There is no special place in the British electoral system for the party leaders. They are candidates only in their constituencies, and in a technical sense one cannot vote for (or against) Boris Johnson unless one is an elector in the constituency of Uxbridge and South Ruislip, and likewise nobody outside Islington North can pass a direct verdict on Jeremy Corbyn. A leader, therefore, can be defeated in their own constituency whatever happens in the rest of the country. Theoretically, their party might win a majority despite the leader’s personal loss, which would create an uncomfortable situation. Constitutionally, the Prime Minister does not have to be an MP, although even in this age in which norms are flouted there is surely a binding expectation that he or she should be in the elected House. The last time that this happened, outside a general election campaign when parliament is in a state of dissolution, was in 1963 when newly appointed Prime Minister Sir Alec Douglas Home was campaigning to win a seat in the Kinross and West Perthshire by-election.

Theresa May is the only new party leader to get a worse result than her party managed nationally

However, this vulnerability is an illusion. Personal defeat hardly ever happens to Prime Ministers and Leaders of the Opposition, even if their parties are badly beaten. The last time a Leader of the Opposition lost his seat was in 1931 when Arthur Henderson lost Burnley; no incumbent British Prime Minister has ever lost his own seat but it happened to John Howard in Australia in 2007. The closest call was in January 1906 when Arthur Balfour, the Conservative leader who had stepped down as Prime Minister the previous month, was defeated in the Manchester East seat he had held since 1885. He resurfaced a few weeks later apparently unchastened in a by-election called for this purpose for the City of London constituency, a Tory sinecure.

Some of the reasons for the leaders’ good electoral fortune are structural. Usually, achieving a leadership position entails long continuous service as an MP, or else a period as an MP for a marginal followed by inevitable defeat and subsequent relocation. In each case, the leader has a safe seat. Edward Heath and Jim Callaghan are the two most recent exceptions – Heath gained Bexley from Labour in 1950 and Callaghan gained Cardiff South from the Conservatives in 1945 and both went on to represent basically the same seat from 10 Downing Street.

We’ll abandon the tabulation in 1979, because the four elections before that all starred the warring duo of Harold Wilson and Edward Heath. As the table shows, a new party leader can expect a small boost in their support in their constituency compared to the national average: par is around 1-3 per cent, and therefore an increase in their majority of about twice that. It seems that the worse the result for the party nationally, the more the party leader’s constituency result overperforms so the stars of this table are Michael Foot and William Hague. The luckless Theresa May is the only new party leader in recent decades to get a worse result in her constituency than her party managed nationally.

It may be that people are pleased when their constituency is ‘put on the map’ by hosting a party leader – it means that the constituency gets more attention than any other safe or semi-marginal seat at election time. If name recognition is a net positive, that obviously helps. It has also often meant, as with the Speaker’s seat, that it gets quietly spoken but influential representation in Parliament. The leader is obliged to delegate most constituency work to his or her office who are probably better at it anyway. It may also be that having a party leader inspires harder work by local party activists and more assistance from outside, particularly if the seat is thought to be in even the vaguest danger.

New party leaders also seem to boost their party’s performance in the regions or nations with which they are identified even if their party does poorly across Britain – Gordon Brown (Scotland), William Hague (Yorkshire) and Neil Kinnock (Wales) all saw their parties do relatively well on home turf. Canny Harold Wilson managed to claim two home regions full of marginal seats – Merseyside and West Yorkshire – and both turned in better than average Labour showings in his first outing in 1964. In 2017, however, there was little evidence of this as the Home Counties produced some mediocre Conservative results and Labour’s strength in London owed much to factors other than Corbyn’s long incumbency as an Islington MP. Boris Johnson does not have a strong regional identity – he has rather turned his back on the London that twice elected him Mayor. If anything, he is a creature of the M40 corridor between the fleshpots of Westminster and the West End and his alma mater of Oxford, via both Uxbridge and his former constituency (2001-08) of Henley.

The most vulnerable party leader in this election is not Johnson, but Jo Swinson

Uxbridge – even coupled with South Ruislip – is not an entirely safe Conservative seat, although 1966 was the last time the Conservatives lost it. Johnson’s majority slipped badly in 2017 and Labour ended up only 10.8 percentage points behind him. This was quite similar to the leads enjoyed by Harold Wilson (9.8%) and Margaret Thatcher (10.4%) in their constituencies as they went into their first elections as party leader. Both enjoyed significant national swings in their favour. Even if the national result is a standstill since 2017 – and if the Conservatives are 10-12 points ahead that would mean the swing in 2019 is on the same scale as 1964 and 1979 – Johnson is safer than he looks on the figures. The Conservatives have no doubt commissioned extensive polling to check this contention. Trying to knock out the party leader is always a temptation for political rivals if the seat looks halfway marginal, but it nearly always fails as it did in Cardiff in 1979 and Folkestone in 2005.

The most vulnerable party leader in this election is not Johnson, but Jo Swinson. The electoral currents buffet third parties more harshly than the big two, and Liberal leaders have been defeated (Archibald Sinclair, 1945) or had uncomfortably tight counts: Jeremy Thorpe in 1970 and Tim Farron in 2017 both saw their majorities reduced to three figures. Swinson’s East Dunbartonshire seat has long had ferociously volatile electoral behaviour (it has elected MPs from all four main Scottish parties since 1983), and she regained the seat in 2017 with a comfortable but not overwhelming majority over John Nicolson, the SNP MP who displaced her in 2015. With the SNP apparently on the up again, and the relatively good state in which Ruth Davidson left the Scottish Conservative and Unionist Party, it might be worth keeping an eye on the suburbs north of Glasgow as well as the ones west of London.