Good Policy Can Lower Violent Crime: Evidence from a Cross-National Panel of Homicide Rates, 1980–97
| Published date | 01 November 2003 |
| Author | Eric Neumayer |
| Date | 01 November 2003 |
| DOI | 10.1177/00223433030406001 |
| Subject Matter | Journal Article |
619
Introduction
The existing empirical literature explaining
variation in homicide rates suffers from two
important shortcomings, one conceptual,
the other methodological.1Conceptually,
there is too little focus on how good policy
can lower rates of violent crime, of which
homicide is the most extreme and by far the
most reliably reported form. The existing,
mostly sociological, literature sees homicide
rates mainly determined by structural deter-
minants such as modernization, population
characteristics, and country-specific cultural
factors. Since these structural determinants
are difficult, if not impossible, to change via
policy, one might draw the misplaced con-
clusion that homicide rates cannot be
lowered via good policies either. Instead, our
study focuses on the effect of good policy on
homicide rates.
The methodological shortcoming that
characterizes many empirical studies of vari-
ation in homicide is that often only multi-
variate ordinary least squares (OLS) analysis
© 2003 Journal of Peace Research,
vol. 40, no. 6, 2003, pp. 619–640
Sage Publications (London, Thousand Oaks, CA
and New Delhi) www.sagepublications.com
[0022-3433(200311)40:6; 619–640; 038284]
Good Policy Can Lower Violent Crime: Evidence
from a Cross-National Panel of Homicide Rates,
1980–97*
ERIC NEUMAYER
Department of Geography and Environment, London School of Economics
and Political Science
This article provides empirical evidence that good political governance and good economic policies can
lower homicide rates. Therefore, violent crime is not simply determined by modernization, population
characteristics, and cultural factors. This result follows from rigorous econometric testing based on a
cross-national panel of homicide data from up to 117 countries over the period 1980–97. Contrary to
most existing studies, which have applied ordinary least squares on data drawn from one time period
only, this analysis uses a fixed-effects estimator with fully robust standard errors. A fixed-effects esti-
mator elegantly controls for time-invariant determinants, such as cultural factors, and allows the
pooling of homicide data from otherwise incompatible sources. This is complemented by random-
effects estimation in sensitivity analysis. The results suggest that economic growth, higher income levels,
respect for human rights, and the abolition of the death penalty are all associated with lower homicide
rates. The same is true for democracy at high levels of democracy. The transition from autocracy to
democracy is likely to be accompanied by a rising homicide rate, however, until full democracy has
been reached. Results also indicate that policies aimed at improving equity have no effect on violent
crime. In particular, there is evidence that the positive effect of income inequality on homicide rates
found in many studies might be spurious. The results reported here are strikingly similar to those found
for the causes of civil war.
* E-mail address: e.neumayer@lse.ac.uk. Helpful comments
from three anonymous referees and the associate editor,
Han Dorussen, are gratefully acknowledged. The data used
in this article as well as the routines used to generate the
results with Stata 7 can be found at http://www.
prio.no/jpr/datasets.asp.
1For an overview of this literature, see Neapolitan (1997a)
and LaFree (1999).
68S 01neumayer (ds) 3/10/03 1:20 pm Page 619
on one-period data is used, which fails to
exploit fully the potential of quantitative
analysis. Few studies employ a fixed-effects
estimator on panel data in order to control
for unobserved country heterogeneity bias in
the estimated coefficients. Such bias can arise
if country characteristics are either imposs-
ible to quantify or unobservable and if the
explanatory variables are correlated with
these characteristics. Given that underlying
time-invariant country characteristics that
are difficult or impossible to measure are
likely to be correlated with variables used for
explaining variation in homicide rates, the
use of a fixed-effects panel estimator gains
great importance. In addition, usage of a
fixed-effects estimator enables us to pool
homicide data from two different sources
that should not be pooled in any analysis that
exploits between-country variation in the
dependent variable, as OLS does.
We assess the impact of good policy on
homicide in terms of political governance
(democracy, respect for human rights, and
absence of the death penalty), economy
(economic growth and income levels), as well
as equity (welfare expenditures, economic
discrimination of minorities, and income
inequality). We find that good policies in
terms of political governance and the
economy can lower homicide rates, but not
good policies in terms of equity. Therefore,
this article provides evidence on how policy-
makers can achieve a more pacified society.
There are striking similarities between our
results and the ones found in recent expla-
nations of the incidence of civil war. For
example, Blomberg & Hess (2002) find that
recessions raise the probability of civil war
much as they raise the homicide rate in our
analysis. Hegre et al. (2001) find that harsh
autocracies and coherent democracies have
few civil wars, while our results suggest that
they have fewer homicides than intermediate
regimes. More generally, Collier & Hoeffler’s
(2002) finding, that civil wars can be
explained better by economic characteristics
such as low average income levels rather than
grievances such as income inequality, mirrors
our result that low income levels are associ-
ated with high homicide rates whereas
income inequality, economic discrimination
of ethnic minorities, and social welfare
expenditures are not related to the extent of
homicide. This suggests that some of the
same processes that drive violent crime are
also responsible for the outbreak of violence
in the form of civil war. Also, in many coun-
tries, people feel threatened by violent crime
as much as by internal or external threats of
military violence. This supports the call for a
broad and encompassing conceptual frame-
work for the analysis of security (Buzan,
Wæver & de Wilde, 1998).
Explaining Variation in Violent
Crime: Traditional Theories
The traditional, mostly sociological,
approach to explaining variation in violent
crime is based upon modernization, oppor-
tunity, and cultural theory.2As we will see, it
leaves little space for good policy to lower
rates of violent crime.
Modernization Theory
Modernization theory based on Durkheim
(1964) is the most famous sociological theory
explaining variation in homicide rates. The
social and economic transformation a
country goes through in the process of
development and modernization is seen as
having important consequences for violent
crime rates. Traditional forms of status
relations, role allocations, social organization,
and control are disrupted and ultimately
destroyed as masses of people leave their rural
homes and flood into the anonymous urban
journal of PEACE RESEARCH volume 40 / number 6 / november 2003
620
2Due to the applied focus of this article, the various
theories can merely be sketched here. More detailed dis-
cussions can be found in Neuman & Berger (1988) and
Neapolitan (1997a).
68S 01neumayer (ds) 3/10/03 1:20 pm Page 620
To continue reading
Request your trial