Although individuals and governments stand to benefit dramatically from crime reduction, actual rates of crime and imprisonment vary widely across industrialized democracies. For example, the USA has over two million people in prison, whereas Japan has significantly fewer prisoners (UNDP 2007; Lacey 2010). One general observation of the literature on imprisonment is that neither democratization nor economic prosperity alone is sufficient to achieve a low level of crime or incarceration (Wacquant 2009). Firstly, low levels of incarceration do not necessarily reflect low crime rates.  In many developing countries, they simply reflect a high degree of impunity. Higher levels of imprisonment may reflect the 1) number of criminalized activities, 2) incidence of crime, 3) length of incarceration, 4) rates of imprisonment, and 5) extent to which the state actually punishes criminals rather than letting some or most of them go free (Nadanovsky and Cunha-Cruz 2009; Wilkinson and Pickett 2010).

Secondly, incarceration is not only a measure of law enforcement, it is also a penal regime connected to the overall political economy of a country (Cavadino and Dignan 2006). Recently, for example, several scholars have linked the neo-liberal “variety of capitalism” (Hall and Soskice 2000) found in English-speaking countries settled by Anglo-Saxons to a harsher penal regime of mass incarceration than that found in the coordinated market economies of continental Europe, East Asia, and Scandinavia (Wacquant 2009; Lacey 2010).

Scandinavian countries in particular have attracted attention for having the lowest average rates of imprisonment among industrialized democracies. One line of argument is that they have fewer criminals, because their comprehensive “universal welfare states” (Esping-Andersen 1990) have explicitly aimed to achieve full (one hundred per cent) employment for the entire male and female labor force (Tilton 1992; Whyman 2003). Although these countries have not always succeeded in maintaining full employment (Pontusson 2005), it is certainly possible that state efforts to increase employment have reduced crime as studies have repeatedly found unemployment to correlate with the incidence of crime (Western and Pettit 2010).

Three other aspects of the universal welfare state may also help to curb crime. Firstly, comprehensive national health insurance reduces undertreatment of mental illnesses and significantly reduces medical costs for those with low income or for those who are unemployed. Lack of health insurance has been found to positively correlate with imprisonment rates (Selke and Andersson 2003). Secondly, the universal welfare state’s provision of extensive and free education appears to also reduce crime. Studies have shown that more educated citizens commit relatively fewer crimes, whereas young men with lower levels of completed schooling commit most crimes (Western 2007). Thirdly, the efforts of a universal welfare state to provide quality education, child care, and health insurance to poor children as well as health care and employment to their parents may play a powerful role in reducing child abuse, ignorance, and disempowerment, the very factors that often lead to crime as young men get older (Loury 2010).

While Scandinavian countries stand out for their universal welfare states and low levels of crime and punishment, it is also true that Scandinavian societies have historically been rather ethnically homogeneous. This may   have helped to keep crime rates low, as it may be easier to maintain a high level of social trust and mutual respect among like-minded peoples. In ethnically or racially divided societies, a caste system may develop where  “superior” groups resent or hate “inferior” groups, thus creating an environment of animosity, distrust, and exploitation (Bangura 2006). When racial segregation, inter-ethnic animosity, and economic stratification are combined with limited social mobility, it can be an incubator of crime and harsh penal regimes as in the USA and South Africa (Wacquant 2009;Wacquant 2010). For example, studies have found imprisonment rates within the USA to increase with the proportion of black residents (Selke and Andersson 2003) and the strength of right-wing political parties (Jacobs and Carmichael 2001).

Because economic systems are dependent on political decisions, it may be possible to trace a significant share of incarceration differences across societies to political institutions that provide (or fail to provide) space for the inclusion of the lower classes to organize and participate in political decisionmaking. Lacey (2010), for example, argues that proportional representation (PR) electoral systems make democracies more inclusive, leading to greater redistribution and social equality which, in turn, reduce crime. Similarly, Jacobs and Kleban (2003) find that the inclusive procedure of social corporatism also reduces incarceration rates, while Arend Lijphart finds “consensus” democracies that combine PR elections with corporatism put fewer people in prison than “majoritarian” democracies (Lijphart 1999, p. 298) with adversarial single-member district (SMD) election systems and fragmented, non-corporatist, interest group representation.

The literature reviewed in the previous section leads us to hypothesize that different types of democracy may be closely related to the incidence of crime in a society. In particular, one might expect more inclusive democracies to have the least crime, because they might also have the least number of dissatisfied individuals. This distinction leads us to a comparison of two different varieties of democracy as defined below.

Firstly, an “adversarial democracy” can be defined as a democracy in which competitive veto points, individualism, and clientelism are dominant features of the political system. As Mansbridge (1981, pp. 469-70) argues, “the fully developed adversary theory of representation rejects the assumption of an underlying common good…voters pursue their conflicting individual interests by making demands on the political system, while politicians, also pursuing their own interests, adopt policies that buy them votes.” This variety of democracy is relatively exclusive for three reasons. Firstly, the institutional structure incentivizes a combative approach to politics whereby influence is gained by simultaneously excluding one’s (perceived) opponents. Secondly, the political process is highly commodified, allowing those with the greatest ability to pay for lobbyists, access, advertising, and campaigns to have a distinct advantage over others. Thirdly, due to the highly competitive and costly nature of this system, many come to feel that political participation is useless because of the high likelihood of being vetoed at some point in the process. Therefore, much of the population gives up on participating in formal political processes which result in special and moneyed (i.e., elite) interests often winning out over general public interests.

A “social democracy,” by contrast, can be defined as a democracy that emphasizes proportionality, egalitarianism, and collective veto points. As Meyer (2007, p. 3) points out, “social democracy…insists that democracy and its associated charter of rights must be extended into the social and economic spheres as well. Specifically, the social order must meet higher standards of democracy by allowing for well-regulated participation, a legal claim to social security, a distribution of wealth and income that takes justice into account, and a democratic state, the regulative and distributive policies of which accord with all of these values.” Compared to an adversarial democracy, a social democracy is relatively inclusive for three reasons. Firstly, the institutional structure incentivizes cooperation, compromise, and inclusion rather than exclusion of one’s (perceived) opponents in political deliberation. Secondly, as an inclusive system, the political process is relatively de-commodified, making political influence less dependent on one’s ability to pay for lobbyists, access, and campaigns. Thirdly, due to the more accessible and cooperative nature of this system, people are more likely to feel that political participation is worthwhile because they might actually have some influence. More of the population participates in trying to influence the political system and, as a result, general public interests more often win out over special and moneyed interests in social democracies than in adversarial democracies.

While many institutional analyses highlight the role of “veto players” (Tsebelis 2002), the two definitions above incorporate the notion of “veto points” building on the work of McGann (2006), Birchfield and Crepaz (1998), and Crepaz (2001) who distinguish “competitive veto points” that  make policy changes difficult (such as presidentialism, constitutional rigidity, judicial review, and bicameralism) from ”collective veto points” which refer to power sharing and proportionality within a legislature. As McGann (2006, p. 180) points out,

From this perspective, “it is very easy to explain why supermajoritarian checks and balances increase inequality. Such institutions allow minority groups that lose from redistribution to veto it” (McGann 2006, p. 196).

Comparing these two varieties of democracy leads us to several hypotheses. Firstly, we might expect democracies characterized by adversarial institutions to have more criminals if these systems generate more inequality and, therefore, have more “losers” who decide to resort to crime out of desperation. Although “adversarial democracy” is perhaps best understood as a radial concept, it can be identified by the presence of multiple adversarial  political institutions. Among these institutions we will focus on two that are particularly important: uninominal elections and presidentialism.

The first relevant institution, the electoral system, takes on an adversarial character when it is uninominal (i.e., based on single-member districts) and uses a plurality selection rule as in the UK and many former British colonies  (including Canada and the USA) (Cole 1999; Finer 1975). Often referred to as a “first-pass-the-post” (FPTP) system, these electoral contests are  adversarial because only one candidate wins in each district while all of the other candidates lose (Norris 2004). The election itself, therefore, functions as a competitive veto point because all candidates (and the voters and the issues they represent) are vetoed at the time of election except for the one candidate who happens to gain a plurality of the votes. If, for example, the winning candidate gets 40% of the vote while three others respectively get 30%, 20%, and 10%, then 60% of the electorate will be unrepresented by a candidate or political party of their choosing. Not only does the election itself create many losers, but the policies that result may also take on the same character. Furthermore, in uninominal systems, electoral competition over time in each district typically centers around two candidates (Duverger 1954), thereby excluding all other possible choices. Yet, as Emerson (2007, p. 4) exclaims, “Democracy, I repeat, is for everybody. The democratic process, therefore, should not be a means by which some come to gain dominance over others (as in a majoritarian structure), but rather a process in which all (or nearly all) come to an accommodation…So any voting procedure must surely offer the voter a choice of more than two options!”

Compared to FPTP systems, an electoral system takes on a more inclusive character under high magnitude, multi-member districts using listbased proportional representation (PR). For example, if the district magnitude is ten and the winning party gets 40% of the vote while three others get 30%, 20%, and 10%, respectively, then 100% of the electorate will be represented by a candidate or political party of their choosing with the first party getting four seats and the remaining parties having respective seat allocations of three, two, and one. In other words, more parties, and hence more perspectives, can be included in a parliament using a proportional electoral system (Norris 2004). Moreover, compared to uninominal elections, PR has been linked to the election of a greater share of representatives from demographic groups typically under-represented in politics including women, young adults, working-class people, and ethnic minorities (Joshi 2012a, 2012b; Norris 2004, 2008). Thus, public policies are likely to incorporate the views and interests of more segments of the population than under an adversarial system because PR often brings a larger share of perspectives and interests into the legislative arena.

Precisely because these conditions facilitate cooperation and compromise, there may be a greater likelihood that public policies will benefit the general public, including the lower classes. This in turn should reduce inequality and crime, which leads us to our first hypothesis (H1) that there will be fewer criminals in countries with proportional representation elections than in countries with uninominal electoral systems.

A second institution that fosters adversarial politics is presidentialism. In presidential systems there is a division of power between the executive and legislative branches of government, creating an inherently adversarial relationship as these two branches of government are pit in competition against each other for influence and power (Shugart and Carey 1992). The adversarialism of presidentialism is often further exacerbated by the fact that the president, being independently elected, feels a sense of entitlement to rule on his/her own without the consent of the national assembly. Hence, the president and the assembly typically have a combative relationship whereby  presidents often resort to a variety of means to augment their power and bypass the legislature such as through proclamations, executive orders, decrees, memoranda, national-security directives, and legislative signing statements (Nelson 2008, p. 155). Scholars have also found presidents more likely to instigate wars or states of emergency to expand presidential power as the incumbents become addicted to power (Gulieyev 2009; Linz 1994).

As Linz (1994) famously argued, presidentialism generates an inherent problem of “dual legitimacy.” Because both the assembly and the president are separately elected, it is not clear in whom the voters have vested greater power in the case of a conflict arising between the two. Perhaps due to this problem, presidential systems have been found to collapse into dictatorship more often than parliamentary democracies. Linz (1994) discovered a 77% breakdown rate for presidential democracies compared to 33% for parliamentary democracies, while Stepan and Skach (1993) found pure parliamentary regimes to have a 61% survival rate compared to 20% for pure presidential regimes. In a parliament there is a clearer chain of delegation whereby voters select members of parliament who, in turn, choose a prime  minister and cabinet to govern.

Presidents are also elected in a uninominal process reflecting the adversarialism inherent in any single-member district election. A president may even come to power with only a plurality of 30% or 40% of the votes, but the system is “loser loses all” for defeated presidential candidates and the large number of voters they represent (Linz 1994, p. 14). Moreover, the winner-take-all character of presidential elections often result in a form of “delegated democracy” whereby “candidates compete for a chance to rule virtually free of all constraints save those imposed by naked, noninstitutionalized power relations. After the election, voters/delegators are expected to become a passive but cheering audience of what the president does…The president and his most trusted advisors are the alpha and the omega of politics” (O’Donnell 2009, p. 37). In many presidencies, cabinet members are also unelected officers chosen by the chief executive (Nelson 2008, p. 19). Meanwhile, even defenders of residentialism acknowledge that presidencies are associated with fewer multi-party coalitions and less programmatic parties (Cheibub 2007).

By contrast, parliamentary democracies are less adversarial when it comes to the relationship between the executive and the legislative functions of government. This is so because the same party or coalition controls both branches, and power sharing among multiple parties in an executive cabinet facilitates compromise and cooperation. Voters are also more likely to know beforehand the policy agenda of a potential government because parliamentary systems encourage partisan voting (Gerring and Thacker 2008). In other words, the focus in elections is more on collective entities (parties) as opposed to individuals (presidential candidates), leading to more  programmatic parties and a more programmatic focus of policy discussion (ibid.). From this discussion, we conjecture that the combative nature of presidentialism may lead to a smaller share of the population obtaining desirable policies than under parliamentarism because the competitive veto structure will allow defenders of the status quo more opportunities to veto proposals for progressive redistribution. This is likely to be less beneficial to those belonging to the lower classes who, out of frustration and desperation, may resort to crime. This brings us to our second hypothesis (H2) that there will be more criminals in countries with presidentialism than in countries with parliamentarism.

Our third and final hypothesis (H3) is that countries with highly inclusive social democracies will have fewer criminals than in adversarial democracies. As Friedman and Hochstetler (2002, p. 23) point out, in an “adversarial democracy,”

In other words, while adversarial democracies are based on the principles of competition and combat, thus producing winners and losers, social democracies are based more on the principle of inclusion and shared security via universal rights and entitlements with the goal of making everyone a winner.

There are two particular means by which we expect social democraciesto reduce crime: social equalization and risk-mitigation. The first factor is important because there is less demand for crime when all members of society have access to a decent basic living standard. When the gap in living standards is low, there may also be less feeling of relative status deprivation among those with low status (Wilkinson and Pickett 2010). Yet, even in prosperous societies, there may be high levels of poverty and disaffection if the gap between the haves and have-nots of material wealth and status is great. The quality of life and the life chances an individual experiences will  depend upon the opportunities she or he has access to within her/his society. When these opportunities are uneven and unequal, it can lead to social problems such as crime, drugs, divorce, unemployment, and mental illness. Researchers have consistently found strong correlations between higher levels of inequality and higher incidences of mental health problems, drug use, obesity, educational failure, teenage births, violence, imprisonment, and punishment (ibid.).

As mentioned above, social democracies specifically aim to increase social equality (Joshi and Navlakha 2010; Tilton 1992) while seeking to deepen democracy by extending political inclusion to the lower classes (Berman 2007). They also tend to pursue progressive taxation requiring greater social contribution from those with excess to guarantee equal  opportunities and social mobility to those born at the bottom of the wealth pyramid. This is particularly important, for children need to be able to access decent nutrition, childcare, education, and a safe environment free from criminal gangs, pollution, noise, and violence. As more children are entitled to these protections and security, a greater share of the population can gain the opportunity to study, acquire skills, work hard, and find gainful employment.

The second and related element through which social democracies can reduce crime is through risk-mitigation via social safety nets. Safety nets prevent people from facing the fallout of structural injustices. This is particularly important in cases such as an environmental catastrophe or exogenous economic shock where the victims are not at fault (Meyer 2007). Thus, even those members of the population who are employed and healthy know that they will be able to sustain their needs if they were to lose their jobs or fall ill due to factors beyond their control (Einhorn and Logue 2003). The presence of safety nets can hence potentially improve the overall physical, mental, and emotional security of the population.

Without social equality and risk-mitigation, however, a country may find a relatively large portion of its population in deprivation, despair, and with increased motivation to engage in criminal behavior. Compared to social democracies, this is more likely to be the case in adversarial democracies where political institutions exclude or minimize participation of the lower classes and, hence, retard the development of social safety nets and full employment policies. If the state provides insufficient welfare support to its citizens and does not actively promote full employment, its people (especially young men) will have more incentives to turn to gangs, drugs, or other criminal activities to prop up their identity and livelihood. Including groups at the bottom of the socio-economic hierarchy into democracy may, therefore, be especially pivotal for crime reduction because most imprisoned criminals come from lower strata of the population.

I will now proceed to test whether these two varieties of democracy systematically vary in their rates of imprisonment. First, I employ a structured, focused comparison of four Nordic social democracies (Denmark, Finland, Norway, and Sweden) and four English-speaking adversarial democracies (Australia, Canada, UK, and USA). I chose to compare these states for several reasons: Firstly, the four social democracies rank at the top of Meyer’s social democracy index (Meyer 2007, p. 213) as the most “highly inclusive social democracies” among twenty long-standing member countries of the Organization for Economic Cooperation and Development (OECD). Secondly, the four adversarial democracies rank towards the bottom of Meyer’s index. Three are classified as “less inclusive social democracies” (Australia, Canada, and UK) and one as an “exclusive  democracy” (USA). Thirdly, as OECD members, these countries have comparable data available and are all fairly rich and stable, allowing us to control for the impact of economic prosperity and political development, factors that otherwise vary greatly across countries.

As shown in table 1, the difference between these two varieties of democracy on rates of imprisonment is stunning. The Nordic social democracies average 75 prisoners per 100,000 people while the Anglo-American neo-liberal democracies average 274 prisoners, or almost four times as many. The USA, as the most exclusive and adversarial of the adversarial democracies, stands out with its exceptionally high rate of imprisonment at 738 prisoners per 100,000 people, but even if the USA is excluded as an outlier, the other adversarial democracies still have an incarceration rate that is roughly 60 per cent higher than that of the Nordic social democracies.

As mentioned above, there are fundamental differences in the political institutions of these two varieties of democracy. While highly inclusive social democracies use proportional representation (PR) elections, adversarial democracies use single-member districts (SMD). Adversarial democracies themselves can then be further divided into moderately adversarial democracies (those with parliaments) and deeply adversarial democracies (those with presidents). As shown in table 2, both proportional representation (PR) electoral systems and parliamentarism correlate with lower levels of imprisonment across these countries. Theoretically, we would expect a greater level of inclusion from PR electoral laws because they should provide more institutional space for the inclusion of the lower classes in political decision-making. Notably, PR systems grant lower-class political parties a greater chance of representation in national and sub-national legislatures than the winner-take-all SMD electoral system, because parties can win assembly seats even with less than a majority of voter support in an electoral district.

Partly due to this fundamental difference in electoral systems, there is a large divergence in the party systems as measured by the “effective number of parliamentary parties” (ENPP) (Laakso and Taagepera 1979; Norris 2009). Whereas the social democracies average seven parties in parliament, the adversarial democracies average less than half this amount (3.3 parties). With more parties in parliament, there is a greater likelihood that lower-class parties will also be included, and among our sample countries those with high ENPP are, in fact, more inclusive of lower-class political parties. Moreover, there is a clear negative correlation between the number of parties in parliament and the average imprisonment rates across these countries. As mentioned earlier, the inclusion of multiple parties facilitates building consensus because the parties must compromise to pass legislation. It also provides an incentive to develop inclusive policies such as universal welfare that can guard against the ill effects of unemployment and deprivation, factors that may lead to crime. The adversarial states, on the other hand, generally have only two or three major political parties and are more likely to experience single-party rule without the need to compromise with the “opposition.”

A second relevant dimension of the electoral system relates to regulations concerning political party and campaign financing. In all of the Nordic social democracies, the government is the primary source of political finance as shown in table 2. Rules vary from country to country, but the presence of public finance allows lower-class political parties to compete on a more even footing with political parties representing the upper class. In adversarial democracies, there is less prevalence of public campaign finance as can be seen in the case of the United States where political finance is largely in private hands. This makes it more difficult for lower-class parties to compete in the absence of strong labor unions.

A third relevant dimension of the electoral system concerns the eligibility to vote and the procedures for voter registration. These rules have a major impact on who participates in running for office and who participates in voting for candidates. Nordic social democracies are particularly inclusive in orientation. As displayed in table 2, they all make voter registration universal and automatic. In the case of Anglo-American adversarial states, however, only two out of four make voter registration automatic. These divergent electoral laws lead to major differences in political participation. The four Nordic social democracies displayed in table 1 average roughly 79 per cent voter turnout, whereas the four adversarial democracies have less than 64 per cent voter turnout as a percent of the voting age population (VAP). Yet even this figure overstates the case because Australia has compulsory voting. Among the three remaining adversarial democracies, voter turnout is only 57 per cent, and many of the people who do not turn out to vote in these countries belong to the lower social classes (Lijphart 1997).

The combination of a proportional representation (PR) electoral system with parliamentary government and public funding of political parties presumably gives representatives of the lower classes more of a chance to enter into the legislature and the ruling coalition. This makes for a more inclusive democracy compared to adversarial systems where lower-class parties may have inadequate or no representation. Lower-class parties are also likely to be under-represented when voter registration is not automatic and when there are few or no regulations on private financing of political campaigns. In these cases, voters may not even bother to register or to vote. With the exception of Australia, this phenomenon regularly occurs in the English-speaking adversarial democracies.

As more inclusive, consensus-oriented political systems, the Nordic social democracies have also been able to generate universal welfare states. As a result, their income inequality is low as represented by an average Gini  coefficient of only 25.6. The adversarial democracies have minimal to moderate welfare states and a much higher Gini of 36.2. These levels of inequality correlate highly with imprisonment rates (r = 0.73) across these countries as shown in table 3. As mentioned earlier, consensus-building, social corporatism, and welfare states appear to play a key role in contributing to crime reduction. Lower-class political participation plays an  important role not only in parliamentary parties but more broadly in political organization, cohesion, and solidarity of the lower classes as well. This is apparent in labor unionization rates, a primary indicator of lower-class organization.

When labor unions are weak and less able to influence politics through the vehicle of a working-class political party as in adversarial democracies, the state has little pressure and less incentive to invest in improving the working and living conditions of the lower classes. One symptom of this process may be a high level of immigration into the country. Rather than employers agreeing to pay higher wages or taxes to benefit unionized workers as in social democracies, they can recruit employees from overseas to reduce their labor costs in adversarial democracies. The business classes can thus both avoid paying higher taxes that could fund improvements in the welfare of the lower classes as well as take advantage of low-cost politically un-organized immigrants.

Table 3 shows how opportunities for lower-class political parties to be included in democratic politics can dramatically transform society by reducing crime. In social democracies, labor union density is high. The majority of workers can belong to unions because the business community does not have hegemonic control over the government. As a result, workers can join labor unions free of intimidation and harassment. With high levels  of unionization, workers are also able to demand higher taxes on the rich and a higher level of publicly funded social services. In turn, high levels of social services and welfare transfers can cut the poverty rate in half.

As shown in table 3, on average 69 per cent of the workers in the Nordic social democracies belong to a labor union compared to only 23 per cent of workers in the English-speaking adversarial democracies. The labor unionization rate is exceptionally low at 12% in the USA, the country with the highest incarceration rate. By contrast, the proportional electoral system in Nordic countries has created space for political parties allied with labor unions to come to power, stay in power, and be formidable opposition groups within the Parliament when out of power. As a result, these governments have been able to collect on average 46% of their GDP in tax revenue.  Resistance from the much stronger business community in the Anglo-American governments has stifled tax efforts so that tax collection is only 33% of GDP and even less (28%) in the USA.

The wealthy are able to have more power in the English-speaking adversarial democracies because of disproportional and adversarial electoral laws and laws that create dependence on private funding and private media for political campaigns. Under this system, even if a lower-class political party gets 30%-40% of the votes in an electoral district, it usually gets no seats. As a result, adversarial democracies levy relatively low taxes on the wealthy and, therefore, have less revenue to spend on comprehensive social programs that can eliminate, or at least reduce, the root causes of crime.

The difference in taxation rates means that the Nordic countries can support comprehensive welfare states whereas Anglo-American public services are more likely to be minimal, miserly, and means-tested (Esping-Andersen 1990) even though both types of states averaged roughly US$40,000 in per capita income in 2008. Whereas public social spending makes up 26% of the GDP in Nordic social democracies, it is only 18% in  adversarial democracies and less than 15% in the USA (Pontusson 2005). With such a minimal welfare state, poverty rates in these adversarial democracies are over 12%, which is more than double that of the Nordic social democracies where poverty rates are under 6%. Lastly, the poverty rate in the most adversarial country, the USA (17.1%), is more than quadruple the rate in Denmark (4.3%).

From a broader perspective, it seems that presidentialism and a disproportional electoral system may facilitate the disempowerment of the working class and the lower strata in adversarial democracies. Presumably, this structural factor helps create a larger breeding and recruiting ground for potential criminals in that over ten percent of the population is below the poverty line. An additional factor may be the high level of immigration into adversarial democracies. Low-strata individuals may be prone to blame immigrants for their problems rather than the business community and upper classes who may, in fact, be largely at fault for depriving them of political space, disempowering them politically and economically, and intentionally bringing in immigrants as a way of keeping overall wage pressures down.

Moreover, in adversarial democracies where restrictive and disproportional political institutions tend to marginalize or exclude lowerclass parties, the level of union membership may also remain low. When the workers have less power, capital dominates. As a result, tax collection is lower, as is the level of publicly funded social services for the lower classes. In this system, workers are at a disadvantage and more likely to fall into poverty because there is less of a safety net to prop them up. This approach has a “winner-take-all” quality for those with great wealth because they pay relatively lower taxes and do not have to bear the burden of lifting up what becomes a considerably large pool of disadvantaged, impoverished, and disillusioned people at the bottom. Ignored by what they perceive to be callous and indifferent elites, some sections of the large lower strata, especially a portion of the young males, become involved in criminal activities. This happens throughout the Anglo-American states where punishment becomes a means of “disciplining the poor” to accept their position within the socio-economic and racial hierarchy of society  (Wacquant 2009).

In this regard, the USA stands out as a deeply adversarial democracy, one that has experienced explosive growth in imprisonment since the 1970s. In the year 2000, 32.4% of black men in the USA aged 20 to 40 who had dropped out of high school had spent time in prison (Western 2007). Such a high level of incarceration has led to a vicious cycle whereby those incarcerated typically have no employment while imprisoned and few  employment opportunities upon their release. Their families and children live on the margins of poverty, physical and emotional abuse is high, and recidivism is also high, increasing the likelihood of these children falling into crime. This is even more so the case due to the criminalization of drug use, which is one of the few means poor minorities have at their disposal to escape (temporarily) the difficulties of living at the bottom of a stratified social, economic, and racial hierarchy.

Thus far, our structured, focused case comparison has found initial confirmation in support of our three hypotheses regarding the positive impacts of parliamentarism, PR elections, and social democracy on reducing crime. To conduct a larger test of these hypotheses, we now utilize multiple regression analysis on the twenty industrialized democracies for which Meyer (2007) has tabulated a comprehensive “social democracy index.” As shown in figure 1, these countries vary from a rate of 60 to 200 prisoners per 100,000 population, with the exception of the USA where the rate is over 700. We use Meyer’s index because it is the only major cross-national measure of the degree to which countries are social democratic. The index includes nine dimensions of social democracy: social and economic rights, universalistic social welfare state, social expenditure, coordinated market economy, co-determination, relative poverty rate, social stratification in the education system, labor-force participation rate, and income equality.

As shown in table 4, our regression models use imprisonment rates (as a share of the population) as the dependent variable. In model 1, we included three variables related to each of the hypotheses stated above. Both the social democracy index (a continuous variable with a minimum of 0 and maximum of 25) and parliamentarism (a dichotomous variable) were statistically significant and negatively related to (i.e., reduce) imprisonment while electoral system type (dichotomously defined as partly to fully proportional or fully SMD) was statistically insignificant. To control for other possible factors, we included in model 2 an array of control variables taken from the World Bank (2012) for population density, degree of urbanization, and per capita income as measured by purchasing power parity. Once again, parliamentarism and social democracy were negatively correlated with imprisonment, a finding that was statistically significant. Lastly, we included an additional control variable for the total number of residents living in metropolitan areas in model 3. This allowed us to control for the total  population level of a country, as small countries may have an advantage in reducing crime. In this model, parliamentarism was still highly statistically significant, while the social democracy index fell to a level just below conventional levels of statistical significance (p < .11). In this model, as in model 2, the rate of urbanization (positive) and per capita income (negative) were also statistically significant.

In this paper, we found general support for the notion that highly inclusive social democracies have fewer criminals than less inclusive adversarial democracies. The scope conditions of this inquiry, however, were limited. We only examined industrialized democracies, and the conditions present in these countries may not parallel those of developing and nondemocratic countries. Therefore, further research is needed to see whether these findings also apply to less developed states.

Nevertheless, among the industrialized democracies analyzed here, the most striking finding was the strong and positive correlation between presidentialism (a strong indicator of adversarial democracy) and high rates of imprisonment. Conversely, we found social democracy to correlate negatively with imprisonment. We also found states with smaller populations to generally have a smaller share of prisoners. Although our findings regarding uninominal and proportional electoral systems were not statistically significant, we did find overall support for the notion that adversarial democracies have more crime, whereas states with parliaments and a social democratic orientation have less crime.

One possible implication of this study is that states which focus on developing inclusive democracies can contribute positively to crime reduction and democratic deepening by providing more political space and opportunities to the lowest classes to participate in making the rules of the game for society. On the other hand, established democracies that put up barriers against the inclusion of the lower classes and lower-class political parties may witness significantly more incarceration and higher rates of crime.

The key to democratic deepening may, therefore, not be pro forma elections alone, but a deep level of inclusion of all strata of society into the political process (Heller 2000). Importantly, there must be opportunities for the lower and marginalized strata in society to participate actively in governance so that people from the lower classes can be elected into positions of power or be able to influence those in power. This may not only be crucial for raising the living standards of those at the bottom, but also for increasing social respect and opportunities for the lower classes. Including groups at the bottom of the socio-economic hierarchy into democracy may be especially pivotal for crime reduction because most crimes involving long-term prison sentences are committed by those from lower socio-economic strata. Hence, a key reform for political leaders and activists in industrialized democracies may be to transform their political institutions to be more inclusive of the lower classes.