Who Carries the Party on Facebook? Leaders, Party Pages and the Polls in Britain, 2010 to 2020

Is a British party's Facebook audience built around its leader or its party page, and does either track the polls? I use 89,110 posts from the pages of eight parties and eighteen of their leaders between June 2010 and October 2020, which drew 218 million interactions, and match them to monthly averages of 2,767 national voting-intention polls. Leaders carried very different shares of their party's engagement. Nigel Farage at the Brexit Party took 79%, Leanne Wood 75%, Jeremy Corbyn 74% and David Cameron 71%, while most Liberal Democrat and Green leaders took a quarter or less. Facebook's audience was a poor picture of the electorate. UKIP and the Brexit Party each took about a third of interactions in the months they were active while polling under 10%, and the Conservatives took a fifth while polling 36%. Within a party, engagement and polls moved together: month-on-month changes in a party page's interactions went with same-month changes in its polling (0.43 points per log point), and higher-engagement months within a party and year went with higher polling. I find no evidence that engagement moved first, though I cannot rule out a short lead. Leader pages added nothing to the party page as a signal, and in periods when more of a party's engagement went to its leader the party polled lower, by 0.7 points for every 10 points of leader share. This association is slow-moving and depends on Labour under Corbyn and the Brexit Party. Election campaigns lowered the leader's share by about 6 points, because party pages doubled their output, although each leader post gained more than each party post. Changes of leader did not reliably lift the party page. Leader posts drew more love reactions and fewer comments and shares than party posts. Facebook engagement in this period tracked a party's standing at the time; it was badly calibrated across parties, and leader-centred engagement went with weaker polling.