increases in an area, an effect called
safety in numbers (ABW, 2016; ECF, 2012; Fyhria, Sundføra, Bjørnskaua,
& Laureshynab, 2017; NACTO, 2016).
As Fig. 11 illustrates, U.S. urban regions with active mode shares
Fig. 7. Vehicle mileage versus traffic deaths (FHWA and CDC data). exceeding 10% average about half the per capita traffic fatality rates as
Per capita traffic fatality rates tend to increase with per capita vehicle-miles in those with active mode shares under 5%. Marshall and Garrick (2011)
U.S. Metropolitan regions. also found that, all else being equal, total traffic fatality rates decline
with increased bicycling mode shares in U.S. cities. Murphy, Levinson,
other studies indicate that traffic casualty rates are much lower in and Owen (2017) found that in 448 Minneapolis city intersections,
compact, multi-modal neighborhoods than in sprawled, automobile- individual pedestrians' motor vehicle crash risk declines as pedestrian
dependent areas (Ewing & Hamidi, 2014; Garrick & Marshall, 2011; traffic increases. Cities that expand their sidewalk and bicycle lane
Welle et al., 2018). networks tend to experience increased walking and bicycling activity
These graphs illustrate simple correlations that may overlook con- and reduce crash rates (NACTO, 2016; Tasic & Porter, 2018).
founding factors. More sophisticated analyses that account for various Various factors help explain these reductions in total crashes asso-
demographic, geographic and economic factors show statistically- ciated with more active transport:
strong positive relationships between mileage and traffic deaths. For
example, Ahangari, Atkinson-Palombo, and Garrick (2017) used 1997 • Safer travel conditions. Both active mode travel and safety tend to
to 2013 data to capture the effect of seven traffic risk factors: exposure, increase with improved sidewalks, crosswalks, cycling facilities,
travel behavior, socioeconomics, macroeconomics, safety policies, and streetscaping, traffic speed control and education programs.
health care quality. Their results indicate that two variables, Vehicles • Complementary factors. Many factors that encourage walking and
per Capita and Vehicle Miles Traveled (VMT) have the strongest impact bicycling, such as connected streets, higher parking and fuel prices,
on per capita traffic fatality rates. Similarly, accounting for various and compact development, also tend to increase safety.
geographic and demographic factors, Yeo, Park, and Jang (2015) found • Reduced total travel. A short active mode trip often substitutes for a
that each 1% increase in per capita VMT is associated with a 0.549% longer automobile trip, for example, walking or biking to local shops
increase in traffic deaths, and comprehensive analysis by Ewing et al. rather than driving to regional shopping centers. Active modes
(2016) found that, normalizing for other factors, each 1% increase in substitute for chauffeuring trips, which often involve empty back-
VMT is associated with 0.3% increase in per capita traffic deaths. hauls, so each mile walked or bike reduces two vehicle-miles. Since
Since about two-thirds of casualty crashes involve multiple vehicles, most public transit trips involve walking and cycling links, im-
and crash rates increase with traffic density (vehicles per lane-mile), proving their conditions can increase transit travel.
changes in total vehicle travel can provide proportionately larger ca- • Reduced vehicle ownership. Improving travel options can allow
sualty changes, particularly in areas with denser traffic (Vickrey, 1968). some households to reduce their vehicle ownership, which leverages
Edlin and Karaca-Mandic (2006) found that each 1% increase in total additional vehicle travel reductions.
vehicle travel increases total crash costs by substantially > 1% in vir- • Increased driver caution. More walking and cycling in an area
tually all U.S. states, and by 3.3–5.4% in dense states such as California. causes drivers to become more aware and cautious.
This indicates that vehicle travel reductions provide external safety • Less high-risk driving. Improving travel options allows young, old,
benefits by reducing risk to other road users, so people become safer if impaired and distracted travellers to reduce driving, increasing the
their neighbors drive less. effectiveness of safety programs such as graduated licenses, senior
driver testing and anti-impaired and distracted driving campaigns
3.2. Quality of Transport Options (Greenwood & Wattal, 2015).
• Reduced risk to other road users. Pedestrians and cyclists impose
The quality of non-auto mobility options can affect crash rates less risk on other road users.
(Stimpson et al., 2014). Fig. 8 shows per mile and per trip crash rates by
mode. Considering both passengers and non-passengers, transit travel 3.3. Transportation Prices
had the lowest total death rate. Bike, motor-cycle and walk have rela-
tively high death rates per mile but impose little risk on others, and Several studies using various analysis methods and data sets in-
since walk trips tend to be shorter than motorized trips, their per trip dicate that more efficient transportation pricing, such as road tolls,
crash rates are similar to auto travel (ABW, 2016). parking fees and fuel price increases, reduces traffic casualty rates
Fig. 9 illustrates the relationship between transit travel and death (Litman, 2014). A comprehensive study of 14 industrialized countries
rates for U.S. urban regions. Regions where residents average > 50 found that a 10% gasoline price decline caused road fatalities to in-
annual transit trips have about half the fatality rates as regions where crease 2.19% (Ahangari, Outlaw, Atkinson-Palombo, & Garrick, 2014).
residents take fewer than 20 annual trips. This represents a small Similarly, Burke and Nishitateno (2015) found that a 10% fuel price

130
T. Litman Research in Transportation Business & Management 29 (2018) 127–135

Fig. 8. Crash rates by mode (NHTSA data)).

Public transit has the lowest total (occupant and external) casualty rate. Auto (cars and light trucks) have moderate crash rates, about a quarter of which is external.
Bike and walk have relatively high per mile crash rates, but their trips are short and impose little external risk, so their total per trip death rates are not much higher
than driving.

Fig. 9. U.S. traffic fatalities versus transit trips (FTA, 2012; NHTSA, 2012).
This figure illustrates total (pedestrian, cyclist, automobile occupant and transit
passenger) traffic fatality rates and average annual transit trips for 35 large
North American cities. As transit travel increases, traffic fatalities tend to de- Fig. 10. Youth and total traffic fatality rates compared to transit travel (CDC,
cline significantly. Cities with > 50 annual transit trips per capita have about 2012).
half the average traffic fatality rate as regions with < 20 annual trips per capita. This figure illustrates total per capita traffic death rates and per capita transit
trip for the 50 largest U.S. metropolitan statistical areas, for youth
increase typically reduces traffic deaths by 3–6%, and estimate that (15–25 years) and total resident. Youths average about twice the traffic fatality
rates as the total population. Both youth and total fatality rates tend to decline
removing global fuel subsidies would reduce approximately 35,000
with increased transit ridership.
annual road deaths worldwide.
U.S. studies find similar results. Paul Leigh and Geraghty (2008)
estimate that a sustained 20% gasoline price increase would reduce vehicle's premiums more accurately reflect its crash costs, and motorists
approximately 2000 annual U.S. traffic deaths plus 600 air pollution have a new incentive to reduce their mileage and therefore crash risk.
deaths. Grabowski and Morrisey (2004) and Grabowski & Morrisey, Green, Heywood, and Navarro (2015) found that London's conges-
2006) estimate that each 10% fuel price increase reduces total traffic tion charge significantly reduced central area vehicle travel and crash
deaths 2.3%, with larger decline for drivers aged 15–21. Morrisey and rates. Within the 8-square-mile charging zone, vehicle travel declined
Grabowski (2011) find that a 10% U.S. fuel price increase reduces 14% and traffic accident rates by 22% (from 4.51 to 3.51 per million
fatalities by 3.2–6.2% with the largest percentage reductions among 15- vehicle-miles), and traffic casualty (injury or death) rates declined 25%,
to 17-year-old drivers, and a 10% beer tax increase reduces motor ve- indicating that the higher travel speeds enabled by reduced congestion
hicle fatalities by 17–24 year old drivers by approximately 1.3%. Stu- do not increase crash severity. Crash rates also declined 16% in areas up
dies by Chi, Cosby, Quddus, Gilbert, and Levinson (2010), Chi, Porter, to four kilometers outside the charging zone.
Cosby, and Levinson (2011) and Chi, Quddus, Huang, and Levinson
(2013) indicate that U.S. fuel price increases reduce both per capita and 3.4. Land Use Development and Roadway Design Factors
per-mile crash rate, so a 1% reduction in total VMT reduces total cra-
shes more than a 1%, with particularly large reductions in youth and Ewing et al. (2016) found that at the U.S. county level, accounting
drunken driving crashes. Analyzing three million vehicle-years of in- for various geographic and demographic factors (land use density and
surance claim data, Ferreira and Minike (2010) found that annual crash mix, block size, roadway connectivity, Walkscore, household size, em-
rates and insurance claim costs tend to increase with annual vehicle ployment and income, race fuel price and climate factors) dispersed,
travel, and so recommend distance-based insurance pricing, so each sprawl land use development is associated with lower per capita rates of

131
T. Litman Research in Transportation Business & Management 29 (2018) 127–135

polycentric design, increased population density and less low-density

sprawl, improving transportation network connectivity, more public
transit facilities, and grade-separated highways. They conclude that
these safety gains result primarily from reductions in per capita vehicle
travel and traffic speeds. They estimate that, all else being equal, a 10%
increase in urban density or the spatial distribution of employment
reduces fatal crash rates by > 15%, a 10% increase in network con-
nectivity increases traffic safety 4.13%, and a 10% increase in public
transit supply reduces fatalities 8.28%.
All else being equal, neighborhoods with more connected and multi-
modal streets tend to have lower traffic casualty rates (Ewing &
Cervero, 2010; Marshall & Garrick, 2011). Mohan, Bangdiwala, and
Villaveces (2017) found that more roadway junctions and fewer kilo-
meters of arterial grade roadways are associated with lower motor-
vehicle and pedestrian traffic death rates.

4. How Common Planning Practices Can Increase Risk

Fig. 11. Active commute mode share and traffic deaths (Census and CDC Data).
This figure illustrates all mode traffic fatality rates and active (walking and
Many conventional transportation and land use planning practices
bicycling) commute mode shares for major U.S. urban regions. Total per capita
can increase total vehicle travel and crash risk (DeRobertis et al., 2014;
death rates tend to decline as active mode shares increase.
Dumbaugh & Rae, 2009). For example, roadway expansions, develop-
ment policies that separate land uses, minimum parking requirements
minor “fender bender” crashes, but significantly higher rate of fatal in zoning codes, and unpriced on-street parking tend to increase motor
crashes, due to the combination of more total motor vehicle travel and vehicle travel (CARB, 2014). More dispersed development, wider roads,
higher traffic speeds in dispersed, automobile-oriented areas. Similarly, and higher traffic speeds also discourage walking and bicycling, which
accounting for demographic and geographic factors (income, fuel prices further increases vehicle travel and reduces the safety in numbers effect.
and compactness) in 147 U.S. urban regions, Yeo et al. (2015) found Some practices intended to increase traffic safety can increase total
that per capita traffic fatality rates increase with sprawl, apparently due crashes by increasing total vehicle travel. For example, since grade-
to a combination of increased vehicle travel, higher traffic speeds and separated highways have low per-mile traffic fatality rates, transpor-
slower emergency response. Similarly, Ahangari et al. (2017) found tation agencies often justify road widening, straightening, grade se-
that traffic death rates decline with urban densities. paration, hierarchical street systems that force traffic onto higher-speed
Najaf, Thill, Zhang, and Fields (2018) find that an urban area's per arterials, and expanded clear zones for safety sake, but such treatments
capita crash rates decline with more job-housing balance, more often cause motorists to drive farther and faster, which can increase

Table 2
Fifteen new paradigm safety strategies (CARB, 2014; Litman, 2018).
Strategy Traffic safety impacts Crash rate reductions

Shorter term (less than three years)

Transit service improvements (more routes, Reduces vehicle travel directly, and often leverage additional Each 1% transit ridership gain typically reduces traffic
frequency, etc.) reductions casualties 1% or more
HOV and bus traffic priority Reduces automobile travel and encourages transit and Can reduce affected traveler's crash rates 10–30%, and
ridesharing total rates 1–5%
Active transport improvements (better sidewalks, Reduces walking and bicycling crash rates, and total per capita Can reduce active travel risks and total crash casualty
bikelane, etc.) crash rates rates 5–10%
Expanded carsharing services Reduces crashes by reducing car ownership Reduce total crashes 0.3–3%, with larger reductions in
denser areas
Raise fuel taxes to fully finance roadway costs, or as Reduces total vehicle travel and traffic speeds A 50¢ per gallon tax should reduce crash casualty rates
a carbon tax 4–12%
Efficient parking pricing (motorists pay directly for Charging motorists directly for parking typically reduces Each 10% increase in the portion of parking that is
using parking spaces) affected trips 10–30%, and may reduce vehicle ownership efficiently priced reduces crash casualties 1–3%
Congestion pricing (road tolls under congested Reduces crashes by reducing automobile use, particularly in Reduces affected area crash casualty rates 15–30%
conditions) large cities
Distance-based vehicle insurance and registration Reduces vehicle use, especially higher risk driving Reduces affected vehicles' crashes by 10–20%
fees
Commute trip reduction programs Typically reduces affected commute trips 5–30%, and may Can reduce affected commuters' crashes casualty rates
cause some vehicle ownership reductions 5–30% and total crashes 0.5–3%
Mobility management marketing Encourages travellers to use non-auto modes Can reduce affected households' crashes 5–10% and total
crashes 3–6%

Longer term (more than three years)

More comprehensive and multi-modal planning Supports more multi-modal transport planning and considers Can lead to large vehicle travel and crash reductions
safety impacts
More connected and complete streets Reduces crash frequency and severity by reducing traffic Can reduce local crash casualty rates 10–30%
volumes and speeds
Reduced parking requirements Reduces crashes by reducing vehicle ownership and use Can reduce affected area's crash casualty rates 5–15%
Urban rail and Bus Rapid Transit Reduces crashes by reducing vehicle ownership and use, and Can reduce crash rates 30–60% in affected areas and
traffic speeds 10–30% region-wide
Smart Growth and Transit Oriented Development Reduces crashes by reducing vehicle ownership and use Typically reduces crash casualty rates 30–60% in
affected areas

New paradigm safety strategies reduce total vehicle travel and traffic speeds.

132
T. Litman Research in Transportation Business & Management 29 (2018) 127–135

Table 3
Comparing analysis frameworks.
Factor Old New

Units of exposure Distance-based units (e.g., casualties per 100 million vehicle-miles or Per capita (e.g., casualties per 100,000 residents)
billion vehicle-kilometers)
Perspective Internal (user) impacts, such as casualties to vehicle occupants Internal and external impacts, such as casualties to vehicle occupants and other
road users
Scope of impacts Traffic crash costs Traffic crash costs and other economic, social and environmental impacts
Scope of analysis Direct impacts only Direct and indirect impacts, including short- and long-term effects on vehicle
travel and risk exposure

The new traffic safety paradigm is more comprehensive and integrated.

total crash casualties (Garrick & Marshall, 2011; Karim, 2015; Noland & etc.), are costly and provide few other benefits, while most new para-
Oh, 2004). As a result, roadway improvements often provide smaller digm safety strategies provide significant co-benefits. For example,
overall safety benefits than predicted. This is not to ignore the benefits improving non-auto modes, efficient transport pricing, Smart Growth
these planning practices may provide, but it is important to consider the and TDM programs tend to reduce congestion, infrastructure costs,
additional crashes they can cause in their evaluation. consumer costs, pollution emissions, as well as improving mobility
options for non-drivers and public health.
5. New Paradigm Safety Strategies These factors can significantly affect planning and traffic safety
program decisions. For example, when comparing roadway expansion
Table 2 summarizes the expected impacts of fifteen vehicle travel or public transit improvements as possible congestion reduction stra-
reduction (also called transportation demand management or TDM) tegies, conventional evaluation usually ignores the additional risk to
strategies, based on previously described research. Of course, actual pedestrians and cyclists caused by wider roads and higher traffic
impacts may vary depending on specific conditions. speeds, additional crashes that result if roadway expansions and dis-
persed development induce additional vehicle travel. Such impacts are
invisible when projects are evaluated using distance-based vehicle
6. New Paradigm Analysis Methods
crash rate data. The new paradigm recognizes these additional risks and
the additional benefits provided by more multi-modal planning and
How impacts are analyzed can significantly affect planning out-
compact development.
comes. A solution that seems effective and beneficial evaluated one way
Most current traffic safety programs reflect the old paradigm (Sung,
may seem ineffective and harmful if evaluated using different metrics
Mizenko, & Coleman, 2017). For example, the 2015 Traffic Safety Facts
and perspectives (DeRobertis et al., 2014). Table 3 compares old and
Report (NHTSA, 2016) shows casualties per 100 million vehicle-miles
new paradigm analyses frameworks. The old paradigm favors auto-
but not per capita, and the USDOT's safety performance indicators are
mobile-oriented safety strategies because it uses distance-based ex-
all distance-based (USDOT, 2017). Few safety programs support vehicle
posure units which ignore the additional crashes that result from in-
travel reduction strategies, and few traffic safety reference documents
creased vehicle travel, ignores the risks that motorized travel imposes
provide guidance on their evaluation or implementation (APTA, 2016).
on pedestrians and bicyclists, and undervalues vehicle travel reduction
co-benefits. Because it applies more comprehensive analysis, the new
paradigm favors more multi-modal planning, Smart Growth and TDM 7. Obstacles and Criticisms
programs.
The old safety paradigm focuses on crash costs, the new paradigm This new traffic safety paradigm faces various obstacles. Many sta-
considers additional costs and benefits. This is important because keholders are unfamiliar with these issues: transportation professionals
planning decisions often involve trade-offs between traffic risk and generally ignore the additional crashes caused by planning decisions
other impacts such as mobility, affordability and environmental that stimulate vehicle traffic, or the potential safety benefits of vehicle
quality. A traffic safety strategy is worth less if it conflicts with other travel reduction strategies. TDM strategies are generally intended to
planning goals, for example, if it increases consumer, government or reduce congestion and emissions; safety benefits are often overlooked.
environmental costs but can be worth far more if it also helps achieve Few traffic safety guidance documents consider vehicle travel reduction
other planning objectives. New analysis tools can help evaluate these strategies, and those that do provide little guidance for their evaluation
trade-offs. and implementation.
Various studies estimate motor vehicle costs (DfT, 2017; Advocates of the old traffic safety paradigm point out that most
Kockelman, Chen, Larsen, & Nichols, 2013; Litman, 2009). A major crashes involve special risk factors, such as youth, senior, impaired or
government study estimated that in 2010, U.S. traffic crash costs distracted driving. From this perspective it seems inefficient and unfair
averaged $1000–3500 per motor vehicle (Blincoe et al., 2015). This to reduce total driving for safety sake, since this seems to punish all
indicates that crash costs are smaller than annual vehicle ownership drivers for errors made by an irresponsible minority. However, even a
costs (financing, depreciation, insurance, registration fees and sched- perfect driver who never errors increases safety by reducing their
uled maintenance, which average about $4000), about equal to total mileage and therefore their chance of being the victim of other drivers'
non-residential parking (the 2–6 off-street parking spaces per vehicle mistakes, and most drivers make small errors that can contribute to a
provided at worksites, shops and other destinations), are about equal to crash, such as driving a little faster than optimal for safety. As a result,
vehicle operation costs (about $2000 for fuel and tire wear), and larger travel reductions tend to provide proportionately larger crash reduc-
than residential parking (about $1200 for a garage or carport), roadway tions, particularly in urban areas (Edlin & Karaca-Mandic, 2006).
costs (which average about $826 per vehicle), traffic congestion (esti- Vehicle travel reduction strategies are sometimes criticized as social
mated at $475 per vehicle), and motor vehicle pollution costs (esti- engineering, with the implication that they force travellers to use un-
mated at $360 annually). desirable mobility options, but such arguments are generally false.
Conventional traffic safety strategies, such as additional vehicle Many TDM strategies improve travel options or provide positive in-
safety features (air bags, rear vision camera, etc.) and traffic safety centives to reduce driving, so travellers benefit overall. Surveys indicate
programs (sobriety checks, new driver testing, advertising campaigns, that many people would prefer to drive less, rely more on alternative

133
T. Litman Research in Transportation Business & Management 29 (2018) 127–135

modes, and live in more multimodal communities. For example, the are convenient, comfortable and affordable. In response, many com-
National Association of Realtors National Community and Transportation munities are implementing more multi-modal planning, pricing re-
Preferences Survey (NAR, 2017) found that most respondents like forms, Smart Growth policies, and TDM programs. This research sug-
walking (80%), about half like bicycling, more than a third (38%) like gests that these strategies also provide significant traffic safety benefits
public transit travel, and nearly 60% report being forced to drive due to that should be considered in their evaluation.
inadequate alternatives. New paradigm strategies respond to latent
demand for non-auto travel. References
Some strategies, such as new urban rail systems, may seem costly
considering just their traffic safety impacts, but provide other important ABW (2016). Bicycling and walking in the U.S. benchmarking report, alliance for biking &
benefits including reduced traffic and parking congestion, infra- walking. http://abw.nonprofitsoapbox.comhttp://abw.nonprofitsoapbox.com/
resources/benchmarking.
structure savings, user savings and affordability, improved mobility for Ahangari, H., Atkinson-Palombo, C., & Garrick, N. (2017). Automobile dependency as a
non-drivers, improved public fitness and health, energy conservation barrier to vision zero: Evidence from the states in the USA. Accident Analysis and
and emission reductions. Considering all impacts new paradigm safety Prevention, 107, 77–85. https://doi.org/10.1016/j.aap.2017.07.012https://bit.ly/
2IMCfdc.
strategies are often very cost effective. Ahangari, H., Outlaw, J., Atkinson-Palombo, C., & Garrick, N. (2014). An investigation
Critics could argue that these strategies' safety impacts are difficult into the impact of fluctuations in gasoline prices and macroeconomic conditions on
to predict, but research described in this article can be used to model road safety in developed countries. Transportation Research Record, 2465 TRB (www.
trb.org); summary at https://bit.ly/2zcwq1i.
how various policy and planning decisions affect travel activity and APTA (2016). The hidden traffic safety solution: Public transportation. American Public
crash rates. Such models are probably as accurate as those used to Transportation Association (www.apta.com); at https://bit.ly/2bYqQpr.
predict conventional safety strategy impacts; in fact, current models Beck, L. F., Dellinger, A. M., & O'Neil, M. E. (2007). Motor vehicle crash injury rates by
mode of travel, United States: Using exposure-based methods to quantify differences.
often exaggerate conventional strategies' net safety gains by ignoring
American Journal of Epidemiology, 166(2), 212–218. https://doi.org/10.1093/aje/
induced travel and offsetting behavior effects (Rudin-Brown & Jamson, kwm064 ; at https://bit.ly/2zPxc8v.
2013). More research is justified, but sufficient information is available Blincoe, L. J., Miller, T., Zaloshnja, E., & Lawrence, B. (2015). The economic and societal
to make reasonable predictions of new safety strategy impacts. impact of motor vehicle crashes, 2010. (Revised), Report no. DOT HS 812 013. National
Highway Traffic Safety Administration ; at https://crashstats.nhtsa.dot.gov/Api/
New paradigm traffic safety strategies may seem outside traffic Public/ViewPublication/812013.
safety program scope, but this is an arbitrary distinction. Conventional Burke, P. J., & Nishitateno, S. (2015). Gasoline prices and road fatalities: International
safety programs influence road and vehicle design standards, law en- evidence. Economic Inquiry. https://doi.org/10.1111/ecin.12171 ; at http://bit.ly/
1QBY62Z.
forcement, business regulations, and social marketing, there is nothing CARB (2014). Research on impacts of transportation and land use-related policies. California
inherently different about multi-modal planning, TDM and Smart Air Resources Boardhttp://arb.ca.gov/cc/sb375/policies/policies.htm.
Growth. CDC (2012). Motor vehicle crash deaths in metropolitan areas—United States, 2009. Center
for Disease Control (www.cdc.gov); at www.cdc.gov/mmwr/preview/mmwrhtml/
mm6128a2.htm.
8. Conclusions and Recommendations Chi, G., Cosby, A., Quddus, M. A., Gilbert, P., & Levinson, D. M. (2010). Gasoline prices
and traffic safety in Mississippi. Journal of Safety Research, 41(6), 493–500 ; at http://
nexus.umn.edu/Papers/GasPricesAndTrafficSafety.pdf.
After a half-century of progress, traffic casualty rates have ended Chi, G., Porter, J., Cosby, A., & Levinson, D. (2011). A time geography approach to un-
their downward trend and recently started to increase. This suggests derstanding the impact of gasoline price changes on traffic safety. TRB (www.trb.org); at
that conventional safety programs are becoming less effective so new http://nexus.umn.edu/Papers/TimeGeography.pdf.
Chi, G., Quddus, M. A., Huang, A., & Levinson, D. (2013). Gasoline price effects on traffic
strategies are needed to achieve ambitious safety goals.
safety in urban and rural areas: Evidence from Minnesota, 1998–2007. Safety Science,
Recent research improves our understanding of traffic risks and 59, 154–162 ; at http://bit.ly/2nkESVx.
potential new safety strategies. Numerous studies using various DeRobertis, M., et al. (2014). Changing the paradigm of traffic impact studies: How ty-
methods and data sets indicate that exposure, the amount that people pical traffic studies inhibit sustainable transportation. ITE Journal, 84(5), 30–35
(www.ite.org), May, pp. 30–35.
travel, is a critical risk factor; all else being equal, an change in vehicle DfT (2017). Transport analysis guidance. UK Department for Transport (www.dft.gov.uk);
travel causes similar changes in crash rates. Since most casualty crashes at www.gov.uk/guidance/transport-analysis-guidance-webtag.
involve multiple vehicles, even a prefect driver increases safety by re- Dumbaugh, E., & Rae, R. (2009). Safe urban form: Revisiting the relationship between
community design and traffic safety. Journal of the American Planning Association,
ducing mileage because this reduces their chance of being a victim of 75(3), https://doi.org/10.1080/01944360902950349 Summer. ; at http://actrees.
another drivers's mistake. org/files/Research/dumbaugh_urbanform.pdf.
A paradigm shift is needed to apply this knowledge. The old traffic ECF (2012). Safety in numbers fact sheet. European cycling federation (https://ecf.com). at
https://ecf.com/files/wp-content/uploads/ECF_FACTSHEET4_V3_cterree_
safety paradigm assumes that most crashes are caused by special risks, SafetyNumb.pdf.
such as youth, senior, impaired and distracted driving, and so favors Edlin, A., & Karaca-Mandic, P. (2006). The accident externality from driving. Journal of
safety programs that target special risk factors. The new paradigm re- Political Economy, 114(5), 931–955 ; at http://papers.ssrn.com/sol3/papers.cfm?
abstract_id=424244.
cognizes that all vehicle travel incurs risk, so policies that stimulate
Ewing, Reid, & Cervero, Robert (2010). Travel and the built environment: A meta-ana-
vehicle travel tend to increase crashes and vehicle travel reductions lysis. Journal of the American Planning Association, 76(3), 265–294. at https://bit.ly/
increase traffic safety. This expands the range of potential safety stra- 1FLv9bB (Summer) .
Ewing, R., & Hamidi, S. (2014). Measuring urban sprawl and validating sprawl measures.
tegies to include more multi-modal planning, pricing reforms, Smart
Metropolitan Research Center, University of Utah (http://mrc.cap.utah.edu); at
Growth development policies, and TDM programs. These new strategies https://bit.ly/2I6StdG.
tend to provide significant co-benefits, including consumer savings, Ewing, R., Hamidi, S., & Grace, J. (2016). Urban sprawl as a risk factor in motor vehicle
congestion reductions, improve mobility for non-drivers, and increased crashes. Urban Studies, 53/2, 247–266. https://doi.org/10.1177/0042098014562331
; at https://bit.ly/2L9zGQT.
public fitness and health. As a result, more comprehensive evaluation Ferreira Jr, J., & Minike, E. (2010). A risk assessment of pay-as-you-drive auto insurance.
tends to support new paradigm strategies. Department of Urban Studies and Planning, Massachusetts Institute of Technology
The new paradigm faces various obstacles, including many stake- (http://dusp.mit.edu); at www.clf.org/wp-content/uploads/2010/12/CLF-PAYD-
Study_November-2010.pdf.
holders' preferences for targeted safety programs and aversion to ve- FHWA (annual reports) (2015). Highway Statistics. Federal Highway Administration
hicle travel reduction strategies. However, new paradigm strategies (www.fhwa.dot.gov); at www.fhwa.dot.gov/policyinformation/statistics/2015/
actually complement existing programs, which become more effective, fi200.cfm.
FTA (2012). National transit database, federal transit administration (www.fta.dot.gov).
equitable and acceptable if implemented with improved mobility op- at www.ntdprogram.gov/ntdprogram.
tions that help higher-risk travellers reduce their driving. Fyhria, A., Sundføra, H. B., Bjørnskaua, T., & Laureshynab, A. (2017). Safety in numbers
This is not to suggest that all motor vehicle travel should be for cyclists—Conclusions from a multidisciplinary study of seasonal change in in-
terplay and conflicts. Traffic Analysis and Prevention, 105, 124–133. https://doi.org/
eliminated for safety sake. However, surveys indicate that many people 10.1016/j.aap.2016.04.039.
would prefer to drive less and rely more on alternatives, provided they

134
T. Litman Research in Transportation Business & Management 29 (2018) 127–135

Garrick, N. W., & Marshall, W. (2011). Does street network design affect traffic safety? https://doi.org/10.1016/j.aap.2017.06.004 ; at http://bit.ly/2uO2mta.
Accident; Analysis and Prevention, 43(3), 769–781. https://doi.org/10.1016/j.aap. NACTO (2016). Equitable bike share means building better places for people to ride. National
2010.10.024 ; draft at http://bit.ly/2Dw9Tlx. Association of City Transportation Officials (https://nacto.org); at https://bit.ly/
Grabowski, D. C., & Morrisey, M. A. (2004). Gasoline prices and motor vehicle fatalities. 2LgfFuT.
Journal of Policy Analysis and Management, 23(3), 575–593 (www.appam.org/ Najaf, P., Thill, J.-C., Zhang, W., & Fields, M. G. (2018). City-level urban form and traffic
publications/jpam/about.asp). safety: A structural equation modeling analysis of direct and indirect effects. Journal
Grabowski, D. C., & Morrisey, M. A. (2006). Do higher gasoline taxes save lives? of Transport Geography, 69, 257–270. https://doi.org/10.1016/j.jtrangeo.2018.05.
Economics Letters, 90, 51–55 ; abstract at www.sciencedirect.com/science/article/pii/ 003.
S0165176505002533. NAR (2017). National community and transportation preferences survey. National
Green, C. P., Heywood, J. S., & Navarro, M. (2015). Traffic accidents and the London Association of Realtors (www.nar.realtor); at www.nar.realtor/reports/nar-2017-
congestion Charge. Lancaster University ; at www.lancaster.ac.uk/staff/greencp/ community-preference-survey.
papers/CongestionChargeMar2015.pdf. NHTSA (2012). Traffic safety facts. National Highway Traffic Safety Administration (https://
Greenwood, B. N., & Wattal, S. (2015). Show me the way to go home: An empirical in- nhtsa.dot.gov). at https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/
vestigation of ride sharing and alcohol related motor vehicle homicide, Fox School of 812032.
Business Research Paper No. 15-054; at. https://doi.org/10.2139/ssrn.2557612. NHTSA (2016). 2015 motor vehicle crash data from FARS and GES. National Highway
Hsu, J. (2017). When it comes to safety, autonomous cars are still ‘teen drivers’. Scientific Traffic Safety Administration (https://nhtsa.dot.gov); at https://crashstats.nhtsa.dot.
American (www.scientificamerican.com); at http://bit.ly/2j9gFPT. gov/Api/Public/Publication/812384.
Hughes, B. (2017). A comprehensive framework for Future road safety strategies, PhD Noland, R. B., & Oh, L. (2004). The effect of infrastructure and demographic change on
Dissertation. Curtin University (https://curtin.edu.au); at https://espace.curtin.edu. traffic-related fatalities and crashes: A case Study of Illinois county-level data.
au/handle/20.500.11937/59647. Accident Analysis & Prevention, 36(4), 525–532 (www.elsevier.com/locate/aap), July
Kaohsiung EcoMobility Festival (2017). Kaohsiung strategies for the future of urban 2004. ; at http://bit.ly/2DnyAQq.
mobility. EcoMobility Festival (www.ecomobilityfestival.org); at https://bit.ly/ NSC (2017). Road to zero. National Safety Council (www.nsc.org); at https://bit.ly/
2OMwsDD. 2h2AyY9.
Karim, D. M. (2015). Narrower Lanes, Safer Streets, CITE annual conference. ; at http://bit. OECD (2015–2016). OECD factbook, organization for economic cooperation and devel-
ly/1GKiICo. opment (www.oecd.org). at www.oecd.org/publications/oecd-factbook-18147364.
Keeney, T. (2017). Mobility-as-a-service: Why self-driving cars could change everything. ARC htm.
Investment Research (http://research.ark-invest.com); at http://bit.ly/2xz6PNV. Paul Leigh, J., & Geraghty, E. M. (2008). High gasoline prices and mortality from motor
Kockelman, K., Chen, T. D., Larsen, K., & Nichols, B. (2013). The economics of transpor- vehicle crashes and air pollution. Journal of Occupational and Environmental Medicine,
tation systems: A reference for practitioners, TxDOT Project 0–6628. University of Texas 50/3, 249–254 ; at www.ncbi.nlm.nih.gov/pubmed/18332774.
at Austin (www.utexas.edu); at www.utexas.edu/research/ctr/pdf_reports/0_6628_ Koopman, Philip, & Wagner, Michael (2017). Autonomous vehicle safety: An inter-
P1.pdf. disciplinary challenge. IEEE Intelligent Transportation Systems Magazine, 9(1), at
Kockelman, K., Avery, P., Bansal, P., Boyles, S. D., Bujanovic, P., Choudhary, T., ... https://bit.ly/2uJho1j.
Stewart, D. (2016). Implications of connected and automated vehicles on the safety and Rudin-Brown, C., & Jamson, S. (2013). Behavioural adaptation and road safety: Theory,
operations of roadway networks. University of Texas Center for Transportation evidence and action. CRC Presswww.crcpress.com.
Research (http://ctr.utexas.edu), for Texas DOT; at http://library.ctr.utexas.edu/ctr- Stimpson, J. P., et al. (2014). Share of mass transit miles traveled and reduced motor
publications/0-6849-1.pdf. vehicle fatalities in major cities of the United States. Journal of Urban Health: Bulletin
Kok, I., Zou, S. Y., Gordon, J., & Mercer, B. (2017). Rethinking transportation 2020–2030: of the New York Academy of Medicine. https://doi.org/10.1007/s11524-014-9880-9 ;
The disruption of transportation and the collapse of the internal-combustion vehicle and oil at http://link.springer.com/article/10.1007%2Fs11524-014-9880-9.
industries. RethinkX (www.rethinkx.com); at http://bit.ly/2pL0cZV. Sung, J., Mizenko, K., & Coleman, H. (2017). A comparative analysis of state traffic safety
Litman, T. (2009). Transportation cost and benefit analysis. Victoria Transport Policy countermeasures and implications for progress “toward zero deaths” in the United States.
Institute (www.vtpi.org/tca). National Highway Traffic Safety Administration (www.nhtsa.gov); at http://bit.ly/
Litman, T. (2014). How transport pricing reforms can increase road safety. Traffic Infra 2zQ1bcQ.
Tech, 68–71 (http://emag.trafficinfratech.com); at www.vtpi.org/TIT-pricesafety. Tasic, I., & Porter, R. J. (2018). Modeling spatial relationships between multi-modal
pdf. transportation infrastructure and traffic safety outcomes in urban environments.
Litman, T. (2018). A new traffic safety paradigm. Victoria Transport Policy Institute (www. Safety Science, 82, 325–337. https://doi.org/10.1016/j.ssci.2015.09.021.
vtpi.org); at www.vtpi.org/ntsp.pdf. USDOT (2017). Safety performance management (safety PM). US Department of
Litman, T., & Fitzroy, S. (2016). Safe travels: Evaluating mobility management traffic safety Transportation (www.transportation.gov); at https://safety.fhwa.dot.gov/hsip/spm.
benefits. Victoria Transport Policy Institute (www.vtpi.org); at www.vtpi.org/ Vickrey, William (1968). Automobile accidents, tort law, externalities, and insurance: An
safetrav.pdf. economist’s critique. Law and Contemporary Problems, 33, 464–487. at www.vtpi.org/
Marshall, W. E., & Garrick, N. W. (2011). Evidence on why bike-friendly cities are safer vic_acc.pdf.
for all road users. Environmental Practice, 13/1 , March; at https://bit.ly/2Li14in. Welle, B., Sharpin, A. B., Adriazola-Steil, C., Bhatt, A., Alveano, S., Obelheiro, M., ... Bose,
Mohan, D., Bangdiwala, S. I., & Villaveces, A. (2017). Urban street structure and traffic D. (2018). Sustainable & safe: A vision and guidance for zero road deaths. World
safety. Journal of Safety Research, 62, 63–71. www.sciencedirect.com/science/ Resources Institute (www.wri.org); at www.wri.org/publication/safe-system.
article/pii/S0022437516305357. Wismans, J., et al. (2017). Economics of road safety–What does it imply under the 2030
Morrisey, M. A., & Grabowski, D. C. (2011). Gas prices, beer taxes and GDL programmes: agenda for sustainable development? Tenth Regional EST Forum in Asia (http://bit.ly/
Effects on auto fatalities among young adults in the US. Applied Economics, 43(25), 2mHZs1p); at http://bit.ly/2iLrgBT.
3645–3654. https://doi.org/10.1080/00036841003670796. Yeo, J., Park, S., & Jang, K. (2015). Effects of urban sprawl and vehicle miles traveled on
Murphy, B., Levinson, D. M., & Owen, A. (2017). Evaluating the safety in numbers effect Traffic fatalities. Accident Analysis and Prevention, 16(4), 397–403. https://doi.org/
for pedestrians at urban intersections. Accident Analysis & Prevention, 106, 181–190. 10.1080/15389588.2014.948616.

135