STV as proportional representation

Is the single transferable vote a form of proportional representation? Right now I believe it is. A reviewer once said the fight had been won. Either way, this issue stops some of my talks dead in their tracks. Some are bothered that PR might coexist with candidate-based ballots.

Yet among the countries (and subnational units) we usually think of as having PR, early adopters often went for candidate-based ballots (e.g., Switzerland). A working hypothesis is that STV is just one of several PR systems whose debut predates the mass party. PR systems like these find appeal where politicians, for whatever reason, want some level of distinction from their party’s brand.

This page exists to collect my evolving thoughts on the issue. I plan to update it as needed.

Regardless of whether STV is “PR,” one thing is certain: it definitely is not a plurality or majority voting rule. If a party is a team that seeks control of government, it needs to be very popular to win that under STV.


How liberals ended PR in the US

Proportional representation is a mostly left-wing cause in the US. Some see it as a path to bigger Democratic House delegations. Others see it as a way out of the Democratic Party, period. Much liberal-wing anger centers on the party’s ties to Wall Street. If we had PR, the story goes, the liberal wing would seat its own party. If not, it might at least scare the Clinton wing into moving left. And the affinity between PR and left politics might draw on a myth, neatly summarized below:

Proportional representation systems were tried earlier in the past century and then discarded precisely because they favored minority representation (racial and left wing/socialist) too much.

I’ve found evidence that the most liberal Democrats were actually PR’s worst enemies. Yes, racially and economically liberal. I’m talking about the AFL and/or CIO and Young Democrats. At roughly the same time they were pulling the Democratic Party leftward, they were working to repeal PR in at least three of the cities that had it.

The urban PR “spells” chart, updated

If you are reading this, you know that 24 U.S. cities tried the STV form of proportional representation. Here is an updated chart summarizing those episodes.


The changes include:

1) Cropping to 1965, since events thereafter are basically chartjunk.

2) Adding a failed 1959 repeal attempt for Worcester, MA.

3) Changing the dates for Oak Ridge, TN, which evidence suggests to have emerged with STV a decade later than I thought. Note that the Oak Ridge council was “advisory,” whatever that means.

4) Adding Norris, TN, which I had not included because I could not find the dates of any STV elections. This Town Council was also “advisory.” Elections were held annually from 1937 through 1945. I cannot find evidence of an STV election post-1945. The federal government sold Norris to a developer in 1948, and Tennessee granted Norris its own charter in 1949.

5) Date-stamping, since this chart may evolve gain.

Please share any suggested, further changes. Wouldn’t it be nice if our state and local politics were better documented?

STV is also totally compatible with polarization

On Friday, I showed NOMINATE scores for Cincinnati City Council members elected under permissive rules. Polarization was super low* for some years because, in each of those years, a lone wolf was able to play coalition kingmaker in a hung parliament.**

Now look at 1947, which is new in this plot, and watch those parties diverge. Yup. STV is totally compatible with high levels of polarization. (I never said it wasn’t.)



Strategic behavior in subaltern DC elections?

[Update: The original map had the wrong data in the wrong districts. I used a bad shapefile. The substantive conclusions, insofar as any existed, don’t change. Thanks to @20002ist of 6C04 for catching the error.]

You might not expect low-turnout, low-salience, non-partisan, local elections to have the properties of organized politics. But the 2014 DC Advisory Neighborhood Commission (ANC) races seem like they did. There was more competition than you might expect, and it looks organized.


More on the map later. For now: yellow (uncontested), gray (noncompetitive), green (two-way competitive), dark green (three-way competitive). Heavy lines are ANC boundaries, and thinner lines are ANC single-member districts.