Explaining the 2011 Federal Election III: The Money and the Metrics
[Welcome, National Newswatch readers!]
Just before the last federal election, we looked at a series of baseline performance metrics for the parties as they prepared to do battle. Now that the full datasets are finally in, we can come back and score exactly how each one did.
What with the end of May deadline for riding associations to file their 2011 returns, followed by the party annual returns for 2011 at the end of June, and most candidate returns for the 2011 election now filed, if not yet reviewed, it's been a busy time for your friendly data-entering blogger.
Here's a report on what datasets have been added to the site's database in the past few weeks, and how they help us understand the organizational back-story of the May 2011 federal election.
2011 GE Central Spending and Rebates
Elections Canada has posted the amounts of the rebates (aka "reimbursements") paid to the headquarters of the registered parties who qualified for one, and I've now entered them into the Pundits' Guide database, visible on the Parties page if you pick either a party or an electoral event.
To recap: a registered party is eligible for a reimbursement of 50% of its election expenses (though only the paid ones, so excluding those based on market values of donated goods and services, or other non-monetary amounts), if it achieves at least 2% of the valid ballots cast nationally, or 5% of the valid ballots cast in the ridings in which it fielded candidates (the rule that makes it possible for regional parties to be reimbursed). These are the same threshold criteria as apply for receiving the quarterly per-vote subsidy as long as it's alive.
Rebates can't be paid until Elections Canada has reviewed a party's Election Return, so we can assume they've also reviewed and verified those returns (to the extent they're allowed to under the law as currently written), and we can therefore report those numbers as well.
Here are all those values for the 2011 GE, in order of rebate size.
Registered Party Expense Limits, Election Expenses, Paid Expenses and Central Rebate, 2011 General Election
|* if different from Total Election Expenses
** Unknown. The Chief Electoral Officer must have revised their submitted expenditures upwards slightly, as the BQ reported paid election expenses of $5,339,922.89 + a non-monetary discount of 3,138.00 for a total of 5,343,060.89 on its election return. But this is less than the $5,344,678 in paid election expenses Elections Canada reports having paid a rebate against, and which is also reported on the Bloc's 2011 financial statements. Anyways, the difference is less than $2,000.
Unlike 2008, when the NDP said that it had the capacity to spend the limit if it wanted to, but then didn't, in 2011 it actually did spend close to the maximum. Only the Bloc Québécois spent a higher percent of its limit, but then they had a fight on their hands this time — from the NDP.
And unlike 2008, when the Liberals pulled back spending severely in the final weeks of the campaign, finishing up at 73% of the limit, this time they were back to a fully-funded campaign, on a par with the Conservatives (who themselves have learned the hard way to leave a bit of wiggle room), at 93%.
And, finally, unlike 2008, when the Green Party spent some 14% of the national limit on their central campaign, this time they managed only 9.3%, a drop of some $870K mostly shaved off the advertising budget.
2011 GE Candidate Spending
Next up are the candidate returns. The 1,587 candidates (1,526 endorsed by the registered parties plus 61 non-aligned and independents) had four months from Election Day to file their financial returns or ask for an extension, which given most of that time straddled the summer months last year, 651 of the 1,587 did.
As of today, 1,564 of the 1,587 candidates have returns posted on the Elections Canada website, of which 1055 have already been reviewed by the Election Finance group (that's 1055 down, 509 to go, fellas; keep plugging away!). Here's how they break down by party, in order of return filing completeness.
Candidate Financial Returns by Posted Status and Reviewed Status (as of July 8, 2012), 2011 General Election
|Cand $$ Returns||Posted|
We can see that there's nearly a full set of returns from each party's candidates, and that just over two-thirds of them have been reviewed by Elections Canada. Many are NIL returns, however, and many others show very small amounts of spending as we'll see in a second, so some of that reviewing work has been less onerous than in other cases.
The unsubmitted returns are from Conservative Ronald Leung in Burnaby-Douglas, BC, New Democrats Dave Nickle in Peterborough, ON and Mary Trapani Hynes in Don Valley East, ON, three Green candidates from Québec, and nine mostly name-on-ballot Liberals (1-NS 4-QC, 3-AB, 1-BC) but including the star recruited candidate Professor Stephen Randall in Calgary Centre-North, AB and another university professor for whom the party had high hopes, William Hogg in Compton-Stanstead, QC.
Remembering that just before the election, we used the threshold of 50% spent on a local campaign as a hint of how many seats a party was targetting, now that we have most of the candidates' returns filed, we can revisit that metric to see what kind of ground organization the parties had in place.
Posted Candidate Financial Returns by Percent of the Local Candidate Expense Limit Reportedly Spent (as of July 8, 2012), 2011 General Election
|Number of Party Candidates Spending
What % of the Local Expense Limit (2011 GE)
|0%||< 10%||>= 10%||>= 25%||>= 50%||>= 75%||>= 90%|
This table shows the number of candidates by level of spending. So, for example, of the 306 Conservative Party candidates who have filed a return, we see that 14 of them spent under 10% of their local candidate expense limit (including 1 candidate who reported no spending at all), while 292 Conservative candidates spent 10% of the limit or greater, including a full 102 of them who spent 90% of the limit or better. 228 of them spent 50% or better (down slightly from 246 in 2008; not shown).
The data show clearly just how targetted the NDP's effort was. The party had a short-term goal of getting riding associations to raise more of their election requirements in the pre-election period through the "Local Victories Challenge", and a longer-term goal of encouraging those ridings to spend a greater proportion of the limit. It partially met the first goal (80 ridings transferred $20K or more to their candidates, up from 68 in 2008; data not shown), but not the second (the party's candidates combined again spent just 25% of their combined ceilings; meaning that $21M was again left on the table, presuming it could be raised; see below).
Of the 147 NDP candidates who spent less than 10% of their limit, to no-one's surprise 60 of them were in Québec including many of those who were subsequently elected, with another 13 of them in then-no-hoper ridings in Atlantic Canada, 44 in Ontario, 20 in Alberta, and 20 others elsewhere across the west.
The 14 low-spending Conservative candidates were also in Québec, while the 33 low-spending Liberals were split evenly between Québec and mainly rural western ridings (plus Saskatoon-Rosetown-Biggar, SK).
The data also demonstrate starkly just how targetted the Green Party's efforts were. Only one riding besides Elizabeth May's seat of Saanich-Gulf Island, BC put on a fully-funded campaign (deputy leader Adriane Carr, who in fact outspent May 98.6% to 92.8% during the campaign period in return for 4th place and 15.4% of the vote in her riding of Vancouver Centre BC), and just 17 other candidates managed more than a token effort, at 25% of the limit or higher (12-ON, 2-BC, 2-AB, 1-YT).
The Bloc showed a pattern of greater spending in more seats in 2011 than 2008, another fact supporting the contention that they fully realized their vote to be under threat of erosion from the NDP (even if not the full extent of the resulting wave). The Liberals also maintained local spending at 2008 levels, though down from when they last won government in 2004.
[Indeed, while the first obvious correlation to try and make against candidate spending is candidate vote-share, I'm not sure that's the best predictor any more. I'll test some other variables, but I bet we find a better correlation with other concepts that get at the degree of competitiveness, or the extent of a perceived threat, and/or the perception of the potential for gains or losses. More work than I have time for today, but definitely some concepts worth tagging for later exploration.]
2011 GE Outcomes and the Changing Contest
What kind of success did that money buy them? Here are the updated graphs on vote outcomes from last year.
First, the total number of votes received, a metric that David Akin argued last spring represented a "structural trend" in the NDP's favour relative to the Liberals:
Second, we can look at the number of ridings in which each party obtained 30% of the vote or better, a rough indicator of the number of seats in which it was competitive in a multi-party, first-past-the-post system (though note that a party's riding vote-share in one election can represent a ceiling on its riding vote-share in the next one, if it's already in broad decline).
Let's look at the distribution in more detail:
|# Cls 3rds||4||4||5||8||1||1||1||1||4||9|
|# < 10%||37||91||65||2||261||297||4||10||7||23|
|# >= 10%||270||217||243||306||41||7||71||65||300||284|
|# >= 15%||226||159||144||267||9||3||67||59||280||259|
|# >= 20%||183||119||82||216||3||1||66||50||246||245|
|# >= 30%||115||77||54||141||1||1||54||22||197||213|
|# >= 40%||69||22||33||93||0||1||38||0||134||161|
|# >= 50%||17||2||7||36||0||0||13||0||80||107|
By this criterion:
- the Liberals were competitive in 77 / 308 seats in 2011 (18 in the Atlantic provinces, 10 in Montréal, 4 in Ottawa plus another 2 in Eastern Ontario, 27 in Greater Toronto, 5 in Southwest Ontario, 1 in Northern Ontario, and 10 out West)
- the Bloc were competitive in 22 / 75 Québec seats (5 in the east end of Montréal, another 3 in the Montréal suburbs, 3 each in the Gaspésie and northern Québec, and a smattering of 8 other ridings throughout the province)
- the Green Party was only competitive in its leader's seat, and only significantly influenced the outcome in 2 other ridings (Yukon at 18.9% and Vancouver Centre at 15.4%)
In contrast, both the NDP and the Conservatives were much more broadly competitive across the country, the NDP in just under half the ridings and the Conservatives in just over a two-thirds.
But were they competitive with each other? To answer that, we need to look at the changing contest.
[Click on image to open a full-sized version in a new window]
The largest number of ridings in 2011 were Conservative-NDP / NDP-Conservative contests (148/308), which represented an increase in that category over 2008 when there were 71 such contests (a figure that had been growing over several elections). The Conservatives traded Surrey North, BC with the NDP in return for Sault Ste. Marie, ON and Elmwood-Transcona, MB; and otherwise the NDP replaced the Liberals in 48 other seats, 4 of which it picked up (Pontiac, QC which had been a Cons-Lib contest, along with Lib-Cons races in the Rouge River & Southwest seats in Scarborough, ON and Esquimalt-Juan de Fuca, BC), while the Conservatives replaced the Liberals in a further 13 (the core urban seats of St. John's East, NL, Ottawa Centre & Hamilton East-Stoney Creek in ON, and Vancouvers East & Kingsway in BC, along with Acadie-Bathurst, NB, 6 Northern Ontario ridings, and Churchill, MB).
On the other hand, the largest number of ridings in 2008 had been Conservative-Liberal / Liberal-Conservative contests (134/308). By 2011, that number had continued to shrink to just 79. The
6 [UPDATE: 5] additions to the group comprised 4 Atlantic seats not targetted by the NDP (Labrador, Humber-St. Barbe-Baie Verte, and Random-Burin-St. Georges in NL, and Sydney-Victoria on Cape Breton; along with Halifax West, where the NDP placed a very close 3rd , and Kingston & the Islands in ON, which had been heavily targetted by the Conservatives). [a slip of the finger in my dataset mis-characterized Kingston in 2008; sorry]
Meanwhile, an insignificant group of ridings in 2008, the NDP-BQ / BQ-NDP contests, suddenly became the third-largest grouping in 2011, accounting for 45 ridings rather than the lone 3 in 2008 (Gatineau, Repentigny, and Rivière-du-Nord, QC).
NDP-Liberal / Liberal-NDP contests represented 30 ridings again in 2011, though most of them were not the same as those from the same category in 2008. The NDP picked up a further 6 seats from the Liberals in this group (St. John's South-Mt. Pearl and Dartmouth-Cole Harbour down east, and their 4 south Toronto pickups), along with 6 other former BQ-Lib / Lib-BQ ridings, and they took over from the Bloc as second-fiddle in a further 2 seats (Bourassa and Papineau).
Of course, the Conservatives' majority hung on the 25 Liberal-Conservative contests from 2008 that it flipped the other way in 2011, of which 20 were found in Ontario, along with 2 in New Brunswick, and 1 each from Manitoba, BC and Yukon.
In a climate where the Conservatives seem to be dropping in support, and the NDP apparently gaining, those are the kind of seats that offer the Liberal Party its best first step on the road back. And luckily for the Liberals, the upcoming two and possibly three by-elections will be taking place in seats that have been Conservative-Liberal contests and are likely to stay that way in the foreseeable future (
Durham, Etobicoke Centre, and Calgary Centre – though the NDP has dipped its toe into that one too to test the waters with a workshop there on "uniting the progressives" hosted by Nathan Cullen this Wednesday).
For the Conservatives, they cannot safely assume that their opponent last time in a seat will be their strongest competitor next time. But in the 25 seats that gave them their majority last time, fewer fit the profile of the kind of seat the NDP can leap over the Liberals in. Which means that the old battle royale between the two old-line parties is not set for the dustbin of history just yet.
UPDATE: As I would have seen, if I'd taken a second to double-check my assumption, the NDP came second in Durham, edging the Liberals 21 to 18 in a campaign with minimal resources; not that the Liberals exactly spent the limit either, though they ran the former Chief Operating Officer of the Toronto Board of Trade, Grant Humes, as their candidate last time. An alert regular reader has drawn our attention to this boo-boo, but I wonder if my mistake doesn't point to another hunch worth considering for a second.
Eyeballing the chart for Durham, it's a pretty reasonable assumption that the Conservatives gained there at the expense of the Liberals over the long-term, while the short-term collapse of the Greens put the NDP temporarily ahead of the Libs. My commenter would have to agree that Durham would not normally have the geographic or demographic profile of the kind of ridng the NDP would target. Provincially the riding shows the more typical set of results. It seems to me that if the NDP wanted to pick a seat where it could take over from the Liberals as the natural alternative to the governing Conservatives, this riding would not figure on the top of anyone's list in Ontario. Federally, the Conservatives won every poll except the Mobile Polls, which it split with the NDP (one loss, one tie). It bears watching to see if some dramatic moves in candidate search can change my perception on this one, though.