Doing it properly for JR

John Rentoul tweeted this on 15 January and commented on my reply:

I’m sure he knows what an x-axis is and the limitations in that respect of the chart he tweeted, especially for periods when elections are spaced other than 4/5 years apart – like now. Anyway, this is possibly a slightly better way of presenting the same data for the vote share of the two main parties:

Really, it should be a histogram for each election, the data having no meaning in the intervals between elections. But joining up the data points appears to reveal a cyclic pattern of several election intervals duration.

Superimposing the percentage of the total vote which goes to parties other than the Conservatives or Labour seems to indicate the end in 2017 of what had been a very long-term trend. It was in the 1970s when the non-Labour, non-Tory share of the vote was last as low as in 2017. There were five general elections in the ten years from October 1964 to October 1974 - if there were a cycle, it was a short one. But it might be unwise to rule out a revival in Liberal Democrat and UKIP fortunes in the next few years.


Note

This is the full link to Cowling's data: drive.google.com/file/d/0B5Ik-g