RESEARCH AND PRACTICE

Undocumented Immigration, Drug Problems, and
Driving Under the Inﬂuence in the United States,
1990–2014
Michael T. Light, PhD, Ty Miller, MS, and Brian C. Kelly, PhD
Objectives. To examine the inﬂuence of undocumented immigration in the United
States on 4 different metrics of drug and alcohol problems: drug arrests, drug overdose
fatalities, driving under the inﬂuence (DUI) arrests, and DUI deaths.
Methods. We combined newly developed state-level estimates of the undocumented
population between 1990 and 2014 from the Center for Migration Studies with arrest
data from the Federal Bureau of Investigation Uniform Crime Reports and fatality information from the Fatality Analysis Reporting System and the Centers for Disease
Control and Prevention Underlying Cause of Death database. We used ﬁxed-effects
regression models to examine the longitudinal association between increased undocumented immigration and drug problems and drunk driving.
Results. Increased undocumented immigration was signiﬁcantly associated with reductions in drug arrests, drug overdose deaths, and DUI arrests, net of other factors.
There was no signiﬁcant relationship between increased undocumented immigration
and DUI deaths.
Conclusions. This study provides evidence that undocumented immigration has not
increased the prevalence of drug or alcohol problems, but may be associated with reductions in these public health concerns. (Am J Public Health. Published online ahead of
print July 20, 2017: e1–e7. doi:10.2105/AJPH.2017.303884)

D

riving under the inﬂuence (DUI) and
drug problems remain pressing public
health issues and foci for prevention and intervention efforts. Since 1990, nearly 1 million persons in the United States have died of
drug overdoses and alcohol-related crashes.1,2
Each year, alcohol-impaired driving fatalities
account for more than one third of drivingrelated deaths in the United States.3 The
economic costs of drug use and drunk driving
total more than $240 billion annually.4,5 In
addition, the rate of drug overdose fatalities
increased 137% between 2000 and 2014 and
more people died of drug overdoses in 2014
than in any previous year on record.6 These
public health trends have coincided with
a substantial inﬂux of undocumented immigrants, from a population of 3.5 million in
1990 to an estimated 10.9 million in 2014,7
sparking controversy as to whether undocumented immigration is related to the
prevalence of drug problems and DUI.8,9

Published online ahead of print July 20, 2017

AJPH

Indeed, within political discourse on undocumented immigration, public health and
safety concerns regarding drug problems and
drunk driving have received considerable
attention from public ofﬁcials. In defending
Arizona’s controversial SB 1070 law, which
required law enforcement ofﬁcers to determine an individual’s immigration status
during a stop, in 2010, then-Governor Jan
Brewer stated “. . . the majority of the illegal
trespassers that are coming in the state of
Arizona are under the direction and control of
organized drug cartels, and they are bringing
drugs in.”10 Similarly, in 2006, an Iowa
congressman claimed that 13 US persons died

every day as a result of undocumented drunk
drivers.8 Such controversies have achieved
national headlines as presidential candidates
have also made claims that undocumented
immigrants increase drug activity.11,12
However, despite public comments and
debate on these issues, empirical research
on the public health consequences of undocumented immigration is lacking because
of the scarcity of reliable data on the undocumented population.7 We ﬁll this gap by
utilizing newly developed state-level estimates of the undocumented population between 1990 and 2014 to provide the ﬁrst
investigation into the relationship (if any)
between undocumented immigration, DUI,
and drug problems.
The literature on undocumented immigration and health primarily focuses on health
care access and utilization.13,14 As a result,
evidence associating undocumented immigration to trends in DUI and drug problems is
indirect and often mixed. On the basis of
current debates and the existing literature,
there are 2 plausible paths relating undocumented immigration to DUI and drug
problems, one that considers a facilitating
relationship, and the other that considers an
inhibiting association.
Regarding the ﬁrst pathway, several factors
are important to consider. First, theories of
addiction suggest that chronic stress plays an
important role in substance abuse,15 and research demonstrates a relationship among
undocumented status, higher stress, and decreased mental health.16 Thus, because undocumented immigrants often experience

ABOUT THE AUTHORS
At the time of the study, all of the authors were with the Department of Sociology, Purdue University, West Lafayette, IN.
Correspondence should be sent to Michael T. Light, 8128 William H. Sewell Social Sciences Building, 1180 Observatory Drive,
University of Wisconsin-Madison, Madison, WI 53706 (e-mail: mlight@ssc.wisc.edu). Reprints can be ordered at http://www.ajph.
org by clicking the “Reprints” link.
This article was accepted April 23, 2017.
doi: 10.2105/AJPH.2017.303884

Light et al.

Peer Reviewed

Research and Practice

e1

 RESEARCH AND PRACTICE

stress from fear of exploitation and deportation, they may use substances to reduce
negative affect, in line with the stress–coping
model of addiction.17 Second, undocumented immigrants display risk factors for
alcohol and drug dependence, including
poverty and low educational attainment, at
higher rates than native-born groups.18,19 In
line with economic theories of criminal behavior and Goldstein’s tripartite model of the
drug–crime nexus,20,21 undocumented immigrants may turn to illegal economic pursuits, such as drug trafﬁcking, because of
formal exclusion from the labor market
combined with their lack of economic assets.
Moreover, a sizable proportion of illegal
drugs in the United States, particularly cocaine and heroin, is smuggled across the
US–Mexico border,22 potentially relating
undocumented immigration ﬂows to drug
trafﬁcking, as has been argued within the
political sphere.10–12 Taken together, these
considerations support a hypothesis that undocumented immigration may increase the
frequency of DUI and drug problems.
In contrast with this hypothesis, it is also
plausible that undocumented immigration
may reduce the prevalence of DUI and drug
problems. Undocumented immigrants are
often motivated by economic opportunities
for themselves and their families, and thus
may be selected on attributes (e.g., ambition)
that predispose them to less criminal involvement and healthier behaviors.23 In line
with this view, research indicates that immigrants in general are less likely to abuse
drugs and alcohol than are their US-born
counterparts24,25 and these protective effects
extend to the children of immigrants.26
Research suggests that protective inﬂuences
may stem from living in immigrant enclaves,
which, because of processes of chain migration, are often characterized by dense networks of immigrant families and neighbors
who reinforce social cohesion and provide
social control to monitor criminal and risky
behaviors.23
For example, research suggests that
living among higher concentrations of recent immigrants is protective against substance use for Mexican youths,27 notable
ﬁndings given that the majority of undocumented immigrants are from Mexico.28
In addition, awareness of police surveillance
and the related threat of criminal deportation

e2

Research and Practice

Peer Reviewed

may reduce the likelihood of consumption
of illegal drugs and drunk driving by undocumented immigrants, especially in public
locales such as bars and nightclubs, or reduce
the likelihood of driving at all. Taken
together, this body of literature supports
a hypothesis that undocumented immigration may decrease the frequency of DUI and
drug problems.
Within our evaluation of contrasting hypotheses regarding the inﬂuence of undocumented immigration on drug and
alcohol problems, we considered 2 behavioral
processes through the outcomes under consideration. The ﬁrst outcome, DUI, allows us
to consider the health behaviors of undocumented immigrants on wider community health and safety. As a focal concern
regarding undocumented immigration and
DUI has centered on the potential for drunk
driving on the part of undocumented immigrants themselves, we consider this direct
behavioral impact on public health. By
contrast, the focal concern on drugs is that
undocumented immigrants affect the behaviors of others in the community by providing opportunities for illegal drug use. For
this reason, the drug outcomes permit us to
consider the inﬂuence of undocumented
immigration on the behaviors of others and
wider public health.

METHODS
We tested the contrasting hypotheses with
data from multiple sources at the state level
from 1990 to 2014. Data for DUI and drug
arrest came from the Federal Bureau of Investigation Uniform Crime Reports, which
compiles data from more than 18 000 law
enforcement agencies representing 98% of
the population in 2014.29 Data on alcoholrelated crash fatalities came from the National
Highway Trafﬁc Safety Administration’s
Fatality Analysis Reporting System, and drug
overdose data are from the Centers for Disease Control and Prevention (CDC) Underlying Cause of Death database. The
Fatality Analysis Reporting System and CDC
data sets are nationwide censuses. Information
on undocumented immigration comes from
2 of the most reliable sources of data on the
unauthorized population: the Center for Migration Studies and the Pew Research Center.

Light et al.

In addition, we collected data on an array
of socioeconomic, demographic, and criminal justice characteristics from the US Census,
the National Prisoner Statistics, and the
Federal Bureau of Investigation Police Employment data. Census measures came from 1
of 3 sources: the 5% Public Use Microdata,
the American Community Survey (ACS), or
the Current Population Survey. The ACS is
an annual nationally representative survey of
approximately 3 million US households,
generating estimates of social and economic
characteristics of the US population. Response to the ACS is required by law, ensuring
high response rates (approximately 96%).30
The Current Population Survey, like the
ACS, is a nationally representative survey of
approximately 60 000 households, used to
calculate employment and labor market statistics. Table 1 displays the sources, properties,
and descriptive statistics for all variables.

Dependent Variables
We used 4 dependent variables. We
measured problems related to drugs by using
the rate of arrests for drug violations (per
100 000) and the rate of unintentional drug
overdose deaths (per 100 000). We used 2
measures because arrests reﬂect a mixture of
use and policing behavior. Overdose fatalities,
by contrast, are unlikely to track changes in
policing, and thus more accurately reﬂect the
prevalence of heavy drug use.31 Using the
same logic, we measured alcohol-impaired
driving by using the DUI arrest rate (per
100 000) and the rate of fatalities involving
drunk driving (per 100 000), deﬁned by the
National Highway Trafﬁc Safety Administration as a fatality (driver, passenger, or pedestrian) involving a crash in which a driver’s
blood alcohol content is 0.08 or greater. We
calculated rates for each dependent variable
by using the total population; however, results were unchanged when we calculated
rates with the adult population (aged ‡ 18
years), driving population (aged ‡ 16 years),
or population of legal drinking age (aged ‡ 21
years). We logged all outcome measures to
reduce positive skewness (analyses using
unlogged outcomes available in Table A,
available as a supplement to the online version
of this article at http://www.ajph.org).
As shown in Table 1, because of a small
amount of missing data in Uniform Crime

AJPH

Published online ahead of print July 20, 2017

 RESEARCH AND PRACTICE

TABLE 1—Summary Statistics for State-Level Dependent and Independent Measures, in an Examination of Undocumented Immigration and
Drug Arrests, Driving Under the Inﬂuence Arrests, Drug Overdose Mortality, and Driving Under the Inﬂuence Mortality: United States, 1990–
2014
Measures

Drug Arrest, Mean (SD)

Drug Mortality, Mean (SD)

DUI Arrest, Mean (SD)

DUI Fatality, Mean (SD)

Dependent variables
Drug offense arrest rate per 100 000 population (logged)a

5.8 (0.7)
2.0 (0.7)

Unintentional drug overdose mortality rate per 100 000
population (logged)b
DUI offense arrest rate per 100 000 population (logged)a

5.8 (0.8)

DUI fatality rate per 100 000 population (logged)c

1.5 (0.5)

Focal measures, estimated proportion of population
Undocumented immigrants—Center for Migration Studies

1.9 (1.8)

1.9 (1.7)

1.9 (1.8)

1.9 (1.7)

Undocumented immigrants—Pew Research Center

2.1 (1.8)

2.1 (1.7)

2.1 (1.8)

2.1 (1.7)

Covariates
Legal immigrants, proportion of the populationd
Poverty, proportion of the populatione
Low educational attainment (aged > 25 y without high-school

6.4 (4.8)

6.5 (4.8)

6.4 (4.8)

6.5 (4.8)

13.2 (3.7)
15.4 (5.2)

13.2 (3.7)
15.4 (5.2)

13.2 (3.7)
15.4 (5.2)

13.2 (3.7)
15.4 (5.2)

10.6 (10.2)

10.7 (10.7)

10.6 (10.2)

10.7 (10.7)

degree), proportion of the populatione
Racial composition: non-Hispanic Black, proportion of the
populatione
Unemployment, proportion of the populationf

5.8 (1.9)

5.7 (1.9)

5.8 (1.9)

5.7 (1.9)

Employed in manufacturing industry,f %

12.7 (5.3)

12.6 (5.4)

12.6 (5.3)

12.6 (5.4)

Employed in managerial or professional occupations,f %

31.9 (5.7)

31.9 (5.7)

32.0 (5.7)

31.9 (5.7)

Age structure: 18–24 y,e %
Population density per square mile,g no.

9.9 (0.9)
235.9 (718.3)

9.9 (0.9)
274.3 (944.2)

9.9 (0.9)
235.7 (719.5)

9.9 (0.9)
274.3 (944.2)

Incarceration rateh,i per 100 000 population

385.0 (182.5)

386.5 (191.1)

384.6 (182.5)

386.5 (191.1)

Police ofﬁcers per 100 000 populationa

213.9 (60.0)

216.0 (67.3)

213.7 (60.0)

216.0 (67.3)

1210

1259

1206

1259

Number of state-years

Note. DUI = driving under the inﬂuence. Number of observations varies because of differences in missing information across dependent variables.
a
According to Uniform Crime Reports.
b
According to Centers for Disease Control and Prevention Underlying Cause of Death database.
c
According to Fatality Analysis Reporting System.
d
According to US Census Bureau/Public Use Microdata.
e
According to US Census Bureau/American Community Survey.
f
According to US Bureau of Labor Statistics.
g
According to US Census Bureau.
h
According to Bureau of Justice Statistics National Prisoner Statistics.
i
Incarceration information is missing for Washington, DC, after 2001, when the city abandoned its prison system, and for 1 year in Arizona (2013).

Reports statistics, the number of state-years
varies between the mortality and arrest
variables. For both DUI and drug arrests, 4%
of state-years are missing. To ensure comparability across samples, we show descriptive statistics of all covariates separately
for each outcome. With 1 exception (population density), sample means do not vary
by more than 1%, providing little evidence
that the samples are biased by missing cases.
We also reran the analysis without the
population density measure as a robustness
check, and the results were substantively
identical.

Published online ahead of print July 20, 2017

AJPH

Undocumented Immigration
The focal measure in the analysis is the
proportion of the total population that is
undocumented. Although state-level undocumented estimates are available from the
Center for Migration Studies for the full study
period, Pew estimates are only available intermittently up until 2005. For this reason,
our main analysis utilized Center for Migration Studies data and we replicated the results
with Pew data. The use of data from 2 independent sources, which used disparate
methodologies, serves as a robustness check.

Pew Research Center estimates. The Pew
counts are perhaps the most widely utilized
estimates of the unauthorized population. In
calculating the undocumented estimates, Pew
uses a residual methodology based on Census
data. The residual method and its variants are
widely used and accepted as the best estimates
of the undocumented population.18 Stated
brieﬂy, this method involves ﬁrst subtracting
the number of authorized immigrants from
the total foreign-born population, and then
uses probabilistic methods based on demographic, social, economic, and geographic
characteristics to classify the residual as either

Light et al.

Peer Reviewed

Research and Practice

e3

 RESEARCH AND PRACTICE

lawful or undocumented. For example, those
in the residual with occupations that require
legal status or receive public beneﬁts restricted
to legal residents are classiﬁed as lawful.
Independent research using various triangulation methods, including death and
birth records, substantiates the general accuracy of the Pew estimates.32 Pew reports
state-level unauthorized estimates for 1990,
1995, and 2000, then 2005 to 2014. We used
linear interpolation to account for missing
estimates in the intervening years.
Center for Migration Studies estimates. Like
Pew, the Center for Migration Studies uses
the residual method based on Census data but
improves upon this technique by accounting
for components of population change (e.g.,
emigration rates, undercount rates, removals,
adjustments to lawful status, and mortality
rates) and calculating independent population controls by country of origin for undocumented residents.7 This methodology
has been empirically vetted in peer-reviewed
research and shown to produce smaller ranges
of sampling error than alternative sources.7,33

Covariates
We included a host of state-level sociodemographic characteristics that may confound associations between undocumented
immigration, DUI, and drug problems. Of
particular importance, we included a measure
of the legal immigrant population (proportion
of the population that are lawful immigrants)
to separate the effects of undocumented
immigration from general immigration
trends. Drawing on previous research, we
adjusted for known correlates that capture the
entanglement of poverty, racial composition,
and public health problems. These include the
poverty rate,34 the percentage non-Hispanic
Black, and the proportion of the population
older than 25 years without a high-school
degree (i.e., low educational attainment). We
include measures of the labor market and
macro-economic climate: unemployment
rate,35 percentage of people in the manufacturing industry, and the percentage of
people employed in managerial or professional occupations. Given the comparatively younger age structure of the
unauthorized population,18 we included
a measure for the proportion of the population aged between 18 and 24 years—a life

e4

Research and Practice

Peer Reviewed

course period when substance use peaks.36
Lastly, we accounted for differences in population density and 2 salient measures of law
enforcement, the incarceration rate and the
police per capita.

Analytical Method
We leveraged the longitudinal nature of
our data by employing ﬁxed-effects regression methods to estimate the effect of
undocumented immigration on DUI and
drug problems. The principal strength of this
approach is that we can treat each state as its
own statistical control.37 As such, we
accounted for between-state differences and
examined only within-state variation over
time, thus eliminating the effects of timestable measured and unmeasured confounding variables, including cross-state dissimilarities in arrest and fatality reporting.
We also included year ﬁxed effects to
account for unmeasured trends that inﬂuenced DUI and drug problems nationally. A
strength of ﬁxed effects models is the ability to
address issues of measurement error in the
undocumented estimates. For example, to the
extent that there is a national pattern of systematic under- or overcounting of the undocumented population, the model
parameters are adjusted for this by the year
ﬁxed effects. Moreover, the state ﬁxed effects
account for unique challenges to estimating
the unauthorized population within each
state. In all models, we reported standard
errors clustered by state and lagged the independent variables by 1 year to ensure that
changes in undocumented immigration
precede any ﬂuctuations in DUI and drug
problems.

RESULTS
Table 2 reports associations between undocumented immigration and drug arrests,
drug overdose fatalities, DUI arrests, and DUI
deaths. For each outcome, we report the full
analytical speciﬁcation, inclusive of all covariates and state and year ﬁxed effects. To gauge
both statistical and substantive signiﬁcance,
we report standardized (B) and unstandardized (b) regression coefﬁcients.
Beginning with the pattern in the ﬁndings, the direction of the coefﬁcient for

Light et al.

undocumented immigration was negative for
all outcomes. For drug arrests, drug overdose
deaths, and DUI arrests, these associations
were statistically signiﬁcant. Moreover, an
examination of the standardized coefﬁcients,
which account for differences in metrics
across independent variables, suggests that
unauthorized immigration is a comparatively meaningful measure in these models.
To offer a substantive interpretation, the
results with unlogged outcomes showed that
a 1-unit increase in the proportion of the
population that is undocumented is associated with 22 fewer drug arrests (per 100 000),
0.64 fewer drug overdose deaths (per
100 000), and 42 fewer DUI arrests (per
100 000). For DUI fatalities, the coefﬁcient
was small and not signiﬁcant, suggesting that
the frequency of drunk driving deaths is
generally unaffected by unauthorized
immigration.
To aid interpretation, we provide a graphic
display of the effects of undocumented immigration on drug and alcohol outcomes in
Figure 1, based on the results from Table 2.
Taken together, the ﬁndings offer no evidence that unauthorized immigration increases the rate of DUI or drug problems.
Rather, the pattern in the data suggests that
undocumented immigration reduced the
prevalence of DUI and drug problems over
the past 25 years or had no effect.
We ran additional analyses to test for
confounding by other measures in our analysis. Speciﬁcally, we ran models in which we
only included the unauthorized measure with
the state and year ﬁxed effects. For drug
arrests, overdose deaths, and DUI arrests,
increased undocumented immigration was
signiﬁcantly associated with reductions in
drug and alcohol problems. For DUI deaths,
the relationship was negative but not significant, in line with results reported in Table 2.
We also investigated the impact of outliers by
replicating our analysis with robust regression
models, which omit observations with
Cook’s distance greater than 1 and weights
the data so that observations with smaller
residuals are given greater weight in the
analysis (reported in Table B, available as
a supplement to the online version of this
article at http://www.ajph.org). The results
showed the same patterns, suggesting that the
ﬁndings were not driven by outliers in the
data.

AJPH

Published online ahead of print July 20, 2017

 RESEARCH AND PRACTICE

TABLE 2—Fixed Effects Models Examining the Relationship Between Undocumented Immigration, Drug Problems, and Driving Under the
Inﬂuence at the State Level: United States, 1990–2014
(ln) Drug Arrest Rate
(n = 1164)
Variable

(ln) Drug Fatality Rate
(n = 1210)

(ln) DUI Arrest Rate
(n = 1160)

(ln) DUI Fatalities
(n = 1210)

b (95% CI)

B

b (95% CI)

B

b (95% CI)

B

–0.20 (–0.35, –0.05)

–0.55

–0.13 (–0.19, –0.06)

–0.32

–0.17 (–0.30, –0.04)

–0.36

–0.0001 (–0.05, 0.05)

–0.06 (–0.17, 0.05)

–0.46

–0.07 (–0.12, –0.02)

–0.48

–0.01 (–0.09, 0.07)

–0.06

–0.01 (–0.04, 0.03)

0.00 (–0.03, 0.03)

0.00

0.01 (–0.01, 0.02)

0.04

0.01 (–0.02, 0.04)

0.03

–0.01 (–0.01, 0.00)

–0.04

–0.01 (–0.05, 0.03)

–0.07

–0.02 (–0.03, 0.00)

–0.12

0.02 (–0.01, 0.06)

0.14

–0.02 (–0.03, –0.01)

–0.21

Percentage Black

0.05 (–0.10, 0.21)

0.87

0.03 (–0.03, 0.10)

0.50

0.17 (–0.01, 0.34)

2.08

–0.01 (–0.06, 0.04)

–0.20

Unemployment

0.00 (–0.05, 0.05)

0.00

0.03 (0.00, 0.05)

0.07

–0.01 (–0.05, 0.04)

–0.02

–0.03 (–0.05, –0.01)

–0.13

Manufacturing

0.04 (0.00, 0.08)

0.32

–0.01 (–0.03, 0.01)

–0.11

0.00 (–0.05, 0.05)

–0.01

–0.01 (–0.03, 0.00)

–0.16

Managerial or professional
Aged 18–24 y

0.00 (–0.04, 0.03)
0.05 (–0.03, 0.13)

–0.02
0.07

0.00 (–0.02, 0.01)
–0.05 (–0.10, 0.01)

–0.01
–0.06

0.00 (–0.03, 0.02)
0.04 (–0.04, 0.12)

–0.03
0.04

0.00 (–0.01, 0.01)
0.04 (0.00, 0.07)

–0.01
0.07

Focal variable: undocumented

b (95% CI)

B
–0.0003

immigration—Center for Migration Studies
Covariates
Legal immigration
Poverty
Low educational attainment

–0.06

Population density

0.006 (0.002, 0.010)

6.51

–0.002 (–0.003, –0.001)

–2.65

0.005 (0.002, 0.008)

4.53

0.000 (–0.001, 0.001)

–0.24

Incarceration rate

0.000 (0.000, 0.001)

0.14

0.0004 (0.0003, 0.0009)

0.12

0.000 (–0.001, 0.001)

0.10

0.000 (0.000, 0.000)

–0.06

Police per capita

0.002 (–0.001, 0.004)

0.17

0.000 (–0.001, 0.001)

0.05

0.001 (–0.003, 0.004)

0.04

0.000 (–0.001, 0.000)

–0.07

Constant

1.00 (–2.57, 4.56)

0.90 (–0.73, 2.54)

Yes

Yes

–0.47 (–4.22, 3.28)

3.57 (1.99, 5.15)

Speciﬁcation and summary information
State ﬁxed effects

Yes

Yes

Year ﬁxed effects

Yes

Yes

Yes

Yes

R2

0.57

0.91

0.69

0.90

Note. CI = conﬁdence interval; DUI = driving under the inﬂuence; ln = natural log. All independent variables are lagged by 1 year. Standard errors are corrected for
clustering. Number of observations varies because of differences in missing information across dependent variables.

Given the inherent difﬁculties in estimating the undocumented population, we
provide a robustness check by replicating our
main ﬁndings with Pew estimates of the
unauthorized population (Table C, available
as a supplement to the online version of this
article at http://www.ajph.org) by using the
same analytical speciﬁcation shown in Table
2. Despite the use of a somewhat dissimilar
methodology as the Center for Migration
Studies, the pattern of results from this alternative analysis with Pew data was substantively identical to our main ﬁndings:
increased undocumented immigration was
associated with statistically signiﬁcant decreases in drug arrests, drug overdose deaths,
and DUI arrests. For DUI fatalities, the coefﬁcient was negative but not signiﬁcant.

DISCUSSION
Political rhetoric on the public health
consequences of undocumented immigration

Published online ahead of print July 20, 2017

AJPH

has far outpaced empirical research. This
study takes a step toward ﬁlling this gap by
providing the ﬁrst longitudinal assessment
of the relationship between increased undocumented immigration, DUI, and drug
problems. We tested 2 competing hypotheses. The ﬁrst hypothesis drew from the
stress–coping model of substance use and
economic theories of crime, which suggested that undocumented immigration may
increase public health problems such as DUI
and illicit drug use.19–21 We found no evidence to support this hypothesis. The second
hypothesis tested an alternative contention
emerging from the existing literature on
selective migration and the healthy immigrant hypothesis, which suggested that undocumented immigration may lead to
decreases in these public health problems.23–27
We found stronger evidence to support
this hypothesis, with unauthorized immigration associated with signiﬁcant reductions
in drug arrests, DUI arrests, and drug
overdoses.

Importantly, we considered 2 behavioral
mechanisms by which these outcomes may
emerge; on one hand, undocumented immigrants may have a direct impact on public
health via their own health behaviors
(DUI), whereas on the other hand, they
may potentially have an impact on public
health by inﬂuencing the behaviors of
others in the community (drugs). We found
evidence for both mechanisms. Regarding
illicit drug use, our ﬁndings challenge claims
that increased prevalence of drug problems
results from unauthorized immigration. We
also found evidence that undocumented
immigration may be associated directly with
public health via behaviors of immigrants,
given the results for reduced DUI arrests.
Undocumented persons may be less likely
to drive after drinking, or drive at all, because of fear of police surveillance and
deportation. For similar reasons, they may
socialize and drink primarily within immigrant enclaves, obviating the need to
drive.

Light et al.

Peer Reviewed

Research and Practice

e5

 RESEARCH AND PRACTICE

Predicted Rates per 100 000

600

400

200

0
-2

0

2

4

6

Change in Proportion Undocumented
DUI arrests

Drug arrests

Drug deaths

DUI deaths

Public Health Implications

Note. DUI = driving under the inﬂuence.

FIGURE 1—Predicted Relationship Between Undocumented Immigration and Drug Arrests,
Driving Under the Inﬂuence Arrests, Drug Overdose Mortality, and Driving Under the
Inﬂuence Mortality: United States, 1990–2014

Strengths and Limitations
Our analysis had several strengths that
bolster these conclusions. First, we triangulated the ﬁndings by using different
measures of DUI and drugs. The chief advantage was to ensure that the pattern of
results was robust from multiple approaches
and across data sources with distinct aims and
reporting procedures. Second, we utilized
state ﬁxed effects, which afford considerable
analytical leverage to understand the associations between undocumented immigration
and rates of DUI and drug problems by focusing on within-state changes. This method
removes the effects of time-invariant causes of
drug and alcohol problems (measured or not)
that potentially confound the undocumented
immigration–DUI or –drug relationship. We
provided an additional analytical advantage
by lagging the independent variables to ensure the time ordering of the relationship.
However, we also tested this decision by
rerunning the analysis without lagged predictors (Table D, available as a supplement to
the online version of this article at http://
www.ajph.org). It is possible that unauthorized immigrants have a contemporaneous effect on drug and alcohol problems,

e6

Research and Practice

Peer Reviewed

causal inference, we cannot fully attribute
causal relationships within this design.
A ﬁnal limitation concerns measurement
error in the unauthorized estimates. We
minimized this concern in several ways. First,
the use of state and year ﬁxed effects accounts
for any national miscounting, as well as
unique challenges to estimating the unauthorized population in each state, as long as
measurement error is stable over time. Second, we leveraged 2 of the most widely used
sources of the undocumented population.
Indeed, these are sources referenced by those
who claim the size of the undocumented
population is too large.39 The fact that our
results align with use of either set of estimates
buttresses our ﬁndings.

particularly if unauthorized immigration is
associated with drug smuggling. However,
we found no evidence for this view.
Rather, the ﬁndings align with those reported in the main analysis: unauthorized
immigration is negatively associated with
drug arrests, overdose deaths, and DUI
arrests, and is not associated with DUI
fatalities.
Despite the analyses’ strengths, there are
limitations. Notably, like all macrolevel
health research,38 the ecological fallacy is
important to consider. When we use statelevel data, we are unable to discern whose
behavior accounts for these patterns. For
example, undocumented immigration may
be associated with decreased DUI and drug
arrests if law enforcement directed attention
toward other matters as the unauthorized
population increased. Alternatively, unauthorized immigrants may be less likely to
abuse drugs and alcohol, as the literature
indicates occurs for immigrants more generally.25,26 In short, future research is needed
to elucidate these mechanisms, a task beyond
the scope of the data currently available. A
second consideration is that, although ﬁxed
effects help ameliorate some concerns about

Light et al.

Whether undocumented immigration
jeopardizes public health is an important
question at the center of contemporary debates on unauthorized immigration. These
debates are certain to continue, but they
should do so informed by the available evidence, and our study speaks directly to this
controversy. Our study does not contradict
assertions that drugs are smuggled across the
US border (they are) or that individual undocumented immigrants have been arrested
for drunk driving (they have). Our ﬁndings
do, however, signiﬁcantly undermine arguments that the public is at greater risk for DUI
or drug problems as a result of undocumented
immigration. If anything, they suggest the
opposite. Thus, public resources should be
directed toward effective prevention and
intervention efforts to reduce drug and alcohol problems, such as those supported by
the CDC, rather than channeled to concerns
unsupported by empirical evidence.
CONTRIBUTORS
M. T. Light originated the study, conducted the analysis,
and wrote the ﬁrst draft of the article. T. Miller was responsible for data collection and data management. B. C.
Kelly helped develop the theoretical framing and contributed to drafting the article. All authors interpreted the
ﬁndings together and edited multiple drafts of the article.

ACKNOWLEDGMENTS
The authors thank Shawn Bauldry for valuable comments
on earlier drafts of this article.

HUMAN PARTICIPANT PROTECTION
This study used administrative records and not human
participants.

AJPH

Published online ahead of print July 20, 2017

 RESEARCH AND PRACTICE

REFERENCES
1. National Highway Trafﬁc Safety Administration. Fatality Analysis Reporting System. Available at: https://
www-fars.nhtsa.dot.gov/Main/index.aspx. Accessed
November 30, 2016.
2. Centers for Disease Control and Prevention, National
Center for Health Statistics. Multiple cause of death data
1990–2014. 2016. Available at: https://wonder.cdc.gov/
mcd.html. Accessed November 30, 2016.
3. National Highway Trafﬁc Safety Administration.
Trafﬁc safety facts 2014: A compilation of motor vehicle
crash data from the Fatality Analysis Reporting System
and the General Estimates System. Washington, DC:
Department of Transportation; 2015.
4. Blincoe L, Seay A, Zaloshnja E, et al. The economic
impact of motor vehicle crashes. Washington, DC:
National Highway Trafﬁc Safety Administration; 2002.
DOT HS 809 446.
5. US Department of Justice. National drug threat
assessment 2011. Johnstown, PA: National Drug
Intelligence Center; 2011.

17. Wills TA, Filer M. Stress–coping model of adolescent
substance use. In: Ollendick TH, Prinz RJ, eds. Advances
in Clinical Child Psychology. New York, NY: Springer US;
1996: 91–132.

36. Black SA, Markides K. Acculturation and alcohol
consumption in Puerto Rican, Cuban American, and
Mexican-American women in the United States. Am J
Public Health. 1993;83(6):890–893.

18. Passel JS, Cohn D. A Portrait of Unauthorized Immigrants
in the United States. Washington, DC: Pew Research
Center; 2009.

37. Allison PD. Fixed effects regression methods for
longitudinal data using SAS. Cary, NC: SAS Institute Inc;
2005.

19. Szaﬂarski M, Cubbins LA, Ying J. Epidemiology of
alcohol abuse among US immigrant populations. J Immigr
Minor Health. 2011;13(4):647–658.

38. Catalano R. The health effects of economic insecurity.
Am J Public Health. 1991;81(9):1148–1152.

20. Becker G. Crime and punishment: an economic
analysis. J Polit Econ. 1968;76(2):169–217.
21. Goldstein PJ. The drugs/violence nexus: a tripartite
conceptual framework. J Drug Iss. 1985;15(4):493–506.
22. US Department of Justice Drug Enforcement Administration. 2015 National Drug Threat Assessment Summary. Washington, DC: DEA Strategic Intelligence
Section; 2015.
23. Sampson R. Rethinking crime and immigration.
Contexts. 2008;7(1):28–33.

6. Rudd RA, Aleshire A, Zibbell JE, Gladden RM. Increases in drug and opioid overdose deaths—United
States, 2000–2014. Am J Transplant. 2016;16(4):
1323–1327.

24. Alvarez J, Jason LA, Olson BD, Ferrari JR, Davis MI.
Substance abuse prevalence and treatment among
Latinos and Latinas. J Ethn Subst Abuse. 2007;6(2):
115–141.

7. Warren R, Warren JR. Unauthorized immigration to
the United States: annual estimates and components of
change, by state, 1990 to 2010. Int Migr Rev. 2013;47(2):
296–329.

25. Abraído-Lanza AF, Chao MT, Flores KR. Do healthy
behaviors decline with greater acculturation? Implications
for the Latino mortality paradox. Soc Sci Med. 2005;61(6):
1243–1255.

8. Beirich H. Getting immigration facts straight. Southern
Poverty Law Center. 2010. Available at: https://www.
splcenter.org/ﬁghting-hate/intelligence-report/2010/
getting-immigration-facts-straight. Accessed November
30, 2016.

26. Ortega AN, Rosenheck R, Alegria M, Desai RA.
Acculturation and the lifetime risk of psychiatric and
substance abuse disorders among Hispanics. J Nerv Ment
Dis. 2000;188(11):728–735.

9. Barton E. America’s heroin epidemic fueled by ﬂood of
illegal immigrants. The Daily Caller News Foundation.
2015. Available at: http://dailycaller.com/2015/09/14/
americas-heroin-epidemic-fueled-by-ﬂood-of-illegalimmigrants. Accessed November 30, 2016.
10. Jacobson L. Arizona Gov. Brewer says “majority” of
illegals are “drug mules.” POLITIFACT. 2010. Available
at: http://www.politifact.com/truth-o-meter/statements/
2010/jun/30/jan-brewer/arizona-gov-brewer-saysmajority-illegals-are-drug. Accessed November 30,
2016.
11. Benen S. Cruz sees border wall as solution to drug
abuse. MSNBC. 2016. Available at: http://www.msnbc.
com/rachel-maddow-show/cruz-sees-border-wallsolution-drug-abuse. Accessed November 30, 2016.
12. Greenberg J, Qiu L. Trump says heroin pours across
the southern border. POLITIFACT. 2016. Available at:
http://www.politifact.com/truth-o-meter/statements/
2016/oct/19/donald-trump/trump-says-heroin-poursacross-southern-border. Accessed November 30, 2016.
13. López-Cevallos D. Are Latino immigrants a burden to
safety net services in nontraditional immigrant states?
Lessons from Oregon. Am J Public Health. 2014;104(5):
781–786.
14. Arredondo A. Health coverage for undocumented
immigrants: barriers on the Mexican side for effective
coverage. Am J Public Health. 2016;106(12):e4–e5.
15. Sinha R. Chronic stress, drug use, and vulnerability to
addiction. Ann N Y Acad Sci. 2008;1141(1):105–130.
16. Gonzales RG, Suarez-Orozco C, Dedios-Sanguineti
MC. No place to belong: contextualizing concepts of
mental health among undocumented immigrant youth in
the United States. Am Behav Sci. 2013;57(8):1174–1199.

Published online ahead of print July 20, 2017

39. Federation for American Immigration Reform. Illegal
aliens taking US jobs. 2013. Available at: http://www.
fairus.org/issue/illegal-aliens-taking-u-s-jobs. Accessed
November 30, 2016.

AJPH

27. Kulis S, Marsiglia FA, Sicotte D, Nieri T. Neighborhood effects on youth substance use in a southwestern
city. Sociol Perspect. 2007;50(2):273–301.
28. Krogstad JM, Passel JS, Cohn D. 5 Facts about illegal
immigration in the US Pew Research Center. 2016.
Available at: http://www.pewresearch.org/fact-tank/
2017/04/27/5-facts-about-illegal-immigration-in-theu-s. Accessed March 1, 2017.
29. US Department of Justice, Federal Bureau of Investigation. Crime in the United States. Available at:
https://ucr.fbi.gov/crime-in-the-u.s. Accessed June 4,
2016.
30. American Community Survey response rates.
Washington, DC: US Census Bureau; 2013.
31. Moffatt S, Wan WY, Weatherburn D. Are drug arrests
a valid measure of drug use? A time series analysis. Policing.
2012;35(3):458–467.
32. Van Hook J. Counting 11 million undocumented
immigrants is easier than you think. The Conversation.
2016. Available at: http://theconversation.com/
counting-11-million-undocumented-immigrants-iseasier-than-you-think-67921. Accessed November 30,
2016.
33. Warren R. Democratizing data about unauthorized
residents in the United States: estimates and public-use
data, 2010 to 2013. J Migr Hum Secur. 2014;2(4):305–328.
34. Markides KS, Ray LA, Stroup-Benham CA, Trevino
F. Acculturation and alcohol consumption in the
Mexican-American population of the Southwest United
States: ﬁndings from H-HANES, 1982–1984. Am J Public
Health. 1990;80(suppl):42–46.
35. Welte JW, Barnes GM. Alcohol and other drug use
among Hispanics in New York State. Alcohol Clin Exp
Res. 1995;19(4):1061–1066.

Light et al.

Peer Reviewed

Research and Practice

e7