Polls this Summer

A conversation I had today told me that wrong impressions of what the polls have said this year are rampant.

Essentially, the status quo was upset in early April by the Brault testimony at Gomery. Liberal support began to plunge but was arrested by Paul Martin's speech on April 21. Wild fluctuations in party preference continued as various dramatic political developments unfolded - Martin's deal with the NDP, David Kilgour leaving the Liberals, Belinda Stronach joining - culminating in the squeaker of the confidence vote on May 18. After that we essentially witnessed a return more or less to the status quo ante.

It is all laid out in a graphic here. Note that I have used a three poll rolling average to smooth fluctuations (which I think are meaningless). My source for all these numbers was the table here.

What isn't shown is that after the May confidence vote, Harper's personal numbers plummeted while Layton's soared.

More on OH-2

I have posted before about the by-election in Ohio’s 2nd Congressional district (here and here) and how it demonstrated the potential of the internet to marshal support. The blog MyDD has published a map of contributors to the campaign of Democrat Paul Hackett, who lost narrowly but may run next year for Senate. I think it illustrates well how the internet and the blogosphere is potentially a tremendously important political force. While its efficacy has been demonstrated in the U.S., we have yet to see anything comparable in Canada.

Liberal Caucus and Conservative Ads

The newspaper today brought word from the Liberal caucus meeting in Regina that their principal strategist, David Herle, sees the real prospect of a parliamentary majority emerging from the next election, while conceding that the sponsorship scandal is hurting in Quebec. This implies winning an additional 20 seats minimum in the rest of Canada, a very tall order. Herle makes some suggestions about how this might be done. Let me comment on them:

1. “Make real gains” in Ontario. Given the pattern of recent polls, in my view this is all but impossible. To do so would require raising the Liberal popular vote above the 45.6% received last time. In most recent polls 46% is roughly the highest level of support the Liberals have achieved. For most of its electoral history Ontario has been a three party province. The elections of 1993, 1997 and 2000, when the Liberals won almost all the seats in Ontario, are, in the broader run of history, exceptions not the rule.

2. Win between eight and ten seats in the Prairies. Let us assume he means Manitoba, Saskatchewan and Alberta. In the latter, they will be lucky to win back the two they won in 2004, especially given the fact one of the MPs, David Kilgour, has left the Liberal caucus to sit as an independent. Most polling in Manitoba/Saskatchewan has not been kind to the Liberals. The most recent Ipsos Reid poll reported the following: Conservative 34, NDP – 32 and Liberal -29. This would produce a net gain of one for the Liberals, the riding of Kildonan-St. Paul in Winnipeg. Doesn’t look like eight to ten to me.

3. There is a “paradigm shift” in B.C. and it “could move to us”. This is the most plausible of his assertions. The Liberals have quite consistently been ahead of the Conservatives and the NDP in polls in B.C. However, some in the spring were taken just before and just after the provincial election in which the provincial Liberals received 45% support. This almost certainly meant some spike in federal Liberal support. There is also a previous pattern of Liberal support in B.C. being higher between elections than on voting day. Nevertheless I don’t dismiss out hand his claims here. B.C. with its growing urban and increasingly ethnically diverse populations is changing from the B.C. of yesteryear. A strong Liberal result here at the expense of the Conservatives, who won 22 of 36 seats with just 36% of the vote in 2004, is conceivable. It is also equally plausible that there will be a strong NDP showing in B.C. that would limit Liberals gains.

David Herle has a reputation of being arrogant. And perhaps the news report doesn’t do justice to his presentation, given that later in the day he denied the report, embarrassed by the implication that the Liberals could ignore Quebec. The problem with the denial is that if you believe his scenario to be true, then a majority with zero gains in Quebec would be a possibility. Whatever the context, he does seem to be working with some very optimistic assumptions.

--------------------------------

On another front, the Conservatives are running some ads in Ontario that were clearly intended for airing in the last election. You can watch them on the web here. I would give them about a B- overall. They have an amateurish look to them with MP’s and Stephen Harper delivering scripts to a shaky hand held camera (to give it a faux cinema verité style) in a slightly wooden manner.

A few observations:
1. The health care ad, although ostensibly about waiting lists, appears to me to be intended to offer reassurance that the Conservatives will protect public health care.
2. The ads feature a number of women MPs. The visual grammar here implicitly addresses a Conservative weakness among female voters without explicitly talking about the issue. See, we wouldn’t really re-introduce a new abortion law with all these professional looking women in our caucus, would we?
3. Despite their weaknesses, the ads do have specific messages targeted at an audience of target voters the Conservatives hoped to win over including parents using private daycare, immigrants and suburban families.
4. There are moments when Harper comes across as just a bit too partisan/angry – reinforcing a real image problem he has.
5. The slogan ‘Stand up for Canada’. Not original. Used by another man of the political right, George Wallace, in 1968 as ‘Stand up for America’.

Dog Days of August Polls

New SES and Strategic Counsel polls were published today. While SES President Nik Nanos suggests his poll would mean a Liberal minority, my own calculations in fact produce a Liberal majority with the Grits at 164, the Conservatives all the way down to 62 and the NDP at 27. However, the Strategic Counsel numbers show a tighter outcome than SES, so, in combination I would guess the Liberals are actually still in minority territory.

The SES leadership numbers put Stephen Harper in free fall finishing behind Jack Layton in third place while Strategic Counsel reports that the sponsorship scandal, his principal political asset, is now a key issue for only 2%. Take this latter fact with a grain of salt. I am sure the Conservatives will be courteous enough to remind us all again and again about it in the closing weeks of the next campaign.

What seems to have happened is a kind of negative honeymoon where, having failed to defeat the Liberals in the Commons, the public then turns on the Conservatives and punishes them for their failure.

Cook on OH-2

I received a commentary on OH-2 from Charlie Cook via an email subscription today. I would describe Cook as a right-leaning (and inclined to view things from a traditional perspective), but scrupulously non-partisan political analyst. Here is the key quote:

There is no one reason why this special election was so close. There was the fact that Schmidt was a less-than-stellar candidate who survived an 11-way primary where the two front-runners killed off one another. Her campaign was equally unimpressive and reluctant to go on the attack against Democratic lawyer and Iraq War veteran Paul Hackett -- who in turn was a far better candidate than Democrats usually land in such overwhelmingly Republican districts. Hackett's ability to make this race a referendum on Gov. Bob Taft -- whose administration is now engulfed in a major state government scandal -- was also important. Hackett's message was reinforced by an ad run by the Democratic Congressional Campaign Committee linking Schmidt and the politically besieged governor.

But while all of these factors contributed greatly to the surprisingly close outcome, it is clear that there is deeper significance to this race than some Republicans admit. There are those close to the White House who remain dismissive of the election's overall significance, but to ignore the warning signals this race has given off is to tempt the gods. Just as odd-year gubernatorial races often (though not always) foreshadow subsequent national results, special elections can be a harbinger for what the upcoming national election will hold.

Let's hope so.

OH-2 and the Internet

The Republican, Jean Schmidt, won this August 2, 2005 Congressional by-election, but the Democrat Paul Hackett came very close, losing 51.7% to 48.3%. Now there is talk of him as a Senate candidate in Ohio in 2006.

Overall I think the result could be significant as an indicator of changes coming, especially given Bush’s falling polls. However, the Ohio Republican Party is beset by scandal – summarized well in this Charlie Cook analysis about OH-2 – so its broader importance is less clear. There has quite a bit of good discussion of the OH-2 election in the blogosphere. I recommend the following summary from the Emerging Democratic Majority blog, where one of the key points is about the new power of the liberal blogs:

One of the key lessons of Hackett's near-win is the power of the liberal blogosphere in raising needed funds for individual campaigns. Lead by The Swing State Project, liberal bloggers raised an estimated $500,000 for Hackett, two-thirds of his campaign budget of $750,000, according to the WaPo article.

Another lesson for Dems is that Hackett's impressive tally was boosted by his refusal to water down his criticism of the Administration's Iraq policy or tone down his anti-corruption message.

Victory for Howard Dean
One of the big winners out this contest was Howard Dean, both as DLC chair, and for his organization, Democracy for America, which raised and donated large sums to a grateful Mr. Hackett.

The traditional approach to politics on the part of establishment Democrats has been to concede the bastions of the other side, and focus resources and effort on swing districts. By that standard they would have ignored OH-2. But the partisan Democratic bloggers, and Howard Dean, want to carry the fight to red state America so they pitched in here. The OH-2 battle is a classic example of the divergences, where the Democratic Congressional Campaign Committee, run by Clintonista Rahm Emanual, entered the contest very late after the Republicans, worried about losing, started to invest in their campaign. Even so the DCCC post-mortem earned them some pointed criticism in the blogosphere because the post did not emphasize that Hackett had enthusiastically attacked Bush’s Iraq policy.

The tension between the bloggers and the more traditional Democrats is analyzed extraordinarily well by Los Angeles Times columnist Ron Brownstein in a column for the National Journal that was reprinted here in this Daily Kos posting. In my opinion this is a seminal piece of analysis about a key current trend in American politics.

The key difference between the netroots activists and the more cautious traditionalists is summarized this way:

The Democratic Internet base cradling that trigger does not speak with one voice. But the emerging generation of online Democratic activists, many of them young and shaped by the bruising partisan conflicts of the past decade, seems united most by the belief that the quickest way for Democrats to regain power is to confront Bush more forcefully and to draw brighter lines of division between the Democratic Party and the GOP.

OH-2 will not be last chapter in this debate.

The more things change....

There were two provincial polls released last week in Ontario, by Ipsos-Reid and Environics. Both suggest there is a virtual tie right now between the Liberals and Conservatives. There is a discussion of their implications in this column by the Toronto Star’s Ian Urquhart. As he reports, polls conducted a couple of months ago showed the Liberals ahead but they came after a relatively popular budget. These polls more likely tell us something about the longer-run trend.

To me the interesting thing is that they would almost certainly produce a minority government with the NDP holding the balance of power. Averaging the two polls and applying my seat forecaster would produce a minority PC government with 51 seats, 38 Liberals and 14 New Democrats.

It has all happened before. Ontario was governed by a Conservative minority led by Bill Davis from 1975 to 1981. And his principal secretary in the years just after he won back his majority in 1981 was - current PC Leader John Tory.

Polls in mid-summer

There may be nothing much going on in federal or provincial politics at the moment but the pollsters have been in the field, and there were four federal polls published in July – two by Pollara, one by Environics and one from Decima.

It seems to me that when there is little news to influence preferences what we see is a reflection of medium or long-term preferences. We even see the Paul Martin tormentors complaining about him not providing grist for the mill by keeping a low profile.

Since elections actually occur in the context of a massive amount of communication both paid messages and free media, the current situation is quite artificial. And it can be misleading – think of how in the summer of 1993 many thought Kim Campbell and the then Progressive Conservative Party would cruise to victory in the fall.

Nevertheless this period provides a context worth knowing and there are a few observations to be made:

1. After a spring beating by Gomery hearings, the Liberals are moving gradually upward in Quebec. See especially the table on Quebec in this Decima news release. Given that there is generally a pro-Bloc tilt in the polls – the so-called hidden federalist vote – it actually suggests that Liberals may not actually lose that much ground in Quebec when the election is finally held.

2. There is persistent Conservative weakness in BC. Liberal support may be getting boosted a little by the proximity to the spring provincial election but by July that effect should be fading. I can’t say I understand this but it is potentially significant and being overlooked.

3. The leadership numbers in the Environics poll are interesting. Clearly Jack Layton is making a good impression on the electorate and Stephen Harper is not. My own sense is that Martin, whose numbers continue to fall, has nevertheless arrived at a sort of grudging acceptance on the part of the electorate, and could even benefit in the coming months from the enormous fall in expectations about his performance that accompanied his first year in office.

4. I have calculated seat numbers for the Pollara and Environics polls. I need all the regional numbers and Decima does not provide them. I average the three polls and ran them through my seat calculator. The detailed numbers are here. Compared to the last election, we find the Bloc and the NDP up, the Conservatives down and the Liberals about the same – losses in Quebec are offset by gains in B.C. I am doubtful that the Liberals are as strong and the NDP as weak in Manitoba/Saskatchewan as the projection suggests but that is what comes out.