Abstract
Using data from the 2012 AmericasBarometer survey collected in nine Latin American countries (n = 14,705), this study examines two theoretically relevant potential sources of punitiveness: economic anxiety and fear of crime. Focusing on these two sources, we explore whether the public opinion dynamics often highlighted by punitiveness scholars also apply to the Latin American context and can thus be of value to explain recent movements towards punitive policies in that region. Generalized Structural Equation Modeling (GSEM) and bootstrapping are used to assess the direct effects of perceived national and personal economic insecurity on punitive sentiments as well as the indirect effects of these attitudes on punitiveness through fear of crime during a time of reduced economic growth, increased economic inequality, and harsh criminal justice policies. Results show that economic anxieties are positively associated with fear of crime, though their effects on support for increased punishments are mixed. Additionally, the effects of economic insecurity on punitiveness are partially mediated by fear of crime, supporting theoretical notions that insecurity produces fear of crime and subsequently influences punitiveness.
Recent decades have seen notable shifts in public policy toward punitive forms of crime control, especially in industrialized Western countries (Beckett and Sasson, 2000; Garland, 2001). Research has documented that these “get tough” policies are largely supported by the public, and scholars speculate that “penal populism” may be as much a response to fear of crime as an attempt to confront criminal activity (Bottoms, 1995; Enns, 2014; Roberts et al., 2002). There has also been an increase in public favor toward more punitive policies in Latin American countries (Swanson, 2013). Within this area of research, scholars have noted that a variety of potential individual and contextual factors are antecedents to punitive attitudes (Cullen et al., 2000; Unnever and Cullen, 2010).
One possible source of support for punitive policies is “diffuse anxieties” about societal or personal economic deterioration (Garland, 2001; Gottschalk, 2014), and these anxieties may translate into fear of crime (Britto, 2013; Costelloe et al., 2009; Dowler, 2003; Holloway and Jefferson, 1997). Individuals’ anxieties about their national and personal economic situations might become conflated with concerns about crime, leading to more punitive views of crime and criminals. Indeed, Garland (2001) and Simon (2007) argue that the cultural and economic anxieties brought on by late modernity have produced a “culture of fear,” and this fear has translated into punitive sentiments and the subsequent adoption of strict crime control legislation. Similarly, Wacquant (2010) emphasizes that fear of crime is a product of not only crime rates but also diffuse fears about personal and national economic situations, which are the result of the expansion of neoliberalism and flexibility within the labor market.
Corresponding with these theoretical expectations, prior empirical work has established that economic situations and fear of crime are directly and independently related to punitiveness (e.g., Costelloe et al., 2009; Dowler, 2003; Hogan et al., 2005; Ousey and Unnever, 2012). Additionally, previous research has acknowledged that those in poor or deteriorating economic situations are more likely to be fearful of crime (Britto, 2013; Hale, 1996; McKee and Milner, 2000; Panatazis, 2000; Rader et al., 2012). Despite evidence of these relationships, no prior study has examined whether the direct effects of economic insecurity on punitiveness are driven by an indirect effect of fear of crime. Further, little research on these issues has been conducted outside the US, and the direct and indirect effects of economic anxieties and fear of crime on punitiveness have not yet been examined within the theoretically relevant context of Latin America. This lack of research is surprising given that this region recently has been marked by economic instability (Bértola and Ocampo, 2012; Flores-Macias, 2012) and has experienced a publicly supported movement toward more punitive crime control (Holland, 2013; Hume, 2007; Swanson, 2013).
Using data collected in nine Latin American countries, the current study situates fear of crime within the context of economic insecurity to determine whether the effects of economic anxieties on punitiveness operate through fear of crime. This article begins by providing a theoretical framework describing the possible role of fear of crime in the relationship between economic anxiety and punitive attitudes. This work is then connected to the existing body of literature as well as within the context of Latin America. Next, the hypotheses of the study are presented, followed by a discussion of the data and research methodology, the results of the analyses, and the implications of the findings for theory and research.
Theoretical framework
Scholars have theorized that anxieties related to personal or national economic insecurity might lead to punitive sentiments (Garland, 2001; Greenberg, 2002; Melossi, 1985, 1993), and several theoretical rationales might explain this hypothesized relationship. For example, research shows that the public may believe that unemployment and crime are linked and that harshly punishing offenders is a necessary response to increases in crime (Box, 1987; Melossi, 1985; Simon, 2007). Therefore, in contexts of economic insecurity, the general public may more strongly support punitive criminal justice strategies to combat crime. Alternatively, some scholars speculate that the presumed “crime problem” might be used by the public as an outlet for misplaced fears and worries, and criminals may be identified as scapegoats for other more complex social issues (Chancer and Donovan, 1994; Gans, 1995; Tyler and Boeckmann, 1997). Indeed, for individuals who perceive personal or societal economic insecurity, criminals may “provide these emotions with a readymade, deeply unpopular, target population against whom they could be directed” (Garland, 2001: 133). This is in addition to Garland’s assertion that there has been a decline in rehabilitation as a criminal justice philosophy. In his view, the fear of crime that stems from deteriorating social and economic conditions and the solidification of an individualistic ideology, which emphasizes individual responsibility in late modern societies, undermined the public’s support for the penal-welfare state and allowed for the construction of the poor and the criminal as feared “others.” These disadvantaged groups in society are blamed for their own fate and thus deserving of harsher punishment and social exclusion instead of support and inclusion from a welfare state (Garland, 2001).
Beyond economic insecurity, fear of crime may also be a source of punitiveness. Since the 1970s, crime has increasingly become a concern for citizens of Western countries, and the resulting fear of crime has been recognized as a prevailing trend (Flanagan and Longmire, 1996; Hough and Roberts, 1998; Roberts et al., 2002). Garland (2001) argues that one of the consequences of fear of crime is increases in punitive sentiments, illustrated by a movement toward punitive legislation aimed at controlling crime. He posits there is a clear link between fear of crime and punitiveness, as many strict penal policies were explicitly created to temper the public’s fear. Ultimately, Garland asserts that, at the national level, responding to fear of crime and support for the victims of crime has become a legislative priority, which was translated into the crime control policies of the “get tough” era.
Garland (2001), Wacquant (2010), and others provide possible rationales for expecting indirect effects of economic anxieties on punitiveness via fear of crime. Garland (2001) argues that cultural transformations in late modernity, including capitalist competition and struggles for political and social equality emerged simultaneously with increased fear of crime, often disproportionate to actual crime rates. In fact, according to Garland (2001), the culture of control—and the punitive shift in the criminal justice system it encompassed—were eminently linked to the sharp social and economic changes that characterize societies in late modernity and the conservative and free-market views that emerged as a response to them in the 1980s. Thus, while Garland acknowledges that the increase in crime was a driving force of the punitive turn, he also asserts that neither changes in crime nor the political and media rhetoric were enough to fully explain this sharp transformation in the crime control field. The culture of control is based on a different social sensibility product of late modernity, in which crime and fear of crime take a central role. The economic changes that led to the fragmentation of the labor market and the decline of the welfare state, the restructuration of family life, the process of suburbanization and transformation of the ecology of the city, the emergence of electronic media, and the democratization of cultural and social life all contributed to generate a different social sensibility about crime. New broad social insecurities were at the center of this new form of social life and fueled a punitive sentiment in the public. The new collective experiencing of crime is that mediated by a new ontological insecurity experienced by middle class families. These new insecurities are produced by a fragmented labor market, a more unstable professional career, and the retreat of the welfare state, all which undermine the sense of control that middle-class families used to experience and the safety net they relied on. These diffuse anxieties thus change individuals’ sensibilities about crime and fuel their fear of crime.
Wacquant (2009) is more adamant in emphasizing the role of neoliberalism and the fragmentation and precarization of social and economic life it generates as the main driver of an increased fear of crime. In his view, the increase in fear of crime (and the punitive sentiments it engenders) observed in contemporary societies is unrelated to changes in crime in itself. Rather, it is the expression of the increasing economic anxieties that fuel punitive sentiments. Thus, “the return of the prison to the institutional forefront of advanced society over the past quarter-century is a political response, not to rising criminal insecurity, but to the diffuse social insecurity wrought by the fragmentation of wage labor and the shakeup of ethnic hierarchy” (Wacquant, 2010: 198). Punitive crime control policies are thus an instrument to strengthen the neoliberal state in an era of economic deregulation, not, as Garland (2001) suggests, a mere “acting out” of a weakened state visibly incapable of producing order.
In essence, political figures use fear of crime to justify imposing punitive measures to protect against a specific type of offender (see also Chevigny, 2003; Holloway and Jefferson, 1997; Simon, 2007). These theorists, therefore, make clear the implications of heightened fear on the relationship between economic anxieties and punitive attitudes as well as the relationship between fear of crime and punitiveness; however, previous research has failed to explore the possible mediating effects of fear of crime on the relationship between economic anxieties and punitiveness. The primary aim of the present research is to investigate these direct and indirect effects.
Prior research
Economic anxieties and punitiveness
Despite the strong theoretical connection between perceived economic insecurity and punitiveness, relatively few empirical examinations of this relationship have been conducted, and the findings from this body of work are mixed. Among the studies that show a positive effect, Hogan et al. (2005) found those who expected to be worse off financially in the next year were more punitive than those who did not, though this effect was only observed among respondents who were not White males. The research by Costelloe et al. (2009) also indicated that expected personal economic deterioration increased punitiveness for White males, particularly those with lower levels of education. In a study of attitudes among the British public, King and Maruna (2009) observed those who saw the national economy as troubled were more punitive than those who did not. Finally, in their cross-national analysis of European nations, Ousey and Unnever (2012) found that respondents who believed their country’s employment situation would soon worsen were more supportive of tougher punishments for criminals.
Several other studies investigating the link between economic anxieties and punitive attitudes reported null findings or unexpectedly negative relationships. For instance, Johnson (2001) found that a perceived decline in one’s personal financial situation was not related to support for the death penalty or harsher criminal courts. Unnever et al. (2008) also analyzed US national data and observed that personal financial harm was not predictive of support for harsher sentences for corporate crime. Recent studies by Britto and Noga-Styron (2014) and Kornhauser (2013) similarly demonstrated that concerns regarding one’s future economic situation were not associated with supporting capital punishment. In addition, Useem et al. (2003) discovered that those who viewed their personal financial situations as deteriorating were unexpectedly less likely to favor punitive measures and criminal justice spending. Similarly, Lehmann and Pickett (2017) also found expected personal, vicarious, and societal economic insecurity to be negatively associated with support for the death penalty.
Fear of crime and punitiveness
The relationship between fear of crime and punitive sentiments has also received empirical attention. Some research has reported that those who are more fearful of crime are also more punitive (e.g., Hogan et al., 2005; Johnson, 2006; Rankin, 1979; Sprott and Doob, 1997; Thomas and Foster, 1975; Tufts and Roberts, 2002; Unnever et al., 2008; Wu et al., 2011). For instance, Costelloe et al. (2009) found that fear of crime is a consistent predictor of support for a wide range of punitive criminal justice policies. Dowler (2003) reported similar findings, showing that those who are more fearful of crime are more likely to hold 11 different punitive attitudes, including thinking the government should focus on punishment, opposing parole and early release for good behavior, and opposing shorter sentences. Other research also demonstrated that those who fear crime are also more likely to think that adult sentences are too lenient (Sprott and Doob, 1997) and to think that prisons have the primary purpose of punishment over rehabilitation (Langworthy and Whitehead, 1986). Even when controlling for other emotional reactions to crime such as anger, fear of crime remains a significant predictor of punitiveness (Johnson, 2009).
In contrast with these studies, however, other research suggests that fear of crime does not consistently impact punitiveness (Chiricos et al., 2004; King and Maruna, 2009; King and Wheelock, 2007; Kleck and Jackson, 2017; Leverentz, 2011; Messner et al., 2006; Stack, 2003; Taylor et al., 1979; Tyler and Weber, 1982; Welch et al., 2011). For example, King and Maruna (2009) observed that fear of crime did not predict support for harsher criminal penalties and that contextual factors such as education and political ideology played a larger role than fear in punitiveness. A comparison of average citizens and court practitioners showed that, while fear of crime does not impact the attitudes of average citizens, it can influence practitioner decision-making (Ouimet and Coyle, 1991).
One difficulty when considering prior fear of crime research is the conceptualization and operationalization of fear of crime. Fear of crime is a notoriously difficult notion to conceptualize and is particularly complex to measure. In fact, the emergence of “fear of crime” as an important object of study and policy is in itself historically situated. As Lee’s (2001) genealogy of its emergence suggest, the notion of “fear of crime” helps to objectivize and dissect a collective subjectivity. Thus, “fear of crime” becomes a social problem that gains reality on its own and is further used by politicians to advance their punitive agenda, which further gives visibility to the problem of “fear of crime” it creates.
Conceptually, understanding how the criminological study of “fear of crime” has shaped social and political representations of such fear is thus paramount. In a methodological vein, researchers differ in their operationalization and measurement of fear of crime. Thus, studies measure it in such varied ways that the validity of fear of crime measures is called into question (Lane et al., 2014). When measuring fear of crime, scholars measure perceived risk or fear of victimization (Ferraro, 1995) or emotional responses or reactions to crime (Garfalo, 1981). Because there is no universal conceptualization of fear of crime, and studies vary in the types of questions used to measure fear of crime (Lane et al., 2014), generalizing about the fear of crime literature as a whole is difficult. For example, some studies ask about fear of particular crimes (e.g., Warr and Stafford, 1983; Wilcox Roundtree, 1998), others focus more on perceptions of neighborhood safety (e.g., Erksine, 1974; Garofalo, 1979), and other still focus on fear in particular locations. Despite the various ways of understanding the concept of fear of crime, prior research shows that with regard to punitiveness, under certain circumstances, a relationship can exist between fear of crime and punitiveness.
Economic anxieties, fear of crime, and punitiveness
Although economic anxiety and fear of crime have been shown to be independent sources of punitive attitudes, there is also evidence that these two factors are linked, and some studies suggest that economic insecurities or a lack of economic resources can be sources of fear of crime. Indeed, previous research has found that those who make less money are more fearful of crime than wealthier individuals (Hale, 1996; McKee and Milner, 2000; Panatazis, 2000; Rader et al., 2012), though there are exceptions (e.g. Franklin et al., 2008; Schafer et al., 2006; Skogan and Maxfield, 1981). The relationship between poverty and fear of crime has also been explored, and studies find that potential precursors to fear of crime include poverty (Delone, 2008; Will and McGrath, 1995) and neighborhood-level disorder (e.g., Brown and Benedict, 2004; Chiricos et al., 1997; Covington and Taylor, 1991; Franklin et al., 2008; Gray et al., 2011; Hinkle and Weisburg, 2008; Roundtree and Land, 1996; Scarborough et al., 2010). While this prior research has demonstrated that links exist among objective economic conditions, fear of crime, and punitive attitudes, it has failed to consider the subjective evaluations of economic worries that are likely theoretically relevant (Garland, 2001; Wacquant, 2010). The current study extends this line of inquiry by assessing the extent to which economic insecurity has an indirect effect on punitiveness through fear of crime.
The Latin American context
Latin America is a unique setting in which to explore economic anxieties, fear of crime, and punitive attitudes. Regarding the region’s economic situation, Latin America experienced notable growth between the 1950s and 1970s because of increased reliance on import substitution industrialization (Thorp, 1998; Torche, 2010). However, socio-political turmoil and the rise of authoritarian governments in the 1970s and 1980s led to the adoption of neoliberal economic reforms (Babb, 2013; Leiva, 2008), which, combined with excessive borrowing, a rise in international interest rates, and the world recession, resulted in widespread economic decline in the 1980s. To confront these issues, Latin American governments implemented a variety of structural measures, including spending cuts, deregulation, privatization, and currency devaluation (Edwards, 1995; Morley, 1995; Stallings and Peres, 2000). These strategies had widespread negative consequences, including increased unemployment, spikes in poverty rates, and declining wages (Psacharopoulos et al., 1995). With increasing democratization after the start of the 21st century, newly elected left-leaning governments implemented redistribution policies, which produced economic growth and reduced poverty (Flores-Macias, 2012; Grugel and Riggirozzi, 2012; Spronk, 2008). By 2012, however, economic growth had stalled, and a push for wealth distribution across social classes was again at the center of political discussions (Bértola and Ocampo, 2012; Reyes and Sawyer, 2016).
Fear of crime in Latin America
Latin America also represents a useful context in which to study crime-related anxieties. Increases in crime and violence beginning in the 1980s have become a defining characteristic of Latin America (Davis, 2006), and, as a result, fear of crime is high in many Latin American countries (Chevigny, 2003; Dammert, 2012), even those such as Chile with relatively low crime rates (Dammert and Malone, 2003). For example, Argentina exhibited a sharp increase in levels of fear of crime, from 20% in the 1980s to 80% in 2009 (Kessler, 2011), and there are high levels of fear of crime in Mexico as many Mexicans do not feel safe in their neighborhoods and protect their homes with locks or barred windows (Vilalta, 2010). Studies of fear of crime in Latin America show that these heightened fears have implications for criminal justice related outcomes such as reduced trust in criminal justice institutions, including the courts and the police (Dammert and Tobar, 2018; Malone, 2010). Similarly, victimization has also been shown to decrease trust in local government institutions such as the local police (Corbacho et al., 2015) and fear of crime is an important consequence of crime that, according to Latin Americans, undermines the social fabric (Varela and Schwaderer, 2010).
Although pervasive, fear of crime in Latin America is a complex phenomenon. According to a qualitative study by Kessler (2011), exploring Argentinian’s worries about crime, individuals do not fear all crime, rather they fear specific random crimes they fear they may become a victim of. As evidence that this fear may be a result of the perception among Latin Americans that they have a high chance of victimization, data from the Latinobarómetro survey (2016) shows that 41% of Latin Americans worry “almost all of the time” about becoming victims of a serious crime. There are complexities between individual regions and locations, yet scholars have concluded that fear of crime is a widespread problem in Latin America with repercussions for a variety of societal issues (Dammert and Tobar, 2018). Prior research on fear of crime in Latin America is more exploratory, and what these studies do not emphasize is the role of more theoretically driven and constructs aimed at capturing social subjectivities such as economic anxieties that may be relevant for understanding fear of crime.
Fear of crime also has important implications for public punitiveness, as this public insecurity, amplified by exaggerated media reports of violent gang activity, has led politicians to implement mano dura (i.e., “iron fist”) crime control policies, which have included “broken windows” policing, limits on offenders’ due process, and harsh anti-crime legislation and sentences (Holland, 2013; Hume, 2007; Krause, 2014; Swanson, 2013). Although there is little evidence to suggest these policies effectively reduce crime and may instead exacerbate disparities in criminal justice contact (Holland, 2013; Hume, 2007; Peetz, 2011), they have been implemented in many Central and Latin American countries (Bonner, 2016; Holland, 2013; Wolf, 2017). The public largely supports these punitive policies, though variation in public punitiveness has been linked with crime salience and victimization (Bateson, 2012; Malone, 2010; Seligson, 2003). However, the effects of economic insecurities on punitiveness have not been explored in this region.
Latin America and the punitive turn
Most of the literature depicting the punitive turn has focused on Western industrialized countries, especially the United States and England, and centered mostly on policies enacted in the last quarter of the 20th century and the early years of the 21st century (Garland, 2001; Simon, 2007, Wacquant, 2009). Many scholars now suggest that the punitive turn may be over and that countries like the United States (where punitive policies expanded in unique ways) may now be entering in a period of decarceration and downsizing (Clear and Frost, 2015; Petersilia and Cullen, 2015; Tonry, 2014). Public sentiments seem to echo these processes, with several studies emphasizing the publics’ mixed views regarding punishment and their considerable support for rehabilitation and other non-punitive interventions (Jonson et al., 2013; Pickett, 2016; Sundt et al., 2015; Thielo et al., 2016).
In Latin America, this retreat in punitiveness seems less clear. Many countries implemented post-neoliberal reforms during the first one and a half decades of the century. Even if not without contradictions, many of these reforms were aiming to counter previous policies of economic privatization and deregulation by promoting social and economic inclusion and democratic reforms (Ruckert et al., 2017). However, recent political events are showing a shift back to neoliberal and authoritarian government policies across the region (Scoones et al., 2017). In Brazil, for example, this includes the recent inauguration of a far-right President, Jair Bolsonaro (Londoño and Andreoni, 2018), and subsequent shift back to punitive crime control against “urban criminals.” Therefore, while the majority of Brazilian citizens may be in favor of protecting human rights, societal level political shifts and increased fears of violence by marginal others may be responsible for the appearance of punitiveness (Paes-Machado and Vilar Noronha, 2002; Wacquant, 2003). In Argentina, support for governmental violence toward out-groups has increased following the election of right-leaning President Mauricio Macri in 2015 (Esperanza Casullo, 2016).
Prior studies of public opinion show that Latin Americans list combatting crime as either the first or second most important public issue (Dammert, 2012), but others find greater support for social policies. Argentinians in particular report supporting more progressive crime controls and rank punitive policies as the third most supported type of policy behind social policies (i.e., education and employment) and legal (i.e., better law enforcement and police reform) (Kessler, 2011). There is also evidence that they are more concerned with perceived moral decline more than criminal insecurities (Kessler, 2011), further highlighting the complexities of causes and consequences of fear of crime in Latin America. Despite this example of decreased punitiveness, scholars recognize there is a lack of knowledge and research about the social, political, and economic implications of fear of crime in Latin America (Dammert and Tobar, 2018).
In light of this discussion, the current study examines two hypotheses. Hypothesis 1: Economic anxieties and fear of crime will exert direct effects on punitiveness. Hypothesis 2: There will be indirect effects of economic insecurity on punitiveness through fear of crime.
Data and methods
This study uses data from the 2012 AmericasBarometer survey, a biennial survey collected in North, Central, and South American and Caribbean countries by the Latin American Public Opinion Project (LAPOP). 1 Using the most recent national census to create strata based on the sizes of municipalities, urban/rural areas, and regions, respondents are selected using a multistage cluster design. This survey design produces samples that are representative of the non-institutionalized voting-age population of each country. Face-to-face interviews are conducted with one individual per household, and quotas for age and sex are used to avoid sampling error due to differential nonresponse. Respondents are asked about experiences, perceptions, and attitudes about a range of political, economic, and criminal justice topics. The 2012 AmericasBarometer survey was administered to 41,632 respondents in 26 countries. However, due to inconsistencies in the questionnaires across countries, only respondents from the nine countries (i.e., Argentina, Belize, Bolivia, Brazil, Colombia, Costa Rica, Mexico, Paraguay, and Uruguay) in which all relevant survey items were included will be assessed (n = 15,145). The descriptive statistics for the study variables are presented in Table 1.
Descriptive statistics: AmericasBarometer Survey, 2012.
SD: standard deviation.
Note: Country dummy variables not displayed.
Dependent variable
The dependent variable used in this study is support for harsher punishments, which is taken from a survey item asking, “In your opinion, what should be done to reduce crime in a country like ours?” The two response options include “implement preventative measures” and “increase the punishment of criminals”. Respondents indicated support for one method or, without prompting from the interviewer, both strategies. Those who favored increased punishment rather than preventative measures or both policies were coded as 1. 2
Independent and mediating variables
The primary predictor of interest is economic anxieties, which are measured using four survey items about perceptions of national and personal economic conditions. First, poor national economy was measured using the question, “How would you describe the country’s economic situation?” Responses were coded on a 5-point ordinal scale ranging from very good (=1) to very bad (=5). Second, to capture economic deterioration, respondents were asked whether they perceived a worsening national economy using the survey item, “Do you think the country’s current economic situation is better than, the same as, or worse than it was 12 months ago?” Respondents were coded as 1 if they viewed the economy as getting worse (0 = the same/better). Participants were also asked about their personal economic situations. The first (poor personal economy) asked how respondents would describe their overall economic situation (1 = very good, 5 = very bad). Finally, a dummy variable indicating whether respondents saw themselves experiencing a worsening personal economy over the past 12 months, and those who answered in the affirmative were coded as 1. The correlations among these variables are low, which suggests that multicollinearity is not an issue.
The present study also examines the mediating influence of fear of crime, which was measured by assessing the extent to which each respondent feels safe in his or her neighborhood. Responses were coded on a scale of 1 (=very safe) to 4 (=very unsafe). 3
Control variables
Control variables that were identified as potential sources of punitiveness and that might also be correlated with economic anxieties and fear of crime were included. First, the race or ethnicity of the respondent is controlled using five mutually exclusive dummy variables: White, Mestizo, Indigenous, Black/Mulatto, and Other with White as the reference category. Black and Mulatto respondents were combined into a single category due to the small number of participants in each category (3.17% and 5.79%, respectively). Other demographic controls include gender (male = 1), age in years, and education represented by the number of years of schooling completed (0 = none, 18 = 18 or more). Participants who reported that they had ever been married and those who have any children were coded as 1 on the marital status and parental status variables, respectively.
To capture respondents’ socioeconomic status, we include a dummy variable indicating whether the respondent receives monthly financial government assistance (yes = 1) as well as a measure that reflects how many products and services out of 14 possible items (i.e., television, refrigerator, landline telephone, cellular telephone, vehicle, washing machine, microwave oven, motorcycle, indoor plumbing, indoor bathroom, computer, internet access, flat panel TV, and a sewage system connection) are in the respondents’ home. These measures are preferable to a monthly income item because differences in currencies across the nine countries can compromise the reliability of the cross-national comparisons. Political ideology was measured on a scale of left (=1) to right (=10). Two ordinal measures capture respondents’ personal religious salience (1 = religion is not important at all to 5 = religion is very important) and religious attendance (1 = never or almost never, 5 = more than once per week).
Due to evidence that victimization is associated with punitiveness, especially in international contexts (e.g., Hanslmaier, 2013; Wu et al., 2011), dummy variables are used to control for any personal crime victimization in the past 12 months (=1) and any vicarious crime victimization (i.e., of another person living in the respondent’s household) in the past 12 months (=1). News media consumption captures how often respondents pay attention to the news via the TV, radio, newspapers, or the internet (1 = never, 5 = daily). This study also accounts for the extent to which respondents trust the criminal justice system of his/her country (1 = not at all, 7 = a lot) and also have trust in the judicial system, that is, how much faith the respondent has that the judicial system would punish someone guilty of robbing or assaulting the respondent (1 = none, 4 = a lot). Finally, to capture general social mistrust, we control for whether respondents believe that they have trustworthy neighbors (1 = very untrustworthy, 4 = very trustworthy). Finally, to account for between-country differences, each model includes country fixed effects dummy variables with Brazil as the reference category. The descriptive statistics and coefficients of these latter variables are not presented in the tables.
Little’s (1988) test was conducted, and the results indicated that the missing data are not missing completely at random. Therefore, missing cases were imputed through multiple imputation using chained equations and 20 imputations (White et al., 2011). The imputation model included all variables in the analyses, though cases missing on the dependent variable were dropped before the analyses were conducted. The final imputed sample includes 14,705 cases.
Analytic strategy
Generalized Structural Equation Modeling (GSEM) is used to examine the direct and indirect effects of the four types of economic anxiety (poor national economy, worsening national economy, poor personal economy, and worsening national economy) on punitiveness. GSEM is useful for these analyses because the direct effects of the four measures of economic anxieties on fear of crime and punitiveness are modeled simultaneously. A logit model is estimated with support for increased punishment regressed on fear of crime, economic anxieties, and the controls, and a linear regression simultaneously models the relationship between economic anxieties and fear of crime. Using the first imputation, bootstrapping is used to test the statistical significance of the indirect effects of economic anxieties on punitiveness as it operates through fear of crime. A single GSEM is estimated for all nine of the countries together. Finally, sensitivity analyses were conducted to assess the robustness of the results.
Results
Generalized Structural Equation Models
The findings reveal the presence of direct effects of economic insecurity on punitiveness. As shown in Table 2, one national and one personal measure of economic insecurity have significant effects on punitive attitudes. Indeed, respondents who believe that the national economy is poor report greater support for increased punishment (b = 0.053, p < .05). However, those who report a poor personal economic situation are less supportive of punitive policies (b = −0.058, p < .05). The findings also indicate that fear of crime is associated with increased punitiveness, and those who have higher levels of fear are more punitive than those with lower levels of fear (b = 0.069, p < .001). Table 2 also presents the effects of economic insecurity on fear of crime, and the findings indicate that three of the four economic anxiety measures are positively associated with this outcome. Specifically, those who think that the national economy is poor (b = 0.050, p < .001), that the national economy is worsening (b = 0.035, p < .05), and that one’s own economy is poor (b = 0.036, p < .001) report higher levels of fear of crime. In short, then, the findings suggest that insecurity is associated with both fear of crime and punitiveness, and there is evidence that fear of crime increases punitiveness as well.
Direct effects of economic anxieties on support for increased punishment.
DV: dependent variable; SE: standard error.
Note: N = 14,705. Unstandardized regression coefficients presented. Country dummy variables not displayed.
*p < .05, **p < .01, ***p < .001.
The hypothesis that economic insecurity will have an indirect effect on punitiveness through fear of crime was estimated using bootstrapping, and results from these analyses are displayed in Table 3. The confidence intervals that do not contain 0 indicate significant effects. Both poor national economy and poor personal economy have significant indirect effects on punitiveness through fear of crime. Those who think that the national economy is poor hold more punitive attitudes, though a poor personal economic situation is negatively associated with punitiveness. At the same time, both measures of economic anxiety are positively associated with fear of crime, and fear of crime is, in turn, positively associated with punitiveness. These direct and indirect relationships are presented in Figures 1 and 2.
Indirect effect of economic anxieties on support for increased punishment.
MV: mediating variable; DV: dependent variable; SE: standard error.
Note: N = 14,705. Unstandardized bootstrap estimates and 95% confidence intervals presented. Control variables not displayed.
*p < .05.
Direct and indirect effects of poor national economy on support for increased punishment. Note: Unstandardized regression coefficients are presented. The covariates are not shown. *p < 0.05.
Direct and indirect effects of poor personal economy on support for increased punishment. Note: Unstandardized regression coefficients are presented. The covariates are not shown. *p < 0.05.
Sensitivity analyses
Additional analyses were conducted to explore the robustness of these findings. The same models were estimated using ordinary least-squares (OLS) regression and logistic regression, allowing for separate estimates of the relationship between economic insecurity and punitiveness as it operates through fear of crime. Specifically, OLS regression was used to determine the influence of economic insecurity on fear of crime, and logistic regression was used to examine the effects of economic insecurity on punitiveness using models that both included and excluded fear of crime. The results of this stepwise approach were approximately the same as the results found using GSEM, with two exceptions. First, when estimated separately, there is no significant effect of poor personal economy on punitive attitudes either directly or indirectly through fear of crime, and second, the effect of worsening national economy on punitiveness is significant. Overall, multiple analyses confirm that the perception of a poor national and personal economy exerts direct and indirect effects on punitive attitudes.
Discussion and conclusion
There are strong theoretical rationales connecting economic anxieties and fear of crime to punitiveness among members of the public (Garland, 2001; Wacquant, 2010), and some prior research has suggested that these two factors may be independent sources of punitive attitudes (e.g. Costelloe et al., 2009; Dowler, 2003; Ousey and Unnever, 2012; Unnever and Cullen, 2010). This research contributes to this line of inquiry by examining fear of crime as a potential mediator in the relationship between economic anxieties and punitiveness. Additionally, this study also extends this literature to include Latin America and explores whether the recent movement toward punitive policies might be explained by public sentiments and fears. Indeed, analysis of survey data collected at a time of reduced economic growth and increased economic inequality in countries with popular mano dura criminal justice policies can provide some insights into the influence of public opinion on more punitive policies (Bértola and Ocampo, 2012; Grugel and Riggirozzi, 2012; Holland, 2013; Krause, 2014; Nivette, 2016; Reyes and Sawyer, 2016; Swanson, 2013; Wolf, 2017).
The results from the study support and extend prior research on the social sources of punitiveness, and several conclusions may be drawn from the findings. First, the perception that the national economy is poor or worsening and that one’s personal economy is poor are positively and directly related to fear of crime. This is consistent with prior research that shows those who are less economically secure and those who think there are poor societal economic conditions are more likely to fear crime (Hale, 1996; McKee and Milner, 2000; Panatazis, 2000; Rader et al., 2012), confirming a link between economic anxieties and fear of crime. Second, fear of crime has a direct effect on punitiveness, as those who report higher levels of fear of crime are more likely to support harsher punishments. This finding is also consistent with prior research, showing that those more fearful of crime are also more punitive (Costelloe et al., 2009; Dowler, 2003; Langworthy and Whitehead, 1986).
A third key conclusion from this study is that economic anxieties have direct effects on punitive attitudes, though the direction of this relationship depends upon the type of economic anxiety considered. Those who think that the national economy is poor are more punitive, while those who believe that their personal economy is poor are less punitive. This negative relationship, while counterintuitive, may be due to the fact that those who feel anxious about their own situations may see themselves as potential targets of punitive policies, which disproportionately affect the poor (Wacquant, 2010). Reporting a similar finding using data from the US collected during the Great Recession, Lehmann and Pickett (2017) speculate that economic insecurity may reduce perceived state legitimacy (see also Wozniak, 2016) and may increase fellow-feeling and altruism, both of which could lead to reduced support for punitive measures and might even increase public support for social welfare programs instead (see also Unnever and Cullen, 2009).
Finally, these analyses also show that economic anxieties have a consistently positive indirect effect on punitiveness through fear of crime. Economic anxieties and fear of crime are factors that contribute separately to punitiveness, but a portion of the effect of economic anxiety operates through increasing fear of crime. These empirical findings about national economic anxieties support the theoretical notions of both Garland (2001) and Wacquant (2010) that economic insecurity breeds fear of crime and, subsequently, influences punitiveness. In his discussion of middle-class reactions to crime, Garland (2001) suggests that, in recent decades, some have responded to the deterioration of the national economic condition through a hardened view of crime and criminal actors. However, at the other end of the spectrum, they respond to economic crises by considering them a “fact of life” and adapting with “measured stoicism” (Garland, 2001: 156). In other words, these individuals may not express a desire for increased punishments. This theoretical rationale may explain why Latin Americans’ concerns about the national economy increase punitiveness but measures of personal economic insecurity decrease it. Perhaps, in the eyes of the Latin American public, a deteriorating national economic system is more indicative of diffuse social anxieties brought by late modernity than personal economic standing, the latter of which may promote an emotional connection and identification with potential targets of social control. Thus, rather than continuing to study these two predictors of punitive attitudes as independent mechanisms, future researchers should examine the complex relationship that may exist between these potential sources of punitiveness.
This study is not without limitations. The dependent variable, support for increased punishment represents only one type of punitiveness. There may be a wide range of indicators of the overarching concept of punitiveness, and this study is only able to explain one of them. Although this study does not show an effect of economic anxiety of all types of punitiveness, it does provide insight into one example of it—support for increased punishments.
Similarly, the fear of crime measure only encompasses one aspect of fear of crime: perceived neighborhood safety. Feeling safe in one’s neighborhood is only one type of fear, and it is possible that some people may feel safe in their neighborhood, but fear crime in a more diffuse way. This is important to note because how fear of crime is measured has been shown to impact the strength of a study’s findings (Collins, 2016). Despite the limited scope of the fear of crime measure, these results provide insight into the effects of economic anxiety on punitiveness as they operate through one specific type of fear of crime, and the mechanisms present here may also extend into broader fears, or perhaps operate differently depending on the type of fear. In addition, Dammert and Tobar (2018) use this same measure from the AmericasBarometer survey to assess the state of fear of crime in Latin America noting that few countries have developed victimization surveys that can be used to provide a clear picture of this issue.
Lastly, the economic anxiety measures may be imprecise. The measures used in this study ask respondents about their perceptions of current personal and national economic situations as opposed to focusing on worries about the future. It may be the case that individuals assess their current economic situations as poor or worsening but are optimistic about the future. When interpreting these results, it should be considered that these findings are based on anxieties that may have been present at the time of the interview, and not anxieties that may manifest in the future.
Despite these limitations, this study provides useful insights about the nature of the relationship between economic insecurity, fear of crime, and punitive attitudes. Prior research has studied economic anxiety and fear of crime as two separate sources of support for punitive policies. The current study reveals that these two concepts work in tandem to produce increases in punitiveness, making explaining punitive attitudes more complex than previous studies imply. Without finding ways to increase individuals’ economic security and decrease their fears, support for punitive legislation will continue, even if policy moves away from harsher penalties toward more rehabilitative policies.
In summary, economic anxieties about the state of the national economy predict punitive attitudes, and this relationship partially operates through fear of crime. Punitive sentiments in Latin America are connected to both economic and crime-related anxieties, and these two factors work in conjunction with one another. However, future research is needed to better understand what social and cultural conditions might produce this relationship within the Latin American context. Further, prior research has demonstrated that public opinion can influence punitive crime control policies in an American and Western European context (Enns, 2014; Roberts et al., 2002), indicating a need for more research to understand how sources of punitiveness operate in Latin America and other international contexts.
Footnotes
Acknowledgments
The authors thank the Latin American Public Opinion Project (LAPOP) and its major supporters (the United States Agency for International Development, the Inter-American Development Bank, and Vanderbilt University) for making the data available.