Designing the Perfect Election system

Designing the Best Election System

Hello, Genossinnen und Genossen.

Last week we discussed the truly abysmal state of democracy (here defined as the proximity between people's votes and the actual result measured in legislature seats) in last month's British election. According to Political Scientist Robert Dahl, Britain is one of only four nation-states among the ranks of democracies and semi-democracies to use simple plurality, single member districts as the method for choosing their legislature. (The United States, France, and Canada are the others). To recap our problems with this system, let's briefly recount the offenses committed by First-Past-the-Post (FPTP) single member district systems

1. Huge numbers of votes are wasted,- extra votes for the winner beyond the second place finisher's total+1 are wasted as the range of outcomes does not change with the margin of victory, and no vote that went to someone other than the winner has any effect on the outcome anyway.
2. A Party/ideology's strength is determined as much by geography and gerrymandering as by popularity, which can produce gross over or under - representation as the case may be- see my previous article for examples from the recent British election
3. The system forces out minor parties by confining representation to groups that command local pluralities, rather than a consistent share of the electorate spread across the entire country, i.e. a plurality party in 1 district gets more representation than a party that has 20% support everywhere but no pluralities. This has the following effects
   1. - A subsidy to the largest one or two parties
   2. Marginalizing the expressed views of even sizable minorities
   3. creating a chilling effect on that expression, by making voting for alternatives meaningless except as a protest
   4. - One positive effect- every locality has at least one legislator dedicated to advancing that region's interests.  This is called Dyadic Representation

It is particularly revealing that whenever we've been in position to dictate another country's constitution (Post-war Japan, Germany, Italy, several others) we've spared them the farce that is our voting system, acknowledging that political science has advanced since our constitution was written.

Now let's talk fixes, and more broadly what our ideal election system would look like. For the purposes of this discussion, I'll be ignoring the titanic power of mega-donations as well as geographic imbalances (equal representation for unequal population) and focusing on how votes are cast and counted.

There are three general alternatives to FPTP

1. Ranked or instant runoff voting- you list candidates in order of preference, the candidate with the fewest first place votes is eliminated, with those votes going to their second choices, and so on until only one candidate remains. Decent at avoiding the chilling effect, but does lead to a lot of shenanigans, as you sometimes have to figure out the optimum sequence to get your slate represented- if there are multiple candidates you could support, you want to make sure one of them isn't eliminated early! This is fun to game out but it doesn't really help us with problems 1 or 2, does it?
   
2. List or Proportional Representation Voting. I LOVE this- it works best the larger the constituency voting as one- ideally the entire country would vote in one election for this to work. Each party nominates a list of candidates in preferential order, up to the total number of seats in the legislature/Congress/Parliament. (Parties can nominate these candidates in different ways, commonly through a national-level primary, which is how I would do it). Some more progressive countries like Sweden require quotas of women's or minority representation among candidates on this list, some even mandating that the subsidized demographic be listed every x spaces on the list to make sure that an unpopular demographic isn't confined to the bottom of the ticket, where they're unlikely to prevail in this case.  There may be other qualifiers- a party may require its candidates to be affiliated with a trade union, or have served in the military, or in some cases profess a particular religion.
   
   
   Voters then cast ballots for this list, and the seats are awarded on a proportional basis among all parties that clear a certain threshold. I've heard of thresholds being set from .4% to 20%, though I'd personally recommend pegging it between 5-15%- too high a threshold fails to address problems 1, 2 and 3, and too low a threshold incentivizes the creation of myriad fringe parties which can then extort the government for their support as seen in the Israeli Knesset and in the Weimar-era German Reichstag.
   Done properly, list voting solves problems 1, 2 and 3. Every vote (except those cast for truly marginal parties) matters equally in determining the composition of the legislature, giving politicians incentives to campaign everywhere, as well as being its own reward, and parties are represented regardless of their ability to win a plurality anywhere. There are a couple of problems with a pure list system though- first, we lose advantage 4- Dyadic Representation.  With a national election, we lose the benefit of each local community having a representative to further local aims in the national legislature.
   The other problem isn't unique to list voting, but it is amplified in this case: you have to choose between denying unaffiliated voters the right to choose candidates IN THE NOMINATING PROCESS, NOT THE GENERAL ELECTION, but still, that is a problem, and undercutting party responsibility by letting people who are unaffiliated with a party's ideas vote on who its candidates will be. Between these two problems, I would move for a closed primary for two reasons (There are ways to mitigate the exclusion of independents, one of which is already happening under a PR system*), and my support for party responsibility- the extent to which a party is about ideas rather than personality, which would be undercut by allowing casually informed independents to influence the nominating process.
   *The list system creates openings for far more views to be represented, so many independents will see something they like under a list system

3. Finally,a multi-member district system. (MMD). This can operate the same as FPTP or Ranked voting except that the top x finishers are awarded seats where x>1. This mitigates problem 1 and solves problem 2 while not sacrificing advantage 4 or alienating independents, but leaves 3 unsolved. Also, being a district system, it is still somewhat contingent on accidents of geography. This is definitely my favorite for incentivizing campaigning everywhere because of the myriad scenarios in which parties should be paying attention to each district. Consider a slightly competitive district (say under a system with 3 members per district) that tends towards one party- the second and third largest parties have every reason to fight for whatever seat they can get in the district, because they can “win” without getting a plurality vote. Or consider a district overwhelmingly loyal to one party- that party has an incentive to compete for every vote, because if they nominate multiple candidates, they might walk away with more than one representative. Meanwhile, if they do this, smaller parties still have a decent shot at a win. Of course, this could be very subject to manipulation (It may conceivably be more valuable to deny a certain opponent a single seat than to win three for your own comrades), so I don't recommend doing this on its own. It is the only meaningful progressive voting reform currently adopted by an American state (Vermont) and has improved the situation a bit. You're still subsidizing the largest parties based on location though, but if you must do a district system, this is how to do it.

Everybody got that? Now we're getting a bit more interesting, with the model first developed by West Germany and Allied diplomats in the late 1940s

Let's hear it for

MIXED MEMBER PROPORTIONAL REPRESENTATION!- the best system I've seen anybody actually try.

What is it, you ask? Why is our normally calm, emotionally distant hostess describing it so dramatically, let alone talking about herself in a hypothetical third person?

Well, it is just that damn cool, that's why. Basically, the citizens vote in two elections simultaneously, with separable votes. One for a party list, and one for a candidate in a single member district. The threshold in the German example is 5%, and I like it. There are two twists to make this work- there are more legislative seats than there are districts, and any candidate who wins a district is bumped to the top of her party's preferential list, thus mostly preserving the advantages of a PR system while still letting districts do their thing and get dedicated local representation into the legislature. It's also sexy to independents who may want to ticket split, or even vote for a candidate without a party affiliation.

There is still one big problem though- Problem 1. It is very easy for a larger or sectionally stronger party to win more seats than it is entitled to by proportion of the vote. When this happens, the extra seats aren't taken from anyone else, but they are added in as what are called “overhang” seats, thus subsidizing the biggest party.

I think we can fix this- MMPR mitigates problem 1, solves problems 2 and 3, and leaves advantage 4 completely untouched, while giving independents a new option of choosing a party and a candidate who may not necessarily be on the same side. We can further mitigate problem 1 by taking this to the next level, and making the districts be multi-member! This means that each party has incentive to compete for every vote, everywhere for the list, and for any district where they are polling at close to 3^rd place or better. You have all the advantages of a list system, while dividing the advantage won from the district system among the top three parties in a regionally homogeneous country, or far more in a regionally heterogeneous country. What this means is way, way more votes affect the outcome, which means more constituencies will be represented, both in their ability to actually elect people or parties of their choice, and in the attention paid to them during campaign season. What this leaves us with is a system that represents more people's wishes, gives politicians more reason to seek votes in new areas, and can also preserve party responsibility with closed primaries.

To Recap, a Mixed Multi-Member Proportional Representation system (MMMPR patent pending:P), in which voters get to vote twice- once for their district's representative, and another time for a party list would represent far more ideas and constituencies in government, and make sure everyone's vote truly matters in a way inconceivable under FPTP Single Member. We should all do what we can to get this adopted in as many states as possible- it's even possible (not plausible, sadly) that with the repeal of one federal law from 1960, states could choose to do this for their house delegations too!  But even at the state level, there's so much to do, and as long as we're giving this whole democracy idea a shake, we should make sure as many votes as possible actually matter.

Solidarität, Genossinnen und Genossen