Improving New York City's
Community School Board Elections
Testimony to the Citywide Community School Board Elections
Committee on December 2, 1997
By Robert Richie, Executive Director, The Center for Voting and Democracy
(The New York City Community School Boards have been elected by the choice
voting method of proportional representation since 1970. The elections have been
controversial. Rob Richie of the Center for Voting and Democracy recently
testified before a task force charged with making recommendations about changes
in the elections, including a possible elimination of the choice voting method.
Note: the Center does not take a position on the school boards. Rather, we seek
to make it clear that problems that exist with these elections have nothing to
do with the method of election --a method which in fact produces more diverse
representation than likely could be achieved by any other election method.)
Good evening. My name is Rob
Richie. I am executive director of The Center for Voting and Democracy. Based in
Washington, D.C., the Center is a non-profit, non-partisan organization that
researches and disseminates information on electoral systems that promote voter
participation and fair representation. I have both a professional and personal
interest in your city's Community School Board elections.
On the personal side, my great
uncle was George Hallett, who worked at the Citizens Union for many decades.
George played a leading role in the adoption of proportional representation for
city council elections in New York City in 1936 and a leading role in several
successful defenses of that system before its 1947 repeal. He also was the
primary architect of the proportional representation election law for the
Community School Boards.
On the professional side, the
Center is the nation's leading national resource on proportional representation
systems. In my capacity as executive director since its founding in 1992, I
- worked with the staff of Congresswoman Cynthia McKinney in developing the
Voters' Choice Act (HR 3068), a bill to allow states to use proportional
- addressed the Voting Section of the U. S. Department of Justice, Texas
Commission on Judicial Efficiency, the annual meetings of the National
Association of Counties and the National Conference of State Legislatures and
many foreign visitors through the United States Information Agency;
- testified in special sessions before charter commissions in Nassau
County (NY), Miami Beach (FL), Cincinnati (OH) and Detroit (MI) and advised
charter commissions and elected officials in several other cities;
- published commentary in the New York Times, Washington Post,
Roll Call, Nation, Social Policy, Christian Science Monitor and many
other publications and been a guest on numerous radio and television programs,
including C-SPAN and MSNBC. In the appendix is an article I co-authored for the
most recent Federal Elections Commission Journal on "Alternative Electoral
Systems as Voting Rights Remedies."
The Center takes a special
interest in the school board elections. Although most mature democracies use
proportional systems, relatively few localities and states have experience with
them in the United States. New York City is one of only two cities to use the
choice voting system -- one that we have begun to call choice voting. This fall
I met with your task force's director Alan Gartner and helped link the task
force with scholars who will help study results from the 1996 school board
Going into those 1996 elections,
the Center received grants totaling $30,000 from the Rockefeller Brothers Fund,
National Non-Partisan Voter Registration Campaign and Fund for the City of New
York to promote increased understanding of choice voting and informed
participation in the school board elections. We produced: a 15-minute video that
aired numerous times on New York cable stations; a public service announcement
that aired on commercial television stations; a series of radio public service
announcements in English, Spanish, Cantonese (Chinese), Creole, Korean and
Russian; information for our web page http://www.igc.org/cvd/ (see appendix);
and materials designed to assist journalists in providing improved media
coverage of choice voting.
Although disappointed in the
record-low turnout in 1996, I believe the most important reasons for this low
turnout had nothing to do with the method of election. At the same time, the
record-high representation of racial and ethnic minorities had everything to do
with the choice voting system. I am sure you will hear a great deal about both
of these issues. What I would like to focus on tonight is:
1) how representative of racial
and ethnic diversity the school boards generally are, when measured by the only
standard that should be used for measuring representativeness in elections --
2) issues to consider when
exploring alternative systems, with a special emphasis on limited voting and
3) my recommendations for
Background on Choice Voting
As you know, the school
boards are elected by voters in their school district every three years.
Candidates run at-large for nine seats, and voters rank candidates they like in
order of choice on paper ballots. Choice voting is used in a variety of
elections around the world. Among them: city council and school board elections
in Cambridge, Mass.; parliamentary elections in the Republic of Ireland; local
elections in northern Ireland; senate elections in Australia; nearly all student
council elections in British universities; and, yes, nominations for the Academy
awards. Other systems of proportional representation are used in most mature
democracies; every nation in Europe now uses proportional representation for at
least one of its national elections.
Choice voting receives a great
deal of criticism in New York City, sparking frequent calls for its elimination
by many reform organizations and journalists. I believe that most of this
criticism is misdirected. The following points should be made about choice
ï Choice voting is a simple
voting process despite negative publicity to the contrary. As one
slogan goes, it is as easy as 1, 2, 3. Voters rank the candidates they like in
order of choice -- in New York City on paper ballots, marking a "1" by
their first choice candidate, a "2" by their second choice and so on.
That is all there is to it for voters, and easy enough that 98% of voters in the
1993 elections cast valid ballots.
The vote-counting process
is more complicated -- in part due to its newness, as baseball or football is
complicated to newcomers -- but the principle is simple: a voter's ballot is
counted toward the voter's highest-ranked candidate who can win with that vote.
Most voters help elect their first choice -- meaning that their other choices
don't factor into the results -- but many voters' top-choice candidate is either
too unpopular to win or popular enough to win without their support. These
voters' ballots are counted toward their next-choice candidate. Given that a
voter's lower choice can never cause a higher choice to lose, voters are free to
rank as many candidates as they want, and sometimes a fifth, sixth or lower
choice actually is the first candidate a voter can help elect.
ï Choice voting guarantees
that nearly all voters (close to 90%) will help elect a candidate -- usually
one's first choice in a large field. No other voting system used in the United
States comes close to this guarantee. Nine of ten voters will officially help
elect someone; most of them and most of the remaining 10% also will rank winning
candidates on their list who did not need their vote to win. Over 70% of voters
regularly elect their first choice in the elections, quite high given voters'
wide range of choices.
If voters know the candidates
and vote according to their preferences, they will win fair representation
within a 10% margin of error, no matter their ideology, race, ethnicity or
gender. It is no accident that the school boards have more women (51%) and
fairer racial and ethnic representation than any of New York City's other
elected offices -- a fairness obscured by some observers who contrast board
composition with student population rather than voting-age
This very fairness, however, can
mean that a majority of voters cannot necessarily oust a board member these
voters find offensive or ineffective as long as this board member has
significant core support. However, a majority can prevent such a board member
from being part of the governing majority. The principle of choice voting is
that the majority will earn the right of decision at the same time that
substantial minorities earn their right to representation.
ï Fair results in choice
voting are entirely tied to voter participation and voter education, which
is thus the source of its potential empowerment and potential problems. If a
group of voters making up 10% of the overall electorate vote for a certain
candidate or candidate slate, these voters will elect one of nine seats. If a
group of voters making up 51% of the electorate votes for a certain candidate
slate, these voters will elect five seats. Every new voter could lead to more
seats to be won, which provides a great incentive for higher voter turnout.
Conversely, low voter turnout
among a certain group of potential voters will mean that particular group will
not elect the candidates it potentially could elect. In addition, lack of
information about candidates will prevent voters from voting effectively. Many
current problems with boards are tied in part to low voter participation and to
lack of information about candidates.
Choice Voting in New York
One might fairly ask if
choice voting is such a good system, why so few participants in the Community
School Board elections seem to like it.
* Ignorance of candidates and
function of boards: One problem is that many voters and potential voters do
not know the candidates in the Community School Board elections, which keeps
them away from the polls or keeps them from voting with much coherence. Many
also do not know the function of school boards.
* Election administration
problems: Candidates and those most involved in the school boards often have
to put much energy into their campaigns and election day efforts. They then
generally have to go through weeks of work after the elections in order
to observe the ballot-count. As will be discussed below, this long ballot-count
is unnecessary -- but understandably a sore point with many candidates and
* "Urban myths"
about choice voting: A third problem is due to the many mythologies about
the voting process -- that the system is impossibly complicated, that ranking
more candidates often keeps voters' first choice from winning, that small groups
of voters can manipulate their ballots to win far more seats than they deserve
and so on. These myths -- all of the above are inaccurate -- scare some voters
away from the polls and confuse many of those who do vote. Low voter turnout and
confusion at the polls is a problem because if voters do not participate or vote
effectively, they will not win their fair share of seats.
Choice voting in fact has a
built-in incentive to promote higher voter turnout than
"winner-take-all" systems used in most elections in the United States:
as explained above, results are tied to turnout, which encourages voters to
participate and candidates to mobilize voters. In Cambridge (MA), voter turnout
in local elections with choice voting is higher than turnout in similar
communities using winner-take-all elections. Cincinnati had voter turnout of
over 70% in its last choice voting election for city council in 1955 (when an
African-American gained the highest number of first choice votes and nearly
became mayor, which was the major reason the city's Republican machine was able
to organize a successful repeal in 1957 after three decades of successful
elections). On a national level, political scientists have found that
proportional representation systems likely increase voter to participation by
10% to 15%.
Once voters understand the power
choice voting gives them to elect their favorite candidates, they have every
incentive to vote and get the representation and policies they seek. In fact,
despite New York's problems with it, voter turnout in the 1993 school board
elections -- probably in part as a result of the Rainbow curriculum -- was the
highest it had been since 1970 and higher than winner-take-all elections for the
Los Angeles school board. Washington, D.C. today is holding a special citywide
election for one of our 13 city council seats; turnout may drop below the 5%
turnout registered for a citywide election earlier this year.
The Voting Rights Act and Fair Representation
Some proposals for changing
the school board elections have been made that I frankly consider irresponsible.
A bill was debated seriously in the New York legislature a couple years ago that
called for going to traditional at-large elections, in which voters would cast
equally weighted votes for up to nine candidates. Such a system would never pass
muster with the Department of Justice, which must pre-clear election law changes
in three of the city's five boroughs because of Section 5 of the Voting Rights
Act (the other two boroughs -- as with the entire country -- is covered by
Section 2 of the Voting Rights Act, meaning lawsuits could be filed against
unfair systems by a protected minority).
Given frequent complaints that
the school boards are unrepresentative, some listeners might be surprised that
the Department of Justice would not leap to support any change away from choice
voting. But the blunt truth is that it is very unlikely that any other election
system could have provided the kind of diverse representation of voters provided
by choice voting in elections since it was instituted in 1970.
Enclosed in the appendix is
revealing information from a 1973 memo distributed by Irving Anker,
then-Chancellor of the Board of Education. The memo was prepared by Frederick
Shaw, head of the Bureau of Educational Program Research and Statistics. It
contrasts the racial and ethnic breakdown of the school boards elected in 1973
with the racial and ethnic breakdown of the general population, voting-age
population, electorate and student population. The elected school board members
actually over-represented black and Latino candidates compared to voting-age
population. Blacks and Latinos won 38% of school board seats, in contrast to
being 29% of the voting-age population -- and in contrast to blacks and Latinos
making up only 5% of the city council and 15% of the state legislative seats
from New York. Yet even then the school system had a majority of blacks and
Latinos, and as a result, many school reformers mistakenly believed the results
were not representative.
The pattern has continued
election after election. Representation of blacks, Latinos and eventually Asian
Americans has generally grown steadily, closely matching and often surpassing
voting-age population for those groups. Non-white candidates won 57% of seats in
1996, while non-white voting age population is approximately 50%. Women
candidates from the beginning also have done very well, usually winning at least
40% of seats and winning over 50% in 1989, 1993 and 1996.
The Center did an analysis of
the 1993 and 1996 elections, district-by-district, which is in the appendix. It
demonstrates that not only have the districts been representative of racial and
ethnic minorities citywide, but also within most districts. Please see the
summary charts in the appendix, but note the following findings:
ï Blacks: In 1996,
black candidates were elected in 26 districts and were elected in proportion to
black voting age population in 26 of 32 districts. Citywide, blacks are
represented far above their share of the voting-age population and far above
their share of the electorate in other elections. Note that blacks were 21% of
New York City electorate in the November 1997 mayor's race, according to a New
York Times exit poll, but won 32% of school boards seats in May 1996.
ï Asians: Asians in
1996 were elected in proportion to their voting age population in 30 of 32
districts, as was true in 1993. Asian voters do not make up 20% of voting age
population in any school district in the city, yet Asians have at least one seat
in 7 districts, and 11 of 15 Asian candidates won.
ï Latinos: In 1996
Latino candidates were elected in proportion to their voting-age population in
18 of 32 districts, up from 13 of 32 districts in 1993. Latinos are
under-represented by only one seat on 12 of the 14 school boards where they do
not have proportionate representation. Overall, Latinos won 21.5% of school
board seats in 1996, up from 17% in 1993 -- mirroring the rise in the Latino
share of the New York City electorate from 13% in the 1993 mayor's race to 20%
in the 1997 race.
But all these numbers don't
necessarily convince parents that the boards are producing representative
results. And indeed the student population of the city's school continues to
differ significantly from the voting population, with far more students of color
than white students. This discrepancy helps explain the fact that minorities are
actually over-represented on the school boards, as parents are somewhat more
likely to participate than non-parents. But non-parent voters have every right
to participate in elections, and they must be counted.
Thus, in 1994 Asian Americans
made up 33% of the student population in District 2, but won only one of nine
seats in both 1993 and 1996. Why? Asians that year were only 10.6% of the
voting-age population in District 2, and almost certainly below 10% of
registered voters. Queens is another good example for Asian Americans -- and one
underlining how well the choice voting system is representing minorities. Asians
in 1994 made up 33% of students in districts 25 and 26. Yet that year they were
only 19% of the voting-age population (VAP) in district 25 and 11% of VAP in
district 26 -- and less than 10% of registered voters. Still, Asian candidates
won three of nine seats in both districts in 1996 -- reflecting in part the fact
that over half of parent registered voters were Asian in these districts, and
Asian voters probably turned out in disproportionately high numbers.
As a final point about voting
rights, let me point out that the boards reflect a very complex diversity.
Eleven of 32 district boards have representatives from at least three of the
city's four major racial and ethnic groupings -- whites, blacks, Latinos and
Asian Americans. Such diversity would be very hard to replicate with even
gerrymandered single-member district plans in these districts.
Other Proportional / Semi-Proportional Systems
I understand that the task
force is looking at other proportional and semi-proportional systems. The two
other major such systems that can be used in non-partisan elections are limited
voting and cumulative voting. Both are currently in use in a number of
jurisdictions around the country. Please see my article from the Federal
Elections Commission Journal on "Alternative Electoral Systems as Voting
Rights Remedies" in the appendix for more information on these systems,
including full descriptions of how they work. Additional information in the
appendix about these systems includes: a chart on cumulative voting and racial
representation in localities around the country; a piece contrasting different
systems; a question/answer piece that I wrote for the Southern Regional
Council's Voting Rights Review in 1995; and a chart on "thresholds
of representation" using different systems.
As an overall comparison, choice
voting has a surface complexity, but an underlying simplicity that improves
chances for accurate representation and reduces demands for complex voting
strategies. Limited voting and cumulative voting are far more transparent in how
they work, but have more complex strategic consequences for both individual
voters and for parties (or slates).
The "thresholds of
representation" chart shows that cumulative voting and choice voting have
the same threshold of representation for winning one seat (as does limited
voting when candidates have only one vote). But because of the mechanism of the
"single transferable vote", choice voting is likely to represent well
every increment of voters -- meaning a group of voters making up 40% of the
electorate is likely to win 40% of seats. Cumulative voting and limited voting
are "all-or-nothing" systems, and thus far more votes can be wasted.
With cumulative voting, a group of voters making up 40% of the electorate might
win 20% of seats or might win 70% of seats depending on how the votes were
spread among the candidates.
I understand that the task force
may take a particularly close look at limited voting. The threshold of
representation for limited voting is determined by the formula:
Limited Voting Threshold (to win
Votes + Seats
Note that this threshold rises with the number of votes a voter has. In a
nine-seat school district in which voters had five votes, the threshold would be
5/14, or 36% -- high enough that the Department of Justice would be unlikely to
pre-clear this change in any of the seven school board districts where
Asian-American voters are now electing candidates of choice and in the many
districts where black or Latino voting-age population is below 30%.
Without touching on the full
range of possible reforms of the election process, I would like to propose
consideration of several possible reforms, some of which go particularly well
1. Sub-districts, with a proportional system: One possible reform is
creating smaller constituencies within school districts. Districts could have
three, three-seat districts, for example. Unlike the Marchi commission proposal,
however -- a proposal that almost certainly would not have been pre-cleared by
the Department of Justice -- I would suggest that a proportional system be
maintained in such districts. Choice voting could be maintained as one option,
or limited voting (with one vote) and cumulative voting could be considered --
they are easier to operate in constituencies with fewer seats. Regardless, the
threshold of representation would rise to 25% with a three-seat district, but
minority voters still might have a fair chance to elect candidates -- most
school districts probably have relative concentrations of different groups of
minority voters in different parts of the school district.
Candidates could campaign in a
smaller area, and voters would have fewer candidates to consider. On the other
hand, drawing sub-districts would be expensive and an administrative headache,
and arguments can be made that such a change might lower voter turnout. If
sub-districts were used, my one strong suggestion is that a proportional system
2. Ballot information: The school board elections are non-partisan.
Without the guide of party labels, voters have to work hard to know something
about candidates -- which can be particularly difficult given the low degree of
media coverage of the elections.
In 1974, in his testimony on
"Use of Proportional Representation for Community School Board
Elections" to the New York State Charter Commission for New York City (see
appendix for full testimony), George Hallett of the Citizens Union suggested
that candidates be allowed to have a designation on the ballot identifying a
slate or an organization -- not political party. Such information would make it
easier for voters to identify allies and harder for candidates to run
3. Media education: The Center for Voting and Democracy's project for the
1996 elections could be greatly expanded upon with staff members of the Board of
Education and outside consultants. Much more work could be done in particular
with journalists, who often explain the system poorly or inaccurately. New York
has a wide array of media outlets, with several daily newspapers, many borough
and neighborhood papers and numerous radio, cable and television news programs.
Journalists have a great capacity to reach potential voters, but unfortunately
as far the school boards, this potential has been largely unrealized.
Many journalists actually
contribute to the climate of defeatism surrounding the school board elections by
exaggerating choice voting's complexity. A quick survey of print coverage in
1993 revealed that the Daily News called the system
"incomprehensible," the Post called it "complicated,"
the New York Times ran a front page story calling it
"complicated" in a headline and Newsday called it
"outdated." The Village Voice made a special effort to spur
people to the polls, but inaccurately described how to vote. The Times
ran a editorial on the day of the election -- a day when many readers were
making a final decision about whether to participate -- that was headlined
"School Board Sham," included an inaccurate description of how the
system works and concluded the system was "complex and confusing."
Even when reporters correctly
describe how to vote, they do not explain how it provides important
opportunities for voters to win fair representation -- thus doing little to
inspire voter turnout. Aggressive outreach to journalists, from education
reporters to editorial boards, should lead to much improved coverage. Better
media coverage would reach a broad range of New Yorkers who are involved in the
school board elections and are in a position to the improve the climate of the
elections, from currently misinformed community leaders to potential voters. One
of the most important audiences would be election officers, who are now not
required to attend training sessions on choice voting and far too often
misinform voters right at the polls. News stories reported on such incorrect
instructions, as does Edward Stancik in his critical report on the 1993 school
4. Community education: Going hand-in-hand with a media education
campaign could be a campaign to get information to community organizations
already involved with the school board elections. A range of non-partisan
educational materials -- print, video, short pamphlets, posters and the like,
with some material geared for English-speakers and some geared to immigrants
speaking other languages -- could be created for community organizations to use,
adapt and distribute as they see fit. These organizations could be asked to help
the project improve the material and to find opportunities to put the
information into the broader context of the significance of the boards;
potential voters must not only understand the power of their vote, but the power
and potential power of the boards that will be elected.
Many parents and potential
voters are much more likely to respond to information from community
organizations than citywide media. Today, as Donald Murphy of People About
Changing Education told the Daily News in 1993 "people are so
cynical and filled with feelings of hopelessness in places like District
12." As the Daily News editorialized after the 1993 elections,
"the system is a turn-off to parental involvement." If parent
organizations working directly in these demoralized districts had more and
better material on choice voting, they would have a better opportunity to
inspire their supporters to find hope in the election process.
5. Ballot-counting modifications: Current ballot-counting rules for
choice voting are based on the count being carried out by hand. Now that the
city has licensed a computer program to do the ballot-count -- once information
from ballots is entered into the computer -- there is no reason to continue the
exact method now used. Without going into great detail, let me say that I
recommend that the method of allocating surplus votes from a winning candidate
be changed to "the mathematical model."
In the current law, some
ballots are transferred at full value to next choices on these ballots.
In the mathematical model all ballots from the winner are transferred to
next choices at an equally reduced value. This change makes the count
precise, meaning that the results will be exactly the same no matter in what
order ballots are counted.
The law also could be modified
to allow voters to cast "tie" votes. The Center has no official
position on this potential change, but I would be happy to answer questions
about it. Tie votes would virtually eliminate invalid ballots -- which were 2.5%
in 1993 -- but make the ballot-count more complicated to explain.
6. Election Administration / Vote-By-Mail / Voter Guides: Election
administration of choice voting has not been a success. The City of Cambridge in
November 1997 finished its choice voting election ballot-counts for city council
and school board elections in one day, using scannable ballots and a computer
program. For its choice voting elections for its national parliament, Ireland
usually finishes its count by hand in one or two days. New York City, however,
generally has taken several weeks to complete ballot-counting in all school
districts. I am pleased to see that the Board of Elections has taken important
steps to improving its administration of the elections, but hope it continues.
The Center for Voting and
Democracy has done some limited research into vote-by-mail elections -- e.g.,
elections in which polling places are not open, but voters have a certain number
of weeks to mail or drop off a ballot that has been mailed to them. More and
more cities and states are using vote-by-mail elections, and New York City's
Voter Assistance Commission gave the idea a favorable recommendation in 1994
testimony to the New York Task Force on Voter Registration. Steps would need to
be taken to ensure ballot security and prevent corruption, but such steps
already must be taken with absentee voters. Other jurisdictions using
vote-by-mail have also developed techniques for dealing with these concerns.
One of the clearest findings of
vote-by-mail elections is that they boost turnout for exactly the kind of
election we are discussing -- local elections held when not much else is on the
ballot. Studies of vote-by-mail election in Oregon show turnout for such local
elections sometimes triples. It is easier for voters to cast ballots, and easier
for the media and civic organizations to encourage higher participation because
voters have two-to-three weeks to cast ballots.
Vote-by-mail also would allow
the Board of Elections to use "marksense" ballots -- as now used in
choice voting elections in Cambridge (MA) -- which then could be used in
conjunction with ballot-scanning machines to record choice voting ballots
accurately and quickly. This would be a great relief to candidates and their
supporters who no longer would have to spend weeks monitoring a ballot count.
Vote-by-mail elections are also
significantly cheaper than polling place elections -- a 1994 U.S. General
Accounting Office study found that the cost per ballot in vote-by-mail elections
had been about one-third cheaper than the cost of conventional elections. Some
of the savings perhaps could go to taxpayers, but it also could go to voter
guides to be sent out with ballots. These voter guides could have information
about the choice voting system -- or other system, if a change is made -- and
about the candidates. Better informed people are more likely to vote -- and more
likely to feel good about their participation.
Thank you, and I look forward to