What’s wrong with Popper’s case against PR? Everything.

Gavin_Hamilton_-_the_oath_of_brutusThe Oath of Brutus – painting by Gavin Hamilton

Last week, the National Post ran an op-ed by Peter Shawn Taylor explaining the celebrated philosopher Karl Popper’s case against proportional representation. I didn’t find Taylor’s piece convincing, but I was sufficiently intrigued to take a look at the argument in Popper’s own words, in the essay “The Open Society and Its Enemies Revisited”. I’m not sure I should have bothered; Popper himself was even less persuasive. His argument is fatally compromised by contradictions and factual errors.

To give a little background, Popper is today best known for his work in philosophy of science, and in particular for his theory (“falsificationism”) that science proceeds exclusively by the deductive process of falsification. According to this view, all scientific theories can be categorized as either falsified or as possibly true, but never as confirmed or even probably true. While Popper’s falsificationism has few (if any) defenders in contemporary philosophy of science, his ideas have had a profound impact on the field, and the central idea is a staple of basic science education. But although this is undoubtedly Popper’s most widely known philosophical legacy, he also made noteworthy contributions to other areas of the discipline, including political philosophy.

Popper’s political philosophy, as Taylor describes it, treats political questions similarly to scientific questions in the sense that democracy isn’t (or shouldn’t be) about installing good rulers, it should be about removing bad ones. “The new problem, as distinct from the old ‘Who should rule?’,” Popper wrote, “can be formulated as follows: how is the state to be constituted so that bad rulers can be got rid of without bloodshed, without violence?” And Popper’s answer? By adopting “the principle that the government can be dismissed by a majority vote.” This principle, he claims, has been adopted by all modern democracies.

Whoa, hold on a second. There are many democracies where that principle has not been adopted—including Canada and its provinces. The federal and provincial governments in Canada aren’t elected or dismissed by voters at all. Perhaps Popper means that a governing party can be removed from power if a majority of voters vote against it? If so, he’s still wrong; governing parties rarely win a majority of votes.

One might reasonably object that not voting for the governing party isn’t the same as voting against it. But this objection isn’t available to Popper, because as we’ll see shortly, his case against PR assumes that those things are the same. If not voting for a party is the same as voting against another party, then governments in Canada are normally formed by or composed of parties that a majority of voters have voted against. This means that the design of Canadian democracy already violates the Popperian principle that a majority vote should be sufficient to result in the termination of a government (even if only indirectly by the loss of its parliamentary majority).

Let’s suppose, however, that this principle has been adopted by all democracies, as Popper claims. Well, some democracies use proportional representation. It follows that there are democracies that use proportional representation and have adopted the principle that a majority vote should be sufficient to dismiss a government. But Popper’s argument against proportional representation turns on the claim that PR is incompatible with the principle that a majority vote should be sufficient to terminate a government. This is a giant, double-barrelled contradiction: Popper must simultaneously claim that PR is and isn’t compatible with his majoritarian* principle, or he must simultaneously claim that all democracies observe the majoritarian principle and that some do not.

Perhaps Popper can avoid the charge of contradiction by denying the claim that some democracies use proportional representation, at least as he uses the term. That wouldn’t be too far off, because when Popper gets around to explaining PR, he complains that PR allows voters to choose only parties rather than people to represent them. This is true only under a system of closed party lists in a single national district, and as far as I know, only one country in the world uses that system (Israel). If that’s what Popper means by PR, then his argument might not be self-contradictory, but it also wouldn’t be relevant to the electoral reform debate in Canada, where closed lists in a single national or provincial district aren’t under consideration.

Popper’s argument could only have any relevance for us if he intended it to apply to PR more broadly, including systems like MMP. So let’s assume that was his intention. How, in his view, would PR violate the majoritarian principle? Popper’s answer is that a party that has just been rebuked by voters can remain in power by assembling a working majority by calling on the support of another, even less popular party to form a working majority. This violates the majoritarian principle; a party that receives less that a majority of votes is (according to Popper) a party that is opposed by a majority of votes. And if a party that is opposed by (for example) 60% of voters (i.e. a party that receives 40% of the vote) is able to arrange the support of a party opposed by 80% of voters, it does not thereby acquire majoritarian legitimacy even though the alliance would be sufficient for a parliamentary majority.

As noted above, however, our electoral system routinely and systematically generates violations of Popper’s majoritarian principle. In fact, some supporters of PR object to the status quo on precisely these grounds. Where they differ from Popper is on the idea that a vote for a given party is a vote against all other parties. If a party that received 40% of the vote forms a coalition with a party that received 20% of the vote, the result is a coalition that is collectively supported by 60% of voters—a decisive majority.

This response seems a bit simplistic to me. After all, if I like apples in my fruit salad and you like pears, it doesn’t follow that we will both enjoy fruit salad with apples and pears. So it’s true that formation of a coalition may very well end up angering or alienating supporters of the participating parties. But this is true of compromise in politics more generally, and so by itself it is no reason to worry about government by coalition.

In any case, the point that no electoral system on offer registers opposition to this or that party or candidate is valid. Election results only allow for the determinate registration of support for this or that party. A coalition of parties representing 60% of the vote may be preferred by anywhere between 100% to 0% of voters. All the election result allows us to say for sure is that the coalition is composed of parties supported by 60% of voters. Although this doesn’t satisfy Popper’s majoritarian criterion, it does satisfy a majoritarian criterion. And in a parliamentary democracy with an electoral system that only allows voters to register support and not opposition, it is the only majoritarian criterion that can feasibly be applied. PR would satisfy this criterion, whereas the status quo does not. So Popper had a point about the importance of setting up the rules of democracy so that governments can be removed when they lose majority support. But on further analysis, this point turns out to imply that we ought to adopt PR rather than stick with the status quo—precisely the opposite of the conclusion Popper was trying to reach.

*Note on terminology: Confusingly, “majoritarian” is sometimes used by political scientists to refer to a family of electoral systems that tend to produce legislative majorities—even when (as in the case of first past the post) those majorities can be achieved without a majority of votes. In this post, I’m not using majoritarian in this technical sense.

Some thoughts on the election in Vancouver

Vancouver City Hall photographed by Donn B. A. Williams

After I got home from the Cambie Report‘s election party, I jotted down a few thoughts on the results in Vancouver, which I’m re-posting here for the record and to maybe spark some discussion.

1. Some (including me) speculated that this would be the year of the independent in Vancouver on account of a) the number of very strong independent candidates running, b) established parties like the NPA and Vision seeming to fall into disarray, and c) the proliferation of small parties with virtually no name recognition. In retrospect it should have been obvious that the sheer number of contenders for council (exacerbated by the random ballot order) would lead voters to rely more heavily on party label heuristics and shut out the independent council candidates altogether (along with the many upstart parties).

2. Given the proven value of controlling established party labels, I expect there will be more internal competition over control of the established party labels in the period between elections going forward.

3. Going into the election my view of Vision was that it had developed into a pure Gregor Robertson vehicle, and I expected the party would do very poorly when he announced he wasn’t running for re-election. But they had some strong candidates and I thought the Vision label might still count for something, so I’m surprised by their near-total annihilation. They’re done, forever. I expect Allan Wong will find another party or end up sitting as an independent before his term is up.

4. The ballot order effect (again, especially pronounced because of the number of candidates in contention) seems to have been strong. For example, why did Boyle do so much better than Yan? Why did Reddy take the school board seat for OneCity, and not the incumbent, Bercic? Why did Parrot, whose personal campaign seems (to me) to have been nonexistent, take the school board seat for COPE instead of Day, who did run a noticeable personal campaign? It seems the winners in all these examples benefited from a higher ballot position. Results like these serve to underline what we already knew: randomizing the ballot order isn’t much help in neutralizing the ballot order effect if all the ballots have the same random order! Numerous municipalities have demonstrated that they’re on board with the principle of randomized ballot; to really tackle the problem that the randomized ballot was meant to address, though, they need rotating randomized ballots. This next step falls within the jurisdiction of the province, not the municipalities.

To clarify, I’m not saying that the winners mentioned above are in any way less fit for office or less deserving of victory. My view is that the number of fit and deserving candidates in this campaign far exceeded the number of offices to be filled. My point is that in some cases, ballot order seems to have made a difference in the outcome for these candidates—sometimes a decisive difference.

5. The housing crisis was supposed to be the big issue in this election, and it was, but it seems like voters collectively opted not to do much about it—not with the tools available to municipal governments, anyway. One interpretation would be that people are basically fine with the housing situation after all, but I don’t think that’s likely. I think it’s more likely that people are (rightly or wrongly) looking for solutions led by the provincial and federal governments before endorsing bold action at the municipal level. This implies that housing will be a bigger issue for the region’s voters in the upcoming federal election than it would have been if the municipal election had turned out differently.

6. Although Stewart’s victory was extremely narrow (he beat Sim by 0.57% of the vote), the combined vote for inclusive urbanist mayoral candidates was overwhelming and no party left standing at the end of the campaign can afford to ignore this in their conduct on council in the next few years.

What the feds can do right now about the notwithstanding clause

Screen Shot 2018-10-15 at 4.50.56 PMThe only way to be sure

In a previous post, I argued that unlike reservation (which allows a lieutenant governor to submit a provincial bill to the federal government for approval before granting royal assent), the federal power of disallowance (which allows the federal government to annul a provincial law within one year of the bill receiving royal assent) is not forbidden by convention. Briefly stated, the argument was as follows:

1. If disallowance was forbidden by convention in 1992, it would have been struck from the constitution in the Charlottetown Accord.
2. The Charlottetown Accord did not strike disallowance from the constitution.*
3. Disallowance was not forbidden by convention in 1992. (from 1 & 2)
4. Nothing has happened between 1992 and 2018 to indicate that a new convention has been created that would forbid the power of disallowance from being exercised.
5. There is currently no convention forbidding disallowance. (from 3 & 4)

My purpose in making this argument was simply to point out that a bit of conventional wisdom (no pun intended) on a relatively obscure part of the constitution might be mistaken. Even if convention permits disallowance, it seems as though currently prevailing political attitudes do not. Those attitudes might change in the future, just as they have in the past. But for the foreseeable future, I argued, it seems very likely that the power of disallowance will be kept behind glass.

I’ve changed my mind. The federal government should break the glass right now and declare that it will disallow any provincial law that uses the notwithstanding clause.

Following elections in Ontario and Quebec, most Canadians live under provincial governments that are prepared to use the notwithstanding clause to override constitutionally protected rights and freedoms. While it’s true that the notwithstanding clause is as much a part of the constitution as the rights and freedoms it is used to override, many Canadians have assumed that it too is to be kept behind glass. If the notwithstanding clause is to be used at all, it is certainly not to be used frivolously. The actions of the Ford and Legault governments, however, show that governments are in fact prepared to use the notwithstanding clause frivolously. To the extent that support for the notwithstanding clause remaining on the books was conditional on the understanding that it would only be used with care, that support should now be withdrawn, and any use of the notwithstanding clause should be vigorously opposed.

One way to oppose the use of the notwithstanding clause would be to withdraw one’s support from politicians that use the notwithstanding clause, and transfer one’s support to politicians that also oppose its use. Another way would be to amend the constitution so as to take this tool out of the government toolbox altogether. Unfortunately, neither of these options can deliver relief in the short term. To deliver relief in the short term, the federal government should adopt a uniform policy of disallowing any law incorporating the notwithstanding clause.

Saying that the notwithstanding clause ought to be repealed is equivalent to saying that it ought never to be used. If the notwithstanding clause ought never to be used, then there is no reason to tolerate its use at any time. Disallowance gives the federal government the ability to prevent the notwithstanding clause’s use. So if the notwithstanding clause ought to be repealed, then any use of the notwithstanding clause should be disallowed while we work towards striking it from the constitution altogether.

*Updated on November 28 2018: To clarify, the Charlottetown Accord would have removed mention of disallowance from the constitution. But the text it substituted for the provisions on disallowance provided for a de facto attenuated form of disallowance: the governor-general (acting on advice of the prime minister) could direct a lieutenant governor to withhold assent from a bill, thus preventing it from becoming law.