The Impact of Minor Parties on Electoral Competition: An Examination of US House and State Legislative Races

The Impact of Minor Parties on Electoral Competition: An Examination of US House and State Legislative Races William M. Salka Professor of Political Science Eastern Connecticut State University Willimantic, CT 06226 salkaw@easternct.edu 860.465.4609 Abstract*: Minor parties have long operated on the fringe of the American electoral scene. As a result, little investigation has been done on the impact of these minor party candidates in legislative elections. Outside of academe, however, minor parties have been gaining increased attention as American voters become more disillusioned with the options offered by the two major parties. This preliminary study seeks to examine the impact of minor party candidates on legislative elections in the states and US House of Representatives, in an effort to fill this gap in the literature. Using elections to the legislature in fifteen states and the US House during the 2000s, the study explores the impact that minor party candidates have on the vote shares of major party candidates. Specifically, the paper finds some evidence to support the theory predicting that votes for a minor party candidate are more apt to reflect opposition to the dominant major party in the chamber, rather than voter support for a minor party s platform. *Paper prepared for presentation at the Western Political Science Association Annual Meeting, March 22 24, 2012.This paper is a preliminary exploration of this particular issue and the results and conclusions are subject to change. Any comments or suggestions would be appreciated. 1

Minor political parties have long operated on the fringe of American electoral politics. Despite the fact that they have existed since the 1820s, running for office at all three levels of government and in all regions of the country, these parties are rarely able to field candidates that can compete with the major parties in the American two-party system (Herrnson 2002; Gillespie 1993). As a result, states have traditionally imposed barriers on who can be listed on the ballot, in an effort to limit the number of candidates participating in each election. Proponents of restrictive ballot access laws argue that such limits are necessary to prevent excessively long and confusing ballots. Minor party advocates, on the other hand, have long argued that their lack of success stems partly from discriminatory state laws, suggesting that these laws are written by elected officials from the major parties who seek to maintain their two-party monopoly on elective office (Winger 1997). Despite the persistent presence of minor parties in the United States, little research has been done on the actual impact of minor party candidates on elections. One popularly held notion is that minor party candidates occupy the extremes of the ideological spectrum, thus, stealing votes from the closest major party candidate. The fact that there are typically minor parties at both extremes may explain efforts by Democrats and s in state legislatures to curtail minor party candidates access to the ballot. While this popular notion of vote taking may seem to accurately explain the 2000 presidential election, it may not actually explain electoral support for minor party candidates in less visible elections lower on the ballot. Given methodological difficulties in determining why individual candidates receive the votes that they do, this popular notion that minor party candidates take votes from their ideologically closest major party opponents has gone largely untested. 2

The present study seeks to explore this question using legislative elections in fifteen states and the U.S. House of Representatives from 2002 to 2010. Using the vote shares of individual and candidates to the lower chambers in fifteen state legislatures and the US House, OLS regression will be used to determine the impact of minor party candidates on the vote shares of their major party opponents. The popularly held notion would lead to a prediction that conservative minor party candidates will take votes away from s while liberal minor party candidates will take from Democrats. The counter theory tested in this paper suggests that relatively few American voters actually align themselves with minor parties, and that most votes cast for minor party candidates reflect opposition to the major party in power or alienation from the two-party system, rather than support for the minor party. As a result, the presence of a minor party candidate in a particular race should result in a lower vote share for candidates from the major party in that chamber, regardless of whether the minor party candidate is liberal or conservative. The reader should be cautioned that this paper represents the early stages of this project, with more analysis yet to come. Minor Party Support in a Two-Party System For decades, the dominant understanding of voter behavior came from the so-called Michigan School, which was launched with the publication of The American Voter (Campbell et al. 1960). While this seminal work delved into a number of complex topics, the central argument held that voters develop their party identification while coming of age, largely through the political socialization of their parents and surrounding environment. Once party identification was solidified during an individual s formative years, this attachment was likely to remain stable throughout that individual s life. Given that most voters did not develop a well defined or understood ideology, party identification became the primary means for making political 3

decisions. The Michigan School, then, predicted that most voters would identify with a political party relatively early in life and cast their votes for that party s candidates from then on. While the Michigan School has been challenged in recent decades, largely for its inability to explain split-ticket and retrospective voting, the theory may remain useful in explaining voting behavior in lower visibility races, like those for state legislatures and even the U.S. House. The Michigan School has also been challenged for its inability to explain the rise of voters who were unwilling to identify with a party, preferring instead to classify themselves as independents or unaffiliated. Many scholars today argue that voters are more apt to pay attention to the perceived performance of the incumbent or party in power, particularly as it applies to economic conditions; and reward or punish those incumbents for their past performance (Fiorina 1981; Nadeau and Lewis-Beck 2001). Evidence suggests, however, that many voters are likely to rely more heavily on party preferences, even if those preferences are weak, in less visible races lower on the ballot (Flanigan and Zingale 1994). When voters have little information about candidates, many tend to return to a form of straight ticket voting. While there has been considerable investigation into voter support for major party candidates, the literature on minor party voting is rather thin. This is largely due to the relative weakness of minor parties in America s two-party system. It has long been understood that single member, simple plurality election systems, such as those used for most U.S. legislatures, produce relatively stable two-party systems (Duverger 1954). Given that a candidate must win more votes than all others in each district in order to gain representation in the legislature, ambitious parties try to appeal to the largest bloc of voters found near the center of a more or less normally distributed uni-dimensional ideological continuum of preferences. This strategy provides room for two major parties to compete for voters in the center of the spectrum, with 4

one leaning left of center and the other to the right. Other parties, then, are forced to occupy the ideological fringes on either side of the major parties. Voters likewise tend to act strategically, casting their votes for one of these centrist parties, with the knowledge that only a candidate from one of the two major parties is likely to win. Votes for peripheral parties are often perceived as wasted, or at least ineffective, given the low probability that a minor party candidate will receive more votes than candidates from the major parties, which would be required to win a seat. Thus, minor parties tend to attract limited attention, campaign contributions and electoral support, as voters may fear that a vote for a minor party is at best wasted, or at worst, will help the major party furthest from their preferences win the election. Those studies that have examined electoral support for minor parties have tended to focus on the top of the ticket, higher visibility races like those for president, the U.S. Senate or governor. Perhaps the most comprehensive work in this area was conducted by Rosenstone, Behr and Lazarus (1996). These authors examined support for minor party candidates in presidential elections from 1840 to 1992. They concluded that support for minor party presidential candidates was mostly a function of the quality of those candidates, amounting largely to name recognition, and perceived failures or inadequacies of the major parties. Minor party presidential candidates tended to do best when voters believed the major parties were not meeting their expectations; i.e. the major parties seemed unresponsive to key issues, the economy was performing poorly, or candidates from both major parties seemed inadequate. Other studies have supported the findings of Rosenstone, et al. (1996), giving particular focus to voter dissatisfaction with the major parties and declining levels of trust in government among American voters (Peterson and Wrighton 1998; Howell 1994; Chressanthis and Shaffer 1993; Chressanthis 1990; Elliot, Gryski and Reed 1990; Howell and Fagan 1988). These studies 5

suggest that support for minor party candidates may be less a reflection of support for the actual platforms of those minor parties and more an expression of dissatisfaction with the constraining nature of the American two-party system. A growing number of voters feel that the U.S. system is broken, with the two major parties hopelessly trapped in gridlock, preventing them from addressing important issues of the day. This frustration with the viable major party candidates offered by the two-party system may also be partially responsible for the growing number of unaffiliated voters in recent decades (Flanigan and Zingale 1994). If a voter is dissatisfied with the major parties and their performance in government, that voter can most easily show her discontent by refusing to register with either party, abstaining from the election, or voting for a minor party candidate, even if they know little about the minor party that is on the ballot. The evidence presented in the literature to date suggests, then, that electoral support for minor party candidates might be less an expression of support for that minor party, and more a vote of protest against the U.S. two-party system in general, and the performance of the two major parties specifically. This is particularly likely in lower visibility races where voters have little to no information about the candidates on the ballot. Further, if voters are using their vote for a minor party candidate to express dissatisfaction with the political system in general, it seems plausible to expect those votes for a minor party candidate would most likely come at the expense of the candidate who is a member of the majority party at that time. The scenario might go something like this. Picture a state where a large majority of voters tend to support one major party over another, such as in Connecticut. Given that a relatively large majority of Connecticut voters tend to lean toward the Party, most of those voters are likely to vote for the candidate in the average legislative race where no minor parties are on the ballot. This option is preferable, even if a voter is dissatisfied with the political system, given the voter s 6

socialization in favor of the Party and against the s. When a minor party candidate is on the ballot, however, the voter is given an outlet to express their dissatisfaction; allowing them to vote against the system without having to vote for the Party. It is also likely that a left leaning minor party will have arisen in that state, in an effort to capture the votes of those dissatisfied with the Party. Thus, in states where one party tends to dominate, we might expect that the largest minor party will occupy that side of the spectrum, and votes in favor of that minor party s candidates will come at the expense of the dominant major party. In states where partisan leanings are more evenly divided, we might expect more balanced minor party representation, and support for a minor party candidate may simply come at the expense of the party that is in power at the time of the election. Overall, then, it is expected that votes in favor of minor party candidates are less an indication of support for that party s platform for most voters, and more of an expression of dissatisfaction with government, the two-party system and the major party in power. Data and Methods This study is designed to examine the impact of votes in favor of minor party candidates on the vote share of candidates from the major parties. The central question is whether the presence of a minor party candidate results in a lower vote share for candidates in the state s majority party, or the major party candidate closest to the minor party. While future work will examine the impact of specific minor parties on and vote shares, the present study will only statistically examine the impact of any minor party candidate on votes for the major parties. Data were drawn from all two-party contested lower chamber elections from 2002 to 2010 in fifteen states, and US House elections in all states. The state legislatures included are 7

California, Colorado, Connecticut, Florida, Idaho, Indiana, Kansas, Maine, Michigan, Minnesota, North Carolina, New York, Oregon, Texas and Utah. With Maine, only elections from 2004 to 2010 were examined, as that state did not redistrict until 2004. While these states represent all regions of the U.S. and vary in their degree of legislative professionalism, they were primarily selected because they had the highest proportion of races with minor party candidates across the country. Lower chambers were used instead of upper chambers in an effort to increase the number of districts with minor parties in the analysis. In all, thirty-two OLS models were run, examining the impact of vote shares for Democrats and s in the fifteen states and the U.S. House. The dependent variable in each model, then, is the district level vote share for the or candidates. The states were divided into four categories: strong, strong, or leaning, and leaning based on the average legislative majorities in the years examined. Strong states are those in which the Party held 60 percent or more of the seats in the lower chamber, on average, over the years examined. Strong states had average majorities of 60 percent or greater over those years. Leaning states were those where the majority averaged less than 60 percent of the seats; and in all leaning states except Maine, both parties held a majority prior to at least one election. The independent variable of interest, then, is a dummy variable indicating whether or not a minor party candidate is running in each district. Based on the theory discussed above, it is expected that the presence of a minor party candidate will have a negative and statistically significant impact on the vote share of candidates in those states classified as strong, while not significantly impacting the vote shares of candidates. The 8

reverse is expected in strong states, with the presence of a minor party candidate having a significant negative impact on vote shares. Again, these hypotheses are based on the theory that weakly attached voters, even those that are dissatisfied with the political system, will tend to vote for the candidate from the majority party in each state. This type of voting, in fact, is what contributes to that party holding a majority in the state. Given their political socialization to favor one major party over the other, most voters may be disinclined to switch sides completely and vote for the opposing major party. When given a third option, however, such as a minor party candidate on the ballot, weakly attached voters who feel alienated from the political system may choose to express their dissatisfaction by voting for a minor party candidate. Such a vote might represent a protest against the status quo, without requiring a voter to support the opposing major party candidate. Data on legislative races were gathered from the Secretary of States offices in each state. Data on U.S. House elections were gathered from the Clerk of the House. With the U.S. House races, elections are separated into those years in which the Party held the majority (2002 2006) and those when the Democrats were in power (2008 2010). As with the legislative races, it is expected that the presence of a minor party candidate will have a statistically significant and negative impact on the vote share of candidates in those elections when the s held the majority, with smaller or no impact on vote shares. The opposite is expected in the election years when the Democrats held the majority, with the presence of a minor party candidate having a greater impact on the vote share of candidates in those years. In order to have fully specified models, a number of additional independent variables were included. In the models below, dummy variables were included to indicate the presence of 9

a or incumbent in each district, as this would certainly impact the vote share received by either candidate. A number of district level demographic variables were also included to control for the effects of demographic variance across districts within each state, or across congressional districts. These variables are the percentage of residents in each district with a four year college degree, the percentage below the poverty line and the percent white. Data for these variables were obtained from Lilley et al. (2008) for legislative races and CQ Press (2003) for congressional races. Dummy variables indicating the year in which each election was held were also included in each model. Finally, some effort must be made to control for the specific minor parties that run candidates. It is plausible to expect that relatively liberal states, with routine majorities, are likely to be home to left-leaning minor parties such as the Green Party. Conversely, conservative minor parties, such as the Libertarian or Constitution Parties, might be more common in more conservative states. Table 1 contains data on minor party candidates running in the lower chamber and congressional races during this time period. Columns 3 and 4 show the number of Libertarian Party and Green Party candidates in each chamber, as these are by far the most common minor parties across these races. These data will be used to interpret the results presented below. (Insert Table 1 somewhere here) FINDINGS: State Legislatures Table 2 presents the findings for the states that had strong majorities throughout the 2000s. These states are California, Connecticut and New York. It should be noted that the models in all of the following tables seem to be performing well, given that the control variables that have been well established as influential in the literature, are generally statistically 10

significant and in the expected directions. With regard to the variable measuring the impact of a minor party candidate in races in these three states, the hypothesis is supported for Connecticut and New York. Both states had large majorities in the 2000s, and the presence of a minor party candidate was associated with a statistically significant decrease in the vote share for candidates. The same variable, however, seemed to have no impact on the vote shares of candidates in both states. An examination of the data presented in Table 1 indicates that both states had relatively low proportions of Libertarian candidates, which might have been expected to take votes from s. Neither state, however, had a majority of candidates from the Green or Working Families Parties either (Connecticut: 46%, New York: 33%). These findings suggest that minor party candidates may be providing a sort of protest option for voters who would otherwise have supported candidates. The findings for California, however, are opposite of what was expected. Minor party candidates in California seem to have little impact on the vote share for candidates from the majority. Instead, those minor party candidates seem to be taking votes from candidates in the minority. This finding may make sense, however, when considered with the data in Table 1, which indicates that a large majority (86%) of minor party candidates in California were Libertarians. (Insert Table 2 somewhere here) Table 3 illustrates similar findings in states with strong majorities throughout the 2000s. Four states: Florida, Idaho, Kansas and Utah were classified as strong states, with the results from three of those states supporting the hypothesis. In Idaho, Kansas and Utah, the presence of a minor party candidate was associated with a statistically significant decline in vote share for candidates, while having no significant impact on 11

candidates. This again indicates that minor party candidates may be offering an alternative to voters that would otherwise be inclined to support a member of the majority, but instead chose to support an alternative party candidate. It should be noted, however, that Libertarian and other conservative party candidates represent the vast majority of minor party candidates in each state (Idaho: 84%, Kansas: 84%, Utah: 97%) and there were no Green Party candidates on the ballot in any of these states. As with California among the strong states, Florida is an anomaly among strong states. The presence of a minor party candidate in Florida seems to detract from the candidates, while having little impact on candidates. This finding is particularly interesting given the small number of Green Party candidates on the ballot in Florida and the relatively high proportion of Libertarian Party or Tea Party candidates (67%). (Insert Table 3 somewhere here) Of the remaining eight states, four were classified as leaning and four as leaning. The results for those states leaning are presented in Table 4. Of these four states, the results from Indiana provide the strongest support for the hypothesis. The Party held a majority in the Indiana house prior to four of the five elections. These majorities were small, however, with the Democrats holding an average of 51 percent of the seats across those years. Yet, the presence of a minor party candidate seemed to detract from the vote share of candidates, while having little effect on s. The statistically significant impact on vote shares exists in Indiana despite the fact that there were no Green Party candidates in any of these races, but, 90 percent of the minor party candidates were Libertarians. 12

The results from Maine and Colorado do not seem to support the hypothesis. In Maine, where the Party held a majority prior to all four post-redistricting elections, averaging 56 percent of the seats, candidates from both major parties were negatively affected by the presence of a minor party. However, the effect size was substantially greater for candidates. If voters were using support for minor parties as a way to protest the majority party, then, candidates should have seen a greater negative impact when a minor party candidate was on the ballot. This is especially true given the fact that there were no Libertarian candidates on the ballot in Maine, but 67 percent of minor party candidates were from the Green Party. The findings for Colorado are more comprehensible. Despite the fact that the Democrats held a majority prior to three of the five elections, the presence of a minor party candidate had a negative impact on vote shares, while having no impact on those for candidates. Again, the opposite should have been found if voters were using their support of minor party candidates as a protest against the majority party. However, 91 percent of the minor party candidates in Colorado were Libertarians, and those Libertarian candidates may have been taking votes away from their ideologically closer opponents. Finally, the presence of a minor party candidate in North Carolina did not seem to have an impact on the vote shares of candidates from either major party. This result may be due to the fact that, while minor party candidates are relatively common in races for the lower chamber in North Carolina, those candidates tend to win relatively few votes, having a limited impact on the vote shares of the major party candidates. (Insert Table 4 somewhere here) 13

The results from the four states that were classified as leaning are more supportive of the hypothesis (Table 5). In three of those states, Michigan, Oregon and Texas, the presence of a minor party candidate was associated with a statistically significant decrease in vote share for Party candidates, while having little impact on the vote shares for candidates. In all three of these states, Libertarian candidates represented a majority of all minor party candidates on the ballot (Michigan: 62%, Oregon: 67%, Texas: 97%). In Minnesota, the presence of a minor party candidate had a negative impact on the vote share for both major parties, but the effect size was slightly larger for candidates. As with the other three leaning states, Libertarian candidates constituted a majority of all minor party candidates (73%). (Insert Table 5 somewhere here) FINDINGS: U.S. House The findings for congressional races provide only limited support for the hypothesis. Again, congressional races were divided into two categories, those where the Party held the majority in the U.S. House prior to the election (2002 2006), and those that were held when the Democrats were in the majority (2008 and 2010). It was expected that candidates would lose vote shares to minor party candidates when the chamber was under control; while candidates would suffer the greatest losses in vote shares while they were in power. The results presented in Table 6, however, suggest that this was not the case. Instead, when the Party held the majority in the House, the presence of a minor party candidate was associated with statistically significant declines in vote shares for candidates, while seeming to have little effect on vote shares. This finding suggests that voters were not supporting minor party candidates as way of protesting the 14

majority party in the chamber. Further, Green Party candidates made up only 11 percent of minor party candidates in those elections, thus, it does not appear as though a large proportion of Green Party candidates were taking votes away from Democrats. After the Democrats won a majority in the House as a result of the 2006 election, the results presented in Table 6 indicate that both major parties lost vote shares to minor party candidates in the 2008 and 2010 elections. Thus, the negative and statistically significant relationship between the presence of a minor party candidate and vote shares remained after the Democrats took control of the chamber, but minor party candidates began taking vote shares from candidates as well. The significance of the minor party variable in the vote share model may be attributable to the increase in the number of Libertarian candidates in those races, as Libertarians rose from 37 to 58 percent of all minor party candidates. DISCUSSION Of the fifteen state legislatures examined in this preliminary study, the findings from ten of those states seem to support the hypothesis proposed in this paper, that voters may view minor party candidates as an outlet to express their frustration with the majority party in the chamber. The findings from Connecticut, New York and Indiana provide the most support for this hypothesis. In Connecticut and New York, votes for minor party candidates clearly came at the expense of those in the majority, despite the fact that Green Party candidates were not the majority of minor party candidates in those races. Similarly in Indiana, votes for minor party candidates came at the expense of candidates in the majority, even though the vast majority of minor party candidates were Libertarians. These findings provide evidence that at least in these three states, votes for minor parties seem to take votes away from the majority 15

party, even when attempting to control for the specific minor parties fielding candidates in those races. Seven other states, Idaho, Kansas, Michigan, Minnesota, Oregon, Texas and Utah, all had findings that seemed to support the hypothesis, in that the presence of minor parties tended to be associated with lower vote shares for the majority Party in each state. These findings may, however, also be explained by the fact that the Libertarian or Constitution Parties were the largest minor parties in each state. The presence of conservative minor party candidates could explain the lower vote shares for candidates in these instances. The findings from the five other state legislatures did not seem to support the hypothesis. The fact that minor party candidates seemed to take votes from candidates in California and Colorado, despite majorities, is likely explained by the high proportion of Libertarian candidates on the ballot in those states. The lack of a significant impact of minor parties on either major party in North Carolina is possibly explained by the relatively low vote shares that minor party candidates receive in that state. Florida and Maine seem to be anomalies and deeper exploration is needed to understand the findings there. Similarly, with the U.S. House races, the evidence presented above suggests that voters are not casting votes for minor party candidates as a response to the majority party in that chamber. Instead, state level effects should be explored. Future research will delve deeper into this issue by controlling for the different minor parties on the ballot in these states. For example, statistical analysis will examine the effects of specific minor party candidates on and vote shares by distinguishing between the different minor parties in each state. 16

REFERENCES: Campbell, Angus, Philip Converse, Warren Miller and Donald Stokes. 1960. The American Voter. New York: Wiley. Chressanthis, George A. and Stephen D. Shaffer. 1993. Major-Party Failure and Third-Party Voting in Presidential Elections, 1976-1988. Social Science Quarterly, 74(2): 264-273. Chressanthis, George A. 1990. Third Party Voting and the Rational Voter Model: Empirical Evidence from Recent Presidential Elections. Public Choice, 65(1): 189-93. CQ Press. 2003. Congressional Districts in the 2000s. Washington, DC: CQ Press. Duverger, Maurice. 1954. Political Parties: Their Organization and Activity in the Modern State. New York: Wiley. Elliot, Euel, Gerard S. Gryski and Bruce Reed. 1990. Minor Party Support in State Legislative Elections. State & Local Government Review, 22(3): 123-31. Fiorina, Morris P. 1981. Retrospective Voting in American National Elections. New Haven: Yale University Press. Flanigan, William H. and Nancy H. Zingale. 1994. Political Behavior of the American Electorate, 8 th Ed. Washington, DC: CQ Press. Gillespie, J. David. 1993. Politics at the Periphery: Third Parties in Two-Party America. Columbia, SC: University of South Carolina Press. Herrnson, Paul S. 2002 Two-Party Dominance and Minor-Party Forays in American Politics. In Paul S. Herrnson and John C. Green, eds. Multiparty Politics in America, 2 nd Edition. New York: Rowman & Littlefield Publishers, Inc. Howell, Susan E. 1994. Racism, Cynicism, Economics and David Duke. American Politics Quarterly, 22(2): 190-207. Howell, Susan E. and Deborah Fagan. 1988. Race and Trust in Government. Public Opinion Quarterly, 52(3): 343-50. Lilley, William, III, Laurence J. DeFranco, Mark F. Bernstein, Karl L. Ramsby. 2008. The Almanac of State Legislative Elections: Voting Patterns and Demographics. Washington, DC: CQ Press. Nadeau, Richard and Michael S. Lewis-Beck. 2001. National Economic Voting in U.S. Presidential Elections. Journal of Politics, 63(1): 159-81. 17

Peterson, Geoff and J. Mark Wrighton. 1998. Expressions of Distrust: Third-Party Voting and Cynicism in Government. Political Behavior, 20(1): 17-34. Rosenstone, Steven J., Roy L. Behr and Edward H. Lazarus. 1996. Third Parties in America, 2 nd Ed. Princeton, NJ: Princeton University Press. Winger, Richard. 1997. Institutional Obstacles to a Multiparty System. In Paul S. Herrnson and John C. Green, eds. Multiparty Politics in America. New York: Rowman & Littlefield Publishers, Inc: 159-171. 18

Table 1: Minor Parties in Legislative and Congressional Races, 2002 2010.* Races with Minor Party Candidate Strong California 133 (35.8%) Connecticut 85 (17.7%) New York 112 (21.7%) Strong Florida 61 (24.9%) Idaho 31 (15.7%) Kansas 62 (18.6%) Utah 102# (37.6%) Leaning Colorado 57 (22.1%) Indiana 60 (19.4%) Maine 54 (9.7%) North Carolina 45 (14.8%) Leaning Michigan 172 (32.5%) Minnesota 108# (16.6%) Oregon 51 (22.3%) Texas 133 (40.9%) U.S. House 1073 2002-2006 (54.8%) U.S. House 444 2008-2010 (57.9%) Races with Libertarian Party Candidate 114 (85.7%) 22 (25.9%) 9 (8%) 41# (67.2%) 26# (83.9%) 52 (83.9%) 99# (97.1%) Races with Green Party Candidate 17 (12.8%) 39# (45.9%) 37# (33%) Races with Other Minor Party Candidate 2 (1.5%) 24 (8.2%) 66 (58.9%) 4 16 (6.6%) (26.2%) 0 5 (16.1%) 0 10 (16.1%) 0 3 (2.9%) 52 (91.2%) 0 5 (8.8%) 54 0 6 (90%) (10%) 0 36 18 (66.7%) (33.3%) 45 0 0 (100%) 107 (62.2%) 24 (14%) 41 (23.8%) 79# 28 1 (73.2%) (25.9%) (0.9%) 34 2# 15 (66.7%) (3.9%) (29.4%) 129 4 0 (97%) (3%) 394 117 562 (36.7%) (10.9%) (52.4%) 258 107 79 (58.1%) (24.1%) (17.8%) *Percentages in the Races with Minor Party Candidate indicate the proportion of all races with at least one minor party candidate on the ballot. Percentages in the remaining columns indicate the proportion of candidates from that party or group out of all minor party candidates. #In Connecticut and New York, the Working Families Party is included with Green Party candidates. In Florida, the Tea Party is included with Libertarian candidates. In Idaho and Utah, the Constitution Party is included with Libertarian candidates. In Minnesota, Independence Party candidates are included with Libertarian candidates. In Oregon, the Progressive Party is included with Green candidates. 19

Table 2: Impact of Minor Party Candidates on Major Party in State Lower Chambers with Strong Majorities, 2002 2010. California California Connecticut Connecticut New York New York Minor Present -1.16 (1.03) -3.88*** (1.04) -3.72*** (0.90) 0.26 (0.88) -4.59*** (0.93) -0.03 (0.02) Incumbent 9.85*** (1.1) -9.47*** (1.11) 6.88*** (0.92) -6.77*** (0.9) 11.15*** (0.99) -9.78*** (0.99) Incumbent -10.89*** (1.26) 11.22*** (1.27) -12.9*** (1.1) 13.13*** (1.1) -11.79*** (1.14) 13.0*** (1.13) Education 0.97*** (0.07) -0.97*** (0.07) -0.02 (0.05) 0.02 (0.05) 0.37*** (0.04) -0.37*** (0.04) % Below Poverty 0.97*** (0.11) -0.97*** (0.11) 0.23* (0.12) -0.29** (0.11) 0.60*** (0.07) -0.59*** (0.07) % White -0.36*** (0.04) 0.36*** (0.04) -2.7*** (0.03) 0.27*** (0.03) -0.32*** (0.02) 0.31*** (0.02) 2002-2.49 (1.44) 2.78 (1.45) 1.2 (1.0) -1.1 (0.98) 2.67** (1.1) -1.76 (1.1) 2004-0.93 (1.41) 1.09 (1.43) 5.0*** (1.05) -4.53*** (1.03) 5.13*** (1.1) -4.78*** (1.1) 2006 1.21-1.08 7.1*** -7.41*** 6.64*** -5.95*** (1.42) (1.44) (1.02) (1.0) (1.1) 2008 3.33* -3.77** 6.85*** -7.13*** 6.52*** (3.33) (1.44) (1.02) (1.0) (1.1) N 371 371 480 480 517 517 Adj. R².749.739.758.774.867.862 * p < 0.05; **p < 0.01; *** p < 0.001. (1.1) -5.66*** (1.1) 20

Table 3: Impact of Minor Party Candidates on Major Party in State Lower Chambers with Strong Majorities, 2002 2010. Florida Florida Idaho Idaho Kansas Kansas Utah Utah -2.63** (1.3) -15.1*** (1.8) 4.31** (1.6) -0.17 (0.1) 0.31 (0.14) 0.46*** (0.06) -5.77** (2.0) -1.87 (2.0) -5.79** Minor Present -4.75*** (1.4) 0.86 (1.4) 0.38 (1.5) -4.9*** (1.5) -1.2 (1.3) -3.73** (1.3) -1.9 (1.3) Dem. 11.98*** -12.1*** 9.36*** -9.14*** 13.69*** -13.3*** 14.85*** Incumb. (1.6) (1.6) (1.6) (1.7) (1.4) (1.4) (1.8) Rep. -6.17*** 6.0*** -3.91*** 4.1*** -5.96*** 6.24*** -4.49*** Incumb. (1.2) (1.2) (1.2) (1.2) (1.2) (1.2) (1.6) Education 0.04-0.02 1.08*** -1.06*** 0.001 0.01 0.2* (0.1) (0.1) (0.1) (0.1) (0.06) (0.06) (0.1) % Below 0.31* -0.30* 1.19*** -1.18*** 0.34** -0.33** -0.38** Poverty (0.15) (0.15) (0.15) (0.15) (0.12) (0.1) (0.15) % White -0.12*** 0.12*** 0.08-0.09-0.15*** 0.14*** -0.49 (0.03) (0.03) (0.06) (0.06) (0.04) (0.4) (0.07) 2002 4.64** -4.2* 4.3** -4.5** 7.21*** -7.4*** 6.2** (1.75) (1.77) (1.7) (1.7) (1.6) (1.6) (2.0) 2004 3.0-2.8 4.2** -4.07* 5.66*** -5.55*** 2.8 (1.8) (1.8) (1.6) (1.6) (1.6) (1.6) (2.0) 2006 7.92*** -7.83*** 7.2*** -6.89*** 9.07*** -8.71*** 5.45** (1.7) (1.68) (1.7) (1.7) (1.5) (1.5) (2.0) 2008 4.42** -4.47** 2.94-2.82 6.5*** -6.57*** 6.97*** (1.6) (1.6) (1.6) (1.6) (1.5) (1.4) (1.8) N 245 245 197 197 333 333 271 271 Adj. R².502.472.615.616.540.551.577.583 * p < 0.05; **p < 0.01; *** p < 0.001. (1.9) -6.89*** (1.8) 21

Table 4: Impact of Minor Party Candidates on Major Party in State Lower Chambers Leaning, 2002 2010. Colorado Colorado Indiana Indiana Maine Maine North Carolina North Carolina Minor Present -0.13 (1.2) -3.38** (1.2) -3.54** (1.2) -1.06 (1.2) -4.81*** (1.2) -13.5*** (1.3) -0.34 (1.3) Dem. 6.41*** -5.87*** 9.36*** -9.07*** 8.92*** -8.86*** 6.41*** Incumb (1.1) (1.1) (1.4) (1.4) (0.9) (0.9) (1.1) Rep. -7.48*** 8.06*** -7.4*** 7.58*** -10.1*** 10.36*** -7.24*** Incumb. (1.2) (1.2) (1.5) (1.4) (0.9) (1.0) (1.2) Education 0.86*** -0.85*** -.11 0.13-0.05 0.05 0.35*** (0.07) (0.07) (0.07) (0.07) (0.05) (0.05) (0.07) % Below 0.50*** -0.49*** 0.62*** -0.62*** 0.03-0.05 0.54*** Poverty (0.11) (0.1) (0.1) (0.1) (0.08) (0.1) (0.14) % White -0.50*** 0.51*** -0.32*** 0.31*** -0.31** 0.26* -0.34*** (0.04) (0.04) (0.03) (0.03) (0.1) (0.1) (0.03) 2002-0.07 0.56 2.0-1.26 # # 4.1** (1.5) (1.5) (1.5) (1.5) (1.4) 2004 2.45-1.46 3.37* -2.9* 4.93*** -5.33*** 6.1*** (1.4) (1.5) (1.4) (1.4) (1.0) (1.0) (1.3) 2006 4.48*** -3.62** 7.56*** -7.34*** 7.2*** -7.1*** 7.56*** (1.4) (1.4) (1.4) (1.4) (1.0) (1.0) (1.3) 2008 4.63*** -3.92** 7.46*** -7.05*** 7.1*** -7.86*** 7.6*** (1.3) (1.4) (1.4) (1.4) (1.0) (1.0) (1.2) N 254 254 310 310 558 558 304 304 Adj. R².749.735.714.706.497.536.682.680 #Maine did not redistrict until the 2004 election, so the 2002 election is not included. * p < 0.05; **p < 0.01; *** p < 0.001. -2.4 (1.3) -6.4*** (1.1) 7.17*** (1.2) -0.35*** (0.07) -0.54*** (0.14) 0.34*** (0.03) -4.05** (1.4) -5.99*** (1.3) -7.56*** (1.3) -7.72*** (1.2) 22

Table 5: Impact of Minor Party Candidates on Major Party in State Lower Chambers Leaning Majorities, 2002 2010. Michigan Michigan Minnesota Minnesota Oregon Oregon Texas Texas Minor Present -1.3 (0.9) -1.83* (0.9) -3.5*** (0.02) -4.16*** (0.8) -1.07 (1.4) -2.72* (1.4) -0.16 (1.0) Dem. 10.87*** -10.7*** 10.5*** -8.77*** 9.82*** -9.57*** 13.61*** Incumb. (1.0) (1.0) (0.9) (0.9) (1.5) (1.6) (1.3) Rep. -10.6*** 10.8*** -6.79*** 8.21*** -7.97*** 8.18*** -2.87* Incumb. (1.1) (1.1) (0.9) (0.9) (1.4) (1.4) (1.2) Education 0.06-0.06 0.24*** -0.27*** 0.83*** -0.82*** 0.02 (0.06) (0.06) (0.04) (0.04) (0.08) (0.08) (0.07) % Below 0.51*** -0.52*** 0.60*** -0.66*** 0.58*** -0.59*** 0.15 Poverty (0.1) (0.1) (0.07) (0.07) (0.15) (0.2) (0.1) % White -0.46*** 0.45*** -0.31*** 0.32*** -0.25*** 0.25*** -0.27*** (0.03) (0.03) (0.02) (0.02) (0.05) (0.05) (0.03) 2002 4.14** -3.63** 2.16* -1.88** -0.06-0.05 7.64*** (1.3) (1.3) (1.0) (0.9) (1.7) (1.8) (1.4) 2004 5.52*** -5.2*** 5.58*** -5.26*** 3.93* -4.06* 7.22*** (1.3) (1.3) (0.9) (0.9) (1.8) (1.8) (1.5) 2006 10.36*** -10.1*** 7.79*** -7.29*** 5.59*** -5.74*** 8.88*** (1.3) (1.3) (0.9) (0.9) (1.6) (1.7) (1.4) 2008 8.99*** -8.95*** 5.1*** -5.0*** 4.74** -4.64** 8.26*** (1.3) (1.3) (0.9) (0.9) (1.8) (1.8) (1.4) N 530 530 652 652 229 229 325 325 Adj. R².766.761.736.749.696.699.689.680 * p < 0.05; **p < 0.01; *** p < 0.001. -2.85** (1.0) -13.4*** (1.3) 3.14** (1.3) -0.02 (0.07) -0.15 (0.1) 0.27*** (0.03) -7.48*** (1.4) -7.19*** (1.5) -9.22*** (1.4) -8.12*** (1.4) 23

Table 6: Impact of Minor Party Candidates on Major Party in the U.S. House of Representatives, 2002 2010. Control 2002-2006 Control 2002-2006 Control 2008-2010 Minor Present -2.2*** -0.59-2.12*** (0.53) (0.52) (0.63) Dem Inc 17.5*** -16.41*** 12.75*** (0.86) (0.85) (0.99) Rep Inc -12.03*** 13.22*** -11.76*** (0.85) (0.83) (1.0) Education 0.23*** -0.25*** 0.26*** (0.04) (0.04) (0.04) % Below Poverty 0.47*** -0.49*** 0.23* (0.09) (0.09) (0.11) % White -0.09*** 0.09*** -0.26*** (0.02) (0.02) (0.02) 2002-5.26*** 5.88*** (0.64) (0.63) 2004-3.38*** 3.85*** (0.63) (0.62) 2008 8.49*** Control 2008-2010 -1.81** (0.64) -11.94*** (1.0) 12.5*** (1.0) -0.25*** (0.05) -0.24* (0.1) 0.25*** (0.02) -8.36*** (0.62) (0.61) N 1073 1073 767 767 Adj. R².779.786.763.754 * p < 0.05; **p < 0.01; *** p < 0.001. 24