Paul's Internet Landfill/ lj/ MMP 102: Coalitions vs. Minority Governments

MMP 102: Coalitions vs. Minority Governments

One of the most frequent criticisms of proportional representation is that it leads to unstable "permanent minority" governments:

Which parties get those seats [list seats] will be based on proportional party vote, a system which could eliminate the winning party's margin of victory. This, of course, would lead to more minority governments, so they would have to resolve the confidence vote issue or hold elections every year and a half or so.

--Robert White, "Reforms would undermine our democracy", Letter of the Day, The Record, April 23 2007.

As no party in Ontario has obtained more than 50 per cent of the vote in a provincial election since 1937, MMP would just about guarantee minority governments.

-- Ian Urquhart, "Radical voting proposal gains steam", Toronto Star, Feb 22 2007.

These criticisms contain grains of truth. It is true that no party in Ontario has earned a legitimate majority government since 1937, and it is true that under MMP most "phony majorities" would disappear. But there is trickery in this criticism, because it confuses the idea of minority governments under first-past-the-post (FPTP) with the coalition governments that are the norm under proportional systems.

Minority governments are very different beasts from coalitions. In a minority government, a single party attempts to govern despite having fewer seats than is strictly required to force through legislation. Thus the government has to do a lot of ad-hoc cajoling with opposition parties to pass legislation. Since many votes are "confidence votes", this puts the governing party at great risk, because if a confidence vote fails the government falls. That's the story, anyways. In practice, it appears that the greatest threat to minority governments might be the ruling party itself -- according to (Dobell 2000) six of nine minority governments from Confederation to Joe Clark's government fell because the ruling party engineered a non-confidence vote. (*) Paul Martin's government's fall was engineered by the opposition, so we are looking at 6 out of 10 governments dying by the ruling party's hand.

Why would a ruling party engineer the fall of its own government? The answer is the same reason that an opposition party would make the government fall: our first-past-the-post voting system. FPTP tends to give huge seat bonuses to parties with a plurality of the popular vote -- 40% of the vote usually translates to 60% of the seats. Accordingly, a small shift in the popular vote can vault an opposition party to majority status, or could upgrade a minority government to a majority. In FPTP minority governments, there is huge incentive for parties to pull the trigger and force elections as soon as they think that they can win the election. Elections are risky operations (public opinion gets really volatile during campaigns) but with FPTP there is a promise of a huge payoff.

Contrast this to governments in proportional systems. In a proportional system, a five percent shift in voter opinion translates to (approximately) a five percent shift in seat totals, which makes it harder for parties to catapult themselves upward. The big payoffs of FPTP evaporate, but election risk remains. Thus, the incentive to call elections of convenience goes way down.

Another fundamental difference has to do with the composition of coalition governments. In minorities a single ruling party tries to govern alone. In coalitions, the party with the most seats finds coalition partners to sign formal agreements to govern together. This could take the form to support the government for a limited time (as the Ontario NDP and Liberals agreed to in 1995) or it can take the form of cabinet sharing, where the smaller coalition parties get cabinet seats and portfolios.

Cabinet sharing is one of the great strengths of coalition governments. For one thing, it gives all coalition partners some power in the government, which makes it less likely that the governing parties will break the coalition. Moreover, it adds a layer of opposition to the government. In addition to the formal opposition parties who are not in government, the smaller coalition partners can act as "watchdogs" for the larger coalition partner, so that the large partner cannot get too crazy or arrogant. (Unfortunately, it also means that it is easier for the partners to dodge blame for their mistakes..)

So what's the catch? What's up with Italy and its annual elections? The catch is that government in proportional systems do tend, on average, to be shorter-lived than governments in FPTP systems. (See LCC 2004, pp. 141-142 for claim and footnotes) However, the difference in government lifespan does not appear to be as dramatic as opponents of proportionality would have you believe. Coalition governments do fall more frequently than majority governments, but they are significantly longer-lived than minority governments.

As far as I can tell, there are two situations in which coalitions fall -- one good, one bad.

The good reason to break a coalition is if the government is broken and people are clamouring for change. For example, if one of the coalition partners engages in a big scandal and is shown to be corrupt, the other coalition partners might break the coalition to keep the stench of scandal away from them.

The bad reason to break a coalition is if it was doomed to be dysfunctional from the beginning. The best coalitions are formed between two or at most three parties who are ideologically close and who each have a fair share of seats. Letting really tiny parties into your coalition is a terrible idea. So is trying to form a coalition with your sworn enemy, or with seven or eight other parties. And guess what? Italy and Israel both suffer from these problems! Governments in these countries can't form stable coalitions because there are too many small parties, so their governments fall all the time. Other countries with proportional systems tend to face less insane situations, and their governments tend to survive longer.

This does not guarantee that Ontario will end up with coalitions. A party can try to run a minority government under proportional representation. Depending on how stupid they are, some parties might just try (especially in the first few Ontario elections, when parties are still getting used to the implications of proportionality). Nonetheless, I am pretty confident that Ontario governments under MMP will tend to be coalitions, and that this spectre of "permanent minority governments" is false. When MMP opponents make their arguments, ask yourself: are they talking about minority governments, or are they talking about coalitions?

(*) Dobell and Wikipedia disagree on the number of minority governments Canada has had federally. Hrm..


(Dobell 2000): P. Dobell, "What Could Canadians Expect from a Minority Government?" (2000) 1:6, Policy Matters, pp. 4-10

(LCC 2004): Law Commission of Canada, Voting Counts: Electoral Reform for Canada, Queen's Printer of Canada (2004). ISBN J31-61/2004E (and available online).

Livejournal URL:

Mood: uneasy