James and Tomilea Allison

The sheer mechanics of voting bear much of the blame for our government’s failure to represent the voters. When 90% of us want single-payer universal health care, why do our elected representatives fail to deliver time after time? Is our “democracy” so very different from that of other countries, whose governments have delivered the kind of health care system American voters want? The same question applies to many other aspects of American life: our criminal justice system; climate change; the infrastructure; mass transit; affordable housing; public education; income inequality; war and peace. Are we so different? Yes and no. This is one of those things where seemingly small differences can have very big effects, including the disenchantment of the voting
public.

Let’s start small, with the 9-member city council of the city of Bloomington, Indiana. That body is usually 100% Democrat, although the popular vote is about 1/3 Republican. Such disproportionate representation is not what we mean by representative government. The problem there is the winner take-all system in each of six council districts, and three at-large districts. If the voters were distributed at random throughout the city, one would expect exactly what happens: Each of the six districts would have a majority of Democrat voters, and so would the city at large, so every one of the nine council seats would likely go to a Democrat. This result comes directly from our conventional winner-take-all mechanics, where the candidate with half of the votes plus 1 wins all of the contested seats (one), and the loser gets the rest (zero).

But who says we have to use the customary American winner-take-all gadgetry? Elsewhere we see proportional representation by party, and more representative government. For example, suppose Bloomington had three council districts instead of nine, with three council seats in each district. That way, if Democrats won 2/3 of the votes in a particular district they would win only two of the three contested seats instead of all three, as they would under winner-take-all. Over time, each party’s share of the council seats would approximate its share of the popular vote.

And who says we have to use a two-party system? The question becomes more acute as both parties somehow drift away from popular sentiment. If that’s how you feel, you might take a look at ranked-choice voting, which loosens the straitjacket of the two-party system. As it happens, both the virtues of ranked-choice voting and the evils of the gerrymander revolve around the concept of the wasted vote, so that is where we start.

There are two kinds of wasted vote. One kind is any vote cast for the losing candidate. The other kind is any vote in excess of the number needed to win—if majority rules, that would be half of the votes cast, plus one. Any additional voters for the winning candidate might as well have stayed out of the election. And the logic of gerrymandering for partisan advantage is simple: Set up the election districts in such a way that your opponent is likely to waste more votes than you do. Typically this amounts to drawing the district maps so as to pack your opponent’s voters into as few districts as possible. As a result, your opponent will typically win only a few districts, but those few wins will be lop-sided, and in what we call “safe districts.” In contrast, your own party will win considerably more districts, but by relatively small margins of victory, in what we call “competitive districts.”

There are several notable consequences of partisan gerrymandering. For one thing, the
candidate fortunate enough to live in a safe district need never worry about the competition. For another thing, the constituents in that safe district have no particular reason to concern themselves with government. They needn’t bestir themselves on election day, because they think the incumbent is almost sure to return to Congress or the state house, whether the individual constituent bothers to vote or not. In a safe district, they think, no one voter is crucial. In addition, their representative will
end up sitting on the minority side of the legislature anyhow, no matter how big the district turnout might be. Because the purpose of the gerrymander, and its effect, is to subvert representative government in favor of minority rule. It is that effect, amplified by modern computational tools, that has driven so many states at last to raise judicial challenges to their gerrymandered districts. Such districts are plainly unconstitutional violations of representative government, where a portion of the voters control a disproportionate share of the legislature.

How could we draw the districts more fairly? We had best leave that job to nonpartisan
political cartographers armed with computers, specially designed software, and data from past elections. But we might well suggest that they do their best to minimize wasted votes.

Ranked-choice voting brings its own kind of focus on the redemption of wasted votes. For example, take the method adopted by the voters of Maine in a 2016 referendum bitterly opposed by the state’s governor, legislature and Supreme Court, re-adopted in a subsequent referendum, and finally scheduled for use in Maine’s 2018 elections for state and federal offices. No other state has gone so far in the use of this method, often called “instant runoff.”

Suppose we have four candidates on the ballot for mayor. Each voter ranks the four
candidates in order of preference. If a candidate receives a majority of first-preference votes (50% plus 1), that candidate is declared the winner. Otherwise, the candidate with the fewest first-preference votes is dropped from the race. However, that eliminated candidate’s ballots stay in play. Specifically, the second-preference votes on those ballots are bumped up to first, and all of the ballots are then re-tallied. The same procedure is followed, as needed, until a winner is declared. A hypothetical example:

First Tally

Candidate	First-Preference	Votes
A	475	46.34
B	300	29.27
C	175	17.07
D	75	7.32

Second Tally

Candidate	First-Preference	Votes
A	525	51.22
B	325	31.71
C	175	17.07

What happened here is that on the first tally no candidate had a majority of the first-preference votes, so no winner was declared. Accordingly, Candidate D, with the smallest number of first-preference votes, was eliminated. However, on all of that candidate’s first-preference ballots the second-preference candidate was bumped up to first, and all of the 1,025 ballots were tallied again. The second tally showed that Candidate A had gained 50 first-preference votes, for a total of 525—a clear majority of 51.22%, and a clear win for Candidate A. Thus, among the 75 voters who originally chose Candidate D as their first preference, 50 also chose Candidate A as their second preference. It was those 50 who put Candidate A over the top in the second tally. The other 25 had chosen Candidate B as their second preference. And we can imagine a subsequent tally that followed a different path, one that might have declared Candidate B the winner instead of Candidate A.

But why bother? Isn’t it true that if we had used the old fashioned most-votes-wins procedure, Candidate A would have won anyhow, on the very first tally, with 46.34% of the popular vote? Yes, but that would have given most voters, 56.66% of them, reason to feel that they had wasted their votes completely in a losing cause. In this particular example, but generally too, the ranked-choice method gives the same outcome, but better vibes. With our ranked-choice method, 51.22% of the voters can claim to have taken part in the choice of Candidate A as the next mayor. In addition, consider the voter who thinks her true favorite, say the Green candidate, has little chance of winning. That voter can rank her true love first, to show the flag and nurture the movement, but also rank a more politic favorite second without jeopardizing the election. Thus can third parties grow. “This time I got my second choice. Maybe next time, my first.”

As of November 2017 multiple cities in five states had adopted and implemented rank-choice voting in municipal elections (California, Colorado, Maine, Maryland, and Minnesota). In four additional states rank-choice voting had been adopted but not yet implemented (Florida, Michigan, New Mexico, and Tennessee). Not all versions focus on votes wasted on losing candidates. In Cambridge, MA, where ranked-choice voting is used to pick a 10-member city council, the focus is on wasted votes cast in excess of the number needed to win. And in a place like Cambridge, with so many seats in play, a
relatively small coalition can claim a seat on City Council with an informative role, if not a dominant one.

None of the alternative voting methods is perfect. Indeed, Kenneth Arrow won a Nobel Prize for his mathematical proof that no voting system could satisfy all of the requisite criteria. But all such reforms—to slay the gerrymander, reverse Citizens United , resist voter suppression—share the implicit recognition that representative government, always under attack, is a worthy ideal with many defenders and many means of defense.

References

Ballotpedia: Ranked Choice Voting (RCV). Retrieved April 13, 2018 from ballotpedia.org/Ranked-choice_voting_(RCV)

Fair Vote: Ranked Choice Voting/Instant Runoff. Retrieved April 13, 2018 from http://www.fairvote.org/rcv#how_rcv_works

Jan Devereux: About our Ranked Choice Voting System. Retrieved April 13, 2018 from jandevereux.org/about/about-our-ranked-choice-voting-system/
Note: The jandevereux link gives a clear explanation of the system used for decades in Cambridge, MA, to elect the 10 members of the City Council.