Essays on the Economics of Crime: Determinants of Crime in an Urban Context

Abstract

In this thesis, I draw on these theoretical models from the criminological literature and on the empirical tools from the field of economics to analyze crime from the perspective of economics.

I use the city of Barcelona, known worldwide for its ability to attract tourists, as my area of study. To the best of my knowledge, no similar crime research has previously been undertaken in this city and, therefore, the results of this thesis should provide interesting outcomes for policy makers..

In Chapter 2, entitled "Should football teams be taxed? Determining crime externalities from football matches", I begin by demonstrating the economic importance of Football Club Barcelona. I then analyze the effect of Football Club Barcelona matches on crime from a spatial perspective. That is, I first evaluate the effect of the number of spectators on crime (thefts and assaults) by comparing crime rates on match days, both home and away, and days that are very similar in all other characteristics apart from the fact that no match has been played. I analyze two types of crime given that their determinants are likely to be very different: first, I focus on thefts, where the concentration of people in space and time may introduce incentives, or reduce costs, to potential offenders; and, second, I focus on assaults (fights and brawls in the main), whose drivers would appear to be more closely related to hooliganism, i.e., unlawful behavior related to football matches, such as fights, drunken disorder or damage to the belongings/property of others. In the next step, I analyze the impact of a football match on the spatial distribution of crime, by carrying out an Exploratory Spatial Data Analysis (ESDA) of the census tracts around the stadium on football match days and on days free of football. Employing econometric models that account for the positive skewed distribution and the over dispersion of the data (e.g. negative binomial regressions), I analyze how the scheduling of a football match can modify the distribution of crime in the city of Barcelona. In the case of thefts, the results indicate an increase in the number of crimes for the whole city of Barcelona on home match days, especially, in those census tracts that are within a 1-km radius of the stadium. This suggests that despite the increase in the number of police officers deployed around the stadium, pick pockets are attracted to crowds where the rewards are likely to be higher and the probability of being detained lower. These results are confirmed by the placebo test that shows a lower number of crimes are recorded in the census tracts around the stadium when Football Club Barcelona plays away. In the case of assaults, a similar spatial pattern to that described for thefts is found, although the overall impact for the whole city is not significant. This result suggests that there is a marked displacement effect towards the census tracts around the football stadium from other parts of the city. This phenomenon would seem to reflect the hooliganism that is present in and around most football stadiums in Spain. As such, the results obtained in this chapter provide public administrations with the opportunity to raise their revenue levels by taxing the crime externalities generated by football teams.

In Chapter 3, entitled "How time shapes crime: the displacement effects of football matches in the city of Barcelona", I purposefully omit the spatial dimension of crime in order to focus on the temporal. I begin by undertaking an analysis of the temporal profile of crime for the city of Barcelona. That is, I carry out a very specific analysis of different types of crime and their unique temporal patterns. I specifically observe temporal crime patterns on a daily, weekly and monthly basis to determine whether they actually follow a clear pattern over time. I then conduct an hourly analysis to examine the impact of a major event, in this case a football match, on different types of crime. The results show that the football matches played by Football Club Barcelona reduce the levels of certain types of crime during the period of the game due to the incapacitation effect, i.e., potential criminals are incapacitated by the fact of their watching the game. Moreover, I find a reduction in those crimes typically reported only by the police, including driving crimes and drug consumption, due to what I identify as a substitution effect, i.e., police officers substitute their duty of reporting crimes for that of safeguarding citizen security. The consequence is an apparent fall in these types of crime, although the reality may be quite different. Finally, the results also show that in the hours leading up to a game and in the hours following the final whistle there appears to be significant increases in certain types of crime such as pick pocketing and violent crimes.

Chapters 2 and 3 give more evidence on the determinants of crime and specifically on crime behavior modifications (as regards both time and place displacements) attributable to major events such as football matches. The results obtained provide police agencies with a better understanding of the way in which criminals shift their decisions regarding the commission of crimes in relation to major events: on the one hand, when criminals decide to commit their crimes on days when a football match is being played; and, on the other, where criminals choose to commit these crimes on match days. Placing these results into the context of public economics, football matches generate negative externalities in the form of higher levels of crime, which serves as a justification for policy makers to find new ways of financing the public sector and compensating the costs of these events. Moreover, and given that tackling crime is one of the main goals of police agencies, Chapters 2 and 3 also shed light on criminal behavior. Specifically, the results reported in these chapters help understand the way in which criminals modify their target preferences (both in terms of time and space) and, therefore, how the police might best deploy police officers in time and space. However, since the early eighties there has been an increasing interest among police officers to ensure that that people not only feel safer, but that they have an enhanced perception of their own safety (Cordner, 2010). Thus, the reduction of the fear of crime has been a major objective of the police given its impact on individuals’ well-being. For instance, a robbery not only has an impact on the victim itself, it also affects the individuals that witness the act (and those who subsequently know about it) since they are likely to feel unsafe and to modify their behaviour accordingly. Networking and social interactions can spread this insecurity among individuals consequently affecting individuals’ well-being. If the authorities are incapable of ensuring that people perceive their personal integrity as being guaranteed and that they live in a safe environment, public efforts and resources devoted to crime prevention and control may well not be assessed as fulfilling their primary objectives.

In this sense, assessing the determinants of fear of crime and of crime risk perception are crucial for the public sector as they tackle the potential negative effects on individuals’ well-being. Community policing is the policy that has been used since the early eighties in an attempt at assessing crime risk perception determinants as well as the potential factors that may reduce crime. It consists of a more human approach to society as the police seek to get closer to the real problems and fears of the citizens. The level of crime risk perception is not only determined by crime or individual characteristics such as gender or age, but also by the characteristics of the neighborhoods. The broken window thesis (Wilson and Kelling, 1982) links three important concepts in neighborhoods: disorder, fear and crime. Specifically, this thesis states that the link between the three concepts may start with a minor disorder such as a broken window. If left unchecked, it will generate the perception that no one cares about it. Hence, this minor disorder may generate increasing levels of fear. Levels of distrust among the neighbours consequently rise and they start to behave differently - staying at home more and socializing less with each other. In turn, this leads to a reduction in natural surveillance permitting further disorder and minor crimes.

In sum, crime risk perception is known to be an important determinant of individuals’ well-being. Therefore, it is crucial, especially for governments, to understand its determinants and those (public) policies that can reduce it. Among those policies, public resources devoted to police forces emerge as a key instrument not only to tackle criminal activity but also to impact on citizenship crime risk perception. In this framework, the aim of Chapter 4, entitled "When police patrols matter. The effect of police proximity on citizens' crime risk perception", is precisely to analyze the determinants (both individual and neighborhood) of citizens’ crime risk perception for the city of Barcelona focusing on the effect of police proximity and taking into account spatial aspects of neighborhood characteristics. We measure, according to the main theoretical theories, how the simplest contact of a police officer with citizens’ can affect their level of crime risk perception. In this sense, we analyze how police policies consisting of an approach to citizens, may be effective in terms of reducing peoples’ crime risk perception. After controlling for possible problems of endogeneity of police forces and crime risk perception and the potential sorting of individuals across neighborhoods, our results show that crime risk perception is reduced when non-victims (randomly) interact with police forces. Moreover, neighborhood variables, such as proxies for social capital and for the level of incivilities, as well as individual characteristics have an impact on individuals’ crime risk perception.

La tesis titulada 'Essays on the economics of crime: Determinants of crime in an urban context' tiene el objetivo de analizar los principales determinantes de la delincuencia en el contexto urbano. En concreto, analiza en profundidad los efectos de los partidos de fútbol jugados por el Fútbol Club Barcelona en la delincuencia con un objetivo tanto de análisis espacial como temporal. Adicionalmente, se analiza otra dimensión de la delincuencia: la percepción del delito. Para tales objetivos, se divide la tesis en 3 principales capítulos que reflejan 3 artículos que pueden ser leídos de forma separada o conjunta.

En el primer capítulo, se analiza el efecto que los partidos de fútbol tiene sobre la distribución espacial de la delincuencia en la ciudad de Barcelona. Para tal objetivo, se utiliza una base de datos que por una parte, contiene todos los partidos de fútbol jugados por el F.C.B (tanto los jugados en casa como los jugados en otros campos de fútbol). Adicionalmente, y como gran novedad, se utiliza una base de datos única en España que contiene todos los delitos registrados por los Mossos de Esquadra i la Guardia Urbana desde 2007 a 2011 con información específica de donde han ocurrido, cuando han ocurrido y de que tipología delictiva se trata. Esta base de datos, permite realizar un análisis primero para el conjunto de toda la ciudad de Barcelona, y luego para aquellas secciones censales cercanas al estadio del Campo del F.C.B. La metodología utilizada consiste en un análisis exploratorio de datos espaciales mediante funciones de densidad de Kernel y la comparación de estos valores cuando existen partidos de fútbol jugados en casa y fuera de casa. De forma adicional, se añade un análisis de regresión que consiste en una primera estimación con datos de serie temporal a nivel de día para analizar los efectos de los partidos de fútbol sobre los hurtos y los delitos violentos en toda la ciudad de Barcelona. A continuación, también se realiza una estimación espacial que consiste en la estimación del efectos de los partidos de fútbol sobre los delitos violentos y sobre los hurtos a nivel de sección censal y día.

En el segundo capítulo de la tesis, se realiza un análisis de los efectos de los partidos de fútbol sobre el desplazamiento de hasta 8 tipologías delictivas en la ciudad de Barcelona. El análisis se realiza mediante los datos descritos anteriormente. La metodología utilizada consiste tanto en un análisis descriptivo de todas las tipologías delictivas al nivel de hora, día, semana y mes, como en un análisis de regresión a nivel hora de los efectos de los partidos de fútbol sobre el desplazamiento de los delitos en el tiempo. Los resultados reflejan que ciertas tipologías delictivas parecen aumentar después del partido de fútbol mientras que otras, como el consumo de substancias estupefacientes y los delitos contra la seguridad vial, parecen disminuir.

Finalmente, el último capítulo analiza, mediante datos de la encuesta de victimización de Barcelona, que determina la percepción del delito de los individuos. Analizo desde los determinantes individuales como sexo, edad, nacionalidad o renta, hasta los determinantes de barrio. Es decir, variables relacionadas con el entorno de orden y del aspecto del barrio. Finalmente, en este mismo capítulo se analiza como el simple contacto con la policía puede incrementar o disminuir el nivel de percepción del delito. La metodología utilizada es un análisis multinivel logístico que tiene en cuenta tanto el orden de la variable dependiente como las variables independientes a dos niveles.