The results of our latest survey of Party members, published yesterday, appear to have produced an interesting reaction.
This week’s survey asked for the first time how many members have already voted. Seventy-one per cent of those on our panel say they have cast their vote, which if the voting intentions are accurate would make it mathematically impossible for Jeremy Hunt to win via a late surge.
Shortly after that finding was published a range of leaked official turnout figures started to crop up. Beth Rigby of Sky News was told the figure was ‘less than half’ by three sources, including one who claimed the figure was lower than 33 per cent. The BBC’s Laura Kuenssberg had been tipped off sufficiently firmly to assert that ‘Fewer than half of Tory members have so far voted in the leadership contest and sent back their ballot to party HQ – the assumption that they would all make up their mind in a flash has turned out to be wrong’. Francis Elliott of The Times has also been told ‘fewer than half’.
In short, there is quite some discrepancy. At one end is our survey figure of 71 per cent. At the other end is that Rigby source claiming somewhere below 33 per cent. And the other Rigby sources, Kuenssberg’s source and Elliott’s source, who say ‘fewer than half’ are in the middle somewhere – let’s assume around 40-49 per cent.
The reasons such a discrepancy might arise are interesting in their own right, but the truth is also politically important. It alters the tone and nature of the rest of the contest, if you believe either that most selectors have voted or most are still up for grabs.
The source of the numbers is key. It seemed likely from Rigby and Kuenssberg’s reports that their figures had come from inside the Conservative Party’s structure. Electoral Reform Services are the outside company contracted to run the leadership ballot, and while the election is formally overseen by the 1922 Committee, ERS’ contract is with – and bills paid by – the Conservative Party itself. So it wouldn’t be surprising to learn that ERS would provide regular progress reports on the running of the ballot to its client – not, of course, on how people are voting (the votes for each candidate are yet to be counted), but on whether people are doing so, whether ballot papers are being successfully received by post, and so on. On initially hearing the BBC and Sky numbers yesterday, I assumed that the figures were from just such a progress report, and were therefore most likely to be leaking from somewhere inside CCHQ or somebody in turn briefed by them.
Elliott’s report in The Times today confirms this assumption to be correct, specifying the source as ‘the internal turnout assessment passed to CCHQ from the Electoral Reform Society’.
By contrast, the ConservativeHome survey is a survey of Party members on our panel – 1,319 of whom answered the turnout question.
Anecdotally, we have other sources who echo it. An experienced organiser within the Johnson campaign tells us that in their area the Get Out The Vote operation has so far turned out 75 per cent of Johnson supporters. A Cabinet minister who has been following their local members’ decision-making estimates association turnout to be 80 per cent. A senior member of the voluntary party estimates the national figure to be around 70 per cent.
Unsurprisingly, we believe our figure to be closer to the truth than reports of only a third, or a minority, of votes having been cast, and it seems that various people closely engaged with the process tend to agree.
But the discrepancy still exists, and must be accounted for. How has it arisen, and might it be possible to navigate the various numbers to get at what is really going on?
We can dismiss the baseless allegations of untruth that have become all-too common. We do not know if any of the journalists reporting the contents of an ERS briefing have seen a document, or simply been told of it, but there’s no reason to believe that they are doing anything other than accurately reflecting information from sources they trust. Let’s engage with all the numbers on the basis of good faith.
Looking at our figure first, are there factors which could lead the ConservativeHome survey figure to be too high?
Bluntly, yes: it’s a survey, not a weighted poll, and by definition a Party member reading this site and subscribed to our panel is likely to be somewhat more politically engaged than the average member. Plus, we’re sending them regular surveys about the leadership election, which could spur some to vote by the simple effect of reminding them.
We won’t be catching negative answers from people who are ill, on holiday, et cetera. And anyone getting two ballot papers – as a member of two associations – but obeying the rules and only voting once will appear as a voter in our numbers but would only appear as 50 per cent turnout (one vote cast, the other not) in the ERS/CCHQ figures.
But even after considering those selection effects, the fact remains that our survey’s findings about opinions within the Conservative grassroots tend to map pretty closely to YouGov’s polling of the membership, so the panel doesn’t seem to be so wildly disproportionate as to account for discrepancies as large as those listed above.
So might there be factors which make the reported ERS figures an underestimate of the true turnout? Again, yes there are.
First, the ERS reports to CCHQ are effectively sampling an earlier stage of the election than our survey. It’s a postal ballot, so included in our figures are people who have recently posted their vote who won’t appear on the ERS tally until their ballot papers have been delivered, separated from personal data (eg the donation slips which were sent out at the same time) and tallied up. There could be a lag of two or three days in that process, which is not inconsiderable in the course of a week’s voting time.
Then there’s the question of how often the ERS submit these reports, and what data they are compiled from. If they’re daily, do they use the tally from the previous day’s postal delivery? Or are they less than daily? Again, this is a question of when these snapshots effectively date from.
We also don’t know when the reports being cited were submitted to CCHQ – they might be from yesterday (ie Wednesday’s tally data) or earlier. Indeed, that could even account for the difference between ‘less than a third’ and ‘under half’. If Rigby’s lower end source was citing earlier numbers than those who gave a mid-range number to her, Kuenssberg and Elliott, they could both be accurate but for different points in the last week – just as our survey, conducted on Wednesday, will include voters who won’t make it into the ERS tally until today or tomorrow.
There’s another effect that I suspect is at play. We’ve all put a letter in an envelope, stamped and addressed it, then left it on the side until we next know we’ll be going past a post box. There are likely to be quite a lot of Conservative leadership election votes in exactly that limbo right now. For good reason they won’t appear on the ERS tally of votes received, but I’d guess quite a few of those voters would regard their vote as having been ‘cast’ – on the basis that they’ve put the X in the box and it’ll be sent in very soon. They aren’t in the ballot box, but they’re out of contention for the candidates to win over – take your pick of whether they should be counted as having voted or not.
In short, it seems likely that our figure might be over by a bit, but that the low-ball claims are likely under by a decent bit – or, in the case of the lowest, by a lot. They aren’t necessarily untrue; instead, in effect the point in the race they illustrate is earlier on than the snapshot provided by the survey.
Of course, in the long-run this will prove academic. But for now it matters – and it’s worth noting that currently the interests of both leadership campaigns and CCHQ itself are all aligned in emphasising that turnout is lower than expected. Hunt and Johnson must activate their supporters as much as possible and avoid either depression or complacency setting in, while the Party’s authorities want to deliver a high-turnout leadership election to display their own effectiveness and deliver the new Prime Minister the largest possible grassroots mandate. Those conditions, more than anything else, underlie this debate on where the race currently stands.