Canadian elections are getting more competitive

Tracing 120 years of election results to understand why fewer seats are “safe” today

Before I begin, thank you for reading my Substack. If you’ve been enjoying it so far, please consider taking 60 seconds to forward this email, or click this link to email my subtack to three people you know who might enjoy it. You can also share this post on social media. I’d greatly appreciate it.

This edition is inspired by J.D.M. Stewart’s excellent new book The Prime Ministers, and in particular his chapter on the 1984 election—Brian Mulroney’s sweeping victory. It is the last time any federal leader won a majority of the national vote, and Mulroney’s Progressive Conservatives won roughly 75% of seats—the second-largest seat share in Canadian history after Diefenbaker’s 1958 victory.

2015 was around the time I really got into Canadian politics, and was the first election I followed closely. The 2015 election resulted in a “landslide” majority for Justin Trudeau; his Liberal Party won 54% of the seats in the House of Commons. Reading about Brian Mulroney’s landslide victory, I could not help but think how difficult it would be for any federal party leader to win three-quarters of the ridings in Canada today (or, close to it).

The data backs this up. Six of our last eight federal elections have produced minority Parliaments. No party has crossed the 60% seat threshold in my lifetime.

Intuitively, many Canadians feel this. What is less obvious is how much more competitive our elections have become, at both national and riding levels—and why.

Declining Majorities

As a starting point, I calculated the average share of seats won by the winning party in every federal and provincial election since 1900. The graph below compares the average share of seats won in the three most recent election cycles to the historical average before them.

Across almost every jurisdiction, the winning party today captures a meaningfully smaller share of available seats.

In federal elections since 1900 the winners averaged 55% of seats before 2015, but just 48% in the three most recent elections.

At the provincial level, the declines are similarly pronounced:

• BC: from 65% historically → 54% in the three most recent provincial elections

• Alberta: from 78% → 64%

• Quebec: from 72% → 62%

• Atlantic Canada: average decline of 14 points

Only two provinces buck the trend: Saskatchewan and Manitoba. Ontario has uniquely remained flat, largely because Doug Ford’s elections have been decisive outliers in an otherwise competitive period.

Looking at the decade-by-decade averages federally and in Canada’s four largest provinces, the pattern is even clearer.

Nationally and in Canada’s largest provinces, the trend is quite linear; winners are gradually winning fewer seats.

Riding Fights

Even beyond aggregate seat totals, the competitiveness inside ridings has increased dramatically*.

*If this is as counterintuitive to you as it was to me, I provide a possible explanation in the conclusion.

To dig into riding levels results, I scoured the internet for riding level results of three historical elections that are comparable to the 2025 federal election.

The elections I chose as comparisons for 2025 were 2008, 1997, and 1979. I wanted to understand how riding-level outcomes differed historically in elections where the winning party captured a similar share of seats. In 2025, the Liberals won 49% of seats; the winning parties in 2008, 1997, and 1979 won 46%, 51%, and 48% respectively. These elections therefore provide appropriate comparisons, and importantly, they are spaced apart by roughly equal intervals.

The histogram below shows the share of ridings in each election decided by margins smaller than the value on the y-axis.

Consider ridings decided by less than a 10-point margin (the top bar):

• 1979: 29%

• 1997: 24%

• 2008: 26%

• 2025: 36%

Almost more remarkably, the number of blowouts has collapsed. In 1979, 33% of ridings were won by margins of at least 40% (the sum of the bottom four bars). In 2025, that figure fell to just 22%.

The below graph further exemplifies this trend; the average margin of riding level victory has fallen from 24% in 1979 to 20% in 2025, and the median has fallen from 20% to 15%.

The Why

There are several possible explanations. Two of the most intuitive—third-party growth and declining turnout—appear necessary but not sufficient (politicos may get this extremely inside baseball joke).

It stands to reason that more parties winning meaningful chunks of the vote would lower overall seat counts for the winner. And there has indeed been an increase in the share of seats won by parties finishing outside the top two. However, the increase has not been uniform; it varies by region and by decade:

A rise in third/fourth/fifth-party strength exists, but it isn’t consistent enough to fully explain the tightening margins.

A variable that does appear to at least somewhat account for the increased competitiveness of our elections, at least federally, is turnout.

Lower turnout may explain why the closeness of our elections has risen, but the correlation is not strong enough to be the only major factor.

The More Interesting Whys

Increased partisanship is likely a key driver in the increased competitiveness of our elections. As politics becomes more tribal, fewer Canadians view multiple parties as palatable options, and campaigns increasingly operate over a shrinking pool of swing voters. With soft supporters growing more limited, and turnout declining, campaigns have shifted toward mobilizing core constituents rather than trying to build broad, cross-party appeal. The strategic emphasis today is often less about persuading swing voters and more about maximizing turnout among core supporters.

Returning to the broader question, one could easily assume that elections are becoming tighter because there are fewer “winnable” ridings for each party; I think the opposite may be true. The decline in riding-level margins suggests that parties are competing in more seats than ever before. A fragmented media environment, the ease of cross-country travel, and most importantly, the ability to micro-target advertising, mean campaigns can now tailor messages to distinct sub-regions, demographics, or communities. The campaign that resonates in downtown Vancouver no longer needs to be identical to the one that works in suburban Mississauga or rural Nova Scotia.

Even further, campaigns used to target messages by geography, out of necessity, and the campaign with the most broadly appealing message won. Now, messages are targeted by identity variables (i.e.: what TV channel you watch, what car you drive, your level of education). Ridings are no longer treated as monoliths, and campaigns are ignoring fewer and fewer ridings. As sophisticated targeting and optimization deepens, more ridings become viable targets, creating tighter elections across the full system—not just nationally, but within individual ridings.

Taken together, competitiveness may be rising because while partisanship is increasing and turnout is dropping, campaigns are becoming more sophisticated at finding and mobilizing voters in every riding across the country.

Tell me what I’m missing in the comments!