Tuesday 18 July 2017

Election 2017 analysis - part 2

A post-election analysis from Charles Britten

Thanks once again to Ben for allowing me access to his blog. This piece can be considered a follow-up to the article I wrote ahead of the election. Hang that landslide my pre-election blog began with a mention that the situation was comparable with 1923: After several years in coalition with the Liberals, the Conservatives had just come into power on their own as a majority government for the first time in many years. But then the prime minister was replaced and his successor sought a new mandate to help push through an agenda that would alter Britain's economic relationship with the rest of the world. The result? A hung parliament. At the time, I suggested the only feature of that election likely to repeat itself this time round was Labour' seat tally, which in 1923 had been 191. After all, a landslide looked on the cards and, even with all the usual caveats, this year's May local elections did nothing to change the outlook. Therefore, the eventual result may leave this commentator looking as embarrassed as the rest, but I must protest: It is a fact that I wrote a detailed blog for this site before the 2015 election explaining how the electoral volatility that has become such a feature of British politics in recent decades has made frequent hung parliaments more likely, much in the manner of the 1885-1929 period when six out of 13 UK elections brought such an outcome.

I therefore plead glorious prescience. Maybe, just maybe, I know what I'm talking about. Drilling down into the 1885-1929 period, one difference with the present should be noted. That era brought hung parliaments, but interspersed these with a series of large majorities; the smallest any government had in that time was 72. If the current period is to be similar, the small Tory majority of 2015 may turn out to be a particularly anomalous result. In short, this year's election was called because it appeared a landslide was on the cards. It still could have been a substantial Tory win but for some extraordinary turns of events, and it may be that the next few elections will follow this pattern of either someone winning big, or nobody winning. The campaign factor If Labour's 1983 manifesto was dubbed "the longest suicide note in history" by Gerald Kaufman, so the words of the Conservative MP for Ribble Valley Nigel Evans that "we shot ourselves in the head" will resonate. The fiasco over social care was simply the worst feature among many of a manifesto that has been swiftly dismantled. Undoubtedly the Conservatives failed to win because of a poor manifesto and a dismal campaign. Labour, by contrast, were seen as having performed very well, Diane Abbott aside. A strong pitch to the youth vote and a highly active social media campaign - albeit one that may have irritated far more people than it converted - helped create a positive perception. If nothing else, it looked energetic, organised and passionate, helping to conceal the reality of a Labour Party full of sitting MPs who loathed Corbyn and hoped to survive the election to see him off afterwards. The fact that the Conservatives did not draw attention to this fact at every opportunity was a key reason why they missed this electoral open goal. 

The question is, what difference did any of this make? In the end, despite all their egregious errors, the Conservatives still managed to get 43.5 per cent of the vote in Great Britain (42.4 when Northern Ireland is included). On either count this was still their best figure since 1983. If their support had dropped a bit from the high 40s, this alone could not account for Labour's rise from poll ratings in the mid-20s to 41 per cent of the GB vote (40 per cent across the UK as a whole). It may make a lot more sense to say that, for all that Theresa May's falling ratings did not massively impact Tory support, they may have had the impact of ensuring that Corbyn did not look half so bad by comparison as he once did. This may be one factor, but not the only one, that bolstered the Labour vote. Conversely, it might be argued that, given how bad the Tory campaign was, Labour missed a golden chance because of its leader and policy platform. So far, only Nottingham East MP Clive Lewis has put his head above the parapet to say so. But that is an issue Labour will need to confront - if it ever can snap out of the delusion evident in its spurious claim to be the "real winner" of the election. 

Did the polls get it all wrong? Before considering the details of the results, it is worth examining the opinion polls, now as much a source of debate as any of the issues. One plausible reading of the situation is that the polls were wrong all along and the Conservatives were never really 20 points ahead, meaning the race was always going to be tighter than they thought. This theory could be supported by noting the numerous polls showing a consistently substantial Tory lead. Even if this had shrunk markedly during the campaign, several established pollsters like Ipsos Mori and NOP had the Conservatives well ahead on the morning of June 8th. In the past the polls had either been accurate or had overestimated Labour support. This was true in 2015 and even more so in 1992, the two most recent Conservative majorities, when the polls had been firmly forecasting a hung parliament. This time, they exaggerated Tory support, probably because, having been wrong in the opposite direction in the past, they overcompensated in trying to correct against a built-in Labour bias. To say they got it wrong again is to understate the point: to exaggerate a Conservative lead was an unprecedented occurrence. However, not all pollsters got it wrong. Some, like relative newcomer Survation and YouGov, were very accurate. Indeed, YouGov's startling prediction of a hung parliament two weeks out turned out to be almost spot on. Their sophisticated, detailed methodology and large sample size may be the model for all pollsters to follow in the future. Having put all the polling data together, the Electoral calculus site had given a final forecast of a Conservative majority of 66. However, it also noted that the margin of error gave a range that, at one extreme, would give the Conservatives only 314 seats, Labour as many as 269 and the SNP as few as 34, all very close to the actual figures. They were also right for the other parties bar the Liberal Democrats, who they tipped to win eight at the most (even here, however, they saw Nick Clegg's defeat coming). While this figure was set at one end of a scale, the exit poll was very close to being right, but with a caveat: it tipped the Conservatives to make no gains from Labour in England, but several in Wales, as well as holding seats such as Gower and Cardiff North. In the end, they made net losses in Wales, yet made gains in England, with these errors largely cancelling each other out. 

The overall accuracy of the exit poll belied the fact that there were some extraordinarily close results. In 2015 the lowest majority was 27. This time it was two (Fife North East), and four other seats were won by less than 25 votes. While only 12 seats were won and lost by less than one per cent in 2015, this time the total was 30. Of these, no less than eight were held by the SNP. Notably, the Conservatives would only have needed to win those seats they were up to half a percent shy of victory in to gain a majority. Factual though this may be, however, it should not be emphasised too much. After all, Theresa May did not call the election in the hope of getting a tiny majority by being lucky in the most marginal seats. Nonetheless, a one per cent swing would have given the prime minister an increased majority.

The youth factor

A key factor in the Labour vote surge was, it was claimed, a large jump in turnout among voters aged 18-24, most of whom supported Labour. Undoubtedly, there is some evidence of an effect, but how much? Firstly, an early tweet claiming that the youngest demographic in the electorate had seen a 72 per cent turnout was inaccurate. Later polling showed it was in the mid-fifties. Nonetheless, that is at least ten per cent above the estimated figure for 2015. Moreover, the specific promise most likely to appeal to young students and graduates - the abolition of tuition fees - does appear to have resonated, with Labour seeing some of their most startling swings in areas with high student populations, such as Canterbury, Chester (which had been the number one Tory target seat) and Birmingham Edgbaston, which has gone from being the middle-class marginal the Conservatives always held until 1997 and could ever quite win back, to a genuinely safe Labour seat. Against that, the 'youth effect' has not been seen everywhere. Many of the seats with higher than average numbers of young voters showed little sign of an unusually large swing to Labour. Moreover, the overall impact on turnout was very limited, as this rose from 66 per cent to 68 per cent, itself a continuation of the post-2001 trend for turnout to rise slightly with each passing election and still low by historical standards. 

The higher youth vote was also much too small to account for all of Labour's increase in vote share. What is evident is that there is now a clear correlation between age and voting behaviour. It always was true that younger people tended to lean more to the left and older people to the right, but the general tendency has become much more pronounced. It remains to be seen how the Conservatives respond in policy terms to the perceived need to make a more compelling offer to younger voters, but this will undoubtedly play a part in shaping its future policy-making. The Brexit double whammy Labour's lead among younger people and graduates contrasted with that of the Conservatives among the older and less-skilled, reflecting the split that had characterised the electorate in the EU referendum. Overall, the Conservatives did better in 'leave' areas of England, while Labour thrived in 'remain' areas. In terms of electoral geography, this helps explain the poor correlation between the seats gained and lost by the main two parties and their target lists. Rather than voting along the lines they had in 2015, many voters switched according to their referendum choices. The Conservatives had a very low correlation indeed. Of the 19 seats they gained, plus their by-election win in Copeland, just eight were in their top 50 targets, and only one of them was among the dozen gains from the SNP. For Labour, the correlation was strong in the most marginal seats - they took 14 of their top 20 targets - but lower down the list things became rather more peculiar. Thus it was that while they could not take Morley and Outwood (6th on the list with a majority of 0.9 per cent), they were able to take Portsmouth South (87th), Battersea (88th) and Canterbury (104th) from the Conservatives. While Labour were able to defend several marginals in London (eight seats they won in 2015 were in the top 25 Tory targets), they were also able to make four gains in the capital. Like London, Labour also performed very well and made gains in other strong remainer cities like Bristol and Brighton. 

Early polling evidence that the Liberal Democrats would not gain much from the 'remain' vote was only partially true, but what is clear is that many remainers flocked to Labour, despite the fact that the party's basic policy of leaving the single market and customs union, and ending free movement, was basically indistinguishable from the Conservative platform. Indeed, perhaps some who voted Labour hoping for a different kind of Brexit are now wondering what they did after Corbyn sacked three frontbenchers for supporting Chukka Umunna's Queen's Speech amendment calling for Britain to stay in the single market. For those whose biggest concern was avoiding a hard Brexit, a vote for the Liberal Democrats or even the Greens would have been infinitely more appropriate than one of either of the two main parties. The fact pollsters were wrong in predicting remainers would stick with their 'normal' party was a blow to the Conservatives. Moreover, it was only slightly mitigated by their relative success in 'leave' areas of England. The big hope had been that they would hoover up most of the UKIP vote, including many former Labour voters, thus winning scores of seats in traditional Labour heartlands. In the event, the broad picture across the country was that Labour actually picked up more of the vote falling away from UKIP, unless a more even split was compensated for by switchers from elsewhere, plus extra votes from young people. However, what was more notable than the general trend was the capricious way in which the UKIP vote unravelled in specific areas. For example, Walsall North fell to the Conservatives because they picked up most of the UKIP vote. An identical shift would also have secured Walsall South. Instead, Labour harvested most of the old UKIP vote and made a marginal seat into a safe one. Similarly, in the East Midlands those unexpected Tory gains in former Labour bastions like Mansfield and Derbyshire North East did not show the whole picture. In most Derbyshire seats the old UKIP vote went to the Conservatives, but not in High Peak, which Labour gained. And in Wales, which voted for Brexit, Labour got back most of the votes they had lost to UKIP in 2015, making gains from the Tories as a result. Thus the Conservatives suffered a Brexit double-whammy. In remain areas they suffered, while leave areas turned out out to be much less consistently favourable than hoped. This shows that gambling on winning an election based on a major shift in voting allegiances from left-right to national-international was a risky step into the unknown for Mrs May. In the event, the fact that 'leave' voters in general and former UKIP voters in particular did not flock to her cause in the numbers anticipated meant she would never get the huge landslide she hoped for, although it took the other failings in the campaign to deny her a majority at all. 

The Liberal Democrat recovery - of sorts

If the remain vote did not flock to the Liberal Democrats, it did at least help produce a 50 per cent increase in seats, despite an overall loss of vote share. The two gains in London brought Vince Cable and Ed Davey back into the fold, while Bath may be the start of a comeback in the party's former stronghold in the south-west. Tactical voting against the Conservatives in remainer areas clearly played a part in this, as it did in Scotland as unionist voters sought the best-placed party to unseat sitting SNP members. However, the successes of the party south of the Severn-Wash line and north of the Tweed contrasted with a very poor performance in Wales, the Midlands and the north. Tim Farron barely held onto his seat, Nick Clegg was one of two members unseated by Labour in Yorkshire, Southport went to the Conservatives and Ceredigion to Plaid Cymru, whose final total of four seats - notwithstanding the failure to snatch ultra-marginal Ynys Mon from Labour - was much better than some pre-election polls had forecast. The election does leave the Liberal Democrats with a few more target seats to go at, particularly traditional areas like rural Scotland and south-west England, but little prospect yet of returning to the levels of strength seen between 1997 and 2010. However, if Brexit proved as much an impediment as a benefit in leave-voting areas like Cornwall this time, the focus on other issues after Britain leaves the EU may work in the Liberal Democrats' favour in future.

The Scottish situation

If Brexit had set the agenda elsewhere, in Scotland it was Nicola Sturgeon's bid to use the EU vote to push for a new independence referendum that focused Scottish minds. Apart from the fact that many SNP members, voters and even some MSPs were pro-Brexit, the EU vote made almost no difference to sentiment about the union. Most Scots might prefer EU membership, but leaving the UK to chase it is another matter. Thus it was in Scotland that the Conservatives, pitching themselves as the most unambiguously pro-union party, were able to make most progress and recorded their best result since 1983. However, here as elsewhere Labour did better than expected, with some large swings to pick up several seats and come close in others. Indeed, most of the several very marginal seats the SNP hold are now particularly vulnerable to Labour attack, notwithstanding the anticipated boundary changes before the next election. Like the Liberal Democrats in England, the SNP has historically been under-represented in seat by spreading their vote too thinly. By hitting almost 50 per cent last time they solved the problem at a stroke, but there is a tipping point where their vote is in the 30s and the Labour and Tory vote not much less, leaving a string of very marginal seats, including several three-way fights. The SNP could easily lose many more next time. 'Peak Nat' has indeed passed and whatever other problems Theresa May now has, the prospect of losing Scotland off the back of Brexit is not among them. 

The next election

If you had thought Mr Corbyn had gone far enough with claiming victory in an election he still lost, he showed up at Glastonbury to prove everyone wrong. Having turned up at an event run by a millionaire to tell people who could afford to pay hundreds of pounds to see millionaires performing how capitalism had failed them, he then declared he would be prime minister within six months. Time for a reality check: The arrangement with the DUP, controversial though it may be for various reasons, will keep the current government in place for two years. The duration of the deal is no coincidence; it means they can get on with the Brexit process and then, after March 2019, try to find a new prime minister who is better at campaigning than being bloody difficult, before going to the country again. That next election could be every bit as unpredictable as the last two, and the big danger for the Conservatives is that Brexit will go badly, the economy will struggle in the aftermath and they will get punished at the polls. Against that, it is hard to see them campaigning as badly again next time. What could make the next election especially hard to call is the possibility of a Brexit unwind. This would involve voters who made their 2017 choice based on 'leave' or 'remain' sentiment focusing on different priorities. Scotland may provide a template for this, with the 2015 election shaped by the previous year's referendum. The passing of 'Peak Nat' was evident this time, although the result was still the SNP's second best ever. Back in England, whatever the overall result, nobody should be surprised if the startling results of Mansfield and Canterbury are both reversed at the next time of asking. It might seem paradoxical that at a time when the two largest parties have secured their highest joint share of the vote since 1970 and so much talk is of a return to 'two tribes', we actually have a situation where regular hung parliaments are now a reality. After all, Labour has failed to reach 300 seats on three successive occasions, while the Tories have won one majority in the last six elections. Despite their low vote shares, the other parties still collectively won 66 seats, enough to keep the electoral arithmetic precarious. This is why large majorities are the most likely alternative to hung parliaments; to win in such circumstances generally requires one of the two main parties to heavily outperform the other. 

Two years ago I wrote about how the process of partisan dealignment - the unravelling of the old class-based ties between voter and party - had prompted more people to vote for parties other than Labour and Conservative. However, the fluidity of allegiance has also led to more switching between the two, and for varying reasons. Nowadays there is little class-based difference between the two parties; the greater contrasts are found in age and, to a lesser extent, other demographics closely linked to which side of the Brexit debate people tend to fall on. If these played a big part this time - allied to Labour's pitch towards young, metropolitan graduates who were strongly pro-remain - there is no reason to suppose that this will still apply in a post-Brexit world. But the question of what and who will provide the compelling answers to the various problems of today remains to be seen. That alone does not mean more hung parliaments are bound to happen, but such outcomes are more likely when neither party has an obvious winning formula to hand. Some, of course, will deny that 'Peak Corbyn' has arrived and declare that it is inevitable he will win next time. Others might consider how the same was said about Scottish independence, now a fast-receding prospect as the SNP wanes. Historical inevitability may appeal to keen adherents of Marxism with their insufferable sense of entitlement, but nothing is inevitable about the British political scene now. To make assumptions about what is certain and what is not smacks of hubris and arrogance. Exactly this was the undoing of Theresa May. Those treating Labour's defeat as a stepping stone to future victory would do well to heed the lesson, but they almost certainly won't. In these uncertain times, perhaps predicting nobody will win is simply the safest bet.