The double incumbency effect

Welcome to the 141st edition of The Week in Polls (TWIP), which takes a look at a 1980s psephological fad that may be due a comeback in this Parliament - the double incumbency effect.

Then it’s a summary of the latest national voting intention polls and a round-up of party leader ratings, followed by, for paid-for subscribers, 10 insights from the last week’s polling and analysis.

This week, that includes the unpopularity of many of the causes Nigel Farage is associating himself with.

Before that, a raise of the eyebrow at the new MRP from More in Common, which the pollster has written up with such precision as, “Labour would lose 87 seats to the Conservatives, 67 to Reform UK, and 26 to the SNP” before, much further down the page, adding a crucial caveat that, “We also don’t have assumptions about tactical voting in this model as we would during a General Election campaign”. If you make the methodological choice to downplay a major factor in determining seat numbers, should exact seat numbers from the model be written up in that sort of way? (Luke Tryl’s response to that is here.)

JL Partners also has a seat projection out, based though not on MRP but on extrapolating from council by-elections. I’ll return to the pros and cons of both their and the More in Common approach in January. And before that - a happy new year to everyone.

---

Want to know more about political polling? Get my book Polling UnPacked: the history, uses and abuses of political opinion polling.

---

The double incumbency effect

Image by Pexels from Pixabay

One of the more niche fads of the 1980s, less widespread than Cabbage Patch Kids, shoulder pads or the Rubik’s Cube,¹ was the ‘double incumbency effect’.

The 1983 general election saw the Liberal/SDP Alliance nearly outscore the Labour Party in the popular vote but fail massively when it came to seat numbers, with nearly all the Labour-to-SDP defectors losing their seats. As a result, when people started analysing the results and wondering what the future held for the parties, one of the pieces of psephological smartness to cite was how the double incumbency effect meant that the seat-by-seat results were (even) worse for the SDP than they looked.

This is because if an incumbent MP loses their seat, then at the following general election not only does their own party lose the benefit of their previous incumbency vote,² but also the rival party which won the seat benefits from any incumbency vote that the new MP has built up.

So poring over the details of the 1983 results, the reality was worse for the Alliance as in many of the seats it came closest to winning, a double incumbency effect would hit it at the next general election: the loss of a defeated Alliance MP’s incumbency vote plus the build-up of a rival party’s incumbency vote in that seat.

Measuring the size of such an effect is difficult. Partly that is for theoretical reasons. What counts as an MP’s incumbency effect? It is really an amalgam of a direct personal effect - people like the person and so are more likely to vote for them - and also an indirect organisational effect - the presence of an active MP helps build up a more active and effective party organisation in the seat.

Muddying the theoretical waters further is the distinction between the personal vote for an MP and the targeting activities of a party. If, say, national HQ pours direct mail into an MP’s constituency, that may produce a boost to the result in that seat which looks like an incumbency effect but is really a campaign intensity effect.

Partly too it is due to a polite, mild-mannered but long-running debate among political scientists - with a niche side-beef between political scientists and Liberal Democrats - about how big such incumbency effects really are.

Decades back, political scientists used to be fairly dismissive of MPs claiming a personal vote, saying any such effect was tiny. More recently, there is a widespread acceptance that this effect is real, but the new conventional wisdom pegs it at a fairly small number.

Speaking on the Trendy podcast in March 2024, for example, Professor Sir John Curtice said:

1,000 votes is probably the maximum you can get out of it [in a Westminster Parliamentary seat] so … if you think you might lose a seat by 1 or 2 per cent, it might just rescue you.

However, other political science evidence, such as the work by Professor Justin Fisher, points to significantly larger constituency-specific campaign effects, whether they are due to incumbency effects or particularly vigorous target seat campaigns.

(Digression: if you were an election agent for the general election and have not yet returned the survey that Justin and his colleagues send out, please do! These agent surveys create an important strand of knowledge about what actually goes on in elections, and with what effect.)

Then there is the (one way) side-beef from Liberal Democrats, who look at figures such as that 1,000 votes one and say, can I introduce you to Helen Morgan, Tim Farron or Norman Lamb,³ for example? Indeed, I have a whole presentation for academic audiences about this issue, and the problems faced with explaining Liberal Democrat election results if you start with believing figures such as that 1,000 votes one. Plus it is rare⁴ to find in academic literature a clear distinction between those three different effects: the personal vote, the indirect organisational benefit and the campaign effect.

But while the size of such double incumbency effects may be hard to calculate, it is likely they will join the discussion of ‘don’t knows’ as a topic from the past which returns for this Westminster Parliament as essential to understanding what is going on.

That is for two reasons. First, thinking about the future for the Conservatives and the extent to which any recovery for them has to run through seats they lost to the Liberal Democrats in July. That Cheltenham result, for example, looks even harder to overturn when you factor in the loss of Alex Chalk’s incumbency effect next time⁵ and the scope for Max Wilkinson to build up his own.

Second, thinking about the future for Labour and the large number of seats it won with a relatively low national vote share. As Compass points out:

131 Labour seats were won by a margin of fewer than 5,000 votes, with 103 secured by less than 5% of the vote.

Or the talk of Labour’s majority being akin to a sandcastle, liable to be swept away.

But… perhaps those seats will turn out to be less fragile than those figures and analogies may make it appear, due to the double incumbency effect.

It is a reason certainly to hope that high-quality MRPs, fully taking into account factors such as tactical voting,⁶ return as it will be them, rather than the national polls, which tell us if the double incumbency effect will, like many of those 1980s fads, remain firmly in the past or, like Rob Lowe, make a comeback that obsesses politicos.⁷

---

Landslide: The Inside Story of the 2024 Election

---

Voting intentions and leadership ratings

Here are the latest national general election voting intention polls, sorted by fieldwork dates:

Next, a summary of the the leadership ratings, sorted by name of pollster:

For more details, and updates during the week as each new poll comes out, see my regularly updated tables here and follow The Week in Polls on Bluesky.

For the historic figures, including Parliamentary by-election polls, see PollBase.

---

Catch-up: the previous two editions

How useful is Uniform National Swing (UNS)?

The return of the don't knows?

---

My privacy policy and related legal information is available here. Links to purchase books online are usually affiliate links which pay a commission for each sale. Please note that if you are subscribed to other email lists of mine, unsubscribing from this list will not automatically remove you from the other lists. If you wish to be removed from all lists, simply hit reply and let me know.

---

Starmer still comes out on top in best PM polling, and other polling news

The following 10 findings from the most recent polls and analysis are for paying subscribers only, but you can sign up for a free trial to read them straight away.

1. You only have to beat who you are up against. Which is why the big net unpopularity of both Donald Trump in 2016 and Boris Johnson in 2019