John Curtice: An era has passed with the death of Sir David Butler

An era has passed with the death of Sir David Butler earlier this month. Best known to the public as a key face on BBC election night television from 1955 to 1979, to his fellow political scientists he was the father of psephology. Nobody did more to advance the academic study of how Britain’s electoral system worked, how election campaigns were fought, or why voters behaved as they did.

He was also my supervisor while I was a graduate student. One of the risks that any senior academic who supervises graduate students potentially faces is that their previous work is questioned or overturned by those they are helping steer into the academic world. In David’s case, it was he who, unwittingly, set me off in that direction. 

The defeat of Jim Callaghan’s Labour government in a vote of confidence in March 1979 precipitated a somewhat early general election. Since 1945 every election had been graced by a Nuffield election study entitled ‘The British General Election of….’. David had contributed a statistical appendix to the very first study and had been the principal author or co-author of every book since 1951. He now gave me the task of co-authoring the appendix to the 1979 study, a task that entailed detailed analysis of the constituency by constituency results.

Shortly after the election, David took me with him to a small private academic conference at the University of Sheffield (his willingness to drive the length and breadth of the country at some speed was legendary). At the conference, David noted that although Margaret Thatcher had secured an overall majority of 39, the Tory advantage was much less than might have been anticipated given that the Conservatives were as much as seven points ahead of Labour.

In his early writing on Britain’s electoral system, David had demonstrated that Britain’s single member plurality electoral system conformed to a ‘cube law’, and that this made it highly likely that whoever came first in votes would have a clear overall majority in seats. But if the cube law was still working, Margaret Thatcher should have had a majority of more like 129 rather than 39.

Struck by that remark, I went off to read the literature on the cube law. This, I discovered, argued that how the electoral system operated depended on the geography of party support. If the cube law was to work, around 30 per cent of all seats needed to be marginal between Conservative and Labour. However, although unnoticed at the time, the number of marginal seats had fallen by October 1974 to little more than half that figure, a change that had persisted in 1979. The cube law was no more.

It was a measure of the quality of the man that not only did David not question his young graduate student’s finding, but even went on to embrace it. One implication was that hung parliaments in which no party secured an overall majority were now more likely, and David wrote in the 1980s about how British constitutional practice might adapt to the prospect of governing without a majority.

In the event this work was before its time – thanks to a sequence of landslide elections, a hung parliament eventually only happened in 2010 (and then again in 2017) – but David had prepared the ground 25 years previously.

The 1979 election proved to be David’s last as BBC TV’s election-night psephologist. However, it was my introduction to the world of election television. In his contribution to the 1945 Nuffield election study, David had invented ‘swing’, the average of the change in the Conservative and Labour percentage shares of the vote since the previous election.

His trade mark contribution to election night television had become to announce after every constituency result what the ‘swing’ was and how it fitted into the pattern of results so far.

Originally, the calculation was made by an army of students equipped – before the days of calculators or computers – with slide rules. By 1979 the BBC had its own mainframe computer that could calculate the swing at speed – and indeed many other statistics besides. But it was somewhat temperamental, and during rehearsal had occasionally broken down.

I was David’s back-up plan should the BBC computer go down on the night. Sitting behind him and armed with a calculator I had programmed for the purpose, my job was to enter each constituency result as it was broadcast and get the swing promptly to David. In the event the computer held up all night and my effort was redundant – but it was the start of what is now more than 40 years of broadcasting experience. 

David never lost his interest in the welfare of his former students. Many years later, on the night of the 2015 election, David, now past 90, was behind the scenes at the BBC and joined the psephological team in the gallery. He had a worried look on his face – and quietly asked whether the exit poll (which, contrary to the opinion polls, was forecasting a Conservative majority and a near complete sweep for the SNP) was right.

As it happened our luck was in. However, David had not forgotten how unforgiving the television lights could be if things went wrong. May he now rest in peace.