The Paradox Of Proportional/Party Voting Vs. Winner-Take-All Voting

By David Grace (

My recent column on proportional/”party” voting (Another Way To Neutralize Gerrymandering. Don’t Change The Map. Change How The Votes Are Counted) versus winner-take-all voting drew quite a bit of interest and some thoughtful comments.

I think this is an issue that deserves further discussion.

At-Large Elections Vs. District-By-District Elections

Multi-candidate elections can either be “at large” or “district-by-district.”

For example, people might elect the members of a seven-person city council in a city-wide vote where each citizen would distribute seven votes among a list of more than seven candidates, or the city might be divided into seven districts with each person voting for one candidate to represent their own district .

In a city-wide election system, a majority of the city’s voters elect all the candidates and the minority of voters elect no candidates, that is, a block of 51% of the voters would elect 100% of the councilmen.

To avoid this total control by one group of voters, many municipal elections are held on a district-by-district basis, but that doesn’t necessarily solve the problem of some voters having electoral power that is not commensurate with their numbers.

Manipulating Election-District Boundaries

If an interest group is able to control how the district maps are drawn, it can manipulate the district boundary lines so that the members of their group are elected in numbers that are disproportionate to their size.

When state legislatures controlled by one party manipulate legislative district boundaries to give an election advantage to the members of their party, that’s called “gerrymandering”.

Proportional/Party Voting

One tactic employed to keep a party from electing more candidates than its actual numbers would otherwise entitle it to is some form of proportional voting that attempts to match the number of candidates elected over the entire city, county or state with each party’s proportional share of the total votes cast.

So, in terms of mechanisms employed to limit the power of a group to gain governmental power in excess of the size of their membership we go from

  • Electing all candidates at large where 51% of the voters elect 100% of the candidates, to
  • Electing one candidate per district where the district boundaries can be drawn so as to favor one group over another, to
  • Electing one candidate per district where district boundaries are neutrally drawn, to
  • Electing candidates over an entire city, county or state in proportion to the votes cast for each of the parties to which those candidates belong.

The Problem With District-By-District Voting

Let’s look at the fictional state of Columbia with five Congressional districts. In the example in my previous column, the statewide party registrations were:

  • 40.6% GOP
  • 46.4% Democratic
  • 10% independent
  • 2% Libertarian
  • 1% Green

What If District Registration Mirrored Statewide Registration?

Suppose we could magically make each of the five districts exactly mirror the statewide party registration. In that Utopian case both parties would be competitive in each district with the 10% independents being able to swing each election to either party.

But, suppose the statewide registration numbers were different. Suppose they were:

  • 41% GOP
  • 51% Democratic
  • 6% independent
  • 1% Libertarian
  • 1% Green

If the districts exactly mirrored the above table of statewide party registrations then we would expect to see Democrats win every one of the five districts and elect 100% of the congressmen in spite of representing only 51% of the registered voters.

The Paradox of District-By-District Voting Vs. Proportional Voting

Proportional-representation voting and winner-take-all voting each have pros and cons and neither one is perfect. Which one you prefer is something of a subjective judgment call.

If we used district-by-district, winner-take-all voting under the above registration example, the majority of the voters in each district would get the representative they preferred, but the overall result would be that barely more than half the voters would have picked all five candidates and barely less than half the voters would have elected no candidates at all.

On the other hand, if the districts don’t mirror that statewide registration then it’s possible that a party representing a minority of the registered voters might end up electing a majority of the candidates.

So, here’s the paradox of Proportional-Representation Voting versus Winner-Take-All Voting:

  • Under proportional-representation voting, if the Democrats receive 51% of the vote in each of the five districts they will still win no more than three of the five contested seats.
  • The downside is that the two seats won by the Republicans will go to candidates who received a minority of the votes cast in the districts in which he/she was a candidate.
  • On the other hand, if you have winner-take-all voting, each district will get a representative that the majority wants, but the majority party might elect all five candidates and the minority party will get no representation at all.

So, what’s more important to you:

  • A majority in each district getting the representative they picked but with the minority in each district getting no representation at all, or
  • The majority and minority each getting representation proportional to their numbers but with some districts being represented by candidates who received a minority of the votes cast in those districts?

In A Perfect World . . .

Yes, in a Utopian world each district would be 40% Republican, 40% Democratic and 20% Independent and the best person from either party would be elected. But it doesn’t work like that in the real world.

Whether by circumstances or by design, representation in districts will not generally mirror the ratio of statewide party registration and, by their very nature, winner-take-all district elections can (and often do) result in the over-representation of a minority of the voters.

Even worse, if party registration in districts does mirror statewide party registration and if one party has a material registration advantage over another, then that party will win elections far in excess of its share of the electorate.

So, again, what’s more important:

  • Districts getting representatives that were approved by a majority of their voters with little or no representation going to the minority party, or
  • Electing candidates in proportion to party membership but with some districts getting representatives who received a minority of the votes cast in their district?

Which alternative is your choice?

Only A Small Number Of Independent Voters Might Actually Go Either Way

I know that people who register as Independents make up a larger proportion of the population than I’ve included in my example, but studies have shown that most Independents lean either Democratic or Republican and that they almost always vote for the party they lean toward.

Only about 12% of Independent voters are actually “in play” in most elections.

Of those 12%, you’re likely to see 4% go to each major party with only the remaining 4% actually switching one way or the other. In other words, you’re likely to see that 12% actually voting someplace in the range between 4% Republican and 8% Democrat and 8% Republican and 4% Democrat.

In my example above I lumped the Demo-leaning independents in with the Democrats and the GOP-leaning Independents in with the Republicans and I increased that true, floating 4% to 6% who might actually go either way.

Party Primaries Exacerbate The Problem Of Minority Control

Would reducing the power of political parties result in elections that were more focused on the strengths and weaknesses of individual candidates instead of their adherence to party ideology?

This problem of a minority of voters having power beyond their numbers is even more extreme if party candidates are chosen in party primaries.

Only about 20% of party members actually bother to vote in the primary. So, if a district has 330,000 registered voters, of which about 168,000 are Democrats (51%) then only about 34,000 Democrats will likely show up to vote in the party primary. Of that 34,000, a little over 17,000 votes will be enough to pick the Democrat candidate which candidate will likely win the general election in that 51% registered Democrat district.

Essentially, these 17,000 dedicated, hardcore party members who show up for the primary election will almost always end up picking the Congressman who is elected from that district holding 330,000 total registered voters.

Eliminate Party Primaries?

One way to reduce the power of political parties and the control of those parties by their ideological, hardcore members would be to eliminate party primaries altogether.

What if we said that anyone who got nominating-petition signatures equal to at least 2% of 330,000 registered voters (6,600 signatures) would qualify for the preliminary candidate list?

In order to avoid having so many candidates that no one would be able to make an informed choice among them, we would put the incumbent and the next five candidates with the most nominating signatures on the final primary ballot.

The candidates could call themselves whatever they wanted, Democrat, Republican, Libertarian, Independent, Communist, whatever.

Each citizen would get two votes, a first choice and a second choice. Their first choice candidate would get two votes and their second choice candidate would get one vote. The two candidates with the highest vote totals would appear on the general election ballot.

This sort of system would put the two most popular candidates on the ballot, would reduce the power of the party insiders, and make incumbents less vulnerable to the demands of the most extreme members of their party.

A Thought Experiment

I offer this thought experiment: What if we had applied some variant of this system to the 2016 Presidential election?

Suppose we had put all the Republican and all the Democratic candidates on one primary ballot in each of the first five primary elections and told every primary voter to pick their first and second choices? Suppose that we then took the top five or six vote-getters from those first five primaries and put them on the ballot for the rest of the primaries, again under a “pick two” system?

Do you think that Trump and Hillary would have been the top two candidates when all was said and done?

Would their respective parties have picked the actually primary winners as their standard-bearers in the general election?

In a system like that I really wonder if either Hillary or Trump would have made the cut.

— David Grace (

To see a searchable list of all David Grace’s columns in chronological order, CLICK HERE

To see a list of David Grace’s columns sorted by topic/subject matter, CLICK HERE.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store