MMP Disproportionality, part 1: The issue of local seats

September 28th, 2007

MMP Disproportionality, part 1: The issue of local seats

By David Pal // 3 Comments

If you follow the literature and discussion on MMP, you'll notice that the words "roughly" and "approximate" are used quite often. You might hear how a party's share of seats would be "roughly equal" to its share of the party vote. There are a number of reasons for the frequent appearance of these adjectives, some of them small, mathematical and relatively insignificant. You have for instance, the rounding up of decimal points and the imprecise distribution of remainders of seats -- not exact mechanisms, but likely not severe enough to wildly throw off election results.

There are however, two key aspects of MMP that could and likely will affect election results, producing what is called "disproportionality," where a party receives more or less seats than their popular support merits, an obviously undesirable state under a system designed to achieve the opposite. The first is the presence of 90 first-past-the-post local ridings. While keeping local seats is meant to preserve local representation and ensure continuity with Ontario's electoral tradition, first-past-the-post local ridings are inherently disproportional -- they still allow for parties to receive a greater or lesser number of seats than their popular support merits, and for majority governments to be formed by parties with less than 50% of the party vote.

Take a look at this scenario drawn up by the Citizen's Assembly:

Party A has won 55 local seats, 42.64% of the vote, while winning only 39.14% of the party vote. They end up receiving 3 more seats than their popular support merits, while the other parties end up getting a smaller percentage of the seats in the legislature than they deserve. These extra seats are called "overhang" seats. While this is a clearly disproportional result, Party A wouldn't lose any of those extra seats, regardless of how slim their margin of victory might have been.

The MMP systems in Germany and New Zealand compensate for the disproportionality produced by the local ridings by temporarily expanding the legislature to ensure that the results are proportional. Parties that receive more seats than their popular support merits would keep them, but the other parties would receive extra seats to ensure that they have a proportional number. The Citizen's Assembly chose not to include this feature in the recommended system, although it's not inconceivable that it could be included in the final legislation if the referendum is successful.

It's also important to consider how vote-splitting will affect the appearance of proportionality. Since voters might vote for one of the big three parties with a good chance of winning locally with their local ballot and a smaller party with their party ballot, major parties might legitimately receive less of the popular vote than the local vote. Indeed, this often happens to larger parties under MMP systems.

Such a gap doesn't mean that the parties in question lack support and thus don't deserve all of their seats, but having two ballots allows voters to express their support in two distinct ways. This is a unique form of voting produced by MMP. At the same time, this still throws off the results for the other parties, leaving them with a disproportionally smaller share of seats in the legislature. Having balance seats would compensate for the effects of this new kind of voting.

Another lingering issue is the possibility of having a majority government elected without majority support. Currently, a party needs to win 54 seats to have a majority government. Under MMP, a party would need to win 65 local ridings in order to achieve majority status, around 72% of the seats. That's a landslide victory, but nowhere near impossible or unprecedented -- David Peterson's Liberals took 73% of the seats in 1987 and the PCs were at 72% and over in 1951, 1955, and 1959, all with their popular support below 50%. Based on past trends, a party would likely need at least 46% of the local vote and a fortunate geographic distribution of support to achieve this under MMP.

In the next post, we'll look at the second main way disproportionality is produced under MMP -- the Hare Formula that is used to determine how many seats a party deserves.

illustration from Archives of Ontario: F 1125-1-0-0-132

Interact

3 Comments
Add a Comment

Rate this post

Loading ...

Comments

Neither the author nor Spacing necessarily agree with the comments posted below. Spacing reserves the right to edit or delete comments entirely. See our Comment Policy.

I'm not actually concerned by the overhang, or the possibility of a majority government if a party gets over 46% of the vote (though more likely it would need even more). In a MMP system with perhaps 5 viable parties, that would be a huge plurality - one that would merit the ability to implement a platform without having recourse to other parties. I don't think we need to worry about exact proportionality - we would already be far more proportional than in the past, and the (remote) possibility of a majority in fact moderates some of the concerns people have about MMP undermining the ability to run an effective government. My own personal preference would have been to have more constituencies and fewer list places so that the possibility of a majority was increased for a party that got over 46%.