This time, it’s not just a simple swing for a Labour majority

In the early 1950s, David Butler, an academic at Nuffield College in Oxford, discovered that, at the time at least, swings between parties tend to be broadly consistent across the UK. On the back of that, the “swimgoneter” was invented as a graphical device to illustrate how many seats the main parties would get in the House of Commons, given a particular amount of swing.

One of the caveats that comes with a swingometer is the assumption that swing is uniform across every constituency across the country. That, of course, is never the case, but the idea is that some seats will be gained in excess of the average swing while other seats are lost that fall short, and those two roughly balance out. When that balancing out does happen, that allows the swingometer to become a decent tool for predicting seats, and for predicting how much of a swing a party will need to gain, or lose, an overall majority in Parliament.

The swingometer for Great Britain suggests that, on the current national poll average, Labour would barely win a majority. (Neither Labour nor Conservatives contest seats in Northern Ireland.)

The current swingometer for Great Britain suggests that Labour needs a swing of around 13% from the Conservatives to win an overall majority in the Commons. If you average the most recent opinion polls from each of the major UK pollsters, you’ll find a swing slightly larger than that at 13.4%, which, on a uniform swing across all GB seats, would put Labour on 330 seats. While that is an overall majority, and would mean a swing larger than Tony Blair’s swing in 1997, it wouldn’t give a Prime Minister Starmer much of a margin of error to govern as he’d want to.

But as we saw in the recent local elections, the swing is not going to be uniform in every part of the country. While not every part of the country voted this week (not a single vote was cast in Wales, Scotland or London, for example), we did see a significant variance in terms of council seats gained and lost in England, and in particular who the beneficiaries of these Conservative losses were. That calls into question the whole theory of uniform swings.

Some regions saw Labour benefit from Conservative losses, while other regions saw the Liberal Democrats benefit.

That suggests that national swing is not enough to work out Labour’s prospects for government in next year’s general election. We need to drill down into the swings in the various regions, and see if any variance in swings is going to have a significant impact. And when we look at the crosstabs in the polls, we see quite a significant variance in swings; the swing in the Midlands is averaging around 5% higher than the national average, while it’s 2% lower in the North of England and 3% lower in Wales.

The opinion polls are showing a significant variance in swings between different regions of the country.

Going back to the original theory of the swingometer, the extra gains that Labour makes in the Midlands should be offset by the seats they miss in the North and in Wales. So if we apply a uniform swing in each region of England, does this happen?

Labour gains a lot more seats than they miss when we look at regional swings, as opposed to national swings.

When we look at the regional swings in England, Labour finishes with 12 more gains from Conservatives than they would if we looked at regional swings alone. This is not entirely surprising when you scroll back up and look at the swingometer itself, where you’ll see an imbalance of seats on either side of the 13.4% swing arm, with more seats up for grabs on a larger swing than lost on a smaller swing. In this case, the imbalance tells us that a variance of swings is likely to result in a better result for Labour than a uniform national swing, as they’ll gain more seats in areas with the excess swing than they’ll miss in other areas.

Additionally, looking at regional swings removes a distortion that national swing introduces with regards to regional parties, specifically the Scottish National Party and Plaid Cymru. The current poll average suggests a 10 percentage-point decline for the SNP in Scotland, but given that they only stand in Scotland, this translates to a decline of only one percentage-point across Great Britain. That understatement of the true extent of the SNP decline means the national swing can’t be used to project seats in Scotland, and a similar argument can be made to Wales as well.

So when we talk about the national swing and the national leads that one party has over another, we need to exercise a degree of caution. In some circumstances, this being one of them, a variance in the swing can help or harm parties more than a uniform national swing, and it’s critical to see whether such a variance exists, and what its impact could be, before jumping to conclusions about what the next government is going to look like.