Abstract
I extend Sutton’s political economy of punishment thesis by including a composite measure of legal and criminal justice practices that are hypothesized to be related to incarceration rates. I test the association between a punishment regime–type variable composed of the nature of prosecutorial discretion, judicial power, and the nature of lay participation with incarceration rates in 15 Western countries between 1960 and 2010. The effects of the punishment regime variable are tested in several time series cross-section models that correct for autocorrelation and heteroscedasticity and include the effects of the leading political–economic correlates of punishment and homicide rates. The results show consistent significant findings for the punishment regime–type variable, union power, and corporatist bargaining, with mixed support for political party power, inflation, and unemployment rates. These findings provide evidence for the existence of multiple cultures of Western punishment. In the end, I suggest that multidisciplinary cross-national partnerships incorporating researchers and governmental entities are needed to develop large comparative criminal justice databases to enhance understanding of comparative crime and justice.
Keywords
The second half of the 20th century has been described as an era of mass incarceration (Garland, 2001), mass probation (Phelps, 2013), and mass social control (DeMichele, in press). These terms refer to the peculiar punitive shift starting in the mid-1970s in which several countries—but, most dramatically, the United States and other common law countries—became more reliant upon criminal justice mechanisms to respond to social marginality (Beckett & Western, 2001; Garland, 2001; Simon, 2007). Garland’s “culture of control” thesis has framed explanations of comparative punishment for some time. But more recently, scholars have questioned whether there is one “culture of control” or if it is more appropriate to investigate “multiple cultures of control” (DeMichele, 2013; Lacey, 2008; Nelken, 2009). In this article, I build on Sutton’s political–economic model of punishment by including a legal institutional measure combining prosecutorial discretion, lay participation, and trial structure with a pooled data set of 15 Western countries. 1
Despite the massive growth in corrections populations, there is little research explaining the differences in the incarceration rates among Western countries. For instance, what accounts for the United States having between 4 and 8 times the average incarceration rate as Germany and the Netherlands, respectively, between 1960 and 2010? And, further, why does Germany have nearly twice the average amount of incarceration as the Netherlands over this same period (Lappi-Seppälä, 2011)? The leading scholarship on punishment focuses on differences in political economics as countries with strong left-oriented political parties (e.g., Social Democrats), high union membership (Greenberg, 1999; Jacobs & Kleban, 2003; Lappi-Seppälä, 2011; Sutton, 2004, 2013), and strong welfare systems are associated with less incarceration (Downes & Hansen, 2006; Garland, 1985). I contribute to this scholarship by suggesting that countries can be grouped into three punishment regime types that are rooted in different legal cultures (i.e., substantive, formal, and mixed). The common law countries (i.e., Australia, Canada, New Zealand, United Kingdom, and United States) consist of a populist regime, continental Europe (i.e., Austria, Belgium, France, Italy, Germany, and the Netherlands) fits into a bureaucratic regime, and Scandinavia 2 (i.e., Denmark, Finland, Norway, and Sweden) is characterized as a collective regime.
The theoretical approach adapts a Weberian typological method popularized by Esping-Anderson’s (1990) finding of three worlds (or regimes) of welfare capitalism supported by specific political–economic frameworks. His welfare state regime types provided comparative researchers with a typological method and theory to explain different approaches to social policies. Later, Cavadino and Dignan (2006) grouped countries according to political–economic structures using Esping-Anderson’s welfare state regime–type variables to explain incarceration rate variation. They found higher levels of neoliberalism (i.e., common law countries) are associated with a greater preference for exclusionary policy choices (e.g., low welfare and high incarceration), whereas other types of political–economic preferences support inclusive policies (e.g., social democracy and corporatism). The regime-type approach builds on Lacey’s (2008) recognition that historically embedded legal institutions structure the everyday decision-making powers of judges, prosecutors, police, and others by providing the bureaucratic definitions to the roles, rules, and incentive structures for legal actors.
In this article, I use time-series cross-section (TSCS) regression models of 15 Western countries from 1960 to 2010 to examine the plausibility of a regime-type explanation of punishment. First, I review prior comparative research explaining incarceration rate variation to demonstrate the importance of political–economic structures and build mostly from Sutton’s (2004) political–economic model. Second, I sketch out the punishment regime–type explanation that I have presented previously using descriptive comparative data (DeMichele, 2013). Next, I describe the data and analytic strategy before presenting the findings. In the end, I suggest that a regime-type approach of punishment is a potential framework to study how legal families (e.g., common and civil) are enacted by broader social structures (e.g., political economics) that create organizational contexts to supply the rules for how practitioners are to carry out their jobs (e.g., arrest procedures, prosecutorial and judicial discretion, and correctional guard training requirements). These organizational rules become institutionalized (Scott, 2001) and determine the decision-making capacity of criminal justice and legal practitioners who have consequences on the type (Whitman, 2003) and amount of punishment (DeMichele, 2013; Lacey, 2008; Nelken, 2009; Savelsberg, 1992).
Prior Research: The Politics of Inequality and Punishment
Explanations of the social structuring of punishment focus on the structural preconditions that frame policy choices used to organize class relations (Lappi-Seppälä, 2011). Punishment is linked with broader social arrangements that shape the institutional bases of persistent inequality (Western & Beckett, 1999) as the poor, uneducated, and weak are disproportionately incarcerated (Western, 2006). Rusche and Kirchheimer (1939) provided one of the first comparative political–economic models of punishment. They demonstrated significant correlations between the nature of the labor market (i.e., the value of laborers) and the type of punishment (e.g., moving from corporeal to incapacitation). Subsequently, I briefly sketch more recent research advancing a political–economic framework to explain incarceration rates.
Political Party Power
Political party power is essential to explanations of punishment. Political parties are the central mechanism of interest formation and they develop support for specific causes, groups, and policies that reflect their ideological orientation. Sutton (2000) found right-oriented parties to be a significant predictor of punishment with a sample of five common law countries and mixed support for Swank’s (2002) measure of the proportion of total cabinet seats held by left political parties related to lower incarceration rates. There is little research on the effects of political party power and comparative incarceration rates, but U.S. research (e.g., Yates & Fording, 2005; Jacobs & Jackson, 2010) and Sutton’s (2000, 2004) findings suggest a negative impact from working class or left political party strength (e.g., Social Democracy).
Union Strength
There is little known about how union strength influences incarceration rates. Structurally, it is reasonable to suspect that union strength will hold incarceration rates down as they reflect political, cultural, and institutional arrangements favoring the working classes. However, criminologists have rarely studied the relationship between unions and punishment. Western and Rosenfeld (2011) provide a compelling argument for how unions fit within the moral economy, and there are few governmental policies more influential to inequality than the nature of punishment. Sutton (2004) was the first criminologist to include measures of union strength in tests of incarceration rates and found a negative relationship. He argued that (Sutton, 2004, p. 177) unions are “political actors inimically opposed to the logic of the market … and discourage policies that are unfriendly to their constituents.” Union strength is related to a host of social measures improving working class economics including improved wages (Boreham, Hall, & Leet, 1996), lower inequality (Kenworthy, 2002; Western & Rosenfeld, 2011), and lower unemployment (Compston, 1997), and it is likely that union strength will be related to less reliance on incarceration.
Welfare
Beckett and Western (2001) in the United States and Cavadino and Dignan (2006) with a comparative sample have demonstrated how welfare and punishment interact to form an overall social policy regime. These studies suggest that some countries rely on inclusive policies and treatment for marginalized groups, whereas other societies develop combined social crime policies that are exclusionary in nature. Downes and Hansen (2006) found a negative correlation between welfare expenditures as a percentage of gross domestic product (GDP) and incarceration rates with a cross-sectional sample of 18 Organization for Economic Cooperation and Development countries. Sutton (2000) found a significant negative relationship with total welfare expenditures using five common law countries from 1955 to 1985, and he found limited support with a measure of unemployment expenditures for 15 countries from 1950 to 1990 (Sutton, 2004).
Unemployment
Traditional economic explanations of punishment have focused on the relationship between unemployment and punishment (Chiricos & DeLeone, 1992). Several criminologists have found significant positive relationships between unemployment rates and prison populations in Australia (Braitwaite, 1980), Canada (Greenberg, 1977), England and Wales (Sabol, 1989), Italy (Melossi, 1985), and the United States (Jankovic, 1977). Research does not provide a conclusive answer regarding unemployment with Parker and Horowitz (1986) summarizing previous studies as providing little evidence to support any relationship, but Inverarity and Grattet’s (1989) summary reaching the opposite conclusion that unemployment rates have a direct inverse relationship with incarceration rates, net of crime rates. In comparative studies, Jacobs and Kleban (2003) and Sutton (2004) find mixed support for relationships between unemployment and incarceration rates. It is likely that the influence of unemployment is washed out when other political–economic measures are included in models.
Economy
Previous attempts to model incarceration rates find mixed support for economic variables. Economic growth, inflation, and wage negotiation (i.e., corporatism) are interrelated, but there are few comparative studies considering how they are related to punishment. It is generally accepted that as GDP rises, unemployment will decrease (Okun, 1962), but others (Moosa, 1997) have found the GDP-employment relationship more nuanced. GDP is expected to be negatively related to incarceration due to the potential to decrease unemployment and affectively improve the lives of wage earners and decrease income inequality. Previous research finds a less clear-cut relationship. Jacobs and Kleban (2003) found GDP to be a predictor of increased incarceration, and Sutton (2004) found little effect for measures of economic growth, and when GDP was significant, it had a negative association with incarceration. GDP has received little investigation within the comparative punishment literature, but I expect a negative effect on incarceration rates, given the relationship with unemployment.
Inflation and corporatism are two additional economic variables that have found support in comparative punishment research (Cavadino & Dignan, 2006; Lappi-Seppälä, 2011; Sutton, 2004). Inflation is a measure of the relative price of goods and services within an economy that adjusts over time, and it is a by-product of economic growth that can have both positive and negative effects on the economy. Contrary to classical political–economic models of punishment (Rusche & Kirchheimer, 1939), Sutton (2004, 2013) found inflation drove punishment downward because he theorized that inflation helps maintain the real incomes of most workers and pensioners during economic downturns.
Corporatism is a form of interest representation (Schmitter, 1974) that is an institutionalized pattern of economic policy formation (Lehmbruch, 1979). Comparative researchers find corporatism provides the structural frameworks by which economic policy is negotiated for the entire economy (Kenworthy, 2002), and it is essential to understanding inequality, stratification, and social policy, and it is likely to have a negative impact on incarceration rates (Cavadino & Dignan, 2006; Lappi-Seppälä, 2011).
Punishment Regime Types
In a previous article, I presented a full treatment of the three punishment regime types that incorporate several deeply embedded historical processes (DeMichele, 2013). Here, I offer a shorter description of the punishment regime–type logic. Figure 1 is a conceptual model of how the punishment regime types are hypothesized to influence incarceration rates. The model is meant only as a heuristic as such sociocultural processes are far more nuanced than can be demonstrated in a one-dimensional graphic.

Conceptual model of punishment regime types.
Culture is at the left of the model and provides the values and meanings that shape the preferences of social policies about how to treat the disadvantaged. Several sociologists of punishment have emphasized a cultural framework supporting different levels of punishment. This includes Garland’s (2001) argument that cultural changes in the United States and the United Kingdom fostered an unparalleled spike in incarceration during the latter part of the 20th century. Feeley and Simon’s (1992) “new penology” and Simon’s (2007) “governing through crime” provide additional support for the notion of changes to U.S. cultural meanings of crime, criminals, and crime control affecting the amount of punishment. These new cultural ethos became embedded throughout social policy to support a stricter response to poverty and inequality (Fording, Soss, & Schram, 2011).
The types of law (e.g., substantive and formal) and preferences for treating inequality (i.e., inclusive and exclusive) are enacted by specific forms of political, economic, and social systems as the research above speaks to regarding punishment (e.g., Jacobs & Kleban, 2003; Lacey, 2008; Sutton, 2004). These structures are carried out by formal criminal justice and legal organizations—such as the courts, law enforcement agencies, and prisons—that provide the role definitions and rules that practitioners must follow as they do their jobs. This perspective builds on a longline of sociological scholarship conceptualizing organizations (e.g., bureaucracies) as small social structures that have consequences on employee behavior and personality (e.g., Merton, 1957). Criminal justice practitioners are embedded within rule-bound contexts in which their behavior must reflect the official definitions of the role they occupy. Simply, all of the countries in this analysis have police who arrest; prosecutors who bring charges; and judges who sentence, lay participation in hearings, and correctional facilities to incarcerate people, but the way these activities are completed varies considerably between countries. 3
The punishment regime–type explanation draws upon three underlying aspects of the law: discretion, styles of courtroom interaction, and insulation from the public. Table 1 provides a description of the punishment regime types to identify aspects of the three legal families (i.e., common, civil, and Scandinavian). Legal scholars have discussed the differences in these legal approaches at length (see Glendon, Gordon, & Osakwe, 1982; Zweigert & Kotz, 1998), but there is little discussion of how legal types influence the amount of incarceration or other criminal justice outcomes. 4 In Weber’s (1967) Sociology of Law, he divided Western law into substantive and formal types of rationality to highlight the differences in how the law is developed (see Damaska, 1975; Kalberg, 1980; Kronman, 1983; Trubek, 1986). Common law incorporates a substantive form of rationality that favors external values in the law to provide more flexibility to change the law to reflect social, moral, and political values through case-based approaches. Civil law, on the other hand, has a formal type of rationality that relies on strict adherence to code-based laws that are developed by legislative bodies (Merryman, 1985; Weber, 1967).
Punishment Regime Type Characteristics.
The differences between substantive and formal rationality inform the amount of discretion afforded to legal and criminal justice actors (Savelsberg, 1992). There is more discretion in decision making for common law practitioners, and Simon (2007) demonstrated that prosecutorial discretion played a large role in U.S. incarceration growth. Given that civil and Scandinavian legal rules are stricter, practitioners are seen as bureaucrats with little ability to influence the law and legal outcomes (Lappi‐Seppälä, 2011). In Whitman’s (2003) historical treatise of U.S., French, and German criminal justice systems, he identified the importance of how restricted discretion provided criminal justice practitioners with protection from the public’s desire for harsh punishments (Savelsberg, 1992). In countries that do not insulate criminal justice practitioners, laws have a greater chance of being quickly developed to satisfy the public’s punitive desire. 5
The legal systems have different ways to incorporate lay participation in which civil and Scandinavian courts use lay judges, whereas common law courts use juries (Langbein, 2005). Upon first glance, there may appear to be little difference between these approaches, but they suggest a different conceptualization about lay input and professionalism of court actors (Alexander, 2002; Mattei, 2002). Lay panels are typically elected and sit for a period of years in which they are able to interact with professional judges and gain some legal knowledge from the informal socialization (Hans, 2003; Leib, 2008). Juries, on the other hand, sit for a single case, have no legal training, and are part of the reason common law courts have such extensive procedural rules (Langbein, 2005). 6
A third mechanism supporting the punishment regime–type explanation is the adversarial nature of the law. Adversarial legalism includes strong litigant activism, lawyer domination, and dispute resolution as a specific way of thinking about the law, of making legal rules, and ways of deciding cases (Kagan, 2001). Adversarial legalism is a battle against the state and defense in which the judge acts as the referee, inquisitorial systems incorporate the judge more as the conductor of the investigation, and the trial in which he or she is to determine which legal code to apply (Hodgson, 2002; Kagan, 2001). To European audiences, the adversarial approach found in common law countries can appear unfair, with one French Minister of Justice, Elisabeth Guigou, stating: The adversarial system of justice is by nature unfair and unjust. It favors the strong over the weak. It accentuates social and cultural differences, favoring the rich who are able to engage and pay for the services of one or more lawyers. Our own system [French] is better, both in terms of efficiency and the rights of the individual. (found in Hodgson, 2002, p. 231)
The level of discretion and adversarialism shape the nature of practitioner insulation from public opinion or what is often referred to as popular punitivism (Bottoms, 1995; Pratt & Marie, 2005). In some ways, popular punitivism is inherent to the common law legal framework as extralegal value considerations are embedded in the historical foundation of substantive rational legal types (see Langbein, 2005; Savelsberg, 1992; Weber, 1967). The punishment regime–type framework is intended as a way to measure the effect the differences in legal and criminal justice practices have on the variation in criminal justice outcomes. Simply, it might not be what practitioners do that is so different (e.g., arresting, prosecuting, and incarcerating) but rather how practitioners carry out their jobs that matters for criminal justice outcomes (Lacey, 2008; Lappi‐Seppälä, 2008; Nelken, 2009).
Data and Method
I developed a time-series cross-sectional data set of 15 Western countries from 1970 to 2010, including Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, the Netherlands, New Zealand, Norway, Sweden, the United Kingdom, and the United States. The procedures used here follow comparative sociological work (Brady, Beckfield, & Seeleib-Kaiser, 2005; Hicks, 1999; Huber & Stephens, 2001) and follow Beck and Katz’s (1995) suggestion to use ordinary least squares with panel-corrected standard errors (OLS-PCSE), a first-order correction for panel autocorrelation (PSAR1), and correction for panel heteroscedasticity. The analytic approach follows previous research and treats the data set of 765 country years as cases in a pooled time series with 51 years and 15 countries. Table 2 provides the names, measurement, and sources for the dependent variable and each of the independent variables.
Descriptions and Sources of Variables.
Note. GDPPP = gross domestic product at purchasing price parity.
Dependent Variable
The dependent variable is the proportion of adults incarcerated per 100,000. 7 This is a standard metric of the amount of incarceration in a society that takes into consideration the overall population, and it is readily available in a time-series format. The incarceration rates were gathered from Sutton’s (2004) data set that included the years 1970–1990, and individual country reports were used to include the additional years from 1991 to 2010.
Independent Variables
The central variable of interest is the legal institutional variable measuring one of the criminal justice policy regimes—the populist regime. This variable is composed of measures of the use of juries (0 = no, 1 = yes), the amount of prosecutorial discretion (0 = none, .5 = limited, 1 = high), and trial type (0 = unitary trial structure, 1 = separate guilt and sentencing phases). The summed scores range from 0 to 3. Higher scores represent a form of punishment associated with higher incarceration rates. This variable is nearly time invariant (Plumper & Troeger, 2007), within country (σ = .11, σ2 = .01) 8 because these criminal justice and legal practices have changed little over time within countries but provide larger between-country variance (σ = 1.21, σ2 = 1.46).
Analysis
The analysis includes 15 countries over 51 years to fit several pooled time-series models. This data structure creates potential for heteroscedasticity and correlation of errors. OLS regression assumes that error structures are spherical—homoscedastic and independent. TSCS data must be corrected for these potential problems. The countries included in the analysis were selected because of their similarities in economic, political, social, and other fields. These similarities have the potential to autocorrelation between and within countries over time. These assumptions are more than a mere nuisance—error structures are important to consider in TSCS (Beck, 2001). OLS without corrections will lead to inefficient estimation and, most problematically, reduced standard errors that foster inflated significance tests (Type I error). It is difficult for comparative scholars to assume that cross-national data will be arranged such that error variances are constant and that all error processes are independent. To adjust for these panel error assumptions, Parks (1967) developed an approach using generalized least squares (GLS) regression that adjusted for heteroscedasticity and autocorrelation. Beck and Katz (1995) proposed an alternative estimator than the Parks–Kmenta estimator because it was found to significantly underestimate variability in finite samples. Beck and Katz suggested utilizing panel-corrected standard errors and use OLS estimation.
I estimated numerous regression models testing the adequacy of different estimators to address the assumptions mentioned previously. 9 This presentation is confined to the OLS-PCSE with panel-specific AR1 adjustments controlling for panel-level heteroscedasticity. 10
The basic TSCS regression model is:
The dependent variable yit
is estimated with a β0 constant term and a set of time-varying covariates (β1
xit
+ β
k
xkit
). The regime-type variable is a (nearly) time-invariant measure (γzi
), ui
is the unobserved country-specific effects, and ∊
it
is the overall residual. Log transformations were used for incarceration rates, homicide rates, unemployment rates, union density, and welfare spending to adjust for outliers. The political party variables are smoothed using a moving average from t − 2 to t − 1, which follows convention in comparative political research (e.g., Hicks and Swank, 1992). I follow Sutton (2004) to express GDP and inflation as measures of proportion change (ln(Xt/Xt − 1)) to estimate economic growth (GDP) and change in the relative worth of money (inflation). The full estimation model includes lagged, natural log, and moving average (ma) indicators with signs indicative of expected relationships.
Findings
In Figure 2, the prison rates are plotted by regime type and countries over the entire period. The top left panel displays the incarceration rates per 100,000 for the populist regime type

Prison rates per 100,000 by punishment regime type, 1970–2004.
The bureaucratic punishment regime type (top right panel) has a more tempered incarceration rate that rarely reaches 100 per 100,000 and maintains a mean incarceration rate of 76 per 100,000. The regime-type approach does not rule out within regime variation, and the Netherlands presents an interesting case as it had the lowest incarceration rate with <20 per 100,000 during the 1970s and the highest incarceration rate (>130 per 100,000) by 2004 for the bureaucratic regime type. The bottom left panel displays the time series for the collective regime type that has the lowest average incarceration rates
Table 3 provides the regression results for four regression models. I start the estimation with a baseline model including only the populist regime variable and homicide rates to estimate the effects of the regime variable independent of the structural variables. Model 1 shows that the populist regime variable has a significant positive effect on incarceration rates, with no effects found for homicide rates. On average, a 1-unit change in the populist type measure is associated with about a 12% increased logged incarceration rates per 100,000 ((eb − 1) × 100). Model 2 incorporates the political parties, union density, and welfare spending variables. Political parties appear to have limited impact on incarceration rates, with left parties having a minor negative impact (p < .10), and welfare spending unrelated to incarceration rates. Union density has a significant and strong effect on incarceration rates with a 1% increase in union density, on average, related to a 0.28% reduction in logged incarceration rates.
Panel Corrected Time-Series Cross-Section Regression Models of Log Incarceration Rates, 1960–2010.
Note. GDPPP = gross domestic product at purchasing price parity.
† p < .10, *p < .05, **p < .01, ***p < .001.
In Model 3, I drop the political variables and enter the economic variables. The populist regime variable changes little from the previous two models and remains positive and significant, but homicide rates have a positive significant effect on incarceration rates. Inflation and unemployment are in the hypothesized direction, negative and positive, respectively, and significant at the lower p < .10 confidence level. As the power of money weakens (i.e., inflation increases), the incarceration rates on average increase, and as the rate of unemployment grows (on average), incarceration rates also increase. In Model 3, linear and quadratic measures of corporatism are significant in the expected directions. The linear measure is significant and negative and the quadratic term is significant and positive, which supports a U-shaped curvilinear relationship. In countries with the highest measures of corporatism (i.e., Scandinavian), incarceration rates trend downward, and countries with the lowest corporatist scores incarceration rates are held lower (Sutton, 2004).
Model 4 is the full model including all variables and there are few major changes from the previous models. The effects of homicide rates are stronger in Model 4 than the previous estimates and remain positive and significant (p < .05). The populist regime variable decreases slightly but remains positive and significant (p < .001) and is indicative in an 11% increase in average incarceration rates. Political parties and welfare spending have no effect, with insignificant estimates that are close to zero. The findings do provide additional support for understanding the relationship between unionization and punishment with strong, significant effects (p < .001). The effects of the economic variables change little with inflation remaining negative and significant at the p < .10 level, and GDP and unemployment rates have no effect. Both the linear and quadratic corporatist measures weaken slightly, but they remain significant in the expected directions to provide support for a U-shaped curvilinear effect. These findings suggest the plausibility of a regime-type framework and support previous political–economic explanations of incarceration. The regression models show that the populist punishment variable and political and economic institutions that support the working classes are strongly associated with lower incarceration rates.
Conclusion
In this article, I have contributed to comparative criminal justice research by incorporating a measure of the differences between substantive, formal, and mixed legal types. This research complements the dominant political–economic explanations of incarceration variation (e.g., Greenberg, 1999; Jacobs & Kleban, 2003; Sutton, 2004). But this approach also suggests the rich theoretical and empirical opportunities available by systematically uncovering specific practical differences between legal and criminal justice systems (DeMichele, 2013; Lacey, 2008; Nelken, 2009). Substantive rational types of law are designed inherently to allow the inclusion of social, economic, and ethical criteria into legal rules and decision making (Savelsberg, 1992). This, essentially, means that the common law is constructed to support popular punitivism, which has taken hold with the greatest force in the United States. The United States may provide a perfect constellation of characteristics that drive mass social control (DeMichele, in press) that differentiates it from other common law countries: weak protection for working classes, limited collective wage bargaining, intense racial divisions, judicial and prosecutorial elections, and neoliberal political views.
Comparative crime and justice research should continue to investigate how political–economic institutions interact with criminal justice and legal practices. The regression models provide key findings about the association between political–economic institutions and incarceration. Left political party power has little direct impact on incarceration rates (at p < .10), and any impact found in Model 2 (Table 3) is negated when the corporatist variables are included in the analysis. In unreported regression models, the political party variables were included without union density, welfare spending, and corporatism, and left party power maintained a significant negative relationship at the lower confidence interval and right party power is insignificant. This suggests that the lack of left party power effect is not a matter of multicollinearity with related political–economic institutions. Additional comparative research is needed to disentangle the complex relationship between political parties and incarceration.
The direct effect of working class protection appears to be a more important set of social protections from high incarceration. The effects for union density and corporatist bargaining provide important findings suggesting a strong link between modern labor relations and punishment. Union power and strong corporatist bargaining reflect institutionalized norms over inequality as part of the moral economy and provide a set of practices that define equitable economic distribution (Elster, 1989; Thompson, 1971; Western & Rosenfeld, 2011). Unions and corporatism moderate inequality through higher wages and social policies that support universal health care and the creation of Medicare (Derickson, 1994), general support for redistributive welfare policies (Kenworthy, 2002), and it appears by holding incarceration downward (Sutton, 2004). Of course, teasing out exactly how unions and corporatist bargaining impact incarceration rates is difficult, but it is plausible that the indirect effects of increasing wages and reducing wage gaps have a positive impact. More qualitatively, however, a contrast between Christie’s (2000) observation that Norwegian correctional unions were a driving force against increasing the use of prisons, and Page’s (2011) observation that California correctional unions successfully lobbied for increased sentences provides some insight.
This research comes with some limitations and weaknesses. First, an obvious critique would focus on the specific indicators for the populist regime type. Juries, for instance, are seldom used, whereas plea bargaining in most common law countries and summary proceedings in many civil law countries account for the bulk of convictions. However, the point to the punishment regime–type measures is to uncover legal and criminal justice practices that are institutionalized features of different types of legal families and to determine whether they are related to criminal justice outcomes. The rules defining prosecutorial discretion among the different legal families result in very different criminal investigations, negotiations, and likely convictions. These rules become the scripts or organizing logics for practitioners in different parts of the world. Second, focusing on incarceration rates leaves out considering a host of other outcomes including pretrial or remand custody, probation and parole, admissions, and length of stay. Incarceration rates, admittedly, are a rather crude approximation of the relative stock of incarceration, but they tell us little about broader punishment patterns.
My response to both of these criticisms focuses on the nature of data availability and the need for cross-national partnerships. Nelken (2009) mentioned the difficulty with comparative research because it is nearly impossible for one person to be an expert in more than one or two countries. I suggest that broad multidisciplinary research and governmental collaborations are needed to develop large data repositories that provide a wealth of historical and comparative data. The Bureau of Justice Statistics made an initial attempt at this during the late 1990s (Farrington, Langan, & Tonry, 2004), and new initiatives partnering BJS with other national research bodies (e.g., Scandinavian Research Council for Criminology) could provide the traction needed to disentangle comparative crime and punishment. This approach could contribute to theory and methods but also uncover evidence-based practices in one country that might fit in other national contexts.
This research responds to Lacey’s (2013, p. 349) recent call for comparative punishment to focus on institutional variation and draw “on historical and comparative methodologies” that incorporate “law, sociology, and political science.” I expanded Weber’s types of legal thinking to include a Scandinavian type of law and identified specific legal and criminal justice practices that differ across the regime types. Durkheim’s concept of a stability of punishment is implied as the descriptive data show that countries cluster on incarceration rates within allowable ranges used to maintain normative order throughout the second half of the 20th century. The effects of the punishment regime variable were tested in four TSCS models that adjusted for autocorrelation and heteroscedasticity. The regression analyses included several leading political–economic correlates of punishment and homicide rates to provide conservative tests of the regime variable. In addition to the strong findings for union density and corporatism, the analyses suggest that there are specific legal and criminal justice institutions that contribute to incarceration rates.
Footnotes
Appendix
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
