Support 100 years of independent journalism.

Does Labour’s rising vote share suggest the polls are wrong?

I've looked at some of the thought-provoking responses to my piece. 

By Stephen Bush

What’s up with Labour’s increasing vote share? I wrote at length about the increase in Labour’s poll rating since the election was called last week, which attracted a number of interesting responses, which I thought I’d go over here.

The end of multi-party politics

Labour are putting on extra votes if the polls are to be believed, but so are the Conservatives. Why? Because Ukip seem to be in an advanced state of collapse and the Liberal Democrats have fallen back since the local elections, boosting both parties, though the Conservatives are overwhelming beneficiary.

This is the most straightforward interpretation of what’s happening, as it is exactly what we’re seeing in the polls.

Labour MPs are fighting good local campaigns

Select and enter your email address Quick and essential guide to domestic and global politics from the New Statesman's politics team. A weekly newsletter helping you fit together the pieces of the global economic slowdown. The New Statesman’s global affairs newsletter, every Monday and Friday. The best of the New Statesman, delivered to your inbox every weekday morning. The New Statesman’s weekly environment email on the politics, business and culture of the climate and nature crises - in your inbox every Thursday. Our weekly culture newsletter – from books and art to pop culture and memes – sent every Friday. A weekly round-up of some of the best articles featured in the most recent issue of the New Statesman, sent each Saturday. A newsletter showcasing the finest writing from the ideas section and the NS archive, covering political ideas, philosophy, criticism and intellectual history - sent every Wednesday. Sign up to receive information regarding NS events, subscription offers & product updates.
  • Administration / Office
  • Arts and Culture
  • Board Member
  • Business / Corporate Services
  • Client / Customer Services
  • Communications
  • Construction, Works, Engineering
  • Education, Curriculum and Teaching
  • Environment, Conservation and NRM
  • Facility / Grounds Management and Maintenance
  • Finance Management
  • Health - Medical and Nursing Management
  • HR, Training and Organisational Development
  • Information and Communications Technology
  • Information Services, Statistics, Records, Archives
  • Infrastructure Management - Transport, Utilities
  • Legal Officers and Practitioners
  • Librarians and Library Management
  • Management
  • Marketing
  • OH&S, Risk Management
  • Operations Management
  • Planning, Policy, Strategy
  • Printing, Design, Publishing, Web
  • Projects, Programs and Advisors
  • Property, Assets and Fleet Management
  • Public Relations and Media
  • Purchasing and Procurement
  • Quality Management
  • Science and Technical Research and Development
  • Security and Law Enforcement
  • Service Delivery
  • Sport and Recreation
  • Travel, Accommodation, Tourism
  • Wellbeing, Community / Social Services
I consent to New Statesman Media Group collecting my details provided via this form in accordance with the Privacy Policy

Always-readable Phil Burton-Cartledge made this suggestion and goes into further detail on his blog, that Labour incumbents are by and large fighting campaigns based on their own local records. The polls, which consistently put Labour at least 14 points behind the Conservatives, mean there is little fear of Corbyn.

Labour is in a perfect storm: people who like Corbyn can vote for him, people who don’t like him don’t fear him.

Recent pollsters have a pro-Labour slant

I don’t mean that in a conspiratorial way. Pollsters have what are known as “house effects”: Throughout the last parliament, Panelbase tended to produce higher scores for Ukip than their competitors – their last three before May 2015 produced Ukip scores of 11, 17, 16 and 17 per cent – and the same tends to be true of all the major pollsters.

Throughout this – admittedly rather shorter – parliament, four pollsters have tended to produce the worst Labour scores: ICM, ComRes, IpsosMORI and Kantar/TNS. ICM and Kantar have continued to put out worse figures for Labour, MORI haven’t put out a poll since before the local elections. Of the most “Labour hostile” – I cannot emphasise enough I am talking about the effects of the pollsters, not the politics of the actual people who work there – pollsters, only ComRes has published a poll showing an increase in the Labour vote outside of the margin of error.

This is perfectly plausible, though we obviously have insufficient data to know whether or not this is the case.   

The polls are wrong

The thing about political polling is that people only really care if you spot the result and are willing to forgive you getting the score wrong.

Take the 1992 election: taken together, the polls taken the month before the election forecast a hung parliament with Labour in the box seat as far as coalition or minority government was concerned. But the pollsters overestimated the Labour vote by five points, and underestimated the Conservative vote by three points. The result was a narrow Tory victory.

Fast-forward to 1997. The polls forecast a huge majority for Labour. But the pollsters overestimated Labour by close to five points, though they got the Conservatives about right. (There was, of course, some sharp variation, with some pollsters underweighting the Conservatives and some overweighing. Every pollster overestimated the Labour vote to a degree.) But the result was a huge majority for Labour, so no-one really noticed.

Could 2017 be a similar polling disaster? It’s worth noting that if all you knew about Britain in 2017 was the history of polling, you’d probably bet on the polls overestimating Labour, as did in 2015, 2005, 2001, 1997 and 1992. Forget the questions about the political context – whether Labour are polling 48 percent, 35 per cent or 31 per cent, no-one has ever gone broke betting that the party will do worse than its polls suggest.