Impact of U.S. Runoff Elections on Racial Minorities and Women:

An Analysis and Comparison with Instant Runoff Voting

August 2000

by Rob Richie and Caleb Kleppner
Center for Voting and Democracy

Overview

Nearly all southern states use two-round runoff elections for their federal and state elections and for many of their local elections. Two-round runoffs are relatively common in other nations (most presidential elections around the world use runoffs) and for local elections in other parts of the country, but no states outside the South (and some bordering states like Oklahoma and Texas) use runoffs for federal and statewide primaries. Given the looming factor of race in voting patterns in the South and the fact that runoff elections by definition help those in the majority, there have been understandable concerns about the impact of runoffs in the South. Several challenges have been brought under the Voting Rights Act, while some states have debated and made changes in recent years.

This memo summarizes findings from a 1991 book, Runoffs Elections in the United States, by Charles Bullock and Loch Johnson. This book is the most comprehensive assessment that we have found of runoffs in the United States, although at the same time, Bullock was a witness for the state of Georgia when it defended its runoff system in a voting rights challenge. We report Bullock and Johnson's findings on who wins runoff elections, how runoffs affect voter turnout and how runoffs affect the election of racial minorities and women. We also summarize Bullock and Johnson's discussion of three legal challenges to runoffs on voting rights grounds.

Given that there does seem to have been an adverse impact on black candidacies and women in some runoff elections, we then explore the possible differences between the use of two-round runoffs and a one-round "instant runoff." Both approaches can increase the threshold of support necessary to win, as opposed to plurality elections that set no threshold (but where a candidate stills needs a majority to be sure of winning), but instant runoff may be better for racial minorities due to factors involving campaign costs, voter turnout and racial polarization. We also contrast instant runoff voting with plurality voting, looking in particular to vote-splitting that can occur with plurality voting if "too many" candidates run and, correspondingly, the issue of suppression of potential good candidacies in plurality systems to avoid such splitting of the vote by "spoilers."

Finally, we present the results from primary runoffs in several black-majority congressional districts in 1992 -- ones that all worked to the advantage of black candidates -- and summarize recent interest in instant runoff voting around the country.

Description of Data Used in Analysis

The bulk of the data in the Bullock/Johnson book consists of the results of 1,222 primary runoffs for state legislative, federal and statewide offices from 1970-1986 in the ten southern states that use primary runoffs and for which election data was available. The states were: Alabama, Arkansas, Florida, Georgia, Louisiana, Mississippi, North Carolina, Oklahoma, South Carolina and Texas. The authors also analyzed data from a 1953 study on runoff primaries from 1920-1948 and studies on black and female candidate success in local runoff elections. Note that all of this data is relatively old; more recent analyses would be helpful, given partisan shifts in the South -- particularly in congressional elections -- in the last decade.

Factors Affecting Frequency of Runoffs in Southern Primaries

Summary: Runoff elections generally are infrequent, but the frequency rises dramatically in open seat elections, in elections with more than two candidates and in statewide elections.

Frequency of runoffs and number of candidates: Between 1970 and 1986, 10% of the partisan primaries that were studied resulted in a runoff, with little variation by state or little change over time. Fully a quarter of all contested primaries resulted in runoffs (176 out of 717 primaries). Note that because many of these contested primaries only had two candidates (in which a runoff election only could be a triggered by a write-in candidacy), the percentage of primaries resulting in runoffs likely was substantially higher than 25% in those primaries contested by three or more candidates.

Frequency of runoffs and open seats: Open seats were much more likely to require runoffs than seats in which incumbents sought reelection. Nearly 41% of open seats resulted in runoffs (107 of 262 races), while only 4% of races with incumbents led to runoffs (69 of 1,632 races).

Frequency of runoffs and prestige of office: Bullock and Johnson found that the competitiveness of nominations was related to the prestige of the office: higher-level offices had runoffs more frequently. Runoffs were required in 34% of governor's nominations (34 of 100 races), 24% of U.S. Senate nominations (24 of 102), and only 7% of U.S. House nominations (118 of 1,1672).

Factors Affecting Who Wins Runoffs: First-Round Results and Incumbency

Summary: The first-round leader had a significant edge in runoffs, but this edge was smaller in elections for more prestigious offices. Even though incumbents still won most runoffs, there was modest anti-incumbent bias in runoffs. This bias may have been more pronounced in local elections, although data was limited.

First-round results and runoff success in primaries: The first-round leader won the nomination 70% of the time in partisan primaries. This tendency varied little among states, but it did vary according to office. The more prestigious the office, the less impact that the first-round results had on the runoff: 54% of first-round leaders won in U.S. Senate primaries, 56% in primaries for governor, 61% for primaries for lieutenant governor and over 70% for primaries for other statewide offices, state legislature and U.S. representative. Note that in primaries for U.S. Senate and governor, the first-round result almost had no bearing on who won -- the second-place finisher won four of every nine primary runoffs in these elections.

Incumbency and runoff success: There was a modest anti-incumbent bias in runoff elections. Incumbents who led after the first-round went on to win the nomination two-thirds of the time (65%), while non-incumbent first-round leaders won in three-quarters of the runoffs (74%). This difference was affected only slightly by the fact that incumbents are more likely to be forced into runoffs in the more prestigious offices. (The authors found 253 primaries in which incumbents were forced into primary runoffs; 19 were for statewide and U.S. House races.)

Local elections: The data set used by Bullock and Johnson was very small for local elections, but their limited information indicated that the anti-incumbent bias may have been higher in local elections. In Dallas, Fort Worth and San Antonio, incumbents won only two of six runoffs (33%).

Civic leaders in Los Angeles have reported a similar pattern to our Center. Their impression is that incumbents who lose typically lose in a runoff after leading the first round. In San Francisco, on the other hand, the first-round leader has won every single runoff for the past 25 years (from 1974 to 1999) -- in 14 runoffs for mayor, district attorney and the Board of Supervisors. These San Francisco victors have included both winning and losing incumbents.

Effect of Runoffs and Plurality Elections on Voting Behavior and Candidacies

Compared to plurality elections, runoff elections seem to give more candidates an incentive to run, even if these candidates may not expect to win. Candidates can win plurality elections with a smaller threshold (although they can't be sure of winning unless they can achieving a majority of the vote), but on the other hand, groupings of voters can fear splitting their vote among too many candidates or "wasting" their vote on a candidate they believe cannot win. Given that the runoff mechanism lessens concerns about "spoilers" (unless spoiling is defined as forcing a runoff), an increase in candidacies is not surprising. Runoffs also give more voters an incentive to vote for their favorite candidate in the first round rather than casting a strategic vote for a frontrunner. As a result, Bullock and Johnson conclude that more serious candidates run in jurisdictions that use runoffs than in places that use plurality elections. They write:

"In crowded fields under majority-vote rule, some voters will support their most-preferred candidate in the initial balloting, even if they foresee little likelihood of this individual winning. The vote does this in the expectation of a second primary, in which it will be possible to support the successful candidate. Without the possibility of a second primary, a larger share of the electoral is likely to line up behind on of the top two candidates rather than waste a vote on a contender who has little prospect of prevailing in the primary. In support of this proposition, Bradley Canon has shown that plurality states tend to have fewer serious candidates - defined in terms of the share of the vote received than do states in which a runoff is possible." (p. 105, Runoffs Elections in the United States)

Voter Turnout in Runoffs

Voter turnout generally declines between the first-round of a primary runoff and the second round. Voter turnout declined in 67% of all state and federal runoffs, with the decline being particularly steep for runoffs for low-profile offices. The party, race and gender of candidates seems to have an impact on turnout. The authors conclude that the impact of candidate race and gender is too small to indicate that voters flock to the polls to help elect or defeat black or female candidates, but we will seek to study this question more thoroughly.

Decline in runoff turnout and level of election: The declines in voter turnout between the first-round and the second-round runoff was particularly steep in runoffs for some lower-profile offices, but there also were some surprises among prestigious offices such as congressional races. In races for governor, lieutenant governor, state senate and state house, turnout increased over the first-round primary in about a third (ranging from 31% to 38%) of the runoffs. For elections for other statewide executive offices, for the U.S. Senate and for the U.S. House, turnout increased in fewer races: about 20% in those categories. The authors theorize that turnout goes up less frequently in congressional races due to personal contact, but that does not seem entirely persuasive. (Note: the number of races under consideration is low for the Senate races: there was data from only 24 U.S. senate primary runoffs, compared to data from more than 1,000 state legislative runoffs.)

Turnout and party of candidates: Democratic primaries had higher turnout in runoffs than Republican primaries, with runoffs in Democratic primaries averaging 96% of first-round turnout compared to 75% for Republican primaries. This difference perhaps could be tied to Democratic primaries likely being seen as more decisive for determining who would win the office in the general election; Democrats in this period dominated most southern elections.

Turnout and race of candidates: The race of candidates also appear to affect runoff turnout. Bullock and Johnson write, "Contests involving black candidates tended to stimulate greater participation than did all-white runoffs." (p. 148). When a black candidate is in the runoff, for example, turnout increased in the runoff more frequently than when the runoff is all-white (55% of races with a black candidate, compared to 33% for all-white runoffs).

Turnout tended to rise more often when whites beat blacks in runoffs than when blacks beat whites. When whites defeated blacks in runoffs (as happened in 27 instances), turnout rose in 64% of the races and averaged 103% of the first-round turnout. When blacks defeated whites (as happened in 11 instances), turnout rose in 44% of the races and averaged 90% of the first-round turnout. Cause and effect are unclear and merit further examination. Turnout and gender of candidates: The turnout in male-female runoffs is similar to that in male-male runoffs. When women candidates won, turnout increased in 37% of the races, and turnout averaged 90% of the first-round turnout. When men won, turnout increased in fewer races (30%) but on average was higher (97% of first-round turnout).

Runoffs and Race of Candidates

Summary: The authors present mixed evidence on the impact of runoffs on black electoral success. On one hand, they conclude that the presence of a runoff requirement in local elections has no effect on the percentage of city councilors who are black. On the other, whites were much more likely to win runoffs between black and white candidates in primaries for both state and local offices. The success of black candidates in U.S. House primary runoffs in black-majority districts suggests that this may be linked more to the question of whether blacks or whites are a majority in the primary electorate rather than the demands of winning a runoff election. As Julian Bond, NAACP chair and former Georgia legislator, once wrote: "The runoff actually discriminates against the numerical minority, black or white, in elections in which there is racial polarization."

Black candidates in state and federal primary runoffs: In primary elections for state and federal office, white candidates won runoffs between black and white candidates far more frequently than black candidates won in these white-black runoffs. 90% (18 of 20) of white candidates who led in the first round went onto win the nomination, while only 50% (9 of 18) of black candidates who led the first round won the runoff. In these 38 instances of white-black runoffs, black candidates would have won 18 times under plurality rules, but won only 11 times with runoffs. (Note. Of the black primary leaders, two were running for statewide office.)

In contrast, in runoffs between candidates of the same race, first-round leaders in all-white and all-black primary runoffs won at very similar rates: 70% and 71%, respectively.

Although the number of black-white runoff cases is small, these findings support the contention that runoffs are racially discriminatory. Anecdotal evidence suggests that the discriminatory effect diminished from the 1970s to the 1980s, and Bullock and Johnson conclude that there is insufficient data to know why the elections rates of white and black first-round leaders differ. The trends over time need to be analyzed more vigorously, and updated as best as possible with data from the 1990s, when there were far more black-majority state and congressional legislative districts (see the appendix for initial findings that indicate black candidates were assisted by runoff elections in black-majority districts).

Black candidates in local election runoffs: According to Bullock and Johnson, "blacks were as likely to serve on city councils in majority-vote cities as in plurality cities." (p. 117) They also point out that the percentage of black city councilors is greater in the West and Midwest in cities with runoffs than in cities using plurality. (They do not mention the relative voting-age population of different racial groups in these cities, however -- important information to have, given the bias that runoffs give to whichever side is in the majority.)

Evidence from Georgia counties paints a similar picture about black-white runoffs as with state and federal primary elections, however. Black candidates who led in the first round won only 50% of the time against white candidates, but white candidates who led in the first round won 84% of the runoffs against black candidates.

Runoffs and Gender of Candidates

Summary: Nationally, the presence of a runoff requirement generally does not affect the percentage of women in local elections, but the evidence is more mixed for state elections and, based on a relatively few instances, runoffs have had a clear adverse impact against women candidates in runoffs for the most prestigious offices.

Women candidates in primary runoffs: In runoffs in all primary elections, women who lead men after the first round generally win (72% of time). They win less frequently than men who lead women (who win 82% of the time), but more frequently than men who lead men (who win 67% of the time). If a woman has a strong lead after the first round (defined as having won at least 40% of the first-round vote and leading by 5% or more), they win about the same rate as men -- 86%. It is the "weak" first-round leaders among women who do not do as well as "weak" male first-round leaders -- winning 61% of runoffs against men compared to the "weak" male first-round leaders winning 80% of their runoffs against women.

But women have won nominations far less frequently than men in elections to the most prestigious offices and when women trail after the first round. Only 29% of women who led primaries for statewide office went on to win the nomination (two of seven races) and in only three cases out of twenty-one did women overcome second place finishes for statewide, U.S. House and state senate races. In races for the state house, women overcame second place finishes more frequently, 12 of 66 races (18%).

In spite of these apparently clear differences, Bullock and Johnson conclude, "Contrary to conventional wisdom, no evidence exists beyond the episodic and anecdotal that runoffs discriminate against women who win a plurality in the first-round of a primary. Women who finish second in the first-round, however, face stiff odds in the runoff election." p. 69 In the wake of a recent general growth in the number of women in office around the country, analysis of runoff elections and gender of candidates in the 1990s would be very helpful.

Women in cities with runoffs: Women make up approximately 20% of city councilors across the country. Bullock and Johnson report that in the Northeast, South and West, there either is no difference or more women serve in cities with runoffs, but in the Midwest, more women serve on city councils when elected by plurality than runoffs (23% vs. 17%).

Legal Challenges to Runoffs

Summary: Bullock and Johnson describe in detail three legal challenges to runoff elections pursued under the Voting Right Act. The case against runoffs has been persuasive to several judges, but no challenge ultimately has succeeded. Their discussion of the concerns raised by candidates of color in New York City provide some of the best insights into the possible value of instant runoff voting as an alternative runoff mechanism.

Butts v. the City of New York: New York City's runoff primary for mayor: In 1969 a relatively fringe, weak candidate named Mario Proccacino won the Democratic nomination for mayor with 33% of the vote. Proccacino went on to lose to incumbent John Lindsay in the general election. The state legislature and governor then approved a 40% runoff threshold for New York City's citywide primary nominations. The new law resulted in several citywide primaries in subsequent elections.

In 1984, in Butts v. the City of New York, the runoff was challenged in U.S. District Court on voting rights grounds. One leading plaintiff was Herman Badillo, a Latino candidate who lost a runoff election in the Democratic primary for New York City's mayoralty in 1973 (after trailing 34% to 29% in the first round, Badillo lost 61% to 39% in the runoff against a white candidate). In the trial, Badillo testified "There are two types of campaigns. One is a media campaign based on money. The other one is a street campaign based upon your physical presence in the neighborhoods... The media campaign can be a short campaign, and that's what the runoff is geared for... If you are dependent on dinners and cocktail parties [rather than big individual contributors], by definition a short campaign less than three weeks would make it very difficult to raise significant amounts of money."

Bullock and Johnson commented on Badillo's complaints with runoffs: "The brevity of the runoff period also encourages last-minute negative campaigning, claimed critics of the runoff provision. Certainly, anti-Badillo literature with racist overtones was distributed to voters during the second primary in 1973, giving him precious little time to refute charges that he favored quotas in hiring and education and was anti-Jewish. As Badillo recalled, 'You have the kind of panic that you can set off on a one-to-one campaign where there is a limited period of time because you don't have enough time to overcome this, you don't have enough time to conduct an investigation.'"

In 1985 U.S. District Court Judge Charles L. Brieant, Jr. ruled against the runoff. He concluded that, in the presence of racially polarized voting, the runoff law enhances the effectiveness of negative racial campaign tactics during a three-week runoff campaign. Since minority candidates tend to have less access to campaign contributions, they are less able to pay for sufficient media to respond to racial attacks. This gap in ability to raise funds thus dilutes their vote. Judge Brieant wrote: "Historical evidence of past inequality when combined with evidence of lower socioeconomic status and lower rates of voter registration, supports a finding that a plaintiff class member has less opportunity to nominate and elect a candidate of choice under a statute which requires a superplurality in the first-round (40%), or the burdens associated with running two primary campaigns." (p. 88)

The city and state argued on appeal that there had been no racial intent in the adoption of the runoff; it was simply designed to avoid the nomination of fringe candidates such as Proccacino. The purpose of the runoff was to encourage coalition-building and avoidance of the dangers of single-issue voting. The goal was to elect a candidate who was ideologically and politically accountable to a broad spectrum of the members of the party. They also showed that no minority candidate had ever been denied the nomination after finishing first in the first-round of a primary, and that no black and Latino state legislators had voted against the bill instituting the runoff.

In a 2-1 decision, the U.S. Circuit Court reversed the Butts ruling and reinstated the runoff. In Judge Oakes' dissent to this decision, he highlighted that no minority candidate had ever been elected to any office affected by New York's runoff provision. He argued that the 40% threshold gave minorities less opportunity to nominate and elect candidates of choice because of the brevity of the campaign, the estimated half-million dollar expense of the campaign and the difficulty of responding to racial appeals in such a short time.

In 1989, black candidate David Dinkins won the New York City's mayor race without a change to the runoff law. His success seemingly has quieted challenges to the runoff law in New York City, but in 2001, there likely will be a decisive runoff once again in the Democratic primary in the mayor's race.

Note that all of the factors raised by Judge Oakes, along with several complaints with runoffs raised by Badillo, would be addressed by instant runoff voting because it would avoid having to hold a second election.

Whitfield v. Democratic Party of the State of Arkansas -- Runoffs in Arkansas primaries: In a 1988 decision, Whitfield v. Democratic Party of the State of Arkansas, Judge Garnett Thomas Eisele upheld the use of runoffs in primaries in Arkansas. Unlike in New York City, where no minority had won the plurality but lost the runoff, there were four instances in which a black led the first-round of a primary but lost in the second-round runoff. In addition, no black had been elected to a countywide office in Phillips County since the turn of the century. To demonstrate discriminatory intent, plaintiffs pointed to the 1983 enactment of a West Memphis runoff law after a black won the mayor's race with a plurality.

Judge Eisele conceded that some legislators may have supported the new law to block the election of black candidates, but the overall legislative intent was simply to promote majority rule. He rejected as too speculative the notion that a black might lead a first-round of primary if the white vote were split among several candidates but then lose the head-to-head runoff against a white. Judge Eisele reasoned that reverting to a plurality nomination might result in fewer candidates running for office.

In a 2-1 panel decision, the Eighth Circuit Court reversed Whitfield by finding that runoffs diluted the black vote in primaries in Phillips county, but it left intact the runoff primaries in the other 74 counties. Supporting the use of runoff primaries, the dissenting judge argued that the runoff could provide an incentive for reducing racially polarized voting because candidates would have to campaign in the runoff to expand their support and to develop biracial appeals.

The full panel of the Eighth Circuit deadlocked 5-5 without comment on Whitfield, which had the effect of vacating the 2-1 decision and restoring the runoff primary in the one place (Phillips County) in which it had been invalidated.

Jeffers v. Clinton: Runoff in general elections in Arkansas: In 1990, in Jeffers v. Clinton, plaintiffs challenged runoffs in general elections in Arkansas, highlighting four cases involving three cities in which the legislature approved statutes requiring majority victories after black candidates won pluralities with less than a majority. The U.S. District Court of Arkansas upheld the runoff requirement for partisan primaries. It also ruled on the use of runoffs in general municipal elections, finding that the intent to discriminate was so convincing that future efforts to require a majority vote in general elections must be precleared. It did not, however, throw out any existing runoff requirements.

Instant Runoff Voting: How it Works and Evidence of Rising Interest

Summary: Instant runoff voting (IRV) is a one-person, one-vote system that achieves the goal of runoff elections -- clear support by substantial numbers of voters - in a single election. Used for national elections in Australia and Ireland and mayoral elections in London, IRV has recently gained support in several American states and localities.

How IRV works: Instant runoff voting (IRV) is a one-person, one-vote system that achieves the goal of runoff elections -- clear support by substantial numbers of voters - in a single election. Because it increases voters' choices without increasing the number of elections, IRV has particular currency in the United States today -- when simple "either-or" choices are being replaced with a more complex array of choices in most aspects of American's lives. As opposed to traditional American elections where voters cast a vote for just one candidate, IRV allows voters to express a range of preferences by ranking candidates in order of choice: "1" for a first choice, "2" for a second "runoff" choice, "3" for a third choice and so on.

Ballot-counting in instant runoff voting simulates a series of runoff elections, although with just one candidate eliminated in each round. If no candidate receives a majority of first choices, the instant runoff takes place. The last place candidate (or several candidates, if a minimum level of support is set) is eliminated. All votes are then recounted, with ballots that listed the eliminated candidate as a top choice now being counted for whichever candidate is the next choice on those ballots. This process of eliminating the last?place candidate and retabulating the votes continues until one candidate reaches a majority.

Rising interest: Used for national elections in Australia and Ireland and mayoral elections in London, IRV has recently gained support in several American states and localities. Among them:

1. Vancouver (WA) and Santa Clara County (CA): after charter commissions recommended amending local charters to allow IRV, voters approved the change in stand-alone ballot measures in 1998 and 1999;
2. Vermont: an impressive coalition supports IRV for statewide elections, including the Vermont Grange and state branches of Common Cause, PIRG and League of Women Voters, and IRV legislation has bi-partisan support, including from Governor Howard Dean;
3. Austin (TX): a charter commission has recommended a stand-alone amendment to enact IRV for city elections;
4. New Mexico: in 1999, the state senate voted to place a constitutional amendment to enact IRV for all state and federal offices on the ballot, and a more modest measure to allow IRV for these offices and local elections has good prospects for passage in 2001;
5. Alaska: legislation to enact IRV for most state and federal offices was introduced in 1999, and a citizen's initiative with similar provisions will be on the ballot in 2002 after the legislature has a chance to act on the measure in 2001.

Contrasting Instant Runoff Voting and Two-Round Runoffs

Summary: Two-round runoffs and instant runoff voting (IRV) both establish thresholds of support necessary to win, but (IRV) ensures that this threshold will be reached in a single election. Avoidance of a second round of voting saves candidates and jurisdictions money and is more likely to produce strong winners. Producing a winner in one election would seem to have particular benefits for advocates of minority voting rights.

General comparisons: Two-round runoffs and instant runoff voting (IRV) both establish thresholds of support necessary to win, but IRV ensures that this threshold will be reached in a single election. Avoidance of a second round of voting has several benefits. Among them:

1) Campaign finance reform: IRV makes candidates less reliant on wealthy contributors. Candidates in runoffs -- particularly in large jurisdictions -- typically must raise substantial money for a second campaign. They often have little time between rounds of election, giving an advantage to candidates who can raise large amounts of money quickly.

2) Lower election administration costs: IRV saves jurisdictions money on election administration because they do not have to conduct two rounds of election. Holding a statewide runoff election in North Carolina costs taxpayers nearly $4 million. Citywide runoff elections cost taxpayers close to $10 million. Many jurisdictions may need to incur a one-time cost in adapting voting machines to what is required for conducting IRV elections, but after that change, will have no additional costs.

3) Election of stronger winners: IRV is more effective in producing the goal of strong winners. First, IRV ensures the decisive election occurs when turnout is highest. In two-round runoffs, turnout in the second round often is much lower than in the first round. (In one dramatic example, turnout dropped in half in Georgia's U.S. Senate race from the general election in November 1992 to the runoff election in December 1992. In addition, candidates must more clearly appeal to the supporters of eliminated candidates, as these supporters might not return to vote in a runoff -- in party primaries, having this broader appeal is all the more important for general election success.

Voting rights and IRV: There are three areas where IRV would seem to be preferable to two-round runoffs for advocates of minority voting rights, even if the threshold of support to win is the same.

1) Campaign finance: Racial minorities and candidates with strong support among racial minorities generally are less likely to have access to funds for campaigns, particularly under time pressure associated with runoffs. As Herman Badillo testified in the New York City Butts case (see above), "If you are dependent on dinners and cocktail parties [rather than big individual contributors], by definition a short campaign less than three weeks would make it very difficult to raise significant amounts of money."

Bullock and Johnson provide two additional relevant quotes (page 103). Victor McTeer, a Mississippi attorney, testified before a House Judiciary subcommittee in 1985 that: "When poor black candidates are required to finance two campaigns rather than one, get out the vote twice rather than once, aid and provide illiterate black voters with voting assistance on two occasions rather than one, the resulting drain of finances and other resources makes the dual primary a real threat to equal participation in both party and general electoral politics."

Michael Thurmond, a black member of the Georgia Assembly, in the same hearing testified against runoffs on similar grounds: "Elections under the runoff provision are like a gun fight at the OK-Corral. By the time you shoot your way through the first election, you're all out of ammunition -- with another election to go just for the nomination!"

2) Voter turnout: The disparity in turnout between the first and second round of a runoff could also have an adverse impact on the chances of candidates with the backing of racial minorities, at least under certain circumstances. The clearest case would be when a higher-profile, up-ballot race drew minority voters who may not return to the polls for a second round. The U.S. Senate runoff in Georgia provides a good example. In November 1992, Democrat Bill Clinton had a narrow win in Georgia, but Democratic incumbent senator Wyche Fowler fell just short of winning 50%. In the December turnout, Republican Paul Coverdell won, even though his vote total was barely half that it had been in November. In 1996, Harvey Gantt won a contested U.S. Democratic primary in the first round in North Carolina, thanked in large part to a high turnout among black voters. Any minority-backed candidates in down-ballot races who were forced into runoffs were affected by the lower turnout of black voters in the second round -- among candidates who lost in the second round after stronger first-round performances were a black woman running for secretary of state and a Native American running for a U.S. House nomination.

The impact of turnout differential should be studied more exhaustively before reaching definitive conclusions about its impact on minority candidates. For example, lower turnout in the second round could work to the advantage of minority-backed candidates in jurisdictions where racial minorities have a strong grassroots organization.

3) Polarizing campaigns: IRV creates more obvious incentives to build electoral coalitions that can cross racial lines. These incentives have been recognized internationally. For example, Bosnia will use IRV for the first time in its next presidential election, with the goal of encouraging coalition-building in an ethnically divided nation. There is anecdotal evidence that campaigns in southern runoffs between white and black candidates can become more racially polarized than might occur if the elections took place in one round.

Imagine, for example, an election for sheriff with one black candidate and two white candidates in which it wasn't clear that a candidate would reach the required threshold based on first-choice votes alone. If the black candidate were to finish third, the second choices of that candidate's supporters would determine which white candidate won. White candidates would necessarily need to curb racial appeals.

In fact, there was just such a combination of candidates in New York City's Democratic mayoral primary in 1997, but under the dynamics of a traditional runoff election, eventual winner Rush Messinger did little to reach out to supporters of black candidate Al Sharpton. This failure contributed to a very low black turnout in her general election defeat against Republican Rudolph Giuliani. (The black share of the general election electorate in New York City fell from 27% in 1993 to 22% in 1997.)

Bullock and Johnson's comments in relation to the New York City Butts case (see above) are pertinent. They write that: "Certainly, anti-Badillo literature with racist overtones was distributed to voters during the second primary in 1973, giving him precious little time to refute charges that he favored quotas in hiring and education and was anti-Jewish. As Badillo recalled, 'you have the kind of panic that you can set off on a one-to-one campaign where there is a limited period of time.'"

References

Bullock, Charles S., III and Loch K. Johnson. Runoffs Elections in the United States (1991, University of North Carolina Press)

Works cited by Bullock/Johnson

Ewing, Cortez A.M. 1980. Primary Elections in the South: a Study in Uniparty Politics. Norman: University of Oklahoma Press, 1953. Reprint. Westport, CT: Greenwood.

Bullock, Charles S., III and Susan A. MacManus. 1991. "Municipal Electoral Structure and the Election of Councilwomen." Journal of Politics 53:76-89.

Bullock, Charles S., III and A. Brock Smith. 1990. "Black Success in Local Runoff Elections." Journal of Politics 52:1205-22.

Blacks and Electoral Success in Congressional Runoffs and Elections in the 1990s

Every black-majority congressional district now has a black representative, including all those with runoff primaries, except for Pennsylvania's first Congressional District (CD-1), where white candidate Robert Brady won with 74% in his initial election. In 1992 there were several noteworthy cases where black candidates won runoffs in congressional primaries in newly created black-majority districts, sometimes after trailing white candidates in the first round. Black candidates won every black-white runoff in Democratic primaries in black-majority districts, including three nominations in which white candidates won a plurality in the first round.

1992 runoff primaries between black and white candidates: Following is a report on all black-white runoffs in Democratic primaries in black-majority congressional districts in 1992. Black candidates were successful in each of these runoff, and every black candidate improved their performance over the first round.

• In Florida CD-23, Alcee Hastings trailed white opponent Lois Frankel 35% - 28% after the first round, but won 58% - 42% in the runoff.
• In North Carolina CD-1, Eva Clayton trailed white opponent Walter Jones 38% - 31% in the first round, but won 55% - 45% in the runoff.
• In Georgia CD-2, Sanford Bishop trailed white incumbent Charles Hatcher 40% - 21% in the first round, but won 53% - 47% in the runoff.
• In Georgia, CD-11, Cynthia McKinney led her white opponent George DeLoach 31%-25% in the first round and increased her margin in the runoff, winning 56%-44%.
• In Florida CD-3, Corrine Brown comfortably won her runoff against white opponent Andrew Johnson 64%-36% after leading 43%-31% in the first round.

 

Runoffs between black candidates in 1992: There were two runoffs in black-majority districts between two black candidates.

• In Alabama CD-7, Earl Hilliard defeated Hank Sanders 50%-50% (a 700-vote margin) after leading 31%-24% in the first round. Turnout dropped significantly; Hilliard had fewer votes in the second round than he won in the first round.
• In Louisiana CD-4, Cleo Fields won 74%-26% in a general election runoff against Charles Jones (apparently black, although unclear in our source) after leading in the first round 48%-14%.

1993 special election: In 1993, there was a special election for Mike Espy's seat in Mississippi CD-2. A blanket primary was held instead of partisan primaries. In a general election runoff with white Republican Hayes Dent, Bennie Thompson won 55%-45% in the runoff after trailing 34% to 29% after the first round.

Reviewing black electoral success in congressional elections: It is likely that if primary runoffs had been used in all federal and gubernatorial primaries, black representation would not be lower than today. Even though nearly all of these offices outside the south are elected with plurality rules in primaries, no U.S. Senator or governor is black or Latino, and only three U.S. House Members are African-Americans who first were elected in white-majority districts. Those three black Members of Congress are Julia Carson (IN), Barbara Lee (CA) and J.C. Watts (OK). Lee won a comfortable majority in her initial primary win in 1998. Carson won her initial primary in 1996 over a white opponent by a plurality -- winning 49% to 31% -- but with enough support to indicate that she likely would have won in a runoff. Republican Watts narrowly won a runoff in his initial primary in 1994 after leading 49% - 35% in the first round.

Race and Gender in Runoffs: Tables

From: Bullock and Johnson, Runoff elections in the United States (1991)

Table 2.16 Success Rate for Women in Runoffs: Controlling for First-Round Finish in Primaries and Level of Office, 1970-1986 (Number of Cases in Parenthesis)

 

 

Table 4.1 Success Rate of Blacks in Runoffs: Success of First-Round Leaders in Primary Runoffs (controlling for race), 1970-1986

 

 

Notes:

1. Because of the small number of blacks in runoffs for higher office, the authors did not report on the relationship of the office to the black success in runoffs.
2. The pattern is similar in local Georgia runoffs: in black-white runoffs, blacks who led the primary won 50% of the 36 races and white primary leaders won 84% of 37 races.

Quotes about Runoffs and Minority Candidacies and Representation (from Bullock/Johnson)

Julian Bond, NAACP chair and former Georgia legislator, wrote: "The runoff actually discriminates against the numerical minority, black or white, in elections in which there is racial polarization." p. 73

Harold Stanley wrote that black candidates might "split the black vote and allow a white candidate to gain a plurality nomination." p. 73

Referring to majority black districts, the (unnamed) chair of the black caucus in South Carolina state house said, "Things have changed [since reapportionment and black districts] so that runoffs now can work in our favor." p. 73

Herman Badillo, a Latino candidate who lost a runoff election in New York City's mayoral race in 1973 61%-39% after trailing 34%-29% in the first round, in testimony in the 1984 case against runoffs (detailed below): "There are two types of campaigns. One is a media campaign based on money. The other one is a street campaign based upon your physical presence in the neighborhoods... The media campaign can be a short campaign, and that's what the runoff is geared for... If you are dependent on dinners and cocktail parties [rather than big individual contributors], by definition a short campaign less than three weeks would make it very difficult to raise significant amounts of money." pp. 85-86

Bullock and Johnson, commenting on Badillo's complaints with runoffs: "The brevity of the runoff period also encourages last-minute negative campaigning, claimed critics of the runoff provision. Certainly, anti-Badillo literature with racist overtones was distributed to voters during the second primary in 1973, giving him precious little time to refute charges that he favored quotas in hiring and education and was anti-Jewish. As Badillo recalled, 'you have the kind of panic that you can set off on a one-to-one campaign where there is a limited period of time because you don't have enough time to overcome this, you don't have enough time to conduct an investigation.'" p. 86

Victor McTeer, a Mississippi attorney in testimony before a House Judiciary subcommittee in 1985: "When poor black candidates are required to finance two campaigns rather than one, get out the vote twice rather than once, aid and provide illiterate black voters with voting assistance on two occasions rather than one, the resulting drain of finances and other resources makes the dual primary a real threat to equal participation in both party and general electoral politics." p. 103

Michael Thurmond, a black member of the Georgia Assembly, testifying against runoffs on similar grounds: "Elections under the runoff provision are like a gun fight at the OK-Corral. By the time you shoot your way through the first election, you're all out of ammunition -- with another election to go just for the nomination!" p. 103

Bullock and Johnson: "It is true that a greater number of blacks members of Congress have been elected from states without a runoff; however, in both runoff and non-runoff states, election to the U.S. House of Representatives seems to hinge primarily on the existence of a majority black district." p. 104