Discussion paper No. 925 – “Down and out in Italian towns: measuring the impact of economic downturns on crime” looks at the effect of local economic conditions on crime in Italian towns in the wake of the 2007-2009 economic slump. De Blasio and Menon show that the economic downturn led to a significant increase in crimes that did not require specialized criminal skills or tool – for example, thefts – while more human and physical capital intensive crimes - such as robberies - actually declined.
Their empirical model is:
CRIMEi,t.r = a + β ECONACTi,t +Sn (dn Xi,t) +gi+ρrt + εi,t,r
where i indexes spatial units, t years, and r regions; ECONACT is a proxy for economic activity in the area and Xi,t indicates time-varying covariates at the level of geographic aggregation. They have a (2004-2009) panel with area (gi) and time-region (ρrt) fixed effects. The variables are taken in logs making the coefficient β the elasticity of crime to economic conditions.
The spatial units are Local Labor Markets (LLM) as defined by the Italian National Institute of Statistics which are aggregations of two or more neighboring municipalities based on daily commuting flows from place of residence to place of work as recorded in the Italian population census. In 2001, 686 of them were defined and they had an average population of 83,084.
The local economic activity measure is the sum of the total revenues (known as fatturato) of all private plants located in the LLM. The crime data is from a dataset on criminal facts known as the “Sistema di Indagine” (Investigation System or IS), which supplements the information collected in victim reports with that on criminal facts that the police department collects in its day-by-day investigation activity which the authors argue can be considered as less spoiled by the underreporting issue that can affect the empirical analysis of crime. IS criminal facts are available at the municipality-level for the years 2004-2009 and include 34 categories of crime. The authors focus on thefts, extortions and robberies (for economic-related crimes) and on murders, assaults and sexual crimes (for the non-economic related ones).
The results? A 1 percent decrease in local economic activity resulted in a 0.6 percent increase in thefts and a 1 percent increase in extortions. However, the impact of the crisis on economic-related offenses that are more involved and require some crime-specific human and physical capital, such as robberies is negative. They also find that the slowdown did not have any effect on noneconomic-related crimes, such as murders, assaults or sexual crimes. Moreover, they found the intriguing result that in regions where organized crime was more pervasive – the impact of the slump on crime was more muted.
When they split the sample into LLMs belonging to regions where the control of organized crime was more prevalent - Campania, Sicilia, Calabria, Puglia – versus the remainder, they find that the regions where organized crime was pronounced found no effect of the economic slump on economic-related crimes. They offer two potential explanations for this:
1. Public sector employment is more important in the Italian south and therefore the weight of the public sector in organized crime regions is more significant than elsewhere: as a result, the public sector might act as an economic buffer that reduces the impact of the slump on criminal activity. However, the authors find that controlling for LLMs with the highest share of public sector employment did not change the results.
2. Organized crime acts as an illegal monopolist that raises entry barriers for criminal activity: in areas controlled by the Mafia, Camorra or Ndrangheta it is more difficult (and even dangerous) to turn to illegality when the reserve value of legal opportunities decreases.
I guess some of my suggestions regarding the paper's methodology would include laying out more explicit discussion of the differences between "theft" and "robbery". As well, I was unclear as to whether the private plants being referred to in the measure of local economic activity were all private firms or simply industrial firms. If the latter, since the Italian south has less industrial activity relative to the north, it could make this measure a poorer proxy for economic activity in the south.
The authors conclude that national and local authorities should add criminality to the long list of social problems that can arise and need to be tackled during economic downturns. Of course, this is a bigger challenge for government than one might think given that the results suggest organized criminals are better at fighting local crime than the government authorities. What I found intriguing is that the results suggest pervasive organized crime served as a barrier to opportunistic crime in the wake of an economic downturn – something one would think is what the public security apparatus of government should be doing.
One simple definition of government is a legal institution that has a monopoly on coercive force. Organized crime can be seen as an attempt at illegal entry into the activities of established government by establishing parallel private institutions that provide revenue generating “local public goods” with their own local monopoly on coercive force. That organized crime can be more effective at combating the criminal upsurge from economic downturns than government is really for me a very interesting result.