The end of Liz Truss
...is probably nigh
It is now approaching a habit for me to sit down after a new British prime minister or American president takes office and get some idea of how long they are likely to last in the job.
As per Philip Thetlock’s Superforecasting method, this entails checking over the track record of previous office holders to figure out the base rate for survival – usually until the next election. And yet in the case of Liz Truss, the as of writing current British prime minister, the task has looked increasingly superfluous.
On the Tuesday that this letter is published, it will be Truss’s 43rd day in office. It may even be that by the time you read it her resignation has been submitted. If so, she will be the shortest serving British premier, behind even George Canning who was in office for 119 days, if you include the last one on which he snuffed it.
I began drafting this during the Conservative conference in Birmingham. “Apprentice losers’ cafe vibes,” was the verdict of pundit Jordan Tyldesley, and it is fair to say the vibe hasn’t improved since, even with the oddly jovial Jeremy Hunt replacing Kwasi Kwarteng as chancellor.
And the polls. Right now Britain Predict’s modelling has the Labour party winning 509 MPs out of the 650 that sit in the House of Commons. At least one poll implies that the Liberal Democrats could become the official opposition, with the Tories relegated to 22 MPs, behind both the Libs and Scottish National Party.
Such an outcome is unlikely, I think. For one, and as ex-pollster Martin Boon notes, there is the potential for a shy Tory effect here. Replacing Truss with an leader who is even slightly plausible will likely recover a few despairing Conservative voters, especially those who viscerally hate Labour.
The wider reason is one that you’ll get bored of hearing if I keep doing these: remarkable events are unlikely by definition. If somebody tells you that the world’s longest standing party of government is about to die, you should assume this is unlikely based on the information just given to you.
All that said, Truss’s exit this side of Christmas feels inevitable to anyone reading the news right now. Sam Freedman, a former civil servant who has a good Substack, summarised the sense of things when he tweeted that Truss would “be gone in days”1.
Similarly vague predictions waft through newspapers and various Twitter feeds, much as they did for most of Boris Johnson’s time as prime minister.
Indeed, one can compare Truss’s predicament directly to Johnson’s back in September 2019, when he faced a gridlocked parliament with no obvious route to get Brexit done. The New Statesman’s Jonn Elledge said it was “plausible” that Johnson could become the shortest-serving British prime minister – “perhaps not probable, but certainly not crazily unlikely”.
In the end, Johnson soldiered on for 3 years and 44 days. This is close to the median for British prime ministers, as evinced by this handy Wikipedia list. Even among prime ministers after the Second World War, there are four prime ministers who bowed out in a shorter period.
In the end Johnson’s premiership was unusual for its chaos, or at least the public nature of said chaos, but in terms of length it was pretty average.
Not that it would have been especially unusual for Johnson not to fight the next general election once in office. Post-war prime ministers have bowed out before the next general election around a third of the time they might have fought it, as I noted back in May 20212.
My normal starting point for these forecasts would be the inverse of that stat: around two-thirds of the time, the sitting prime minister will fight the next general election. I would have stuck close to that number when Truss took office had I published a forecast.
And yet to look at the press coverage, the polls and Truss’s public performances, the main thing keeping her in office is seemingly the sheer ridiculousness of replacing a prime minister appointed only six weeks ago. There is simply no way that there is a 67% chance of her staying in office until the next general election.
And yet, I don’t think a swift exit is inevitable either. Yes her prospects look terminal. The plotting among backbenchers is so blatant it could not fairly be described as a conspiracy. She has no authority.
But trumping all this, whoever replaces her will probably lose the next election – hardly an enticing prospect for the ambitious types who want to become prime minister.
It is also worth remembering the contingency that led to Johnson being ousted: a government whip getting blasted and allegedly groping some men. Had Chris Pincher caught Covid the day before his big night out, Johnson could well be in office still.
For Truss’s part, she gives every sign of wanting to keep going. Sacking Kwarteng would have been irrational if she was planning to resign imminently. And she just told the BBC she wishes to contest the general election.
Current betting odds give Truss only a 10% chance of leading the Tories into that. (Although there is a bet with implied probability of 95% that she won’t.)
I suspect this underestimates her odds of continuing. One of Tetlock’s key forecasting insights was that the base rate (historical trend) was a better guide than insider chatter, which ultimately is what a lot of the above is.
Contrasting this, I wonder whether there has been an increase in leadership volatility in the past decade in British politics (or indeed Westminster system politics more broadly). Turnover has certainly been high of late, but it feels premature to declare this a trend rather than a short-term Brexit effect.
Opinions will vary on how to synthesise all the above. I suspect that Truss has a rather better chance of lingering onto the next general election than current events suggest: something in the range of 20-30%. Call it 25% if you want a single figure.
Put another way, I think it about three times more likely that she’ll be replaced. What I’d happily wager is that Labour will win the next general election outright. It is a good time to be Keir Starmer.
Taken literally this is a non-prediction prediction, but I’m sure he meant the rest of her tenure wouldn’t be measured in weeks.
With Johnson having not fought the next general election, it’s now 26 times out of 40 that a prime minister has fought the next general election since the Second World War, or 65% of occasions.