Sunday, September 26, 2010

Trends still favour New Brunswick PCs

The race in NB tightened up and the final polls are based on surveys more than a week old.  There is a good summary of all the details over at It nevertheless looks like the PCs will wind up on top.  The excellent blog Politics from a New Brunswick Perspective foresees a tight race and a narrow PC victory. 

Wednesday, September 22, 2010

The politics of gun control

The gun registry lives on.

While one can understand the intense media interest in this, the commentary of the pundit class does drive one to distraction.  There are two particularly idiotic examples here and here (both of which ask us to accept that despite the Conservatives being poised to win the vote for many months, losing actually helps them). The clip of Harper on tonight's news says otherwise.

The whole issue matters much less electorally than meets the eye. To accept that it matters, one must believe there are a large number of voters in the constituencies of the NDP and Liberal members who switched their vote tonight to support the registry who, a), did not vote Conservative last time, but b), feel so passionately about this issue that they will hold that concern all the way to the next election, ignore all the other issues that come along, and support the Conservatives due to this consideration either primarily or alone.  Was this a big vote determining issue in 2006 in these places or 2008?  TC does not think so.  In any case most of the rural seats inclined to go Conservative are already safely in that party's camp.

The Liberals recognized the second time around the issue would not hurt them and could give them an advantage by being unified while the NDP was divided, all the while driving soft urban small 'l' liberal votes their way.  The wonderful Pundit's Guide has a detailed listing of the Liberal and NDP members who voted last time with the Conservatives to kill the gun registry.  If you look closely at the results in just these constituencies, there are no more than three or four that could be classified as both rural and holding real potential as Conservative gains in a new election.  The most obvious candidates are the NDP held ridings of Western Arctic (3.8% lead over the Cs) and Skeena (13.8%) and, among the Liberals, Malpeque (4.9%) and Yukon (13%).

There are a few other contests where the Conservatives might think about a challenge but in most constituencies, the principal challenger to the NDP is the Liberal candidate and vice-versa, or, as in the case of Esquimalt-Juan de Fuca and Welland the ridings are not rural (and presumably therefore not obsessed with gun control) - Esquimalt has a rural component but is mostly suburban Victoria while Welland is mainly industrial and urban.

In the end the vote was a victory for Jack Layton (who I suspect had at least one more vote in reserve).  The potential for a significant setback for the NDP was there, so he should be wiping his brow.  The Liberals might well have been able to use the gun registry vote as a tactical voting metaphor in urban areas in the next election.  In this case the difference between winning and losing was crucial.  In addition, losing is not good for the government (except possibly in fund raising if one accepts their claims) - there is always at least a brief negative honeymoon associated with a strategic defeat in a contest that they had expected to win.

Notwithstanding all the ballyhoo, TC does not expect this to be an issue or key voting consideration in the next federal election.

Sunday, September 19, 2010

U.S. Midterms - What are the Democrats Prospects?

Most accounts of the Democrats prospects suggest a bleak midterm election result.  Until far right Christine O'Donnell (who once claimed to have dabbled in witchcraft) won the Republican nomination in Delaware over a moderate, the conventional wisdom was increasingly suggesting the Republicans could win both the House and the Senate.

I have followed all this but noted months ago that surveys reported again and again that the Republicans were more disliked than the Democrats.  The assumption has been that the declining economy and slumping income meant that this would be a very bad year for Democrats.  However, never before in a situation like this has the opposition party been so disliked.

Now there is an intriguing new analysis from Barry Pump, a grad student in political science at the University of Washington, that suggests we should indeed pay close attention to the survey evidence on the unpopularity of the Republican brand.

The argument is simple:
I’ve been looking at the low Republican favorability rating for a long time and always suggesting to friends who ask me about it that it’s more important than we think. People don’t vote for people they hate. But they may hold their nose and vote for people they dislike less.
And from a second post: theoretical argument is that people don’t vote for parties they hate. They vote for the party they dislike the least relative to the other party. The leading counter argument, and the one that has pretty much dominated the narrative is different. That argument goes: because the economy is terrible, people will vote for the opposition party regardless of how much they dislike them. In other words, it’ll be a referendum election on the economy, and the economy stinks so Democrats will lose a bunch of seats, probably even the majority in the House. 
In order to predict how well parties should perform in the upcoming election analysts typically use polls on the generic Democrat or Republican ballot question. But Pump believes that more attention should be paid to the net favourability polls which report that the Republican Party is far more disliked than the Democratic Party.  Look at the bars on the far right of this graphic (Republicans are red) for the most recent party standings on this:

Mr. Pump compared the predictive value in past elections of the favourability surveys with the generic ballot. He acknowledges that there are data limitations on the analysis but concludes:

So what can we take away from this little discussion?
First, we’ve never been in a situation until now — as far as we have data to show it — where both parties were disliked but one party was disliked far more than another. We’ve also never been in a situation where the difference between the favorability rankings of the two parties was as great as it is now. (That’s from the first graph.)

Second, we’ve yet to be in a situation until now — as far as we have data to show it — where the favorability rankings of the two parties were so discordant with the generic ballot.

Third, given this new territory and the uncertainty of estimates made so far away from election day, analysts don’t know as much as we’d like about where things stand in US politics.

Fourth, there appears to be a puzzle about voter decision making embedded in all of this: under what conditions will voters choose a party they profoundly dislike over the party they merely dislike? And is a dismal economy one of those conditions?

And fifth, I’m inclined to think that while Democrats will lose many seats on Election Day, those losses will be tempered by the fact that the Republican Party “brand” has been deeply tarnished over the last six years and many voters don’t think the party is ready to govern again.

Wednesday, September 15, 2010

New Brunswick Election - Conservatives likely to win

Both the blog ThreeHundredEight and Politics from a New Brunswick Perspective have reacted to the latest polling from New Brunswick that shows the Conservatives surging to a 12 point lead to suggest that the province is heading for a Conservative government. TC agrees that this is the point at which the campaign is taking on a clear direction and the outcome really isn't now in doubt.  Premier Shawn Graham is headed for defeat.