When I wrote this based on Johnson winning a November general election, someone asked me whether it was a prediction or a warning. It was both. I still think it is the most likely outcome..Here are my thoughts that led me to that conclusion, which of course may be completely wrong.
Putting Brexit to one side, Johnson needs a larger majority to govern. So an election sometime in 2019 seems very likely. It is also clear he wants this to be a ‘people versus parliament’ election, where of course Johnson represents the people and parliament ‘is colluding with the EU’ to block No Deal. He has hit the ground running with various popular measures. So the key question is when in 2019 the election will be.
There seem to be three possibilities.
An election before 31st October
An election announced before 31st October that takes place after that date
An election in November or December
An important factor governing this choice is that Johnson will avoid an election after we crash out with no deal, unless he is completely deluded about what crashing out means. Even if he blames the EU for everything that goes wrong, voters will not take kindly to a government that played down the consequences of crashing out when crashing out turns out to have severe consequences.
This means that if he intends to leave with no deal, his only choice is (1). But he needs a pretext, and that pretext would have to be MPs instructing the PM to get an extension if he fails to agree a deal. He could ask parliament on 19th September to call an election to be held on 24th October, for example, but that would rely on parliament acting very quickly to rule out crashing out. Johnson also misses out on a crowd pleasing budget. It would be odd to base a complete strategy on parliament moving quickly, but nevertheless I think it has to be the second most likely outcome, particularly if he thinks leaving by 31st October is critical in suppressing the Brexit Party vote. .
Leaving this possibility to one side, we can rule out Johnson ignoring an instruction from parliament to extend the A50 deadline, because it involves holding an election during the chaos of crashing out. Which suggests all the talk of governments of national unity is a red herring, but for rather different reasons than Stephen Bush gives here. Labour may still call for a vote of no confidence (VONC), but Tory rebels are likely to vote it down while they pursue other avenues to stop no deal. Corbyn's choice about whether to call a VONC is lose/lose: if he doesn't he gets blamed and if he does and fails it will be spun as Tory MPs not wanting to support a Corbyn led government.
If we rule out the improbable early election option (1), and we also assume Johnson will not want to hold an election after we crash out, then this means Johnson will accept an instruction from parliament to obtain an extension before 31st October. That would be obtained on the assurance of an imminent general election, and the EU would almost certainly accept this. Johnson gets his people versus parliament election.
This has the disadvantage, of course, that Johnson will have failed to keep his promise of leaving by 31st October, which risks a revival of the Brexit Party. He could respond that he has been forced to do so by an instruction from parliament which it would be irresponsible of him to ignore. His election campaign would be that the EU would have offered an alternative to the backstop, but this was undermined by MPs ruling No Deal out.
As calling a general election before 31st October to be held after that date (option 2) mixes the task of campaigning with having to ask the EU for an extension, a November election seems the most likely outcome, although far from certain.
There are another set of possibilities where Johnson really wants a deal, which is based - as Aleks Eror suggests - on something like a backstop for Northern Ireland alone. This is what the EU originally suggested of course, but no doubt Johnson could dress it up as a result of his tough talking, and much of the press would back him up. This leads to the same place. Parliament would reject the deal with the ERG and DUP in the lead, but he could fight an election on this basis and use a victory as a mandate.
If there is a November election, a Johnson victory seems the most likely option, although again far from certain. While Johnson taking over as PM has squeezed the Brexit Party’s support, Labour has so far only received a small recovery in its vote since the European Elections and the Liberal Democrats remain strong. Of course that may change before any election, but I would say the balance of probability is that this will fail to stop many Remainers in key Lab/Con marginals from voting LibDem or Green rather than Labour.
The irony of all this, if I'm correct, is that Johnson is looking to his rebel MPs to stop a no deal Brexit. A failure to do so, if this was accompanied by a failure to negotiate a new deal with the EU, would require either fighting an election after the chaos of crashing out, or Johnson choosing rather than being instructed to get an extension, as May did in March. If that happened then Johnson would suffer the same fate as May, and the Brexit Party would revive. Only in those circumstances might we not see an election in 2019.