Crime against trading: The case of cargo theft in São Paulo state, Brazil

Crime against trading: The case of cargo theft in São Paulo state, Brazil Marcelo Justus¹*, Vania Ceccato²; Tulio Kahn³, Gustavo C. Moreira⁴

Abstract

The aim of this chapter is to provide a general picture of cargo theft in Brazil, focusing on the nature of the crime and its temporal-spatial distribution in São Paulo, one of the most economically dynamic states of the country. The study makes use of multiple datasets from 2001 to 2016. Findings show signs that cargo theft has indirect connections with retail as most stolen goods are foods, drinks, electronics, tobacco and pharmaceuticals. Most incidents occur during weekdays when most retail and service establishments are operating. Although the São Paulo state capital shows the highest levels of cargo theft, it is in non-metropolitan areas that records of this offence are on the rise. Intra-state patterns show a convergence tendency of cargo theft levels over time between São Paulo’s metropolitan area and non-metropolitan municipalities. The chapter finalizes suggesting a research agenda to improve the understanding of the phenomenon and its prevention.

Keywords: retail crime, robbery, space-time trends, cargo insurance.

1 Introduction

Cargo theft is a major problem throughout the world. Some countries have shown a continuously increase, among them, Brazil, which is one of the countries that presents the highest risk to cargo security in the world. Cargo theft imposes an estimated direct

* Corresponding author.

Email adresses: mjustus@unicamp.br (Marcelo Justus), vania.ceccato@abe.kth.se (Vania Ceccato), tulio.kahn@hotmail.com (Tulio Kahn), gustavocmoreira@usp.br (Gustavo C. Moreira)

1 Center for Social, Urban and Regional Economics – Institute of Economics, University of Campinas, São Paulo, Brazil.

2 School of Architecture and the Built Environment, Royal Institute of Technology – Stockholm, Sweden.

3 Researcher and consultant for public safety issues.

cost to the Brazilian economy of 442 million USD per year and this amount has been increasing in the last decade (FreightWatch, 2014).

Although cargo theft is defined according to the Brazilian penal code as “subtraction of goods for resale when they are being transported” this offence, contrary to many other countries (see e.g. Burges, 2013), is more than a crime against property. In Brazil, cargo theft is highly violent as many of these property crimes result in deaths, almost always with firearms. There are indications that this crime is not distributed homogenously across the country, being concentrated in the most economically developed regions, especially in the state of São Paulo, where half of all cases currently takes place (National Public Security System Information, SINESP, 2016). Despite the gravity of the problem, little is reported about this phenomenon in the international literature.

In this chapter, we make use of multiple datasets to investigate the overall nature of cargo theft in the country, and focus the analysis on the specific case of São Paulo state – which is one of the most cargo theft targeted states of the country – for which there are reliable and recent data. This aim is achieved by reporting the nature of cargo theft and assesses its temporal and spatial distribution. More specifically, we detect areas where cargo theft increased the most in São Paulo state.

This chapter is organized as follow. Section 2 presents a brief literature overview regarding cargo theft. Issues of data availability and quality as well as methods are presented in Section 3. In section 4, the nature and space-time trends of reported cargo

theft and its potential causes are then discussed. In Section 5, the chapter ends proposing a research agenda and concluding remarks.

2 Literature overview

This study draws from three complementary theoretical perspectives: rational choice theory (Becker, 1968), routine activity theory (Cohen and Felson, 1979) and situational crime prevention (Clarke, 1983). Their principles are discussed in the sections below.

2.1 Definition of cargo theft

In this study, we adopt the Brazilian definition of cargo theft, which is “subtraction of goods for resale when they are being transported” (Law number 8.072 of 25th July of 1990) even though this crime may also associated with a number of other offences Cargo theft can involve many types of criminal activities, such as theft, robbery, burglary, fraud and may also lead to violence. In the USA, Burges (2013) indicates that cargo thefts are often non-confrontational and non-violent, involving mostly the theft of the cargo. The theft of cargo includes, according to FBI’s reporting system as goods, “chattels, money, or baggage that constitutes, in whole or in part, a commercial shipment of freight moving in commerce, from any pipeline system, railroad car, motor truck, or other vehicle, or from any storage facility or wharf, or from any aircraft … any freight consolidation facility” (Coughlin, 2013:8).

2.2 The nature of cargo theft

The targets are usually products considered as CRAVED goods (Clarke, 1999). Some of the key attributes of hot products are their value, size and portability. These attributes are summarized by CRAVED, an acronym referring to hot products that are concealable, removable, available, valuable, enjoyable and disposable. They can easily be sold for a high price on the black market, guaranteeing an attractive return. In Brazil from 2006 to 2009, these include electronic equipment, famous brands of clothes and shoes, perfumes, jewels, cigarettes and pharmaceutical products. Cargo thieves from all over the world seem to be unanimous in their interest for these items (Burges, 2013). Besides of high return, Burges (2013) reports that less than 4% of cargo stolen is recovered in the United States. According to Moreira and Carvalho (2011), this recover rate is around 10% to 20% in Brazil. It is noted that cargo theft is an attractive criminal activity since it is combines a high return and low risk of failure. In these circumstances, the economic theory proposed by Becker (1968) can help in providing some interpretation. The rational choice theory postulates that the potential offender evaluates his or her own risk

before making a decision to commit a crime. According to the economic theory of crime, this alternative essentially depends on two factors: monetary return from crime and the probability of failure. Assuming mobility of criminals and economic rationality, crimes occurs in localities where there is higher expected utility. It is expected that the two determinant factors of the decision to commit crime from an economic angle can

significantly differ among places and help explain its geography. This may explain why the modus operandi of cargo theft varies significantly from one country to another (Ekwall and Lantz, 2015; Burges, 2013). Ekwall and Lantz (2013) showed that in Europe, the Middle East and Africa, these crimes are often opportunist in nature, occurring when vehicles are parked in unsafe regions and with no contact with the driver.

Burges (2013) suggests several examples of how this crime is carried out in the world. In the United States, cargo theft is almost never associated with violence. In Europe, however, violence is more common, in the form of intrusion, pilferage, and hijackings.

In Mexico, this is an extremely violent crime, which is carried out by gangs. Brazil, Guatemala and Venezuela are the countries where the operating methods are most dangerous, with the use of heavy arms and corruption; it can also be linked to international organized crime (Oliveira and Martins, 2014).

2.3 Temporal dimension of cargo theft

Cargo theft only happens when the necessary conditions are present at a particular place, at a particular time: unattended loaded trucks parked in an isolated area are detected by motivated thieves. According to routine activity theory, most crimes – including cargo theft – depend on the interrelation of space and time of offenders’ motivation, suitable targets and absence of responsible guardians (Cohen and Felson, 1979). People’s activities and daily habits are rhythmic and consist of patterns that are constantly

repeated. Variations of crimes during weekdays, weekends, and seasonally reflect these changes in people’s routine activity. Thus, knowing when cargo theft occurs is an important information to all involved in retail, specially operators of supply chain companies, since it indicates when more resources has to be added to prevent that their goods are left unattended.

Situational Crime Prevention are crime prevention strategies that are used to reduce the

criminal opportunities (Clarke, 1983). Such strategies include ‘hardening’ of potential targets, improving surveillance of areas that might attract cargo theft, and deflecting potential offenders from settings in which crimes might occur (e.g., by limiting access of such persons to storages, harbors or where cargo might be placed. Yet, locking down facilities is not the biggest challenge, Coughlin (2013) suggests that the biggest challenge appear during the release of the freight for transit, while in its transit, and during receipt at arrival. As far as the time of day that these events occur, Rick (1995) found that in England, Scotland and Wales, cargo theft occurs mainly between midnight and 8:00 am, when drivers are elsewhere and their trucks are parked in company warehouses or places that are not very safe. This, however, seems to be specific to the region, as using data for Europe, Middle East, and Africa, Ekwall and Lantz (2013) showed that the majority of cases occurred during business hours. In Sweden, theft of trucks was concentrated in afternoon hours, especially between 16:00 and 17:00 while theft of cargo of trucks in the night hours, around 00:00-01:00 (Ceccato, 2015).

There is also a weekly variation to when these events occur. Here, the literature converges to the fact that these crimes are concentrated on weekdays, Tuesdays, Wednesdays and Thursdays (Rick, 1995; Ekwall, and Lantz, 2013; Burges, 2013). The low level found at the weekend is related to the fact that few vehicles operate during this period. In Sweden, Ceccato (2015) reports that a fourth of all events happen on Fridays and that overall, cargo theft from trucks was concentrated in weekdays. However, in the United States, researchers found a concentration of cargo theft on the weekends but also in 3-day holiday weekends (Burges, 2013; Coughlin, 2013). With regards to seasonal variations, Ekwall and Lantz (2013) found that cargo theft were concentrated in the winter in 82% of the regions in Europe, Middle East and Africa. Another observed seasonal effect is the steep fall in cargo theft after Christmas. This is because sales tend to fall significantly after this date, meaning there is a lower demand for goods. In Sweden, for instance, it was observed differences between summer and autumn; summer had 23 thefts from trucks a day while in the autumn, an average against 32 (Ceccato, 2015).

2.4 Spatial dimension of cargo theft

Cargo theft has long been thought of as a crime that knows no geographic boundaries In the United States, Coughlin (2013:13) showed evidence that cargo “theft is predominately concentrated in populated areas (big cities and metropolitan areas) within

those states as a result of the routing of state freeway networks through major hubs”.

California has by far reported the most cargo theft activity in North America followed by Texas, New Jersey, Georgia between 2009 and 2011. In other words, cargo theft takes place essentially in more dynamic economic regions, where many industries and logistics operators are found (Ekwall, 2009; Burges 2013). It is in these places where most valuables and interesting targets are, not less cargo. In Sweden, Ceccato (2015) found clear differences in the spatial pattern of cargo theft, especially along the main road networks/hubs: 61% of cargo theft in trucks happened in larger urban areas (Stockholm, Gothenburg and Malmö), 23% in accessible rural areas and 16% in remote rural areas.

Municipalities with “resting areas” for trucks are more targeted by cargo theft. This location has also been common in the United States for cargo thefts. For example, Coughlin (2013) found that truck stops, carrier facilities and parking lots concentrated most cargo theft locations in that country.

Historically, criminals operate mainly in their own areas. Nevertheless, there is some evidence of migration into regions where this activity offers a greater probability of success. This has also been observed by Ekwall (2009) in Sweden. A number of methods exist for moving cargo that it has been stolen, from e-fencing to exporting to reproducing it into the legitimate supply chain (Burges, 2013).On analyzing the spatial distribution of this crime, Hesseling (1994) argues that “... increased pressure from law enforcement agencies and increased security-hardening measures by the supply chain industry have

caused gangs to operate across larger geographical regions”. In other words, there is a displacement effect when there is increased law enforcement and other crime preventive measures are in place – a factor that might help explain changes in the geography of cargo theft over time.

3 Framing the case study

3.1 The study area

Figure 1 shows the study area in Brazil: the state of São Paulo. It is worthwhile to emphasize that São Paulo is the namesake of a state composed of 645 municipalities, a metropolitan region, the Greater São Paulo (composed of 39 municipalities including the state capital) and the São Paulo municipality, which is the state capital. According to Statistics Brazil (IBGE, 2016), São Paulo municipality, which is the state capital, has 11.2 million inhabitants and accounts for about 12% of Brazil's GDP and 36% of the state production of goods and services. São Paulo metropolitan region is one of the ten largest metropolitan regions in the world (United Nations, 2014).

Figure 1 – About here

The state of São Paulo is the largest economic and industrial hub in the Southern hemisphere, the largest business center in Latin America and also the capital of

innovation and technology in the region, generating alone about a third of all the wealth produced by Brazilian economy (IBGE, 2016). Historically, São Paulo is the state with the highest GDP in Brazil. The state has a wide range of activities, varying from the most traditional to the modern and sophisticated technology activities. The state of São Paulo has 42 millions of citizens and 35% of the industry and 34% of the services of the Brazilian market (IBGE, 2016). According to the Federation of São Paulo’s State of Commerce (2016), the retail generated 2.5 million jobs and a revenue in 2015 of 512.8 billion Reais ($ 155 billon), accounting for 30% of total sales in the country. A third of this revenue comes from food sector followed by car dealership with 12%, and then pharmaceuticals 8% and 7% are electronics.

A large part of the cargo in Brazil is transported using roads. The roads with higher quality in the country are also located in the state, which has 34,000 kilometers of paving, as well as the largest and most modern Brazilian port (Harbor of Santos), located 80 km from the capital. In the state there is also the largest Brazilian cargo terminal, 90 km from the capital, in Campinas. About 40% of all cargo imported in the country comes through this terminal (São Paulo Global, 2016).

Cargo theft is not an isolated security problem in São Paulo state, where almost 500,000 crime occurrences were registered in the second quarter of 2016 by the SSP-SP - Public Security Secretariat of São Paulo (SSP-SP-2016). Of these, 62% were crimes against property and 26% were crimes against people. As far as the crimes against

property are concerned, 88% referred to theft or robbery, giving a total of over 250,000 occurrences. The most commonly used method is theft, where there is no contact with the victim, which accounts for 61% of the cases. During this period, there were 2,246 occurrences of cargo theft in the state, representing around 0.74% of property crimes and 2.3% of thefts. Despite relative small numbers, cargo theft accounted for 295,85 millions of Reais ($ 176 million), in 2011 and in the country overall, this offence is estimated to cost around 1 billion of Reais ($ 442 million).

3.2 Data and methods

As anywhere in the world, long term and reliable statistics of cargo theft are problematic.

In Brazil, cargo theft statistics are rare and has started to be more systematically collected across the country since 2000s. Total number of cargo thefts is available yearly for the whole country between 2008 and 2015 from Fórum Brasileiro de Segurança Pública – FBSP and SINESP. However, the data quality is limited since the information contained in this database lacks systematic recording both temporally and spatially.

Better data quality is found within state level. São Paulo state for instance stands out by having one of the most reliable and complete databases on cargo theft in the country (see SSP-SP, 2016). The most reliable time period of this data is between 2006 and 2011, as the dataset was created as a result a collaborative agreement established between the São Paulo’s Federation of the Cargo Transportation - FETCESP, and Public Security

Secretariat in São Paulo State – SSP. The database contains information about the type of product stolen, the value of the stolen cargo, where it happened (name of motorways), the time the cargo theft occurred (hour, day and year).

At municipal level, the data is only available in SSP-SP database during the period from January 2006 to June 2016. As complementary source, information from a technical report elaborated by Coordenadoria de Análise e Planejamento da Secretaria de Segurança Pública de São Paulo – CAP/SSP (2009) was used as reference. Data contained twenty-seven thousand reported cargo thefts from 2006 to 2009, with more detailed information such as the type of crime, o modus operandi, the amount of cargo, type of stolen goods.

It is noteworthy to emphasize that because they refer to statistics derived from police records, official data only reveal crimes actually reported to the competent authorities and duly recorded, meaning that they underestimate the actual crime level.

Underreporting of crimes can be more or less pronounced according to the features of the locations where they take place, and it is usually more common for crimes such as theft and extortion. Cargo theft is a good crime indicator due to the relatively high value of the stolen goods, possible vehicle robbery aside the stolen cargo, and their insurance coverage (cargo and/or vehicle), most of such occurrences result in a police report. Even so, we have to assume that the underreporting of cargo theft is time and cross-section invariant. Another relevant issue is that cargo theft should be ideally measured in relation

to the amount of cargo transportation vehicles (truck, pickup trucks etc). However, there is no available data to make such standardization. Thus, as the measurement of cargo theft rate per population is not appropriate, we analyzed the absolute number of occurrences reported to law enforcement agencies.

We performed a descriptive analysis of cross-section data and time series of cargo theft. The goal of the analysis has been to identify time-space patterns in cargo theft, and then by comparing the Brazilian case with those found in the international literature, suggest potential mechanisms behind cargo theft. Based on these diagnostics, we suggested directions for future studies. The lack of detailed and systematic data of cargo theft has limited a more quantitative approach to the analysis, which was originally intended in this chapter. In order to support the temporal analysis, we used data from FETCESP de 2006 a 2011 e da SSP-SP de 2012 a 2015 to build a monthly time series of cargo theft in São Paulo state. Moreover, the sample composed of 46 quarterly observations for state of São Paulo between the 2005 first quarter and the 2016 second quarter from SSP-SP was also used. It is worthwhile to emphasize that for our proposals we considered the state divided in three regions: The Greater São Paulo excluding São Paulo city (henceforth GSP), São Paulo municipality ('city'), and non-metropolitan municipalities of the state. With the SSP-SP data, we calculated an index based on the first quarter of 2005 (2005Q1). The aim was to assess the evolution of this phenomenon from the beginning of our series. The index for the period is defined as:

(1)

In the following analysis, we made a logarithmic transformation of cargo theft data.

According to Morettin and Toloi (2004), this transformation is useful to remove tendency and stabilize variances over time. Figures presented in this chapter were created using statistical packages (the Stata, R and Gretl, see StataCorp, 2013) and desktop Geographic Information Systems (GIS) (Mapinfo 11.0). A georelational database was created linking data to each municipality, which facilitated the spatial analysis of cargo theft across the state. In that particular analysis, we used data from SINESP, CAP-SSP (2009) and from SSP-SP.

4 Results and Discussion

4.1 The nature of cargo theft: Brazil and focus in São Paulo state

In this section, we use data from SINESP, FETCESP (2006-2011) and the technical report elaborated by CAP/SSP (2009) as a basis for the analysis. According to SINESP, in 2014, there were 17,500 incidents recorded in Brazil and half occurred in the São Paulo state. Furthermore, the total cargo theft in Brazil is estimated to be around 1 bilhão de reais ($ 442 million), and one third of this is there.

In Brazil, the most targeted product types for cargo theft are foods, drinks, electronics, tobacco and pharmaceuticals. These goods are easily sold to consumers in legitimate stores or in the black market, sometimes beyond the national borders (Oliveira and Martins, 2014). This is indirect connection between cargo theft and retail. Since the cargo is a product of commercial transaction between two or more economic agents, then the cargo theft can be approached as a type of crime against trading and service activities. Ekwall (2009) and Burges (2013) indicate that there is a unanimous preference for so-called “hot products” which are high-technology goods with a high black market value. These are, however, certain targeted products that are particular to Brazil, when compared to the rest of the world. Moreover, regional differences create different demands for different products and the needs of each market have a bearing on the process. For example, Rick (1995) found that the majority of stolen products in Sweden in 1994 were materials used in civil construction, brought about by the real estate boom in the country. These demands also vary over time. In Sweden, high-technology goods but also fuel from trucks are common in Sweden than in the past (Ceccato, 2015).

Cargo theft (about 70%) in Brazil results in ‘express kidnapping’ of the driver and other people such as driver’s assistants, which means that they are kept captive as the crime takes place. It is also observed that the modus operandi of criminals depends on where the theft is committed. Compared to the cargo thefts that occurred in the city, crimes on highways are mostly characterized by frequent use of big guns, vehicle

robbery aside the stolen cargo, and greater number of criminals and vehicles in the action. Arguably, these features are present in organized crimes, in other words, crimes which are executed by criminal organizations (CAP-SSP, 2009). The use of severe violence is a phenomenon that more often occurs in Brazil, Venezuela and Guatemala (Burges, 2013). In Europe, Burges (2013) reports that the use of violence is positively related to the cargo value. For this reason, England is the country with more aggressive approach to the whole continent. Spain, in turn, is the country with the most varied methods. In other countries, such as the United States and the United Kingdom, violence is seldom used (Rick, 1995). Once again, regional differences bring particular features to this crime’s modus operandi. Burges (2013) argues that these depend primarily on each country’s crime culture, as well as on law enforcement for violence and use of weapons.

In São Paulo state, almost three quarters of cargo theft are carried out by three or fewer people (CAP/SSP, 2009). For the period 2006-2009, almost 80% of cases were performed by kidnapping the driver while he was driving the vehicle. As in England, cited by Burges (2013), the use of violence and large weapons tend to be proportional to the cargo value.

In São Paulo, the most stolen goods are food cargo from 2006 to 2009 in urban areas.

Although not of the highest value, foods account for approximately 25% of total goods stolen during the period. Only the second most stolen goods are electronics – around 10.4%. The third highest cargo theft risk is for pharmaceutical products, which account

for about 7.6% throughout the four years. On highways, the most targeted by cargo thieves were metallurgical products, machines and equipment, stationery goods and vehicle parts. The fuel, metallurgical products and machines and equipment were the three most targeted on highways, especially in non-metropolitan areas. Regarding the values of the stolen cargos, average values were smaller and greater for crimes which occurred on urban areas and highways, respectively. In general, the value of thefts on highways is much higher than the values stolen within cities. Moreover, the values are greater in non-metropolitan cities than the values in the GSP (CAP-SSP, 2009).

Recent data from SINESP indicate that, only in 2015, at least 17,852 thefts cargo occurred in Brazil, where São Paulo was the top ranking for cargo theft in the country accounting for 47.5% of the total reported incidents. The state of Rio de Janeiro was the second in rank with 40.5% of the total crime that was reported, and Minas Gerais represented 3.4% of total cargo theft reported in this year. Therefore, 91.4% of incidents occurred in only three states located in southeast region – the richest Brazilian region.

São Paulo, Rio de Janeiro and Minas Gerais are the top three because they are places where the greatest cargo on Brazilian highways circulates. Consequently, demand for cargo insurance is higher in these states than the others Brazilian states (Gameiro and Caixeta-Filho, 1999). In the next section we discuss more in detail the geography of cargo theft as well as its temporal signature.

4.2 Temporal and spatial trends in cargo theft in São Paulo state

Most cargo thefts in urban areas are concentrated during the hours between 6:00 and 16:00, which corresponds to the operation time of most retail and services activities.

Almost three-quarters of crimes are concentrated during business hours. However, cargo theft on highways has distinct pattern. Two-thirds of highway crimes occurred from 20:00 to 8:00, especially between 22:00 and 6:00. The most risky hours were between 4:00 and 6:00, when the circulation of cargos is at its peak. Concerning the days of the week with higher risk of cargo theft, most incidents occur during weekdays on which most retail and service establishments are operating. Tuesday, Wednesdays and Thursday account, altogether, for about 60% of total crimes. On the contrary, Monday is the business which less crime occurs, around 2% of cargo thefts. This evidence is corroborated by the results reported by Rick (1995), Ekwall and Lantz (2013) and Burges (2013). These studies also observed that most the cargo theft occurs on Tuesday and Thursday. In Brazil, there is no drop in number of cargo theft in Saturday. This fact can be related to the work regime of drivers in Brazil with regard to the hours of work. In general, they work 4 hours on Saturday, when most retail and service establishments are operating too.

Figure 2 – about here

Cargo theft are often committed between March to May and between October and December (Figure 3). There is an average of 634 events per month, compared to an average of 586 events per month for the rest of the year. There are, therefore, troughs during the months of June to September, and January and February. This seasonality corroborates the results of Ekwall and Lantz (2013), who found the same temporal cargo theft dynamic in Europe, the Middle East and Africa. In addition, this Post-Christmas hypothesis has also been confirmed in Brazil. Cargo theft falls during the months after this date, reflecting the lower demand for goods.

Figure 3 – about here

- the overall spatial pattern of cargo theft in São Paulo state

Cargo theft in São Paulo is concentrated in urban centers, often in the most economically dynamic areas (Figure 4). Approximately 82% of the crimes from 2006 to 2009 occurred within cities, while about 18% occurred on highways around the cities of the state. Most targeted cross-country highways in São Paulo are the Dutra, Anhanguera and Régis Bittencourt. These highways altogether account for about half of cargo thefts within São Paulo. However, although less frequent, cargo thefts on highways are very distinct from the crimes that occur in urban areas where retail activities and population also are concentrated (CAP-SSP, 2009). Figure 4 shows the percentage of total cargo theft by

municipalities of São Paulo state from 2013 to 2015. The area painted in dark blue is the state’s capital accounting for 59% of the average number of cargo theft reported in the state.

Figure 4 – about here

The state’s capital and other 38 municipalities, which formed the GSP, altogether account for about 77.4%. Clearly, there is a cluster of cargo theft around the São Paulo municipality. Figure 4 also indicates signs of two possible underlying factors behind the spatial distribution of cargo theft in the state of São Paulo. First, the majority of crime occurs around the main highways and around the ports of the state. Second, there is a suggestive relationship between cargo theft, urbanization rate and high level of economic activity, especially around the São Paulo city. The incidence follows the route of main highways, which connect the São Paulo city to other major urban centers of the state, but also to other state capitals such as Rio de Janeiro (Brazilian Southeast) and the state of Paraná (Brazilian South). Tietê-Paraná Waterway (with 2,400 kilometers) is a link from important areas of the non-metropolitan areas of São Paulo state to other Brazilian states and to neighboring countries. It is connected to the state’s highway and railroad networks. Thus, there is more transport facilities, especially for cargo, making easier the national and international trade. Obviously, more cargo means more targeted to be stolen

by motivated criminals. A positive relationship between cargo theft and high economic dynamic of regions also was found elsewhere (e.g., Ekwall, 2009; Burges, 2013).

- Intra-state patterns of cargo theft

In this section, intra-state patterns of cargo theft are analyzed here using data from SSP-SP divided in three areas: The Greater São Paulo (GSP) - the whole metropolitan area (39 municipalities), São Paulo city (São Paulo municipality only) and non-metropolitan municipalities (606 municipalities). In the first three months of 2005 (2005Q1) to the last three months of 2016 (2016Q2), São Paulo municipality accounts for about 60.3% of total cargo theft incidents within the state, the GSP accounts for about 23.4%, and the 606 other municipalities altogether account for 16.3%. By building an index based on 2005Q1 (Figure 5a), we observed that reported cargo theft increased over time in São Paulo, specifically after 2008. However, the increase is higher in the 606 non-metropolitan municipalities.

Figure 5 – about here

More interestingly, as soon as the number of reported cargo theft was transformed in logarithm (Figure 5b), a convergence tendency of levels of cargo theft seems to have occurred after the last quarter of 2010 in the series of the GSP and non-metropolitan

municipalities. São Paulo city (the municipality alone) has, as it could be expected, a different and independent path of cargo theft increase and/or temporal variation from the rest of the municipalities of São Paulo state. Following Justus and Santos Filho (2011) based on the economic theory proposed by Becker (1968), it is possible to suggest an economic interpretation for this convergence tendency of thefts levels in the GSP and other non-metropolitan municipalities. According to Justus and Santos Filho (2011) assuming perfect mobility of factors in criminal activities among localities – GSP and other non-metropolitan municipalities – it is possible that a portion of the convergence can be result of displacement to commit crime. According to the economic theory of crime, this choice essentially depends on two factors: monetary return from crime and the probability of failure (e.g. being caught committing the crime). The net monetary

return is mainly determined by the expected return and planning, execution and opportunity cost of crime. The theory of rational choice predicts an unequivocal positive relationship between the magnitude and value of crimes and the expected return.

Concerning the cost of planning and committing crime, the theory predicts an inverse relationship with the quantity of committed crimes. As to the risk of failure, the theory suggests that this becomes higher when the efficiency of laws and police increase. Note that the perception of greater risk of failure is determined by the awareness of the risk of being caught, convicted and paying for committed crime (Justus and Scorzafave, 2014).

In sum, these two determinant factors - monetary return from crime and the probability of failure - can significantly differ among regions of the state analyzed in this study.

Figure 6 - about here

In order to better understand these variations, the non-metropolitan municipalities were assessed in more detail. Most cargo thefts are concentrated in areas characterized by grandes centros de distribuição ou comercialização de mercadorias no Estado e, consequentemente, com um número maior de veículos de carga circulando. Figure 6 shows the time path of the series in nine departments (or divisions) from 2006Q1 to 2016Q2. Most of the cargo theft occurs in the municipalities of Campinas, Santos and Piracicaba, which indicates two opposite axis, one for the interior of the state (Campinas and Piracicaba) and the other towards the most important harbor of the country (Harbor of Santos). In this period, the three regions altogether account for approximately 74.6%

of total reported cargo theft within the state (35.2%, 19.2% and 20.1%, respectively) and are concentrated where the major transportation infrastructure is located. For example, in Campinas are found the most modern highways in the state and also Brazilian territory.

The largest cargo airport in Latin America (Viracopos International Airport) is also localized in this region. Furthermore, a railroad network connected to the Santos harbor.

The region has an important technological center in Latin America and it is in this region

(especially in Campinas city) that the subsidiaries of some of the largest multinational companies in the world are located, and that producing and/or selling the goods most craved by cargo thieves. The Department of Santos boasts South America’s largest port named Santos Port, which historically handled about one quarter of Brazil’s trade flow.

São Sebastião harbor, which account for about 7.3% of total cargo theft from 2005Q1 to 2016Q2, is also an important chanel to national and international trade. It is noteworthy that these ports also serve the states of Minas Gerais, Mato Grosso, Mato Grosso do Sul, Goiás, Paraná and Rio de Janeiro. In sum, such regional infrastructure and supply of targets have created the necessary conditions for a rise in cargo theft in non-metropolitan areas.

This development has multiple causes but can certainly be associated with inherited high economic growth rates experienced by non-metropolitan municipalities during this period in comparison with GSP and São Paulo capital (IBGE, 2016), leading to a supply of potential targets. Yet, this is of course an empirical question worth to be tested. A complementary hypothesis is that that carriers which work in the metropolitan areas, where there is higher number of cargo theft started during this period to adopt more efficient preventive cargo theft strategies, such as “target-hardening”: tagging goods, chips geolocation hidden in goods, armed escorts, cargo fractionation. These precautions have decreased the attractiveness of targets for theft because the risk to be caught

becomes higher. Ekwall (2009) noticed similar process in Sweden. Author wrote that cargo theft crime had been displaced from urban areas to less urban areas, mainly because more effective prevention methods had been adopted in big cities. Burges (2013) also exemplify that implementation of safety preventive measures and intensification of law enforcement was forcing agents to operate in other geographical regions. Moreover, it is also possible that police actions targeting the GSP have contributed to the displacement of cargo thefts to non-metropolitan municipalities, facilitated by high quality system of roads and highways, including by the inauguration of express ring roads around the São Paulo municipality.

5 Conclusions and looking ahead: A research agenda and policy implications

The cargo theft is concentrated in the most economically dynamic regions of Brazil, as the states of São Paulo, Minas Gerais and Rio de Janeiro.In 2014 this phenomenon generated a direct cost estimated at 442 million dollars, resulting from more than 17,000 recorded crimes. Half of these crimes occurred in the São Paulo State, which accounts for one third of the national GDP. In this state, in turn, 60% of cargo theft occurred in São Paulo municipality between 2005 and 2016. Most crimes occur during the weekdays and between 9:00 and 16:00, when trade is in operation and there is a greater flow of vehicles on roads and in urban areas. Moreover, there is a lower incidence of this crime after Christmas, between January and February. In this region, the prevailing modus

cargo. The use of violence and of large arsenal of weapons is generally proportional to the cargo value and occurs most often on the main state and national highways.

This study provides evidence of the nature of cargo theft, the recent space-time trends in the country and with especially focus on the state of São Paulo. Its main contribution is that of shedding light on the potential causes of the convergence process in levels of cargo theft between metropolitan and non-metropolitan areas. At the core of this explanation are the two determinant factors of the decision to commit crime monetary return from crime and the probability of failure, which from an economic

angle can significantly differ among regions of the state analyzed in this study. There are three plausible situations: equal risk of failure and different net monetary return; equal net monetary return and different risk of failure; and different net monetary return and risk of failure. Assuming mobility of criminals and economic rationality, crimes will occur in localities where there is higher expected utility – an economic concept, which determined by both factors.

Although São Paulo capital shows the highest levels of cargo theft, it is in non-metropolitan areas that records of this offence are on the rise. This fact was observed especially in three areas: Campinas, Santos and Piracicaba – these areas together constitutes of a strong economic corridor that canalizes products from inland towards the foreign market and vice-versa through the international harbors, especially the most important one in Santos. There is a clear change in the time series behavior after the

2010 fourth quarter for this corridor. A possible reason is that cargo theft increased more in non-metropolitan municipalities because of higher cargo circulation on the state’s highway (more targets in transit, more cargo theft opportunities). Another alternative is that cargo theft increased more in non-metropolitan cities because the risk of failure of theft operations in the GSP increased as a result of implementation of effective crime prevention initiatives. Although there is still no empirical evidence in the literature giving support these hypotheses, we believe that this study contribute to the subject as a first step in this direction. Data permitting, future studies are essential particularly to empirically test the above stated hypotheses.

For future research there is a need to match types of thefts with the situational conditions of crime at detail level. It is desirable studies that can combine evidence from offenders’ modi operandi, cargo type and situational conditions of cargo theft. This is particular important when multiple criminal organizations are set up to commit a sequence of different crimes, having cargo theft as the ultimate goal. A relevant question to be answered here is whether goods from cargo theft vary by destination. Fundamental is also to learn from freight companies that work with supply chain-preventive safety measures; whether and how technology are being put in practice to prevent cargo theft.

In terms of situational conditions, more than identifying regions and particular time windows that crime happens, the next step should include analysis of particular environments where high concentrations of cargo thefts by road segments are found, then

followed by identification of factors that help explain these cargo theft concentrations at micro-level.

From the evidence generated in this chapter, policy makers can allocate resources more efficiently in situational crime prevention in the most targeted municipalities and regions and at most critical periods of time. Furthermore, if this kind of crime continues to rise, it is because there is an expanding demand for these products on the black market. More extensive enforcement in trade regulations, mainly in electronics and food sector could also inhibit the number of cargo theft. Last but not least, it is noteworthy to emphasize that since an efficient and effectively implemented public security policies have a bearing on the rational behavior of potential criminals, then effective national wide public security policies devoted to the problem are fundamental to reduce the high number of cargo theft in Brazil, especially in São Paulo.

6. References

Becker, G. S. (1968). Crime and punishment: an economic approach. The Journal of Political Economy, 76(2), 169-217. doi: 10.1086/259394

Burges, D. (2013). Cargo theft, loss prevention, and supply chain security. Amsterdam, Netherlands: Elsevier.

CAP-SSP Coordenadoria de Análise e Planejamento da Secretaria de Segurança Pública de São Paulo. (2009). Crimes de Carga no Estado de São Paulo 2006 a 2009. Retrieved from http://www.ssp.sp.gov.br/

Ceccato, V. (2015). Thefts of goods transport in Sweden. Washington, DC: Annual Conference of American Society of Criminology, Washington.

Clarke, R. V. (Ed.) (1983). Situational crime prevention: Its theoretical basis and practical scope. Chicago: University of Chicago Press.

Clarke, R. (1999). Hot Products: understanding, anticipating and reducing demand for stolen goods. London: Home Office.

Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: A routine activity approach. American Sociological Review, 44.

Coughlin, J.J. (2013). Cargo crime: security and theft prevention. Boca raton: CRC press.

Ekwall, D. (2009). The displacement effect in cargo theft. International Journal of Physical Distribution and Logistics Management, 39(1), 47-62. doi:

http://dx.doi.org/10.1108/09600030910929183

Ekwall, D. & Lantz, B. (2013). Seasonality of cargo theft at transport chain locations.

International Journal of Physical Distribution and Logistics Management, 43(9), 728-746. doi: http://dx.doi.org/10.1108/IJPDLM-06-2012-0175

Ekwall, D. & Lantz, B. (2015). Cargo theft at non-secure parking locations. International Journal of Physical Distribution and Logistics Management, 43(3), 204-220. doi:

http://dx.doi.org/10.1108/IJRDM-06-2013-0131

FETCESP - Federação das Transportadoras de Carga do Estado de São Paulo. Cargo Theft’s survey in Brazil [Data file]. Retrieved from http://www.fetcesp.net

Federation of São Paulo’s State of Commerce (2016). Pesquisa conjuntural do comércio varejista. Disponível em: http://www.fecomercio.com.br/pesquisas/

FreightWatch International Supply Chain Intelligence Center (2014). 2013 Global cargo

theft threat assessment. Retrieved from:

https://www.naed.org/NAEDDocs/Research/Legal%20Issues/FreightWatch%202013%2 0Global%20Cargo%20Theft%20Threat%20Assesment%20Full_0.pdf

Gameiro, A. H. & Caixeta-Filho, J. V. C. (1999). Demanda por seguro e o desaparecimento de cargas no transporte rodoviário brasileiro. Pesquisa e Planejamento

Econômico, 29(3), 445-462. Retrieved from:

http://repositorio.ipea.gov.br/bitstream/11058/3414/5/PPE_v29_n03_Demanda.pdf

Hesseling, R. (1994). Displacement: a review of the empirical literature. Crime Prevention Studies, 3, 197-230. doi: 10.1.1.375.7985

IBGE - Brazilian Institute of Geography and Statistics (2016). Regional Accounts.

Retrieved from: http://www.ibge.gov.br/english/.

Justus, M. & J. I. Santos Filho. (2011). Convergência nas taxas de crimes no território

brasileiro. EconomiA, 12(1), 131-147. Retrieved from:

http://www.anpec.org.br/revista/vol12/vol12n1p131_147.pdf

Justus, M. & L. G. Scorzafave (2014). Underreporting of property crimes: An empirical economic analysis. Economic Analysis of Law Review, 5(2), 271-284. doi:

http://dx.doi.org/10.18836/2178-0587/ealr.v5n2p271-284

Moreira, J. C. S. & Carvalho, J. L. F. (2011). Investigating the cargo theft on brazilian highways from the perception of the actors involved in the problem. Gestão e

Regionalidade, 27(79), 1-14. Retrieved from:

http://spell.org.br/documentos/ver/3334/investigating-the-cargo-theft-on-brazilian-highw ays-from-the-perception-of-the-actors-involved-in-the-problem

Morettin, P. A. & Toloi, C. M. C. (2004). Análise de séries temporais. Editora Edgard Blucher Ltda. 554 p.

Oliveira, E.N. & Martins, J.H. (2014) Globalização e mercados transnacionais ilícitos na América Latina. Conexão Política, 3(2),11-26.

Rick, B. (1995). The Nature and Extent of Heavy Goods Vehicle Theft. (Crime Detection and Prevention Series: Paper 66.

São Paulo Global (2016). General and business informations about the State of São Paulo. Disponível em: http://www.São pauloglobal.sp.gov.br/eng-index.aspx.

SINESP - Sistema Nacional de Informações de Segurança Pública. (2016) [Data file].

Retrieved from https://www.sinesp.gov.br

SSP-SP - Secretaria de Segurança Pública de São Paulo (2016). Estatísticas trimestrais.

Retrieved from: http://www.ssp.sp.gov.br/novaestatistica/Trimestrais.aspx

StataCorp (2013) Stata Statistical Software: Release 13. College Station, TX: StataCorp LP; R is a free software environment for statistical computing and graphics, available for

download at https://www.r-project.org/; and Grelt is a free and open-source software package, available for download at http://gretl.sourceforge.net/

United Nations. World Urbanization Prospects (2014). Retrieved from:

https://esa.un.org/unpd/wup/Publications/Files/WUP2014-Highlights.pdf. 32 p.