Just a follow-up to yesterday’s post on whether Saanich-Gulf Islands is the best choice for Elizabeth May.

The 2008 race in Saanich, of course, was complicated by the fact that NDP candidate Julian West dropped out due to controversy (but too late to remove his name from the ballot). So Dippers voted in large numbers for Liberal Briony Penn. Greens also voted in large numbers for Penn, both because it was clear once West dropped out that Penn was the main challenger, but also because Penn herself was a former Green with a following, and probably in part due to the Liberal/Green (Dion/May) agreement and May’s comments that appeared to support strategic voting. Due to these many factors, it’s clear that the 2008 results were heavily skewed toward Penn and the Liberals in a way they would not be under normal circumstances. And, of course, despite all the stars seemingly aligned for Penn, CPC incumbent Gary Lunn still won by 4 points.

What’s important to note is that so go the Conservatives in B.C., so goes Lunn — that his results are typically on par with the general trend in B.C. — 2004 (36.2% CPC vs 34.6% Lunn), 2006 (37.3% CPC vs 37.2% Lunn), 2008 (44.4% CPC vs 43.4% Lunn). Typically, in Saanich, the NDP and Liberals are a couple points below their provincial average (for example, in 2006, 27.6% LIB vs 26.1% Saanich and 28.5% NDP vs 26.5% Saanich), while we might expect the Greens to almost double their provincial average (for example, in 2006, 5.3% GPC vs 9.9% Saanich). This trend is also observed in 2004, although it was slightly more skewed for the Greens since they ran a fully-funded campaign for the first time in the history of the party — we might have expected the Greens to win around 12-13% based on party support at the time but they won almost 17%; as a result, Lunn and the Liberals each won about 1% less than might otherwise be expected and the NDP about 3% less (which shows that the Greens draw from all 3 parties, but the NDP most heavily).

Here is a comparison of recent results, plus my adjusted (projected) had things followed the typical pattern, as well as current projections, both on a typical basis and making adjustments for May to give a best-case scenario.

Lunn won about 800-900 more votes than we might otherwise expect (no doubt NDPers or Greens who shifted, but towards Lunn not Penn). Another 1,300 or so voters stayed home (probably mostly NDPers, since their candidate dropped out). So we might reasonably estimate that roughly 10,000 NDPers and 4,000 Greens voted for Penn. This is of course speculative and could be verified with detailed polling but are reasonable assumptions. So we might expect that that the Greens would have received 16-17% (similar to 2004) under normal circumstances.

So what’s the current situation? An aggregate of recent polling data shows roughly the following in B.C.: CPC 35.7%, LIB 28.3%, NDP 22.3%, GPC 12.9%. Of course, we need to make some minor adjustments to this based on the differences we typically see between polling and voting patterns (this reflects the strength of GOTV operations, tendencies of some demographics to vote more than others, strategic voting, last-minute shifts, etc). Using the same methodology, here’s a comparison of the final aggregate polling averages vs. actual results: for 2008 — polling 38.4% CPC vs. 44.4% actual (+6.1); polling 23.3% LIB vs 19.3% actual (-4.0); polling 25.3% NDP vs 26.1% actual (+0.8); polling 11.9 GPC vs. 9.4% actual (-2.5), and for 2006 — polling 36.4% CPC vs 37.3% (+0.9); polling 30.3% LIB vs. 27.6% actual (-2.7); polling 26.7% NDP vs. 28.5% actual (+1.8); 6.0% GPC vs. 5.3% actual (-0.7). No doubt 2008 was a bit more favourable for the CPC than normal and that the 2006 shifts are probably more in line with expectations, but the recent pattern is clear: the CPC goes up at the expense of the LIBs and the NDP goes at the expense of the GPC. So making adjustments, we might estimate current support in B.C. at: CPC 38.7%, LIB 25.3%, NDP 24.8%, GPC 10.4%.

Using this estimate and running the projection, we might estimate current support in Saanich at: CPC 37%, LIB 22%, NDP 22%, GPC 18%. So the Greens with a typical candidate might need to close a gap of about 12,500 or so votes.

But we should expect May to deliver personal votes. A reasonable comparison might be to look at what Ed Broadbent did in Ottawa Centre in 2004 — he delivered a 8% premium above and beyond what we might otherwise have expected using standard projections. While a candidate can deliver a much higher premium in a by-election (as May did in London North Centre), since the government is not on the line and more votes are personal votes, but not in a general election. So if we give May the benefit of the doubt that she can draw as many personal votes in Saanich as Broadbent did in Ottawa Centre, then May could deliver an 8-point bump of personal votes. This also includes the benefit of spending to the maximum.

The Greens tend to draw more heavily from the NDP, and very little from the CPC (this was confirmed by an analysis I had commissioned when I was GPC National Campaign Chair). So we might estimate an 8-point bump for May would result in a 5-point drop for the NDP, 2-point drop for the Liberals and 1-point drop for the CPC. Again, detailed polling can give us precise numbers for Saanich, but this is a reasonable estimate. So, the current absolute best case scenario for the Greens, assuming a fully-funded campaign and May able to deliver personal votes is: CPC 36%, GPC 26%, LIB 20%, NDP 17%. Not bad, but even under a best-case scenario, there is still a gap of about 6,500 votes. And typically, Green support drops as election day approaches, so baseline Green support here is probably over-estimated. To close the gap will require an strong local operation, that hits the ground running, and builds upon work already done. Unfortunately, as I previously mentioned, the Greens currently don’t have a robust ground operation in Saanich and haven’t extensively canvassed the riding (unlike Guelph).

The above is, of course, speculative with plenty of assumptions, but it gives you a pretty good idea of where the Greens stand in Saanich-Gulf Islands and the magnitude of the challenge ahead if they hope to unseat Gary Lunn.

Update: respected University of Victoria political scientist Richard Pilon tells the Tyee: “I’m not sure that Saanich-Gulf Islands is the best choice for Elizabeth May.”