Are soft NDP voters switching to save Liberals?

In my earlier preview of September elections, I noted that the surge in polls for the Conservatives in the Canadian election might lead to the NDP losing votes. It is possible this is happening now.

In recent days, the Liberals have returned to a narrow lead in votes and strengthened their existing seat lead, according to the CBC poll tracker. At the same time, there is a notable dip in NDP votes and seats.

We need to be careful about inferring individual change from aggregate trends. But the most likely cause of what the poll tracker and its seat estimator are picking up is softer NDP voters worried about a Conservative plurality.

Note that this would be strategic voting, but not based on district-level expected outcomes (“coordination”), rather on national-level expectations. “All politics is national”, as Taagepera and I put it in the title of our chapter (10) on predicting district patterns from the Seat Product Model in Votes from Seats.

Here are screen shots from the poll tracker on the morning of 12 Sept.

Votes:

Estimated seats:

The challenge for voters who prefer the NDP over Liberals but are motivated more by stopping the Conservatives is that some of them may be in districts where the Conservatives wouldn’t have won anyway. But some of these voters may help the Liberals win a majority of seats—the poll tracker shows this outcome back within its 95% confidence interval. Yet the sort of voter I am describing wound surely prefer a Liberal minority with a strong NDP third-party caucus.

Getting just the right amount of strategic voting is hard when seats are determined one-by-one, but voters key mostly in national expectations. Yet this is exactly the best available information, which voters tend to employ in choosing voting strategy, according not only to the Seat Product Model, but also Richard Johnston’s The Canadian Party System.

It appears the Liberals have regained their stalled momentum and thus Justin Trudeau just might get what he was seeking after all. On the other hand, a short-term trend need not continue, and at the moment the most likely result still seems to be a Liberal plurality of seats.

Elections in September, 2021–campaigns matter

It won’t be quite like September, 2005, back when the virtual orchard was just a sapling, but somewhat like that September sixteen years ago featured several interesting elections, this month also looks great for election watchers.

In September, 2005, we had elections in three major examples of mixed-member systems: Japan, New Zealand, and Germany. (As I look back, I see I wrote several times about Afghanistan’s election that month; I am guessing there will no longer be a need for new Afghanistan elections plantings. I also see lost of posts about a hurricane disaster in New Orleans. Some things do recur, though fortunately in this case not on as horrific a scale, though bad enough.)

In September, 2021, we have the California recall. I don’t have anything at the moment to say about that beyond what I’ve already said. We have Germany, again, with its general election on 26 Sept. And we have Canada, on 20 Sept. There are also various other elections this month, but these are the two I will focus on here.

The German election for this September has been known about for a long time, as it is occurring on schedule (unlike the one in 2005). The Canadian one, on the other hand, is a snap election as one would not have been due till 2023. Both of these elections are going to become case studies in how campaigns really can matter.

For months it has seemed certain that in Germany, the ruling Christian Democratic/Christian Social “Union” bloc (CDU/CSU) would again come out on top, with the Greens in second place and a likely new coalition partner. Then a funny thing happened: the Social Democratic Party (SPD), for years seemingly finished as a major party, started to surge and seemingly now has pulled ahead of the CDU/CSU. The range of coalition possibilities is suddenly rather large, with some novel possibilities in the cards.

The election is also notable compared to past German elections in that the incumbent Chancellor (prime minister), Angela Merkel (CDU) is not a candidate to remain head of government. In fact, it seems likely that the campaign has caused voters to reckon with the with the less than inspiring leadership of the party’s chosen successor to Merkel (Armin Laschet), to feel less than sure they are ready to make the Greens potentially the largest or even second largest party, and to have turned to the SPD and its new leader, Olaf Scholz as the potential safe pair of hands to lead the government. These three parties seem certain to be the big three, but with the largest still below 25% in polls, it will probably take three parties to forge a majority coalition (taking the Union as one for purposes of government-formation, even though it is actually two parties). Unless all three govern together, or the current no-so-grand coalition clears a majority fo seats and continues in power, only with the SPD on top, it is going to take some combination involving the Free Democrats (FDP) to have a majority coalition. The post-election bargaining will be interesting (FDP with either Greens or SPD is not “natural”), and thus the election will really matter for which combines are possible and how much bargaining power each party has.

In Canada, the Liberal Party of incumbent Prime Minister Justin Trudeau looked safe to be not only the largest party, but also to win a majority of parliamentary seats. In fact, given that the election timing was the government’s choice, that is precisely why it is happening–to convert a minority Liberal government into a majority Liberal government. Then Trudeau called the election and a funny thing happened: his party started sinking in the polls and the Conservatives appear to have caught up. In votes, that is.

As in 2019, the Conservatives could lead in votes and still come second in seats. The Conservatives probably need to win the votes by more than a couple percentage points to have a reasonably good chance at a plurality of seats, due to their inefficient vote distribution across the country. A majority Conservative government probably requires that party to keep adding support at a rapid pace, and that may be happening. Yet if that continues, some voters would probably dessert the New Democratic Party (NDP), currently running at around 20%, in favor of the Liberals. Given the FPTP electoral system, of course, such NDP desertion for the Liberals would not be guaranteed to help the latter at the aggregate seat-winning level, although it probably would do so. (Canada’s Green Party, meanwhile, is in a shambles, largely because its black, female, Jewish leader was not sufficiently anti-Israel for others in the party.) One thing seems safe to say at this point: A Liberal majority has become rather unlikely.

That’s the funny thing about campaigns. Sometimes they actually matter.