Fri 19 May 2017 4:49PM

Sightline’s Guide to Methods for Electing Legislative Bodies

AZ Adam Zielinski Public Seen by 399

New article by Kristin Eberhard on multi-winner proportional representation systems.



Adam Zielinski Fri 19 May 2017 4:51PM

It’s funny how with single winner elections for an executive position, one of the big concerns is negative campaigning, and the virtue of candidates trying to appeal to a wider range of voters.

But then with multi winner, proportional representation elections, this concern not only goes out the window but gets flipped 180 degrees on its head, with the desire to ensure representation of ideological minorities.

The assumption becomes that only minorities will vote for minorities, and those in the majority will never vote for minorities, and candidates and parties trying to appeal to a broader range of voters and avoid negative campaigning is a pipe dream or non issue in multi winner proportional representation systems.

in reality these issues are two sides of the same coin. If you want to have a voting system that encourages candidates to appeal to a wider range of voters than its narrow base, then there is a trade off in terms of not being able to strictly guarantee it will be able to win a seat on just the strength of it’s own narrow base.

If you want to have a system that guarantees that ideological minority candidates can win a seat with only support from their narrow base supporters, then you need a voting system that does not encourage or require candidates to appeal to a broader audience.

Multi-Winner Score Runoff and Re-weighted Range Voting offer a better balance between these two competing values. Candidates are encouraged to at least attempt to appeal to a broader range of voters, otherwise they might not win a seat. However if they do so then they are very likely to win a seat and the system is very likely to be proportional, certainly much more so than majoritarian voting.


Jeremy Macaluso Fri 19 May 2017 11:27PM

The concern doesn't go out the window, it just mostly gets solved by allowing multiple winners. The goal is to have as many people be represented as possible. With single-winner elections, this means that the single winner has to appeal to as many people as possible. With multi-winner elections, different people can be represented by different winners, so individual candidates don't need to appeal to everyone. The goal is still to have as many people be represented as possible, but the majority generally does not need to worry about being represented, while a minority does.


Clay Shentrup Sat 20 May 2017 3:58AM

The theme throughout is: homogenous legislatures including only, say, white men with a narrow range of political ideologies or life experiences, produce poor results for a diverse electorate, while diverse legislatures, including people with many different life experiences and political perspectives, produce better results.

This is not an empirical scientific statement, but an intuitive feel-good statement. While it's true that diversity of opinions improves information, it also can create gridlock and inefficiency. So when you're analyzing something so complex and important as an electoral system, you want to use data, not intuition.

Warren Smith, objectively a genius Princeton math PhD who has truly revolutionized the field of voting, including proportional methods, did such an analysis by citing the preeminent experts both for and against PR. While he came out slightly in favor of PR, it was a well informed analysis with important caveats. It is also important to note that the advantage of PR over Plurality Voting may be more about Plurality Voting's specific flaws than about the distinction between PR and single-winner methods. We don't have such an analysis on PR vs. Score/Approval/Condorcet/etc., so we really don't know.

The United States and Canada illustrate the problem: both countries are racially, ethnically, culturally, and economically diverse. Yet elected officials at the federal, state, and provincial levels are disproportionately wealthy white men.

Calling non-PR-ness a "problem" is a problem. You haven't definitively shown it leads to less representative legislation.

One experiment you could do is public opinion polling on legislation passed by the Australian Senate and House, to see which produces more popular legislation. You could be surprised to see that the House (IRV) does better than the Senate (STV/PR). Maybe not, but the point is you cannot just take a wild guess and present that as a fact.

If a computer drew compact districts, but one district encompassed an urban area that was 71 percent Democrat, and the surrounding four suburban districts were each 51 percent Republican, the map would look well-proportioned and logical, but it would not be very “small d” democratic.

If this happened to a handful of districts out of scores or hundreds, I doubt it would be much of a problem. In fact, this doesn't even really make an argument that it's particularly undemocratic.

There is no fair way to draw single-member district lines. Single-winner districts prioritize geography above all else

Non-sequitur. This assumes the definition of "fair" is "proportional", which is debatable to say the least.

Clearly, proportional election methods win.

Very little direct evidence was cited for this claim. The anti-Gerrymandering properties of PR were far and away the strongest.

the best systems in Cascadia are those that provide diverse representation

The article provides no evidence of this.

Cumulative Voting already has a track record in dozens of American jurisdictions

Cumulative Voting is very nearly pointless. Money quote:

The trouble with cumulative voting is that, strategically speaking, it is the same thing as plurality voting. A voter is strategically foolish – "wastes" part of her vote – if she gives anything less than her full point-count to just one candidate.

Majoritarian methods have problems; Proportional methods have solutions

The bias here is cringe inducing. People turn to think tanks like Sightline to hear nuanced unbiased information that tells all sides of the story. There are incredibly well informed people who've studied this subject for decades and don't support PR, for a host of evidence-based reasons. It is just not this cut and dried.

Cities and counties could introduce Reweighted Range Voting or Multi-Winner Score Runoff Voting and potentially achieve proportional representation.

The word "potentially" is unwarranted here. RRV actually satisfies a proportionality theorem and is arguably more proportional than any other method in use in the world today.

Party List Voting is the most proportional of election methods.

As I explained to Kristin via email, this is simply unfounded. RRV can be much more proportional, because it recognizes that voters are not perfectly slotted into a single party. If you vote Dem=5, Green=4, GOP=0, that's more information than just casting a vote for Dem. A simple party list vote captures less information, so there's more "rounding error".

Proportional election methods lead to better representation

Little to no evidence provided for this. Just the assumption than diverse legislatures are better.

Again, these Sightline pieces are fine for getting lay readers interested in these issues. But they lack deep expertise and in some ways confuse as much as they illuminate.

If you want to understand voting systems, you will be much better served by reading about the issue from election experts like Warren Smith, the Center for Election Science, the various PR experts cited in Warren Smith's analysis, etc.


Clay Shentrup Sun 21 May 2017 4:21AM

Clearly, proportional election methods win.

Peter Hain wrote this book:

Proportional Misrepresentation: The Case Against PR in Britain

About the author
Peter Hain was Labour MP for Neath from 1991 to 2015 and held a number of senior posts in Tony Blair and Gordon Brown's governments. After his South African family was forced into exile in 1966, he became a leader of the Anti-Apartheid Movement in Britain, also obtaining degrees at Queen Mary College, London, and Sussex University. Over nearly 50 years in UK politics Hain is known for favouring candour over caution, and plain speaking over political spin. He has published twenty books including his memoir, Outside In as well as pamphlets and media articles, appearing widely on radio and television.

Now here's someone whose spent literally on the order of 1000 times more hours of his life studying PR, and strongly disagrees with Kristin's assertion. So no, it is not "clearly" the case that PR wins.

Of course Hain could be wrong. But making that case requires evidence, something the Sightline piece almost entirely lacked. It would be one thing if Kristin had read his book and then responded to his specific arguments. But you cannot just cite basic facts that Hain and other PR detractors were well aware of, and argue that they "clearly" lead to a different conclusion. No, it is not clear. It's super complicated. Your job is to acknowledge that complexity and cut through it, leaving the readers with some unbiased perspective.


Jeremy Macaluso Sun 21 May 2017 7:31PM

I read Smith's article which includes a summary of Hain's critiques. Most of them seem to be or stem from the lack of accountability that many proportional systems have. I know that proportionality has been defined mathematically. Is there an accepted mathematical definition of accountability for multi-winner elections? If not, do you know of any articles that describe previous attempts to define it mathematically?


William WAUGH Mon 26 Jun 2017 11:21PM

I suggest that the form of proportional representation that is the most proportional is “Choice of Representation”. Each citizen is accorded a right to choose her representative. Each representative votes on legislative questions with the proxies of those who chose that representative. If there isn’t room in the legislative chamber to seat all the representatives, the less powerful ones might have to vote by wire from home or pass on proxy to someone having a seat. Localities could choose the variant most appropriate for their respective situations. https://www.facebook.com/profile.php?id=195236817222399&ref=br_rs