Friday, July 23, 2010

On proportional representation (2): the case against

Introduction

Last week, I argued for the claim that proportional representation (PR) is preferable to single-member plurality (SMP) as a system of electing representatives to a democratic assembly. The central reason for this conclusion was that PR is the default system of elections, given that it is the system which ensures a strong connection between the expressed preferences of the voters and the representatives in the assembly (calling this the "strong connection" thesis in what follows). In this post, I consider nine objections to this conclusion, all of which try to identify important principles which PR allegedly cannot accommodate, but SMP can. What I'm looking for here is whether these principles are sufficiently important to outweigh the principle which supports the conclusion that PR is preferable to SMP.

(1) PR leads to unstable governments/difficulty forming governments

The idea here seems to be that stability is an advantage in assemblies or in governments. It's not clear, though, what's supposed to be so good about stability. I presume it's something like predictability: if the assembly/government is stable, then we can make reasonably good predictions about what they're going to do in any given case. If that's right, then I don't see how this can override the importance of the strong connection thesis. The problem is that if stability is more important than the strong connection thesis, then the system of elections is fundamentally paternalistic. That is, if people make the "wrong" choice -- a choice which leads to instability -- then this choice should be "corrected" by the biases of the system itself. If, however, one has followed me to this point and accepted that the strong connection thesis is important, there is really no basis for the claim that a choice should be corrected in this fashion. This sort of paternalism is inconsistent with giving the strong connection thesis any importance beyond some minimal limit.

If stability is important, then it is less important than the strong connection thesis. So, if stability is to be achieved, it must be achieved by people choosing in such a way that a stable assembly/government results. Anything less would just demand stability -- and, if we're going to go down that road, why have elections at all?

(2) PR allows fringe elements to get elected

This criticism can come from the left, usually against racist or fascist parties, and from the right, usually against socialist or communist parties. The underlying principle seems to be that being mainstream is an advantage, and we want to preserve mainstream candidates and parties in the assembly.

What's good about being mainstream, though? Presumably, being mainstream means that lots of people agree with one's views or ideology. Else, one is fringe. But once we see that, it becomes obvious how wrong-headed this criticism really is. You can't know what's mainstream or fringe until you know what people prefer. It might turn out that a fascist party/candidate or a communist party/candidate has more supporters than had previously been believed. This shows that the mainstream isn't where one might have thought -- or hoped -- it would be.

So, if the judgement about what's mainstream and what's fringe has this content, it actually supports, rather than undermines, PR. PR allows us to see clearly, in the assembly, where the mainstream is and where the fringes are. The judgement might have a different content, though. It might by a covert attempt to privilege conforming to a certain set of attitudes or ideals. That is, the "mainstream" are those who agree with me, and the "fringe" are those who don't. Like (1), though, this is deeply inconsistent with giving the strong connection thesis any importance at all. If it's legitimate to sweep away some preferences because they aren't the "right" ones, then it's hard to see why we would respect any preferences at all.

(3) PR allows fringe elements to dominate governmental negotiations

If the underlying principle here is that negotiations being dominated by the mainstream is good, then this falls prey to the same objections as those raised above. Furthermore, it's obviously a rather silly principle -- it amounts to endorsing mob rule, where the biggest elements should get their way, and the smallest should get nothing.

But, more subtly, the underlying principle may be an expression of the idea that weight of voice in negotiations should be proportional to one's actual support in the legislature. If that's the idea, then this supports PR. It's not clear how we could ensure that voices in negotiations are proportional without first getting the members of the assembly to be proportional to the voices (i.e., expressed preferences) of the voters. If we don't do that, then we may as well follow the Athenians and choose an assembly by lot: the privileging of one set of ideas, through affording them seats not earned through earning votes, is basically arbitrary.

Even more subtly, however, the underlying principle may be an expression of concern that, since smaller parties have less to lose for digging in their heels and insisting on getting everything they want, and bigger parties have more to lose, then smaller parties will necessarily have more sway than is proportional to their size in the assembly. I'm inclined to deny that this is actually a problem, though, rather than a reactionary fear.

These people from the "small parties" have been elected, which means some voters out there prefer them. If bigger parties don't want to have to give in to these smaller parties, then what follows is not that we should gimmick the rules to make it possible for big parties to ignore the small ones, but to insist that the big parties should work their butts off to lure more voters to their side. The fact that smaller parties may be able to bargain hard shows only that the big parties didn't do a very good job in their campaigns. Why should the system protect big parties who can't get themselves enough support?

(4) PR governments deadlock

The underlying principle seems to be that making decisions quickly is a good thing. Well, I agree: it's true making decisions quickly is a good thing. But making the right decisions is better. And making decisions for the people that are actually for the people -- and not for the party, or for the next election, or for the post-political career -- is even better than that. "Getting things done" is not what we in philosophy call an "all things considered" good. That is, it's not a thing that's good, no matter what. It's only good in certain circumstances. For this particular thing -- making decisions quickly -- my sense is that it isn't good unless the decisions are ones that are made for the people, and are the right decisions. (If you don't quite agree with me: long-form census, changing the national anthem, and prorogation.) Besides, there's actually no guarantee that an SMP system doesn't deadlock; the US government is elected under SMP -- the House, the Senate and the President -- and they deadlock all the time. Efficiency isn't really something that's got anything to do with the way we select representatives, but with the powers that the representatives have, and they way they approach their dealings with each other.

(5) PR encourages fragmentation of big-tent parties

It's not at all clear what the objection here is. It might be osmething like community is a good thing. But one hardly needs a big formal system to compel someone to be in community with others -- what one needs is an attitude of openness, negotiation, etc. Big-tent parties tend to function by suppressing disagreements which really should be hashed out, in order to help people reach common ground. So, I'm not at all clear how doing away with big parties is a bad thing.

Furthermore, it's important to note that big parties are a relatively recent invention in at least the Canadian parliamentary tradition. The British parliamentary system, to which we are the heirs, didn't have "parties" in any sense we'd recognize until about the mid-19th century, when the Tories and Whigs coalesced into formal organizations rather than loose collectives. So, Parliament has worked without parties of any kind, never mind big ones -- why can't it work again?

(6) Look at (fill in the blank) -- Israel, Italy, etc.

These comparisons really aren't principled. They usually amount to dishonest drive-by smears. Drive-by because the actual problem with (fill in the blank) is often left unexplained -- we're just supposed to recoil from (fill in the blank) instinctively -- and dishonest because the problems with whatever fills in the blank usually have nothing to do with PR. Take Israel, for example. One odd thing elections to the Israeli Knesset do is set the threshold of votes you need to get to get elected very low -- only 2% of the votes cast. Hence, very small parties can get representatives in. Hence, the Knesset has 12 parties splitting 120 seats. Elections are also conducted such that one voters for a party, and not for an individual. I'm not suggesting either feature is necessarily bad, but they are both unusual. Given that, it's dishonest to pretend that any objection to Israel's system of elections must be because it's a PR system. It differs from the Canadian system in other ways, too.

The same could be said for another usual fill to the blank, Italy. Italy is notoriously corrupt and the bureaucracy is highly inefficient. Those are clearly bad things, but they have little to nothing to do with the system of voting, and don't apply to anything like the same extent in Canada.

(7) PR prevents voters from directly choosing their legislators

The underlying principle here seems to be that directly choosing representatives is a good thing. I do tend to agree; systems which elect members based on a list formulated by the party worry me. But directly choosing reprsentatives is not barred by PR, only by some forms -- namely, those which allot some seats in the assembly to representatives selected from a list compiled by the parties.

Even allowing that, there's still a sense in which such a system -- such as Mixed-Member Proportional, or MMP -- does allow for directly electing representatives. Non-list representatives are directly elected in the same way representatives are directly elected under MMP. And, list representatives are directly elected in the sense that votes were cast which do count towards their election. The only sense in which there is not a "direct election" of the representative is that the individual is not chosen by the voter -- the party chooses. But, let's be realistic here: the party always chooses who you can vote for. (Even in jurisdictions with a primary system, although those do seem to make it slightly more likely that someone outside the central party structure can stand for election.) Having representatives elected off a list just makes the process a little more transparent.

So, unless one is willing to advocate something very radical -- abolition and prohibition of political parties, say, such that all candidates standing for election have not been put there by any party structure -- candidates and thus representatives will always be partly chosen by the party. This holds under SMP just as it does under PR.

(8) The difficulty in removing an entrenched party from power quickly

The underlying principle seems to be that being able to quickly remove a party from power is a good thing. As a criticism, however, this is fatuous nonsense. If a party's support collapses precipitously, they can be removed under PR as easily as under SMP -- lost support still equals lost seats. Indeed, it's probably easier, as more votes make a difference under PR. Under SMP, support can drop hugely but, if the geography is favourable and the vote against the party in powersplits, that party can still hang on to lots of seats -- possibly even a majority in the assembly.

To my eye, the objection seems to really be motivated by the desire to "throw the bums out" -- and to read the sentiment "throw the bums out" into the voters as a whole -- when, really, there's just a small tick in general support for the party. Under SMP, small ticks can make a huge difference. In the '93 election, Jean Chr├ętien gained 10% in support and 94 seats, amounting to a 113% increase in seats, while Kim Campbell's PCs lost 15% of the vote and 167 seats, amounting to a 98% drop in seats. That a 10% increase equals a 113% increase, and a 15% drop equals a 98% drop. There was no real "throw the bums out" sentiment -- more of a "we're not terribly pleased with these bums, overall, but would like to keep some of them around nonetheless". It just didn't exist. And yet, the results in terms of seats suggest otherwise.

If the bums deserve to be thrown out, they need to drastically decline in support. Just because you want the guys in your colour shirts to win is no reason to to allow the assembly's composition to swing around as madly as this.

By the way: wasn't one of the alleged advantages of SMP that it would produce stable assemblies...?

(9) Difficulties in understanding a more complex voting system

The underlying principle seems to be that it's better to have a system that's simpler, so that voters can understand it.

I'm not sure that it's true that voters can't understand a PR system. We are, after all, talking about people who can grasp the complexities of, say, FIFA's Laws of the Game, or complex statisical calculations regarding the likelihood of a hockey goalie's making a save. So, it's not that people can't grasp complex systems. It's more that they may not care to grasp the system of elections. But, to my eye, this suggests that the the problem isn't that PR is too complicated, but that SMP is discourages people from caring about voting as such. After all, if you know your vote might not make a damn bit of difference, why would you bother learning how the system works? So, I suspect that a system which made an individual vote more valuable would actually encourage people to know how the system works -- so they could use their now more valuable vote most effectively.

(10) SMP is old

People do actually make this argument. They usually refer to the grand British parliamentary tradition, or some such, but this is what it amounts to. Do I really need to explain the difference between being old and being good?

Oh, all right: creationism, vitalism, caloric fluid, and phlogiston. All old. All wrong.

Conclusions

Overall, I'm not persuaded that there is any goal sufficiently more important than matching the expressed preferences of voters which would justify retaining SMP over PR. If there is such a viable possibility, I don't know what it is. Certainly the ones I've canvassed here are pretty dismal failures, turning on either rejecting the value of voting as such, failing to actually show SMP is superior to PR, or being simply ridiculous.

Next week: Filling in some of the holes here. Why is PR better than a preferential system (e.g., instant run-off voting).

2 comments:

Wilf Day said...

Excellent reasoning and summary.

One footnote: you skip over the argument about thresholds. Israel's is "very low" as you say -- only 2%. But why do we need any threshold at all? Most of your arguments rebutting the criticisms of PR could also rebut the need for a threshold. Germany's Weimar Republic had no threshold. The Netherlands still doesn't.

The answer is practical: A coalition of six parties tends to be unworkable. In Weimar Germany the major centre-left parties formed what became known as the "Weimar Coalition" -- the only government in sight. No coherent democratic opposition alternative. But no coalition can govern forever. With similar fragmentation in Canada in 1935, Mackenzie King's winning slogan was "It's King or chaos." But when Weimar fell apart, they got Hitler. Consider the 2006 Netherlands election: 10 parties. Many voters, bored with perpetual centrist governments, had swung to the Socialist Party to try to get a frankly left-wing coalition. Labour, the Socialists and the Green Left had 65 seats, but a majority is 76. Add two small progressive parties (Democrats '66 and the Christian Union) and you have 74. You also need the two young women from the brand-new Animal Rights Party. But all recent Dutch cabinets have consisted of only two or three parties. They could not contemplate six. So they got, once again, the centrist coalition of the centre-left and centre-right.

Which is why the British political scientists helping the Germans in the British zone insert denazification provisions in the new voting system in 1946 -- the ones who invented the MMP model --recommended a high (5%) threshold. Most other European countries have either 5% or 4%. PR produces a chaos of splinter parties? Well, yes, it can, without a threshold.

ADHR said...

I see the concern, but I'm not convinced that the cases justify inserting a threshold. For example, Hitler wasn't able to achieve power just because of issues in the voting system.

The problem I have with thresholds is that the introduction of such a thing works out to a form of gerrymandering: trying to exclude "fringe" parties from the possibility of playing a role in government by playing with the machinery of the voting system. (Which is, in effect, what SMP does.) It also seems to excuse "mainstream" parties from trying to appeal to a broad range of voters -- something they surely have to do to actually be mainstream.

And that's beside the problems associated with where the threshold is set. The number always seems rather arbitrary. Why 4% rather than 4.5%? Or 4.66%? Or a fixed number of votes rather than a proportion? And so on.

Basically, if the voters want a chaos of splinter parties, as far as I'm concerned, that's what they should get. And if representatives can't work together like adults, then voters should have the sense to replace them. (FWIW, I do support recall legislation in order to remove reps in between elections.)