Introduction

What makes individuals in one community commit more crimes than in another? It is a common question in the study of crime. However, from a sociological perspective, this is said to be due to fractures in norms and social cohesion (^{Bouffard & Muftic, 2006}). This is, in general terms, that the deterioration of neighborhood ties, family dysfunction, ethnic heterogeneity, poverty, inequality, the envy effect, and the pressure suffered by certain individual groups are the drivers of crime (^{Merton, 1968}; ^{Agnew, 1992}; ^{Rock, 2012}). The act of committing a crime is described as one, or a set, of unfortunate social circumstances. One way of looking at it is that none of these factors are necessarily mutually exclusive, but rather are different dimensions of the same phenomenon: the need to feel safe.

From the above, ^{Hampson (2013}) argues that the individual needs to be or feel safe. In other words, to safeguard the values that are most appreciated and can make choices. This means, in the most elemental sense, the maintenance of life and being able to do (^{Williams, 2013}). Hence, to appease this insecurity, human being creates social technologies that generate stability and certainty in their daily activities. These are called institutions (^{North, 1990}) and have a fundamental role in building generalized trust (^{Stolle, 2002}). However, when these institutions are dysfunctional, the human being distrusts; and, in many cases, commits a crime. Thus, to understand the relationship between generalized trust - the propensity to crime - community criminal differences, it is necessary to start from the idea that the first search of the human being is the security (^{Williams, 2013}); and therefore, it creates institutions to be able to act confidently. However, when such institutions are perceived as uncertain, generalized trust erodes. Ergo, the propensity to crime is interconnected with generalized trust and with community criminal differences.

The objective of this research is to explore the factors associated with crime propensity, with an emphasis on generalized trust. For this purpose, in the first section, a theoretical link between crime propensity and general distrust is proposed. It is mentioned that the sociological factors of crime gravitate around the human being's inherent need to be or feel safe. An explanation of the sociological theories of propensity to crime, the definition of crime, and the concept and role of generalized trust regarding the intention to commit a crime are offered.

In the second section, the crime incidence in Zapopan is analyzed within the context of the State of Jalisco and the Mexican case. It is important to study the city of Zapopan because it is the largest municipality in terms of extension in the Guadalajara metropolitan area, which in turn is part of the State of Jalisco, one of the three most economically important along with Nuevo León and Mexico City. The following section explains the methodology and data obtained through a survey of 400 households in Zapopan. Answers to questions about age, education, income, gender, life satisfaction, trust in others, the inclination for bribery, propensity for crime, and distrust are explained. Then, through logit models, both positive and negative relationships are obtained to explain crime propensity. In the last section, results and policy implications are offered.

The theoretical relation between trust and crime propensity

Trust provides certainty, enables cooperation, communication, and coordination (^{Putnam, 1993}). ^{Fukuyama (1995}), on the other hand, notes that levels of shared trust on a macro scale generate social well-being. Generalized trust (GT), therefore, is defined as the inclination of individuals to cooperate, although they do not know each other; and it is attributed to good governance (^{Stolle, 2002}) and historical and social processes (^{Putnam, 1993}; ^{Fukuyama, 1995}).

Crime, on the other hand, is a breach of the norm (^{Tappan, 2001}). The study of the propensity to crime, meanwhile, is to understand precisely what motivates an individual to break the law. In this sense, the main traditions of study in sociology are Anomie Theory - General Strain Theory, Control Theory - Economic Theory of Crime, and Social Disorganization Theory. Thus, (1) the Anomie Theory delineates that individuals are motivated to commit a crime because they do not have the means or skills necessary to achieve the goals that society demands of them as ideals (^{Merton, 1968}). Hence, from this perspective, poverty, inequality, and envy effect lead to crime (^{Deller & Deller, 2010}). Another variant is ^{Agnew's (1992}) General Strain Theory, which argues that when people are neglected, rejected, or ignored, they will opt for crime in reaction. (2) Control Theory - Economic Theory of Crime, on the other hand, argues that people commit crimes because it is profitable and enjoyable to do so (^{Rock, 2012}). (3) Social disorganization theory, finally, compares and explains the factors that generate criminal differences between communities (^{Moore, 2019}). In this sense, ^{Ciobanu (2019}) establishes that residential instability and mobility, family breakdown, poverty, ethnic heterogeneity, and population density are causes of crime.

However, there are empirical reasons that invite us to think about a close relationship between generalized trust (GT) and crime. A high degree of social distrust is associated with poverty, economic inequality, unemployment, social marginalization, low schooling, and ethnic and racial heterogeneity (^{Bjørnskov, 2006}); as well as with the poor quality of public services and poor social and distributive policies (^{Fukuyama, 1995}). Likewise, a high criminal propensity is also linked to poverty, economic inequality, unemployment, social marginalization, low schooling, and ethnic and racial heterogeneity (^{Hagan, 1992}; ^{Moore, 2019}; ^{Ciobanu, 2019}); with a weak police and justice system (^{Ehrlich, 1975}; ^{Yamada, Yamada, & Kang, 1993}; ^{Listokin, 2005}); and with community abandonment and detachment (^{Gottfredson & Hirshi, 1990}; ^{Sheley, 2018}).

Indeed, the relationship between generalized trust and crime occurs under the following logic: there is a cognitive process that occurs in the minds of individuals in a situation of poverty, inequality, or rejection; and it is what ultimately leads them to commit a crime. This cognitive component is expected to be distrust. And it is due to the problem of surviving (material aspect) and feeling excluded from a community that provides security. When such distrust is shared with other individuals on a macro scale, there is talk of a generalized distrust that explains the criminal differences between societies. One of the deficiencies in the theories of criminal propensity, from the sociological perspective, is not to abound in this cognitive process.

In the following section, a general description of the city of Zapopan is made and its criminal incidence is analyzed within the context of the state of Jalisco and the country of Mexico, to understand the level of crime in Zapopan as well as its variations in recent years.

The criminal incidence in Zapopan in the context of Mexico

According to the Jalisco Institute of Information, Statistics and Geography (^{IIEG, 2018}), in the document ‘‘Zapopan, municipal diagnosis May 2018’’, the municipality of Zapopan, Jalisco, has an area of 1,017 km2. It borders the municipalities of Guadalajara, San Pedro Tlaquepaque, Tlajomulco De Zúñiga, Ixtlahuacán Del Río, and San Cristóbal de la Barranca. By 2020 the population of Zapopan is projected to be 1,414,972 inhabitants; of which, 689, 327 are men and 725, 645 are women (^{IIEG, 2018}).

On the other hand, the same (^{IIEG, 2018}) outlines that in 2015, 64.6% of the population is in a situation of poverty and some type of social vulnerability. While the other 35.4% is neither poor nor vulnerable. However, considering indicators of marginalization such as illiterate population, lack of public services, and insufficient income, the municipality in question has a degree of marginalization considered ‘‘very low’’ within the national classification. It should be mentioned, however, that in 2017, Zapopan had 50,881 formal companies; predominantly service-oriented (47.2%). Being the food industry, the beer industry, the tobacco industry; and retail, the most outstanding (^{IIEG, 2018}).

Regarding crime incidence, measured through intentional homicides per 100,000 inhabitants during the years 2015 to 2019, Zapopan is below the national average and almost always below the Jalisco average. Table 1 shows this indicator for some states, as well as for Jalisco and the municipality of Zapopan. The information was obtained from the Executive Secretariat of the National Public Security System (SESNSP in Spanish).

It is important to point out that the intentional homicide rate per 100,000 inhabitants is used as an indicator of crime incidence because homicide is a crime that can be measured more reliably through death certificates. Crimes such as theft of a vehicle, house, or passerby are not always formally reported.

Table 1 shows some States in Mexico with more intentional homicides for every 100,000 inhabitants. The states with the maximums for 2019 are Colima, Baja California, Chihuahua, Morelos, Guanajuato, Guerrero, and Quintana Roo. However, we observe that some entities have shown very accelerated growth in the period from 2015 to 2019. For example, Colima has had an increase of 280%, Guanajuato 209%, and Baja California 198%. On the other hand, in the State of Guerrero, even when its crime rate is high, a decrease of 23% has been observed from 2015 to 2019.

The states with the lowest crime incidence in 2019 are Yucatan, Aguascalientes, Coahuila, Campeche, Querétaro, and Durango. Likewise, favorable changes are observed during the period from 2015 to 2019. For example, a decrease of 39% is observed in the rate of intentional homicides in the states of Durango and Yucatán.

Jalisco has an average of 18.1 intentional homicides per 100,000 inhabitants during the years 2015 to 2019, showing an increase of 102% in this period. It should be noted that the numbers of this state regarding this rate are close to the national numbers. To cite an example, this indicator in 2019 was 24.3 for Jalisco, while at the national level, it was 23.2.

Zapopan has maintained an intentional homicide rate below the national and Jalisco levels, especially in the years 2018 and 2019. Likewise, the percentage increase from 2015 to 2019 in Zapopan (42%) is significantly less than that of Jalisco (102%) and the national (75%).

The following section explains the methodology used, based primarily on a survey of 400 Zapopan households, as well as the statistical analysis of the variables involved and the relationship between them.

Methodology and description of data

In this research, factors associated with the propensity to crime are explored in 400 citizens interviewed in Zapopan. In addition to the attributes of the individuals, their propensity to crime is evaluated through an approximate question that is the following: “Sometimes it is necessary to break the law even if we don't like to admit it”.

As the answer is dichotomous "Yes" or "No", the Ordinary Least Squares model is not the most appropriate because it assumes a linear behavior when that of a model with a dichotomous response variable is not. Therefore, it is decided to carry out Logit models that have characteristics that allow better modelling of the binary qualitative dependent variable.

A face-to-face survey was applied to 400 households in October 2018 in the Mexican city of Zapopan, located in the state of Jalisco, Mexico. The respondents were in homes located in 20 representative colonies of the municipality.

Among others, the survey asked questions about age, education, income, gender, life satisfaction, trust in others, bribe propensity, crime propensity, and distrust. Table 2 shows the questions and descriptive statistics of the responses of the citizens surveyed.

The average age of the people interviewed is 45.5 years. 21.3% are between 18 and 24 years old; 29.5% between 25 and 44; 24.5% between 45 and 64; 23.5% between 65 and 80; and 1.3% are over 80. Most respondents have a bachelor's degree (33.3%), high school (28%), and basic education (24.8%). The majority earn between 2,650 and 13,254 Mexican pesos per month (37.5%); and between 13,254 and 26,508 (38.3%). Assuming an average exchange rate of 20 Mexican pesos per US dollar would be between 132.5 and 662.7 dollars per month (37.5%); and between 662.7 and 1,325.4 USD/month (38.3%). Almost half are men (49.8%) and half are women (50.2%).

Concerning life satisfaction, the vast majority declare to be very satisfied (75%) and satisfied (20.3%). No one answers not being satisfied. A large part of the respondents expressed that they should trust in others fairly (34.3%), and more than half indicated that they should not trust in others (50.3%).

71.8% think that bribing is not acceptable. However, 41.5% state that sometimes it is necessary to break the law even if we don't like to admit it, although the vast majority think otherwise (58.5%). Likewise, most respondents justify insecurity due to a lack of education and opportunities, and for necessity (70%); while 16.3% associate it with a cultural issue, and 13.8% state that insecurity is generated because people are bad.

The correlation matrix can be seen in table 3. As expected, a high correlation is observed between age (AGE) and age in ranges (AGERANGE) (0.96), and between distrust (DISTRUST) and distrust in three levels (DISTRUST3L). Out of these two cases, no high correlations were found between the variables. The highest are those between education (EDU) and income (INCOME) (0.47) and between corruption propensity (CORR) and crime propensity (CRIMEPROP) (0.37). In the following section, different logit models will be carried out to obtain the factors that are associated with the crime propensity of the citizens surveyed.

The empirical results section seeks to demonstrate the relationship between the crime propensity and the explanatory variables indicated in the theoretical section, exploring the role of generalized trust.

Empirical results

Logit models are used to explain dependent variables with values of zero or one. In this research, the variable to explain is crime propensity (CRIMEPROP) which takes the value of one when the individual states that sometimes it is necessary to break the law even if it is difficult to admit it, and zero in another case.

Table 4 shows the logit econometric models carried out to observe the factors associated with crime propensity. It is appreciated that significant variables were AGE, education (EDU), propensity for corruption (CORR), and DISTRUST (with four and three levels). AGE was significant with the expected negative sign in three models, in one at 10% and in two at 5%. Education was significantly negative in the five models, in two at 5% and in three at 1%. Corruption propensity (CORR) was used in three models, resulting significantly positive at 1% in all of them. And distrust was used with its original four levels in three of the models, and in three levels in two models. With four levels it was significantly positive at 1% and with three levels was significantly positive at 5%. The percentage correctly predicted (PCP) of the models varies between 60.5% and 71.3%. The models that combine parsimony with a high percentage correctly predicted are two and four.

Table 4 Estimated logit model to determine the crime propensity (CRIMEPROP) 

Source: authors.

Table 5 shows the regression coefficients as odds ratios. It is observed that the odds ratio of the age variable is around 0.98 and 0.99. If the inverses are obtained, it can be concluded that for each unit that increases the age, there is slightly more probability (1.00 - 1.01) that the person is not prone to the crime that it is.

Regarding the education variable (EDU), the odds ratios are between 0.78 and 0.83. If the inverse is obtained, it can be inferred that, on average, an increase in the EDU indicator causes it to be between 1.2 and 1.3 times more likely that the person is not prone to the crime that it is.

On the other hand, the odd ratios of the CORR variable are close to 2, which means that when the indicator of the propensity for corruption (CORR) is increased by one, it is 2 times more likely that the citizen is prone to crime than not.

Regarding the distrust indicator in four levels, it is found that the odds ratios are between 1.37 and 1.43. The above indicates that when the distrust indicator (DISTRUST) increases by one unit, the respondent is more likely to be prone to crime than not between 1.37 and 1.43 times. In the three-level distrust indicator (DISTRUST3L), the odds ratios are between 1.54 and 1.59. And the interpretation is the same but with these ranges.

Table 5 Odds ratios of the logit model to determine the crime propensity (CRIMEPROP) 

Source: authors.

With the study data, it is observed that a person is more likely to commit a crime or break the law when he is younger, has a lower educational level, is prone to corruption, and has a high level of distrust. In the last section, some conclusions are offered on the main findings of this document and the possibilities for generating public policy actions.

Policy implication and conclusions

As mentioned above, this study shows evidence of the impact of a citizen's age, education, proclivity for corruption, and degree of mistrust, on crime propensity or willingness to violate the law.

According to the results found, and in general terms, we can conclude that there is a greater propensity for crime in younger people, with a lower degree of education, less social trust, and with a greater inclination to corruption. In most of the models analyzed, these variables have a significant impact and the goodness of fit of these models is between a correctly predicted percentage of 60.5% and 71.3%

However, beyond the interpretation of the numbers, it is important to describe the policy implications. This is the importance of building trust. This is linked to the notion of being or feeling safe regarding the problem of surviving and being able to make choices (^{Williams, 2013}). For this, the provision of rights of individuals and having reliable political and social institutions that guarantee overcoming poverty, inequality, and unemployment are essential; promoting family and community ties and the political inclusion of minorities; and guarantee access to health services, personal and legal security; and to public services in general. Of all this, however, it is important to place greater emphasis on youth. For which the creation of programs that emphasize the importance of respecting the law and the rights of their peers is also recommended. Education, on the other hand, is a crucial tool in building trust, every time it generates capacities that allow for civic, ethical, and professional development.

Regarding trust in institutions, it is important to punish corrupt acts and increase the cost of crime, as this inhibits the incentives for the law to violate. Increasing the level of social trust is the most relevant factor because it modifies the perception of limitations in citizens and more involvement with their community.