Thursday, March 31, 2011

More Policies: Tories on Trade, Grits on Pensions, Dippers on Corporate Taxes

The second batch of policy announcements is, on the aggregate, somewhat better than the first, but still unimpressive to me.

The Tories reiterate their support for free trade agreements, which I generally support, but offer few details. They also say that under a hypothetical "Liberal coalition," free trade agreements wouldn't happen. Does this mean that if the Liberals win and propose a trade deal, the Tories would side with the NDP and the Bloc in opposing it? To me, this issue doesn't put too much light between the two main parties: perhaps the Conservatives would be somewhat more proactive than the Liberals in seeking trade agreements, but the Grits are hardly protectionists. Still, this is a much better policy than the ill-conceived family tax cut announced earlier.

I have little to say about funding the hydroelectric project in the Atlantic: I don't know enough details to determine whether it's a good project. If it is a worthwhile investment, the appropriateness of federal funding depends on one's own view of the federation. Although given that Hydro-Québec did not get any help from the federal government over the years, the Bloc is totally justified in hammering this move. (Yes, the Bloc is often too whiny, but not on this one.)

The Liberals want to expand the CPP, allow Canadians to invest more with the CPP, and expand the GIS. I'm generally against the first idea, in favour of the second, and neutral concerning the third.

My opposition to an expansion of the CPP is because my generation is already getting a pretty bad deal out of it because we're effectively subsidizing seniors. While CPP long-term real returns are forecast at about 4%, Canadians born between 1970 and 2000 will only get a 2.2-2.4% return on their contributions according to the latest CPP actuarial report (see page 74). Instead of their money doubling every 18 years or so, people under 40 have their money doubling every 30 years. This is because those born before 1950 are able to enjoy returns above 4%. An expansion of the CPP, even gradual (unless it's done over 40 years), will further disadvantage young workers.

Allowing Canadians to invest more with the CPP is a good idea. Many people might like to save more, but do not know how to appropriately invest their savings. Moreover, asset management fees are often ridiculously high. The CPP would provide a low-cost way for Canadians to benefit from professional financial services. The flip side is that it wouldn't be a personalized option, so each citizen will have to assess whether the CPP strategy responds to their needs - but at the very least, having more choice won't hurt. Because additional contributions are capped at the RRSP deduction limit, this policy probably won't put too much strain on the CPP investment board or squeeze the private asset management industry too much. I'd like to see this policy implemented.

Finally, whether you want to expand the GIS is essentially a moral question concerning your attitude toward redistribution and how responsible poor seniors are for their situation.

The NDP keeps on going with economic policies that sound good, but either won't work or aren't cost effective. Why increase the gap in tax rates between large and small companies? Doing so increases the amount of economic distortion (you're essentially penalizing successful firms that become big), and is a woefully inefficient way of doing redistribution. Indeed, corporations aren't people: their owners/shareholders are. Are shareholders of big firms richer than owners of small businesses? Keep in mind that big firms are often largely owned by pension funds, which represent average Canadians. A much more effective way to redistribute is through the personal income tax system. Unfortunately, at the NDP, when populism and progressivism clash, the former often wins out.

The tax credit for creating new jobs will run into trouble too. Do I get the credit if I fire a worker and hire someone else the next day? The next week? The next month? The next year? Also, you would be penalizing firms that held on to their employees through the recession (because they wouldn't be able to get the credit by "creating" jobs) and paying firms that got rid of workers at the first sign of trouble. Guess what they'd do the next time around...

Nanos: Tories' Lead Shrinks to 6.4, but Is Up in Ontario

Today's Nanos tracking poll has the Liberals up by 4% nationally. The Conservatives are also up, but by just 0.7%, so the gap has shrunk by 3.3%. Since the first tracking poll was conducted March 27-29, while this one covers March 28-30, it follows that the measured difference was about 10 points smaller on March 30 than March 27. Such a shift is not quite (but close to) statistically significant for a daily sample of 400.

Before the Grits get too excited nationally, however, they need to worry about Ontario, which is the only region where their support did not increase. Instead, the Tories are up by 4.5% there, implying a 13.5% increase from March 27 to 30. However, due to the limited regional sample size, that change is not statistically significant.

This poll is bad news for the NDP, which is down everywhere, while the Bloc is stable.

The new projection moved seats around for little net change:

CON - 156
LIB - 68
BQ - 52
NDP - 32

My weighted average for the Tory national lead is down slightly, to 13.2%.

Wednesday, March 30, 2011

Forum Research: Tories by 17; Nanos: Tories by 9.7

Since the last projection, regional data from the Harris-Decima poll, as well as numbers from a Forum Research survey (via and from Nanos' daily tracking poll have become available. The lesson from these numbers is not very exciting: things have not moved much during the first weekend of the campaign in any region of the country.

As a result, the new projection does not change much:

CON - 155
LIB - 68
BQ - 52
NDP - 33

The average Conservative national lead is now 13.5%.

Tuesday, March 29, 2011

Policy Choices So Far: Boo, Meh and Whaaaa?

Each of the three major parties has made one big announcement so far in the campaign.

Conservatives: "Stephen Harper's family tax cut" would allow families with children under 18 to split up to $50,000 in income. As in any income splitting scheme, this will only benefit couples where the two people are in different tax brackets. Clearly, this is a highly regressive policy since it will not benefit at all low-income and many middle-income households.

Worse, even in terms of economic efficiency, this policy's effect is at best ambiguous, and in fact probably negative. The tax cut will reduce the marginal tax rate of the first earner in some couples by effectively putting him/her in a lower bracket. But now the second person in the couple faces higher taxes on his/her labour income. For the vast majority of Canadian families, having at least one full-time earner is necessary - that person will have an incentive to work more, but there's only so much overtime you can do. However, for many couples, whether the second partner works is much more of a choice. Under the Conservative plan, there will be a strong disincentive for that person to work. As a result, the net long-term effect of this policy on the economy may well be contractionary. In any case, it is unlikely to be expansionary - which is pretty sad for a tax cut.

Think about it: a proposal that is regressive, contractionary, and costs money. Boo.

Liberals: The Canadian Learning Passport will provide $1,000-$1,500 per year for up to 4 years to students at post-secondary institutions. Before getting too excited, however, parents and youth should note that this grant replaces the education and textbook credits (the tuition credit will stay). For a full-time student, those credits are now worth $465/mo x 8 mos/yr x 15% = $558/yr, or say $500/yr after time discounting because students often don't have the income to benefit from the credit right away. Therefore, the true benefit of the Liberal proposal for most families is only about half the advertised amount.

It is unclear whether this policy is progressive: although low-income families get more per child, children from high-income families are more likely to go to university. It is also unclear whether it is economically efficient - this depends on how many more kids go to a university as a result (I'm assuming that's a good thing), and on economic distortions generated by the cost. Meh.

NDP: Wants to cap credit card interest at prime+5% (8% currently) and to regulate credit card transaction fees. Just the interest rate cap will nab something on the order of $10 billion from credit card companies. Think about how much we'll all save - a wonderful free lunch! Whaaaa?

The credit card industry may generate high profits, but it also requires lots of capital. The question is how much excess profits it generates - i.e. profits over and above what all that capital would generate if invested elsewhere. The answer is, most probably, much less than $10 billion. After all, if offering cards were that profitable, banks would be falling over each other to give cards with better terms in order to attract more customers.

If only credit cards with 8% interest and low transaction fees can be offered, then:
1. there will be a lot fewer of them;
2. those that exist will have little or no rewards, and high annual fees.
Canadians will then have two choices:
- pony up that big annual fee up front; or
- use debit, keep checking their chequing account balance to make sure balance doesn't run low, and go through the hassles and uncertainty of getting a personal loan when a little extra is needed.
My guess is that few would choose the former option, so the NDP proposal would essentially kill credit cards in Canada. Most Canadians would simply end up losing the convenience of revolving credit and probably paying a bunch of extra bank fees instead of credit card interest. The NDP is probably smart enough to know this. Part of the party may even think that getting rid of credit cards is a good thing, though I would hope that the majority is not that paternalistic. Most likely, they are banking (probably rightfully so) on enough Canadians being naïve enough to support this policy, knowing full well that they'll never get a chance to implement it.

Monday, March 28, 2011

Abacus: Tories by 9; Harris-Decima: Tories by 14

Two new polls today, one from Abacus and one, with incomplete regional results, from Harris-Decima. Both these surveys were conducted partly before and partly after the official start of the campaign. The Abacus poll is somewhat encouraging for Ignatieff: a single-digit national gap, with a virtual tie in Ontario and a tie with the NDP for the lead in Atlantic Canada. However, the Harris-Decima poll is pretty bad for the Grits: the 14-point gap is more than twice that measured by the previous H-D poll, and the 14% in Québec is disastrous and suggests that CROP's 11% may not be an outlier after all. The silver lining is that H-D still puts the Tory lead in Ontario at a modest 6%.

The projection moved one Ontario seat from the Conservatives back to the Liberals:

CON - 154
LIB - 70
BQ - 52
NDP - 32

The average Conservative national lead has increased to 13.0%.

One note concerning methodology: I will gradually accelerate the depreciation of polls this week. Normally, polls up to 7 days old are given full weight, while those less than 27 days old are given some weight. I will transition to something like 3 and 13 for the campaign.

Meanwhile, in Québec Politics

The spokesperson of Québec Solidaire, a far-left separatist provincial party in Québec with one MNA and 11% support according to the latest Léger poll, proposes to "limit economic growth." The party also wants a $16/hour minimum wage. That's more than the average wage of all workers in "sales and service occupation" (as defined by Statistics Canada), which comprise more than 25% of all employees in Canada.

The NDP looks positively right-wing by comparison...

Around the Web: Projections at the Start of the Campaign

I thought it might be useful to record projections from various websites reflecting the situation just before the writs were dropped. Links to all of these sites can be found on the left of this page.

Issued after the weekend polls came out:
155 C, 66 L, 52 B, 35 N (Too Close To Call)
152 C, 72 L, 51 B, 33 N (
157 C, 64 L, 53 B, 34 N (Riding By Riding)
156 C, 68 L, 51 B, 33 N (LISPOP)
155 C, 69 L, 52 B, 32 N (Canadian Election Watch)
155 C, 68 L, 52 B, 33 N (average)

Issued on March 25, prior to the weekend polls:
157 C, 69 L, 49 B, 33 N (democraticSPACE)
157 C, 68 L, 51 B, 32 N (LISPOP)
156 155 C, 68 L, 51 B, 33 N (average of 6 projections)

As you can see, all the projections are very consistent with each other. This is not surprising since they are all based on averages of recent polls.

A more dated projection also based on polling averages was posted by CalgaryGrit on March 18: 163 C, 66 L, 55 B, 24 N. (Actually, CalgaryGrit posted ranges, and I took the midpoint of each range.)

There are also projections based on individual polls. Election Almanac posts them on a regular basis. Also, on March 14 today, EKOS released an in-house projection based on their 2/24-3/8 3/17-24 poll: 140 139 C, 86 L, 52 53 B, 29 N, 1 IND. Of course, EKOS' polling data is less unfavourable to the Liberals than most other firms'.

Finally, BC Iconoclast has not issued a projection based on current polls, but instead put up a prediction of the results of the campaign on March 23: 150 C, 79 L, 49 B, 30 N.

Sunday, March 27, 2011

Maps of Canada at the Start of Campaign 2011

Here are the maps of the last projection reflecting pre-writ polls. I will update these maps if there are more polls released that were concluded by March 25.

These riding-specific projections take by-elections into consideration. They do not account for the incumbency factor. For example, Esquimalt--Juan de Fuca is projected to go Liberal, but since it's a tight riding and incumbent Liberal Keith Martin is not running this time, one could argue that it is more likely to go Conservative.

I plan on mapping the final projection of the campaign. If I have time and there is demand, I may also map one or two additional projections during April.

Trend from the Past 20 Months

The start of the 2011 campaign seems to be a good time to review the trends in seat projections since the creation of this blog about 20 months ago.

As you can see, the current Conservative lead is as large as it has been since July 2009. Ever since the prorogation fallout in early 2010, the Tories have been on a generally upward path (although they were flat for most of 2010). The Grits have been generally falling since then, save for a bump during the summer of 2010 that got completely wiped out over the fall. The Bloc has been stable around 50 seats, while the NDP has dropped from around 40 during most of 2010 to the low-to-mid 30s.

I will, of course, start a new graph for this campaign, and hope to post chart updates every week.

Conservatives Start the Campaign in Majority Territory (Angus Reid: Tories by 14; Léger: Tories by 16; CROP: Grits at 11% in Québec)

Life gets in the way sometimes, but here's the post that I promised for last night.

On the first day of the campaign, there was an avalanche of bad polling news for the Liberals: they trail in every region of the country, and by double digits in Ontario. CROP's poll even suggests that they've completely fallen off the map in Québec, though that one looks like an outlier.

In both national polls, the Conservatives are at 39% nationally, and in fact at or above 39% in every region except Québec. Both polls also give the NDP 19%, which is pretty good for them. The NDP's strength implies a division of the left that would allow the Tories to earn a majority while staying below 40%.

The new projection is the worst one for the Liberals since the inauguration of this blog, and has the Tories back (barely) in majority territory. I would like to remind you that these numbers try to reflect the parties' standing during the past week, and are not predictions of the final election result.

CON - 155 (39.0%)
LIB - 69 (26.5%)
BQ - 52 (9.9%)
NDP - 32 (16.8%)
GRN - 0 (6.5%)

The average Conservative national lead is thus 12.6% (numbers don't add up due to rounding). This is only 1.2% more than in the 2008 election, but the Conservative vote is also more efficient than back then due to their progress in Ontario.

Because this is the first projection of the campaign, I will list below the projection for each region of the country:

Atlantic Canada
CON - 16 (39%)
LIB - 12 (34%)
NDP - 4 (18%)

BQ - 52 (39%)
LIB - 12 (19%)
CON - 10 (21%)
NDP - 1 (15%)

CON - 57 (42%)
LIB - 34 (32%)
NDP - 15 (17%)

CON - 23 (53%)
LIB - 3 (21%)
NDP - 2 (19%)

CON - 27 (58%)
NDP - 1 (11%)
LIB - 0 (19%)

British Columbia
CON - 21 (40%)
NDP - 8 (24%)
LIB - 7 (23%)

Clearly, compared to the 2008 results, the most salient changes are Liberal losses to the Conservatives' favour in Atlantic Canada and Ontario. The Grits' popular support has in fact increased across Western Canada, but they were so far out of the game that this rise does not translate into many more seats. The NDP has lost ground everywhere except in Québec, but the losses are small.

It will be interesting to compare these numbers to the final projection of the campaign and to the election results. Maps are coming soon!

Saturday, March 26, 2011

The 2011 Campaign is Underway!

It's official: Canadians will be heading to the polls on May 2. (Well, a minority of Canadians, since minors, people that don't vote and absentee/early voters will, in fact, not be at a polling station on Election Day. Sadly, I will not be part of this minority.)

Will the Conservative teflon coating hold? Will Ignatieff finally tell us what he and the Liberals are now about - and will it be too late? Will the NDP hold on to their near-tripling of seats in the past decade in an election where the economy will likely figure prominently? Will the Greens finally manage to elect an MP? Will the Bloc... Well, we know they'll win 40-55 seats anyway.

Five weeks is an eternity in politics. That's even more true during a Canadian campaign, where changes in the public mood are often large, brisk and unexpected (see 1993, and to a lesser extent, 2006).

I hope to issue new projections every day or two during the campaign. The poll weighting formula will use more rapid depreciation - probably roughly twice as fast as usual.

Later today, or whenever the details of the latest Angus Reid poll comes out, I will issue a regular projection, which will be mapped. For now, here's a writeup of a Segma poll in the five ridings of Quebec City proper. The Conservatives won 3 of these 5 seats in 2008, with an overall 3% edge over the Bloc in popular vote. The survey shows that the Tories retain the lead in only one of these ridings (Louis-Saint-Laurent, held by Josée Verner), and trail the Bloc by 7% citywide.

Friday, March 25, 2011

EKOS: Tories by 7.2

The latest bi-weekly EKOS poll has the Conservatives squarely in minority territory, with a modest lead in Ontario. Not much time today, so I won't elaborate, but here is the latest projection:

CON - 152
LIB - 73
BQ - 51
NDP - 32

The average Tory national lead is back down to 11.3%.

When the campaign officially gets underway, I will say more, and post new maps and a new trend graph. So check back tomorrow!

Thursday, March 24, 2011

Ipsos: Tories by 19 (!)

The latest Ipsos poll is a blockbuster for the Conservatives. They lead by double digits in Atlantic Canada and Ontario, are at a respectable 25% in Québec, and are at over 50% in each of the Western regions. The Bloc remains strong, but for both the Liberals and the NDP, this survey's numbers are a disaster across the board.

Frank Graves of EKOS, however, has suggested that the latest EKOS numbers, which will be released tomorrow, are completely different. So stay tuned...

In the meantime, the aggregate projection moves back up for the Tories, but not quite back into majority territory:

CON - 154
LIB - 72
BQ - 51
NDP - 31

The average national Conservative lead is 12.1%.

Tuesday, March 22, 2011

Harris-Decima: Conservatives Lead by 6 Only; Projection: Back to Conservative Minority

The latest Harris-Decima poll tells the story of a Conservative party that is actually squarely in minority territory. It shows virtual Liberal-Conservative ties in Atlantic Canada and Ontario, much like the polls before the recent flare-up in Tory numbers. Also, this survey suggests three-way race in BC, though I would caution that Harris-Decima often shows lower Tory numbers than other pollsters in BC.

We've had six March polls so far, and while four of them suggest that the Tories are either in or close to majority territory, two instead say that the Conservatives are in familiar minority grounds. Here is the Tory-Grit gap in each of these polls:

EKOS (2/24-3/8, 2892 respondents): 7.4%
Ipsos (3/7-9, 1002): 13%
Angus Reid (3/8-9, 1021): 16%
Léger Marketing (3/7-10, 1494): 13%
Nanos (3/12-15, 1216): 11%
Harris-Decima (3/10-20, 2001): 6%

A similar pattern of EKOS and Harris-Decima showing smaller gaps than other pollsters also emerged in February. Who's right?

In any case, the aggregate projection uses a weighted average of recent polls (9 for this one), and spits out the following:

CON - 152
LIB - 73
BQ - 51
NDP - 32

Not surprisingly, given this latest survey, the Tories are back in minority territory. Their average national lead over the Grits has dropped to 11.4%, the same as in the 2008 election.

Sunday, March 20, 2011

Nanos: Tories Lead by 11; Projection: Conservative MAJORITY

The latest Nanos poll suggests little movement. In Atlantic Canada, the only region where the Liberals vie for the lead, this survey sides with the Tories by a statistically insignificant 5%. In BC, unlike most other recent polls, this survey suggests that the NDP is competitive with the Conservatives (7.8% behind). In fact, this poll, as is often the case with Nanos surveys, is a strong one in general for the NDP: 19.9% nationally, and 23% in Ontario.

The number that shifted the projection slightly to the favour of the Tories is their 12.3% lead in Ontario - yes, that's bigger than their national lead. If the election plays out this way, the Tories would likely win over 60 Ontario seats. As a result, the aggregate projection now has the Conservatives in majority territory for the first time since the creation of this blog:

CON - 155
LIB - 71
BQ - 51
NDP - 31

The average national lead for the Tories is also 12.3%.

If a campaign indeed gets underway, it would be rather ironic: after repeatedly backing off from defeating the government due to weak polls, the Grits and Dippers would head into the election both in their weakest position in the past 20 months.

Thursday, March 17, 2011

Ipsos: Tories Lead by 13

The results of the latest Ipsos survey have been posted on Relative to other recent polls, this one has the Conservatives particularly strong in Atlantic Canada (47%) and BC (50%), the Liberals strong on the Prairies (35%), the NDP strong in Alberta (19%) and the Bloc in a good position (46% in Québec).

There are no changes to the aggregate projection:

CON - 153
LIB - 73
BQ - 51
NDP - 31

The average Tory national lead is 12.1%.

Monday, March 14, 2011

Léger: Tories Lead by 13 Nationally

Léger Marketing has released a national poll confirming the Conservatives' solid national lead. This poll is a very bad one for the Liberals: they trail the Conservatives by 14% in Ontario and the Bloc by 23% in Québec. The only bright spot for the Grits is a 7% lead in Atlantic Canada.

For the Tories, their low 16% in Québec is worrisome, especially because they trail the Bloc by 11% in the Quebec City area, which they won by 7% in 2008. However, their stellar results in Ontario and BC (45% vs. only 13% for the NDP) make this a quite satisfying poll.

The aggregate projection changes little, and the Conservatives are still short of a majority:

CON - 153
LIB - 73
BQ - 51
NDP - 31

The average Tory national lead is now 12.2%.

Thursday, March 10, 2011

Angus Reid: Tories Lead by 16; EKOS: No, It's 7.4

After a week without new polls, we got two today. First, Angus Reid released a survey showing not so much the Tories soaring, but the Grits tanking - to 23%. Despite the ghastly headline, the Liberals are actually doing OK in Québec, also with 23%, compared to a weak 34% by the Bloc. In Atlantic Canada, all parties are below 30% due to a mysterious 20% going for 'Other'. Even with very small Atlantic sample sizes, this is a freak number: random variation has a very small chance of explaining it...

For the Tories, this is a great poll across the board (except in Atlantic Canada): 41% in Ontario, 45% in BC, and even stronger results than usual across the Prairies. They even got 22% in Québec despite refusing to fund the new arena in Quebec City (the poll was conducted March 8-9).

As for the NDP, the 19% in Ontario is good, but their numbers out West are dismal: they would be down to around 6 seats from 15 in the previous election.

Later in the day, EKOS released its bi-weekly poll, which appears to tell a very different story on the surface. However, a good chunk of the difference in headline numbers comes from the Tories not quite having as eye-popping numbers out West as in the Angus Reid poll - this does not have big seat implications.

In all important Ontario, the two polls are consistent with each other: Angus Reid shows a 12% Conservative-Liberal gap, while EKOS shows 7%. Moreover, the two polls agree that the NDP is doing poorly in the West, and that the Bloc has dropped from around 40% to the mid-thirties. The most surprising finding of this poll is a 9% Liberal lead in Atlantic Canada, which differs from what all February polls have shown. Of course, due to small sample size, this remains to be confirmed.

The new projection reflects the weak latest numbers for the Bloc and the NDP. Even as the Liberals bounce back a little, the Tories inch closer to a majority:

CON - 152
LIB - 74
BQ - 51
NDP - 31

The average Conservative national lead is up slightly, to 11.9%.

Thursday, March 3, 2011

Harris-Decima: Conservative Lead at 8

Harris-Decima's latest poll, conducted 2/17-27, shows a solid, but not huge, Conservative lead. This puts it closer to the latest EKOS poll (5% gap, 2/10-22) than the recent Abacus (15%, 2/23) and Ipsos (16%, 2/23-27) surveys. There were few surprising regional results in this HD poll, except for the three-way tie in BC, but then again that's probably due to a small sample.

There are virtually no changes to the national projection:

CON - 151
LIB - 70
BQ - 53
NDP - 34

The Tory national lead decreases slightly, and sits at 11.6%.

Wednesday, March 2, 2011

Ipsos: Tories Lead by 16

The latest Ipsos poll (via puts the Conservatives squarely in majority territory, with 43% of the vote. This is a very strong survey across the board for the Tories, except in Québec where their 19% is still average. For the Grits, this was actually not bad in terms of their support level, and in fact the 27% in Québec is quite encouraging. However, the Tory-Grit gaps, at 12% in Ontario and 19% in Atlantic Canada, would spell electoral disaster for the Liberals.

So where did all the Conservative votes come from? Well, this was a very bad survey for the NDP (13%) and for the Greens (5%), though the latter may be due to a methodology change: as explained by Éric at, Ipsos has opted to stop prompting for the Green Party. My averaging method for the national popular vote has been modified to account for this change.

Not surprisingly, this poll brings the Tories closer to that magic 155-seat mark in the aggregate projection, though they are not there yet:

CON - 152
LIB - 70
BQ - 53
NDP - 33

This is the absolute worst projection for the Liberals since this blog was created in July 2009; it is also just one short of the highest point for the Tories. The average Conservative national lead is at 12.1%.

Also, readers of Éric's blog have probably noticed that he has a new projection model in place. The change in the model gave 4 extra seats to the Conservatives and 9 to the NDP, at the expense of 13 seats for the Liberals. Over the past 1.5 years, Éric's old method has, compared to my projections, consistently disadvantaged the Tories and especially Dippers to the benefit of the Grits. His methodology change makes the seat projections from our two blogs much closer. I feel a little vindicated, but mostly reassured that I was not way off in the left field!

However, it remains true that my methodology depreciates old polls faster than Éric's. This would explain why I peg the Tory lead at 12.1% now compared to 10.4% for