A 1995 Exit Survey of Sixteen Communities
Robert Brischetto
The Voting Rights Act and its grant of authority to the Federal courts to uncover official efforts to abridge minorities’ right to vote have been of vital importance in eradicating invidious discrimination from the electoral process and enhancing the legitimacy of our political institutions...As a Nation, we share both the obligation and the aspiration of working toward this end. This end is neither assured nor well served, however, by carving electorates into racial blocs. (Justice Kennedy, majority opinion, Miller v. Johnson, 6/29/95).
With those words a divided, predominantly conservative Supreme Court declared a black-majority congressional district in Georgia illegally drawn to segregate voters on the basis of race and ruled that drawing electoral district lines chiefly on the basis of race can be presumed unconstitutional, absent some compelling state interest.
Two years ago the high court had called into question a "bizarrely" shaped North Carolina congressional district in Shaw v. Reno and warned that "Racial classifications with respect to voting carry particular dangers. Racial gerrymandering, even for remedial purposes, may balkanize us into competing racial factions; it threatens to carry us further from the goal of a political system in which race no longer matters..."(Shaw v. Reno, 1993).
For voting rights advocates, Shaw and Miller were bitter pills to take. For almost three decades they had been drawing districts chiefly on the basis of race in order to level the playing field and allow minorities an opportunity to elect candidates of their own choice. Indeed, the creation of majority-minority districts largely explains why there are 38 African Americans and 17 Latinos in the House of Representatives today.
Growing Interest in PR Alternative
After Shaw and Miller, the "radical" alternatives to districting presented by legal scholar Lani Guinier, whose nomination as Assistant Attorney General for Civil Rights was withdrawn by President Clinton, are looking increasingly relevant. In the wake of these decisions, voting rights advocates are seeking solutions that would provide better representation for minorities without resorting to racial gerrymandering. Some have turned to voting systems that approximate proportional representation in multi-seat elections: cumulative voting, limited voting and preference voting.
The search for alternatives to districting has engendered a long-overdue national debate on more basic questions about how well our democracy works and how we choose our elected officials. The United States is one of only a few modern democracies that have not adopted some form of proportional representation. As Birmingham civil rights attorney Edward Still puts it: "Surely any majoritarian system that can leave 49 percent of the people...with nothing to show for having gone to the polls except a patriotic feeling is not the answer."
More than a hundred local governments throughout the country--cities, counties and school districts--now employ these alternatives to districting, most in response to vote dilution law suits filed in the past decade. This report focuses on cumulative voting, particularly as it has been applied in cities and school boards in Texas.
The Cumulative Voting Alternative
Cumulative voting allows each voter as many votes as seats to be filled in a given election. In that way, it is the same as simple at-large systems. However, under cumulative voting, a voter may distribute votes among candidates in any combination, even distributing all votes to one candidate.
Although relatively rare, this system is not new to the American political scene. From 1870 to 1980, Illinois elected members of their general assembly by cumulative voting. Each legislative district had three representatives and a voter could cast one vote for each of three candidates, one and one-half votes for each of the two candidates or three votes for one candidate. Cumulative voting has also been used for decades to elect members of many corporate boards of directors.
Within this tradition, during the past decade some three dozen local jurisdictions have adopted cumulative voting as a remedy for minority vote dilution. For blacks in Peoria (IL), Hispanics in Alamogordo (NM), Native Americans in Sisseton (SD), blacks in Chilton Co., Centre, Guin and Myrtlewood (AL) and for Latinos in more than two dozen local jurisdictions in Texas, it is a means of obtaining at least some representation. A federal judge last year was the first to order cumulative voting as a remedy in a case against Worcester County, MD (Cane v. Worcester County ).
Cumulative Voting in Texas
Except for the use of limited voting in the small town of Grapeland, cumulative voting is the only modified at-large system in Texas that currently addresses the problem of minority vote dilution. In 1991, the Lockhart Independent School District settled a case of minority vote dilution by adopting four single-member districts and electing the remaining three board members by cumulative voting. In 1992 cumulative voting was adopted to settle lawsuits against the city of Yorktown and Yorktown Independent School District. Since then at least twenty-four small cities and school districts in the Texas panhandle and the Permian Basin have settled lawsuits with cumulative voting, most of them brought by San Antonio attorney Rolando Rios on behalf of the League of United Latin American Citizens (LULAC).
In Atlanta, an East Texas town of 6,000 near Texarkana, no black candidate had ever won a school board election--until May. In election after election, whenever blacks ran for office, the number of votes they received approximated the number of black voters. And that was always fewer than was needed to win.
Determined to address their lack of representation in school district matters, Atlanta’s black community leaders sought help from local attorney Clyde Lee and the national office of the NAACP. A suit was filed in March of 1992.
In its initial response to the lawsuit, the school board offered five single-member districts with one majority black district and two at-large seats; but the plan did not pass muster with the Justice Department. The plaintiffs produced a seven-district plan with two majority-black districts and the school board promptly rejected it. No agreement was reached and it looked as if the case would go to trial on May 1, 1995.
In mid-March, NAACP national legal director Dennis Hayes traveled to Atlanta with an expert on voting systems, hoping to negotiate a settlement. On March 23, six weeks before the next election, Federal District Judge John Hannah signed an agreement between the two parties for a cumulative voting system.
May 1995 Cumulative Voting Elections Study
The May 6, 1995 election provided a rare opportunity to test the effectiveness of cumulative voting in Texas. Twenty-six small cities and school districts in Texas held elections under cumulative voting that day, most for the first time, all in response to litigation. In sixteen of these jurisdictions, minority candidates were competing against Anglos; in ten jurisdictions, minority candidates did not file.
Fifteen of the sixteen jurisdictions studied had Latino candidates on the ballot. The Hispanic Research Center at the University of Texas at San Antonio conducted polling in these cities and school districts. Bilingual teams of pollsters went to these jurisdictions with bilingual questionnaires to gather data from 3,615 voters on how they cast their ballots, how well they understood and how they evaluated the new system of voting.
Atlanta Independent School District (ISD) held the only election in which a black candidate was running. The Atlanta survey of 569 voters was a cooperative effort by experts for the plaintiffs and defendants. The political science department at Texarkana College conducted the field work.
The study analyzes the exit polls of 4,184 voters in the sixteen jurisdictions in which minorities ran for office under cumulative voting. The study addresses several questions:
- Was there racially polarized voting? Were there clear differences between minority and Anglo voters in their preferred candidates? Did minority voters vote as a bloc?
- Did cumulative voting work to elect minority-preferred candidates? If not, why not?
- Did voters understand cumulative voting?
- Did voters accept cumulative voting?
Racially Polarized Voting
Knowing whether voters polarize along racial lines is pivotal in voting rights cases since in the absence of polarization there can be no claim of minority vote dilution.
In Atlanta ISD white and black voters could not have been much more polarized in their choices of candidates. Veloria Nanze came in last among white voters, but first among African American voters. Fewer than 3% of white voters cast even one of their four votes for Nanze; 94% of all votes cast by blacks went to Nanze.
The same general pattern of polarization between Anglos and Latinos was found in the jurisdictions with Latino candidates, but it was less severe. Except in two cases, Latino candidates were the top choices of Latino voters and ranked last or next to last among Anglo voters.
Threshold of Exclusion
In the worse case scenario of totally polarized voting, one can predict the outcome for a racial or ethnic group under cumulative voting by a simple calculation of the "threshold of exclusion." The threshold of exclusion is the percentage of votes that any group of voters must exceed in order to elect a candidate of its choice, regardless of how the rest of the voters cast their ballots. It is calculated as simply one divided by one more than the number of seats to be filled.
With four seats up in the 1995 election to Atlanta school board, the threshold of exclusion was 1/(4+1) or 20%. That meant that, even if Veloria Nanze did not get a single white vote -- and if white voters were spread evenly among just four candidates -- she could win as long as black voters comprised at least 20% of the total voters and concentrated their votes on her.
Blacks in Atlanta comprised 21% of the voting age population in 1990 and 31% of the voters in 1995, which means that voter turnout among blacks in this election apparently was much higher than among whites. In next year's election, when three seats are open on the school board, the threshold of exclusion will be 25%, and it is likely that blacks will elect another representative (as long two or more black candidates do not split the vote).
| Rank of Minority candidate(s) by: | ||||||||
| Total Candi- dates | Minority Voters | Anglo Voters | Positions Elected | Exclusion Threshold | %Minority of Voters | % Over(+) or Under(-) Threshold | Minority Elected? | |
| Minorities Won: | ||||||||
| Atlanta ISD | 5 | 1 | 5 | 4 | 20% | 31% | +11% | Yes |
| Anton | 8 | 1 | 5 | 3 | 25% | 30% | +5% | Yes |
| Morton | 4 | 1,2 | 3,4 | 3 | 25% | 26% | +1% | Yes (1) |
| Morton ISD | 7 | 1,2 | 6,7 | 3 | 25% | 23% | -2% | Yes (1) |
| Roscoe | 8 | 1 | 6 | 5 | 17% | 17% | 0% | Yes |
| Rotan | 8 | 1,2 | 7,8 | 5 | 17% | 32% | +15% | Yes (2) |
| Rotan ISD | 5 | 1 | 5 | 3 | 25% | 25% | 0% | Yes |
| Yorktown | 2 | 1 | 2 | 2 | 33% | 43% | +10% | Yes |
| Olton | 6 | 1 | 6 | 2 | 33% | 22% | -11% | Yes |
| Minorities Lost: | ||||||||
| Andrews ISD | 7 | 1 | 6 | 3 | 25% | 8% | -17% | No |
| Denver City ISD | 5 | 2 | 4 | 2 | 33% | 4% | -29% | No |
| Dumas ISD | 7 | 2 | 6 | 2 | 33% | 2% | -31% | No |
| Earth | 6 | 1,4 | 5,6 | 3 | 25% | 16% | -9% | No |
| Friona | 6 | 1 | 5 | 3 | 25% | 12% | -13% | No |
| Friona ISD | 6 | 1 | 6 | 2 | 33% | 7% | -26% | No |
| Stamford ISD | 5 | 1 | 4 | 3 | 25% | 7% | -18% | No |
“Minority” refers to African Americans in the case of Atlanta ISD, where Latinos are fewer than 0.5% of voters. In the other 15 jurisdictions, “minority” refers to Latinos; blacks are only 1.2% of voters in these jurisdictions.
The Results Under Cumulative Voting
In the case of Atlanta ISD, cumulative voting worked as it should have for black voters seeking to elect one candidate. The African American community not only elected their candidate with almost no white support, but they voted together, placing almost all their votes on Nanze, who came in a close second among five candidates in a race that elected the top four choices.
In the 15 contests involving Latino candidates, on first glance the results seem mixed: 8 wins and 7 losses. A closer examination of the contests involving Latinos reveals that cumulative voting worked almost precisely as expected in polarized communities. In each of the jurisdictions where Latino candidates lost, there were not enough Latino voters to rise above the threshold of exclusion.
- In Andrews school district, Latinos were 32% of the total population but only 16% of the registered voters and 8% of all voters in the election. Since three seats were up, the threshold of exclusion was set at 25%, not low enough for Latino voters to elect their preferred candidate.
- In Denver City ISD, Latinos were 36% of the total population and 15% of the registered voters, but only 4% of those who came out to vote. Two seats were up in the election, setting the threshold of exclusion at 33%, far above what was possible for Latino voters to elect someone.
- In Dumas ISD, Latinos comprise 33% of the total population, but only 2% of the voters in the May 6 election. There were two seats up for election and the threshold of exclusion was set at 32%.
- In the city of Earth, Latinos are 51% of the total population, 31% of registered voters and 16% of actual voters. With three seats up for election May 6, the threshold of exclusion was 25%.
- In the city of Friona, Latinos are 51% of the total population, 37% of the registered voters and 12% of the voters in the May 6 election. With three seats up, the threshold of exclusion was 25%.
- In Friona ISD, Latinos are 45% of the total population, 21% of the registered voters and 7% of those who came out to vote.
- In Stamford ISD, Latinos are 22% of the total population, 14% of the registered voters and 7% of those who voted.
In hindsight, all seven losses could have been avoided either by lowering the threshold of exclusion or raising the level of minority participation in the election. The thresholds could have been reduced by agreement between the parties designing the cumulative voting system, realizing that the more seats up in an election, the lower the threshold. (In school districts with seven seats, if all seats were up at once, the threshold would be 1/(7+1) or 12.5%.)
The Key Role of Community Organizing
The other strategy for winning is somewhat more difficult: raising the level of voter participation through voter registration and education, minority candidate recruitment and get-out-the-vote efforts. Where minority candidates were elected under cumulative voting, voter education and mobilization were apparent. In Atlanta, blacks launched door-to-door voter education and get-out-the-vote drives in black neighborhoods. In Morton, Roscoe and Rotan, where as many as five jurisdictions had Latino winners, the Southwest Voter Registration Education Project provided training on voter mobilization under cumulative voting. In Yorktown, where Concerned Citizens for Voting had begun mobilizing under their first cumulative voting election in 1992, a Latino was running as an incumbent.
In stark contrast, where Latino candidates lost, minority voter participation was low. The average turnout rate among Latinos registered to vote in the seven jurisdictions in which Latino candidates lost was one-half the turnout rate for non-Latino voters.
Finally, for a group or party to win under cumulative voting in a highly polarized political contest, they must vote together as a group. This may require planning to limit the number of minority candidates so as not to split their votes and placing all their votes on their preferred candidate (this problem is avoided by preference voting). Placing all one's votes on a single candidate, or "plumping," is a practice that may enable minority voters to concentrate the strength of their group's vote and improve their chances of electing at least one candidate of their choice. African American voters in Atlanta, Texas, planned their effort very carefully in only a few weeks by agreeing to field only one candidate and by conducting door-to-door voter-education drives in the black community and get-out-the-vote drives and early voting on election day. The poll found that 90% of blacks in Atlanta ISD "plumped" their votes for Veloria Nanze.
Is Cumulative Voting Understood and Accepted?
Ten of the sixteen jurisdictions were holding elections under cumulative voting for the first time; five for only the second time. Beyond the success of minority candidates, we wanted to know how all voters responded to the cumulative voting system. Did they understand the new voting system? How do both minority and white voters perceive cumulative voting?
Since all sixteen jurisdictions polled had been sued for minority vote dilution, it is likely that white voters harbored much resentment at being forced to adopt a settlement over which they had no control. Yet our exit polls found greater understanding and acceptance of cumulative voting than might be expected. More than nine in ten voters of each ethnic group knew they could concentrate all of their votes on a single candidate. Asked to compare cumulative voting with previous election systems, more said that cumulative voting was easier than said it was more difficult.
There were large ethnic differences in evaluations of cumulative voting with regard to difficulty. More than twice as many minority as Anglo voters felt cumulative voting was easier compared to other elections in which they had voted; even so, fewer than two in ten white voters found this election system more difficult than previous voting methods they had used.
Contrary to our expectations, cumulative voting was not rejected by the majority of white voters. We found slightly more agreement than disagreement among whites with the statement that "the voting system used today gives everyone a fair chance to elect officials of their choice." Almost nine in ten blacks and eight in ten Latinos agreed that it was a fair system. However, there were a number of whites -- between one in five and one in four -- who strongly disagreed with that statement.
Conclusions and Recommendations
Lawsuits were settled with cumulative voting schemes in at least 26 jurisdictions in Texas during the past three years, many without regard for local community input or organizing. We studied 16 of these cities and school districts with exit polls on election day, May 6, 1995, all jurisdictions with cumulative voting where minority candidates were on the ballot. In ten places with cumulative, minority candidates were not even fielded in the 1995 elections. Where there was a minority community initiative and the threshold of exclusion was exceeded, the minority candidate won. In those cases where minority candidates lost, there was no organized effort to get out the minority vote and the threshold of exclusion was too high.
The results of our study show that cumulative voting results in a more diverse city council or school board. In those cases where cumulative voting did not result in minority victories, it was not that the method did not work as predicted, but rather because it was not applied correctly.
The study sheds some light on this little-known alternative to districting and offers advice to those considering its adoption.
- Racial bloc voting was found to be extensive in all the cities and school districts studied. In such a racially polarized context, the traditional winner-take-all at-large election system effectively precludes minority groups from electing candidates of their choice.
- Before fashioning an alternative to districting, such as cumulative voting, one must calculate the relative size of the eligible minority voting group. This proportion determines what “threshold of exclusion” is needed in the settlement. For Latino communities, population figures generally will not be an accurate measure of the size of the Latino vote; a better measure is the count of Spanish surnames on the list of registered voters for the jurisdiction.
- The number of seats up at any one time is crucial to determining a minority group's ability to elect since it is used to calculate the "threshold of exclusion." If terms of office are staggered, the chance that a minority group can elect a candidate of its choice will be diminished.
- The ability to elect candidates of their choice issues from the education of minority voters on how the system can work for them and lots of shoe leather. If there is not sufficient local mobilization to get out the minority vote in excess of the calculated threshold of exclusion, the minority's candidate is not likely to win.
- All of the jurisdictions which have adopted cumulative voting in Texas are small. A modified at-large election system was viewed by election administrators as a far more desirable alternative than carving their small communities into even smaller single-member districts. From this limited field experiment, it is not clear whether single-member districts in larger communities would work as well or better.
- Finally, if any numeric minority -- racial, gender, country club crowd, bubbas, the militia -- meets the magic threshold and votes as a block, they can elect a candidate of their choice.
Perhaps this last point is why the system is so controversial.
Robert Brischetto, Ph.D., principal investigator of the study, is a sociologist in Texas. The study was conducted at the Hispanic Research Center at the University of Texas at San Antonio with support from the Seasongood Foundation of Cincinnati, the Center for Voting and Democracy, Attorney Jose Garza of San Antonio, the Poverty and Race Research Action Council and the Mexican American Legal Defense and Education Fund. Co-investigator was Richard Engstrom, Professor of Political Science at the University of New Orleans.