By: Robbie Rolfe
The results of the Alberta provincial election that took place on April 23, 2012 generated a lot of commentary on the gap between what the polls were saying and what actually happened on election day. While this is an interesting puzzle, it is perhaps more important to look at another gap—namely, the chasm between votes cast and seats won.
There is a large differential between the percentage of the vote some parties received and the percentage of seats they won. The table below shows seat share minus vote share by region. A negative number indicates where a party had a smaller share of seats than votes while a positive number indicates where a party had a larger share of the seats than votes. When it comes to the first and second place parties (the Progressive Conservative and Wildrose parties), the differentials are quite large.
The Wildrose party had strong support across the province. Nevertheless, it came a close second or third in many ridings. To modify the cliché, close only counts in horseshoes and hand grenades, not in the single member plurality races we use to select our representatives in Alberta and across the country. So despite garnering more than a third of the popular vote in the province, the Wildrose party won only a fifth of the seats.
The Progressive Conservative party, on the other hand, won about 23 more seats than it would have if each party had received the same proportion of seats as its proportion of the popular vote. In many cases, the PCs won with a plurality (the most votes) rather than a majority.
Another way to measure the gap between vote share and seat share that typifies single member plurality systems is the least squares index, a common measure of proportionality used in political science. The scale runs from 0 to 100 with higher numbers indicating a less proportional result. This measure allows us to compare proportionality across elections. For the recent Alberta election, its value is about 22. To provide some context, that is a more disproportional result than any Canadian federal election since 1945. (This includes Brian Mulroney’s 1984 victory, where his party won 75% of the seats with 50% of the votes, while the Liberal and New Democratic parties’ seat shares were about half of their popular vote shares.)
Proportionality matters. Arguably, when the distribution of seats in the legislature does not accurately match the preferences of the voters expressed in the general election, it is less likely to pass laws reflecting the diverse preferences of the population. In other words, the single member plurality system we use in Canada tends to distort the representativeness of our legislatures.
To create a better match between votes and seats, we would need to implement some form of proportional representation. In these systems, seats are allocated based on a candidate or party’s share of the popular vote rather than on which candidate comes first in each riding. It may even have other positive effects (for example, it may increase the number of women and ethnic minority representatives in the legislature, two historically underrepresented groups).
Of course, a change of this sort will also have costs. For example, it will likely mean the end of majority government. In order to get a majority in a proportional representation system, one party will have to get very close to half the vote in order to get more than half the seats. This could be costly because majority governments are seen as stable, strong and able to implement their campaign promises. Voters can also identify responsibility easily and hold governments accountable when one party controls the levers of power. It is also easy for the electorate to turn majority governments out in a single member plurality system, as small movements in the popular vote can result in large changes in seats. (Some also argue that majority governments are better for economic performance, though there appears to be no statistically significant link in practice between the electoral system and economic performance. See, for example, Arend Lijphart’s 1999 book, which compares different democratic institutional types.)
Nevertheless, we can discuss and decide on a system that works for us. There are many options we can choose (we can even retain single member districts). We can minimize the costs and maximize the benefits.
We cannot do that, however, without talking about it—it is time to grapple with this elephant in our political living room.