Friday, August 25, 2006

Moving the Party to the Left is not the Right Answer

As the leadership race continues to gather steam and the Liberal Party tries to redefine itself in its current opposition role, there are a lot of questions about what policies the Party should support. Internal divisions are very clear; one needs look no further than the Israel/Hezbollah question to see battle lines drawn in the Liberal Party.

An overarching theme that has attracted a lot of attention is whether the Liberals ought to move further to the left or to the right. Ought they become the party of social justice and big government in an attempt to capture some of the NDP's vote, or ought they become the party of fiscal restraint and small government in order to capture of the Conservatives' vote?

An increasing number of voices within the party have been advocating a move to the left. One of the front runners, Gerard Kennedy, is significantly to the left of most recent Liberal party leaders. Bob Rae, despite his recent right-ward lurch, is a former NDPer. The era of severe fiscal restraint seems to be ending, as candidates speak more and more about the roles government should fill. Indeed, the most right-wing of the candidates, Maurizio Bevilacqua, even spoke early in his campaign about the idea that Liberals should resists temptations to move to the left.

Of course, the overly simplistic analysis of political issues solely on the left/right spectrum is only of limited use, and this I admit freely. At the same time, it is a useful way to discuss issues, and there do seem to exist in most of Canada (at least outside Quebec) certain groups who identify as leftist or rightist. Much of the discourse does indeed happen along this one continuum.

As a (usually) right-of-centre Liberal, I personally hope that the leftward lurch will not come to pass. While this stems largely from my considerations of policy, I also believe that a leftward lean would be disastrous for the party at the ballot box. Why?

When people speak of moving to the left, they imagine that the Liberals will be able to capture most of the NDP's seats and turn them into a virtual non-party. I don't dispute that this could probably happen. But anyone who advocates this has complete lost sight of the fact that a leftward lurch will also cause some of the rightist Liberals to jump ship to the Conservatives. A lot of rightist Liberals voted Conservative in the last election due to dissatisfaction with the the Liberal Party, and as the Conservatives have pursued a relatively moderate course with respect to social policy thus far, the Conservative Party is increasingly coming a close second in many rightist Liberals' minds. A jump to the left, and those Liberals will abandon ship.

So, how much will the addition of the leftists to the Liberal Party help it electorally? Even if the Liberals completely wiped out the NDP and stole all their seats (without losing any others), they would only have 131 seats--24 short of a majority.

But in addition to this, we have to remember that a leftward lurch would cause the Liberals to lose seats. How many? That's not entirely clear. It can be estimated, however, using basic models of voting behaviour, if one makes certain assumptions.

Let's assume that there are three parties, the Conservatives, Liberals, and NDP, all of whom are placed on a continuum from right to left. Let's also assume that voters occupy a similar space on this continuum. Parties announce their policies, thereby placing themselves on this continuum, and voters then choose the party which is closest to them in their preferences.

What happens when a party changes policies? It typically gains supporters and loses supporters. Although it's difficult to tell exactly how this would change, let's assume that voters are distributed uniformly, such that a policy change in the centre would cause the party to lose as many voters as it gains. That is, if the Liberals move to the right and gain 1% of the Conservative's voters, then they would correspondingly also lose 1% of their voters to the NDP. While this is by no means a completely realistic scenario, it is also of some value, especially in the case of a centrist party sitting between a leftist and rightist party.

The following table indicates the results of a model created to examine the changes in seats in the House of Commons given changes in Liberal's placement on the left/right continuum.

The methodology is as follows. The results of the 2006 election were put into a tabular format, with three columns representing the share of the vote captured by the Conservatives, Liberals, and NDP is each of the 233 ridings in Canada outside Quebec. The 0 Liberal Deviation row corresponds to the 2006 election result.

The other rows correspond to the composition of the House of Commons if the Liberals had grabbed different shares of the vote. In the situation of a -1% Liberal deviation, it is assumed the Liberals have moved to the left on the political continuum (making it more attractive to those with left wing preferences), such that 1% of voters move from the NDP to the Liberals, and 1% of voters from the Liberals to the Conservatives. The victors of each riding are then recalculated, in accordance with the proposition that the candidate/party with the highest number of votes in a given riding wins.

Two notes:
a) Quebec is excluded from this analysis, because the federalist/sovereigntist axis is as important--if not more important--than the left/right axis, so analysis along the left/right axis will be completely useless.
b) The "Government" column calculates situations only based on the rest of Canada, as if Quebec did not exist. In fact, because of the presence of the Bloc Quebecois, it would be much more difficult for any government to win a majority of seats. Under more realistic scenarios including Quebec, a Conservative majority would only occur under a -8% deviation, not a -1% deviation as in the table below.

Liberal Deviation Liberal Seats Conservative Seats NDP Seats Government:
-10 66 156 11 Conservative Majority
-9 67 152 14 Conservative Majority
-8 72 147 14 Conservative Majority
-7 75 140 18 Conservative Majority
-6 77 137 19 Conservative Majority
-5 80 131 22 Conservative Majority
-4 81 128 24 Conservative Majority
-3 85 124 24 Conservative Majority
-2 86 121 26 Conservative Majority
-1 89 117 27 Conservative Majority
0 90 114 29 Conservative Minority
1 94 109 29 Conservative Minority
2 94 106 31 Conservative Minority
3 98 101 34 Conservative Minority
4 97 97 38 NDP Supported Liberal Minority
5 97 90 45 Liberal Minority
6 99 85 48 Liberal Minority
7 98 82 52 Liberal Minority
8 99 81 53 Liberal Minority
9 99 78 54 Liberal Minority
10 98 75 59 Liberal Minority

This analysis shows that while there are some seats to be gained by moving to the right, a move to the left is disastrous, as the Liberals would lose seats by making such a move. The NDP seats won would not make up for the seats lost to the Conservatives. Why is this case?

In most ridings where the Liberals have a chance of winning, their main competitor is the Conservatives, not the NDP (who often trail by quite a lot). Thus, a move to the left makes them stronger relative to the NDP, but in most cases they were never in any danger of losing the riding to the NDP. Rather, they were in danger of losing it to the Conservatives.

Under the -5% scenario (i.e. the Liberals move 5% to the left), the Liberals would have lost the following ridings:

Newton--North Delta
West Vancouver--Sunshine
Coast--Sea to Sky Country
London West
Saint Boniface
West Nova
Desnethé--Missinippi--Churchill River
Saint John
Mississauga South
Random--Burin--St. George's

And they would have only gained:
Hamilton East--Stoney Creek
Sault Ste. Marie
Parkdale--High Park

In any event, there's a pragmatic justification for not moving to the left. A leftward lurch will not win the Liberals any elections. It will only strengthen a resurgent Conservatives, who will increasingly attract the moderates.


Post a Comment

<< Home