Enduring Consequences of Right-Wing Extremism: Klan Mobilization and Homicides in Southern Counties

Introduction

Recent research has demonstrated that social movement activism can be linked to a wide variety of outcomes including the development of old-age policy in the United States (Amenta, Carruthers and Zylan 1992), election of blacks to public office in the South (Andrews 1997), hate crime enforcement and reporting (McVeigh, Welch and Bjarnason 2003), and state-level legislation banning same-sex marriage (Soule 2004). To date, research on social movement consequences has focused primarily on social movements' attempts to influence public policy or to win new benefits for constituents. Less is known about how social movements affect social relations within the communities in which they are embedded. Rarely, too, are the consequences of social movements examined over an extended period of time in order to assess the extent to which social relations in the present can be traced to social movement struggles of the past.

In this article, we examine the enduring consequences of Ku Klux Klan activism that occurred in the 1960s in the American South. More specifically, we focus on how Klan activism helped to create conditions in local settings that are conducive to violence, resulting in increased numbers of homicides even decades after the movement declined. Our research is motivated, in part, by prior findings that past racial violence can have lasting consequences. For example, research shows that the lynching of blacks in the late 1800s and early 1900s can be linked, empirically, to higher levels of homicide (Messner, Baller and Zevenbergen 2005), weak enforcement of hate crime legislation (King, Messner and Baller 2009), and the use of capital punishment (Jacobs, Carmichael and Kent 2005). By focusing our attention on a more recent episode of racial conflict, we are able to employ data that include measures of homicide both before and after the historical episode of racial conflict.

We argue that an empirical link between Klan activism and increasing homicide resulted from the ways in which movement activism disrupted community cohesion. We expect that, over time, weakened community cohesion continued to interfere with maintenance of informal social control mechanisms in local settings in ways which, according to criminological theory and research, should lead to higher levels of lethal violence. We expect that this was particularly the case in communities that were faced with the task of integrating newcomers as populations swelled in many “sunbelt” destinations (Tolnay 2003). As Tomaskovic-Devey and Roscigno (1996) show, economic development in the new South in recent decades has not been evenly distributed across geographic regions and this unevenness has been shaped by race and class divisions of the past. Historical conflict should also be relevant in determining how Southern communities fare when it comes to maintaining social order and limiting lethal violence during a period of population growth, economic transformation, changing racial norms and other broad social changes. Perhaps no organization embodies the legacy of racial conflict and violence in the South more than the Ku Klux Klan.

Klan Mobilization in the 1960s

Threats to white privilege posed by the burgeoning Civil Rights Movement and proposed federal legislation saw the Ku Klux Klan reemerge in the 1960s as a violent resistance organization intent on maintaining Jim Crow-style segregation. Similar to the first Ku Klux Klan that rose in the 1860s following the Civil War, the Klan of the 1960s drew its support primarily from white Southerners who were threatened by a challenge to economic, political and social systems that were based on the oppression of blacks.

The movement's strength in the 1960s paled in comparison to the Ku Klux Klan that spread throughout the nation earlier in the century when, from 1915 to 1924, millions of American men and women joined the “Invisible Empire” and openly engaged in rallies, parades and marches that attracted hundreds and sometimes thousands of curious spectators. During the 1920s, Klan leaders diligently sought to recruit middle-class Americans. They addressed a broad range of issues in hopes of attracting support, while attributing much of the blame for problems confronting their constituents to immigration and to Catholic influence (McVeigh 2009). In the 1960s, Klan membership could be counted in the tens of thousands, rather than in the millions, and the organization appealed primarily to white working class men and women who felt threatened by advances in civil rights (Cunningham 2004). Although there was broad support for segregation among white Southerners, middle-class and upper-class whites tended to distance themselves from the Klan in the 1960s and instead resisted integration through institutionalized politics or within organizations such as the Citizens Council that drew support from business leaders and from more affluent whites (McMillen 1971).

Although the 1954 Brown school desegregation decision was a major impetus for the rise of the civil rights-era KKK, it was not until the early 1960s that the Klan coalesced into a formidable, unified movement. In the 1950s, Klan activity was characterized by factionalization and often-disorganized efforts at realignment. Mississippi, well-known as the most violently brutal Klan state in the mid-1960s, had no KKK organization at all prior to 1963 (Dittmer 1995). Klan membership in North Carolina, the state with by far the largest mass mobilization in the 1960s, had fallen to no more than 200 during the late 1950s (U.S. House of Representatives 1967). The only serious region-wide organization, the Georgia-based U.S. Klans, was beset by repeated leadership controversies, including Alabama Grand Dragon Alvin Horn's scandalous marriage in 1957 to a 14-year-old girl and his subsequent incarceration for contributing to the delinquency of a minor. When national U.S. Klans head Eldon Edwards died unexpectedly in 1960, the KKK field was in rudderless disarray (Chalmers 1981).

That changed in 1961, however, when the Robert Shelton-led United Klans of America began a coordinated organizing campaign across the South. By 1963, Shelton's UKA could boast of more than 9,000 members spread over several hundred chapters – referred to, in Klan parlance, as “klaverns” – in eight states. Two years later, a Congressional investigation found that UKA membership was three times that size, with at least 12,000 members in North Carolina alone. By 1966, the UKA had organized more than 350 klaverns across the South, and had also made token inroads into the Midwest and the eastern seaboard (Cunningham 2004; U.S. House of Representatives 1966, 1967).

Klan leaders in the 1960s presented their organization as one that was dedicated to preserving the Southern way of life and to resisting threats from blacks and from outsiders who, they claimed, were using the civil rights struggle to advance an anti-American agenda and stir up trouble in their communities. In some of these communities, local authorities gave the Klan free rein to organize and to employ violent tactics in defense of white supremacy. Violent resistance was a defining characteristic of the 1960s Klan, including the UKA, but of course not all Klansmen engaged in violent acts. Klan leaders characterized their organizations as patriotic fraternal orders prepared to defend the nation from communism and other “un-American” influences. As such, members participated in weekly klavern meetings, public rallies and social events such as Klan turkey shoots and fish frys (Cunningham 2008). The insularity produced by these sorts of activities, ironically, contributed to defensive claims that the Klan functioned to preserve law and order. When FBI agents and the national media descended on the state of Mississippi in 1964 in response to the disappearance of three civil rights workers, Klan propaganda portrayed the victims and other civil rights adherents as communists and criminals. As with brutal lynchings of blacks in the late 1800s and early 1900s, Klan violence, from the Klansman's perspective, represented vigilante justice that was required by the circumstances if law and order were to be upheld and Southern institutions preserved (MacLean 1994; McMillen 1990; Tolnay and Beck 1995).

The Klan's failed attempt to preserve Jim Crow segregation, combined with a concerted effort from the FBI to infiltrate and undermine the Ku Klux Klan and its agenda, led to a rapid decline in membership and a virtual collapse of the movement in the late 1960s (Cunningham 2004). Although the efforts of David Duke and other leaders contributed to the Klan's brief resurgence in the late 1970s (Sims 1996; Berbreier 1998), and while Klan organizations continue to survive into the 21st century (Dobratz and Shanks-Meile 1997; Blee 2002), the Klan has been unable to reestablish itself as a formidable resistance organization. We expect, however, that the KKK's legacy endures in less direct and likely unintended ways. In particular, we argue that the Klan's mobilization in the 1960s altered the community context in local settings in ways that continue to increase the likelihood of lethal violence in those settings.

Community Disruption and Criminological Theory

Scholars studying crime and violence have long recognized the important role that community context can play in maintaining social order. Showing strong links to the writings of Durkheim (1951), for example, Braithwaite argues that communities characterized by communitarianism are best equipped to maintain social order. In communitarian communities, according to Braithwaite (1989), interdependence and shared cultural values weigh heavily on individuals, as they would be more likely to feel a sense of shame if they violated community norms and standards. Along these same lines, researchers have emphasized the extent to which generalized trust within local settings reduces violent crime (Rosenfeld, Messner and Baumer 2001; Sampson 1997). This trust also extends to community institutions. Sampson and Bartusch (1998) demonstrate how particular ecological conditions can produce “legal cynicism” – i.e., dissatisfaction with the police and formal legal processes – that in turn contributes to a rise in crime and disorder.

These theoretical perspectives are highly consistent with the central claims of social disorganization theory (e.g., see Shaw and McKay 1942; Sampson and Groves 1989), a theoretical framework that has been successfully applied in numerous studies of crime and delinquency. According to the theory, weak community cohesion undermines informal social control mechanisms that might otherwise reduce crime. Criminal acts, from this perspective, represent an affront to the community and not just to the individual victims of crime. A high rate of crime suggests that many individuals feel little sense of obligation to the community or to standards and norms that might otherwise constrain deviant behavior (e.g., see Bellair 1997; Lyons 2007; Saegert and Winkel 2004; Sampson and Groves 1989). Empirically, social disorganization theorists have emphasized factors such as population change and racial or ethnic heterogeneity that can make it difficult to form and maintain community solidarity and cohesion. Population turnover, in particular, can exacerbate social disorganization because strong social bonds can take years to form and require frequent meaningful interactions among neighbors and other community members.

The social disorganization perspective has been advanced through quantitative studies of community characteristics, but also through qualitative research that has placed more emphasis on culture and micro-level processes. Small (2002) argues, for example, that cultural characteristics are central to changes in community participation. More specifically, as particular resident cohorts “frame” their neighborhoods, they shape others' willingness to engage in the sort of collective projects that build collective efficacy and participation in community initiatives, thereby affecting possibilities for subsequent contention and crime. Such frames emerge not as the simple coalescence of individual attitudes and values, but because of the efforts of entrepreneurs that mobilize particular schemata in ways that resonate broadly toward shared values and goals.

Identity, Framing and Community Cohesion

We propose that processes associated with mobilizing a racist extremist movement can lead to changes in communities that make them more prone to violence. Social movements engage in an ongoing process of drawing group boundaries, identifying opponents, and defining and redefining the nature of the group and its relationship to other groups in society. Boundary construction plays a vital role in attracting supporters, as individuals are more likely to participate in collective action if they see themselves as being part of a clearly defined group (Gamson 1992). This same sense of belonging also helps to sustain participation and commitment over the long haul (Taylor 1989; Taylor and Whittier 1992). Identity work within a social movement may transform the lives of participants, permanently bringing about changes in their own sense of self and how they should live their lives and engage the broader society (McAdam 1988). Yet identity-based claims and contestation may also lead to lasting change in the political terrain and in the nature of inter-group relations (Polletta and Jasper 2001, p. 296).

Conceivably, the Klan's mobilization in the 1960s could have produced a myriad of enduring consequences, some noteworthy and others not. We expect that violence is one important legacy of Klan activism, and this outcome is due in large part to the way in which the movement mobilized racial identities in the 1960s. In one sense, the group boundaries for the Ku Klux Klan were rather clear. Klansmen were not at all subtle in communicating that they were about organizing whites and whites alone. Whiteness, from the perspective of Klan members and supporters, made them superior to blacks and others not considered to be white. Whiteness also corresponded with rights and privileges that Klan members and supporters believed should be reserved for members of their own race.

As a cultural category, the Klan's version of “authentic whiteness” needed to be continually enacted, through clear and consistent demonstration of allegiance to cultural values centered on segregation. In many cases, spectators were forcibly ejected from Klan rallies when their appearance or manner painted them as suspect “aliens” or “white niggers.”(Alsop 1966; U.S. House of Representatives 1966) Thus, boundaries around racial categories were intentionally drawn tightly, to exclude Catholics, Jews, “Communists” or anyone, regardless of race or religion, who was thought to be allied with blacks and the civil rights struggle (Crespino 2007).

Klan enemies included political and legal elites who failed to protect white privilege, and movement rhetoric thus frequently sought to undermine the legitimacy of established authorities. Similarly, the Klan's strong emphasis on boundaries separating its constituents from other members of American society, and the way that it resorted to violence and intimidation to enforce those boundaries, exacerbated tensions between groups in local contexts, making it increasingly difficult to generate trust and a sense of community cohesion when de jure segregation came to an end. Indeed, Klansmen organized with the specific goal of ensuring that local communities would maintain racial divisions, while also generating fear and suspicion toward any community members who did not share this vision.

In addition to constructing group boundaries, social movement actors are also centrally involved in defining circumstances in ways that will encourage others to join the cause (Snow et al. 1986). There is reason to expect that interpretive frames constructed by Klan activists in the 1960s should have contributed to a general shift in the way in which many Southerners viewed authority. Particularly important, in this regard, is the way in which Klansmen employed defensive rhetoric, claiming that the movement was the only organization in existence to fight un-American influences, to thwart challenges to Southern traditions and values, and to preserve law and order during a time period when, they argued, the federal government sided with law breakers. Through the movement's appeals to Protestant Christianity, Klansmen claimed to be obeying a higher authority and they were not bound to follow the laws of a lawless society being imposed upon them by outside influences and agitators (Crespino 2007).

This type of oppositional framing was, for many, reinforced by historical developments. The Klan's failure to defend Jim Crow in the face of federal legislation that mandated change over the objections of local political elites, coupled with the fact that the FBI played a significant role in bringing about the Klan's demise (Cunningham 2004), further undermined the legitimacy of legal authority in the eyes of many white Southerners. Particularly important, in this regard, is the way in which Klansmen asserted a belief in the right of citizens to take the law into their own hands when law enforcement officials are deemed unwilling or unable to bring “justice” to those who are seen as deserving of punishment. In doing so, they were drawing upon a tradition of vigilante justice practiced in the South – a tradition that included the use of lynching as a means of intimidating and maintaining dominance over blacks (McMillen 1990; Tolnay and Beck 1995; Messner, Baller and Zevenbergen 2005).

Black Southerners also had incentive to take the law into their own hands, but for different reasons than those held by white segregationists. In the decades following Reconstruction, Southern blacks were disfranchised and left powerless within institutionalized politics (e.g., see Redding 2003). Jim Crow customs and laws were enacted to further undermine the capacity of black Southerners to achieve economic prosperity and social equality. Deep-seated prejudice against blacks, combined with threats posed by blacks to white interests, meant that black Southerners would regularly face intimidation, violence and imprisonment on false charges at the hands of white counterparts (McAdam 1982; McMillen 1990).

Blacks, therefore, had to adapt to living under conditions where legal authorities served the white community and often condoned – or directly produced – violent responses to challenges to the racial order. As McAdam (1988) reported in his study of the Freedom Summer campaign, many white civil rights volunteers in the 1964 project were surprised to learn that black Mississippians, rather than passively submitting to oppression, were often armed and prepared to use lethal force to protect themselves and their property. Ties between local law enforcement agents and the Ku Klux Klan further clarified the extent to which black Southerners would have to fend for themselves when it came to protecting themselves from whites, but also when it came to settling internal disputes and conflicts. Black Southerners had reason to fear local authorities and came to understand that legal officials were at best uninterested in serving their law enforcement needs (Beck and Tolnay 1997; Dittmer 1995; Payne 1995). Thus, the Klan's framing of legal cynicism within a tradition of vigilante justice not only influenced the way that white sympathizers came to view citizens' relationships to legal authority, but also helped to undermine black Southerners' investment in the efficacy of legal institutions.

Summary of the Argument

Through processes of boundary construction and framing of collective grievances, Klan activism of the 1960s disrupted community cohesion and undermined generalized trust and the perceived legitimacy of local authority. Indeed, Klan organizations deliberately sought to shatter any semblance of a communitarian spirit that would otherwise inhibit crime, violence and disorder (Braithwaite 1989). While Klan members contributed directly to lethal violence in the 1960s, we believe that the KKK's relationship to violent crime in subsequent decades stems not primarily from the acts of former Klan members and sympathizers, but rather from the ways in which prior Klan activism produced tears in the social fabric that continue to undermine informal social controls that might otherwise constrain generalized violence. Population growth in the latter half of the century likely exacerbated these circumstances. A deep literature on the deleterious effects of residential instability, for example, has shown that resident turnover and significant in-migration decreases the likelihood that communities can maintain a non-factionalized sense of social cohesion and collective efficacy (e.g., see Shaw and McKay 1942; Bursik and Grasmick 1993; Sampson et al. 2002; Xie and McDowall 2008; Hipp et al. 2009; Decker et al. 2009). Managing population growth while still maintaining social control, we expect, would be more difficult in local settings where the Klan had previously played a role in disrupting community cohesion.

Data and Methods

To test our argument, we examine county-level variation in Southern homicides. We limit our analysis to those states where the Klan made a diligent effort to establish local chapters throughout the entire state. That would include counties in the states of Alabama, Arkansas, Georgia, Florida, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee and Virginia (U.S. House of Representatives 1966). We focus on counties as units of analysis because we have argued that Klan activism likely contributed to change in local settings that would continue to influence social behavior long after the movement had itself left the scene. Our focus on counties also allows us to examine variation using numerous cases. This unit is substantively appropriate as well, as many of the Klan's chapters were organized at the county level and counties were especially meaningful political units in southern states during the time period in which the Klan was active (Andrews 2004; Cunningham and Phillips 2007).

To demonstrate the validity of our argument, it is not enough to simply show that homicide occurs more frequently in counties where the Klan had once been active.¹ It is possible that the Klan settled into more violence-prone communities and any current relationship between prior Klan activism and contemporary homicide is spurious. To rule out that possibility, we utilize a longitudinal analysis that includes measures of homicides prior to Klan mobilization, as well as after Klan mobilization. We produce fixed-effects estimates to show how Klan activism is related to changes in homicides within counties over a time span of four decades, while controlling for other relevant variables.

A fixed-effects model is particularly well suited for estimating the consequences of an event – in this case, mobilization of the Klan in the 1960s (Allison 1994). It is important to keep in mind that this model provides estimates of change that occurs within counties over time, rather than estimates of variation in the dependent variable across counties at one moment in time. Estimates are equivalent to those that would be derived if a dichotomous variable for each county were inserted in the model. Particular attributes of each county, therefore, are controlled or held constant and the goal is to identify how change in independent variables within counties is related to the average level of change in the dependent variable within counties over time. An important feature of the fixed-effects model is that it automatically controls for all constant and unobserved differences across cases (Allison 1994).

Because we are interested in estimating the enduring consequences of Klan activism, we examine change in the number of homicides over three different durations. First, we examine change in homicides from 1960 to 1970 in order to assess the short-term change, if any, that resulted from the Klan's presence within a county. To examine medium- and long-term consequences, we also examine change from 1960 to 1980, and change from 1960 to 1990. In each case, we are examining change in the number of homicides from a point (1960) prior to the Klan's mobilization to a point that followed the mobilization and ultimate decline of the Klan.

Much of the publically-available ed by Steven Messner, Luc Anselin, Darnell Hawkins, Glenn Deane, Stewart Tolnay and Robert Baller with funds provided by the National Consortium on Violence Research and the National Science Foundation. As our longitudinal analysis spans from 1960 to 1990, it is necessary to adjust for numerous changes in boundaries for counties and county equivalents in the state of Virginia. Data are aggregated in counties that experienced a boundary change to form new geographic units, or “county clusters.” We follow the clustering strategy used by Messner and his colleagues, creating 11 of these units out of the 28 counties (or in some cases former counties) that were affected by boundary changes.

Homicides

We use the NCOVR data to form a measure of the average number of homicides in a three-year period, centered on the decennial census year (1959-1961, 1969-1971, 1979-1981, 1989-1991). The homicide counts are taken from the U.S. vital statistics compiled from death certificates at the county level by the National Center for Health Statistics (Rosenfeld, Messner and Baumer 2001). Because the distribution of homicides across counties is severely skewed, we use a logarithmic transformation (the log of 1 + the average homicide count).²

Klan Activism

We estimate Klan activism through a measure of whether or not the Klan had an active chapter in the county between 1964 and 1966. The source of these data is a census of Klan units compiled as part of the House Committee on Un-American Activities hearings on the Ku Klux Klan (U.S. House of Representatives 1967). To provide a sense of the Klan's distribution across the region, Table 1 shows the total number of Klan chapters in each state, as well as the percentage of counties that had at least one active Klan organization. Given the ephemeral nature of many self-proclaimed local KKK organizations, we use the UKA – which was by far the largest and most stable KKK group of the era – to represent the overall Ku Klux Klan population. In three cases, however, we merged the UKA data with that of other Klan organizations (in Florida, with the United Florida KKK; in Mississippi, with the White Knights of the KKK; and in Louisiana, with the Original Knights of the KKK), as in each case there was considerable movement of members back and forth between those organizations and the UKA. During this time period, no other Klan organization had more than a handful of units or members. Our measure captures more than 95 percent of the estimated KKK members across the South, and can thus serve as a strong and unbiased proxy of overall Klan organizing. Yet it should be noted that no membership data exist that would allow us to assess the relative size and strength of particular Klan units across counties.

Control Variables

We include control variables measured at four time periods corresponding with our measures of homicide. Because homicides are more frequent in populous counties, we control for the natural log of the total county population. Because we are analyzing change over time, we are particularly interested in determining whether population growth resulted in larger increases in homicides in counties that experienced Klan activism compared to counties that did not experience Klan activism. Prior Klan activism could have interfered with effective integration of newcomers during a time when population size was swelling in many Southern communities. Prior research suggests that economic deprivation can also disrupt community cohesion (Shaw and McKay 1942; Sampson, Morenoff and Gannon-Rowley 2002; Sampson and Raudenbush 1999) and also undermine the legitimacy of legal authority (Grant and Martinez 1997; Braithwaite 1989; Blau and Blau 1982; Logan and Messner 1987).

We use three variables from the NCOVR data to calculate a measure of economic deprivation. One of these is the log of median family income, centered around the mean value for all counties during the decade in question (1959, 1969, 1979, 1989). A second is a measure of income inequality captured by a Gini index of family income inequality. A third is the percent of families below the poverty line (for the 1960 decade, it is measured as the percent below $3,000). Because these measures are highly correlated, and because we are interested in developing a general measure of economic deprivation, we constructed factor scores from a principle components analysis. For each decade, all three items load on a single factor that accounts for, at minimum, 83.4 percent of the variance in the variables (with Eigenvalues ranging from 2.5 to 2.8 across the four decades). The factor score for economic deprivation was calculated using the regression method.

Homicides may also be related to the investment that communities make in police protection. Greater investment in police protection may serve as a deterrent to crime, and it may also tell us something about the value that the community as a whole places on social order. The Census of Governments provides information on the amount of money spent on police protection by local governments situated within each county. The Census of Governments is published every five years. Therefore, we include data from 1957, 1967, 1977 and 1987 to correspond with our four data panels. Because the spending measure is severely skewed, we use the natural log of spending (measured in thousands of dollars). We include two additional measures of community cohesion that may be related to homicide. One is the percent of males ages 14 and older (ages 15 and older for the latter two decades) who are divorced. We use the decennial census to calculate a measure of the percent of homes in a county that are owner occupied.

The racial composition of the county should also be related to homicide and is especially relevant because we are considering the possible influence of Klan mobilization on homicides in Southern counties. Therefore, we control for the percent of the population that is black for each of the four decades.³ Homicide may also be related to age distributions in counties over time. Median age can provide some insight into the ways in which local contexts are changing over time. Some counties, for example, may be aging as younger people move out in search of better economic opportunities. At the individual level, we know that younger people are more likely to commit crimes than older people (Gottfredson and Hirschi 1990), and that may also be reflected in the relationship between homicides in a county and median age.

We also consider the possibility that violent crime may be related to the broader political context. If there is a relationship between Klan activism and homicides, it is important to rule out the possibility that the relationship merely reflects the general political orientation of the local setting. We include a measure of Republican presidential voting and a measure of the percent of votes that were cast for segregationist third-party presidential candidate George Wallace in the 1968 election. We measure Republican voting as the percent of the vote that went to the Republican presidential candidate in the election most proximate, in time, to the decennial year (1960, 1968, 1980, 1988). The Wallace variable is the percent of votes cast for Wallace in each county. Values for this variable in the first period of our panel design are zero for all cases, because that time period precedes the Wallace vote.

Finally, in our longitudinal analyses we also include a dichotomous variable to control for potential time period effects. For example, if there was a general increase in homicides during the 1980 period compared to the 1960 period, we want to control for that in order to ensure that any estimated effect of prior Klan activism does not simply reflect a general increase in homicides across the two time periods.

Data Analysis

Results presented in the first two columns of Table 2 show the estimated effects of the independent variables on the change in homicide frequency from 1960 (prior to the Klan's resurgence) to 1970 (soon after the Klan's decline). In the first model we exclude the measure of Klan activism and we include it in the second model. As can be seen, only two control variables have a statistically significant effect on increasing homicide. Not surprisingly, homicides tended to increase in counties that experienced population growth. We also find that after controlling for other variables, counties tended to have more homicides in the 1970 time period in comparison to the 1960 time period. When we add our measure of Klan activism to the model, we see that net of other variables counties that experienced Klan activism showed larger increases in homicides than counties that did not experience Klan activism.

Results presented in the first two columns of Table 2 indicate that Klan activism, which intervened between the 1960 and 1970 time periods, was one of only a few factors that contributed to an increase in homicides in Southern counties. But was this effect enduring or temporary? To address this question, we examine change in homicides over a longer time span. When we examine change from 1960 to 1980, more of our control variables emerge as significant predictors of increasing homicides. Population increase, of course, continues to have a strong effect on homicide increase. But we also find that homicide tended to increase in counties with increasing proportions of blacks and increases in divorce. Controlling for other variables, homicides were reduced in counties with greater increases in expenditures on police protection. The control for time period indicates that, all else constant, homicides increased in the 1980 period relative to the 1960 period.

An increase in Republican voting (in this case the change in the vote for Nixon in 1960 and the vote for Reagan in 1980) is also positively related to increases in homicide. This is the only time interval in which the variable is statistically significant. We suspect that this finding may reflect a temporary disruption of social cohesion resulting from a dramatic shift in party alliances that took place during this time period as, in many Southern contexts, white voters who were angered by the alliance between the civil rights movement and the national Democratic Party broke with the party. The Wallace vote, on the other hand, is associated with slight reductions in homicide after accounting for the shift in Republican voting. It should be noted, however, that during this time period the shift toward Republican voting was most pronounced in Wallace strongholds.

When we add the Klan activism variable we see that Klan presence continues to have a significant and positive effect on homicide increase when considering the 20-year time interval rather than the 10-year interval. It is also worth noting that including the Klan variable produces a slight reduction in the magnitude of the coefficient for the police protection variable. Although the reduction in the size of the coefficient is not dramatic (moving from significant at the .05 level to just shy of statistical significance (p = .062), it does appear that at least part of the correlation between police expenditure and homicide is attributable to changes that Klan activism brought to some Southern counties.

When we examine change in homicides over an even longer time span of 30 years, we found that increases in population size, increases in the percentage of black residents, and increasing levels of economic deprivation led to increasing homicide, while increasing expenditures on police protection reduced homicides. When we add our measure of Klan activism we have further evidence that the effect of Klan mobilization on lethal violence was enduring. In fact, the coefficient for the Klan variable is even larger than was the case when we examined change from 1960 to 1970. It is also noteworthy that when we add the Klan variable, coefficients for our measures of police protection and economic deprivation are reduced. Once again, we are careful not to exaggerate the magnitude of the change. However, this does suggest that increases in homicide that might be otherwise be interpreted as resulting from limited investment in social control and from economic deprivation (which typically disrupts community cohesion) are in part reflecting ­differences between counties that experienced Klan activism in the 1960s and those that did not. Of particular interest, we also find that the dichotomous variable that captures the time period effect is reduced substantially when the Klan variable is included in the model and is no longer statistically significant. In other words, the increase in lethal violence across the two time periods that is unexplained by our control variables is no longer significant when Klan activism is factored into the equation.

When we examine changes in homicide frequency from 1960 to 1990, we find that Klan activism is associated with an increase of 0.134 in the logged value of the number of homicides averaged across a three-year time interval. Another way of expressing that would be to say that net of control measures and net of all unmeasured constant differences between counties, we would expect to find slightly more than one (1.14) additional homicide per year in a typical county that experienced Klan activism relative to a county that did not experience Klan activism. While the number does not appear large at first glance, it is worth remembering that approximately 45 years have passed since the Klan's 1960s heyday, and the Klan was active in 300 Southern counties included in our analysis. When viewed in that light, the Klan's legacy of violence looms large. Indeed, it could be argued that the Klan activism contributed to as many as 13,000 additional homicides over time.

Integration of Newcomers

In this research, we showed that counties that experienced Klan activism in the 1960s tended to also experience an increase in homicides, and that effect endured for several decades. We also argued that Klan activism in the 1960s may have been particularly problematic in counties that were faced with the task of integrating newcomers during a period of sunbelt transition. All else constant, communities characterized by cohesion and general consensus should be better suited to integrate newcomers in ways that contain violence.

To examine this possibility, we calculate an interaction between Klan activism and our population measure and include it in the analysis of change in population from 1960 to 1990. To facilitate interpretation of the interaction effect, we center the population variable on its mean value. Results are presented in column 1 of Table 3. As can be seen, the coefficient for the interaction term is highly significant, which indicates that the effect of population change on increasing homicide does, in fact, depend upon whether the Klan had been active in the county in the 1960s. The main effect of the population variable shows that population growth has a strong effect on increases in homicide even in counties with no Klan activism. However, the positive and statistically significant coefficient for the interaction term indicates that the effect of population change is substantially stronger within the counties where the Klan had been active. Results also indicate that the Klan activism variable is most strongly related to increasing homicide in counties with higher than average population growth.

If our explanation for this empirical relationship holds merit, we might expect that population growth would also lead to particularly high increases in homicide in counties that display other forms of social disorder and weak community cohesion. Many researchers have identified economic deprivation as an important factor that undermines social order and community cohesion. As can be seen in column 2 of Table 3, an interaction between the population variable and the economic deprivation variable operates in a manner that is similar to the interaction of population and Klan activism. Population change leads to more homicides in counties with average levels of economic deprivation, but that effect grows even stronger in counties characterized by higher levels of economic deprivation. In the third column of Table 3, we see that the interaction between population and Klan activism remains statistically significant when we simultaneously consider the interaction of population and economic deprivation.⁴

Conclusion

Both crime and social movement activism were once broadly understood to have common root causes (Merton 1938; Gurr 1970). In this research, rather than examining how social disorganization can generate activism we have instead considered the possibility that some forms of activism can disrupt social organization in ways that are detrimental to the collective good. More specifically, we have argued that Klan activism in the 1960s changed the communities in which the organizations were present, not only in the short term but also in the long term. The Klan constructed boundaries that exacerbated pre-existing divisions among community residents, and the organization frequently resorted to violence and intimidation to enforce those boundaries. It drew upon and highlighted a historical tradition of vigilante justice in the South, and gave both white and black members of the community reason to view legal authority with fear or suspicion. Because homicide and other forms of violence can cause further damage to social cohesion and can undermine informal social control mechanisms, dynamics set in motion by Klan activism can endure long after memories of the Klan's contributions to disorder have faded.

Our research contributes to a growing literature that examines the consequences of social movement mobilization. Rather than focusing on a particular policy outcome or the achievement of a goal specified by the movement, however, we aim to reveal ways in which social movements can have a broader impact on society. Testing an argument such as ours presents some methodological challenges, however, because many factors other than Klan activism can (and do) contribute to homicides in Southern counties. Here, we have made systematic comparisons across many cases, showing that homicide is, in fact, more frequent in Southern counties where the Klan had once been active. Most importantly, we are also able to show that this reflects a change over time that is not due to Klan organizations initially becoming active in communities that were already characterized by frequent homicide.

By giving serious attention to the timing of events and by employing longitudinal data, we believe we have made a compelling case for our argument about how right-wing extremism can produce subsequent violence that endures decades after the movement has declined. One limitation of our research, however, is that we are unable to make confident claims about the racial identities of the violent criminal offenders. We expect that Klan activism led to increases in homicides with black as well as white offenders. Yet we lack data that would allow us to determine the race of homicide offenders during time periods prior to the Klan's mobilization. In preliminary analyses, we did estimate models with an interaction between our measure of Klan activism and our measure of the percent of the county that was black. The interaction is not statistically significant, which suggests that the effect of Klan activism on homicides does not depend on the racial composition of the county.⁵

During the 1960s, leaders of the Ku Klux Klan drew strength from deep-seated racial prejudice among white Southerners and revived an organization that aimed to preserve white privilege through violent resistance to the civil rights movement and to the dismantling of Jim Crow segregation. Thousands of Americans joined the organization and participated in activities initiated by the Klan. Even more Americans were victimized by the Klan's violence and intimidation. Research on social movement activism has shown how movements can often be agents of progressive change. As we have shown, however, some forms of social movement activism can do lasting harm to individuals and to society.

Notes

References