Why proportionality matters

Prime Minister Birgitte Nyborg of the Moderate Party meets with representatives of her Labour and Green Party coalition partners in the Danish political drama Borgen

Supporters of proportional representation claim that first past the post is defective because it allows gross disproportionality between the popular vote and the distribution of seats in the legislature. As I discussed in my last post, defenders of FPTP may reply that under FPTP there is no such thing as the popular vote, because FPTP only allows for the tabulation of votes by candidate. What supporters of PR mean by “popular vote”, however, is just the share of all votes cast for candidates from each party. The existence of those numbers is undeniable. Of course those numbers are not tabulated for the purpose of determining winners, but that’s precisely why supporters of PR object to FPTP.

Defenders of FPTP may respond that the composition of the popular vote does not provide a straightforward accounting of people’s considered preferences. This is partly because many people are so-called strategic voters. But even the choices of “sincere” voters are shaped by the political environment of which the electoral system is a part. PR supporters assume that the popular vote is a closer approximation of people’s considered preferences than the resulting distribution of seats. But this assumption is unwarranted. Change in the electoral system is likely to cause change in the popular vote. And if the popular vote changes because the rules for translating votes into seats change, this suggests that the popular vote has no special claim to be the authentic expression of the people’s collective will. And if the popular vote lacks this special claim, then it is unclear why we should find disproportionality especially troubling.

In my view, this defence misconstrues the significance that supporters of PR attribute to the popular vote under FPTP. In order for the defence to work, PR supporters must be pointing to the discrepancy between the popular vote and the distribution of seats because they regard the popular vote as democratically authoritative; because the distribution of seats fails to reflect the democratically authoritative popular vote, the mechanism for translating seats into votes must be defective. But PR supporters are not committed to the view that the popular vote alone is democratically authoritative; they are only committed to the view that disproportionality is democratically defective. The challenge for the supporter of PR, then, is to supply grounds for thinking that disproportionality is a defect without attributing special democratic significance to the popular vote.

One possibility is that under any electoral system, both the popular vote and the resulting distribution of seats are democratically authoritative. If an electoral system does not take the popular vote into account in translating votes into seats, it fails to respect the popular vote’s equal democratic authority. FPTP is such an electoral system; thus, it is democratically defective, By definition, only a proportional electoral system can respect both expressions of authority. So if more democracy is better, all things being equal, then proportional representation is preferable to FPTP.

Another reason why disproportionality might be a defect is that it makes cooperation between voters much more difficult. This feature of FPTP is sometimes obscured by critics’ fondness for the argument that FPTP encourages strategic voting. The irony is that although FPTP may very well encourage strategic voting, it also makes “strategic voting” an almost useless strategy. Voters simply do not have sufficiently reliable information about each other’s intentions at the riding level (the only level that counts under FPTP) for strategic voting campaigns to work. However, voters do have sufficiently reliable information about each other’s intentions at regional, provincial or national levels for more limited kinds of strategic voting to succeed — provided that aggregate votes at the regional, provincial or national levels are electorally significant (as they are under PR). This means that under PR, people may vote strategically to ensure a breakthrough for a minor party, boost the standing of a preferred party’s prospective coalition partner, or block an especially disliked party from gaining ground. If the possibility of this kind of cooperation between voters translates into more popular self-government rather than less, then PR is again preferable to FPTP.

Finally, disproportionality might be a defect because it gives voters fewer choices. Defenders of FPTP have argued that this can be a virtue, because disproportionality encourages the formation of ideologically moderate and inclusive big-tent parties. But supporters of PR may respond that the need to form multiparty coalitions in order to win power under PR encourages ideologically moderate and inclusive governments, mitigating the advantage claimed by FPTP. If PR and FPTP both encourage moderate, inclusive coalitions (albeit by different mechanisms), then the fact that PR offers voters more choices can act as a tie-breaker.

At this point, FPTP’s defenders can ask why more voter choice is all that important if it just leads to the same kind of compromises we see in the big-tent parties under FPTP. One reason is that more voter choice means more, smaller political parties. Smaller political parties are likely to be more internally democratic. More internally democratic political parties operating under the discipline imposed by the necessity of interparty coalition-building create more opportunities for meaningful political participation. Increased opportunities for meaningful political participation will improve the quality of our democracy. So, if PR increases voter choice, PR is likely to improve the quality of our democracy.

So as it turns out, one can make a case for proportional representation without mistakenly assuming that the composition of the popular vote is unaffected by the electoral system or that the popular vote is a uniquely authentic expression of the will of the electorate. Concern for proportionality is justified as long as the popular vote has some normative significance. It is further supported by other considerations such as the value of a wider array of choices for voters and a better climate for informed decision-making. And although they are not conclusive, these reasons are at least sufficient to establish a presumption in favour of proportionality. It is therefore incumbent upon defenders of the status quo to provide equally compelling reasons to prefer a system that allows such extreme disproportionality as FPTP.

Why not proportional representation?

“Proportional representation” (PR) refers to a family of electoral systems for parliamentary elections that tend to result in each party receiving a share of seats that at least roughly matches the proportion of all votes cast for that party’s candidates. Proponents of PR tend to take it for granted that proportionality is a desirable goal, and point to the disproportionality of election results under the system currently used for provincial and federal elections in Canada (usually called “first past the post” or FPTP for short). Defenders of the status quo, on the other hand, tend to deny that there is anything objectionable about the lack of proportionality under FPTP.

Defences of FPTP against the charge of disproportionality can be divided into two types. According to the first type, disproportionality is a virtue of FPTP. On this view, disproportionality creates an incentive to build large, inclusive and ideologically moderate parties. Furthermore, the system’s tendency to result in single-party majority governments improves accountability, while its tendency to translate small swings in the party vote into large shifts in party standings makes it easier for voters to get rid of governing parties they don’t like.

The second type of defence — my focus in this post — is simply to deny the salience of disproportionality altogether. At its most extreme, this kind of argument denies that proportionality is conceptually coherent in the context of FPTP. A more moderate approach is to accept the concept of disproportionality but deny its relevance to the evaluation of FPTP.

An example of this second type of defence was recently provided by Brian Marlatt, a former candidate for the Progressive Canadian Party. Marlatt points out that general elections under FPTP take the form of a large number of local elections. Each of these local elections chooses a single winner according to a plurality formula (i.e. the candidate with the most votes wins); there is no separate national election in which votes are cast for parties. The share of the vote captured by each party nationally may provide an interesting piece of trivia, but there is no unfairness in disregarding numbers to which the system does not purport to assign any importance whatsoever.

A similar argument is given by political journalist Dale Smith in his book The Unbroken Machine. Smith argues that critics of FPTP simply do not understand how it works. If they did, they would see that there is no problem with disproportionality between parties’ shares of seats and votes. Under FPTP, elections are about the voters in a riding choosing one representative to send to the capital; they are not about the overall composition of the legislature. Pointing to the disproportionality of election results under FPTP only reveals critics’ ignorance.

One can argue that there is an inherent illogic in the way we look at proportional representation (PR) schemes. For one, it seeks to apply a result that doesn’t reflect the reality of the situation. In the case of a federal election, what is being voted on is who will represent each constituency in the House of Commons. No matter that people might feel they’re voting for the party, or the leader, or whatever other consideration they might have in mind, the ballot they are casting is for the representative in that seat in the House of Commons. Saying that seats need to be apportioned otherwise would indicate that the votes being cast were to determine the composition of the House of Commons in an overall cohesive manner rather than to determine its population of individual MPs. (p. 54)

The defences given by Marlatt and Smith are very puzzling; faced with criticism of FPTP, they simply reply that elections in Canada operate according to the principles of FPTP. This may be a compelling argument against the charge that FPTP fails on its own terms, by promising but not delivering proportional results. But is this the essence of the case for PR? Obviously not. By taking the desirability of proportionality for granted, advocates of PR sometimes seem to assume that FPTP shares the goal of proportionality and is just very bad at delivering it. But this is not an essential feature of the pro-PR position. The only thing that’s essential to the pro-PR position is the belief that representation ought to be proportional. And if a person thinks that representation ought to be proportional, there is no reason why learning that FPTP does not aim at proportionality should make them any more likely to support the status quo. For Marlatt and Smith’s arguments to work, they must provide reasons independent of the constitutive rules of FPTP why a good electoral system needn’t be proportional.

Instead of defending disproportionality directly, both Marlatt and Smith take aim at the measures needed to ensure proportionality. Both writers share the concern that PR systems give too much power to political parties, which are not directly accountable to voters, at the expense of individual politicians, which are. Moreover, PR may increase the centralization of power within parties. Smith worries that under a mixed-member proportional system (MMP, the kind of PR most often prescribed by Canadian reformers), party leaders will have a free hand in selecting people to fill the party list seats, weakening the power of the rank and file in the riding associations.

It’s true that an MMP system could be designed with these flaws. But as Smith makes clear in his own call for reform of the status quo, so can FPTP (an issue I discussed previously here). And these flaws are not essential to MMP; an MMP system could be designed to avoid them (by means of regional list nominations and open list voting, for example). In other words, Smith is comparing the worst possible implementation of MMP with the best possible implementation of FPTP. It should be little surprise — and of even less interest — if MMP comes up short in this kind of comparison. At most, these criticisms can only serve as a caution against viewing PR as a political panacea. They are not effective against the more reasonable view that that PR is simply likely to be an improvement.