How the Tories might end up with just 25 seats in a Canada-style wipeout

The latest opinion polls put the Conservatives on a record low of 20 per cent, warns John Rentoul, raising the spectre of the Canadian Conservatives who turned a majority into just two seats…

Call an election in May or keep digging? Rishi Sunak is less popular now than John Major was ahead of his 1997 defeat (PA)

The 1993 Canadian federal election used to be an obscure quiz question. Now it is a vivid nightmare feared by Conservative MPs, who are all experts in what happened then.

The Conservative government led by Kim Campbell, a prime minister with a majority in parliament, went down to such a catastrophic defeat that it was left with just two seats. There are uncanny similarities with Britain today.

Brian Mulroney, a strong character, had stood down as prime minister when it became clear that he was heading for defeat. He was replaced by Campbell, Canada’s first woman prime minister, who briefly seemed to be a better prospect. She was in office for only five months, putting off the election as long as the law allowed, but she lost popularity during the campaign. The Conservatives were overtaken by an insurgent right-wing party called Reform and suffered from the distorting effect of the first-past-the-post voting system.

Could it happen here? It would require a lot of worse things to happen to Rishi Sunak for a result that bad, but it is no longer outlandish to suggest that the Tories could end up with a result that is worse than their worst defeat since the universal franchise, which was when they won 165 seats in 1997.

It is quite difficult, however, to enter figures for shares of the vote in online models to produce a result as low for the Conservatives as two seats. If you feed the figures from this week’s Ipsos poll, which put the Tories on a record low of 20 per cent, into the Electoral Calculus model, the Tories end up with 25 seats. On the other hand, the model made by Professor Ben Ansell of Nuffield College, Oxford, gives the Tories 99 seats on those assumptions.

Even if you put in a 16 per cent share of the vote, which is what the Canadian Conservatives got in 1993, and assume that all the lost Tory votes go to Labour, Electoral Calculus produces six Tory seats (Ansell says 32).

Of course, the Ipsos result is just one poll, the most extreme of a range of estimates produced by the polling industry. But the other opinion polls are hardly reassuring for the prime minister. If we take the average as the barometer of public opinion on the eve of the Budget, Labour is still 20 points ahead and the Conservatives are heading for a drubbing about as bad as they suffered at the hands of Tony Blair.

The average Labour lead has hardly changed since Sunak became prime minister, but the gap has widened slightly rather than closed, and there is no evidence that another tax cut in the Budget will make a difference.

The prime minister hopes, as Kim Campbell did in Canada, that holding the election as late as possible, or nearly as late as possible, will give people time to feel better off. But the 2p cut in the rate of national insurance in January doesn’t seem to have had any effect. Maybe a second cut, on top of earnings rising faster than inflation for several months, will do the trick.

At some point, it seems likely that the media narrative will turn, and the opinion polls will start to narrow a bit, in which case we may briefly pay as much attention to the most extreme polls in the Tories’ favour as we did to Ipsos this week.

But such a recovery cannot be guaranteed. Some people are still coming off fixed-rate mortgages to find themselves paying substantially more, while most voters still tell pollsters that they feel worse off than they were at the last election, and worse off over the whole of the period of Conservative-led government – even if the official statistics say that disposable income per person went up under David Cameron as prime minister and has recovered from the lockdown recession since then.

In fact, it seems just as likely that things will keep getting worse for Sunak and the Tory party – hence the feverish talk, encouraged by Labour but arising spontaneously among Tory MPs, about a May election.

There are more polling records to break, and the laws of random variation mean that they are likely to be broken. The Ipsos poll recorded the lowest share of the Tory vote since its series began in 1978, but its 27-point Labour lead was not the biggest – and it is the size of the gap that really decides elections. The average Labour lead in opinion polls with fieldwork in the week in which Liz Truss left office was 31 points.

Nor is Sunak yet as unpopular as John Major was in 1994 – although Major recovered a little by the time of the 1997 election, and Sunak is less popular now than Major was as he prepared to be swept away by Blair’s landslide.

This week’s Ipsos poll may be an outlier, but as such it could still be a harbinger of bad news for the Tories to come. And the warning from Canada in 1993 may be an extreme case of the possible effects of first-past-the-post voting, but something like it could easily happen here.