Cumulative Voting and Split Votes
Contrary to the claims of some critics, Lani Guinier advocates a rigorous adherence to democratic principles, as demonstrated in her article in this report. While strongly supportive of Guinier's principles of political fairness and of her analysis of the benefits of proportional voting systems, I have concerns about her focus on cumulative voting as a method of elections.
The Center for Voting and Democracy promotes consideration of cumulative voting as a possible reform, and I agree with Guinier that cumulative voting is an improvement over single-member districts. Nevertheless, cumulative voting has significant drawbacks, particularly compared with the choice voting form of PR used in Cambridge (MA) and several countries. Applying Guinier's own tests of fairness to the two systems, choice voting wins on all three counts.
As stated by Mary Inman -- a recent student of Professor Guinier's -- in a May 1993 University of Pennsylvania Law Review article "C.P.R. (Change Through Proportional Representation): Resuscitating a Federal Electoral System," the key to preference [choice] voting's superiority derives from the power it gives voters to have their ballot transferred to the next candidate on their list if the first candidate on the ballot cannot be helped by the vote.
"Unlike cumulative voting," writes Inman, "[with choice voting] a party need not worry about nominating too many candidates and thereby splitting its vote among them. As long as party members make their sequential votes for candidates within the party, the votes of a party candidate who is unable to surpass the threshold amount needed for election will be transferred to help elect another candidate within the party."
This transfer of ballots means that a particular group of voters with similar interests will not have their votes dispersed among "too many" candidates appealing to that group of voters. Thus, choice voting encourages vigorous competition within voting blocs as well as among voting blocs, and because of the dynamics of the transfer, promotes positive rather than negative competition among such candidates.
Comparing the Systems
Comparing cumulative voting and choice voting in Guinier's three tests of political fairness reveals the following differences:
Does the system mobilize or discourage participation? The
"transfer" component of choice voting means that two candidates who appeal to
like-minded voters will not split the support of these voters and deprive them of
representation. Instead, if the number of these voters warrants only one seat in an
election (e.g., roughly 20% in a race for five seats), then eventually their ballots will
coalesce behind the candidate who has greater support between them. The reason is that
after the weaker candidate is eliminated, the ballots cast for him or her will go to the
other candidate who appealed to that constituency and likely would be those voters' second
With choice voting, there is every incentive for more candidates to run. By bringing more voters to the polls, additional candidates can only help gain representation for their supporters.
With cumulative voting, however, the vote easily could be split between the two candidates, resulting in neither candidate winning. The practical result of this possible split would be that similar candidates would be discouraged from running, and challengers would be discouraged from running against incumbents.
In a five-member council in a city that is 20% African-American, for example, having only one credible African-American candidate would be encouraged, which both would limit competition for African-American votes and put a ceiling on the number of African-Americans elected to the council.
It would be quite possible, then, that cumulative voting could follow the pattern of race-conscious districts, in which the first campaign would generate much more community organizing and voter turnout than future campaigns. Voters would have more opportunities to define their representation with cumulative voting than they do in single-member district races, but they almost certainly would have even more opportunities with choice voting.
Does the system encourage genuine debate or foster polarization?
Because of the vote-splitting problem discussed above and the resulting incentive to limit
competition, cumulative voting would not be as liberating to political dialogue and the
emergence of new political coalitions as choice voting likely would be. Furthermore,
choice voting provides a greater incentive against negative campaigning -- a major point
in Guinier's case against district elections.
With choice voting, it nearly always is necessary for a candidate to reach out to the supporters of other candidates to gain their support because it is rare for someone to win office without winning transfer votes. The result is that candidates are encouraged to run more positive campaigns -- particularly toward those candidates closest to them in philosophy -- in order to avoid alienating supporters of other candidates and have a chance of gaining a high place on those voters' ballots.
With cumulative voting, candidates have every incentive to get all of people's votes for themselves. This easily could lead to negative campaigning -- particularly among candidates similar in philosophy, in a reversal of choice voting. Japan provides a lesson in such campaigning, as its limited voting system -- in which voters have one vote in districts of three-to-five members -- has led both to bitter intra-party competition and to smaller parties avoiding such competition by nominating only one candidate.
Does the system promise real inclusion or only token representation?
Choice voting provides much clearer opportunities to build cross-racial alliances because
voters are free to continue ranking candidates beyond their first choices; they know that
their vote will go to lower choices only if their higher choices do not need the vote.
Thus, candidates have a greater incentive to reach out to other constituencies in the hope
of attracting transfer votes -- both during campaigns and once in office in preparation
for the next campaign -- and voters can more easily rank candidates of other racial and
A study of ballot transfers in the choice voting elections in New York city in the 1940s demonstrates that while the elections provided much more balanced racial and ethnic representation, most people at the same time voted for tickets that reflected their partisan interests more than race and ethnicity.
Choice voting also provides a greater ability to measure changes in the electorate, leading parties to respond quickly to new constituencies. Cincinnati first elected an African-American to its city council in the 1930s when the city's population was barely 10% African-American because a major party nominated an African-American candidate after seeing how strongly he ran in the previous election as an independent.
With cumulative voting, voters must calculate whether giving votes to a lesser choice will cause their first choice to lose. Thus -- as demonstrated in cities like Alamogordo (N.M.) -- voters have an incentive to give all of their votes to one candidate and shy away from potential coalitions across racial or ideological lines. New constituencies are harder to measure, as voters will worry about "wasting" votes on likely losers.
Possible Distortions in Cumulative Voting Elections
Cumulative voting has another serious problem: because of the danger of vote-splitting, it might not represent the population accurately, undercutting both majority rule and representation of minorities. Unlike choice voting -- where fair results are guaranteed as long as voters: 1) turn out to vote; 2) have a full range of preferences; and 3) rank their favorite candidate first, their second favorite second and so on -- cumulative voting can waste votes and thus waste candidacies.
Guinier in a 1991 Virginia Law Review article recognizes that cumulative voting is a semi-proportional system, but then neglects that fact when incorrectly explaining a theoretical election for 10 seats with 1000 voters. She states that using cumulative voting, "only 90 voters would be unrepresented under the most adverse circumstances." While this would be true in a choice voting election, cumulative voting's problem of splitting votes would make it quite unlikely that there would be as few as 90 voters unrepresented.
As an example, consider the possibility of an extremely racially and ethnically divided town of 1000 people who elect 10 city council members, thus creating a threshold of exclusion (the minimum number of voters who have the ability to elect a candidate on their own) at 91 votes. If the town were 60% white, 10% Asian-American, 10% African-American, 10% Cuban-American and 10% Mexican-American, then the city council might theoretically have six white council members and one member each from the four non-white groups.
Choice voting would guarantee such a result in the unlikely event that voters from the different racial and ethnic groups only voted for candidates of the same group. But with cumulative voting, the same voting preferences could result in an all-white council.
First, a disciplined group of 600 white voters could give their slate of 10 white candidates one vote each and totals of 600 votes. Then, if there happened to be two strong African-American candidates, 50 African-American voters might "plump" their 10 votes on one candidate to give her 500 votes, while the 50 other African-Americans might give the other candidate all their votes, giving him also a total of 500.
Such bad luck -- produced by "too many" candidates -- could happen with the other groups as well, resulting in all of the non-white candidates ending up with less than the 600 votes gained by the 10 white candidates.
Cumulative voting may be appropriate in some situations, but choice voting has clear advantages that demand its careful consideration.
Robert Richie is National Director of The Center for Voting and Democracy.
Table of Contents