The impact of cell phones on motor vehicle fatalities

Applied Economics, 2009, 41, 2905–2914
The impact of cell phones on motor
vehicle fatalities
Peter D. Loeba,*, William A. Clarkeb and Richard Andersonc
a
Department of Economics, Rutgers University, Newark, NJ 07102, USA
b
Department of Economics, Bentley College, Waltham, MA 02154, USA
c
Department of Economics, New Jersey City University, Jersey City,
NJ 07305, USA
This article develops a set of models for the determinants of automobile
fatalities with particular attention devoted to the effects of increased
cell phone usage. Cell phones have been associated with both life taking
and life-saving properties. However, prior statistical evaluations
of the effects of cell phones have led to fragile results. We develop
in this article econometric models using time-series data, allowing for
polynomial structures of the regressors. The models are evaluated with
a set of specification error tests providing reliable estimates of the effects
of the various policy and driving-related variables evaluated. The statistical
results indicate the effect of cell phones is nonmonotonic depending
on the volume of phones in use, first having a net life-taking effect, then
a net life-saving effect, followed finally with a net life-taking effect as the
volume of phone use increases.
I. Introduction
The determinants of motor vehicle accidents have
been the topic of interest among economists,
public policy makers and health professionals
for many years. Studies have been conducted on
the determinants of motor vehicle accidents in
aggregate, as well as by components, i.e. automobiles,
trucks, motorcycles, etc. The interest in transportation accidents also led to studies involving railroads,
ships and aircraft as well as accidents due
to the interaction of two or more modes of
transportation. In addition to interest in accidents
themselves, there has been an interest in the
determinants of the outcomes of these accidents, i.e.
injuries, fatalities and property damage.1 Centering
our discussion on motor vehicle accidents, numerous
studies have investigated the effect on motor vehicle
accidents due to: speed, speed variance, alcohol,
speed limits, vehicle miles travelled, measures
of income, unemployment rates, technology
advances, the age of the fleet, population characteristics, police enforcement, seat belt legislation and the
effects of the deregulatory climate which came
about in the 1980s, among others. More generally,
these potential determinants of accidents and factors
reducing accidents may be placed into three categories: those associated with the vehicles themselves,
e.g. technology improvements; those due to the
roadways, e.g. speed limits; and those relating to
drivers, e.g. alcohol consumption, income, seat belt
usage, etc. More recently, the question has arisen as
to the effect of cell phones on motor vehicle accidents.
While it may generally be argued that the probability
of a motor vehicle accident increases with the use
of cell phones by drivers, it is not necessarily
as obvious when considering motor vehicle fatalities.
Some analysts claim that fatalities, like accidents
*Corresponding author. E-mail: [email protected]
1 See Loeb et al. (1994) for a detailed discussion of the determinants of transportation accidents.
Applied Economics ISSN 0003–6846 print/ISSN 1466–4283 online 2009 Taylor & Francis 2905
http://www.informaworld.com
DOI: 10.1080/00036840701858133
increase when drivers use cell phones, due to an
inability of at least some drivers to dial and talk
while driving (similar to an inability to chew gum
and tie one’s shoes).2 In addition, there is evidence
that cell phones reduce the driver’s attention span
and reaction time which in turn increases the
probability of an accident. Further, others have
argued that the distraction of using a cell phone
is not restricted to the time while on the cell
phone itself, but may extend for several minutes
after the phone conversation has terminated.3 On the
other hand, the opposite argument has been made
that cell phones can reduce fatalities, that is, given an
accident, cell phones increase the probability
of obtaining help promptly which may result in
the saving of lives.4 In any case, cell phone use by
the public at large has increased dramatically since
1985. In 1985, there were approximately 340 200
cell phone subscribers. By the year 2004, this number
grew to over 182 140 000 subscribers.5 It probably
would be descriptive to say the growth of cell phone
subscribers was explosive rather than exponential
over time. In addition, Glassbrenner (2005), in
a National Highway Traffic and Safety
Administration (NHTSA) study, estimates that in
2005, 10% of all drivers at any moment during
daylight hours are using a hand-held or hand-free cell
phone. Looking at hand-held phones only,
Glassbrenner estimated that driver hand-held use
rose from 5% in 2004 to 6% in 2005.6 As such, we see
not only tremendous growth in cell phone subscribers
but also an increase in usage of these devices over
time by drivers.
The question then arises as to the net impact
that cell phones have on motor vehicle fatalities.
This study provides some econometric insight into
the effect of cell phone use on motor vehicle fatalities.
This is accomplished using econometric techniques on
an annual time-series dataset covering the period
1975 to 2003. The models developed are subjected
to a set of specification error tests to assure that the
results are statistically viable.7
II. Background
Between 1975 and 1980 motor vehicle fatalities
trended upward (from 44 525 to 51 091) and then
trended downward to 42 643 in 2003.8 The determinants of these fatalities have been the subject
of numerous studies in the past. These factors, as
mentioned above, include socioeconomic factors such
as the unemployment rate, income and population
attributes as well as vehicle and driver-related factors.
These latter factors include, among others, alcohol
consumption, speed limit laws, vehicle miles
travelled, interstate highway mileage, the ratio of
rural to interstate highway mileage, the blood alcohol
threshold indicative of driving under the influence
of alcohol and a time trend.9 Recently, the question
of the potential effect of cell phones on motor vehicle
fatalities has become a significant policy issue such
that three states (New Jersey, Connecticut and
New York) as well as Washington DC ban the use
of hand-held cell phones while driving.
The traditional econometric-based literature has
examined many of the potential factors effecting
motor vehicle accidents and injuries, starting with
Peltzman (1975). Obviously, early studies do not
consider cell phones, since they were not in use
at the time of the studies. In these studies, economic
factors such as measures of income and the unemployment rate have been examined.10 Some have
argued that higher incomes are associated with more
accidents while others have argued that they are
associated with fewer accidents. Since the demand
for safety and driving intensity are both related to
income and may counteract one another, the net
effect needs to be determined empirically.11
2 See, for example, Violanti (1998).
3 See, for example, Consiglio et al. (2003) and McEvoy et al. (2005). 4 See Chapman and Schofield (1998).
5 See Cellular Telecommunications and Internet Association (2005). The growth in cell phone subscriptions is shown in Fig. 2.
6 See Glassbrenner (2005).
7 Cell phones have also been suggested as a potential contributor to disease. This question is not addressed in the present
article. See, for example, Maier et al. (2000) on this matter. 8 See Fig. 1.
9 See Loeb et al. (1994) on this. 10 A time trend has also been used in some studies as a proxy for permanent income, changes in technology or other omitted
factors. See Peltzman (1975) and Loeb (1993, 2001). Concern arises in using both a time trend and measures of income in
models due to high correlation between such variables. Loeb and Clarke (2007) found a correlation between real GDP and a
time trend to be 0.988 and 0.997 when the variables are measured in natural log form for the time period 1970–2001.
11 Prior studies have found a positive relationship between accidents and income using time-series models and a negative
relationship with cross-sectional models. See Loeb et al. (1994).
2906 P. D. Loeb et al.
Unemployment rates have been used in many
studies as a control for economic activity and have
been found to be inversely related to motor
vehicle accidents.12 Miles driven is expected to be
positively related to motor vehicle accidents as
is alcohol consumption by drivers and the
public in general.13 Further, interstate highway
mileage and the relationship between rural and
urban highway mileage have also been investigated
as to their effects on accidents.14 In addition, studies
have suggested that empirical models should be
normalized by population or characteristics of the
population, such as the proportion of the population
aged 65 or older, or the proportion of the population
characterized as youthful.15 The effect of speed and
speed variance has been an issue debated at length in
the transportation literature.16 The effect of speed
limits have also been investigated as potential
contributors to motor vehicle accidents and have
been found in some studies to have some effect on
fatalities.17
Unlike the above-mentioned variables, cell phones
have only recently been investigated regarding their
influence on motor-vehicle-related accidents, and
then rarely in a traditional regression format.
Perhaps the most well-known study to date is that
of Redelmeier and Tibshirani (1997). Using crossover analysis, they analyse 699 drivers involved
in property-damage-only accidents in the Toronto
(Canada) area during the period 1 July 1994 to 31
August 1995. They find, among other things, that the
risk of being involved in such an accident is four
times higher when a cell phone is being used. In
addition, they find that hands-free devices offer no
additional safety as compared to the hand-held
devices. However, they find that 39% of drivers
involved in these accidents used cell phones to call
for assistance after a crash which suggests that cell
phones may be advantageous after a crash. Similarly,
McEvoy et al. (2005) using cross-over analysis and
Australian data evaluated 456 drivers aged 17
or older who owned cell phones and were
involved in crashes which resulted in hospital visits.
Interestingly, they found that the use of mobile
phones up to 10 min before a crash was associated
with a four-fold increased probability of a crash.
Hands-free devices again were found no safer than
hand-held devices. Laberge-Nadeau et al. (2003),
using Canadian survey data (36 078 responses) and
logistic-normal regression models found the relative
risk of accidents is higher for users of cell phones as
compared to nonusers by approximately 38%.
However, the risk diminishes when the models are
more extensive. Another survey-based study using
cross-over analysis was conducted by Sullman and
Baas (2004). Their survey resulted in approximately a
50% response rate and investigated all crashes,
regardless of severity. Once demographic and other
variables were accounted for, they did not find a
significant correlation between cell phone use and
crash involvement. Examining traffic fatalities,
Violanti (1998) using regression analysis finds that
cell phones are associated with approximately
a, ‘nine-fold risk for a fatality over those not using
a phone.’18 This life-taking effect is countered by
Chapman and Schofield (1998) who claim that cell
phones in Australia should be credited with saving
lives. They found that, ‘Over one in eight current
mobile phone users have used their phones to report a
road accident.’19 Referring to the ‘golden hour’
(which is critical to trauma victims) and with
reference to medical and other emergencies over the
last decade they claim that, ‘it seems highly plausible
that many Australians may have had their lives saved
because help was summoned on a mobile phone.’20
An alternative technique to examine the effect of cell
phones on motor vehicle accidents involves the use
of measuring response rates, or reaction time, in
a simulated laboratory situation. Consiglio et al.
(2003) found that both hand-held and hand-free
devices resulted in a reduction in reaction time in a
braking situation. They also found that conversation
both in-person or via a cell phone caused reaction
time to slow while listening to music did not. As such,
there is evidence suggesting that cell phones may have
both life-saving as well as life-taking attributes.
12 See, for example, Evans and Graham (1988), Loeb (1995) and Fowles and Loeb (1995). 13 See Loeb et al. (1994), Fowles and Loeb (1995) and Loeb and Clarke (2007). 14 See Loeb et al. (1994). 15 See, for example, Loeb (1988), Fowles and Loeb (1992), Keeler (1994) and Loeb and Clarke (2007). It should be noted
further that, as in the case concerning real GDP and a time trend, there is also a strong correlation between real GDP and
both population and vehicle miles driven. For the sample period 1975–2003, the correlation between real GDP and a time
trend is 0.99 while the correlation between real GDP and population and vehicle miles driven is 0.995 in both cases. As such,
most models developed do not include more than one of these variables so as to avoid multicollinearity.
16 See Lave (1985), Fowles and Loeb (1989, 1995), Levy and Asch (1989) and Loeb et al. (1994). 17 See Loeb (1993) and Fowles and Loeb (1995). 18 Violanti (1998, p. 522). 19 Chapman and Schofield (1998, p. 5). 20 Chapman and Schofield (1998).
The impact of cell phones on motor vehicle fatalities 2907
The current article evaluates the effect of cell
phones on motor vehicle fatalities using econometric
methods and a time-series dataset covering the period
1975 to 2003. As such, we are able to diminish
statistical problems associated with survey data
and provide an analysis while controlling for traditional factors in such models. To reduce the likelihood that the models suffer from omission
of variables, misspecification of the structural form,
or simultaneous equation bias (which may result
in biased estimates) along with nonnormality of the
error structure and serial correlation, a set
of specification error tests are applied to the
models.21 Only models which are not rejected by
any of these tests are used in the analysis to follow.
III. The Model and Data
The general model
A model of the form
FATTOT ¼ 0 þ 1X þ 2CELLS þ ð1Þ
is suggested where FATTOT is total motor vehicle
fatalities and X is a matrix of socioeconomic and
motor-vehicle-related variables thought to be determinants of the dependent variable. CELLS is
a measure of cell phone availability/usage whose
influence on the dependent variable is to be examined. is a random error term assumed to comply
with the full ideal conditions underlying the classical
linear regression model which would result in Best
Linear Unbiased Estimates of the coefficients
when the model is estimated by Ordinary Least
Squares (OLS). Cell phone usage by drivers is not
measured directly but is proxied by cell phone
subscribers.22 Glassbrenner (2005) has estimated
that approximately 10% of all drivers at any daylight
moment are making use of a cell phone device.
Therefore, the number of cell phone subscribers
should serve as a reasonable proxy for the potential
distracting influence on drivers which may result in
an accident and fatality.23 In the models suggested,
CELLS is entered so that the model appears
as a polynomial of degree 3. This conforms to similar
models in the Industrial Organization literature
examining the determinants of Research and
Development.24
The variables considered for possible inclusion
in X are
UNEMP the civilian unemployment rate.
ETHOTOTAL apparent total per capital ethanol consumption (in gallons)
based on population aged 14
and over.
VEHMI vehicle miles driven (billions).
TOTALMI total highway mileage
(in thousands).
INTERMI interstate highway mileage
(in thousands).
URBANMI urban highway mileage
(in thousands).
RURALMI rural highway mileage (in
thousands).
POP total population (in
thousands).25
BAC a dummy variable indicating the
blood alcohol threshold associated with driving under the
influence of alcohol. State laws
vary over time. A BAC of ‘1’
indicates either a 0.1 or a 0.08
BAC threshold is in effect while
a BAC of ‘0’ indicates no such
legal threshold exists.26
SPEED95 a dummy variable indicating the
repeal of the maximum speed
limit by the National Highway
System Designation Act of 1995
in December 1995 which allowed
states to set their own limits
21These include a Regression Specification Error Test, a Jarque–Bera test and a Durbin–Watson test. 22 In addition, it may capture the effect of cell phone availability to those not involved in an accident on the outcome of an
accident.
23 Again, as the number of cell phones among the general public increases, there may be a reduction in fatalities given an
accident due to the increased speed in which medical assistance may be acquired.
24 See Scherer (1965a, b) and Loeb and Lin (1977). The polynomial function allows for the examination of both the possible
life-saving and life-taking effects of cell phones as the number of cell phones increases.
25 Additional population variables were examined as well where the entire population was decomposed into those aged 16–19,
20–24, 25–44, 45–64, 65 and older. This allowed for examining whether certain groups of the population contributed to
accidents more than others. Results using POP were superior to other variations of population. The alternative results are
available from the authors.
26The BAC data are by state and over time. An annualized value is obtained by weighting each state’s BAC threshold for a
given year by its relative share of the population for the year. We are grateful to Michael Grossman for the data and weighting
mechanism used to generate this variable.
2908 P. D. Loeb et al.
for the first time since 1974.27
SPEED95 is set equal to
0.083333 in 1995, set equal to 1
from 1996 to 2003 and set to zero
for all other time periods.
CELLS estimated number of cell phone
subscribers (in millions).
We anticipate a negative effect on the dependent
variable by UNEMPL and BAC. We expect a positive
effect by ETHOTOTAL, the ratio of rural mileage
to urban mileage (RURALMI/URBANMI) and
CELLS. We do not have strong a priori expectations
regarding the effects of INTERMI and POP.28
The data
Data generally cover the period 1975 to 2003, but may
vary depending on the model and variables included.
The dependent variable (FATTOT) is measured as
total motor vehicle fatalities and is reported in
NHTSA (2005). Vehicle miles driven, measured in
billions of miles (VEHMI), are reported in NHTSA
(2005) as well. The unemployment rate (UNEMP) and
population data (POP) are reported in the Economic
Report of the President (2005). Total apparent per
capita ethanol consumption (ETHOTOTAL) is
reported by the National Institute on Alcohol Abuse
and Alcoholism and is found at: http://www.niaa.-
nih.gov/Resources/DatabaseResources/QuickFacts.
Data on highway mileage, i.e. total mileage
(TOTALMI), interstate mileage (INTERMI), rural
mileage (RURALMI) and urban mileage
(URBANMI) are reported in the US Bureau of
the Census (various years). The speed limit variable
(SPEED95) was generated as a dummy variable
taking on the value of ‘1’ for the years in which the
National Highway System Designation Act of 1995
was in effect and ‘0’ otherwise. In the year 1995, the
variable was set equal to 0.083333 to reflect the proportion of that year the act was in effect. Data associated with the BAC laws are provided by MADD
(http://www.madd.org/laws/) and the US Bureau
of the Census (http://www.allcountries.org/uscensus/
1046_state_legislation_alcohol_and_road_safety.html).
Data for the number of cell phone subscribers (CELLS)
are in millions and are reported by CTIA (2005) for the
years 1985–2004. CTIA does not report the number of cell
phone subscribers prior to 1985 and one could infer that
none existed. However, AT&T put into place a trial
system involving 2000 customers in 1977. As such, rather
than assume there were no cell customers prior to 1985, we
extrapolate potential subscribers for the period 1977 to
1984. Relatively speaking, these do not amount to
significant numbers, but do take account of the fact that
cell phones were in very limited use prior to 1985.29
IV. Model Selection and Empirical Results
Model selection
Equation 1 can be respecified to introduce
a polynomial effect of CELLS on the dependent
variable as
FATTOT ¼ 0 þ 1X þ 2CELLS
þ 3CELLS2 þ 4CELLS3 þ ð2Þ
where X includes various combinations of the
other regressors discussed above.30 Variations of
Equation 2 are estimated by Ordinary Least
Squares (OLS). Under the full ideal conditions,
OLS results in Best Linear Unbiased Estimates
(BLUE). Concern arises due to the possibility of
not only serial correlation, given the use of time-series
data, but also the possibility of biased estimates due
to the potential omission of variables, simultaneous
equation problems and misspecification of the
structural form of the model. As such, the various
specifications estimated were subjected to a set of
specification error tests. These tests are used to detect
violations of the full ideal conditions. They include
the Regression Specification Error Test (RESET)
developed by Ramsey (1974), the Jarque–Bera test
(J–B) for normality and the Durbin–Watson test
(D–W) for serial correlation. RESET examines the
residuals of a given regression for the possibility of
omission of variable(s), simultaneous equation bias
27 An alternative dummy variable to account for the speed limit was also evaluated. SPEED87 indicated the years Congress
allowed states to increase speed limits to 65 mph on rural interstates. These results are available from the authors.
28 As mentioned previously, population and vehicle miles drive were found to be highly correlated with other variables, e.g.
the time trend and other measures of miles driven. As such, most models estimated restricted the number of such variables
included so as to minimize the risk of multicollinearity. A correlation matrix is available from the authors.
29Even in 1985, there were only slightly more than 340 000 cell phone subscribers. See CITA (2005). It should be further noted
that model (1) was expanded to include additional variables to account for the influence of hospital availability, income,
technology and other factors. However, these variables are not included in what follows since their inclusion either did not
add substantially to the results reported or they introduced potential specification errors such as multicollinearity or errors
resulting in biased estimates or nonnormal errors.
30These would include the ratio of RURALMI to URBANMI and Vehicle Miles per Capita (VEHMI/POP), among others.
The impact of cell phones on motor vehicle fatalities 2909
and misspecification of the structural form of the
regressors, any of which may result in biased
estimates.31 Any given specification estimated is
rejected if one or more of the above-mentioned tests
rejects the hypothesis of the presence of the full ideal
conditions. This is a rather severe requirement
imposed on the model selection process as compared
to relying on student t-statistics or R2
’s.32 Those
specifications which are not rejected provide reasonable specifications to evaluate at least from a single
equation perspective.
Empirical results
Models suggested by Equation 2 were estimated by
OLS and examined for specification errors. Table 1
provides summary statistics for variables used in the
equations not rejected for specification errors mentioned above or for multicollinearity.33 Table 2
provides a set of models for FATTOT which were
not rejected for specification errors.
The estimated coefficients tend to be stable
(nonfragile) and consistent across the various specifications. Those reported are not rejected by
the various specification error tests and all have
adjusted R2
’s above 0.91. The coefficient associated
with the unemployment rate is consistently negative
and statistically significant at better than the 0.005
significance level across all the specifications
estimated. The coefficient associated with alcohol
consumption is, as expected, always positive
and statistically significant at better than the 0.005
significance level. The ratio of rural to urban highway
mileage has a positive influence on the dependent
variable as seen across all specifications, as expected,
with the associated coefficients varying in significance
with most near the 0.1 one-tail level. The BAC
coefficients were negative and significant at approximately the 5% or better one-tail level. The interstate
mileage driven and population variables did not have
coefficients which were statistically significant.
The variables of particular interest in this study
are variants of CELLS indicating the potential effect
of cell phone usage on motor–vehicle-related fatalities. The coefficients associated with the variables
which account for cell phone usage are always
nonfragile across specifications and are statistically
significant at the two-tail significance level of 0.05
or better (with the vast majority of these coefficients
significant at the 0.01 level or better). They suggest
that the effect of cell phones on fatalities is positive.
In addition, cell phone results suggest that their
influence on fatalities is nonlinear. The effect on
fatalities increases at a decreasing rate as noted by the
coefficients associated with CELLS and CELLS2
, but
the effect on fatalities is significantly impacted by
CELLS3
. At some point, the effect of cell phones may
actually decrease the number of fatalities. Using
specification (1) from Table 2, we can see that the
marginal effect of CELLS on fatalities depends on
the level of CELLS. Evaluated at the mean of CELLS
indicates a value of 41.575, all else equal. However,
examining specification (1) for the volume of cells
associated with a maximum or minimum number of
fatalities requires that we examine the partial
derivative of FATTOT with respect to CELLS.
Setting @FATTOT/@CELLS ¼ 0 gives rise to a quadratic with the two roots being 34.49 and 97.76.
Evaluating the second derivatives results in
CELLS466.13 is associated with a minimum
and CELLS566.13 is associated with a maximum.
More precisely, evaluating the original specification
(1) for the effect of CELLS on FATTOT, holding
all else equal, reveals an initial maximum at
34.49 CELLS and a minimum at 97.76 CELLS.
As such, the effect of CELLS initially exacerbates
fatalities and then fatalities decline with additional
phone subscribers. However, the analysis must not
cease at this point, for the above maximum and
minimum values are local maxima and minima.
The function then increases without end after
Table 1. Summary statistics of selected variables
Variable name Mean SD
FATTOT 44 356.64 3300.725
UNEMP 6.425 1.466
ETHOTOTAL 2.445 0.24
VEHMI 2049.464 478.932
TOTALMI 3892.643 33.260
INTERMI 44.050 3.43
URBANMI 746.464 83.49
RURALMI 3147.250 57.12
POP 24 9334.5 22 114.98
SPEED95 0.253 0.439
CELLS 26.401 42.235
RURALMI/URBANMI 4.275 0.551
VEHMI/POP 0.008 0.001
BAC 0.725 0.31
31 RESET’s ability to detect misspecification associated with structural form is of particular interest given the polynomial
specification suggested.
32 See Ramsey and Zarembka (1971) on this. 33 A correlation matrix is available from the authors.
2910 P. D. Loeb et al.
Table 2. Regression results for models of FATTOTa
Equation #
Independent
variable
name (1) (2) (3) (4) (5) (6) (7)
Constant 4159.003 (0.247) 69.774 (0.008) 87.115 (0.014) 3346.150 (0.210) 4698.404 (0.269) 9910.879 (1.099) 11 020.58 (1.252)
UNEMP 2241.382 (8.633) 2254.079 (9.014) 2254.109 (9.246) 2248.608 (8.969) 1966.420 (9.692) 1946.920 (10.611) 2178.663 (9.17)
ETHOTOTAL 17 335.02 (3.460) 17 814.51 (3.852) 17 811.30 (4.087) 17 364.19 (3.554) 22 919.23 (6.112) 19 541.31 (5.832) 14 970.70 (3.332)
CELLS 242.780 (3.306) 244.027 (3.407) 244.038 (3.497) 243.911 (3.415) 252.338 (3.318) 181.972 (2.338) 183.362 (2.422)
CELLS2 4.761 (3.838) 4.842 (4.1) 4.842 (4.224) 4.815 (4.081) 4.614 (3.590) 3.77 (3.043) 3.965 (3.270)
CELLS3 0.024 (4.131) 0.024 (4.426) 0.024 (4.570) 0.024 (4.419) 0.022 (3.759) 0.019 (3.41) 0.020 (3.73)
INTERMI 22.947 (0.194) 0.232 (0.003) 55.188 (0.456)
POP 0.014 (0.292) 0.008 (0.224) 0.030 (.0616)
RURALMI/
URBANMI
3158.385 (1.607) 3277.626 (1.745) 3277.854 (1.790) 3223.214 (1.706) 2647.787 (1.478)
BAC 2148.064 (1.998) 1822.012 (1.704)
Adjusted R2 0.918 0.922 0.925 0.922 0.911 0.928 0.932
D–Wb 1.771 1.804 1.804 1.780 1.86 1.839 1.736
J–Bc 2.775 2.719 2.717 2.705 0.844 0.411 1.664
RESETd 1.335 1.291 1.361 1.389 2.353 2.356 1.3
Notes: aNumbers shown within parentheses are t-statistics associated with the coefficients. bDurbin–Watson statistic.
cJarque–Bera test. dRegression specification error test.
The impact of cell phones on motor vehicle fatalities 2911
reaching the minimum.34 This may explain some of
the different results found by other researchers.
As such, we first note a contributing effect
of cell phones on motor-vehicle-related deaths.
This may be due to several factors including
inexperience among drivers using cell phones (much
like not being able to chew gum and do another task
at the same time), the distracting effect of cell phones,
and an insufficient number of cell phones among the
public to afford a faster response time to an accident
due to cell phone calls which might counter
the negative effects of cell phone use. After a critical
amount of subscribers is reached, the life-saving effect
of cell phones among the public may, for example,
increase the likelihood of survivorship in an
accident due to the ability to immediately call for
medical assistance. Hence, the life-saving effects
of cell phones outweigh the life-taking effect after
some point. However, beyond a certain level once
again, the life-taking effect overwhelms the life-saving
effect again. With the explosive growth in the number
of cell phones, drivers have become more accustomed
to using cell phones while driving. This may
reflect a change in attitudes and habits regarding
cell phone usage over time. This change
is documented by NHTSA which reports a 50%
increase in drivers using cell phones in the daylight
hours since 2002.35 An examination of Figs 1 and 2
add to the analysis. Figure 1 indicates that motor
vehicle fatalities tend to trend downward until the
early to mid-1990s. By the year 2000, motor vehicle
fatalities were close to 42 000 and rising. Meanwhile,
cell phone subscribers had surpassed in number the
97.8 million mark. Hence, the visual evidence adds
additional support to the conclusion that cell phones
and their usage above a critical threshold add to
motor vehicle fatalities.36
V. Concluding Comments and Policy
Recommendations
This study examines the effect of cell phones on motorvehicle-related fatalities using econometric models
and specification error tests. The models adjust
for unemployment rates, alcohol consumption, blood
alcohol legislation and other factors as well as cell
phones. Data are obtained for the years 1975–2003
and the models are evaluated using OLS. Only models
which fail to be rejected for specification errors are
considered in the analysis.
As expected, alcohol consumption is shown
to have a positive and statistically significant effect
on fatalities with coefficients which are stable across
specifications. This lends support for public policies
38000
40000
42000
44000
46000
48000
50000
52000
54000
1960 1970 1980 1990 2000 2010
Fatalities
Year
Fig. 1. Motor vehicle fatalities over time
0
40
80
120
160
200
1960 1970 1980 1990 2000 2010
Cell phone subscribers (millions)
Year
Fig. 2. Cell phone subscribers over time
34 See Fig. 3 for the estimated response rate of fatalities to cell phones. 35 Reported by the Associated Press in http://www.msnbc.msn.com/id/8545779/. See Glassbrenner (2005) as well on increases
of cell phones by drivers.
36Models in terms of rates, i.e. fatalities/capita, were estimated as well. The results conform with those presented in Table 2
especially in those models which do not include BAC. These results are available from the authors.
2912 P. D. Loeb et al.
which would result in reduced alcohol consumption.
Such policies may include increased taxes on alcohol
or possibly raising the minimum legal drinking age.37
The ratio of rural highway mileage to urban highway
mileage also has an impact but at a lower significance
level. The unemployment rate, as expected, has
a negative and statistically significant effect on
fatalities. Similarly, the use of BAC laws to define
driving while under the influence of alcohol also
reduces fatalities. As such, reducing the blood alcohol
limit indicating when driving under the influence
may prove to be a valuable policy to reduce motor
vehicle fatalities.
Most significantly, cell phones are found to
have an adverse effect on fatalities initially as
cell phones become more readily available. After
a point, some life-saving effects of cell phones
overtake the life-taking effects. However, with cell
phone subscribers reaching about 100 million or
more, the life-taking effect overtakes the potential
life-saving effect once again. These results are
nonfragile across specifications and are statistically
significant. Given that there are over 219 million cell
phone subscribers in the United States (as of 2006)
leads one to believe that the life-taking effect of cell
phones is greater than the life-saving effect. As such,
policies which would reduce cell phone use by drivers
may be warranted. This might be accomplished
through an appropriate fine structure combined
with active enforcement.
Acknowledgements
The authors are indebted to Michael Grossman,
Jeffrey Cohen and Jason Barr for valuable suggestions. Loeb gratefully acknowledges the research
support of a Rutgers University Research Council
Grant.
References
Cellular Telecommunications and Internet Association
(CTIA) (2005). Available at http://www.mindfully.org/
Technology/2005/Wireless-Industry-SurveyJan05.htm
(accessed 21 January 2009).
Chaloupka, F. J., Saffer, H. and Grossman, M. (1993)
Alcohol-control policies and motor vehicle fatalities,
Journal of Legal Studies, 22, 161–86.
Chapman, S. and Schofield, W. N. (1998) Lifesavers and
samaritans: emergency use of cellular (mobile) phones
in Australia, Accident Analysis and Prevention, 30,
815–9.
Consiglio, Wm., Driscoll, P., Witte, M. and Berg, Wm. P.
(2003) Effect of cellular telephone conversations and
other potential interference on reaction time in a
braking responses, Accident Analysis and Prevention,
35, 495–500.
Economic Report of the President (2005) United States
Government Printing Office, Washington, DC.
Evans, W. and Graham, J. D. (1988) Traffic safety
and business cycles, Alcohol, Drugs and Driving, 4,
31–8.
Fowles, R. and Loeb, P. D. (1989) Speeding, coordination,
and the 55-MPH limit: comment, American Economic
Review, 79, 916–21.
Fowles, R. and Loeb, P. D. (1992) The interactive effect of
alcohol and altitude on traffic fatalities, Southern
Economic Journal, 59, 108–11.
Fowles, R. and Loeb, P. D. (1995) Effects of policy-related
variables on traffic fatalities: an extreme bounds
analysis using time-series data, Southern Economic
Journal, 62, 359–66.
Glassbrenner, D. (2005) Driver cell phone use in 2005 –
overall results. Traffic safety facts: research note,
NHTSA, DOT HS 809967. Available at http://wwwnrd.nhtsa.dot.gov/pdf/nrd-30/NCSA/RNotes/2005/
809967.pdf (accessed 21 January 2009).
Keeler, T. E. (1994) Highway safety, economic behavior,
and driving environment, American Economic Review,
84, 684–93.
Laberge-Nadeau, C., Maag, U., Bellavance, F., Lapiere,
S. D., Desjardins, D., Messier, S. and Saidi, A. (2003)
Wireless telephones and the risk of road crashes.
DOI:10.1016/S0001-4575(02)00043-X.
Lave, C. A. (1985) Speeding, coordination and the 55 mph
limit, American Economic Review, 79, 1159–64.
Levy, D. T. and Asch, P. (1989) Speeding, coordination,
and the 55 mph limit: comment, American Economic
Review, 79, 913–5.
Loeb, P. D. (1988) The determinants of motor vehicle
accidents – a specification error analysis, The Logistics
and Transportation Review, 24, 33–48.
Loeb, P. D. (1993) The effectiveness of seat-belt legislation
in reducing various driver-involved injury rates in
California, Accident Analysis and Prevention, 25,
189–97.
Loeb, P. D. (1995) The effectiveness of seat-belt
legislation in reducing injury rates in Texas,
American Economic Review – Papers and Proceedings,
88, 81–4.
Loeb, P. D. (2001) The effectiveness of seat belt legislation
in reducing driver-involved injury rates in Maryland,
Transportation Research: Part E, 37, 297–310.
0
2000
4000
6000
0 20 40 60 80 100 120 140
Cell phones (millions)
Fatality effect
Fig. 3. Estimated response rate of fatalities to cell phones
37 See Chaloupka et al. (1993) on the effect of alcohol control policies.
The impact of cell phones on motor vehicle fatalities 2913
Loeb, P. D. and Clarke, Wm. A. (2007) The determinants
of truck accidents, Transportation Research: Part E,
43, 442–52.
Loeb, P. D. and Lin, V. (1977) Research and development
in the pharmaceutical industry – a specification error
approach, The Journal of Industrial Economics, 26,
45–51.
Loeb, P. D., Talley, W. K. and Zlatoper, T. J. (1994)
Causes and Deterrents of Transportation Accidents:
An Analysis by Mode, Quorum Books,
Westport, CT.
Maier, M., Colin, B. and Koivisto, M. (2000) The health
hazards of mobile phones, British Medical Journal,
320, 1288–9. Available at http://www.pubmedcentral.
nih.gov/articlerender.fcgi?artid=1127293 (accessed
21 January 2009).
McEvoy, S. P., Stevenson, M. R., McCartt, A. T.,
Woodward, M., Haworth, C., Palamara, P. and
Cercarelli, R. (2005) Role of mobile phones in
motor vehicle crashes resulting in hospital
attendance: a case-crossover study, British Medical
Journal. DOI:10.1136/bmj.38537.397512.55.
National Highway Traffic Safety Administration (2005)
Traffic Safety Facts 2003: A Compilation of
Motor Vehicle Crash Data from the Fatality Analysis
Reporting System and the General Estimates
System, National Center for Statistics
and Analysis, US Department of Transportation,
Washington, DC. Available at http://www-nrd.nhtsa.
dot.gov/pdf/nrd-30/NCSA/TSFAnn/2003HTMLTSF/
TSF2003.htm (accessed 21 January 2009).
National Institute on Alcohol Abuse and Alcoholism.
Available at http://www.niaaa.nih.gov/Resources/
DatabaseResources/QuickFacts/AlcoholSales/default.htm
(accessed 21 January 2009).
Peltzman, S. (1975) The effects of automobile regulations,
Journal of Political Economy, 93, 677–725.
Ramsey, J. B. (1974) Classical model selection through
specification error tests, in Frontiers in Econometrics
(Ed.) P. Zarembka, Academic Press, New York,
pp. 13–47.
Ramsey, J. B. and Zarembka, P. (1971) Specification error
tests and the alternative functional form of the aggregate
production function, Journal of the American Statistical
Association, Applications Section, 57, 471–7.
Redelmeier, D. A. and Tibshirani, R. J. (1997) Association
between cellular-telephone calls and motor vehicle
collisions, The New England Journal of Medicine, 336,
453–8.
Scherer, F. M. (1965a) Size of firm, oligopoly, and research:
a comment, Canadian Journal of Economics and
Political Science, 31, 256–66.
Scherer, F. M. (1965b) Firm size, market structure,
opportunity, and the output of patented inventions,
American Economic Review, 55, 1097–125.
Sullman, M. J. and Baas, P. H. (2004) Mobile phone use
amongst New Zealand drivers, Transportation Research
Part F: Traffic Psychology and Behavior, 7, 95–105.
US Bureau of the Census (various years) Statistical
Abstract of the United States, Washington, DC.
Violanti, J. M. (1998) Cellular phones and fatal traffic
collisions, Accident Analysis and Prevention, 30, 519–24.
2914 P. D. Loeb et al.

---

Get Professional Assignment Help Cheaply

Are you busy and do not have time to handle your assignment? Are you scared that your paper will not make the grade? Do you have responsibilities that may hinder you from turning in your assignment on time? Are you tired and can barely handle your assignment? Are your grades inconsistent?

Whichever your reason is, it is valid! You can get professional academic help from our service at affordable rates. We have a team of professional academic writers who can handle all your assignments.

Why Choose Our Academic Writing Service?

• Plagiarism free papers
• Timely delivery
• Any deadline
• Skilled, Experienced Native English Writers
• Subject-relevant academic writer
• Adherence to paper instructions
• Ability to tackle bulk assignments
• Reasonable prices
• 24/7 Customer Support
• Get superb grades consistently

 

Online Academic Help With Different Subjects

Literature

Students barely have time to read. We got you! Have your literature essay or book review written without having the hassle of reading the book. You can get your literature paper custom-written for you by our literature specialists.

Finance

Do you struggle with finance? No need to torture yourself if finance is not your cup of tea. You can order your finance paper from our academic writing service and get 100% original work from competent finance experts.

Computer science

Computer science is a tough subject. Fortunately, our computer science experts are up to the match. No need to stress and have sleepless nights. Our academic writers will tackle all your computer science assignments and deliver them on time. Let us handle all your python, java, ruby, JavaScript, php , C+ assignments!

Psychology

While psychology may be an interesting subject, you may lack sufficient time to handle your assignments. Don’t despair; by using our academic writing service, you can be assured of perfect grades. Moreover, your grades will be consistent.

Engineering

Engineering is quite a demanding subject. Students face a lot of pressure and barely have enough time to do what they love to do. Our academic writing service got you covered! Our engineering specialists follow the paper instructions and ensure timely delivery of the paper.

Nursing

In the nursing course, you may have difficulties with literature reviews, annotated bibliographies, critical essays, and other assignments. Our nursing assignment writers will offer you professional nursing paper help at low prices.

Sociology

Truth be told, sociology papers can be quite exhausting. Our academic writing service relieves you of fatigue, pressure, and stress. You can relax and have peace of mind as our academic writers handle your sociology assignment.

Business

We take pride in having some of the best business writers in the industry. Our business writers have a lot of experience in the field. They are reliable, and you can be assured of a high-grade paper. They are able to handle business papers of any subject, length, deadline, and difficulty!

Statistics

We boast of having some of the most experienced statistics experts in the industry. Our statistics experts have diverse skills, expertise, and knowledge to handle any kind of assignment. They have access to all kinds of software to get your assignment done.

Law

Writing a law essay may prove to be an insurmountable obstacle, especially when you need to know the peculiarities of the legislative framework. Take advantage of our top-notch law specialists and get superb grades and 100% satisfaction.

What discipline/subjects do you deal in?

We have highlighted some of the most popular subjects we handle above. Those are just a tip of the iceberg. We deal in all academic disciplines since our writers are as diverse. They have been drawn from across all disciplines, and orders are assigned to those writers believed to be the best in the field. In a nutshell, there is no task we cannot handle; all you need to do is place your order with us. As long as your instructions are clear, just trust we shall deliver irrespective of the discipline.

Are your writers competent enough to handle my paper?

Our essay writers are graduates with bachelor's, masters, Ph.D., and doctorate degrees in various subjects. The minimum requirement to be an essay writer with our essay writing service is to have a college degree. All our academic writers have a minimum of two years of academic writing. We have a stringent recruitment process to ensure that we get only the most competent essay writers in the industry. We also ensure that the writers are handsomely compensated for their value. The majority of our writers are native English speakers. As such, the fluency of language and grammar is impeccable.

What if I don’t like the paper?

There is a very low likelihood that you won’t like the paper.

Reasons being:

• When assigning your order, we match the paper’s discipline with the writer’s field/specialization. Since all our writers are graduates, we match the paper’s subject with the field the writer studied. For instance, if it’s a nursing paper, only a nursing graduate and writer will handle it. Furthermore, all our writers have academic writing experience and top-notch research skills.
• We have a quality assurance that reviews the paper before it gets to you. As such, we ensure that you get a paper that meets the required standard and will most definitely make the grade.

In the event that you don’t like your paper:

• The writer will revise the paper up to your pleasing. You have unlimited revisions. You simply need to highlight what specifically you don’t like about the paper, and the writer will make the amendments. The paper will be revised until you are satisfied. Revisions are free of charge
• We will have a different writer write the paper from scratch.
• Last resort, if the above does not work, we will refund your money.

Will the professor find out I didn’t write the paper myself?

Not at all. All papers are written from scratch. There is no way your tutor or instructor will realize that you did not write the paper yourself. In fact, we recommend using our assignment help services for consistent results.

What if the paper is plagiarized?

We check all papers for plagiarism before we submit them. We use powerful plagiarism checking software such as SafeAssign, LopesWrite, and Turnitin. We also upload the plagiarism report so that you can review it. We understand that plagiarism is academic suicide. We would not take the risk of submitting plagiarized work and jeopardize your academic journey. Furthermore, we do not sell or use prewritten papers, and each paper is written from scratch.

When will I get my paper?

You determine when you get the paper by setting the deadline when placing the order. All papers are delivered within the deadline. We are well aware that we operate in a time-sensitive industry. As such, we have laid out strategies to ensure that the client receives the paper on time and they never miss the deadline. We understand that papers that are submitted late have some points deducted. We do not want you to miss any points due to late submission. We work on beating deadlines by huge margins in order to ensure that you have ample time to review the paper before you submit it.

Will anyone find out that I used your services?

We have a privacy and confidentiality policy that guides our work. We NEVER share any customer information with third parties. Noone will ever know that you used our assignment help services. It’s only between you and us. We are bound by our policies to protect the customer’s identity and information. All your information, such as your names, phone number, email, order information, and so on, are protected. We have robust security systems that ensure that your data is protected. Hacking our systems is close to impossible, and it has never happened.

How our Assignment  Help Service Works

1.      Place an order

You fill all the paper instructions in the order form. Make sure you include all the helpful materials so that our academic writers can deliver the perfect paper. It will also help to eliminate unnecessary revisions.

2.      Pay for the order

Proceed to pay for the paper so that it can be assigned to one of our expert academic writers. The paper subject is matched with the writer’s area of specialization.

3.      Track the progress

You communicate with the writer and know about the progress of the paper. The client can ask the writer for drafts of the paper. The client can upload extra material and include additional instructions from the lecturer. Receive a paper.

4.      Download the paper

The paper is sent to your email and uploaded to your personal account. You also get a plagiarism report attached to your paper.

PLACE THIS ORDER OR A SIMILAR ORDER WITH US TODAY AND GET A PERFECT SCORE!!!

---