\ ‘ 1.. v. .. hang «Is... x ‘1’. Va 9‘ a. 37!...‘31MWVJ. 1| ..\ L. w: ; ~ .k .2. - 1 “It i. 4:; J. V t . . , a? t . QR allu.‘ r $9.}..uk A . . ”W 5:61 ‘3'“ 3"" HF ' Mum ’ “13:34“. a!” $1.. ;. v ... .. t. t :rd. :- iffifiwmfla ..,_....wmwfiwwfim%fi£ . _ V . . . .2. ||| ' “€533 LIBRARY 1 co?! Michigan State University This is to certify that the dissertation entitled MAJORITY RULE: DIRECT DEMOCRACY AND MINORITY RIGHTS presented by Daniel C. Lewis has been accepted towards fulfillment of the requirements for the Doctoral degree in Political Science y. a Major grofessor’s Signature —_ 5/2/05, Date MSU is an affinnative-action. equal-opportunity employer PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE JAN am 20131 has. 5/08 K IProj/AccsPres/CIRCIDateDue.indd MAJORITY RULE: DIRECT DEMOCRACY AND MINORITY RIGHTS By Daniel C. Lewis A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Political Science 2008 Abstract MAJORITY RULE: DIRECT DEMOCRACY AND MINORITY RIGHTS By Daniel C. Lewis Does direct democracy endanger minority rights? This question has long been at the center of the debate over direct citizen legislative institutions like the initiative and the referendum. Direct democracy institutions were implemented in American state governments to allow citizens to circumvent corrupt and unresponsive lawmakers. Critics of citizen legislation, however, argue that in circumventing representative government, direct democracy exposes minority groups to potentially tyrarmical policy outcomes favored by only a slim majority of the voters. Empirical research into this question has produced mixed results, with evidence supporting both sides of the debate. This dissertation seeks to more clearly reveal the impact of direct democracy institutions on minority rights through a systematic examination of contemporary minority policies in the American states. It expands beyond the previous literature by accounting for both the direct and indirect impacts of direct democracy on minority rights while providing more direct tests of the tyranny of the majority hypothesis by comparing outcomes in direct democracy states to outcomes in non-direct democracy states. Results from seven specific policies and three broad policy areas suggest that the rights of minorities do tend to be put at heightened risk under direct democracy institutions, but this impact may be contingent on public support for minority groups. Cepyn'ght by DANIEL C. LEWIS 2008 Dedicated to my family Elizabeth, Baby and Nora ACKNOWLEDGEMENTS I would like to express my sincere gratitude to my advisor, Saundra K. Schneider, for her constant guidance and patience throughout my time here at Michigan State University. I would not have been able to successfully navigate the Ph.D. program and find a great professional position without her vital support and wisdom. I would also like to thank the other members of my dissertation committee - William G. J acoby, Richard C. Hula, Charles Ballard, and Elisabeth R. Gerber — who have provided me with the kind of access, constructive criticism and insight that has undoubtedly improved my dissertation and helped me in furthering my professional career. I also need to thank my fellow graduate students in the Department of Political Science for providing scholarly advice as well as rather unscholarly distractions (not to mention the best softball team this side of the Kinesiology Department). In particular, I owe a debt of gratitude to Frederick Wood, David Dreyer, Jeremy Duff, Carl Snook, Dr. Major J .P. Peltier, Sarah Gubala, Judd Thornton, and Shane Singh. Of course, I need to thank my parents, Chief and Her Majesty, for their life-long support, which has allowed me to follow my dreams. Lastly, I am grateful for the love and support of my wife, Elizabeth, throughout this arduous journey. TABLE OF CONTENTS LIST OF TABLES ........................................................................................................... viii LIST OF FIGURES ............................................................................................................ x CHAPTER 1: DIRECT DEMOCRACY INSTITUTIONS & THE THREAT OF TYRANNY ......................................................................................................................... 1 1.1 Direct Democracy & the American States ................................................................ 3 1.1.1 The Impact of Direct Democracy ...................................................................... 9 1.2 Direct Democracy & Minority Rights: Tyranny of the Majority? .......................... 11 1.2.1 Whose Rights? ................................................................................................. 19 1.2.2 Existing Research ............................................................................................. 22 1.3 Goals & Plan of the Study ...................................................................................... 25 CHAPTER 2: DIRECT DEMOCRACY & THE DIFFUSION OF ANTI—MINORITY POLICIES ......................................................................................................................... 27 2.1 Same-Sex Marriage Bans ........................................................................................ 28 2.1.1 The Politics of Same-Sex Marriage ................................................................. 29 2.1.2 Event History Analysis .................................................................................... 36 2.1.3 Results .............................................................................................................. 48 2.2 Official English ....................................................................................................... 52 2.2.1 The Politics of Official English ....................................................................... 53 2.2.2 Event History Analysis .................................................................................... 58 2.2.3 Results .............................................................................................................. 62 2.3 Affirmative Action Bans ......................................................................................... 67 2.3.1 The Politics of Affirmative Action Bans ......................................................... 69 2.3.2 Analysis ............................................................................................................ 71 2.4 Discussion ............................................................................................................... 79 CHAPTER 3: A REPRESENTATIONAL FILTER? THE PASSAGE OF ANTI- MINORITY POLICY PROPOSALS ............................................................................... 81 3.1 Anti-Minority Policy Proposals .............................................................................. 82 3.1.1 Direct Democracy vs. Representative Democracy .......................................... 85 3.2 Direct Democracy & Anti-Minority Proposal Passage ........................................... 96 3.2.1 Results .............................................................................................................. 99 3.3 Discussion ............................................................................................................. 102 vi CHAPTER 4: THE FLIP SIDE; DIRECT DEMOCRACY & PRO-MINORITY POLICIES ....................................................................................................................... 104 4.1 Sexual Orientation & Non-Discrimination Laws ................................................. 106 4.1.1 Direct Democracy and the Adoption of Non-Discrimination Policies .......... 109 4.1.2 Event History Results .................................................................................... 113 4.2 Racial Profiling Bans ............................................................................................ 115 4.2.1 Direct Democracy and Racial Profiling Bans ................................................ 118 4.2.2 Event History Results .................................................................................... 119 4.3 Hate Crimes Laws ................................................................................................. 122 4.3.1 Direct Democracy and Hate Crimes Laws ..................................................... 125 4.3.2 Event History Results .................................................................................... 127 4.3.3 Hate Crime Laws, Sexual Orientation & Direct Democracy ......................... 129 4.3.4 Event History Results .................................................................................... 133 4.4 Discussion ............................................................................................................. 135 CHAPTER 5: CONCLUSION ....................................................................................... 137 5.1 Summary of Findings ............................................................................................ 138 5.2 Public Preferences & Direct Democracy .............................................................. 141 5.2.1 National-Level Issue Attitudes & Direct Democracy .................................... 143 5.3 Conclusion ............................................................................................................ 148 APPENDICES ................................................................................................................ 155 Appendix A: Adoption of Anti-Minority Policies ...................................................... 156 Appendix B: Coding of Direct Democracy Measures ................................................ 157 Appendix C: Principle Components Analysis of Direct Democracy Measures ......... 158 Appendix E: Keyword Search for Anti-Minority Bills, 1995-2004 ........................... 160 Appendix F: Adoption of Pro-Minority Policies ........................................................ 161 Appendix G: National-Level Public Opinion Surveys, Question Wording ................ 163 REFERENCES ............................................................................................................... 165 vii LIST OF TABLES 1.1 Direct Legislation Institutions ........................................................................................ 8 2.1 Ballot Measures Considering Same-Sex Marriage Bans ............................................. 34 2.2 Measures of Direct Democracy ................................................................................... 39 2.3 Correlations between Direct Democracy Measures ..................................................... 40 2.4 Principle Components Analysis ................................................................................... 41 2.5 State Adoptions of Same-Sex Marriage Bans, 1996 — 2005 ........................................ 49 2.6 English-Only Ballot Measures ..................................................................................... 58 2.7 State Adoptions of Official English, 1981-2005 .......................................................... 63 2.8 States with Affirmative Action Bans ........................................................................... 71 2.9 Initiative States vs. Non—Initiative States, 1996-2006 ................................................. 73 3.1 Anti-Minority Proposals by Targeted Group, 1995-2004 ............................................ 84 3.2 Passage Rates of Anti-Minority Proposals by Targeted Group, 1995-2004 ................ 85 3.3 Passage Rates of Anti-Minority Proposals, 1995-2004 .............................................. 86 3.4 Anti-Minority Proposal Passage by Targeted Group, 1995-2004. . . . . . . . . . . ....90 3.5 Passage Rates of Proposals Targeting Foreign Speakers ............................................. 93 3.6 Passage Rates of Anti-Minority Proposal by State, 1995-2004 ................................... 95 3.7 Determinants of the Passage of Anti-Minority Bills, 1995-2004 .............................. 100 4.1 Ballot Measures Addressing Sexual Orientation & Non-Discrimination .................. 110 4.2 Adoptions of Non-Discrimination Measures Protecting Sexual Orientation ............ 114 4.3 State Adoptions of Explicit Racial Profiling Bans, 1999-2005 ................................. 120 4.4 State Adoptions of Hate Crimes Laws, 1981-2005 ................................................... 127 viii 4.5 State Adoptions of Hate Crimes Laws that include Sexual Orientation .................... 134 5.1 National-level Issue Attitudes & Direct Democracy Effects ..................................... 145 5.2 Anti-Minority Policies Adopted from 1980 to 2007 .................................................. 150 5.3 Pro-Minority Policies Adopted from 1980 to 2007 ................................................... 151 Al Same-Sex Marriage Bans, 1995-2005 ..................................................................... 156 A2 Official English Laws, 1981-2005 ............................................................................ 156 D.1 State Adoptions of Affirmative Action Bans, 1996-2005 ......................................... 159 El Non-Discrimination Policies w/Sexual Orientation, 1982-2007 ............................... 161 F .2 Racial Profiling Bans, 1999-2005 .............................................................................. 161 F .3 Hate Crime Laws, 1981-2004 .................................................................................... 162 F4 Hate Crime Laws w/Sexual Orientation, 1984-2005 ................................................. 162 ix LIST OF FIGURES 1.1 State Adoptions of Direct Democracy Institutions ........................................................ 6 1.2 Policy Process under a Representative Democratic System ........................................ 15 1.3 Policy Process under a System with Direct Democracy Institutions ........................... 15 2.1 Initial State Adoptions of Same-Sex Marriage Bans ................................................... 31 2.2 Initial State Adoptions of Same-Sex Marriage Bans, 1993-2006 ................................ 32 2.3 Same-Sex Marriage Bans Passed by Initiative or Referendum ................................... 35 2.4 Estimated Hazard Rates ............................................................................................... 47 2.5 State Adoptions of Official English, 1981-2005 .......................................................... 56 2.6 State Adoptions of Official English Policies ............................................................... 57 2.7 Hazard Rates of Adopting Official English, 1981 ....................................................... 65 3.1: Anti-Minority Proposals Passed, Binary Comparisons .............................................. 87 3.2 Anti-Minority Proposals Passed, By Institutional Arrangement ................................. 88 3.3 Anti-Minority Proposal Passage, by Group & Institutional Arrangement .................. 91 3.4 Predicted Probabilities of Anti-Minority Bill Passage ............................................... 102 4.1 States with Non-Discrimination Laws that include Sexual Orientation .................... 108 4.2 State Adoptions of Non-Discrimination Policies, 1982-2007 ................................... 109 4.3 State Adoptions of Racial Profiling Bans .................................................................. 116 4.4 Yearly Adoptions of Racial Profiling Bans, 1998-2005 ............................................ 117 4.5 States with Hate Crime Laws ..................................................................................... 124 4.6 Yearly Adoptions of Initial Hate Crimes Laws, 1980-2005 ...................................... 125 4.7 States with Hate Crime Laws that Include Sexual Orientation .................................. 130 4.8 Yearly Adoptions of Hate Crimes Laws that include Sexual Orientation ................. 131 5.1 National-level Issue Attitudes .................................................................................... 144 CI Scree Plot ................................................................................................................... 158 xi CHAPTER 1: DIRECT DEMOCRACY INSTITUTIONS & THE THREAT OF TYRANNY In 2006, Michigan voters cast ballots to ban affirmative action programs in the state. The same year, citizens of Colorado voted to amend their constitution to prohibit recognition of same-sex marriages, while Arizona voters passed a constitutional amendment to require virtually all government actions to be conducted in English. Each of these policies restricts the rights of minority groups based on their race, ethnicity, gender, sexual orientation or national origin. Each policy was enacted through citizen legislation rather than through the traditional legislative process. Unfortunately for minority groups in the United States, these three states were not alone in passing policies that restrict minority rights through direct democracy processes like ballot initiatives and referenda. In all, twelve ballot measures in ten states addressed the rights of minority groups in 2006. Eleven of these measures resulted in an anti-minority outcome. These policy outcomes, and other similar outcomes from recent years, have raised concerns about the security of the rights of minority groups in states with direct democracy institutions. This study examines this concern through a series of analyses of contemporary policies that affect minority groups. So, does direct democracy endanger minority rights? Despite the recent attention to this question, it is not a new concern. Rather, this issue has long been at the center of the debate over governmental institutions that allow for direct citizen legislation, like the ballot initiative and popular referendum. Advocates of direct democracy contend that citizen legislation protects the public by allowing them to circumvent corrupt lawmakers that are beholden to special interests (Sullivan 1893). Although anti-minority outcomes are certainly possible under direct democracy institutions, it is argued that these outcomes are no more prevalent than under traditional representative democratic institutions (e. g. Commager 1958). The counter-argument asserts that mass participation exposes minority groups to potentially tyrannical policy outcomes favored by only a slim majority of the voters (i.e. fifty percent plus one). According to this perspective, minority rights are at increased risk under direct democracy institutions because they allow the majority to circumvent the checks and balances of a representative, separated powers system which is designed to encourage deliberation and minority representation. While empirical evidence has been presented that supports both sides of the debate (e. g. Gamble 1997; Hajnal, Gerber, and Louch 2002), this evidence has been limited in its scope and generalizability. In his recent examination of direct democracy and policy congruence Matsusaka (2004) writes: In short, we simply do not have compelling evidence yet on whether initiatives or legislatures pose a greater threat to minority rights, or even if there is a difference. The most defensible position is one of agnosticism: pending more evidence, we just don’t know. (p. 117) This dissertation seeks to address this “agnosticism” in the debate surrounding the impact of direct democracy on minority rights by taking a more systematic examination of contemporary policies that affect the rights of a variety of minority groups across the United States. This study expands beyond the previous literature by accounting for both the direct and indirect impacts of direct democracy on minority rights through examinations of policy outcomes from both citizen legislation and traditional legislation. It also expands upon the previous literature by directly comparing policy decisions in direct democracy states to decisions in non-direct democracy states while examining both anti-minority policies as well as pro—minority policies. The empirical results show strong support for a majoritarian effect of direct democracy, increasing governmental responsiveness to the preferences of the majority. In cases where the majority prefers policies that target the rights of political minorities, direct democracy is shown to have a negative impact on these groups. 1.1 Direct Democracy & the American States Before exploring the effects that direct democracy may have on minority rights, it is instructive to consider how direct democracy became such a prevalent form of policy making in the United States. A central tenet for architects of the newly independent nation and its states was that the authority and legitimacy to govern derived from the consent of the governed. At the same time, these revolutionaries were also concerned with protecting against various forms of tyranny stemming from “the ills of factions”. In designing a representative democratic system of government they were able to link the public, from which the govemment’s authority was derived, to policy makers through regular elections, and thus make elected officials sensitive to public preferences. More direct participation by the public was viewed as unwise and even dangerous to the fledgling republic. Apart from the ratification of constitutional amendments and state constitutions, the first century of governance in the United States was executed solely through representative democracy. By the end of the nineteenth century, however, dissatisfaction with government policies on a range of issues began to sow the seeds of a populist movement to increase public participation in government (Cronin 1989). In an era of rapid industrialization, corporate monopolies, and boom-and-bust cycles in both farming and mining, farmers, labors, miners and other “plain folk” advocated for policies such as free silver coinage, graduated income taxes, public ownership of railroads and single taxes. All of these policies sought to wrest economic power away from moneyed special interests, like trusts and corporations. Unfortunately for populist forces, both the national parties and the legislatures were often controlled by these economic behemoths. Party bosses and machine politics controlled much of the legislative process, especially at the local and state levels. In order to achieve their varied economic reforms, populists soon realized they would first have to pursue political reform. Populist theories of governance assert that government policy should reflect the “will of the people”. Furthermore, this public “will” or preference can be identified through a public vote (Haskell 2001). In short, populists equated the will of the people with the majority position of the voting public. Most farmers, laborers and single-taxers believed that their positions were supported by the majority of the people, and took government inaction on their issues as a sign of how government had been captured by special interests. With the public will effectively subverted by representative government, Populists (and later Progressives) sought to increase public participation in government as a way to force government to reflect the public will. The direct democracy tools of initiatives, referenda and recalls gained prominence in the Populist platform following the publication of J .W. Sullivan’s Direct Legislation by the Citizenship through the Initiative and Referendum in 1893. Sullivan had traveled to Switzerland and became enamored with their system of direct democracy which was modeled on the ancient tradition of Landsgemeinde. annual open-air meetings where all men of the canton would decide the policies of the local government (Schmidt 1989). In addition to Sullivan’s writings, other advocates of direct democracy also touted initiatives and referenda as institutions to circumvent unresponsive legislatures. Nathan Cree claimed that direct democracy would “break the crushing and stifling power of our great party machines,” (1892). Direct democracy reforms were soon viewed as the most viable means to achieve the populists’ varied economic policy goals. By the mid to late 1890’s direct democracy was also viewed as way to implement social policy as well. Women’s Suffrage groups and Prohibitionists soon took up advocacy of initiatives and referenda. What had once been a working class issue had now achieved more broad-based support as more and more groups saw direct democracy as an acceptable means to pursue their own policy goals. Despite the egalitarian rhetoric of the initiative and referendum movement, support for these reforms was also undeniably influenced by self-interest and more narrow policy preferences (Cronin 1989). Nonetheless, the Populist, and now Progressive, movement for “good government” through direct democracy spread throughout the country at the turn of the century, especially in the West and Midwest. The 1890’s saw the first adoptions of ballot initiatives in cities and states. South Dakota became the first state to adopt initiatives and referenda in 1898. Over the next twenty years, eighteen more states followed (see Figure 1.1). The adoptions of direct legislation institutions then ceased until 1959, when Alaska became a state. Following Alaska, four more states and the District of Columbia have also adopted some form of the initiative. As of 2000, these twenty-four states (and DC.) comprised almost half of the population of the United States. At the local level, fifteen of the top twenty most populous cities in the country had ballot initiatives (Matsusaka). Figure 1.1 State Adoptions of Direct Democracy Institutions Adopted _ 1892-1918 Tb ' Adoptedafter1918 D No Direct Democracy Though almost half of the American states allow for direct citizen participation in the policy process through some form of citizen legislation, the arrangements of these institutions varies across the states. There are three basic types of citizen legislation: direct initiatives, indirect initiatives and popular referenda. Direct initiatives are the most analogous to pure democracy; they completely bypass the legislature. Under these institutions citizens can draft policy proposals, petition to place them on the ballot, and then vote to accept or reject the policy. Indirect initiatives are similar, except that they must be submitted to the legislature for consideration before they are placed on the ballot. Popular referenda, meanwhile, do not originate with the citizens. Rather, citizens can petition to have specific, enacted piece of legislation placed on the ballot for the people to reject or accept. In addition to these three basic types of citizen legislation, states also vary in the rules and regulations that concern their direct democracy provisions (see Table 1.1). Some states allow citizen legislation for both statutes and constitutional amendments. Others can only use these mechanisms for one or the other. In terms of content, most states have a single subject rule which limits logrolling phenomena that could occur through the bundling of multiple policies on a single ballot measure. Many states also have firrther content restrictions. For example, Alaska does not allow ballot measures that concern appropriations or revenues. In Nebraska, the same subject cannot appear on the ballot more than once in three years. Wyoming does not allow ballot measures concerning the judiciary. Another area of variation is in the petition requirements. Massachusetts only requires the signatures of three and a half percent of the electorate from the preceding gubernatorial election, while Maine requires ten percent. Most states base their signature requirements on votes cast in the most recent gubernatorial election, which usually has high turnout, but others use the previous general election, whose turnout can vary depending on the races at the top of the ballot. North Dakota’s signature requirements, meanwhile, are based on the resident population. Signature requirements can be further complicated with geographic distribution rules. For example, Alaska requires at least one signature from two-thirds of the electoral districts in the state. Table 1.1 Direct Legislation Institutions Single Content Signature Reggirements State Type Subject Restrictions Statutory Constitutional Alaska I, R A A 10* -- Arizona D, R A 10 1 5 Arkansas D, R 8* 10* California D, R A 5 8 Colorado D, R A 5 5 Florida A --- 8* Idaho D, R A 10* -- Illinois D A --- 8 Maine I, R 10 --- Massachusetts I, R A A 3 + ‘/z*"‘I 3* Michigan D, l, R A 8 10 Mississippi I A --- 12* Missouri D, R A A 5* 8* Montana D, R A A 5* 10* Nebraska D, R A A 7* 10* Nevada D, I, R A A 10* 10* North Dakota D, R A A 2b 4b Ohio D, I, R A A 3 + 3*a 10* Oklahoma D, R A 8 15 Oregon D, R A 6 8 South Dakota D, R 6 8 Utah D, I, R A 5 + 5*a -- Washington D, I, R A 8 -- Wyoming I, R A A 15* -- Notes: D indicates direct initiatives, I indicates indirect initiatives, R indicates popular referenda; Signature requirements are a percentage of a previous election unless other wise indicated * Indicates a geographic distribution requirement ‘ The first figure is the percent needed for a proposal to be referred to the legislature and the second figure is the additional percent needed to place the measure on the ballot if the legislature does not pass it. A b North Dakota’s signature requirement is a percentage of the resident population. These variations in the institutional arrangements are not trivial. They create minimum thresholds of support that can deter or encourage citizen lawmaking. The difficulty in placing a policy proposal on the ballot can affect how much citizen legislation is used in a state, and thus can affect its impact on public policy. The institutional arrangements of direct democracy can also have an impact on the power of the legislature to influence policy relative to its citizens. Bowler and Donovan (2004) have shown that both qualification difficulty and the insulation of the legislature in direct democracy states have significant effects on policy outcomes. 1.1.1 The Impact of Direct Democracy Undoubtedly, the adoption of direct democracy reforms has added new features to the policy process in many of the American states. By allowing direct citizen participation in the state-level policy process, the direct democracy reforms of the past century have most certainly also altered the policy outcomes in these states. This influence on state policy outcomes has important consequences for public policy in the United States. Scholars have noted the resurgence of the American states in policy prominence and institutional capacity in the past few decades (e. g. Mooney 2001; Weber and Brace 1999). Recent trends in federal policy, like the devolution of programmatic responsibility of Medicaid and welfare to the states, have further increased the significance of state-level policy in the lives of American citizens. The upshot of all of this is that direct democracy reforms have influenced important policy outcomes across the American states. The question that follows is: how exactly have these policies been affected? It’s true that populist and progressive groups pursued direct democracy reforms like the initiative and referendum as a means to enact their specific policy goals, but they also saw these institutions as a way to make policy in general more reflective of the public will. While several states used direct democracy to pass progressive reforms like the prohibition of liquor sales and women’s suffrage, other policy areas, from taxes to abortion, have also been addressed through ballot initiatives and referenda. Since 1980, governmental reforms, like term limits and campaign finance laws, and tax policies have been the most common issues addressed by ballot initiatives. Social and moral policies, as well as environmental policies, have also been common subjects of recent ballot initiatives (Bowler and Donovan 2008; Tolbert, Grummel, and Smith 2001). So while direct democracy reforms have certainly served the specific policy goals of the progressive and populist movements, the adoption of citizen legislative institutions have also had a broader effect on virtually all state policy areas. Still, it is not clear that these reforms have brought policy outcomes closer to the public will. If we assume that the majority policy preferences do accurately reflect the general will of the public, then these reforms may be considered successful. Indeed, empirical studies have found that, for many policy areas, outcomes in direct democracy V states are more reflective of the majority preference than in non-direct democracy states (Arceneaux 2002; Bowler, Donovan, and Tolbert 1998; Matsusaka 1995; Matsusaka and McCarty 2001). Matsusaka (2004) shows that, in general, initiatives do produce policies favored by the majority of the population. By circumventing the legislature, initiatives allow citizens to pass policies favored by the majority without obstruction from the legislature or special interests. Research has also shown that initiatives can have an indirect effect on polices enacted by the legislature (Gerber 1996; 1999; Matsusaka and McCarty 2001; Romer and Rosenthal 1979). With the threat of citizen legislation, policies passed by the legislature in direct democracy states tend to be more congruent with the preferences of the majority. 10 Although recent academic studies have supported the claims of the early direct democracy advocates in terms of increased governmental responsiveness to the majority, most of this research does not provide guidance on the normative nature of these effects. Is this type of responsiveness beneficial to American democracy? Are there negative consequences to these reforms? Critics of direct democracy institutions have long contended that a government that is hyper responsive to the public may be worse than a government that is less responsive. From this perspective, the mass public is considered to be a relatively ignorant, irresponsible and capricious group (Cronin 1989). Responsiveness to the masses would only produce rash and unwise policies. Another problem that opponents of direct democracy point to is the assumption that the public will (as defined and identified by populists) equates to the public interest. If the public will is simply the preference of the majority of the public, then minority interests and preferences will be routinely overlooked, and even intentionally targeted. By making government responsive only to the preferences of the majority, minority rights could be put at risk. 1.2 Direct Democracy & Minority Rights: Tyranny of the Majority? Concern about the rights of minorities in democratic societies extends far beyond the direct democracy debate sparked by the p0pulists and progressives early in the twentieth century. In designing America’s democratic institutions, the framers of the Constitution frequently cited the protection of minority rights as justification for their form of representative government. James Madison noted the importance of guarding “one part of the society against the injustice of the other part,” and warned that, “If a ll majority be united by a common interest, the rights of the minority will be insecure.” ([1787] 1999.). He further argues that pure democracy, in which citizens participate directly, cannot cope with the ills of factions because there is no check on the power of the majority to rule at the expense of minorities. With these concerns in mind, Madison vigorously opposed many forms of direct citizen participation, from citizen legislation to the direct election of Senators and the President, and offered representative government as an alternative governmental design that would better protect minority rights. Thus, the mischief of factions was to be thwarted by the filtering processes of representative democratic government. The F ederalists argued that raw public opinion could be “refine[d] and enlarge[d]. .. by passing through a medium of a chosen body of citizens” (Madison [1787] 1999). This representational filter is designed to work by emphasizing deliberation, compromise, and consensus building. Although the partisan politics of contemporary American politics may mitigate some of these filtering mechanisms (e. g. Cox and McCubbins 1993; 2007) to some degree, the legislative process nonetheless continues to stress deliberation, compromise and consensus building in its design. Obviously, legislatures allow minority groups to obtain some degree of representation the policy process, as well as a voice in debate via their elected representatives. In addition, the process itself can create a bargaining environment conducive to cooperation and moderation. Bills face a daunting gauntlet of obstacles on their way to passage. At each point, from committee mark-ups, to bicameral conference reports, legislation can be changed and refined in order to build the necessary consensus for enactment. In general, a fairly large coalition of support is essential to pass anything 12 through this process. Building a large legislative coalition provides incentives to moderate or compromise in order to attract the requisite support for passage. Thus, the traditional legislative process allows for ample consideration of interests on both sides of an issue, which should help to ensure minority representation, and also provides incentives to compromise and moderate (Bell 1978; Cain and Miller 2001). In addition to allowing for minority representation through the electoral connection, representative democracy also provides organized minority groups multiple access points from which to promote their interests in a way that citizen legislation does not allow for. This is not to say that organized interests do not play a significant role in policyrnaking under direct democracy institutions. Studies have shown that interest groups may actually benefit from these institutions (Boehmke 2002; 2005; Smith and Tolbert 2004). Rather, reducing meaningful access by political minorities to the policy process can limit the advancement of their interests. Furthermore, the dominance of citizen groups that mobilize around moral and social issues in direct democracy processes can actually exacerbate the problem of protecting minority rights (Gerber 1999). Another aspect of the representational filter is that the legislative process places a premium on relationship-building. Legislative decision-making is not a one-shot game. Instead, legislators work with each other again and again across a myriad of issues and policies. It would be ill-advised for legislators to completely shut out their minority colleagues on one issue since they may be needed for consensus on another issue. Thus, the legislative process, itself, provides both opportunity and incentives for the kind of deliberation and compromise that should help protect minority rights. In circumventing this representational legislative process, direct democracy affords little opportunity for 13 minority voices to be heard and creates more rigid legislation that requires far less consensus, especially from interested minority groups (Cain and Miller 2001; Eule 1990). Direct democracy also provides opponents of minority rights a relatively easy way to achieve their goals by quickly expanding of the scope of conflict from the legislative arena to the public forum (Haider-Markel and Meier 1996; Schattschneider 1960). Compared to a ballot measure, where millions of citizens may participate, the relatively small arena of traditional legislation allows minority voices to be part of the deliberation and debate. As such, minority groups can often contribute to the policy debate through the filtering processes of representative government. However, in expanding the scope of conflict to the mass level, majoritarian preferences can easily overwhelm any consideration of minority rights. Figures 1.2 and 1.3 illustrate the majoritarian effect of direct democracy through very simplified diagrams of the legislative process. Figure 1.2 shows the policy process in a state with a traditional, representative democratic system. Public preferences (among other inputs) are filtered through representative democratic institutions. The filtering mechanisms of deliberation, representation, and coalition-building all help protect the rights of minorities, albeit imperfectly. Thus, the representative system generates policy decisions that are products of all public preferences and the filtering mechanisms. Figure 1.3, meanwhile, shows the policy process in a state with direct democracy institutions. Under this system there are two paths to reach policy decisions. First, the representative democratic institutions can set policy through a process similar to the one shown in Figure 1.2. Second, citizens can circumvent the legislature and its filtering mechanism and create policy directly. In the second path, the policy outcome is wholly l4 dependent on the preferences of the majority since policy decisions on this path are determined by a plurality vote. This second path is the direct impact of direct democracy on policy decisions where majority preferences are transferred directly into policy decisions. Fi are 1.2 Policy Process under a Representative Democratic System . Legislative Filtering PUDIIC Preferences ‘ Mechanisms Policy Decisions figure 1.3 Policy Process under a System with Direct Democracy Institutions . Legislative Filtering . . . Public Preferences ‘ Mechanisms ‘ Policy Deasrons Indirect Effect: Signaling & Credible Threats M. . eJer'tY Preferences ‘ my 0...... Direct Effect: Ballot Initiatives & Referenda In addition to the direct impact ballot initiatives and referenda can have on policy decisions, their potential use by the public also has indirect effects (as shown by the arrow connecting majority preferences to the legislative filters). This indirect impact works by altering the expectations that legislators have about the policy process and 15 subsequently altering their strategic decision-making (Gerber 1996; Romer and Rosenthal 1979). The influence of direct democracy institutions is felt by legislators in two ways: it can send clear signals about the policy preferences of the majority and it can serve as a credible threat to enact a particular policy. By sending clear signals as to the preferences of the majority, legislators will be pressured to adopt policies closer to these preferences. As an alternative policymaking path, citizen legislation allows the majority to credibly threaten to enact extreme policies. Legislators may respond to this threat by preempting the citizen legislation and passing a more moderate version of the policy. Either way, the use or threatened use of ballot initiatives and referenda forces legislators to be more cognizant and responsive to the majority. Legislators may also move to preempt citizen legislation in order to claim credit for popular policies as well as to mitigate more extreme policy decisions. In altering the behavior of legislators to make them more responsive to majority preferences, this indirect effect of direct democracy also reduces the access that organized minority groups might otherwise have in the traditional legislative process. This indirect impact may serve to undermine the influence of all types of organized interests in the legislative process, including minority interests. Further, this deflation of lobbying power should also extend to the executive agencies that implement policy passed by the legislature. Studies have shown that bureaucracies can serve a vital representational role for minority groups (6. g. Keiser and Meier 1996; Meier 1984; 1993; 1975). Again, without this additional access point, minority interests should be less well represented in states with direct democracy institutions as legislators become more responsive to majority interests. , 16 So, in both the direct and indirect paths of influence that direct democracy can wield, the impact of the preferences of the majority is magnified. The result of this magnified impact is a general majoritarian effect on policy outcomes. Empirical research supports this majoritarian impact, showing that states with initiatives are, indeed, more responsive to public opinion than states without direct democracy institutions (Gerber 1996; 1999). This responsiveness may be beneficial to the public in many policy areas, but in states where minority rights are targeted by the majority this induced policy congruence can be problematic. As Gerber and Hug (2001) stress, citizen legislation by itself doesn’t produce tyrannical outcomes, but coupled with anti-minority public preferences it can be detrimental to the civil rights of minority groups. When the majority preference is at odds with the rights of minority groups, these rights are exposed. One key assumption of the tyranny of the majority argument is that the majority will tend to prefer policies that restrict the rights of minorities; that direct democracy has a negative impact on minorities because majorities often do unite against minority groups. In the cases where this is true, the rights of minorities should be better protected under traditional legislative systems (Figure 1.2) than under systems with direct democracy institutions (Figure 1.3). Thus, the tyranny of the majority argument is a special case of the more general, majoritarian effect of direct democracy institutions. So theoretically, there are compelling reasons to expect that direct citizen legislation would put minority rights in danger — especially when the majority prefers policies that restrict minority rights. However, there also sound arguments that direct democracy can avoid the potential negative impacts on minority groups. Gillette (1988) argues that motivations that draw citizens to the voting booth are usually not narrowly l7 focused on a single issue. Thus, voter behavior will be shaped by multiple considerations rather than a singular focus on their ethnic, racial or other group interests. Empirically, however, it is not clear the extent to which these multiple motivations actually protect against tyranny of the majority. Another potential check on tyrannical outcomes is found in the judicial branch. Ellis (2002) notes the significant role of judicial review of citizen-passed legislation. Eule (1990) stresses that an active judiciary is a vital check on the majoritarian tendencies of direct democracy. Empirical studies have confirmed this role of the judicial branch, finding that the courts are, indeed, actively reviewing and overturning unconstitutional public initiatives (Miller 1999; Qvortrup 2001). Still, as a relatively passive player in the policy process, it is not apparent the extent to which this institution can protect the rights of minorities against a united majority. Courts may be less willing to take counter- majoritarian action and overturn policies passed directly by “the people” than policies adopted by the legislature (Grodin 1988). Furthermore, overturning citizen legislation is particularly difficult in states that select judges through popular elections. In these states, judges may be reluctant to overturn policies enacted by the same electorate they will have to face in order win reelection (Manweller 2005). Despite these potential checks on tyrannical outcomes under direct democracy institutions, theories of filtering through representational democracy, separated powers, and of the scope of conflict still predict that minority rights would suffer under governmental systems that allow citizen legislation. While the courts and the multiple motivating factors that draw voters to the polls both certainly have the potential to check anti-minority policy outcomes, neither directly refutes the theoretical underpinnings that 18 underlie the majoritarian argument. Courts can only act to protect minority rights after a policy has been enacted, and thus the majoritarian tendencies of direct democracy systems should not be altered by an active judiciary. The argument concerning the multiple motivating factors of voters also does not refute the logic of the tyranny of the majority argument. Just because voters are not motivated to vote by a single issue, does not mean they will not cast their ballot for policies that restrict the rights of minority groups. The motivations of voters also do not address the issue of the indirect effect of direct democracy, where just the threat of citizen legislation creates incentives for legislators to be responsive to the majority. Thus, the tyranny of the majority argument remains theoretically sound, and we should expect that states with direct democracy institutions should be more likely to threaten and restrict minority rights. Assuming that the majority of the public will often prefer policies that restrict the minorities, the impact of direct democracy on policy decisions should be to increase responsiveness to these majority preferences. Through both the direct and indirect paths 6f influence, the total effect of direct democracy should have a negative impact on minority rights. 1.2.1 Whose Rights? Throughout this discussion, the terms “minorities” and “minority rights” have been used without a clear explanation of exactly which groups they refer to. In the most basic sense, a minority group is any group of citizens defined by some characteristic, such that the group constitutes less than fifty percent of the population in a political system. This broad definition encompasses a huge range of groups distinguished by any 19 number of characteristics. Traditionally, minority groups have been defined by race, ethnicity and religion. However, other characteristics have also come to define important minority groups in American politics. Indeed, the writings on the factions in the Federalist Papers are aimed at groups defined by their social class and economic characteristics. Madison was concerned with the rights of wealthy elites at the hands of the newly empowered lower class citizens who comprised the majority of the American population. At the most abstract level, almost any characteristic can define a minority group. For example, both vegetarians and vegans constitute minority populations in the United States, but dietary choice is not usually considered to be a defining characteristic of minority groups whose civil rights need protecting. Although vegans may be a numerical minority, their distinguishing characteristic is not one that is typically considered in policy and political discussions. A narrower, more practical definition focuses less on the numerical aspects of minorities, and more on their rights vis-a-vis a politically dominant majority group. Using the term “political minorities”, this perspective defines minorities as groups that are “subjected to social, political and economic discrimination in society” or have been historically subjected to different legal standards than the majority (Kittilson and Tate 2005). For the purposes of this study, the concept of “political minorities” that focuses on discrimination and political rights is most appropriate. In evaluating the whether direct democracy has a negative impact on minority groups, it is necessary to concentrate on the rights of groups that are or have been targeted by the majority. It is exactly these groups whose rights are most endangered and have been widely recognized as needing political 20 protection. A more concrete inventory of the characteristics that are most commonly recognized to define political minorities can be found in state and federal civil rights laws and anti-discrimination policies. These widely protected classes include minority groups defined by race, color, religion, ethnicity, and national origin. More recently other groups have also gained recognition as valid “political minorities”, including groups defined by their sexual orientation, age, gender, and disabilities. Throughout this study, the analyses focus on these political minorities and policy proposals that have an explicit impact on their rights. While there is certainly a wide range of policies that can affect the interests of minority groups, from environmental policy to tax'policy, most do not overtly target the rights of political minorities (Hajnal, Gerber, and Louch 2002). Still, there are a non-negligible number of policies that do have a direct impact on the rights of political minorities, like sarne-sex marriage bans and racial profiling laws. It is exactly these types of policies that critics of direct democracy are referring to in arguing that citizen legislation endangers minority rights.I In empirically assessing this argument, it is important that the tests remain focused on these explicit policies that affect the rights of widely recognized political minority groups. In examining on these types of policies, I am trying to concentrate on those issues that fundamentally impact minority rights, and not on those that more tangentially affect minority interests (Haider—Markel, Querze, and Lindaman 2007). Thus, the policies under examination here all deal with a minority group’s rights in terms of equality under the law, equal opportunity, and individual privacy. I It is also these policies that have been the focus of previous scholarly examinations of this question. So in order to speak directly to this line of research, it is important to assess similar issues. A full discussion of the existing literature follows in succeeding section. 21 1.2.2 Existing Research As discussed earlier, proponents of direct democracy certainly have some persuasive arguments that minority rights are not unduly threatened by citizen legislation, but the most compelling argument in this debate is based on empirical evidence (or lack thereof). General studies of direct democracy concede the potential for civil rights abuses, while noting a lack of empirical evidence to condemn direct democracy institutions as the culprit for impairing minority rights (Cronin 1989; Magleby 1984). More recent examinations of empirical record reach similar conclusions. Donovan and Bowler (1998a; 1998b) find that statewide initiatives targeting gay rights actually pass at a lower rate than other types of initiatives. Hajnal and his colleagues (2002) found that racial and ethnic minorities in California are no more likely to be on the losing side of an initiative contest than are the majority Anglo voters, and conclude that these minority groups are not being systematically targeted through the initiative process. In a study of direct democracy in Switzerland, where three-quarters of the referenda held worldwide were undertaken, Frey and Goette (1998) found evidence that comparatively few measures restricting minority rights have passed by citizen legislation. In general, these studies conclude that direct democracy institutions are not overly susceptible to anti- minority outcomes. At the same time, there is also empirical research that supports the argument hat minority rights are endangered under direct democracy institutions. In her analysis of anti-minority initiatives and referenda in American states and cities, Gamble (1997) finds that seventy-eight percent of these measures passed. This high passage rate dwarfs the thirty-three percent rate for the rest of the initiatives and referendums in the sample. 22 Other recent studies have also found evidence of a detrimental impact of direct democracy on minority rights. Miller (1999) examines restrictive measures passed by citizens in three high-use initiative states. Evidence of tyrannical outcomes is also found in studies of specific anti-minority policies, such as official English language measures (Preuhs 2005; Schildkraut 2001; Tatalovich 1995), affirmative action bans (Chavez 1998), and anti-gay rights policies (Haider-Markel, Querze, and Lindaman 2007). In all, the existing literature has produced a mixed picture of the impact of direct , democracy institutions on minority rights. One shortcoming of most of this research on direct democracy and minority rights, regardless of which side of the debate it supports, is that it only examines measures considered and passed through citizen legislation while omitting traditional legislation from the analyses (Gerber and Hug 2001). As discussed . earlier, direct democracy institutions have both direct impacts on policy decisions through citizen legislation, as well as indirect impact by influencing legislators’ behavior. In order to understand the full impact of direct democracy institutions on minority rights, it is necessary to examine both paths of influence by considering both citizen legislation and traditional legislation. Another limitation of the existing research is also rooted in the singular focus on policies passed by citizen legislation. At its heart, the “tyranny of the majority” argument is a comparative one. Madison ([1787] 1999) advocates for a representative form of government than can provide a “cure” for the ills of factions, whereas pure democracy is unable to cope with these issues. In other words, he argres that representative government can better protect the rights of minorities than a pure democracy. Thus, the most direct test of this argument would be to compare policy decisions in direct 23 democracy governments with those produced by representative governments. If direct democracy institutions do endanger minority rights, then governments with these institutions should be more likely to pass anti-minority policies than those governments that have purely representative systems. In their recent examination of the debate on direct democracy and minority rights, Haider-Markel and his colleagues (2007) contend that this comparative approach should employed be whenever possible, and is a key criteria for understanding whether minorities are endangered by direct democracy institutions. In order to directly test this comparative argument, while at the same time accounting for both the direct and indirect impacts of direct democracy, it is necessary to explore the determinants of the adoption of all types of anti-minority policies, whether they are passed through citizen or through traditional legislation in both direct democracy and non-direct democracy states. The few studies that have used a similar tact have found that direct democracy does contribute to negative outcomes for minority groups. In studies of state adoptions of Official English language measures, Schildkraut (2001) and Preuhs (2005) find that initiatives states are more likely to adopt as the number of foreign-bom population increases. Non-initiative states, meanwhile, show decreasing likelihoods of adoption as the foreigrr-bom population in the states increase. These findings are consistent with the theory that representative governments provide a filtering and representational function that helps to protect minority rights. On the other hand, direct democracy provides a way to circumvent these processes, leading to more majoritarian policy outcomes. 24 While the studies on Official English laws certainly provide evidence of the potential detrimental effects of direct democracy, these effects are contingent on a rising “threat-level” posed by the foreign-bom population. But it is not clear how well this . finding generalizes to other issue areas. These studies examine the rights of linguistic and ethnic minorities. Do these results extend to other types of minorities (i.e. based on race or sexual orientation)? Are less dynamic populations of minority groups also endangered by direct democracy institutions? How do direct democracy institutions impact the adoption of policies that explicitly protect minority rights? 1.3 Goals & Plan of the Study The goals of this study are to directly and systematically test whether minority rights are less protected in states with direct democracy institutions, and to provide a better understanding of the impact that these institutions have on minority groups. Again, this study will extend upon the previous research by examining policy outcomes from states with direct democracy institutions as well as states without direct democracy institutions. The analyses will also consider both the direct and indirect effects of direct democracy by including both outcomes from ballot measures and outcomes from traditional legislative means. Another extension in this research is that it examines multiple policies and minority group types in order to enhance the generalizability of the findings. Finally, this study departs from earlier work by also studying the outcomes of pro-minority policies under direct democracy systems. The plan of the study is as follows: The first analysis focuses on the adoption of specific, contemporary anti-minority policies in the American states. In this section I ask 25 whether direct democracy states are more likely to adopt these policies than states without direct democracy institutions. The second analysis will focus on individual anti- minority policy proposals, asking whether the filtering mechanisms of representative democracy are reducing the likelihood that these proposal become law. In this chapter, I compare the passage rate of anti-minority bills in states with different direct democracy institutions. These first two analyses focus on whether direct democracy increases the likelihood of policy outcomes that restrict minority rights. The final analysis will take a contrasting perspective, examining whether direct democracy also influences the consideration of policy proposals that would protect or enhance minority rights. Similar in design to the first analysis, this section examines specific contemporary policies that protect minority rights. Together these three analyses provide a more complete picture of the effect direct democracy has on minority rights. 26 CHAPTER 2: DIRECT DEMOCRACY & THE DIFFUSION OF AN TI-MINORITY POLICIES One way to evaluate the majority tyranny theory is to examine the diffusion of anti-minority policies across the American states. If direct democracy does endanger minority rights, then states with ballot initiatives and referenda should be more likely to adopt an anti-minority policy than states without direct democracy institutions. With the ability to circumvent the representational filters of traditional democratic government, a united majority should be better able to enact their preferred policy with little resistance from minority groups. In this chapter, I examine the spread of three contemporary anti-minority policies to provide answers to the following question: Are direct democracy states more likely than non-initiative states to adopt a specific anti-minority? The policies under examination here are same-sex marriage bans, Official English laws and affirmative action bans. While the normative nature of these policies is certainly open to debate, each clearly targets a minority group and restricts their rights in some way. Although these three policies all target minority groups, the group that is targeted varies across the policies. Same-sex marriage bans target homosexuals, Official English laws target foreigrr-speakers (usually of Hispanic or Asian descent), and affirmative action bans broadly target racial and ethnic minorities. The variety of groups affected by these three policies should help generalize the findings beyond the specific policies in question. 27 2.1 Same-Sex Marriage Bans The 2004 elections notably included 13 state ballot measures proposing to ban same-sex marriage. Much ink was spilled discussing the impacts these ballot initiatives may have had on voter turnout and vote choice in a highly contested presidential election (e. g. Hillygus and Shields 2005; Lewis 2005), but relatively little attention was paid to the effects these direct democracy measures had on gay rights policies in the American states. In the end, all thirteen constitutional amendments passed, adding Ohio and Oregon to the list of states where same-sex couples are barred from marrying and reinforcing the existing statutory bans of the eleven other states. Large majorities in the thirteen states voted to restrict the rights of the minority gay population their state. So while these ballot initiatives may have spurred the turnout of evangelical voters and pushed “moral values” to the front of voters’ minds in making their presidential choice, the direct democracy tools used in this election also fundamentally changed policy towards homosexuals in these states. On the one hand, the 2004 elections provide a good example to illustrate the tyranny of the majority theory: Thirteen anti-minority policies were proposed and passed with only futile resistance from the targeted minority. On the other hand, only two of these states were enacting their initial same-sex marriage ban, and thirty-eight states before them had already adopted similar policies. Indeed, most same-sex marriage bans have been enacted with traditional legislation. So what exactly has been the role of direct democracy in the adoption of same-sex marriage bans in the American states? This section examines this question from an event history perspective, examining a state’s risk of adopting a same-sex marriage ban in any given year. This approach will evaluate the 28 total effect of direct democracy on the adoption of this anti-minority policy by analyzing policies passed by citizen legislation and traditional legislation while also accounting for other known determinants of policy adoption and diffusion. 2.1.1 The Politics of Same-Sex Marriage Same-sex marriage in the US. has been arguably the most visible issue in the gay rights movement over the past decade. While this issue has been part of the gay rights debate since same-sex couples in Wisconsin and Minnesota were denied the right to marry in 1971, the issue did not shift from the courts to the legislatures until the 1990’s. Early court rulings refused to recognize same-sex marriage, but in the absence of laws that explicitly banned gay marriage judicial decisions often relied on biblical references.2 It was not until 1993, when the Hawaii Supreme Court ruled that prohibiting same-sex couples from marrying may violate the Hawaii Constitution's ban on sex discrimination, that the issue reached the legislative arena and attracted national concern. Alarmed by the prospect of legalized same-sex marriage, opponents across the country moved to pass laws that banned recognition of gay marriages and similar unions. At first these efforts bore little success, but when the federal government enacted the Defense of Marriage Act (DOMA) in 1996 the states soon followed in passing their own versions of the law. While the DOMA is certainly notable for its definition of marriage as existing only between a man and a woman for the purposes of federal law, its most significant impact stems from the exemption of state marriage laws from the “full faith and credit” 2 E.g. Minnesota Supreme Court in 1971, Arizona Supreme Court in 1975 - for more detailed information, see Kranz and Cusick (Kranz, Rachel, and Tim Cusick. 2005. Gay rights. Rev. ed, Library in a book. New York, NY: Facts on File). 29 clause of the Constitution. States are constitutionally required to recognize legal contracts from other states, but the passage of DOMA excused states from recognizing marriages issued by other states. Prior to the DOMA, a state like Utah, which banned same-sex marriage in 1995, would nonetheless be required to recognize same-sex marriages conducted in other states where these contracts were legal. For this reason, the “full faith and credit” clause undermined that ability of states to set their own policy regarding the recognition of gay marriage. With the federal DOMA clearing the way for state legislation and stimulating a form of vertical diffusion (Welch and Thompson 1980), states across the country began considering and adopting what have come to be called “mini DOMAs.” Figure 2.1 shows the timeline of states’ initial adoptions of same-sex marriage bans from 1996 to 2005.3 Sixteen states enacted laws banning same-sex marriage in 1996. Nine more states followed in 1997.4 After the initial rush of legislation the flow of statutes barring recognition of gay marriage slowed. This pattern of yearly adoptions, as well as the cumulative adoption pattern (both shown in Figure 2.2), fit nicely with traditional models of the diffusion of innovation (Gray 1973; Walker 1969). The cumulative adoption line shows an “S-shaped” curve where a few early adopters (in this case, just Utah) are followed by a flurry of adoptions in the next couple years as state policymakers emulate and take cues from one another. The line then flattens as the last few states adopt their own policies. 3 The timeline shows each state’s initial adoption, not subsequent constitutional amendments, because it is the initial adoption that serves to change the status quo policy and restrict the rights of homosexuals. Constitutional amendments passed afterward simply reinforced the existing policy and made it more difficult to overturn, but did not substantively change those states’ policies toward same-sex marriage. 4 A table showing the dates of all the states’ initial adoption of same-sex marriage bans is presented in Appendix A. 30 Figure 2.1 Initial State Adoptions of Same-Sex Marriage Bans AK. AZ. DE: GA, '0, IL, Ks ’ AL, H'tt’ IA, CA, CO, NE*, NV" NH OH* OR* MI, MO, NC, KY WA . wv ' ' 0K, PA, sc, ’ so, TN \ \. / / ./ 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 -J... ‘ / 1 MN, MS, MT, ND, LA, VT TX* CT A *Initial ban was passed by constitutional amendment. "Technically, Hawaii amended its constitution to allow the legislature to pass a same-sex marriage ban, but did not explicitly ban same-sex marriages with theamendment. The legislature subsequently passed a statute banning same-sex marriage that same year. Although new state adoptions slowed to a standstill by 2001, events in 2003 and 2004 soon returned the issue to national attention. In 2003, the US. Supreme Court, in Lawrence v. Texas, ruled that all thirteen remaining state sodomy laws were unconstitutional. That same year, the Massachusetts Supreme Court issued the first of a series of decisions ruling that the state constitution mandated marriage equality. The next year several local officials began actively pursuing recognition of same-sex marriages. San Francisco Mayor Gavin Newsom led this movement, granting over 4,000 same-sex marriage licenses. Local officials in New Mexico, New York, Oregon and Washington soon followed suit. With all the increased activism and swell of media attention, the debate over same-sex marriage spurred a backlash, and new legislative efforts were soon 31 undertaken. Ballots in thirteen states in 2004 included proposed bans on the recognition of gay marriage. All thirteen passed with at least of fifty-six percent of the vote. From 1995 to 2005, forty-two states had banned gay marriage either through constitutional amendments or by state law. Fifteen of those states had passed both a law and an amendment outlawing the recognition of same-sex marriage. Figure 2.2 Initial State Adoptions of Same-Sex Marriage Bans, 1993-2006 o _ v O (D co ‘1 .9 S a) *5 L. O _ (D N .0 E 3 z 9 - I ’4“ 7“ \x ,A—— a” ‘uA———A O ‘ *a- l I l l H 1 993 1 996 1 999 2002 2005 Year —0— Cumulative AdOptions - -Ar - - Yearly Adoptions Most states adopted their initial policies barring recognition of same-sex marriage by following the lead of the federal DOMA in 1996. Utah actually anticipated the federal passage of DOMA, and enacted its initial ban a year earlier in response to the expected decision of the Hawaii Supreme Court to uphold its 1993 ruling. While the influence of federal action in the diffusion of same-sex marriage bans may be fairly obvious, the role 32 of direct democracy is not immediately apparent. So, what was the role of citizen legislation in the spread of same-sex marriage bans from 1996 to 2005? Table 2.1 and Figure 2.3 shed some light on this question by showing every ballot measure that has proposed a ban on the recognition of same-sex marriage. Despite all the media attention to the thirteen states with bans on the ballot in 2004, only Ohio and Oregon were enacting their initial laws against same-sex marriage. The other eleven states were reinforcing previously passed statutes with constitutional amendments. Of the forty-five states that have adopted policies barring recognition of same-sex marriage since 1995, only seven used direct democracy institutions to pursue their initial prohibitions. The other thirty-eight states passed their initial bans through traditional legislative institutions. The use of direct democracy by twenty-one other states was aimed at reinforcing existing policy, not altering policy outcomes. In many cases these reinforcing policies were enacted without circumventing the legislature, and only utilized direct democracy as ratification requirement to amend their constitution. Thus, the majority of these policies were not initiated directly by citizens. Judging from the frequency with which citizen legislation has been used to ban same-sex marriage, it seems as if direct democracy may not have played a large role in states adopting these policies. However, the success of these ballot measures tells a different story. While citizen legislation only makes up fourteen percent of the initial prohibitions of same-sex marriage in the states, just one ballot measure addressing this issue has failed to pass.5 In fact, these types of ballot measures garnered an average of nearly seventy percent of the 5 In 2006, Arizona’s Proposition 107, an initiative to amend the constitution to ban same-sex marriage, was narrowly defeated. This amendment would have reinforced an existing statute banning same-sex marriage. 33 vote, showing strong public support for this anti-minority policy. So even though citizen legislation has not been the most oft-used vehicle to adopt same-sex marriage bans, it has been enormously successful when employed. Table 2.1 Ballot Measures ConsideringSame-Sex Marriage Bans State Year Type of Institution Vote % Initial Ban? Alaska 1998 Legislative Referendum 68.1 Hawaii* . 1998 Legislative Referendum 69.2 A California 2000 Initiative 61 .4 A Nebraska 2000 Initiative 70.0 A Nevada" 2000 Initiative 70.0 A Nevada" 2002 Initiative 66.9 A Arkansas 2004 Initiative 75.0 Georgia 2004 Legislative Referendum 76.2 Kentucky 2004 Legislative Referendum 74.6 Louisiana 2004 Legislative Referendum 78.0 Michigan 2004 Initiative 58.6 Mississippi 2004 Legislative Referendum 86.0 Missouri 2004 Legislative Referendum 70.7 Montana 2004 Initiative 66.6 North Dakota 2004 Initiative 73.2 Ohio 2004 Initiative 61 .7 A Oklahoma 2004 Legislative Referendum 75.6 Oregon 2004 Initiative 56.6 A Utah 2004 Legislative Referendum 65.9 Kansas 2005 Legislative Referendum 70.0 Texas 2005 Legislative Referendum 76.2 Alabama 2006 Legislative Referendum 81.2 Arizona 2006 Initiative 48.2 Colorado 2006 Initiative 55.0 Idaho 2006 Legislative Referendum 63.4 South Carolina 2006 Legislative Referendum 78.0 South Dakota 2006 Legislative Referendum 51.8 Tennessee 2006 Legislative Referendum 81.3 Virginia 2006 Legislative Referendum 57.1 Wisconsin 2006 Legislative Referendum 59.4 A *Hawaii amended their constitution to allow the legislature to pass a same-sex marriage ban, but did not explicitly pass a ban by itself. The legislature subsequently passed a statute banning same-sex maniage that same year. MNevada requires constitutional amendments to pass through the public initiative process twice before it becomes law. 34 Figure 2.3 Same-Sex Marriage Bans Passed by Initiative or Referendum .W/a' I). This brief review of the politics of same-sex marriage policies in the American states presents a mixed picture of the role of direct democracy in this policy area. On the one hand, most states have prohibited same-sex marriage through traditional legislative means. Even when citizen legislation is used, it tends to reinforce existing policy rather than initiating the policy change in the first place. On the other hand, anti-gay marriage initiatives have passed nearly ninety-seven percent of the time. Once on the ballot, it appears that the minority gay population in the state has virtually no chance of defeating these proposals. 35 2.1.2 Event History Analysis So far, it looks as if direct democracy may not be a significant driving force behind the relatively quick spread of same-sex marriage bans in the American states. At the same time, the use of citizen legislation does seem to be a highly effective tactic in pursuing state DOMA legislation. In order to get a better grasp on the effect of citizen legislation on a state’s propensity to adopt same-sex marriage bans, it is necessary to move beyond the direct impact of direct democracy and assess the indirect impacts as well. To effectively assess the total impact of direct democracy, it is necessary to examine policy decisions from both citizen legislation (the direct effect) and traditional legislation (the indirect effect). To that end, this analysis uses an event history approach to examine state adoptions. Event history modeling, alternatively known as survival or hazard analysis, is a good way answer the following question: Given that a state has not adopted a gay marriage ban in previous years, what is the probability that it will do so in that year? To address this question I have collected yearly data on forty-eight states from 1996 to 2005.6 The analysis begins in 1996 because this was the year that the federal government considered and passed DOMA, which paved the way for states to adopt similar marriage restrictions.7 Since some states have not yet adopted same-sex marriage bans, but certainly still have the potential to do so in the years ahead, this data is right- 6 Utah is not included because it adopts its initial gay marriage in 1995, before the passage of the federal DOMA. Nebraska is also excluded from the analysis because the models include controls for party effects and the state has a nonpartisan legislature. Including these states, by extending the analysis back to 1995 and/or dropping the partisan variables from the models, does not change the interpretation of the effects of direct democracy. 7 Although a few states had statutes with language that defined marriage as solely between a man and a woman prior to 1995, these statutes did not address recognition of marriages conducted in other states and thus are not directly comparable to the “mini DOMAs” passed in the past decade. The analysis does control for those states with predating language defining marriage in this way. 36 censored. For example, Wisconsin passed a constitutional amendment to ban same-sex marriage in 2006, but since the analysis ends in 2005 they are coded as not having adopted this policy yet. Another important characteristic of the data is its temporality. With yearly data, there are defined discrete time periods rather than continuous temporality. In order to deal with this right-censored, discrete data, I utilize a Cox Proportional Hazards model which is ideally suited to this type of data. By employing an exact- discrete approximation for tied cases (multiple adoptions in a year), the Cox model is equivalent to a conditional logistic regression model or a fixed effects model that can make interpreting the results more straight-forward. In addition to these advantages, the Cox model allows for analysis of the impacts of covariates without any assumptions about the distribution of the baseline hazard rate. Rather than specifying a particular form of the duration dependency, the Cox model has an unspecified, flexible baseline. hazard. In other words, since this study is primarily focused on the impact of citizen legislation when controlling for several other relevant covariates, and not the on duration dependency of state adoptions, the Cox model is soundly equipped to address the question of direct democracy’s role in the adoption of same-sex marriage bans.8 The dependent variable in this study is a dichotomous indicator of whether a state adopted its initial same-sex marriage ban in a given year. I do not differentiate between traditional legislation and citizen legislation, or between statutes and constitution amendments. The focus here is whether or not direct democracy states are more likely to adopt policies that ban same-sex marriage, not necessarily how the policy is enacted. 8 For more information on this model and many other event history models, refer to Box-Steffensmeier, Janet M., and Bradford S. Jones. 2004. Event history modeling .' a guide for social scientists. Cambridge ; New York: Cambridge University Press. 37 Once a state adopts its initial gay marriage ban, it drops out of the analysis. I confine the analysis to include only the initial same-sex marriage ban, since this is the policy outcome that functions to restrict gay rights. Subsequent reinforcements, often in the form of a constitutional amendment, simply make the existing policy harder to Overturn, but generally do not fundamentally alter the policy of the states toward same-sex marriage. The main explanatory variable for this study measures a state’s direct democracy institutions. The simplest way to measure direct democracy is a dichotomous variable indicating whether the state has direct citizen initiatives —— the most direct form of citizen legislation. Nineteen of the twenty-four direct democracy states have direct initiatives. While this measure has intuitive appeal in its simple interpretation, it does not account for other types of citizen legislation, such as indirect initiatives and referenda. It also cannot . account for other types of variation in the arrangements of the state direct democracy institutions. Thus, measures that do account for this variation should also be considered. One approach to address the institutional variation within direct democracy states, is Bowler and Donovan’s (2004) indices of legislative insulation and qualification difficulty (see Table 2.2).9 The legislative insulation measure ranges from zero to nine, and gauges the extent to which the legislature is isolated from the effects of citizen legislation. California is the least insulated, with institutional characteristics like the inability to modify initiatives, no fiscal restrictions on initiatives and no indirect initiatives. I have reversed the coding for this study, so that higher scores indicate less legislative insulation. The qualification index, ranging from zero to six, measures how difficult it is to qualify for the ballot in each state. Qualification restrictions include 9 Details of Bowler and Donovan’s coding of these indices are presented in the Appendix B. 38 geographic distribution of signatures requirements, the proportion of signature required, and substantive subject matter restrictions (among others). Oregon tops the Qualification Index list with relatively minimal requirements to place a proposal on the ballot. With the reversed coding in this study, higher scores indicate easier qualification requirements. Table 2.2 Measures of Direct Democracy Legislative Qualification Direct . . Initiative Use State Initiatives? '"fu'at'?" D'fficu'fy 1996-2005 ndex Index Wyoming 1 1 1 Maine 2 3 14 Massachusetts 2 4 1 1 Mississippi 3 2 1 Montana A 4 4 16 Nebraska A 4 3 12 Ohio A 4 5 8 Alaska 4 2 15 Missouri A 4 4 1 1 Florida A 5 3 16 Nevada A 5 3 17 Illinois A 5 3 0 Utah A 6 4 4 Washington A 6 4 36 Idaho A 6 5 6 Oklahoma A 6 4 4 South Dakota A 6 5 9 Colorado A 7 6 35 North Dakota A 7 6 11 Oregon A 7 7 56 Arizona A 7 4 19 Michigan A 7 5 9 Arkansas A 8 5 8 California A 9 6 67 *Based on Bowler and Donovan’s (Bowler and Donovan 2004) measures — higher scores indicate less insulation and less qualification difficulty (i.e. fewer restrictions on direct democracy) Another way to account for the differences in the states’ direct democracy institutions is to use a measure of initiative use (Pippen, Bowler, and Donovan 2002). States that use direct democracy more often tend to have easier qualification rules and 39 less legislative modification (Bowler and Donovan 2004). Hence, direct democracy should have larger impact on policy decisions in states the use initiatives more often. For this analysis, initiative use is operationalized as the number of recent ballot initiatives (from 1996 to 2005).10 Since each of these variables is measuring direct democracy institutions in a different way, it is not surprising that they are all highly correlated (see Table 2.3). To avoid statistical problems such as multi-collinearity, I conducted a principle components analysis. Based on this analysis, a single measure of direct democracy institutions in a state was created by combining the four measures of direct democracy weighted by the eigenvectors for the first component. In addition to avoiding statistical problems associated with the high correlations of these four measures, this approach also allows for a relatively parsimonious test of the impact of direct democracy on policy adoption while guarding against spurious results that might plague a binary measure. Table 2.3 Correlations between Direct Democracy Measures Direct Legislative Qualification Initiative Initiatives Insulation Difficulty Use Direct Initiatives 1 '000 Leg's'tt'v" 0.923 1 .000 Insulation Qualification Difficulty 0.861 0.938 1.000 '"'t'at"’° 0.851 0.924 0.9443 1.000 Use '0 In the analysis, the natural log of initiative use is employed to account for the diminishing effect that the number of initiatives has as the variable reaches higher numbers. 40 As shown in Table 2.4, the first component accounts for ninety-three percent of the variance in the four variables.11 The resultant measure ranges from zero, for the twenty-six states that have no direct democracy institutions, to 5.5 for California - a state with direct initiatives, low legislative insulation, easy qualification requirements and high initiative use.12 Given the expectations of how direct democracy institutions should affect public policy, this new measure can be interpreted as the relative impact of citizen legislation on policy decisions in each state. So, if direct democracy does, indeed, have a majoritarian impact on policy adoptions, then the event history analysis should produce a positive coefficient for the direct democracy impact variable - indicating that direct democracy states are more likely to adopt a same-sex marriage ban. Table 2.4 Principle Components Analysis Variables Eigenvectors Direct Initiatives * - ' 0.488 Legislative Insulation 0.509 Qualification Difficulty 0.503 Initiative Use 0.500 Eigenvalue Proportion 3.720 0.930 ” A scree plot, supporting the use of a single measure to account for the variance in the four direct democracy variables is presented in Appendix C. '2 The original scoring based on the principle components analysis creates a measure with a mean of zero. I adjusted the measure for easier substantive interpretation, so that non-direct democracy states would have a score of zero, rather than a negative score. 41 In addition to direct democracy impact variable, I also control for several other potential determinants of state policy adoption.l3 Clearly, public opinion can affect a state’s policy outcomes (e. g. Erikson, Wright, and McIver 1993; Hill and Hinton- Anderson 1995; Norrander 2001). To incorporate this factor into the model, I rely on the measure of citizen ideology developed by Berry, Ringquist, F ording and Hanson (1998; 2003).14 Based on the ideologies of elected officials and their electoral support, higher scores indicate more liberal citizen ideologies. I expect more conservative states to be more likely to adopt same-sex marriage bans. One problem with this variable is that the data have only been calculated up to 2002. To overcome this obstacle, a four year running-mean was used to extend the data to 2005. Erikson, Wright and McIver (2006), in extending their original analysis of state public opinion and policy liberalism, note that state ideology has been remarkably stable in the past twenty five years. Given this stability, using the running-means as a proxy measure of state citizen ideology should serve as an adequate indicator of public ideology for the years 2003 through 2005.‘5 Another important determinant of policy outcomes are interest groups (e. g. Brace 1988; Shipan and Volden 2006). Unfortunately, measuring the strength of interest group communities directly can be extremely difficult, especially when dealing with smaller advocacy communities like the gay rights community and its opponents. The '3 In addition to the explanatory variables presented here, other potential determinants of policy adoption were also included in alternate models. Variables measuring political culture, legislative professionalization, per capita income and region added little explanatory power to the model and did not alter the interpretation of the direct democracy coefficients. '4 I opt for this measure rather than the popular ideology scores developed by Erikson, McIver, and Wright (Erikson, Robert 8., Gerald C. Wright, and John P. McIver. 1993. Statehouse democracy .‘ public opinion and policy in the American states. Cambridge ; New York: Cambridge University Press.) because of the latter’s exclusion of both Hawaii and Alaska. Alternative specifications that use this survey-based measure of ideology show no appreciable differences. '5 A truncated analysis up to 2002 generates very similar results, but leaves an incomplete picture of the spread of same-sex marriage bans since many states adopted their initial bans after 2002. 42 characteristics of the same-sex marriage debate, however, do lend themselves to a more indirect approach. Generally, advocacy for DOMA legislation and gay marriage bans stems from the conservative Christian community (Haider-Markel 2001; 2000). This movement is lead by organizations that are based on public membership. Thus, it is possible to gauge the membership resources available in a state to conservative Christian groups by a simple count of the evangelical population in each state. The rate of evangelicals in a state population should be a good indicator of the strength of the conservative Christian movement in that state. This analysis uses the evangelical rates from the 2000 Religious Congregations and Membership survey.l6 States with higher rates of evangelical Christian membership should have a higher likelihood of adopting same-sex marriage bans. To account for interest group influence and mobilization on behalf of gay rights and against same-sex marriage bans, I take a similar approach. The organization and mobilization of gay rights groups began in large cities like Los Angeles, San Francisco and New York, and has largely been an urban movement. With more organizational resources and visibility in urban areas, gay rights groups should be more successful in metropolitan areas. Indeed, research has shown that Members of Congress with more urban districts tend to be more supportive of gay rights and tend to oppose gay marriage restrictions (Lublin 2005). Thus, I use the urban population rate as a proxy for the resources of the gay rights groups. States with higher rates of urban populations should be less likely to adopt same-sex marriage bans. '6 The data are available from Association of Religion Data Archives at EVELIIICHIKIQLUE and were collected by Association of Statisticians of American Religious Bodies (ASARB). 43 To control for the partisan influences on state policy making, I include two dichotomous variables indicating either unified Republican control of the state government or unified Democratic control of the state government. Since Republicans tend to be more ideologically conservative and enjoy support from the conservative Christian community, I expect Republican governments to be more likely than either Democratic or divided governments to enact same-sex marriage bans. Jointly these two party control variables also account for the differences between unified and divided government. All things equal, unified governments should be more likely than divided governments to pass any type of policy. I also include a measure of party competition. Since I have already accounted for party control of the government, I use a folded Ranney index.l7 Given the broad popular support for same-sex marriage bans, states with higher party competition should be more likely to adopt these measures as parties compete for public support and electoral success (Holbrook and Van Dunk 1993). To account for geographic diffusion impacts, I include a measure of the proportion of adjacent states that have passed same—sex marriage bans. The diffusion of innovation literature suggests that state will be more likely to pass new policies when neighboring states have already passed those policies (e. g. Berry and Berry 1990; Crain 1966). However, in the case of same-sex marriage bans, the geographic diffusion may have an opposite effect. One of the key aspects of the federal DOMA is that it allows states to ignore marriages from other states. If all the surrounding states have already '7 The original index is a measure of Democratic Party control of state government ranging from 0 (complete Republican control) to 1 (complete Democratic control). Its midpoint indicates evenly divided control or even competition between the parties. By “folding” the scale at its midpoint, the index of party control becomes an index of competition. Ranney, Austin. 1976. "Parties in State Politics". In Politics in the American States : a comparative analysis, edited by H. Jacob and K. N. Vines. Boston: Little, Brown. 44 banned same-sex marriage there may be less incentive for a state to adopt a gay marriage ban of its own. I also include several indicators of the legal environment pertaining to gay rights in each state. In particular, I control for whether or not a state has the following: a civil union or domestic partnership law, any criminalization of sodomy,18 a non-discrimination law that includes sexual preference or statutory language that predates the contemporary debate over same—sex marriage which defines marriage as only between a man and woman. Civil unions and legally recognized domestic partnerships are often viewed as equivalent substitutes to marriage,19 so states with these laws should be more likely to adopt a same-sex marriage bans. The presence of sodomy laws should also increase the likelihood of adopting a same-sex marriage ban, but for very different reasons. Sodomy laws are indicators of legal environments that maybe hostile towards gay rights. Non- discrimination laws that include sexual preference, meanwhile, should decrease the likelihood of enacting a same-sex marriage ban since these laws are indicative of a more tolerant legal environment. Finally, a few states have statutory language enacted well before the contemporary debate that describes marriage in such way that indicates that it is a heterosexual arrangement. For example, Wisconsin law states that “marriage is a legal relationship between two equal persons, a husband and wife, who owe to each other mutual responsibility and support...” (WISC. STAT. §765.001(2) (1979)). This type of language could be understandably be interpreted as defining marriage as only between a man and woman. Thus, these states should be less likely to adopt a “mini DOMA”. '8 Following the Lawrence v. Texas decision in 2003, all sodomy laws were struck down. '9 For example, Connecticut’s 2005 civil union bill was signed into law by Gov. Jodi Rell (R) only after an amendment defining marriage as between a man and woman was added. 45 The final two control variables are demographic variables.20 The first addresses the “scale of democracy” argument. Based on Publius’ treatment of the majority faction problem and Hamilton’s contention that majority tyranny could be constrained by the “enlargement of the orbit,” we should expect that larger (more populous) states should be less likely to adopt measures that restrict minority rights. Donovan and Bowler (1998a) find that pro-minority outcomes are significantly more likely in larger places than in smaller governmental arenas. Following their work, I include the natural log of the state population and expect a negative effect on state DOMA adoption. The other demographic control gauges the level of educational attainment in a state as measured by the percent of the population over 25 years old with at least a bachelor’s degree.21 Levels of education have been shown to be consistent determinants of tolerance of minority groups-(e. g. McClosky and Brill 1983). Thus, I expect states with relatively higher numbers of college graduates to be less likely to adopt same-sex marriage bans. As discussed earlier, one advantage of the Cox proportional hazards model is its assumption that time dependency has an arbitrary form. This assumption allows the baseline hazard to vary over time. Due to this aspect of the model, indicators of spurring events or other dichotomous measures of temporality cannot be included. Still, it is reasonable to expect that the court decisions in 2003 and 2004, and the willingness of local officials to grant same-sex marriage licenses in 2004 would spur states to adopt 20 I do not, however, control for race. Although popular accounts of gay rights issues have suggested that race may play a role in attitudes toward homosexuals, scholarly examinations have not found this to be the case (Herek, Gregory M., and John P. Capitanio. 1995. "Black heterosexuals’ attitudes toward lesbians and gay men in the United States". The Journal of Sex Research 32 (2):95.). Rather, religiosity, which can be easily entangled with race, is the driving force behind attitudes towards gay rights and homosexuality (Wilcox, Clyde and Robin Wolpert. 2000. "Gay Rights in the Public Sphere: Public Opinion on Gay and Lesbian Equality". In The Politics of Gay Rights, edited by W. Rimmerman, and Wilcox. Chicago: University of Chicago Press). 2' Educational attainment data drawn from the US. Census’ annual Current Population Survey: http://www.census.gov/population/www/socdemo/educ-attn.html. 46 bans. These temporal affects are evident when examining the estimated hazard functions. Figure 2.4 shows the hazard functions of states with direct initiatives and states without direct initiatives, and a considerable increase in the hazard rate in 2003 and 2004 is clearly visible, suggesting that these judicial events did have an effect of states’ propensity to adopt same-sex marriage bans. Figure 2.4 Estimated Hazard Rates dz- / \ / / // / .5 / '2 / g N r » // .C '0 i=3 // O // O ,/ E ‘— —-+ ’z’ (D /,/’ O-i *— l I l r If I 1 996 1 998 2000 2002 2004 2006 analysis time No Direct Initiatives ————— Direct Initiatives These hazard functions also reveal a substantial difference between the two types of states, providing some preliminary support for the argument that direct democracy increases the propensity to adopt same-sex marriage bans. However, without controlling for the myriad of other determinants of state policy adoption discussed above, more 47 evidence is required. The Cox model allows us to control for these determinants while factoring in the yearly variation in the baseline hazard rates. 2.1.3 Results The results from the Cox proportional hazards analysis are shown in Table 2.5. The coefficients can be interpreted similar to conditional logistic regression coefficients, as the effect of a one unit change in the independent variable has on the log-odds of a state adopting a same-sex marriage ban in a particular year. Consequently, initial interpretations of the results should focus on the direction of the coefficient and its statistical significance. The estimation results clearly support the tyranny of the majority argument. The direct democracy impact coefficient is positive and statistically significant, suggesting that direct democracy states (with an impact score greater than zero) are more likely to adopt a same-sex marriage than non-direct democracy states (with an impact score of zero). Furthermore, this result shows that direct initiative states with relatively high initiative use, less insulated legislatures and easy qualification requirements more likely than other direct democracy states to adopt this anti-minority policy. This positive coefficient meets the expectations of a majoritarian effect on policy adoption and has negative consequences for the rights of homosexuals. Moving down the model, most of the control variables also meet expectations. Citizen ideology has a significant negative coefficient, as hypothesized. On average, liberal states are less likely than conservative states to adopt same-sex marriage bans. The interest group resource variable coefficients are both in the expected direction, but 48 only the urban rate of the state seems to have a marginally significant effect. Still, higher rates of evangelical members, and thus higher resources for conservative Christian groups, tends to increase the likelihood of a state adopting a gay marriage ban in a given year. On the other hand, more urban states, those with more resources for gay rights groups, are less likely to ban gay marriage. Table 2.5 State AdoLtions of Same-Sex Marriage Bans, 1996 — 2005 Variable Coefficient Standard Error j Direct Democracy Impact [+] 0.406“ 0.125 Citizen Ideology H -0.056* 0.025 Evangelical Rate [+] 0.002 0.004 Percent Urban H -0.034# 0.022 Republican Government [+] 2.145" 0.692 Democratic Government [+] 1293* 0.660 Party Competition [+] 9.608" 3.765 Bordering States w/Bans -3.082** 0.935 Civil Unions or Partnerships [+] 2458* 1.074 Sodomy Law [+] 1.676“ 0.675 Non-discrimination Law H -0.616 0.794 Predating Language H -2.518* 1.220 Population (log) H -0.329 0.334 Educational Attainment {-1 .0157" 0.060 Observations 203 Log Likelihood -85.909 Notes: Expected direction of coefficients in brackets; one-tailed tests used where appropriate, two-tailed tests otherwise; # significant at 10%; * significant at 5%; ** significant at 1% 49 This analysis also reveals partisan effects on the propensity to adopt same-sex marriage bans. Republican controlled governments are significantly more likely than divided governments to adOpt these measures. Democratic controlled governments also show a positive effect on adoption propensity. On average, there is no statistically significant difference between the propensity of Republican controlled governments and Democratic controlled governments to adopt a same-sex marriage ban. Given the widespread support of same-sex marriage bans (see Table 2.1) it is not surprising that both parties would move to enact these popular policies. When considered jointly, as the difference between unified and divided government regardless of which party is in control, unified governments do show a significant positive effect on the likelihood of adopting same-sex marriage bans. Although unified government increases the probability of adopting these policies, party competition also has a positive impact. While this may seem like a contradictory statement at first, consider that a state can be very narrowly controlled by one party and still have high levels of party competition. In this case, party competition creates incentives for both parties to cater to the policy preferences of large majorities and enact popular policies like these “mini DOMAs”. The geographic diffusion variable also shows an effect that contrasts with the findings from previous research. As the proportion of neighboring states with same-sex marriage bans increases, the likelihood of a state adopting its own ban actually tends to decrease. As discussed earlier, the federal DOMA excludes marriage laws from the “full faith and credit” clause that forces states to recognize contracts in made in other states. This exclusion creates a motivation for states that view the recognition of same-sex marriages granted in other states as a realistic “threat” to pass their own bans. When 50 most of the surrounding states have already banned same-sex marriage, the “threat” of same-sex marriages from other states may be diminished and the incentive to pass their own policies is lower. The results also show that the legal environment affects a state’s likelihood of adopting a same-sex marriage ban. The substitutive effect of having civil unions or domestic partnership is clear. Those states that recognize same-sex arrangements that are roughly equivalent to marriage increase the odds of adopting a gay marriage ban by over eleven times in all three models. At the same time, a legal climate in which sodomy is criminalized increases the odds of that state adopting a same sex marriage ban by about three in each model. As expected, states with statutory language that can be interpreted to define marriage as only between a man and woman are less likely to adopt another .* redundant law banning same-sex marriage. Finally, the results show that states with more highly educated citizens are less likely to adopt same-sex marriage bans. As discussed earlier, more educated individuals tend to be more tolerant, and thus less likely to prefer policies that restrict rights of minority groups like homosexuals. In all, the results of this analysis support the hypothesis that direct democracy states are more likely to adopt an anti-minority policy like a same-sex marriage ban than states with purely representative democratic governments. While ballot initiatives have only been utilized by six states to enact their initial bans on same-sex marriage, the effect of direct democracy institutions seem to be far more pervasive. This direct effect of direct democracy institutions is also supplemented by an indirect effect. The presence of 51 direct democracy institutions increases the likelihood that a state government will adopt a same-sex maniage ban. 2.2 Official English The evidence presented in the previous section strongly supports the critique that direct democracy endangers minority rights. However, the case of same-sex marriage policy may be difficult to generalize to all anti-minority policies. The diffusion of “mini DOMAs” in the past decade has been supported by large majorities in most states and the policy has been adopted by almost every state in the union. So how will this critique hold up under different circumstances? In the section, I examine another contemporary minority rights issue: Official English. Whereas same-sex marriage bans have been adopted in almost every state, English is the official state language of only about half of the states. Official English policies often target larger minority groups based on ethnicity, language and race rather than lifestyle preference. Despite these differences, theoretically both issues target minority groups, and therefore should be more likely to be adopted in states that have direct democracy institutions. Like the previous analysis, this section will take an event history approach to examine the effect of direct democracy institutions on the likelihood of adopting this anti-minority policy while controlling for the myriad of other determinants of policy adoption. 52 2.2.1 The Politics of Official English Although English has been the most dominant and common language in the country, the United States has never had an official language. While there is evidence that the founding fathers considered the role of language in the formation of the new republic, the language debate did not produce an official declaration of a national language (Baron 1990). This seems to reflect the conclusion that an official language was not necessary for the fledgling nation. English was the de facto language of government and the vast majority of commerce, so it was assumed that it would continue to be the dominant language. At the same time, the country was populated by a diverse array of non-English speaking immigrants whose support was crucial to the new democracy (Schmid 2001). Indeed, many of the important political documents of the time, like the Articles of Confederation and the Federalist Papers, were printed in multiple languages, while the Constitution itself remained silent on the question of 1 language (Kloss 1977). It was not until the twentieth century that American governments began to address the issue of an official language. The first major spurt of language legislation was aimed mainly at restricting the German language during and immediately following World War 1. While this era saw the introduction of “liberty cabbage” instead of sauerkraut and “Salisbury steak” instead of hamburger, widespread language legislation did not flourish until the 1980’s (Crawford 1992). Two events in the Seventies precipitated the Official English movement in the Eighties. In 1974, the US. Supreme Court ruled in Lau v. Nichols that schools that did not provide non-English speaking students the opportunity to “participate meaningfully” were in violation of the 1964 Civil Rights Act, amounting 53 to discrimination based on national origin. In 1975, language provisions were added to the Voting Rights Act of 1965 that required bilingual ballots in communities where non- English speakers accounted for at least five percent of the population. Congress concluded that English-only ballots, coupled with English-only education, were excluding language and ethnic minorities from democratic participation (Schmidt 2000). Together these events were perceived to be a national policy of language pluralism. In a backlash to these policies and to the changing demographic makeup of the country, the contemporary Official English movement was born. It formally began with the introduction of a proposed Constitutional amendment to designate English as the sole official language of the United States by Senator S. I. Hayakawa (R-CA) in 1981. This amendment ultimately failed, but similar amendments have been introduced to Congress every year since then. Senator Hayakawa also co-founded the group US. English in 1983 to lead the Official English movement (Schmidt 2000). While Official English policies have not been very successful at the federal level, groups like US. English have had more success at the state level in the past twenty-five years. Though most state Official English policies are products of the contemporary English-only movement, five states do have policies that predate this period. These policies, however, tend to be more symbolic than firnctional. Louisiana, with its French heritage, officially adopted English as its language of record when it joined the Union in 1812, but retained its francophone rights as well. In the Courts, the legislature most business was conducted bilingually, reflecting the widespread use of French in Louisiana throughout the nineteenth century. Nebraska and Illinois were the next states to adopt an official state language policy in the 1920’s. In an era of “nativism” and “wartime 54 hysteria,” when anti-foreign and specifically anti-German sentiment was at its peak in America, Nebraska amended its constitution in 1920 to declare English the official state language (Tatalovich 1995)”. Illinois followed suit three years later, adopting “American” as its official language. The official language was later changed to English in 1969. Like Louisiana’s language policy, the last two states to adopt English as an official language prior to the contemporary movement also adopted measures that amounted to bilingual or pluralist policies. Massachusetts required English proficiency as a requirement for political participation in 1975, but still maintained a policy of bilingual education for immigrant children. Hawaii’s policy, enacted in 1978, is officially bilingual, recognizing both English and Hawaiian as the official state languages. At the state level, the contemporary Official English movement started with . Virginia’s adoption of its language'law in 1981. In what has been called “nativism reborn” (Tatalovich 1995), this movement has been propelled by negative responses to the growing Spanish speaking population in the United States. Figure 2.5 shows the timing of the centemporary adoptions of Official English. The mid-Eighties witnessed a spike in adoptions, with fifteen states enacting their policies by the end of the decade. A lull in the diffusion of Official English followed until the mid-Nineties. From 1995 to 2002 though, a second wave of adoptions occurred, with eight more states declaring English as the official state language. These two waves are apparent in the cumulative adoption line, as there are two distinct s-shaped curves. 22 The amendment originally declared American as the official language, reflecting the broad nativist attitudes of the time, but in the end English was chosen to avoid confusion. 55 Figure 2.5 State Adoptions of Official English, 1981-2005 to a N o - N (D .9 .9 ‘9 - m ‘— “5 s E 53 - :3 2 ID _l / >; / \\ it! \\ ._ /\ A o - 5’ h/Ka—a—A—l kfi/ \AA‘A’ArA—A-it I I I ' I ' l l 1 980 1 985 1 990 1 995 2000 2005 ’Year _. —O-— Cumulative --‘r-- Adoptions By 2007, twenty-nine states had declared English to be the official state language.23 Figure 2.6 shows the all of states that have an official language. Most of these adoptions occurred after 1980 during the modern Official English movement and have substantive impact on the state government’s use of language in its programs. By contrast, the five states that adopted an official language prior to 1980 have more symbolic language policies, as described earlier. 23 A list of yearly state adoptions is presented in Appendix A. 56 Figure 2.6 State Adoptions of Official English Policies , Since 1990 980 990 <2. W listen; i990 % *Arizona’s strict Official English was found to be unconstitutional and was overturned in 1998. "Hawaii’s language policy is officially bilingual, recognizing both English and Hawaiian. Like the same-sex marriage bans, most Official English laws have been passed through traditional legislation. Table 2.6 shows the English-Only ballot measures considered in the states since 1980. Of the twenty-three states that have adopted an Official English policy, only six used citizen legislation to achieve this outcome. At the same time, the pattern of Official English adoptions is also similar to the same-sex marriage bans in terms its success. Each of the six states using citizen legislation passed the measure; usually by a wide margin. Other English-only policies that have addressed education and voting issues have also met with comparable success. The lone exception 57 was Colorado’s initiative to move from bilingual education to English immersion education in 2002. Again, in examining only the outcomes from citizen legislation, it looks as if direct democracy may not increase the likelihood of a state adopting Official English. As shown in the previous section though, only focusing outcomes of citizen legislation can be misleading. Table 2.6 English-Only Ballot Measures Official English Measures State Year Type of Law Vote For Alaska 1 998 Statute 68.6% Arizona* 1988 Amendment 50.5% Arizona* 2006 Amendment 74.2% California 1986 Amendment 73.3% Colorado 1988 Amendment 61 .2% Florida 1988 Statute 84.0% Utah 2000 Statute 67.2% Other Language Policies State Year Subject of Law Vote For California 1984 Voting 70.7% California 1998 Education 60.9% Colorado 2002 Education 44.6% Massachusetts 2002 Education 68.0% *Arizona’s initial amendment was overturned in 1998 2.2.2 Event History Analysis Again, in order to fully assess the impact of direct democracy on the adoption of Official English laws it is necessary to account for both the direct and indirect effects by examining policy decisions from both citizen legislation and traditional legislation. To that end, this analysis follows the analysis in the preceding section and employs an event 58 history approach to answer the question: Given that a state has not adopted an Official English policy in previous years, what is the probability that it will do so in that year? In the case of Official English, data has been collected on forty-five states from 1981 to 2005.24 The analysis begins in 1981 because this is the year Virginia adopted its Official English law. This adoption was the first language policy enacted in response to a perceived federal policy of language pluralism (Tatalovich 1995). 1981 is also the year that Sen. Hayakawa (R-CA) first proposed an Official English amendment to the Constitution, propelling the issue to the national stage. Not unlike the same-sex marriage ban analysis, this data is yearly and right-censored. As such, a Cox Proportion Hazards model that uses the exact discrete method for ties is well-suited to this policy as well. The dependent variable for this analysis is a dichotomous indicator of whether or not a state adopted an Official English policy in a given year. For the sake of comparability, this study only includes those laws that declare English as the official state language. Other English-only policies aimed at specifically at education and voting are not included. For example, California first enacts an English-only voting policy in 1984 and then enacts an Official English law in 1986. In this dataset, California is coded as a zero in 1984 and remains in the dataset until 1986 when the dependent variable becomes a one. As with the previous analysis, the key independent variable in this analysis measures each state’s direct democracy institutions. Again, I rely on an index of direct democracy impact created from a principle components analysis of the following measures: a dichotomous indicator of direct initiatives, the legislative insulation index, 24 Louisiana, Nebraska, Illinois, Massachusetts and Hawaii are excluded from the analysis because they all adopted an official language policy prior to the contemporary English-only movement. 59 the qualification difficulty index, and a count of the ballot initiatives in a state over the time period in question (1981 to 2005). The index ranges from zero for non-direct democracy states to 5.4 for California.25 The average score among direct democracy states is 3.9. Unlike the previous analysis, it is also necessary to include an interactive term in the model. Similar to the “racial threat” phenomenon in which white voters are more likely to oppose candidates and policies that protect the rights of racial minorities in areas in where these minority populations are high (Blalock 1967; Key 1949; Tolbert, McNeal, and Smith 2003; Tolbert, Grummel, and Smith 2001; Tolbert and Hero 1996), a language-based or ethnic threat has emerged in the area of language policy. Tatalovich (1995) differentiates between states that have high populations of foreign language . speakers in states that have relatively few foreign language speakers. As the population of foreign language-speakers in a state grows, the English-speaking majority is expected to feel threatened and more likely to prefer a restrictive language policy. In addition, the foreign language-speaking population also affects the salience of language policy. For states with high populations of foreign speakers, language policies would have a significant substantive effect on its citizens’ lives. For states with small populations of foreign speakers, language policy is more symbolic, less likely to have as significant substantive impact, and consequently less likely to draw opposition. The policy differences between states with high numbers of foreign language- speakers and states with low numbers have important implications for the effect of direct democracy. Direct democracy states should have increased probabilities of adopting 25 The index is slightly different from the previous analysis due to the differing time periods that went into the initiative use measure. 60 Official English as the number of foreign language speakers increases (Schildkraut 2001). Increased numbers of foreign language speakers in a state should increase both the substantive impact and relevance of language legislation, but without representative filters direct democracy states these language minorities are risk from an increasingly cohesive and relevant majority. In states with traditional representative governments, the likelihood of adopting Official English should decrease as the number of foreign language speakers increases. As language minorities increase in population, their representation in government should also increase, thus decreasing the probability of the states adOpting an English-only policy. Similar studies found empirical evidence to support this interactive relationship (Preuhs 2005; Schildkraut 2001). To account for this minority threat dynamic, I include a multiplicative term between the direct democracy impact measure and the percentage of foreign born * residents in each state. The foreign-bom population serves as a proxy measure for foreign language-speakers in each state since data on the language is not available for most years of the analysis.“ If citizen legislative institutions do endanger the rights of this minority group, then the slope on the combined effect of the direct democracy impact index and the interaction term should increase as the foreigrr-bom population increases. In addition to these key variables, several other potential determinants of policy adoption are also included in these models.27 Like the previous section, this analysis controls for citizen ideology, party control of government, diffusion from neighboring states, population, and education levels. The model also includes two new controls, 26 This data is available from the US. Census Bureau for 1980, 1990, and from 1994 to 2005. Data between 1980 and 1990, and between 1990 and 1994 are interpolated. 27 Again, other potential determinants of policy adoption have been considered in earlier models, but are presented here because they do not provide any additional explanatory power. The other variables Include political culture, income per capita, legislative professionalization and voter turnout rates. 61 reflecting the different factors that might influence language policy as opposed to marriage recognition policy. Economic hardships, according to the conventional wisdom, may lead to people to blame immigrants for these problems, and thus increase the propensity for states to adopt anti-immigrant policies (Schildkraut 2001). Consequently, states with higher unemployment rates should have a higher likelihood of adopting an Official English law. The final control variable, a dichotomous indicator of whether the state is southern, accounts for regional differences. Extending from 1981 to 2005, this study includes an era when the South was considered to be politically exceptional, especially on policies of ethnicity and race. Southern states should be more likely to adopt an Official English law in a given year. 2.2.3 Results The results from the Cox proportional hazards analysis of the adoption of Official English are shown in Table 2.7. Like the previous analysis, the coefficients can be interpreted similar to conditional logistic regression coefficients, as the effect of a one unit change in the independent variable has on the log-odds of a state adopting Official English in a particular year. Though the estimates from the model are fairly consistent with the results from the same-sex marriage ban analysis, this evidence is not quite as clear cut. Since the model includes a multiplicative term between its direct democracy variable and the foreign born variable, it is necessary to consider both the interaction coefficient and the original direct democracy coefficient together. 62 When considered alone, the direct democracy impact coefficient indicates the effect of these institutions when the foreign born population is zero. Thus, it is not surprising that this coefficient has a significant negative effect when there are essentially no speakers of foreign languages in a state. With a completely homogenous population, there is little need for a language policy, and certainly no threat from a minority group. Table 2.7 State Adoptions of Official English, 1981-2005 Variable Coefficient Standard Error Direct Democracy Impact -0.823** 0.238 Percent Foreign Born H -0.280* 0.121 Direct Democracy X Foreign Born [+] O.227** 0.053 Citizen Ideology H -0.018 0.027 Republican Government [+] 1.773* 0.922 Democratic Government H -1.799* 0.791 Party Competition [+] 1.090 4.170 Bordering States w/Official English [+] -5.653** 1.540 Population (log) H -1.005** 0.424 Educational Attainment H -0.125# ‘ 0.093 Unemployment Rate [+] -0.057 0.171 Southern State [+] 6.81 1** 1.302 Observations 783 Log Likelihood -79.299 Notes: Expected direction of coefficients in brackets; one-tailed tests used where appropriate, two-tailed tests otherwise; # significant at 10%; * significant at 5%: ** significant at 1% 63 However, if the population of foreign born residents is greater than zero, the effect of direct democracy cannot be interpreted solely from the direct democracy impact coefficient. Rather, the influence of direct democracy on the adoption of Official English is the additive effect of the direct democracy coefficient and the interaction coefficient. Taken together these coefficients show that states with direct democracy institutions will have an increasing likelihood of adopting an Official English law as the foreign-bom population increases. Conversely, states without direct democracy have a decreasing likelihood of adopting this policy as the foreign-bom population increases. This relationship is more apparent when examining the predicted hazard rates of adopting Official English for a given year.28 It is clear from Figure 2.7 that as the percentage of foreign language speakers increases in states with direct democracy, the likelihood of adoption also increases. When the foreigrr-bom population reaches a critical percentage of the resident population, the propensity to adopt a restrictive language policy is clearly higher for direct democracy states compared to non-direct democracy states. The estimated hazard rates also reveal that high impact direct democracy states have a higher propensity to adopt Official English over states with lower impact direct democracy institutions, once the foreign born population is large enough. This result suggests that as the number of foreign born residents in a state increases, the majority should be increasingly threatened and in favor of restrictive language policies like Official English. The positive slope of the line shows an increasing responsiveness to this threatened majority. Anecdotal evidence also supports this finding. Florida English activists in 1987 contended that Florida citizens perceived a 28 Figure 2.7 shows the hazards rates only for 1981, but other years display similar patterns. 64 “threat” in the increasing numbers of Spanish speaking people in the state (Tatalovich 1995). With Official English legislation stalled in the legislature, “threatened” Floridians circumvented their elected officials and passed the language policy on their own with a ballot initiative. Figure 2.7 Hazard Rates of Adopting Official English, 1981 ('O.‘ Predicted Hazard Rate l I _ I _ I 0 5 '10 - 15 20 25 Percent Foreign Born . Direct Democracy --&-- —-O— No Direct Democracy Impact ___ 3 _,__.,___ Direct Democracy Impact = 5 Figure 2.7 also shows an interesting relationship between non-initiative states and the foreign-bom population. While these states have a relatively higher hazard of adopting Official English when the foreign born population is very low, the hazard rate drastically decreases as the percentages increase. This result supports the concept of representative government providing a filter to protect minority rights. As the foreign born population grows in a state, it should also gain representation in the legislative 65 process. The more representation this minority group has in traditional representative governments, the more the legislative filters should work to protect minority interests and the less likely it is that these governments will adopt anti-minority policies. So despite the increased preference for a restrictive language policy among the English-speaking majority, language minorities are better able to oppose Official English legislation as their numbers grow. This result is also consistent with recent findings that diversity moderates the effectiveness of representative institutions in protecting minority rights (N icholson-Crotty 2006). Moving beyond the direct democracy and foreign-bom variables, the coefficients of the control variables in the model generally meets expectations, with a few exceptions. The first of these exceptions is the insignificant effect of citizen ideology. In each of the models the coefficient is in the expected direction, suggesting that liberal states are less likely to adopt Official English, but it only reaches marginal significance in the fourth model. Despite these insignificant findings, we should be careful in dismissing the effect of citizen ideology on adopting language policies. Remember that the foreign-bom population variable is, in effect, serving as proxy for the public attitudes towards language and may swamp any effect that the more general citizen ideology variable may have. Also, the results show a highly significant difference between southern states and the rest of the country. This blunt dichotomous variable is most likely capturing some of cultural and ideological determinants of policy adoption that would also be picked up by citizen ideology and levels of education. The other exception is the significant negative effect of the regional diffusion variable. As the percentage of surrounding states with Official English increases, the likelihood of adoption actually decreases. 66 Other control variables in the analysis show less surprising results. In each of the four models, there is a significant effect of party control. Republican governments are significantly more likely than both divided governments and Democratic governments to adopt Official English. Democratic governments, meanwhile, are significantly less likely to adopt these measures than either divided or Republican governments. The results also support the argument that states with higher levels of education should be more tolerant of minority groups (including language minorities). Hamilton’s argument about the “enlargement of the orbit” to protect minorities from majority factions is also supported. More populous states are less likely to adopt Official English than smaller states. Overall, the evidence from this analysis supports the tyranny of the majority argument in the case of Official English. While the effect of direct democracy is moderated by the foreign born population in a state, these institutions nonetheless increase the probability of adopting this anti-minority policy. When the issue is made sufficiently salient and the majority English-speaking population is united by the threat of increasing numbers of foreign-speakers, direct democracy facilitates the adoption of Official English. As with the spread of same-sex marriage bans, citizen legislation only directly led to policy adoption in six states. However, when the indirect effect of citizen legislation is also accounted for, it is clear that direct democracy states are higher risk to adopt Official English than non-direct democracy states. 2.3 Affirmative Action Bans So far, this chapter has shown evidence that direct democracy institutions have played significant roles in the adoption two contemporary anti-minority policies that have 67 diffused across many American states. In this section I examine one final anti-minority policy: affirmative action bans. This policy area presents some interesting differences from the other two policies that should help to further test the tyranny of the majority argument. Unlike same-sex marriage bans and Official English, policies banning or severely limiting affirmative action have only been adopted in a small number of states. Another difference is more substantive. While the previous two policies clearly restrict the rights of minority groups relative to the majority group, affirmative action bans arguably could be described as rolling back minority rights to be equal to the rights of the majority group. Due to this difference, it is not surprising that proponents of affirmative action bans have taken up the language of civil rights and equality (Fobanjong 2001). Still, in banning affirmative action plans, states are effectively adopting policies that specifically'target minority groups by restricting the opportunities that these minority groups have for employment, advancement and education. Affirmative action programs were not implemented to grant equal rights, but rather to provide equal opportunities to groups that have been historically discriminated against. Advocates of these programs contend that banning them would leave minority groups without equal opportunities in these areas. Thus, this issue has been considered as a policy that effectively restricts the rights of minority groups and is one that has been studied in terms of the impact of direct democracy institutions (Schrag 1998). So if direct democracy does endanger minority rights, then it should also apply to this case where a united majority is targeting the (arguably enhanced) rights of minorities. 68 2.3.1 The Politics of Affirmative Action Bans Affirmative action policies have long been the center of heated debates, but it is only in the past decade that these policies have been targeted by statutory bans. Affirmative action policies were initially developed in the 1960’s as a way to remedy or “level the playing field” in education and employment after decades of segregation and discrimination.29 Using Title VII of the Civil Rights Act and several executive orders, the Johnson and Nixon administrations utilized policies of affirmative action to remove barriers to equality for minorities. The 1970’s saw affirmative action policies extend from government employment to private employment and education. The minorities included under these affirmative action policies included Afiican Americans, Hispanics, other ethnic minorities, and women. Throughout this period, opponents of affirmative action contended that these policies violated nondiscrimination laws and were essentially government sanctioned “reverse discrimination” This backlash against affirmative action was rebutted by the courts in several decisions (e. g. University of California v. Allan Bakke) in the 1970’s, but found new support in the 1980’s. The Reagan Administration gave a national voice to opponents of affirmative action and through conservative court appointments was able to effectively Whittle away at the scope of these programs over the next 15 years. By the mid-Nineties, however, the debate spread from the courts to state legislative and executive branches which began considering statutory action to eliminate affirmative action policies. In California, Govemor Pete Wilson (R) began to dismantle the state’s affirmative action policies with an executive order in 1995. That same year, 29 For a more detailed account of the history of affirmative action in the United States, see: Anderson, Terry H. 2004. The pursuit of fairness : a history of affirmative action. New York: Oxford University Press. 69 Regent Ward Connerly — a Wilson appointee — led a movement within the University of California (UC) regents to end affirmative action. With two votes, the board of regents eliminated preferences for minorities and women in UC employment and admissions. The next year, Connerly and Californians Against Discrimination and Preferences (CADAP) led a campaign to pass Proposition 209, a ballot initiative that would ban affirmative action by the government “in the operation of public employment, public education, or public contracting.” The initiative passed with fifty-four percent of the vote. Soon after, Connerly helped found the American Civil Rights Institute (ACRI) to continue the anti-affirmative action movement across the nation. Since 1996, seven states have adopted bans on affirmative action. Table 2.6 shows the states that have enacted an affirmative action ban. Four of the states enacted relatively broad bans that apply to employment, education and public contracting. The _ others have more narrow policies, from bans for only police and fire hiring to bans for public employment. In terms of the direct impact of citizen legislation in this policy area, three of the states used citizen legislation to enact their bans. Each of these policies was far-reaching, covering employment, education, and public contracting. Table 2.8 also shows further evidence as to direct democracy’s effect in this policy area. Every state that has adopted an affirmative action ban thus far is also an initiative state. Conversely, there is not a single non-initiative state that has adopted an affirmative action ban. 70 Table 2.8 States with Affirmative Action Bans Type of Initiative State Year Adoption Programs State? California 1996 Initiative Employment, Education, A Contracts Washington 1998 Initiative Employment, Education, A Contracts Colorado 1999 Legislation Employment A Missouri 1999 Legislation Police & Fire A Employment Florida 2000 Executive Order Employment, Education, A Contracts Utah 2003 Legislation Employment A Michigan 2006 Initiative Employment, Education, A Contracts 2.3.2 Analysis With only direct democracy states adopting affirmative action ban, it appears that these institutions once again have a negative impact on the rights of minority groups — this time on a host of racial and ethnic minorities. However, it is important to be cautious before the evidence is more systematically examined and analyzed. In the previous two sections event history analysis was utilized to this end. Unfortunately, using the same analytical strategy in the case of affirmative action bans presents some problems. Since such a relatively small number of states had adopted an affirmative action ban as of 2005, the dataset only includes six adoptions (or failures) out of a possible 455 observations. This creates the potential for biased estimates. The nature of the six states that have adopted bans also presents problems for an event history analysis. Since only six states 71 have bans, many of the control variables are perfect predictors of failure and standard errors cannot be estimated. Due to this obstacle only a very limited and probably underspecified model can be estimated. 30 Finally, since the scope of these affirmative action bans varies quite a bit from state to state, there are issues of comparability in equating each option as the same. Due to these problems, alternative analytical approaches are necessary. To take a closer look at the differences between direct initiative states and non-initiative states, an aggregate comparison of the two types of states is revealing. The argument that direct democracy endangers minority rights does have a few observable implications that lend themselves to this type of analysis. The first can be seen in outcomes — if initiative states do have a higher likelihood of adopting an affirmative action ban, then the policy outcomes should reflect this. As seen in Table 2.9, nearly thirty-seven percent of direct initiative states have banned affirmative action to some extent while not a single non— initiative state has adopted a comparable policy. The second implication is seen in attempts — given the higher likelihood of adoption and the indirect pressure that majorities can apply on the legislature in direct initiative states, non-initiative states should be less likely to propose such policies. Again, the empirics support this expectation. Almost three-quarters initiative states have proposed a measure to ban affirmative action, a significantly higher rate than non- initiative states. Another telling statistic in Table 2.9 is the average number of bills that have been proposed in the states which have considered an affirmative action ban. The difference 30 A Cox proportional hazards models was estimated and produced results consistent with this analysis and the previous analyses, but given the uncertainty surrounding the bias of those estimates the results are not presented here. Those results are shown in Appendix D. 72 between the two types of states is statistically insignificant. While direct initiative states have significantly higher percentages of states with bans and states that have tried to ban affirmative action, the average number of bans proposed in the two types of states is statistically the same. Even though non-initiative states consider roughly same number of anti-affirmative action proposals, they are less likely to actually adopt these measures. Among the states that have considered these policies, half of the direct initiative states passed a law while none of the non-initiative states passed one. This suggests that the representative mechanisms of traditional state governments are working to filter out bills that target minority rights. Table 2.9 Initiative States vs. Non-Initiative States, 1996-2006 Initiative Non-Initiative . States States Difference Percent with Bans 36.8 % 0.0 % 36.8%* Percent Attempted 73.7 % 45.2 % 28.5%* Average Number of 1.9 1.2 0.7 Proposals *Significant at 1% A second analytical approach that closely examines the relevant cases can also shed some light on the role that direct democracy played in the adoption of these policies. Of the seven states with affirmative action bans, four are far-reaching, affecting university admissions policies, public employment policies and public contracting policies. These policies are of particular interest because of their expansive restriction on affirmative action policies. For three of these states — California, Washington and Michigan - the role of direct democracy is clear. Each of these states used ballot 73 initiatives to adopt their affirmative action bans. Led by Ward Connerly and the ACRI, the ballot measures easily passed with an average of fifty-seven percent of the vote. Despite having fairly diverse populations,31 a united majority was able to circumvent the filters of representative government and restrict the rights of minorities in their state. In California, three-quarters of Afiican Americans and Latinos opposed Proposition 209, while two-thirds of white men favored it along with a majority of white women (Anderson 2004). Similar patterns of support were seen in Washington and Michigan (Eason Jr. 1998; Lewin 2006), revealing a united majority of white men and women voting against the preferences of ethnic and racial minorities. The role that direct democracy played in the adoption of the fourth broad affirmative action ban in Florida is less obvious. The policy was issued as an executive order by Governor Jeb Bush (R) on November 9, 1999. The order ended racial preferences in state employment, contracting and university admissions. At the same time, it also proposed a “Talented 20 Plan” to ensure admission to the top twenty percent of each high school graduating class as an altemative policy to help maintain diversity. This plan was subsequently approved by the Florida Board of Regents. Although this policy was enacted unilaterally by the Governor, political context surrounding its adoption included many actors. Earlier that year, Ward Connerly announced a petition drive to get a proposal similar to Califomia’s Proposition 209 and Washington’s I-200 on Florida’s 2000 ballot. Until 1999, Bush had never advocated a plan to end affirmative action in the state. In fact, Bush had built his electoral success, in part, on a strategy of rapprochement with minority and civil rights leaders, and had 3 1 California is the third most diverse state in the country, as measured by Hero’s minority diversity index (Hero, Rodney E. 1998. Faces of Inequality: Social Diversity in American Politics. New York: Oxford University Press). Michigan and Washington rank 22"d and 25m, respectively. 74 opposed Connerly’s efforts (March 1999). However, the specter of Connerly’s proposal on the 2000 ballot put the Governor in a difficult situation. Bush’s brother was running for president in 2000 and Florida was predicted to be an important swing state. The 2000 ballot also included a competitive race for the open US. Senate seat. The Republican Party worried that Connerly’s proposal could stimulate Afiican Americans (who tend to vote for Democrats) to turnout in massive numbers, thus threatening GOP chances in both elections. Both national and state Republicans wanted to remove this issue from the ballot. Further pressure was applied to Bush in early November by the public when newspapers reported that Florida voters supported Connerly’s proposal by a margin of more than two to one. With simultaneous pressures to remove the issue from the 2000 election from the GOP, end affirmative action from the public, and support affirmative action from his own electoral coalition, Bush acted unilaterally to impose a moderated anti-affirrnative action policy. While his Florida One plan ended racial preferences in employment and contracting, it also called on department leaders to voluntarily take diversity into account in their hiring and contracting processes. The order also ended racial preferences in university admissions, but implemented a percentage plan for admissions to mitigate the loss of minority enrollment in these schools. After examining the political context surrounding Bush’s executive order, it is clear that direct democracy played a role in the adoption of an affirmative action plan. This is a perfect example of the indirect impact that citizen legislation can have on the behavior of policymakers. Connerly’s petition drive not only pushed the Governor to address the issue, but with the strong support for the proposal (indicating a united 75 majority) it also pushed him to take action. Normally, the indirect effect of direct democracy is thought as a pressure on legislators. In this case, the ballot initiative put pressure on the elected executive to enact a policy in line with the preferences of the majority at the expense of minority groups. In all four states that have enacted far reaching affirmative action bans, direct democracy has played a prominent role in the adoption process. But these types of policy proposals have not been limited just to direct democracy states. As shown in Table 2.9, over forty-five percent of non-initiative states have also had affirmative action bans introduced in their legislatures, but none of these bills has been successful. If the tyranny of the majority argument is correct, these proposals in states with traditional representative legislatures would have “filtered” them out and protected minority rights. One state, in particular, is illustrative of how the representative filters can work. In New.:lersey, which has seen more proposals a than any other non-initiative state, opponents of affirmative action have introduced two bills each session that would ban these types of policies. One is a simple ban on all preferences programs that are based on race, ethnicity, gender, color or national origin. The other eliminates these programs and replaces them with economic-based preference programs. To date, neither bill has been able to get out of committee. These bills’ failures are somewhat surprising given the relatively high support the bill have in public opinion. The sponsor of the bill, Rep. Michael Patrick Carroll (R), notes that, "sixty percent to seventy percent of the public supports the idea” (Richmond 1996). The failures are also surprising given the high amount of publicity the issue of affirmative action has received in the state, especially prior to 1998. In 1989, the 76 Piscataway, New Jersey school board was cutting personnel and fired Sharon Taxman, a white teacher, instead of a black teacher. Both were hired on the exact same day and received equal evaluations. The board cited a policy of maintaining diversity in letting Taxman go. She sued the school board, claiming reverse discrimination under title VII of the Civil Rights Act. The case worked its way through the courts and made headlines in 1996 when a federal appeals court ruled that the harm imposed to the non-minority by the loss of their job was greater than the benefit of increased diversity. The case made further headlines in 1997 when a civil rights coalition dropped their appeal to the Supreme Court, fearing another that decision like Hopwood v. Texas would fiirther erode affirmative action programs in the US. Over the course of the suit, the Taxman case drew both statewide and national attention. Both the Bush administration and the Clinton ' administration weighed in (on opposite sides) along with other national leaders (Anderson 2004). The case brought the issue of affirmative action to the gubernatorial race and to several legislative races. Despite all the attention to affirmative action in New Jersey and the high public support to end these programs, the bills continue to die in committee. If New Jersey is a case where representative democracy is providing a filter to protect minority interests, then we should expect the state to have two characteristics that would allow these filters to function. First, the state needs to have fairly diverse population so that there are a sufficient number of minorities that would oppose an affirmative action ban. Second, the state’s minority groups should have some degree of descriptive representation in the legislature that would serve their interests. New Jersey meets both of these criteria. In terms of population diversity, New Jersey ranks eleventh in the country. In 2000, the 77 state’s population was nearly fourteen percent black, over thirteen percent Hispanic, almost six percent Asian, and seventy-two percent white. The legislature, although not completely mirroring the population diversity, does provide some descriptive representation. According to the National Conference of State Legislatures, the state legislature in 2003 was thirteen percent black and six percent Hispanic.32 With this amount of diversity both in the population and in the legislature, minorities are able to filter out policy proposals that would restrict their rights. Even in a climate of heightened attention to affirmative action issues and majority public support to end these programs, minority groups have been able to successfully defend their interests in the state government. While the evidence presented in this section is less systematic than the previous two analyses, the results are nonetheless consistent. Even though the diffusion of affirmative actions bans has yet to Spread beyond a handful of states, the influence of direct democracy in this process is clear. Ballot initiatives were used directly to enact three of the four most expansive affirmative action bans. In the fourth state, the threat of direct democracy influenced the governor to take unilateral action to enact a more moderate version of the policy. States without direct democracy institutions, meanwhile, have yet to adopt a single affirmative action ban. To date, seven direct democracy states have adopted policies banning affirmative action in state programs. 32 http://www.ncsl.org/programs/legman/about/demographic_overview.htm 78 2.4 Discussion The analyses in this chapter all consistently produced evidence supporting the thesis that direct democracy endangers minority rights. For each policy, the question was posed, “Given that states had not yet adopted the policy, are direct democracy states more likely to adopt the policy in that year than states without direct democracy institutions?” The results of this examination all answer this question in the affirmative. Whether it’s a same-sex maniage ban, an Official English law, or an affirmative action ban, states with direct democracy institutions are more likely to adopt the policy as it diffuses across the country. While there may be moderating circumstances that affect the public attitudes on a particular issue, as was the case for Official English, citizen legislation consistently increases the ability of a united majority to enact policies that target minority rights. This effect seems to be constant across minority group types, from race to ethnicity to sexual preference. The important factor is the unity of the majority group in its preferences to restrict the rights of minorities in some way. In addition to the consistent support for the main hypothesis, these analyses also revealed some important findings about the nature of the impact of direct democracy on minority-related policies. By examining both the direct outcomes of citizen legislation as well as the traditional legislative outcomes in initiative states, the indirect effect of direct democracy was readily apparent. In only considering the direct outcomes of citizen legislation, it appeared that direct democracy did not play a significant role in the diffusion of same-sex marriage bans and Official English. At best, this more superficial examination produced mixed results. However, once the indirect effects of direct 79 democracy were accounted for (by including the outcomes of traditional legislation in the analysis), the full influence of direct democracy can be appreciated. The measurement of direct democracy institutions that accounts for institutional arrangements and initiative use also presented some interesting implications for how citizen legislative institutions can affect policy outcomes. By measuring direct democracy in multiple ways, the analyses were able to show more than just a dichotomous effect. Clearly, not all direct democracy institutions are the same. The variation in institutional arrangements affects the policy decisions these states produce. In general, states with direct initiatives, less insulated legislatures, easier qualification requirements and relatively high initiative use were more likely to adopt these anti- minority policies 80 CHAPTER 3: A REPRESENTATIONAL FILTER? THE PASSAGE OF ANTI-MINORITY POLICY PROPOSALS The evidence presented in the previous chapter suggests that states with direct democracy institutions are more likely to adopt specific anti-minority policies in a given year than states without direct democracy institutions. As argued in the introduction, the source of this increased propensity for direct democracy states to adopt anti-minority proposals lies in the institutional arrangement that allows citizens to circumvent the filtering processes of representative democracy. Thus, direct democracy states are less likely to reject anti-minority policy proposals than states without direct democracy institutions because these proposals are not subject to legislative filters of the traditional representational process. In the analyses of same-sex marriage bans, official English laws and affirmative action bans it is clear that direct democracy states were more likely to adopt these three policies, but it is still unclear whether this increased propensity to adopt anti-minority policies is a direct consequence of the varying filtering mechanisms of the two institutional arrangements. In using policy adoption as the unit of analysis, the study was not able to examine failed anti-minority policy proposals. If a state does not adopt the policy in a given year, it is coded as a zero, whether or not the policy was even considered. From this perspective, states that reject anti-minority proposals are equivalent to states that don’t even consider them. Similarly, a state that passes a single anti-minority proposal out of ten introduced is equivalent to a state that passes the only anti-minority policy proposal it 81 considers. Again, while event history analysis does provide clear evidence of a state’s propensity to adopt a specific policy, it cannot directly verify the underlying foundation of the tyranny of the majority argument because it does explicitly examine whether the filtering mechanism are actually functioning to protect minority rights. Thus, in order to more directly study this question, it is necessary to consider individual policy proposals. The “tyranny of the majority” argument implies that an anti-minority policy proposal is more likely to pass in a state with direct democracy institutions because it can avoid or undermine the filtering processes of representative democracy. It follows from this argument that states with direct democracy institutions should have higher rates of passage for anti-minority proposals. In this chapter I examine these implications through an analysis of anti-minority policy proposals considered in the American states in the last decade. 3.1 Anti-Minority Policy Proposals In this study, I examine policies that directly and explicitly restrict the rights of political minorities, and not the broader set of policies that impact the interests of minorities. Again, it is the explicitly anti-minority policies that should be most impacted by the majoritarian effects of direct democracy and it is these types of issues that most directly affect the rights of minority groups. From 1995 to 2004, there were three broad types of policy proposals that explicitly aimed to restrict the rights of minority groups in the American states: policies that targeted homosexuals, policies that target limited English proficient (LEP) speakers, and policies that target minorities in general by restricting or banning affirmative action 82 programs. These three categories of anti-minority proposals were considered in most states during this decade. Every state considered at least one anti-homosexual policy. Thirty-four states considered an anti-foreign language policy, and thirty-one states considered at least one general anti-affirmative action policy. Together, these three policy categories encompass the vast majority of all the anti-minority policies considered during this time period. Certainly other anti-minority policies were introduced during this period, but these other anti-minority policies tended to be less generalizable to other states and more idiosyncratic. 33 By focusing on these three categories, this study does not include every single anti-minority policy proposed in the American states over this type period, but it does include the most salient and visible minority rights issues of the period, like same sex marriage bans, official English laws, and affirmative action bans. Again, most anti-minority policies introduced from 1995-2004 do fit into one of these three categories and are covered in this analysis. To identify the these anti-minority policy pr0posals, I conducted keyword searches of the text of all legislative bills and the text of all ballot measures introduced from 1995 to 2004. 34 Table 3.1 shows the types of policy proposals identified in the search by the target group. The most common policy proposals targeting homosexuals were same sex marriage bans, while the most common policy proposal targeting foreign language speakers and minorities were official English laws and general affirmative action bans, respectfirlly. 33 For example, the Arkansas Senate considered a bill to limit parking for people with disabilities in 2003. 3" Keyword searches were conducted through Lexis Nexis State Capital, the National Conference of State Legislature’s Ballot Measure Database, and individual state legislative bill archives. For more details about the search see Appendix D. 83 In total, the search yielded 600 proposals (see Table 3.2). 335 targeted homosexuals, 148 targeted foreign language speakers, and 117 targeted minorities in general. Of these 600 proposals, 82 passed. The passage rate varied across the targeted groups, with anti-homosexual proposals passing nineteen percent of the time and proposals targeting minority groups in general passing about four percent of the time. Table 3.1 Anti-Minority Pmosals by Targeted Group, 1995-2004 Homosexuals Same sex marriage ban* Bar homosexuals from adoption and/or foster care Ban on benefits for ‘ same sex couples Ban on education of homosexuality Ban on recognition of civil unions Ban on sodomj* Foreign Speakers Official English* English-Only education* English-Only ballots English proficiency requirement for public assistance English-Only driver’s license exams *Also includes constitutional amendments Proposals in italics indicate the most common policy proposal for each target group 84 Minorities General affirmative action ban* Ban on affirmative action for educational institutions Ban on affirmative action for public hiring & contracting Ian on group-norming test scores for employment or admissions Affirmative action restrictions Table 3.2 Passage Rates of Anti-Minority Proposals by Targeted Group, 1995-2004 Homosexuals Foreign Speakers Minorities All Groups Proposals 335 148 1 17 600 Passed 64 13 5 82 Percent 19.1 8.78 4.27 13.67 Passed 3.1.1 Direct Democracy vs. Representative Democracy With the anti-minority proposals identified, it is possible to compare the passage rates between states with direct democracy institutions and states without direct democracy institutions. If the direct democracy does reduce the effectiveness of the representational filter and endanger minority rights as critics suggest, then anti-minority , policy proposals in direct democracy states should pass at a higher rate than non-direct democracy states. The four measures of direct democracy employed in the principle components analysis from the previous chapter are utilized individually here to compare passage rates of these anti-minority policy proposals: a dichotomous indicator of direct initiatives, the Legislative Insulation Index, the Qualification Difficulty Index, and the Initiative Use count. In addition, dichotomous comparisons are made between the twenty-four states with any direct democracy institution and the twenty-six states without direct democracy. Using the individual components of the direct democracy impact index allows for a more substantively simple interpretation of these comparisons. Table 3.3 shows the passage rates of these proposals over the various measures of direct democracy discussed above. The first section compares states with direct initiatives to those without direct initiatives. 45 of the 210 proposals considered in the 85 nineteen direct initiative states passed, a rate of over twenty-one percent. States without direct initiatives, meanwhile, only passed 37 of 390 proposals — less than half the rate of the states with direct initiatives. The last column in the table shows that the difference between the two, at almost twelve percent, is statistically significant. Table 3.3 PassaLeRates of Anti-Minority Proposals, 1995-2004 States Number Passed Rate Diffeerenc By Direct Initiatives Direct Initiative States 19 210 45 21.43 % 11.94 %** Non-initiative States 31 390 37 9.49 % By Direct Democracy Direct Democracy States 24 270 53 19.63 % 10.84 %** No Direct Democracy 26 330 29 8.79 % EV Legislative Insulation - O ’ Least Insulated (5 9) 15 145 33. 22.76 /0 6.76 %# More Insulated (1-4) 9 125 20 16.00 % o II- No Direct Democracy (0) 26 330 29 8.79 % 7'21 /° By Qualification Qifficulty Least Difficulty (5-7) 9 91 19 20.88 % 1 88 % More Difficult (1-4) 15 179 34 19.00 % o ** No Direct Democracy (0) 26 330 29 8.79 % 10'21 /° By Initiative Use ‘ 0 High Use (>10) 14 154 30 19.48 /0 0.73 % Low Use (1-10) 9 112 21 18.75 % o 9.47 %** No Use (0) 27 334 31 9.28 /o Notes: Variable score in parentheses**Significant at the 0.01 level; *Significant at the 0.05 level; #Significant at the 0.1 level 86 Using the broader direct democracy indicator to compare passage rates produces similar results. Anti-minority proposals in direct democracy states pass at more than double the rate of anti-minority proposals that are considered in states without direct democracy. Again, the difference in the passage rates between these two categories of states is highly significant. The difference in the passages rates is also evident in Figure 3.1. Figure 3.1 Anti-Minority Proposals Passed, Binary Comparisons Non-Direct Initiative States -. ‘- Direct Initiative O _ States Non-Direct Cli- ul- Democracy States Direct Democracy States nib fi 4. I I 0 .1 .2 .3 Proportion Passed Note: Vertical bars indicate the upper and lower bounds of the 95% confidence interval The last three sections of the table employ the more nuanced measures of direct democracy. Given the small number of states within each scale score, I combine the 87 higher scores into one category, the lower scores into another category, and the non- direct democracy states into another. The differences between these categories are also shown in Figure 3.2. Looking at the legislative insulation categories, there is a marginally significant difference between states with less insulated legislatures and the states with more insulated legislatures; just under twenty-three percent compared to sixteen percent. The difference between states without direct democracy and direct democracy states with relatively more insulated legislatures is also significant, with anti- minority proposals passing at almost twice rate of the former category. Figure 3.2 Anti-Minority Proposals Passed, By Institutional Arrangement g g No DD i Q t f7? g High t o t 05") to _i 5 Low t f ‘r é E‘ No DD : o t 8 3 5.; 3% High 1 £ 1 m ._ :3 D I I 0 LOW i ‘ r 0, None 1 O i .2 a, g 3 Low t k : C — High i o : I T I 0 .1 2 .3 Proportion Passed. Note: Vertical bars indicate the upper and lower bounds of the 95% confidence interval 88 Using the qualification difficulty index produces similar results, but the difference between the two categories of the direct democracy is not statistically significant. Still, the primary difference of interest, between direct democracy states and non-direct democracy states, remains robust. The percentage of anti-minority proposals passed in states with relatively difficult qualification requirements more than doubles the percentage passed in states without direct democracy institutions. Comparing passage rates by the states’ initiative use over the time period reveals a significant difference between states that use initiatives and states that do not. As with the qualification difficulty comparison, there is no significant difference between high use states and lower use states. No matter how direct democracy is measured, the comparisons show that states with direct democracy institutions pass significantly higher percentages of anti-minority proposals than states without direct democracy institutions. However, since the anti- minority proposals in this study include three types of targeted minority groups, it is also prudent to examine whether this result is robust across the three groups. In other words, does the discrepancy in passages rate of anti-minority proposals hold across the different targeted minority groups? I address this question by comparing the passage rates of anti-minority proposals in direct democracy states to the passage rates of similar proposals in states without direct democracy across the three types of targeted groups. The results of these comparisons are shown in Table 3.4 and Figure 3.3. The table also examines the difference between states with direct initiatives and those without direct initiatives. For proposals that target homosexuals, a relatively large difference between the states is evident. Likewise, a 89 significantly higher percentage of proposals that target minority groups through affirmative action restrictions pass in states with direct democracy institutions than in non-direct democracy states. Table 3.4 Anti-Minority Proposal Passage by Targeted Group, 1995—2004 By Direct Initiatives BK Direct Democracy Direct No Direct Direct No Direct Initiatives Initiatives Democracy Democracy Homosexuals Proposals 1 1 1 224 153 182 Percent Passed 31.53 12.95 26.14 13.19 Difference 1859* 12.96* (0.0000) (0.0013) Foreiqn Speakers Proposals 100 '48 61 87 Percent Passed 10.42 8.00 13.11 5.75 Difference 2.42 7.37# (0.3134) (0.0595) Minorities Bills 51 66 56 61 Percent Passed 9.80 0.00 8.93 0.00 Difference 9.80* 8.93* (0.0047) (0.0085) *Significant at the 0.01 level ; #Significant at the 0.1 level; P-values for one-tailed tests in parentheses 90 Figure 3.3 Anti-Minority Proposal Passage, by Group & Institutional Arrangement U, No Direct Initiatives i O i c i. 33 Direct Initiatives 4. G i L it: o g No Direct Democracy H—O : “'cn Direct Democracy : % 1 w . . . . "g No Direct Initiatives .L G i 5 Direct Initiatives : O i (D E No Direct Democracy 1 O : g Direct Democracy : O i a, No Direct Initiatives. CD . .. . 33 Direct Initiatives % O i ,8 No Direct Democrac 2 Direct Democracy 1' 1 : T l l 1 - 0 .1 .3 .4 .2 Proportion Passed Note: Vertical bars indicate the upper and lower bounds of the 95% confidence interval Interestingly, these differences diminish for proposals that target foreign speakers. While the disparity of passage rates between direct democracy states and states without direct democracy is marginally significant (with a p value of 0.0595), the difference is insignificant when compared by the presence of direct initiatives in a state. This may be caused by the minority threat dynamic discussed in the previous chapter. As the literature on language policy in the American states suggests (Preuhs 2005; Schildkraut 2001; Tatalovich 1995), when there are relatively low numbers of foreign speaking individuals in a state, language policy is not a publicly salient issue. However, as the population of foreign speakers increases, the issue of language becomes more salient and 91 the majority may feel threatened. The analysis in of Official English laws shows that among states with high foreign-bom populations, direct democracy states are more likely to adopt these policies than states without direct democracy. This interactive effect between direct democracy and the size of the foreign speaking population on the likelihood of adopting official English should also extend to the passage rates of proposals that target foreign speakers. Table 3.5 shows the differences between the states grouped by their foreign-bom population. In states with low populations of foreign-bom individuals, there is no difference in the percentage of proposals that pass. In the states with higher levels of foreign born individuals, however, the difference between direct democracy states and states without direct democracy is quite large and statistically significant. The difference between direct initiative states and states without direct initiatives also increases, but falls short of traditional levels of significance. Although the dichotomy of low and high foreign-bom populations is a blunt way to assess this interactive effect, the results do support the idea that threat or public salience affects the impact that direct democracy has on language policies and mirrors the results from the previous chapter. One final way to compare direct democracy states to non-direct democracy states is to examine the average percentage of anti-minority proposals that each state passes. Table 3.6 shows the passage rates for each state over this time period. The states in the left column are the states with direct democracy institutions (the middle column indicates whether they have direct initiatives). The averages for each type of state are presented in the bottom rows. From this perspective, there is an even starker difference between the states. On average, direct democracy states pass well over thirty percent of their anti- 92 minority proposals, over twenty percentage points more than states without direct democracy. The gap between direct initiative states and states without direct initiatives is even larger. Table 3.5 Passage Rates of Proposals Targeting Foreign Speakers Foreign-born Foreign-born Population 55% Population > 5% Direct Democracy 9.38 °/o 17.86 % No Direct Democracy 8.51 % 2.50 % . 0.86 °/o 15.36 %* D'fie'ence (0.4472) (0.0140) Direct Initiatives 8.00 % 13.64 % No Direct Initiatives 9.26 % 6.52 % . -1.26 "/6 7.11 % D'fie’ence (0.5727) (0.1666) * Significant at the 0.05 level; P-values for one-tailed tests in parentheses Table 3.6 also reveals that direct democracy states, on average, consider roughly equal numbers of anti-minority proposals as non-direct democracy states. If anything, states without direct democracy consider more anti-minority proposals than direct democracy states. Direct democracy states averaged just over eleven anti-minority proposals during this time period, while non-direct democracy states considered over twelve anti-minority proposals, on average. This is important because it underscores the point that direct democracy states are not adopting anti-minority policies because they are more exposed to these types of proposals. Rather, direct democracy states are actually passing anti-minority proposals at a higher rate. Direct democracy states do not seem 93 more likely to consider anti-minority proposals, just more likely to pass them given the ability of the citizens to circumvent the filtering mechanisms of representative government. In contrast, though non-direct democracy states are considering roughly the same number of anti-minority policy proposals, they pass these policies at a significantly lower rate. This strongly suggests that the representational filters are functioning to protect minority rights in these states. In all, the descriptive data tends to support the findings from the previous chapter. In comparing the passage rates of anti-minority policy proposals, direct democracy states pass higher percentages of proposals they considered than states without direct democracy. This relationship holds across varying indicators of direct democracy and across the three targeted minority groups. 94 Table 3.6 Passage Rates of Anti-Minority Proposal by State, 1995-2004 Direct Democracy States Non-Direct Democracy States Percent Percent State Proposals Passed State Proposals Passed AZ 7 28.57 AL 28 3.57 AR 5 60. 00 CT 4 0.00 CA 17 17.65 DE 5 20.00 CO 18 11.11 GA 14 28.57 FL 2 50.00 HI 21 4.76 ID 3 33.33 IN 7 14.29 IL 3 33.33 IA 17 1 1.76 MI 20 15.00 KS 9 11.11 MO 43 13.95 KY 10 20.00 MT 5 60.00 LA 11 18.18 NE 5 20. 00 MD 8 0.00 NV 3 66. 67 MN 16 6.25 ND 2 100.00 NH 7 14.29 CH 12 16.67 NJ 25 0.00 OK 24 12.50 NM 10 0.00 OR 10 10. 00 NY 29 0.00 SD 4 50. 00 NC 7 14.29 UT 6 83.33 PA 1 5 6.67 WA 21 9. 52 RI 2 0.00 AK 5 60.00 SC ' 22 4.55 ME 5 20.00 TN 4 25.00 MA 1 1 9.09 TX 14 7.14 MS 35 5.71 VT 10 10.00 WY 4 25.00 VA 6 66.67 WV 19 5.26 WI 10 10.00 Direct Non-Direct Democracy 1 1 .25 33.81 * Democracy 12.69 1 1 .63* Average Average Direct Non-Direct Initiative 1 1 .05 3620* Initiative 12.58 13.62* Average Average Note: Direct initiative states in italics; "‘ Significantly different at the 0.001 level from the corresponding average percent in the same row 95 3.2 Direct Democracy & Anti-Minority Proposal Passage While the cross tabulation analysis certainly provides further support for an anti- minority impact of direct democracy and for the protective filtering mechanisms of representative democracy, it cannot account for other determinants of anti-minority proposal passage, like public ideology and party control of the government. Obviously, there are a myriad of factors that can influence whether a single policy proposal is ultimately adopted by the state. In order to assess the effect of direct democracy while also controlling for other possible determinants of anti-minority proposal passage, I estimate a logistic regression model where the unit of analysis is each individual proposal. This approach asks the following question: For any given anti-minority proposal, is it more likely to pass in a direct democracy state or a state without direct democracy? The dependent variable is a binary indicator of whether or not an anti-minority policy proposal was passed. It does not distinguish between ballot measures and legislative bills since the study is trying to gauge the total effect of direct democracy, not just the direct effect. The primary independent variable of interest measures a state’s direct democracy institutions. As in the previous chapter, an index of direct democracy impact, generated from a principle components analysis of the following variables is used: a dichotomous indicator of direct initiatives, the legislative insulation index, the qualification difficulty index, and the use of initiatives in the state over the time period in question. As with the previous analyses I expect a positive coefficient on this variable, indicating that direct democracy states are more likely to adopt anti-minority proposals than states without direct democracy. 96 The other possible determinants of anti-minority pr0posal passage can be grouped into three general categories: governmental attributes, state attributes, and proposal attributes. The governmental attributes include party variables and measures of the institutional attributes of the legislature. ‘To account for party effects, three variables are employed. Party control of the state government is accounted for with two dichotomous indicators of whether the Republican or Democratic Party has control of the legislature and the govemorship. The base category in these models is divided government. I expect Republican governments to be more likely to pass anti-minority proposals and Democratic governments to be less likely to pass these proposals. The folded Ranney Index is also included to account for party competition. The models also account for institutional aspects of the state legislatures, including legislative professionalization,” the size of the legislature, and whether or not one of the chambers has a limit on the number of bills that can be introduced. Larger and more professional legislatures tend to consider more bills in each session and thus should be less likely to pass any single bill. Conversely, legislatures with bill introduction limits should be more likely to pass a bill. In addition to governmental attributes, the model also controls for several state attributes. First, citizen ideology is taken into account using the Berry, et al. measure (1998; 2003).36 More liberal states should be less likely to pass anti-minority proposals. Minority diversity is also accounted for using Hero’s (1998) diversity measure. More diverse states should be less likely to pass anti-minority proposals. Several other 35 Squire Index (Squire, Peverill. 1992. "Legislative Professionalization and Membership Diversity in State Legislatures". Legislative Studies Quarterly 17 (l):69-79, Squire, Peverill. 2000. "Uncontested Seats in State Legislative Elections". Legislative Studies Quarterly 25 (1)2131-146. 36 An alternative specification, using the Erikson, McIver and Wright measure of public ideology (Erikson, Robert 8., Gerald C. Wright, and John P. McIver. 1993. Statehouse democracy .' public opinion and policy in the American states. Cambridge ; New York: Cambridge University Press. was also utilized, producing very similar results. The Berry, et al. measure is presented here because it does not exclude Alaska and Hawaii. 97 demographic characteristics of the states are also included in the model: educational attainment (percent of population with a college degree), unemployment rate, and population (logged). States with more educated populations should be more tolerant, and thus should be less likely to pass an anti-minority proposal (e. g. McClosky and Brill 1983). Economic bad times may increase anti-minority sentiment and thus increase the likelihood of passage (Tatalovich 1995). Finally, following arguments from the Federalist Papers (Hamilton, Madison, and Jay [1787] 1999), larger states should be better able to overcome the problems of majority factions and thus less likely to pass anti- minority proposals. The last group of control variables account for the attributes of the anti-minority proposals themselves.37 Constitutional amendments often require superrnajorities for approval, and should be less likely to pass than proposals that can pass with a simple majority. Thus, the models include the proportion of votes required to pass (0.5, 0.6, .67, or 0.75). Proposals that are competing for passage with several other similar pr0posals should also be less likely to pass. Finally, I control for the minority group that is being targeted by the proposal. In the last decade, many more anti-homosexual policies have been adopted than other anti-minority proposals. Thus, proposals that target homosexuals should be the most likely to pass. Since only a few states have adopted anti-affirmative action policies, these proposals should be least likely to pass of the three targeted groups. 37 Alternative determinants of proposal passage that were considered but not presented here include political culture, income per capita, and region. These variables did not add any additional explanatory power to the models and excluding them did not alter the interpretation of the direct democracy coefficients. 98 3.2.1 Results The results from these models, estimated using a logistic regression with robust standard errors clustered on the state,38 are presented in Table 3.7. Given the binary nature of the dependent variable and the relatively low percentage of proposals that passed, the four models fit relatively well, and are certainly improvements over the null prediction.39 As expected, the direct democracy impact coefficient is positive and significant, providing further support to the findings from the previous analysis. The model shows that states with high impact direct democracy institutions are more likely to pass anti- minority proposals than both states without direct democracy and states with relatively low‘impact direct democracy institutions. The models also show some other significant predictors of anti-minority proposal passage. Republican Party control increases the likelihood of passing these proposals. Legislatures with bill introduction limits are marginally more likely to pass these anti- minority bills. Citizen ideology also has a marginally significant effect in three of the models, suggesting that more liberal states are less likely to pass anti-minority proposals. 38 With a binary dependent variable, logistic regression is appropriate in this case (Long, J. Scott. 1997. Regression models for categorical and limited dependent variables, Advanced quantitative techniques in the social sciences ,' 7. Thousand Oaks: Sage Publications. Since anti-minority proposals introduced in the same state may be related in some unmeasured way, robust standard errors, clustered by state, were calculated. 39 Alternative goodness-of-fit measures, like other pseudo R2 statistics, the Hosmer-Lemeshow )8 and the proportional reduction in error X, all suggest that the four models fit the data well and are improvements over the null prediction of the modal category (not pass). 99 Table 3.7 Determinants of the Passage of Anti-Minority Bills, 1995-2004 Variable Coefficient RobusEtrSCt’a'ndard Direct Democracy Impact [+] 0.208“ 0.083 Republican Control [+] 0.923“ 0.366 it; Democratic Control H 0.079 0.317 g Party Competition H -0.552 2.424 g Professionalization H -1.188 1.921 g Size of Legislature H 0.001 0.003 8 Introduction Limits [+] 0.500# 0.370 Citizen Ideology H -0.018# 0.013 g Minority Diversity H 0.206 1.362 g * Educational Attainment {-1 -0.025 0.043 g Unemployment [+] -0.042 0.130 <0 Population (logged) H —0.1 14 0.283 Supermajority Required {-1 -1.841 2.486 g 3 Similar Bills in Session H -0.367** 0.110 03% Targets Homosexuals [+] 2.125“ 0.561 <1 Targets Foreign Speakers [+] 1.159* 0.612 Constant 0.125 2.907 N 595 595 Log pseudo likelihood -199.660 -199.751 McKelvey & Zavoina R2 0.281 0.281 Note: Expected signs of the coefficient in brackets; # significant at 0.1; * significant at 0.05; “significant at 0.01; one-tailed tests where appropriate 100 In addition, the attributes of the proposal have significant impacts on the likelihood of passage. As the number of similar proposals considered in the same session increases, the likelihood that of any one of them passes decreases. There are also significant differences in the likelihood of passing depending on the minority group that is targeted. Proposals that target homosexuals and foreign speakers are more likely to pass than proposals that target minorities in general. The difference between proposals that target homosexuals and proposals that target foreign speakers is also statistically significant. Even when accounting for other determinants of bill passage like party control and public ideology, direct democracy significantly increases the likelihood of an anti- minority proposal passing. The substantive effect of direct democracy institutions can be more clearly seen by examining the predicted probabilities, as shown in Figure 3.4. Compared to non-direct demOcracy states, a high impact direct democracy state like California is almost three times as likely to pass an anti-minority policy proposal. Though the predicted probabilities never rise above 0.3, a passage rate of 10-30 percent is not particularly low for legislative passage rates in the American states (Squire and Hamm 2005). The important aspect of these findings is not that anti-minority bills have exceptionally high probabilities of passing in direct democracy states, but rather that they have a significantly higher probability of passing compared to similar bills in non-direct democracy states. Again, it is this comparison between the two types of systems that most directly tests the “tyranny of majority” critique. 101 Figure 3.4 Predicted Probabilities of Anti-Minority Bill Passage I". Predicted Probability of Passing — i I . i 0 2 4 6 Direct Democracy Impact Note: Predicted probabilities. generated usingClarify (King, Tomz, and Wittenberg 2000) 3.3 Discussion The argument that direct democracy endangers minority rights, as discussed earlier, is a comparison between two institutional arrangements. It is not that citizen legislative institutions necessarily have a negative impact on minority rights, but rather that they are relatively more likely to have this type of policy effect than representative democratic institutions. It follows that the hypothesis that derives from this argument is, itself, comparative as well. Namely, states with direct democracy institutions are more likely to adopt policies that restrict the rights of minority groups than states without direct democracy institutions. Thus, the most direct way to test this hypothesis is to compare 102 the outcomes of anti-minority policy proposals in governments with the two types of institutional arrangements. As undertaken in this study, the results are unambiguous. From 1995 to 2004 states with direct democracy institutions passed anti-minority proposals at a significantly higher rate than states without direct democracy. By extending the analysis to not only examine instances of policy adoption, but also instances of failed policy proposals, it is clear that anti—minority proposals in direct democracy states fared much better than comparable proposals in states without direct democracy. Furthermore, the finding that states with traditional, representative democracies filter out a higher percentage of anti- minority proposals than their direct democracy counterparts extends across the three targeted minority groups under examination here. By comparing all anti -minority proposals introducedin direct democracy states to the anti-minority proposals introduced in states without direct democracy institutions, the study was also able to account for both the direct and indirect effects of citizen legislation. Where previous research on direct democracy and minority rights tended to focus on the direct outcomes of citizen legislation, this study also examined legislative bills that target minority groups in direct democracy states in addition to the proposals considered by voters on the ballot. In doing so, the total impact of direct democracy institutions on the likelihood of an anti-minority policy proposal passing was clear: the presence of direct democracy in a state increases the probability of passing an anti- minority proposal. 103 CHAPTER 4: THE FLIP SIDE; DIRECT DEMOCRACY & PRO- MINORITY POLICIES So far, I’ve examined how anti-minority policies fare in states with direct democracy mechanisms as compared to similar policies in states without these institutions. In looking at policies that explicitly restrict the rights of minority groups, the results have been unambiguous: States with direct democracy are more likely to pass anti-minority policies than other states. While these results support the tyranny of the majority argument, that the rights minority groups are endangered under pure democracy systems, they do not tell the entire story of minority rights and direct democracy. Just as states can pass laws that restrict minority rights, they can also adopt policies that protect and enhance the rights of minorities. Although direct democracy states pass anti-minority measures at higher rates than non—direct democracy states, and are more likely to adopt specific anti-minority policies, it is not clear from the previous chapters or from the existing literature how pro-minority policies fare under direct democracy systems. The very language of pro-minority policies, like non-discrimination and civil rights, may engender more widespread support for these policies than opposition to anti-minority policies. If this is the case, direct democracy may actually increase the likelihood of passing these types of policies. An important caveat to Madison’s concerns about tyranny of the majority is the extent to which the majority group is unified in purpose against minorities. If pro-minority policy proposals do not unify the majority in opposition, but rather unify them in support, then direct democracy may actually facilitate the adoption of these policies. This could act as a balance to the tyrannical outcomes seen in the previous chapter. Alternatively, the 104 tyrannical effect of direct democracy that was evident in previous chapters may continue to hold for pro-minority policies, which would put the rights of minorities under these systems at further risk. Currently, it is not clear how direct democracy impacts the adoption of pro-minority policies. Little, if any, work has focused solely on policies that protect or expand the rights of minorities. Previous research on direct democracy and minority rights has tended to aggregate both pro and anti-minority policy proposals together and evaluate the direction of the outcomes (Donovan and Bowler 1998a; Gamble 1997; Hajnal, Gerber, and Louch 2002). One recent study of gay rights legislation did separate the two types of policy proposals (Haider-Markel, Querze, and Lindaman 2007). This study found that pro—gay measures tended to outnumber antigay measures in state legislatures, but anti-gay measures outnumbered pro- gay measures on ballot measures. However, in examining the outcomes of these policy pr0posals, the authOrs aggregated both types of gay-related legislation. They then evaluated the direction of the outcome (pro or anti-gay), but did not account for whether the proposal was intended to protect or restrict gay rights. Thus, it remains unclear if direct democracy impacts pro-minority measures differently than anti- minority measures. To evaluate the impact of direct democracy on the adoption of policies that explicitly protect or expand minority rights, I examine four specific pro-minority policies using a similar approach to the analyses in the second chapter on anti-minority policies. The first section evaluates the expansion of state non-discrimination policies to include sexual orientation which protects gay rights. The next section examines racial profiling bans which protects several racial and ethnic minorities. The third section undertakes an 105 analysis of hate crimes laws, which offer expanded protection to most minority groups. The final section examines a particular set of state hate crime laws — those that include sexual orientation. In each section I evaluate the impact of direct democracy on the adoption of these specific policies by comparing direct democracy states to non-direct democracy states. To account for the total effect of direct democracy, including both the direct and indirect effects, the study encompasses pro-minority policies passed through both citizen and traditional legislation. 4.1 Sexual Orientation & Non-Discrimination Laws One of the most fundamental pro—minority policies a state can pass is a non- discrimination law. These civil rights policies outlaw discrimination in employment, housing, education and public accommodations. Everys‘tate has some form of a non- discrimination law. Most states originally barred discriminatory practices based on race, ethnicity, gender and religion. However, since the late 1980’s many states have expanded them to include sexual orientation and sexual identity. By 2007, twenty states had laws on the books that outlaw discrimination based on sexual orientation. The inclusion of sexual orientation in non-discrimination laws was a focal point in the gay rights movement of the 1970’s. The gay rights movement had gained momentum following the Stonewall Riots in 1969, sparking the organization of more than 800 homosexual political groups by 1973 (Button, Rienzo, and Wald 1997). At the national level, congressional supporters sought to expand the 1964 Civil Rights Acts to include sexual orientation, but were unsuccessful. The movement found more success at the local level. In 1973, East Lansing, Michigan became the first city to ban discriminatory 106 practices against homosexuals. San Francisco followed two months later with its own non-discrimination law. Today, over one hundred local governments include sexual orientation in their non-discrimination policies. Despite the adoption of these policies at the local level, however, state and federal lawmakers continued to resist passing similar laws. It was not until 1982 that Wisconsin became the first state to pass a non- discrimination law that extended protection to homosexuals. Seven more years would pass before another state followed suit. After a backlash against the gay rights movement in the late 1970’s, the debate shified away from anti-discrimination policy to center on the AIDS epidemic during the 1980’s. Though no state adopted this type of non- discrimination policy during this period, there was growing acceptance of gay rights in US. By the end of the Eighties, the focus of the movement shifted back towards anti- discrimination policy. From 1989 to 1999, eight states added sexual orientation to their . nondiscrimination policies. Eight more have passed these policies from 2000 to 2007. Figure 4.1 shows the states that have adopted non-discrimination policies that include sexual orientation.40 The patterns of yearly and cumulative adoptions of these policies at the state level are presented in Figure 4.2. The flat line through much of the 1980’s shows the reluctance of state policymakers to follow Wisconsin’s initial adoption. However, a more traditional pattern of policy diffusion emerges in the Nineties, shown by a clear S-shape in the cumulative adoptions from 1989 to 1995. In the late 1990’s the adoption rate slowed to one state every other year, but the more recent increase in the cumulative line suggests a new burst of policy adoption. ’0 Tables showing the yearly adoption of all four pro-minority policies are presented in Appendix E. 107 Figure 4.1 States with Non-Discrimination Laws that include Sexual Orientation 108 Figure 4.2 State Adoptions of Non-Discrimination Policies, 1982-2007 8 s / I ,l ,«v to _ f ‘- _ / .33. {e l (U , ‘75 H]! "6 g _ ,s—e’ g H E X" "‘ z ? / LO — / f / f «t 9 He» + O I I I I 1980 1990 2000 2010 year —0— Yearly Adoptions --+-- Cumulative Adoptions 4.1.1 Direct Democracy and the Adoption of Non-Discrimination Policies From the Figures 4.1 and 4.2, it is clear that the spread of policies to protect homosexuals from discrimination has been relatively incremental since 1980. Nonetheless, forty percent of the states have adopted these policies during this time period. So what role, if any, has direct democracy has played in the diffusion of pro- minority policies? In terms of a direct effect of direct democracy institutions, only six ballot measures addressing the inclusion of sexual orientation in non-discrimination policies have been considered since 1970 (shown in Table 4.1). 109 Table 4.1 Ballot Measures Addressig Sexual Orientation & Non-Discrimination State Year Ballot Measure Outcome Colorado 1992 Amendment 2: Initiative to amend 0 , the state Constitution to prohibit the Passed (53'4 /°)' IDCIUSIOI‘I. of sexual orientation in A nti-Minon'ty nondiscrimination polimes Maine 1998 Question 1: Popular referendum to reject a ban on discrimination based Passed (51.3 %): on sexual orientation passed by the Legislature and signed by the Anti-Minority Governor Maine 2000 Question 6: Legislative referendum to ratify a non-discrimination law Failed (45 %): including sexual orientation that was passed by the Legislature and Anti-Minority signed by the Governor Maine 2005 Question 1: Popular referendum to reject a ban on discrimination based Failed (45 %): on sexual orientation passed by the Legislature and signed by the Pro-Minority Governor ‘ - Oregon 1988 Measure 8: Initiative to revoke a ban Passed (52.8 %) on sexual orientation discrimination in the state executive branch Anti-Minority Washington 1997 Initiative 677: Prohibits discrimination based on sexual orientation Failed (40.3 %) Anti-Minority Source: National Conference of State Legislatures Five of these measures actually sought to restrict the rights of homosexuals by explicitly banning the inclusion of sexual orientation in the state’s non-discrimination policies or by vetoing a policy passed by the government. The only measure considered to include sexual orientation in the state’s non-discrimination law was defeated handily. The only pro-minority outcome was the failed popular veto of Maine’s non-discrimination policy 110 in 2005. This was third time the Maine government had passed this policy, but the previous two popular veto attempts were successful. Thus, for state non-discrimination policies, it seems that the only direct effect of direct democracy may be to inhibit their passage. Beyond the just direct influence of citizen legislation, other evidence also suggests that direct democracy states may be slightly less likely to adopt these laws than non-direct democracy states. Of the twenty states that currently have a non- discrimination policy that includes sexual orientation, only eight of them are direct democracy states. Only a third of direct democracy states have adopted this policy, compared with forty-six percent of non-direct democracy states. While these simple comparisons are far from conclusive, they do support the tyranny of the majority argument. To more closely examine. the relationship between direct democracy and the I, adoption of non-discrimination laws that include sexual orientation, it is necessary to also consider other determinants of policy adoption. As in Chapter Two, I use event history analysis to answer the following question: Given that a state has not adopted a nondiscrimination policy that includes sexual orientation in previous years, what is the probability that it will do so in that year? Yearly data has been collected on forty-nine states from 1982 to 2005."1 The analysis begins in 1982 with the first adoption of a state-level non-discrimination policy that includes sexual orientation by Wisconsin. Since Colorado, Oregon, Iowa and Washington have all adopted these policies after 2005, the dataset is clearly right- “ Nebraska is excluded because of its non-partisan legislature. Models that include Nebraska and are estimated without party variables do not significantly change the results regarding the effect of direct democracy. 1]] censored. Due to the nature of the data, a Cox Proportion Hazards model that uses the exact discrete method for ties is, again, well-suited to address this question. The dependent variable in this analysis is a dichotomous indicator of whether or not a state adopted a non-discrimination policy that includes sexual orientation in a given year. Once a state adopts its initial policy to protect the rights of homosexuals, it is dropped from the dataset. Thus, states like California, Rhode Island and New Jersey, who all initially adopted protections based on sexual orientation and later passed protection for gender identity or expression, are not included in the dataset after their initial adoption. As with the previous analyses, the key independent variable in this analysis measures the impact of direct democracy institutions in the state. Again, this variable was generated from a principle components analysis of four direct democracy measures: a dichotomous indicator of direct initiatives, the Legislative Insulation Index, the Qualification Difficulty Index, and a count of the ballot initiatives in a state over the time period in question (1982 to 2005). A negative coefficient on this variable would suggest another negative impact on the rights of minorities, while a positive coefficient would suggest that direct democracy may have a positive effect in for this pro-minority policy. I also control for several other potential determinants of policy adoption.42 Like the previous studies in this project, this analysis controls for citizen ideology, party, geographic diffusion from neighboring states, population, and education levels.43 Given the minority group in question, I also control for the urban population as a proxy for the organizational strength of the gay rights movement. Urban states should be more likely ’2 Political culture, racial diversity, legislative professionalization, per capita income, and region were included in alternate models but they added no significant explanatory power to the models and are not presented here. 43 The variable indicating unified Republican government is not included in the models because it perfectly predicts failure to adopt the policy and estimates of the standard error cannot be estimated. 112 to adopt protections based on sexual orientation than states with more rural populations. A dichotomous indicator for whether or not that state has a sodomy ban (prior to Lawrence v. Texas, 2003) is also included to measure the hostility of the legal environment towards homosexuals. States with sodomy laws should be less likely to adopt protections based on sexual orientation. 4.1.2 Event History Results The results from the Cox proportional hazards analysis of the adoption of non- discrimination policies that include sexual orientation are shown in Table 4.2. Like the previous analysis, the coefficients can be interpreted similar to conditional logistic regression coefficients, as the effect of a one unit change in the independent variable has on the log-odds of a state adopting this type of non-discrimination policy in a particular year. The results from this analysis are consistent with the findings from the analyses of anti-minority policies in that direct democracy states are more likely to make a policy decision that has negative consequence for the minority group. The coefficient on the direct democracy impact variable is negative and marginally significant, suggesting that direct democracy states are less likely to adopt anti-discrimination policies that cover sexual orientation. Again, this measure of direct democracy institutions also reveals differences among direct democracy states. On average, states with lower impact direct democracy institutions seem more likely to adopt this pro-minority policy than states with relatively higher impact direct democracy institutions. 113 Table 4.2 Adoptions of Non-Discrimination Measures Protecting Sexual Orientation Variable Coefficient Standard Error Direct Democracy Impact H -0.271# 0.167 Citizen Ideology [+] 0.154“ 0.034 Democratic Party Government [+] 0.203 0.653 Party Competition [+] 4.140 4.431 Education [+] 0.031 0.066 Percent Urban [+] 0.050* 0.023 Population [+] -0.350 0.277 Neighboring States Adopted [+] -0.1 17 0.197 Sodomy Law H 0.315 0.711 Observations 1043 Log Likelihood -61.286 Notes: Expected direction in brackets; one-tailed tests used where appropriate, two-tailed tests otherwise; # significant at 10%; " significant at 5%; " Significant at 1% Looking at the control variables, the adoption of this type of non-discrimination policy is driven entirely by citizen ideology and the organization strength of the gay rights movement in the state. As expected, more ideologically liberal states are more likely to adopt protections based on sexual orientation. Similarly, more urban states — where gay rights groups traditionally have the highest levels of organization and support — are more likely to adopt these measures. From both the event history perspective and the descriptive comparisons, direct democracy states are less likely than states with purely representative governments to adopt non-discrimination policies that provide protection on the basis of sexual orientation. Although the ballot measures have only been utilized a relatively few 114 number times to address these issues, and thus have had little direct effect, the overall effect of direct democracy has been to decrease the likelihood of adoption and subsequently decrease the likelihood of legal protections for homosexuals. 4.2 Racial Profiling Bans The preceding examination of non-discrimination polices that include sexual orientation falls in line with the analyses regarding the rights of homosexuals under direct democracy institutions from Chapters Two and Three. Direct democracy states are more likely to produce an anti-homosexual outcome than other states whether they’re considering an anti-homosexual measure or a pro-homosexual measure. It seems that direct democracy has a consistently negative impact on gay rights, but does this consistent effect extend to other types of pro-minority policies? In this section, I examine a policy that seeks to protect the rights of a different set of minority groups: racial profiling. The American Civil Liberties Union (ACLU) defines racial profiling as “the discriminatory practice by law enforcement officials of targeting individuals for suspicion of crime based on the individual's race, ethnicity, religion or national origin” (ACLU 2005). Awareness of this issue was raised in 1998 by a US. Department of Justice investigation of the New Jersey State Police for singling out members of racial and ethnic minorities for traffic and other minor offenses in order to search them for drugs and guns (Racial Profiling Data Collection Resource Center 2007). Following the investigation, several states moved to adopt measures to address racial and ethnic profiling in law enforcement. 115 In general, states have addressed this issue through explicit bans of racial profiling practices and through data collection orders (both voluntary and mandatory). The collection of racial and ethnic data during traffic and pedestrian stops is aimed at assessing the extent of racial profiling problems, as well as monitoring law officers’ behavior to discourage biased policing. Despite this discouragement, data collection policies do not prohibit racial profiling. These policies vary significantly across the states, with some mandating collection by all officers and others asking for voluntary collection from certain jurisdictions. The more explicit prohibitions of racial profiling tend to be more straight-forward and are much more comparable across the states. Thus, for the sake of comparability across states, this study will focus on these explicit bans. Figure 4.3 State Adoptions of Racial Profiling Bans 116 After the Department of Justice’s investigation of the New Jersey State Police, Connecticut became the first state to adopt an explicit racial profiling ban in 1999, with three more states passing bans in 2000. In the next few years, there was a flurry of adoptions, and by 2005 twenty-four states had explicitly banned racial profiling (seen in Figure 4.3). Figure 4.4 shows the yearly and cumulative patterns of adoption, with the S- shaped curve of policy diffusion evident in the cumulative adoptions line. Figure 4.4 Yearly Adoptions of Racial Profiling Bans, 1998-2005 g _ ’I. J’/ o x" N ‘ / / / / (I) a /" (0 l0 ' 1 / 31'5" / u— / 0 fi/ 5 a g / E 53 - / 3 / z I I I I I 1 998 2000 2002 2004 2006 year —O— Yearly Adoptions - -¢~ - - Cumulative Adoptions 117 4.2.1 Direct Democracy and Racial Profiling Bans In the spread of racial profiling bans from 1999 to 2005, citizen legislation has not played a very discernible role. In fact, the direct effect of direct democracy has had even less of an impact on the adoption of racial profiling bans than on non-discrimination policies that include sexual orientation. Not a single measure relating to racial profiling has made it on to the ballot. However, in terms of comparing the adoptions in direct democracy states to other states, there does seem to be a difference between the two sets of states. Contrary to adoptions of non-discrimination policies that cover sexual orientation, racial profiling bans tend to be adopted more frequently by direct democracy states. Almost sixty percent of direct democracy states had adopted an explicit racial profiling ban by 2005, while less than forty percent of the other states had done the same. Of the twenty-four states that have adopted explicit bans, fourteen of them are direct democracy states. On the surface, it looks as if the “tyranny of the majority” argument does not hold in this case. To examine this relationship more closely, an event history analysis is again used to assess the impact that direct democracy institutions have on the likelihood of a state adopting a racial profiling ban, given that it had not already adopted one previously. Yearly data were collected on forty-nine states from 1999 to 2005.44 The analysis begins in 1999 with the first adoption of a state racial profiling ban by Connecticut. Due to the nature of this data, Cox Proportion Hazards models are employed again. The dependent variable is a dichotomous indicator of whether or not a state adopted the policy in a ’4 Nebraska is excluded because of its non-partisan legislature. Models that include Nebraska and are estimated without party variables do not significantly change the results regarding the effect of direct democracy. 118 given. When a state adopts its initial racial profiling ban, it is subsequently dropped from the analysis. The independent variables in these models are the same as in the previous section, for the most part.45 The primary variable of interest is the direct democracy impact measure. The control variables include citizen ideology, party competition and control of govemment, education, population, and regional diffusion. Since racial profiling laws do not affect the rights of homosexuals, the models do not include the sodomy law variable ’ and the percent urban p0pulation variable fiom the previous analysis on non- discrimination policy. The models, however, do include a variable measuring the minority diversity of the state since racial profiling policies directly affect ethnic and racial minorities. Based on Hero’s (1998) diversity measure, this variable gauges how diverse the states’ populations are in terms of Afiican Americans, Hispanics, and Asians , relative to the white population. I expect that higher diversity would increase the base of support for minority protections and should increase the likelihood of the adoption of a racial profiling ban. 4.2.2 Event History Results The estimates from the four Cox Proportional Hazards models are shown in Table 4.3. As with the initial comparisons, the tyranny of the majority argument is not supported in this case. Not only does direct democracy not inhibit the passage of these policies that protect minorities, it actually seems to enhance the likelihood of their ’5 Again, political culture, racial diversity, legislative professionalization, per capita income, and region were included in alternate models for all of the analyses in this chapter. but added no significant explanatory power to the models and are not presented here. 119 passage. The direct democracy impact coefficient is positive and statistically significant. This suggests that the higher the direct democracy impact, the more likely the state will ban racial profiling by law enforcement agencies. Table 4.3 State Adoptions of Explicit Racial Profiling Bans, 1999-2005 Variable Coefficient Standard Error Direct Democracy lmpact[-] 0.371 ** 0.1 15 Citizen Ideology [+] -0.020 0.018 Party Competition [+] -2.120 3.836 Republican Party Government H -1.331* 0.706 Democratic Party Government [+] 0.236 0.552 Minority Diversity [+] -1.044 1.831 Education [+] -0.168** 0.059 Population"[+] ' 0.1-07 0.281 Neighboring States Adopted -0.718** 0.223 Observations 323 Log Likelihood -81.579 Notes: Expected direction in brackets; one-tailed tests used where appropriate, two-tailed tests otherwise; * significant at 5%; ** significant at 1% Though these results may be contrary to expectations derived from the anti- minority critiques of direct democracy, they make more sense in light of one of the key assumptions in this argument. The “tyranny of the majority” argument is essentially a special case of a more general majoritarian effect of direct democracy. It assumes that pro-minority policies would not be favored by a majority of the population, thus making these policies less likely to pass. However, for racial profiling laws this assumption does 120 not hold. First, racial profiling affects many different minority groups that may be able to generate a broader base of support that could approach a majority. Second, in many cases the majority group may not be “united in a common interest” against the minority groups as Madison writes in his discussion of the problems of majority factions (Madison [1787] 1999). In the case of racial profiling, an overwhelming majority of Americans seemed to support banning the practice. In 1999, following several years of news coverage of highway stops of innocent, minority drivers, a Gallup poll found that eighty-one percent of respondents disapproved of racial profiling (Pampel 2004). With such overwhelming support for racial profiling bans, the results from this analysis suggest that the impact of direct democracy institutions is to make government more responsive to public preferences. These results are more in line with a generalized majoritarian impact argument than a strict tyranny of the majority argument. Moving down the models, the significant control variables tend to meet expectations. Republican governments are less likely to adopt racial profiling bans than divided or Democratic governments. States with higher levels of education are more likely, on average, to adopt these policies. Finally, there is a significant and negative regional diffusion effect. As the number of neighboring states with a racial profiling ban increases, the likelihood of a state adopting its own decreases. The evidence from this examination of racial profiling bans in the American states suggests that, contrary to the tyranny of the majority argument, minorities may actually benefit from direct democracy institutions in some cases. In the case of racial profiling bans, which affect many minority groups and has overwhelming public opposition, direct 121 democracy institutions seems to have increased the responsiveness of state government on this issue. Still, on other issues, like same-sex marriage bans or non-discrimination policies, it is evident how this increased responsiveness could be hannfirl to the rights of minority groups - particularly unpopular minority groups. 4.3 Hate Crimes Laws In the first two sections of this study, empirical tests of the impact of direct democracy on the adoption of explicitly pro-minority policies have produced seemingly contradictory results. In the case of non-discrimination policy, which affects the rights of homosexuals, direct democracy decreased the likelihood of a pro-minority outcome. When examining racial profiling bans, which affects the rights of racial and ethnic minorities, direct democracy increased the likelihood of a pro-minority outcome. In this section, I examine another issue, hate crimes, which affects gay rights as well as the rights of racial and ethnic minorities. The Federal Hate Crimes Statistics act of 1990 defines hate crimes as “crimes in which the defendant’s conduct was motivated by hatred, bias, or prejudice based on the actual or perceived race, color, religion, national origin, ethnicity, gender, or sexual orientation of another individual or group of individuals” (U .8. Congress 1992). This differentiation in crimes based on the motivation of the perpetrator is necessary, advocates argue, because bias or hate crimes cause special injury to victims because of the knowledge that their own race (or ethnicity or religion or sexual orientation) was the prime motive for the crime committed against them. Furthermore, hate crimes can send a symbolic message of fear and terror to members of the victim’s community (Altschiller 122 2005). For these reasons, hate crimes are viewed as more egregious than other crimes. States have responded by enacting hate crimes laws which enhance penalties for existing crimes and/or create new categories of crimes like “ethnic intimidation” or “malicious harassment”. Although hate crimes have occurred throughout American history, the differentiation from other crimes motivated by other factors is a relatively recent development. The movement to adopt hate crimes policies began at the convergence of several minority group movements, like the civil rights movement and the gay rights movement, and the victims’ rights movement in the late 1960’s. With this issue convergence, a broad coalition of groups pressed for government action on the issue of hate crimes. In particular, the Anti-Defamation League of B’nai B’rith (ADL) pushed for governments to adopt its model hate crime legislation —- which was unveiled in 1981. That same year, Oregon and WaShington became the first states to adopt a broad hate crime statute that recognized, defined, and respond to discriminatory violence (J enness and Grattet 2001 ). While states have used various approaches to address hate crimes, from the modification of existing statutes to the creation of new and freestanding hate crime statutes, the core policy concept of enhancing penalties for these types of crimes quickly diffused to other states. By 1990, twenty-eight states had enacted a hate crime statute (see Figures 4.5 and 4.6). The pattern of adoptions does not follow the classic s-shape, as shown in Figure 4.6, but rather there seems to be several distinct periods of policy adoptions. The first peaks in 1982 and subsides by 1985. The second peaks in 1989 and subsides by 1991. The most recent period has seen fairly steady pattern, with a few states adopting hate 123 crimes legislation each year until 2004. These three waves of adoption are evident in both the cumulative and yearly adoption curves. By 2007, only five states had failed to enact a hate crime law.“ Figure 4.5 States with Hate Crime Laws (D ’ m Adopted 1981-I985 Adopted 1986-1990 0 |||||||||l||l|||| am 4" A complete list of the yearly adoptions of hate crimes laws in present in Appendix E. 124 Figure 4.6 Yearly Adoptions of Initial Hate Crimes Laws, 1980-2005 o _ l0 {fire-ewe tea—er" 9r ‘ If?” ,4-e’ a .f“ .9 O 4 46’ (I) m g»? “5 / a 7’ s8“ .7 z x at) 2 ~ ,/ gar-9’ W O —I I I I I I I 1980 1985 1990 1995 2000 2005 year —O— Yearly Adoptions --O-- Cumulative Adoptions 4.3.1 Direct Democracy and Hate Crimes Laws Like the diffusion of racial profiling bans, ballot measures have not been used to adopt hate crimes laws in the US. Thus, there does not seem to be a direct effect in this case. Comparing the number of adoptions in direct democracy states to non—direct democracy states also suggests no significant effect of direct democracy on the adoption of hate crimes laws. Almost every state has adopted a hate crime law, regardless of the presence of direct democracy institutions. Of the five states that do not have a hate crime law, three are direct democracy states and two are non-direct democracy states. 125 However, if we break down the adoptions into the three periods (1980-1985, 1986-1990, 1990-2005), there is some indication of differences between the two types of states. In the first period, direct democracy states constituted seven of the ten adoptions, but in the second period they constituted only eight of the eighteen adoptions. Of the last seventeen adoptions, six were direct democracy states. Direct democracy states were among the earliest innovators and adopters of hate crime policies (6. g. OR, WA, CA), accounting for seventy percent of the initial adoptions in the 1980’s. Non-direct democracy states caught up in later periods. Still, it is not clear from this examination whether the early spate of adoptions by direct democracy states was caused by these institutions or by some other policy determinant such as citizen ideology. To more explicitly test the effect of direct democracy on the likelihood of a state adopting a hate crime law, while also controlling for other factors, it is once again necessary to undertake an event history analysis. As with the previous analyses, the dependent variable is a dichotomous indicator of whether or not the state adopted the policy in a given year. Once the state adopts a hate crime policy, it is subsequently dropped from the analysis. Data on forty states was collected from 1981 to 2005.47 The time period under analysis begins in 1981, when Oregon and Washington became the first states to adopt a hate crime policy. The primary explanatory variable of interest is, once again, the index of direct democracy impact based on a principle components analysis of the four measures of direct democracy institutions. The rest of the independent variables follow the models from the racial profiling analysis. ’7 Nebraska is excluded because of its non-partisan legislature. Models that include Nebraska and are estimated without party variables do not significantly change the results regarding the effect of direct democracy. 126 4.3.2 Event History Results The results are shown in Table 4.4. The coefficient on the direct democracy impact variable confirms the slight differences between the two types of states discussed above. While, almost every state had adopted a hate crime law by 2005, direct democracy states show a higher propensity to adopt this policy for any given year when other factors are controlled for. The results also convey differences between the various institutional arrangements of the direct democracy states. States with less insulated legislatures, with easier qualification requirements, and which use ballot measures more often, seem to be more likely to adopt a- hate crimes law in a given year. Table 4.4 State Adoptions of Hate Crimes Laws, 1981-2005 Variable Coefficient Standard Error Direct Democracy Impact[-] 0.226* 0.093 Citizen Ideology [+] i 0.044“ 0.015 Party Competition [+] -3.945 2.414 Republican Party Government H -1.172* 0.607 Democratic Party Government [+1 0.054 0.455 Minority Diversity [+] -4.237** 1.338 Education [+] 0.1 11* 0.055 Population [+] 1 .173** 0.275 Neighboring States Adopted [+] -0.594 0.136 Observations 573 Log Likelihood -1 75.991 Notes: Expected direction in brackets; one-tailed tests used where appropriate, two-tailed tests otherwise; * significant at 5%; ** significant at 1% 127 In addition to the effect of direct democracy, several control variables have significant impacts on a state’s likelihood of adopting a hate crime law. As expected, states with larger, more educated, and more liberal populations are more likely to adopt these policies. Republican governments are less likely than Democratic or divided governments to pass these policies. As in the previous models, the number of neighboring states that have adopted a hate crime law has a significant negative effect. The models do present two surprising results. First, there is a negative effect of minority diversity. States with relatively diverse populations seem to be less likely to adopt a hate crime law while less diverse states are more likely to adopt this policy. Another surprising result is the marginally significant and negative coefficient on the party competition variable. High party competition is usually associated with liberal outcomes as parties compete to provide government goods to ensure reelection (Barrilleaux, Holbrook, and Langer 2002; Key). In this case, however, high party competition decreases the likelihood of a liberal outcome — adopting a hate crime law. The results from this analysis tend to comport well with the results from the racial profiling examination. Like the racial profiling issue, hate crimes laws affect a broad range of minority groups and has a wide base of support. This base may be even wider and include members of traditional majority groups. According to the US. Department of Justice, almost twenty-one percent of hate crime victims in 1992 and just less than twenty percent in 1999 were targets of anti-white bias (1994; 2001). This is the second highest percentage of hate crime victims. The same reports show that white males constitute forty percent of the victims in the most serious hate crimes. Beyond this wide base (and probably due to it) state hate crime laws have also received strong public 128 support. A Gallup poll conducted in 1999 found that seventy percent of respondents would favor a hate crime law in their state. With this strong and wide-ranging support, the positive effect of direct democracy makes sense. Again, these institutions are designed to make government more responsive to the public preferences. In the cases of racial profiling and hate crimes laws, policies that benefit minorities in general, direct democracy can help protect minority groups. 4.3.3 Hate Crime Laws, Sexual Orientation & Direct Democracy So far in this chapter, the examinations of the adoptions of pro-minority policies have produced two distinct results that seem dependent on the policy issue at hand. When the policy is broadly defined to protect a wide range of minority groups (like racial profiling or hate crimes) and has popular support, direct democracy increases the likelihood of a state adopting these types of measures; However, when the policy is more narrowly defined to protect a single, smaller minority group and subsequently receives less popular support, direct democracy may have the opposite effect and decreases the likelihood of a state adopting these policies. Again, these results fit the broader majoritarian impact of direct democracy where policies favored by the majority of the public are adopted no matter how they affect minority rights. In this section, I test this pattern by examining hate crime policies again. In the first analysis I examined the adoption of a state’s initial hate crime policy, no matter how the protected classes were defined. In this analysis, I’ll narrow my focus to hate crime policies that include sexual orientation. With a sole focus on providing protections for a narrower and less publicly supported minority group, attitudes towards this policy should 129 be less positive. If the pattern evident from the first three analyses holds, states with direct democracy institutions should be less likely to adopt hate crime policies that include sexual orientation than other states. Hate crime policies that include crimes motivated by bias based on sexual orientation are far less widespread than hate crimes policies that protect more traditional classes of groups based on race and ethnicity (see Figure 4.7). Figure 4.7 States with Hate Crime Laws that Include Sexual Orientation % . ,, 1964-1994 D Adopted 1995-2005 130 Still, by 2007, there were thirty states with hate crimes laws that cover crimes motivated by a victim’s sexual orientation (HRC 2007). 48 Many of the earliest states to adopt hate crimes policies did not include sexual orientation in their initial policy, but most added it later on.49 The first state to adopt a hate crime policy that included sexual orientation was California in 1984. As seen in Figure 4.8, twelve more states followed in the next decade by adding sexual orientation to their existing hate crimes law or by including this class of crimes in their initial policy. Figure 4.8 Yearly Adoptions of Hate Crimes Laws that include Sexual Orientation a - I :4 /»’ it / 9 g/ 19 a; a) // o a, ’41” ‘E’ ,,+-¢’ 2 53 - / I I T I I F 1 980 1 985 1 990 1 995 2000 2005 year ——0— Yearly Adoptions -—+— - Cumulative Adoptions ’8 Arizona has a hate crime policy that includes sexual orientation for data collection only. For this reason it is not included in the analysis. 49 A list of the yearly adoptions of hate crimes laws that cover sexual orientation is present in Appendix E. 131 A second wave of adoptions followed from 1995 to 2005, punctuated by the high- profile killing of Matthew Shepard, a gay student at the University of Wyoming, in 1998. From 1995 to 2005, thirteen more states included sexual orientation in their hate crimes laws. Interestingly, Wyoming, a state with direct democracy institutions, is one of the few states without a hate crimes law, much less one that covers sexual orientation. Of the thirty states with hate crimes policies that include sexual orientation, twelve are direct democracy states. Only half of the direct democracy states have adopted this policy, while nearly seventy percent of non-direct democracy states done the same. Thus, on the surface it looks as if direct democracy states are less likely to adopt this more narrow pro-minority policy, but obviously firrther testing is necessary before any conclusions can be drawn. To test this difference between direct democracy states and non-direct democracy states, I estimate a Cox Proportional Hazards model. Data on forty-nine states from 1984 to 2005 were collected.50 The dependent variable in this analysis is a dichotomous indicator of whether a state adopted a hate crime policy that includes sexual orientation in a given year. Once a state adopts the policy it is subsequently dropped from the analysis. Each model is estimated using a different measure of direct democracy. The independent variables are the same as the models from the general hate crimes policy analysis, but also include three additional variables. As discussed in the section on non-discrimination policies that cover sexual orientation, the percent urban population of each state is included as a proxy measure of the organizational strength of the gay rights movement. Urban states should be more likely to adopt this policy than 50 Nebraska is excluded from the analysis because of its nonpartisan legislature. Models that include Nebraska and are estimated without party variables do not significantly change the results regarding the effect of direct democracy. 132 more rural states. A dichotomous indicator of whether a state has a sodomy law (prior to Lawrence v. Texas) is included to gauge the legal environment towards homosexuals. States with sodomy laws should be less likely to adopt this policy than states that had repealed their sodomy laws. Lastly, a dichotomous indicator of whether a state had previously adopted a general hate crimes law is included. A state with an existing hate crime law must act to grant these protections exclusively to homosexuals, and thus the policy proposal is focused solely on the rights of a single group. When a state does not have an existing hate crime law, its consideration of a measure that covers sexual orientation is focused simultaneously on several other groups. For this reason, states that already have a hate crimes law that does not include sexual orientation should be less likely to adopt the policy in question. 4.3;4 Event History Results The results from the four Cox Proportional Hazards models are presented in Table 4.5. In this case, the direct democracy impact coefficient is statistically indistinguishable from zero. This suggests that, unlike the previous three pro-minority policies, direct democracy does not seem to affect the likelihood of a state adopting a policy to cover sexual orientation under their hate crimes statute. This null result is surprising, given the majoritarian effects found in the previous analyses of pro-minority policies. This may reflect the contrasting popular support for hate crimes policies and the opposition to homosexuality and gay rights. Given these ambivalent public attitudes towards this policy, there may be no a clear majority to which direct democracy institutions can provide increased responsiveness. 133 Table 4.5 State Adoptions of Hate Crimes Laws that include Sexual Orientation Variable Coefficient Standard Error Direct Democracy Impact H 0.095 0.121 Citizen Ideology [+] 0.032* 0.018 Party Competition [+] 8.946“ 3.373 Republican Party Government H -1.270# 0.789 Democratic Party Government [+] 1.065* 0.478 Education [+] -0.014 0.052 Population [+] 0022 0.288 Percent Urban [+] 0.042* 0.023 Sodomy Law H -0.361 0.489 Hate Crime Law H -0.658 0.524 Neighboring States Adopted -0.142 0.187 Observations 953 Log Likelihood -126.534 Notes: Expected direction in brackets; one-tailed tests used where appropriate, two-tailed tests otherwise; # significant at 10%; * significant at 5%; ** significant at 1% With no significant impact of direct democracy institutions, the adoption of these policies is driven by citizen ideology, party competition and control, and the organizational strength of gay rights groups. States with relatively more liberal, urban citizenry are more likely to adopt hate crimes policies that include sexual orientation. Unlike the previous analysis, party competition increases the probability of a state adopting this policy. Party control also matters: Republican governments are less likely than divided or Democratic governments to adopt this policy. Conversely, Democratic governments are more likely to adopt these types of hate crime laws. There is also may 134 be negative impact of having previously adopted a hate crimes law. States that have previously adopted a general hate crimes law are marginally less likely to adopt a subsequent expansion to include sexual orientation. In assessing the impact of direct democracy on the adoption of hate crime policies that cover crimes motivated by a victim’s sexual orientation, the evidence points to a null effect. There is neither a negative nor positive impact that reflects any increased responsiveness to the majority preferences in this case. This may be due to a public ambivalence about this issue. On the one hand, the very term “hate crime” engenders public support for these types of policies, with general hate crime policies receiving wide public backing. On the other hand, including sexual orientation means extending this policy to cover a minority group that traditionally has received very little public acceptance or support. 4.4 Discussion In these analyses of the adoption of pro—minority policies, the impact of direct democracy has varied across the different policies. While the “tyranny of the majority” argument would predict a negative effect on all four policies, this expectation was only met for one pro-minority policy. This chapter revealed two pro-minority policy areas in which direct democracy actually increased the likelihood of adoption. These results, given the broad scope and the national popularity of the two policies, were suggestive of a more general majoritarian effect. When pro-minority policies are able to draw on a wider base of support, as in the cases of racial profiling bans and general hate crimes laws, direct democracy can aid in protecting the rights of minority groups. 135 The variation in the impact of direct democracy on the adoption of pro-minority policies evident in this chapter makes it clear that pro-minority policies have different dynamics than anti-minority policies. Where the anti-minority policies experienced a consistent negative effect on minority rights in direct democracy states, pro-minority policies tend to have a more general majoritarian effect that does not necessarily always endanger the rights of minorities. In examining these four pro-minority policies, two of the analyses revealed an increase in pro-minority outcomes for direct democracy states that is consistent with a majoritarian effect. Unfortunately without state-level measures of issue attitudes, this argument is difficult to address more closely. Clearly, more work on these issues is needed, especially studies that focus on the interaction between public preferences on particular pro-minority policies and direct democracy institutions. Still, the analyses from this chapter did shed some light on this subject matter. It is apparent that the dynamics of pro-minority policy adoption is highly variable and dependent on the factors like citizen ideology, public perceptions of the policy, and the organizational strength of the groups involved. This chapter also makes clear that the politics of pro-minority policies is different from anti-minority policies, where more consistent impacts of direct democracy are apparent. 136 CHAPTER 5: CONCLUSION Although concern over the rights of minority groups has long been part of the debate on direct or pure democracy institutions, the existing scholarly literature has yet to produce systematic and empirical evidence to either validate or contradict this argument. Indeed, previous research has generated findings that seem to support both sides of the debate depending on the study. On the whole, this line of research has presented a very mixed picture of how direct democracy affects policies that impact the rights of minority groups. On the one hand some work has shown what appear to be tyrannical outcomes that produce anti-minority policies. On the other hand, there is work that suggests that these outcomes are not more prevalent under direct democracy institutions. Throughout this project, I have argued that ambiguous results of most studies on this question are due to two critical limitations. First, they tend to only focus on the direct effects of direct democracy; on the policy outcomes of ballot initiatives and popular referenda. This approach does not account for the indirect effects of direct democracy that would manifest themselves in outcomes from traditional legislation in direct democracy states. In order to get a complete picture of the impact of direct democracy on the rights of minorities (or any other policy area for that matter), it is necessary to examine both citizen and traditional legislations. The second limitation on the existing research, which has led to an unclear view of the impact of direct democracy on minority rights, is the tendency for these studies to solely focus on states with direct democracy institutions. The “tyranny of majority” argument is a comparative critique. The claim is that representative democratic systems 137 are better able to protect minority rights than systems with direct democracy institutions. So in order to evaluate whether this claim is correct, it is necessary to directly compare policy decisions in the two types of institutional arrangements. Taken together, these two extensions — accounting for both direct and indirect effects and comparing direct democracy states to states without direct democracy institutions — have provided for a more complete picture of the of the impact of citizen legislative institutions on minority rights. 5.1 Summary of Findings Utilizing these two critical extensions, the analyses in Chapters Two and Three produced consistent and clear results that support the anti-minority critique of direct democracy institutions. Chapter Two examined the effect of direct democracy on the adoption of explicitly anti-minority policies and found that state with direct democracy . institutions were significantly more likely to adopt same-sex marriage bans in a given year than non-direct democracy states. Furthermore, the event history analysis of the adoption of same-sex marriage bans revealed that the variation in states’ direct democracy institutional arrangements also had an impact on policy adoption. The index of direct democracy impact suggests that states with more insulated legislatures, more stringent qualification requirements and states that use initiative less often are better able to protect minority rights. This anti-minority effect was also apparent in the examination of Official English laws. Following the previous literature on official English (Preuhs 2005; Schildkraut 2001; Tatalovich 1995), this event history analysis showed that among states with high 138 numbers of foreign born residents, those with direct democracy were more likely to adopt this restrictive language policy. When language policy is a salient issue where the majority feels threatened by the prevalence of foreign languages in their state, direct democracy institutions seems make government more responsive to this perceived threat. Put more succinctly, when the majority prefers a restrictive language policy direct democracy institutions increase the probability of state governments to adopting an official English law. ’ The results of the analysis affirmative action bans were also consistent with the tyranny of the majority argument. Only states with direct democracy institutions have adopted this broad anti-minority measure. Meanwhile, states without direct democracy institutions, like New Jersey, have considered this policy but have been unable to pass suggesting that the filtering mechanisms of representative democracy were helping to protect minority rights. While not a single non-direct democracy states has adOpted an affirmative action ban to date, thirty—seven percent of direct democracy states have passed this policy. In the third chapter, I examined individual anti-minority policy proposals and their rates of passage. Again, by examining both traditional legislation and citizen legislation and comparing outcomes across institutional arrangement, the impact of direct democracy was clear and consistent. Direct democracy states had significantly higher rates of passage for policy proposals that targeted the rights of homosexuals, speakers of foreign languages, and minorities in general. Furthermore, the probability of passing any individual anti-minority policy proposal was significantly higher in state with direct initiatives. The effects of the arrangements of direct democracy institutions are also 139 evident from this analysis. Policy proposals in states with less insulated legislatures, easier qualification requirements and higher initiative use all had higher probabilities of passage. The analyses from Chapters Two and Three both produced clear evidence showing the rights of minority group were at higher risk in direct democracy states than in states without citizen legislative institutions. These policies, whether they target homosexuals, speakers of foreign languages or broadly apply to all minorities, have a higher probability of passing in direct democracy states and thus the rights of minorities are at heightened jeopardy in these states. However, anti-minority policies are not the only policies that affect the rights of minority groups. States can also pass legislation to protect minority rights. Little research has focused on the effect that direct democracy has on the adoption of these pro- minority policies. The fourth chapter examined four pro-minority policies —- nondiscrimination laws that include sexual orientation, racial profiling bans, hate crimes laws and hate crimes laws that cover sexual orientation —— to ascertain the impact of direct democracy on the adoption of laws that would enhance or protect the rights of minority groups. The results from these analyses less consistently supported the argument that direct democracy institutions have a negative impact on minority rights. While direct democracy states were shown to be less likely to pass non-discrimination policies that apply to sexual orientation, they were more likely to pass racial profiling bans and hate crimes laws. There was no statistically significant impact on the likelihood of adopting a hate crime law that includes bias based on sexual orientation. I40 Given the consistent results produced in the analyses of anti-minority policies that supported the argument that citizen legislation endangers minority rights, the varied effects of direct democracy on the adoption of pro-minority policies is surprising. However, it is important to carefully consider the roots of this critique. Concern for minority rights under direct democracy stems from the fear that government will be overly responsive to majority preferences to restrict minority rights. Embedded in this argument is the assumption that the majority will prefer policies that restrict minority rights. Again, Madison recognizes this caveat when he argues that minority rights will be endangered by pure democracy “if a majority be united in common interest” (Madison [1787] 1999.). Thus, the tyranny of the majority argument is special case of a more general effect of direct democracy — a majoritarian effect where governments increase their responsiveness to the preferences of the majority. - When the majority of the public is united against a minority group, the rights of this minority are placed at higher risk under direct democracy institutions than under representative democratic institutions. But when the majority of the public is not united against a minority group, we should not expect direct democracy to endanger the rights of that minority group. Further, we should expect that in cases where the majority actually prefers to protect the rights of a minority group, it is more likely that pro-minority policies will be enacted. 5.2 Public Preferences & Direct Democracy In order to further explore whether the more general majoritarian effect of direct democracy can help explain the results of the previous analyses, it necessary to examine 141 public preferences on these issues. If the direct democracy has a broad majoritarian effect, rather than a strict anti-minority effect, then public support for a policy should be associated with a positive impact of these institutions on policy adoption. Likewise, lack of public support for a policy should be associated with a negative impact on adoption. Ideally, this could be tested by incorporating yearly, state-level measure of public attitudes towards the policy in question into the event history models as an interaction term with the direct democracy variables. Unfortunately, this type of yearly, state-level data on issues attitudes is not available.5 I An alternative approach relies on proxy measures of the public’s attitudes toward each policy. In essence, this approach is taken in the analysis of the adoption of official English laws in the second chapter. Based on the minority threat theory, the public should be more in favor of a restrictive language policy, like an official English law, as , the number of foreign language speakers in their state increases. The incorporation of the percent of each state’s population that is foreign-bom in the model acts as proxy measure for public attitudes towards language policy. Using this specification, it is clear that states with direct democracy institutions were more responsive to the majority’s preferences for this type of policy. By indirectly accounting for the public issue attitudes, the analysis was able to demonstrate the majoritarian effects of direct democracy institutions, even in a case where public opinion on the issue varied quite a bit from state to state. The difficulty in this approach is finding suitable proxy measures that could be used to approximate public preferences on each of the policies under examination here. 5 ' This data is not available both in terms of comparable survey question across the fifty states and in terms of similar questions asked on a yearly basis. 142 5.2.1 National-Level Issue Attitudes & Direct Democracy Another way to examine the interplay of public preferences and the impact of direct democracy institutions on policy adoption is to compare the results of the previous analyses with national-level public opinion on these policies. Although this approach cannot account for state-level variations in issue attitudes, it does provide a starting point from which to evaluate the varying impacts of direct democracy across different minority rights policies. Figure 5.1 shows the percent of respondents that have attitudes in line with the anti-minority policies analyzed in the second chapter.52 From 2000 to 2005, a clear majority of Americans believed that marriages between homosexuals should not be recognized by law as valid. The graph also shows that large majorities of Americans favor making English the official language of the United States.53 Finally, Figure 5.1 reveals that more than two-thirds of Americans oppose race or gender based preferences - in hiring. In terms of the anti-minority policies examined in Chapter Two, it is evident that there are large majorities of Americans that support these policies. This high level of support also tends to be consistent over time. On there own, these public opinion data support both a strict anti-minority argument and a more general majoritarian argument. As seen in Table 5.1, the majority support for each of the anti-minority policies is associated with a negative impact on the likelihood of a state adopting these policies in a 52 Data were gathered from Gallup Polls and the General Social Survey. Specific citations and question wordings are present in Appendix F. 5" Differences in the wording of the questions may the cause of the fairly large difference between 1994 and 2000. In both surveys respondents were asked if they favored “a law making English the official language of the United States”, but in 1994 the question also included the phrase, . .meaning government business would be conducted in English only”. Nonetheless, both surveys produced large majorities in favor of an Official English law. 143 given year. Since these policies each restrict minority rights, this positive effect on policy adoption has a negative impact on the rights of these minority groups. Thus, there is majoritarian effect where direct democracy institutions increase responsiveness to the majority’s preferences, but there is also an anti-minority effect where direct democracy institutions are associated with the adoption of policies that restrict minority rights. Figure 5.1 National-level Issue Attitudes o 52 ........................................ O Q. a... «a 8_. .' ’fi’w’ “A.” '73 ___’__;..-n=.‘:_", .... .. .4- .. —.+-i-~ *"‘""' ' O. ’_,.—-—" The?” ‘5 o_"" g—r m L to . o > as u- . c og ;: v C 8 <0 0. 34 o , I I I I I I I 1 994 1 996 1998 2000 2002 2004 2006 Year —O— Same-Sex Marriage Ban ...... §...... AA Ban - Race ——L-- Official English .. -- r;—-- AA Ban - Gender *Exact wording of each question presented in Appendix F. 144 Table 5.1 National-level Issue Attitudes & Direct Democracy Effects Direct Democracy Effects on... . Public Policy Minority Policy Support* Adoption Rights (I) :3 Same Sex Marriage Ban 61 a + - E >5 “CE" Official English 63 / 78 + - .S E. E Affirmative Action Ban 83 b / 71C + - <£ Non-Discrimination Law (0 including Sexual 21d .. _ {’3’ Orientation 0? Hate Crime Laws 70 + + >, fig .5 Hate Crime Law covering 3. Sexual Orientation 56 No Effect N0 Effect 9 ‘ ' , D. . Racial Profiling Ban 81 + + *Cell entries are percents; Actual wording) of survey questions is shown in Appendix F. ’ Average percentage from 2000 to 2005; Average percentage from 1994 to 2006; c Average percentage from 1996 to 2006; d Average percentage from 1982 to 2006 Sources: Gallup Poll 2000-2006, General Social Survey 1982-2006, Washington Post/ Kaiser Family Foundation/ Harvard University, September, 2000. Public opinion pertaining to the pro-minority policies that were examined in Chapter Four shows more variation. While attitudes towards homosexuality have become more positive since 1982 according to the General Social Survey, a large majority continues to believe that it is “wrong” to some extent. Between 1982 and 2006, the average percent of Americans that responded that homosexuality is “not wrong at all” is only twenty-one percent, with a low of twelve percent in 1987 and high of thirty-two percent in 2006. Attitudes towards hate crimes laws, meanwhile, are more supportive. In 145 1999, a Gallup poll found that seventy percent of Americans would favor a hate crime law in their state. A survey conducted by the Washington Post, the Kaiser Family Foundation and Harvard University in 2000 found that a smaller majority of Americans also supported hate crimes laws that cover sexual orientation, with fifty-six percent of respondents in favor of laws that impose additional penalties for crimes motivated by prejudice against gays and lesbians. Finally, a robust eighty-one percent of respondents to a 1999 Gallup poll disapproved of racial profiling by police. Along with the variation in public support for these pro-minority policies, a corresponding variation in the effects of direct democracy on policy adoption is also evident. With such negative attitudes towards homosexuality, it is not surprising that states with direct democracy institutions were significantly less likely to adopt a non- discrimination law covering sexual orientation. As with the anti-minority policies, this. association between public preferences and the effects of direct democracy fits both the general majoritarian argument and the tyranny of the majority argument. Since public support for expanding the non-discrimination law to include sexual orientation is low, direct democracy institutions are shown to increase responsive to the preferences of the majority (against adoption) by decreasing the likelihood of passing the policy. In decreasing the likelihood of adoption, direct democracy institutions simultaneously have a negative impact on the rights of homosexuals, thus fitting the strict anti-minority argument as well. The next three pro-minority policies, however, do not support both arguments. Despite protecting minority rights, hate crimes laws actually enjoy broad support among Americans. Given the strong support of the majority for this policy, direct democracy 146 institutions seems to increase the likelihood of adopting a state hate crimes law as would be predicted by a general majoritarian argument. This result, though, is contrary to the expectations of a strict “tyranny of the majority” interpretation of the impact of citizen legislative institutions on minority rights. Direct democracy institutions actually show a positive, not negative, impact on the rights of minorities in the case of hate crimes laws. The association between attitudes towards racial profiling and the impact of direct democracy institutions lends itself to similar conclusion. In response to majority opposition to racial profiling, direct democracy institutions increase the probability of adopting a racial profiling ban. In this case there is a majoritarian effect, but not the anti- minority impact predicted by a strict “tyranny of the majority” argument. Interpretation of the association between public support for hate crimes laws that apply to crimes motivated by the victim’s sexual orientation and the impact of direct democracy is less straight-forward. While there is majority support for this policy, there was no corresponding significant effect of direct democracy on policy adoption and subsequently no discernible impact on the rights of homosexuals. This result may be reflective of the ambiguous nature of public opinion regarding this issue. While a large majority of Americans support hate crimes laws, a similarly large majority hold negative attitudes towards homosexuality. Without a clear majority preference on this policy, the majoritarian effect of direct democracy may be mitigated. In all, the evidence presented in Table 5.1 supports the more general majoritarian effect of direct democracy. Although the survey data presented here are national-level data, they nonetheless provide compelling support for a more generalizable impact of direct democracy than for a narrower, anti-minority impact of these institutions. Only 147 four of the seven policies showed an anti-minority effect, compared to six that clearly showed a majoritarian effect. Even in the case of hate crimes laws that include sexual orientation there is evidence that the insignificant effects of direct democracy may be due to ambiguity in public preferences. 5.3 Conclusion (Viewing the results of the previous chapter through a lens of public preference strongly suggests the effect of direct democracy on the adoption of policies that pertain to the rights of minority groups is a majoritarian effect that should be generalizable to all types of policies. That is, direct democracy institutions should increase the likelihood of states adopting policies that (have majority public support and should also decrease the probability of adopting policies that havemajority opposition. Obviously, this broader maj oritarian effect needs to be further scrutinized and tested before any hard conclusions can be drawn, but the initial evidence presented on minority rights policies here does fit this argument well. Although the analyses in this study are more supportive of a majoritarian effect than a strict tyrannical effect, this does not mean that the impact of direct democracy on minority rights is negligible or even positive. Rather, the studies have shown that in most cases the resulting policy has a negative impact on minority rights. All of the anti- minority policies under examination here had an increased probability of adoption in states with direct democracy. This impact is evident from Table 5.2 which shows each state’s adoption record on the policies considered in this study. Only states with direct initiatives adopted all three anti-minority policies. Of the twenty-four direct democracy I48 states, only Massachusetts did not pass one of the three contemporary anti-minority policies considered here. Meanwhile, five of the non-direct democracy states did not adopt any one of these polices, and none of these states adopted all three. These differences were also apparent from the analyses of anti-minority bill passage in Chapter Three. For pro-minority policies, direct democracy had a negative impact in one analysis, a positive impact in two others and an insignificant impact on the last. This variation in effects seems to be contingent on public preferences, or the extent to which the majority is “united in common interest” against a minority. As opposed to the anti-minority policies, the varying impacts of direct democracy create a muddled picture of the overall impact of direct democracy on minority rights in terms of the adoption of pro-minority policies. Table 5.3 shows the ad0ptions of the four pro-minority policies examined in this study. There is no clear difference between the two types of states. Each includes several states that adopt all of the policies, as well as a few states that have adopted none. Comparing the patterns visibly evident in Tables 5.2 and 5.3 further supports the argument that direct democracy has had a detrimental impact on minority rights. Whereas the presence of direct democracy institutions is moderately correlated (0.32 to 0.41) with higher numbers of anti-minority policies adopted, a correlation between states with direct democracy states and the number of pro-minority policy adoptions is virtually nonexistent (0.01 to 0.09). 149 Table 5.2 Anti-Minority Policies Adopted from 1980 to 2007 State SS Ban Official English AABan Total Direct Initiatives CA CO FL MO UT AK AZ AR MS MT ND SD Ml WA ID IL ME NE NV OH OK OR WY MA >>>>>>>>>>>> Pb >>>>> DD >>>>> Direct Democracy States >>>>>>>>>>>>>>>>>>>>>> >>>>> DDPDFDD b AL GA IN IA KY NH NC SC TN VA CT DE HI KS LA MN PA TX VT WV WI MD NJ NM * NY RI >>>>>>>§5> Non-Direct Democracy States >>>>>>>>>>>>>>>>>>>>> OOOOOAAAaaeaaaaawmmmmmmmmmoAAAaaaaaammmmmmmmmwwwww * Passed a symbolic or bilingual official language law prior to 1980 150 Table 5.3 Pro-Minority Policies Adopted from 1980 to 2007 Hate HC Law - . Nondiscrimination State Crime Sexual ”32:” Law — Sexual Total Law Orientation Orientation Direct Initiatives CA CO IL MA NV WA FL MO NE ME OR AK MT OK UT AZ ID Ml MS OH SD AR ND WY >>>>>>>>> PFDDDD DPDDDDDDDDDDDDV Direct Democracy States >>>>>>>>>>>>>>>>>>>>> b DDDP >>>>>> P >>>>> >>> CT MD MN NJ RI >>>>>>>>>>>>>>>>>>> >>>>>>>>> >>>>>>> >>>>> Non-Direct Democracy States 2 Z Z I >>>>>>>>>>>>>>>>>>>>>>>> OO-k—tA-sNNNNwwwwwwwwwww-Alshh-hOO—t-b-t-e-r—‘ANNNNwwwwwA-blsbhh 151 These associations are reflected in comparisons of the average number of policies passed between the two groups of states. For anti—minority policies, direct democracy states average significantly more adoptions than non-direct democracy states; 1.8 compared to 1.2. For pro-minority policies, there is no statistically significant difference between the two types of groups. Non-direct democracy states average 2.5 pro-minority policy adoptions while direct democracy states average 2.3 adoptions. Taken together, this suggests that the net impact of the relationships between direct democracy and the adoption of policies that affects the rights of minority groups is negative. Again, while the effect of direct democracy on minority rights seems to be conditional on whether the majority opposes minority rights, I can not overstate the risk that this contingency presents. Though these tyrannical outcomes and anti-minority effects of direct democracy can be described as a special case of a broader, majoritarian effect, this special case is far from rare. US history is. littered with cases in which there is strong and sometimes violent opposition to the rights a variety of minority groups. In contemporary politics, tyrannical outcomes, especially for policies that target homosexuals, are relatively common. This research suggests that in these cases minority rights are put at further risk under governmental systems with direct democracy institutions than under purely representative democratic systems. The results from this study not only showed a strong majoritarian and often anti- minority impact of direct democracy, but it also highlighted the influence that institutional arrangements of these institutions can have on public policy. This speaks directly to the previous literature. Frey and Goette (1998) concluded that the relatively few instances of anti-minority outcomes in Switzerland was a result of the benign impact 152 of direct democracy on minority rights, but this research suggests that these results were more likely due to the arrangements of direct democracy in that country. The Swiss system requires initiatives to be presented to the legislature before reaching the ballot. The legislative assembly is then given up to four years to deliberate and respond, usually offering a competing measure for public consideration. More often than not, the assembly’s counter proposal is passed rather than the public’s proposal. In short, Swiss direct democracy insulates the legislature to a much higher degree than US. direct democracy and allows for much more deliberation and other types of filtering mechanisms. Rather than using the Swiss institutions as an example of the benign effect of direct democracy on minority groups, they are better suited as examples of how incorporating filtering mechanisms, especially legislative insulation, into direct democracy arrangements can protect minorities. Another implication .of this research pertains to the organization and mobilization of minority groups. Under traditional, representative democratic systems, minority groups seem to have a better chance at opposing restrictive policies solely through organization within there own community and by gaining representation in state legislatures. In direct democracy states, however, this tactic may not be sufficient. Minority groups may need to broaden their base of support beyond their members and aim their messages at the general public in order to build mass support for the protection of their rights. With direct democracy increasing responsiveness to public opinion, minority groups cannot just rely on representation in the legislative process. Rather, they’ll need to show public support in order to better protect and enhance their interests. 153 In practice this is often easier said than done. In the case of affirmative action bans, minority coalition groups such as By Any Means Necessary (BAMN) sought to widen their public support beyond racial and ethnic minorities to include women’s groups. Though gender was emphasized in their public campaigns, opposition to affirmative action bans has been unable to garner the numbers necessary to form a majority. Despite tactics aimed at building a cohesive majority to oppose affirmative action bans, the public still preferred the anti-minority position. Thus, when the rights of minority groups are contingent on majority preferences, as is the case in under direct democracy institutions, these rights are clearly at risk. Though impact of direct democracy on minority rights is better characterized as a general majoritarian one, rather than a strict tyrannical one, the analyses presented in this study show a strong potential for policy outcomes that restrict the rights of minorities. Indeed, every anti-minority policy under examination here had a higher likelihood of being adopted in states that allowed citizen legislation. So while there are certainly some caveats in answering the question of whether direct democracy endangers minority rights, most often this question should be answered in the affirmative. 154 APPENDICES 155 Appendix A: Adoption of Anti-Minority Policies Table A.1 Same-Sex Marriage Bans, 1995-2005 State Year State Year Utah 1995 Mississippi 1997 Alaska 1996 Montana 1997 Arizona 1996 North Dakota 1997 Delaware 1996 Virginia 1997 Georgia 1996 Alabama 1998 Idaho 1996 Hawaii 1998 Illinois 1996 Iowa 1998 Kansas 1996 Kentucky 1998 Michigan 1996 Washington 1998 Missouri 1996 Louisiana 1999 North Carolina 1996 Vermont 1999 Oklahoma 1996 California 2000 Pennsylvania 1996 Colorado 2000 South Carolina 1996 Nebraska 2000 South Dakota 1996 West Virginia 2000 Tennessee 1996 Nevada 2002 Arkansas 1997 Texas 2003 Florida 1997 New Hampshire 2004 Indiana 1997 Ohio 2004 Maine 1997 Oregon 2004 Minnesota 1997 Connecticut 2005 Table A.2 Official English Laws, 1981-2005 State Year State Year Virginia 1981 Colorado 1988 Indiana 1984 Florida 1988 Kentucky 1984 Alabama 1990 Tennessee 1984 Montana 1995 California 1986 New Hampshire 1995 Georgia 1986 South Dakota 1995 Arkansas 1987 Wyoming 1996 Mississippi 1987 Alaska 1998 North Carolina 1987 Missouri 1998 North Dakota 1987 Utah 2000 South Carolina 1987 Iowa 2002 Arizona 1988 156 Appendix B: Coding of Direct Democracy Measures Qualification Difficulty Index: Points are added to the index score for a state if: 1) 2) 3) 4) 5) 6) only statutes or constitutional measures are allowed the length of the qualifying period is limited geographic distribution of signatures is required the proportion of voters’ signatures required for qualification is between 7.0 and 10.0 percent the pr0portion of voters’ signatures required for qualification exceeds 10.0 percent there are substantive limits on the subject matter of initiatives *In this study scores are reversed so that higher scores indicate less difficulty. Legislative Insulation Index." Points are added to the index score for a state if: 1) 2) 3) 4) 5) 6) 7) 8) 9) the state has a single-subj ect rule there are limits on the substance of an initiative there are restriction on fiscal initiatives the legislature can amend or repeal a statutory initiative the legislature can repeal initiative statutes without a waiting period the legislature can repeal initiative statutes without a superrnajority the state allows no constitutional amendment initiatives the state allows direct and indirect initiatives the state allows indirect initiatives only *In this study scores are reversed so that higher scores indicate less insulation. Source: Bowler and Donovan (2004) 157 Appendix C: Principle Components Analysis of Direct Democracy Measures Figure C.1 Scree Plot Q- .— m —i (D Q) 2 CD > N - C a) 2" LIJ o - 4 3 I I I I 1 2 3 4 Component 158 Appendix D: Cox Proportional Hazards Models on AA Bans Table D.1 State Adoptions of Affirmative Action Bans, 1996-2005 Variable Coefficient Standard Error Direct Democracy Impact [+] 3.545“ 1.444 Citizen Ideology H -0.453** 0.177 Republican Government [+] 2.908 2.758 Democratic Government H 4.296# 2.252 Party Competition [+] 40.485* 20.925 Minority Diversity {-1 -40.269* 18.301 Educational Attainment H 0.234 0.179 Population (log) H 7.163" 2.627 Observations 455 Log Likelihood —1 1.010 Notes: One-tailed tests used where appropriate, two-tailed tests otherwise; # significant at 10%; * significant at 5%; ** significant at 1% 159 Appendix E: Keyword Search for Anti-Minority Bills, 1995-2004 Search Terms by Targeted Group: 0 Homosexuals: o Homosexual 0 Gay 0 Same-sex 0 Civil union 0 Maniage 0 Foreign Language Speakers: 0 English language 0 Official English 0 English education 0 General Minorities (Anti-Affirmative Action Policies): 0 Affirmative action 0 Racial preferences 0 Discriminate 0 Civil rights Sources: 0 LexisNexis State Capital 0 National Conference of State Legislature’s Ballot Measure Database 0 http://www.ncsl.org/programs/legismgt/elect/dbintro.htm 0 Various State Legislative Archives 160 Appendix F: Adoption of Pro-Minority Policies Table F.1 Non-Discrimination Policies w/Sexual Orientation, 1982-2007 State Year State Year Wisconsin 1982 Nevada 1999 Massachusetts 1989 Maryland 2001 Connecticut 1991 New York 2002 Hawaii 1991 New Mexico 2003 California 1992 Illinois 2005 New Jersey 1992 Maine 2005 Vermont 1992 Washington 2006 Minnesota 1993 Colorado 2007 Rhode Island 1995 Iowa 2007 New Hampshire 1997 Oregon 2007 Table F.2 Racial Profiling Bans, 1999-2005 State Year State Year Connecticut 1999. Maryland 2002 California 2000 New Jersey 2002 Missouri 2000 Utah 2002 Oklahoma 2000 Washington 2002 Colorado 2001 West Virginia 2002 Florida 2001 Alaska 2003 Kentucky 2001 Arkansas 2003 Massachusetts 2001 Illinois 2003 Minnesota 2001 Montana 2003 Nebraska 2001 Rhode Island 2004 Nevada 2001 Kansas 2005 Texas 2001 Tennessee 2005 161 Table F.3 Hate Crime Laws, 1981-2004 State Year State Year Oregon 1981 Vermont 1989 Washington 1981 Connecticut 1990 Alaska 1982 Iowa 1990 Illinois 1982 New Hampshire 1990 New York 1982 New Jersey 1990 Pennsylvania 1982 North Carolina 1991 Rhode Island 1982 Utah 1992 Idaho 1983 South Dakota 1993 Massachusetts 1983 Texas 1993 California 1984 Alabama 1994 Ohio 1986 Mississippi 1994 Oklahoma 1987 Virginia 1994 West Virginia 1987 Delaware 1995 Wisconsin 1987 Maine 1995 Colorado 1988 Arizona 1997 Maryland 1988 Louisiana 1997 Michigan 1988 Nebraska 1997 Missouri 1988 Kentucky 1998 Florida 1989 Georgia 2000 Minnesota 1989 Hawaii 2001 Montana 1989 Kansas 2002 Nevada 1989 New Mexico . 2003 Tennessee 1989 Table F.4 Hate Crime Laws w/Sexual Orientation, 1984-2005 State Year State Year California 1984 Louisiana 1997 Wisconsin 1987 Nebraska 1997 Minnesota 1989 Kentucky 1998 Nevada 1989 Rhode Island 1998 Oregon 1989 Missouri 1999 Connecticut 1990 New York 2000 Illinois 1990 Tennessee 2000 Iowa 1990 Hawaii 2001 New Hampshire 1990 Texas 2001 New Jersey 1990 Kansas 2002 Vermont 1990 Pennsylvania 2002 Florida 1991 New Mexico 2003 Washington 1993 Colorado 2005 Maine 1995 Maryland 2005 Massachusetts 1996 162 Appendix G: National-Level Public Opinion Surveys, Question Wording Iggarding Anti-Minority Policies Same-Sex Marriage Ban: 0 Do you think marriages between homosexuals should or should not be recognized by the law as valid, with the same rights as traditional marriages? - Gallup Polls, January 13, 2000 — April 29, 2005. Official English: 0 Do you favor a law making English the official language of the United States, meaning government business would be conducted in English only, or do you oppose such a law? - General Social Survey, 1994 0 Do you favor a law making English the official language of the United States, or do you oppose such a law? - General Social Survey, 2000 Affirmative Action Ban: 0 Some people say that because of past discrimination, blacks should be given preference in hiring and promotion. Others say that such preference in hiring and promotion of blacks is wrong because it discriminates against whites. What about your opinion -— are you for or against preferential hiring and promotion of blacks? - General Social Survey 1994-2006 0 Some people say that because of past discrimination, women should be given preference in hiring and promotion. Others say that such preference in hiring and promotion of women is wrong because it discriminates against men. What about your opinion -- are you for or against preferential hiring and promotion of women? - General Social Survey 1996-2006 163 Legarding Pro-Minority Policies Non-discrimination Law including Sexual Orientation: o What about sexual relations between two adults of the same sex--do you think it is always wrong, almost always wrong, wrong only sometimes, or not wrong at all? - General Social Survey, 1982-2006 Racial Profiling Ban: 0 Do you approve or disapprove of the use of "racial profiling" by police?” - Gallup Poll, December 9, 1999 Hate Crimes Law: 0 Some states have special laws that provide harsher penalties for crimes motivated by hate of certain groups than the same penalties for the same crimes if they are not motivated by this kind of hate. Would you favor or oppose this type of hate crime law in your state? - Gallup Poll, February 19-21, 1999. Hate Crimes Law covering Sexual Orientation: _ 0’ Do you favor or oppose a federal law that would impose additional penalties on people who commit crimes out of prejudice toward gays and lesbians? - Washington Post/ Kaiser Family F oundation/ Harvard University, September, 2000 164 REFERENCES 1979. Chapter 765: Maniage. In Wisc. Stat. 765. 001. ACLU. 2007. Racial Profiling: Definition 2005 [cited June 2007]. Available from http://wwwacl11.0rg/racial itistice/racialprofiling/2 l 741 res2005 1 l23.html. Altschiller, Donald. 2005. Hate Crimes: A Reference Handbook. 2nd ed, Contemporary World Issues. Santa Barbara, Calif: ABC-CLIO. Anderson, Terry H. 2004. The Pursuit of Fairness: A History of Affirmative Action. New York: Oxford University Press. Arceneaux, Kevin. 2002. "Direct Democracy and the Link Between Public Opinion and State Abortion Policy." State Politics & Policy Quarterly 2 (4):372-3 87. Baron, Dennis E. 1990. The English-Only Question: An Official Language for Americans? New Haven: Yale University Press. Banilleaux, Charles, Thomas Holbrook, and Laura Langer. 2002. "Electoral Competition, Legislative Balance, and American State Welfare Policy". American Journal of Political Science 46 (2):415-427. Bell, Derrick. 1978. "The Referendum: Democracy's Barrier'to Racial Equality". Washington Law Review 54:1-29. Berry, Frances Stokes, and William D. Berry. 1990. "State Lottery Adoptions as Policy Innovations: An Event History Analysis". The American Political Science Review 84 (2):395-415. Berry, William D., Evan J. Ringquist, Richard C. F ording, and Russell L. Hanson. 1998. "Measuring Citizen and Government Ideology in the American States, 1960—93". American Journal of Political Science 42 (1):327-348. Berry, William D., Evan J. Ringquist, Richard C. Fording, and Russell L. Hanson. 2003. 1960-2002 Citizen Ideology Series: ICPSR. Blalock, Hubert M. 1967. "Causal Inferences, Closed Populations, and Measures of Association". The American Political Science Review 61 (1):l30-136. Boehmke, Frederick J. 2002. "The Effect of Direct Democracy on the Size and Diversity of State Interest Group Populations". The Journal of Politics 64 (3):827-844. Boehmke, Frederick J. 2005. "Sources of Variation in the Frequency of Statewide Initiatives: The Role of Interest Group Populations". Political Research Quarterly 58 (4):565-575. 165 Bowler, Shaun, and Todd Donovan. 2004. "Measuring the Effects of Direct Democracy on State Policy: Not all Initiatives are Created Equal." State Politics and Policy Quarterly 4:345-363. Bowler, Shaun, and Todd Donovan. 2008. "The Initiative Process". In Politics in the American States, A Comparative Analysis, edited by V. a. R. L. H. Gray. Washington, DC: CQ Press. Bowler, Shaun, Todd Donovan, and Caroline Tolbert. 1998. Citizens as Legislators: Direct Democracy in the United States, Parliaments and Legislatures Series. Columbus: Ohio State University Press. Box-Steffensmeier, Janet M., and Bradford S. Jones. 2004. Event History Modeling: A Guide for Social Scientists. Cambridge ; New York: Cambridge University Press. Brace, Paul. 1988. "The Political Economy of Collective Action: The Case of the American States". Polity 20 (4):648-664. Button, James W., Barbara Ann Rienzo, and Kenneth D. Wald. 1997. Private Lives, Public Conflicts: Battles over Gay Rights in American Communities. Washington, DC: CQ Press. Cain, Bruce E., and Kenneth P. Miller. 2001. "The Pop'ulisttLegacy: Initiatives and the Undennining of Representative Govemment"; In Dangerous Democracy? The Battle Over Ballot Initiatives in America, edited by H. R. E. a. B. A. L. Larry J. Sabato. Lanham, MD: Rowman and Littlefield Publishers, Inc. Center, Racial Profiling Data Collection Resource. 2007. History of Racial Profiling Analysis 2007 [cited June 2007]. Available from http://wwwracialprofilinganalysis.neu.edu/background/historvphp. Chavez, Lydia. 1998. The Color Bind: Califomia's Battle to End Affirmative Action. Berkeley: University of California Press. Commager, Henry Steele. 1958. Majority Rule and Minority Rights, The James W. Richards Lectures in History. Gloucester, Mass: Peter Smith. Cox, Gary W., and Mathew D. McCubbins. 1993. Legislative Leviathan: Party Government in the House, California series on social choice and political economy ; 23. Berkeley: University of California Press. Cox, Gary W., and Mathew D. McCubbins. 2007. Legislative Leviathan: Party Government in the House. 2nd ed. Cambridge ; New York: Cambridge University Press. Crain, Robert L. 1966. "Fluoridation: The Diffusion of an Innovation among Cities". Social Forces 44 (4):467-476. 166 Crawford, James. 1992. Language Loyalties: A Source Book on the Official English Controversy. Chicago: University of Chicago Press. Cree, Nathan. 1892. Direct Legislation by the People. Chicago: A.C. McClurg. Cronin, Thomas E. 1989. Direct Democracy: The Politics of Initiative, Referendum, and Recall. Cambridge, Mass: Harvard University Press. Donovan, Todd, and Shaun Bowler. 1998a. "Direct Democracy and Minority Rights: An Extension". American Journal of Political Science 42 (3):1020-1024. Donovan, Todd, and Shaun Bowler. 1998b. "Responsive or Responsible Government". In Citizens as Legislators, edited by T. D. a. C. J. T. Shaun Bowler. Columbus, OH: Ohio State University Press. Eason J r., Oscar. 1998. Retrospective on Campaign for Initiative 200. Seattle Post- Intelligenger, November 12, 1998, A15. Ellis, Richard. 2002. Democratic Delusions: The Initiative Process in America, Studies in Government and Public Policy. Lawrence: University Press of Kansas. Erikson, Robert S., Gerald C. Wright and John P. McIver. 2006. "Public Opinion in the States: A Quarter Century of Change and Stability". In Public Opinion in the . States, edited by J. Cohen. Palo Alto, CA: Stanford University Press. Erikson, Robert 3., Gerald G. Wright, and John P. McIver. 1993. Statehouse Democracy: Public Opinion and Policy in the American States. Cambridge ; New York: Cambridge University Press. ‘ Eule, Julian N. 1990. "Judicial Review of Direct Democracy". Yale Law Journal 99 (7)21503-1590. Fobanjong, John. 2001. Understanding the backlash against affirmative action. Huntington, N.Y.: Nova Science Publishers. Frey, Bruno S., and Lorenz Goette. 1998. "Does the Popular Vote Destroy Civil Rights?" American Journal of Political Science 42 (4): 1343-1348. Gamble, Barbara S. 1997. "Putting Civil Rights to a Popular Vote". American Journal of Political Science 41 (1):245-269. Gerber, Elisabeth R. 1996. "Legislative Response to the Threat of Popular Initiatives". American Journal of Political Science 40 (1):99-128. Gerber, Elisabeth R. 1999. The populist paradox .' interest group influence and the promise of direct legislation. Princeton, NJ: Princeton University Press. 167 Gerber, Elisabeth R., and Simon Hug. 2001. "Legislative Responses to Referendum". In Referendum Democracy: Citizens, Elites, and Deliberation in Referendum Campaigns, edited by M. M. a. A. Parkin. Toronto: Macmillan/St. Martin’s Press. Gillette, Clayton P. 1988. "Plebiscites, Participation, and Collective Action in Local Government Law". Michigan Law Review 86 (5):930-988. Gray, Virginia. 1973. "Innovation in the States: A Diffusion Study". The American Political Science Review 67 (4)11174-1 185. Grodin, Joseph R. 1988. "Developing a Consensus of Restraint: A Judges Perspective on Judicial Retention Elections." Southern California Law Review 61 :1961. Haider-Markel, Donald P. 2001. "Policy Diffusion as 3 Geographic Expansion of the Scope of Political Conflict: Same-Sex Marriage Bans in the 19905." State Politics and Policy Quarterly 1 (1):2-26. Haider-Markel, Donald P. 2000. " Lesbian and Gay Politics in the States: Interest Groups, Electoral Politics, and Public Policy." In The Politics of Gay Rights, edited by a. K. D. W. Craig A. Rimmerman, Clyde Wilcox. Chicago, IL: University of Chicago Press. Haider-Markel, Donald P., and Kenneth J. Meier. 1996. "The Politics of Gay and Lesbian Rights: Expanding the Scope of the Conflict". Journal of Politics 58 (2):332-349. - Haider—Markel, Donald P., Alana Querze, and Kara Lindaman. 2007. "Lose, Win, or Draw? A Reexamination of Direct Democracy and Minority Rights". Political Research Quarterly 60 (2):304-314. Hajnal, Zoltan L., Elisabeth R. Gerber, and Hugh Louch. 2002. "Minorities and Direct Legislation: Evidence fi'om California Ballot Proposition Elections". Journal of Politics 64 (l):154-177. Hamilton, Alexander, James Madison, and John Jay. [1787] 1999. The Federalist Papers. Edited by C. Rossiter. New York: Penguin Books. Haskell, John. 2001. Direct democracy or representative government? : dispelling the populist myth, Transforming American politics. Boulder, Colo.: Westview Press. Herek, Gregory M., and John P. Capitanio. 1995. "Black heterosexuals' attitudes toward lesbians and gay men in the United States". The Journal of Sex Research 32 (2):95. Hero, Rodney E. 1998. Faces of Inequality: Social Diversity in American Politics. New York: Oxford University Press. 168 Hill, Kim Quaile, and Angela Hinton-Anderson. 1995. "Pathways of Representation: A Causal Analysis of Public Opinion-Policy Linkages". American Journal of Political Science 39 (4):924-935. Hillygus, D. Sunshine, and Todd G. Shields. 2005. "Moral Issues and Voter Decision Making in the 2004 Presidential Election". PS: Political Science & Politics 38 (02):201-209. Holbrook, Thomas M., and Emily Van Dunk. 1993. "Electoral competition in the American states". The American Political Science Review 87 (4):955. HRC. 2007. Statewide Hate Crimes Laws Human Rights Campaign, 2007 [cited June 2007]. Available from http://www.hrc.org/docri m c11ts/Iratc_crime_laws_20070709.pd f. J enness, Valerie, and Ryken Grattet. 2001. Making hate a crime : from social movement to law enforcement. New York: Russell Sage Foundation. Keiser, Lael R., and Kenneth J. Meier. 1996. "Policy Design, Bureaucratic Incentives, and Public Management: The Case of Child Support Enforcement". Journal of Public Administration Research and Theory: J-PART 6 (3):337-364. Key, V. O. 1949. Southern Politics in State and Nation, Vintage book ,' V-510. New York,: Vintage Books. King, Gary, Michael Tomz, and Jason Wittenberg. 2000. "Making the Most of Statistical Analyses: Improving Interpretation and Presentation". American Journal of Political Science 44 (2). 347- 361. Kittilson, Miki Caul, and Katherine Tate. 2005. "Political Parties, Minorities, and Elected Office". In The Politics of Democratic Inclusion, edited by R. E. Hero and C. Wolbrecht. Philadelphia, PA: Temple University Press. Kloss, Heinz. 1977. The American bilingual tradition. Rowley, Mass: Newbury House. Kranz, Rachel, and Tim Cusick. 2005. Gay rights. Rev. ed, Library in a book. New York NY: Facts on File. 3 Lewin, Tamar. 2006. Michigan Rejects Affirmative Action, and Backers Sue. New York Times, November 9, 2006. Lewis, Gregory B. 2005. "Same-Sex Marriage and the 2004 Presidential Election". PS: Political Science & Politics 38 (O2):l95-l99. Long, J. Scott. 1997. Regression Models for Categorical and Limited Dependent Variables, Advanced Quantitative Techniques in the Social Sciences ,' 7. Thousand Oaks: Sage Publications. 169 Lublin, David. 2005. "The Strengthening of Party and Decline of Religion in Explaining Congressional Voting Behavior on Gay and Lesbian Issues". PS: Political Science & Politics 38 (02):241-245. Madison, James. [1787] 1999. "No. 10: The Same Subject Continued". In The Federalist Papers, edited by A. Hamilton, J. Madison and J. Jay. New York: Penguin Books. Madison, James. [1787] 1999. "No. 51: The Structure of Government Must Furnish the Proper Checks and Balances Between the Different Departments". In The Federalist Papers, edited by J. Alexander Hamilton and M. a. J. Jay. New York, NY: Penguin Putnam. Magleby, David B. 1984. Direct Legislation: Voting on Ballot Propositions in the United States. Baltimore: Johns Hopkins University Press. Manweller, Matthew. 2005. "The Angriest Crocodile: Information Costs, Direct Democracy Activists, and the Politicization of State Judicial Elections". State and Local Government Review 37 (2):86-102. March, William. 1999. Bush Readies Affirmative-Action Policies Under Mounting Pressure. Tampa Tribune, November 7, 1999, l. Matsusaka, J. G. 1995. "Fiscal Effects of the Voter Initiative - Evidence from the Last 30 Years". Journal of Political Economy 103 (3)2587-623. Matsusaka, John G. 2004. F or the Many or the Few: The Initiative, Public Policy, and? . American Democracy, American Politics and Political Economy. Chicago: University of Chicago Press. Matsusaka, John G., and Nolan M. McCarty. 2001. "Political resource allocation: Benefits and costs of voter initiatives". Journal of Law Economics & Organization 17 (2):413—448. McClosky, Herbert, and Alida Brill. 1983. Dimensions of Tolerance: What Americans Believe About Civil Liberties. New York: Russell Sage Foundation. Meier, Kenneth J. 1984. "Teachers, Students, and Discrimination: The Policy Impact of Black Representation". The Journal of Politics 46 (1):252-263. Meier, Kenneth J. 1993. "Latinos and Representative Bureaucracy Testing the Thompson and Henderson Hypotheses". Journal of Public Administration Research and Theory: J-PART 3 (4):393-414. Meier, Kenneth John. 1975. "Representative Bureaucracy: An Empirical Analysis". The American Political Science Review 69 (2):526-542. Miller, Kenneth P. . 1999. The Role of Courts in the Initiative Process. In American Political Science Association Meeting. Atlanta, GA. 170 Mooney, Christopher Z. . 2001. "State Politics and Policy Quarterly and the Study of State Politics: The Editor’s Introduction". State Politics and Policy Quarterly 1 ( 1): 1 -4. Nicholson-Crotty, Sean. 2006. "Reassessing Madison's Diversity Hypothesis: The Case of Same-Sex Marriage". The Journal of Politics 68 (4):922—930. Norrander, Barbara. 2001. "Measuring State Public Opinion with the Senate National Election Study". State Politics and Policy Quarterly 1 (l):l 11-125. Pampel, Fred C. 2004. Racial Profiling, Library in a Book. New York: Facts on File. Pippen, John, Shaun Bowler, and Todd Donovan. 2002. "Election Reform and Direct Democracy: Campaign Finance Regulations in the American States". American Politics Research 30 (6):559—582. Preuhs, Robert R. 2005. "Descriptive Representation, Legislative Leadership, and Direct Democracy: Latino Influence on English Only Laws in the States, 1984-2002". State Politics & Policy Quarterly 5 (3):203-224. Qvortrup, Mads. 2001. "The Courts vs. the People: an Essay on Judicial Review of Initiatives". In The Battle Over Citizen Law Making edited by M. D. Waters. Durham, NC: Carolina Academic Press. Ranney, Austin. 1976. "Parties in State Politics". In Politics in the American States : A ' Comparative Analysis, edited by H. Jacob and K. N. Vines. Boston: Little, Brown. ‘ ‘ 1 Richmond, Kelly. 1996. Assembly Bill Would End Affirmative Action by State. The Record, November 27, 1996, A7. Romer, Thomas, and Howard Rosenthal. 1979. "The elusive median voter". Journal of Public Economics 12 (2): 143-170. Schattschneider, E. E. 1960. The Semisovereign People; A Realist’s View of Democracy in America. lst ed. New York,: Holt. Schildkraut, Deborah, J. 2001. "Official-English and the states: Influences on declaring English the offical language in the United States". Political Research Quarterly 54 (2):445. Schmid, Carol L. 2001. The Politics of Language: Conflict, Identity and Cultural Pluralism in Comparative Perspective. Ox ford ; New York: Oxford University Press. Schmidt, David D. 1989. Citizen Lawmakers: The Ballot Initiative Revolution. Philadelphia: Temple University Press. 171 Schmidt, Ronald. 2000. Language Policy and Identity Politics in the United States, Mapping Racisms. Philadelphia: Temple University Press. Schrag, Peter. 1998. Paradise Lost : California's Experience, America 's Future. New York: New Press : Distributed by W.W. Norton. Shipan, Charles R., and Craig Volden. 2006. "Bottom-Up Federalism: The Diffusion of Antismoking Policies from US. Cities to States". American Journal of Political Science 50 (4):825-843. Smith, Daniel A., and Caroline J. Tolbert. 2004. Educated by Initiative: The Effects of Direct Democracy on Citizens and Political Organizations in the American States. Ann Arbor: University of Michigan Press. Squire, Peverill. 1992. "Legislative Professionalization and Membership Diversity in State Legislatures". Legislative Studies Quarterly 17 (l):69-79. Squire, Peverill. 2000. "Uncontested Seats in State Legislative Elections". Legislative Studies Quarterly 25 ( 1): 131-146. Squire, Peverill, and Keith E. Hamm. 2005. 101 Chambers: Congress, State Legislatures, and the Future of Legislative Studies. Columbus, Ohio: Ohio State University Press. Statistics, Bureau of Justice. 1994. Hate Crimes Reported in NIBRS, 1990-1992, edited by U. S. D. 0. Justice. Statistics, Bureau of Justice. 2001. Hate Crimes Reported in NIBRS, 1997-1999: US. Department of Justice. Sullivan, J. W. 11893. Direct Legislation by the Citizenship through the Initiative and Referendum. New York: True Nationalist Publishing. Tatalovich, Raymond. 1995. Nativism Reborn .7: The Official English Language Movement and the American States. Lexington, KY: University Press of Kentucky. Tolbert, C. J ., R. S. McNeal, and D. A. Smith. 2003. "Enhancing civic engagement: the effect of direct democracy on political participation and knowledge". State Politics and Policy Quarterly 3:23. Tolbert, Caroline J ., John Grummel, and Daniel Smith. 2001. "The Effects of Ballot Initiatives on Voter Turnout in the United States". American Politics Research 29 (6):625-48. Tolbert, Caroline J ., and Rodney E. Hero. 1996. "Race/Ethnicity and Direct Democracy: An Analysis of California's Illegal Immigration Initiative". The Journal of Politics 58 (3):806-818. 172 Walker, Jack L. 1969. "The Diffusion of Innovations among the American States". The American Political Science Review 63 (3):880-899. Weber, Ronald E., and Paul Brace, eds. 1999. American State and Local Politics: Directions for the 21st Century. New York: Chatharn House. Welch, Susan, and Kay Thompson. 1980. "The Impact of Federal Incentives on State Policy Innovation". American Journal of Political Science 24 (4):715-729. Wilcox, Clyde and Robin Wolpert. 2000. "Gay Rights in the Public Sphere: Public Opinion on Gay and Lesbian Equality". In The Politics of Gay Rights, edited by W. Rimmerman, and Wilcox. Chicago: University of Chicago Press. 173 S lIll/llllfliallljflflifllllllllflllllI