Proportional
Representation:
Advance
of Democracy
THEODORE S. ARRINGTON
Professor
of Political Science, University of
North
Carolina at Charlotte
Testimony
during hearings on “H.R. 1173/States’ Choice of Voting Systems Act” before
the U.S. House of Representatives Committee on the Judiciary, Subcommittee on
the Constitution, September 23, 1999
My concern as a political scientist, expert witness, and political activist has
always
been
to further the process of representative government. I ask: what can we do which
will
make the system more effective and accurate in translating votes into seats on
governmental
bodies? I certainly want the majority to rule, but I believe that minorities
are
entitled to a seat at the table, a voice in the legislature. When I say
“minorities”
here
I am not necessarily referring to racial or language minorities which are
protected
by
the Voting Rights Act [of 1965]. The minority that needs representation in any
particular state or region on any particular governing body may be a partisan,
religious, ideological, ethnic, or some other kind of minority. I believe in a full and fair
representation for everybody. And when I speak of “representation” I mean simply
that
voters have somebody they would choose speaking for them in government bodies.
PROBLEMS
WITH SINGLE MEMBER DISTRICTS
One
might ask, “Where’s the beef ?” The single-member district system has
served
our
country well for over two hundred years. In many cases it continues to provide
full
and fair representation for the U.S. House of Representatives, state
legislatures,
and
many local units of government. I would be the first to argue against wholesale
changes
in our methods of election, but the demands of the current era may be different
from
the demands of the past. Specifically I believe that two changes have made
single member
districts
less effective in providing full representation.
First, there seems to be increasing diversity of race and ethnicity, life style,
living
conditions,
religion, ideology, language, and values in America. This change may be a
product
of immigration or perhaps increasing diversity in the media. Second, these
diverse
groups are often not concentrated geographically. Where once people could be
said
to identify their interests by where they lived, this is increasingly uncommon.
Especially
in our urban areas, people with very different values and concerns may be
living
near each other. Although people may group themselves into neighborhoods based
on
shared concerns and values, only in the very largest cities can such
neighborhoods
become
most of a congressional district. In most urban areas many neighborhoods
must
be combined to form a district. Thus the concerns of minorities may be
sublimated
or
submerged.
Single-member districts work well to represent minorities when they are large
enough
and geographically concentrated enough to win at least one district. But
minorities
that are geographically dispersed may not be able to win any district in a state
and
may therefore never be able to elect a candidate of their choice even when they
are
numerous
and cohesive in their voting. Recent changes in court rulings on the
Voting
Rights Act illustrate this point. In the original Gingles
case [Gingles
vs.
Edmisten,
a 1984 U.S. District Court case involving redistricting in North Carolina],
which
defined the application of the Voting Rights Act, the court gave three
“prongs”
which
must be proven to show that an at-large election system is in violation of the
Voting
Rights Act. One of these was that the protected minority must be large
enough
and geographically concentrated enough to form a majority in a single-member
district.
“Geographically concentrated” was not defined. Initially this was
taken to mean that any geographically contiguous area could be a district. Since then, however,
the Shaw case (also litigation from North Carolina), and its progeny have further refined
what
“geographically concentrated” means. In doing so, the courts have
declared
a number of Congressional districts which were designed to provide
representation
for
racial minorities to be unconstitutional.
The practical effect of Shaw
[Shaw
v. Reno (1993),
in which the U.S. Supreme
Court
rejected creating districts based on race to promote the election of minority
representatives]
is to make it impossible in many states to provide representation for
racial
minorities through the drawing of single-member districts. In New York City
or
Atlanta there are enough African- Americans concentrated in one area to form
one
or more Congressional districts in which they could choose their representative.
In
North Carolina, to take a contrary example, this is not the case except perhaps
in
the northeastern portion of the state. In the Piedmont area of North Carolina
there
are
more than enough blacks to form a majority in a congressional district, but they
are
dispersed across several geographically distinct urban areas.
I wish to emphasize here that I am only using the example of racial minorities
because
the
Committee members will be familiar with the developments in this area. What is
true
of racial and language minorities is also true of other kinds of minorities as
well.
A second problem is that the single member district system forces us to choose
between
representation and competition. Whenever someone complains about the
lack
of competition in congressional elections, I remind them that we are electing a
House
of Representatives
not a
House of Competitors.
Nevertheless, competition for
office
is an important value, and the lack of competition for congressional office
should
be a concern for all of us. If one draws a set of districts in such a fashion as
to
form districts that are relatively homogeneous, one will provide for a good deal
of
representation.
The representative will be able to know the interests of his or her constituents
to
a high degree, and to vote in accord with their desires. But such districts
will
naturally be noncompetitive especially in a partisan sense. Competition within
the
primary
rarely if ever provides a substitute for vigorous two-party contests. On the
other
hand, if one draws districts to maximize competition, there will necessarily be
a
very
large minority in each district which feels alienated from the system having
voted
for
candidates who consistently lose. Each district would be diverse and it will be
impossible
for one representative to satisfy more than a large fraction of his or her
constituents
in
voting on controversial issues. The congressman might, however, serve
almost
all of them in terms of constituency service, providing information, etc.
A
third problem with the single-member district system is that the drawing of
districts
has such a drastic impact on the process of translating votes into seats that
the
practice of gerrymandering [manipulating the boundaries of district to favor a
political
party or candidate] has become more of a threat. It has become accurate to
say
that it is the legislators who choose the voters rather than the voters who
choose
the
legislators. Of course, state legislatures draw congressional districts, but the
point is
still
valid as congressmen have friends and allies in the state legislature and exert
much
influence there. The art of gerrymandering has been greatly aided by the rapid
development
of geographic information systems which allow anybody to draw many alternative
districting
schemes in a short period of time. I define gerrymandering in the traditional
way
as being an effort to draw the districts in such a fashion as to misrepresent.
That
is, to distort the process of translating votes into seats so that some group
will
receive more seats than their votes would entitle them to while some other
group
would receive less. Gerrymandered districts might not be oddly shaped.
A related problem with single-member districts is that
one may be forced to decide
which
kind of minority one wants to represent. Drawing the districts one way may
provide
fair representation for urban and rural interests, but provide an unfair
partisan
balance.
Drawing the districts another way makes the partisan divide fair, but
might
result in having all districts dominated by urban interests. Drawing a set of
districts
which represent a diversity of crosscutting interests may be impossible. In
other
words, the district drawer has to decide what kinds of interests deserve
representation
and
which ones do not. Again, the politicians choose the voters instead of
the
voters choosing the politicians. Some individuals argue that one should
draw
districts without regard to any interests —use natural boundaries and be as
compact
as possible. But as Justice White once observed, this can result in districts
which
are egregious gerrymanders or obviously unfair. He called it a “politically
mindless
approach.” Geography is not destiny… at least not any longer in our urban
and
suburban nation.
MULTI-MEMBER
DISTRICTS
Although
there are no multi-member congressional districts, various forms of proportional
or
semi-proportional election systems are in use in local governments
throughout
America. Such systems are also in wide—indeed, almost universal—use
throughout
the world. From the examples of these systems in the U.S., I conclude
that
semi-proportional systems can provide a remedy for some of the problems
noted
above.
First, semi-proportional systems provide a method of representing minorities
that are
numerous
enough to elect a representative but too dispersed geographically to form a
majority
in a single-member district. This is a remedy for the “Shaw
problem.”
But it is
applicable
to much more than providing representation for racial or language
minorities.
There are many examples of partisan or other minorities in regions of
some
states that lack representation. In general, we would expect that multi-member
districts
with semi-proportional systems should provide a more accurate and efficient
method
of translating votes into seats. Diversity in representation could be
enhanced
by multi-member districts and semi-proportional voting systems.
Second,
multi-member districts with semi-proportional voting systems allow us to
combine
both representation
and competition. It is possible for a far larger majority of
the
voters to feel that they supported someone who represents them. Yet competition
is
also
increased. Contrary to the myths that many Americans—including some Supreme
Court
Justices—have about semi-proportional systems, they are not characterized by
“immobilization.”
Rather, such systems are more likely to have vigorous competition
than
single-member districts.
Third, multi-member districts are less subject to gerrymandering than
singlemember
districts,
although the problem does not go away completely. Semi-proportional
voting
systems are “open” in the sense that the voters decide how to group
themselves
politically
in terms of the candidates they support. Voters may decide that they
are
Republicans, or farmers, or Presbyterians when they vote and combine their votes
with
others of a like mind.
SEMI-PROPORTIONAL
ELECTION SYSTEMS
I
am sure that the Committee members have an intuitive grasp of what a
multi-member
district
would be like, but may not have an idea of semi-proportional election systems.
Proportional
methods of election, such as the “single transferable vote” system have
their
advocates,
and I hope some of them testify today. But I think the American system of
government
would be better served with two the semi-proportional systems that I have
studied.
I favor these systems for two reasons. First, because they do not require the
voter
to do anything unusual, allowing a ballot layout which is simple,
straight-forward,
and
familiar to the voter. And, second, existing voting equipment can be programmed
to
count the votes with these methods easily and cheaply.
The first of these methods is called “limited voting.” Suppose a state were
divided
up
into triple-member districts. North Carolina, for example, could easily be
divided
into four three-member districts. Suppose further that each voter could only
vote
for two people. This would provide minority representation in each district.
This
method of election is used by a wide variety of local governments, concentrated
in
North Carolina and Alabama. The second method is called “equal allocation
cumulative
voting.” Suppose again that the state is divided into triple-member
districts.
Each voter has three votes. But if the voter should vote for only two people,
each
would
receive 1.5 votes. If the voter found only one candidate worthy of his or her
vote,
that
candidate would receive three votes. This method was used in Illinois for the
lower
house
of the state legislature until relatively recently, and is used in one form or
another
in
various local elections in the south, especially in Texas. I prefer “equal
allocation”
cumulative
voting to systems which require the voter to allocate the votes two for this one
and
one for that one. Few voters evidently take advantage of this flexibility, and
it greatly
complicates
the layout of the ballot. I prefer systems which allow voters to do exactly what
they
already are accustomed to doing: vote for a candidate or don’t.
Evidence from systematic scientific studies of these voting systems in use in
the U.S.
show
that voters understand how these systems work. These systems provide meaningful
minority
representation, while still allowing majorities to form and be elected.
Competition
is increased as well. In some countries proportional representation
systems
produce too much diversity in the legislature with very small minorities
controlling
policy. This is a problem in Israel, for example. But if the multi-member
congressional
districts
are held to three- or four or even five-member districts, this problem is
avoided.
To win a seat in a three-member district a minority must be of substantial size
and
vote cohesively. Very small minorities will not be systematically represented.
Nor
am I concerned that multi-member districts and semi-proportional systems will
harm
our two-party system. The reasons why we have a two-party system and other
countries
have many parties are not completely understood. In any case, the single member
district
system has not produced a two-party system in Canada or Great
Britain,
and is probably not the major factor in producing our unique party system.
My
own view is that the election of presidential electors by plurality vote within
each
state is the glue which holds our two party system together along with tradition
and
various provisions of state law.
STATES
RIGHTS
I
am well aware of some of the reasons for
the
congressional mandate that states use single-member districts. In the early
1960s
I
was living in New Mexico. At that time
the
state had two democratic congressmen,
both
elected at-large. It was common
knowledge
that this was a form of a gerrymander.
It
was thought that if the state
were
districted into two, the district with
Albuquerque
would elect a Republican.
Indeed,
when the state acquired a third
congressional
seat and was districted, the
Republicans
won two of the three seats, and
have
held at least two since.
However, such abuses can be prevented
without
restricting the states from experimenting
with
multi-member districts with
proportional
or semi-proportional voting
methods.
I recommend that state legislatures
be
required to use proportional or semiproportional
voting
systems if they choose
to
use multi-member districts.
I would oppose requiring states to use
multi-member
districts or any particular
form
of voting. Our states are the great laboratories
for
reform. This is one of the most
important
features of our federal structure.
The
Congress should free the state legislatures
to
experiment with these proportional
or
semi-proportional systems. If the experiment
proves
successful, other states will follow.
If
the experiment is not successful it will
be
abandoned. I recommend that we allow
states
the right to use a variety of methods to
choose
congressmen. No more than this is
needed,
no more is wise at this time.