polling station

For the fourth in our series focusing on constituencies to watch during the general election, we take a look at the Deputy Prime Minister’s seat

Three constituency polls over the last year have shown Labour beating Nick Clegg in his Sheffield Hallam constituency. Back in May, former Lib Dem Lord Oakeshott commissioned a poll by ICM that gave Labour a nine point lead. In November, Tory polling guru Lord Ashcroft’s poll by Survation initially showed a small lead for Clegg, but having re-checked the figures more recently, this poll too gave Labour a lead, albeit a smaller one of three percent. Finally, Survation carried out another poll in the seat last month for Unite, giving Labour a ten percent lead.

Below is a graph showing all the polls in Hallam since the general election.


The constituency

Instinctively, this sounds surprising: Clegg had almost a 30% majority in 2010, the constituency has never returned a Labour MP and Labour has been in third in every general election since 1983.


Sheffield Hallam takes in the western side of the city, as it climbs up the slopes of the Pennines. About half the seat is actually in the Peak District national park, although the vast majority of voters live in the suburban western fringe of the city itself. Those suburbs include Crookes, Crosspool, Dore, Ecclesall, Fulwood, Loxley, Millhouses, Stannington and Topley.

Outlying Pennine villages include High Bradfield, Dungworth, Ringinglow and Worrall, scattered amongst the river valleys that flow into the River Don further east in the city.

Not all the seat was traditionally Hallamshire, the county within a county that once covered most of what is now within Sheffield’s boundaries. Dore and Totley to the south were part of Derbyshire until 1934.

Loxley deserves a special mention – traditionally, Robin of Locksley (i.e. Robin Hood) was born in this village that has long since been absorbed within the city.


Sheffield Hallam is far from the highrise towers and former steelmills of the Steel City. Whilst not all the seat is not rich – parts of Crookes and Stannington are certainly not wealthy – overall Sheffield Hallam is one of the most affluent constituencies outside of the South East and it has the 70th highest median income of the 650 in the country – that is wealthier than Tunbridge Wells or David Cameron’s Witney. It has the lowest level of child poverty of any constituency in the land.

It is certainly one of the most highly educated seats in the nation: 60% of those of working age have a degree – that’s more than Cambridge. In 2001, the constituency had more people classified as professionals of any in the UK.

What about the students?

A lot of the Liberal Democrats’ success here since 1997 has been put down to the student vote. Whilst there is a fairly large student population (18% of adults are in fulltime education) and Crookes ward in particular has plenty of houses let out to students, this is less than before the 2010 boundary changes. Those changes removed Broomhill ward, which includes the main campus of the University of Sheffield, and replaced it with Stannington ward to the north.

Now, it is Sheffield Central next door that is the student hotbed: 39% of adults there are in fulltime education.

The Liberal Democrats certainly needed student votes to win the seat for Richard Allan in 1997 and to secure Clegg’s first election in 2005. But they are probably equally glad that it is no longer quite the factor there that it used to be now that student votes tend to be a lot less favourable for them.

Even with the loss of those student numbers in 2010, Clegg’s victory was emphatic: 53.4% of the vote, with the Conservatives (who dominated the seat until 1997) thirty points behind and Labour on a mere 16%.


Since 2010

Unsurprisingly, every one of the constituency’s fifteen councillors in 2010 was a Liberal Democrat. Given their national polling decline, what is more surprising is that thirteen of them still are: all the councillors have now stood for election since then. Four of the five wards still only have Liberal Democrats, whilst Crookes ward (the most urban) has two Labour councillors but still elected a Liberal Democrat one in 2012.


Labour comes reasonably close in one other ward, Stannington to the north. In fact, they only fell five votes short of victory in 2011. But they are a long way further back in Ecclesall and Fulwood wards, getting about half the Liberal Democrat vote in the latter.

They are even further behind in Dore and Totley ward, where they are still in third place behind the Conservatives. The Tories still have a strong showing there, Sheffield’s wealthiest and the last ward to have Conservative councillors, although that was some time ago now. In 2012, they came 7% behind the Liberal Democrats.

The map below shows the relative strength of Labour and Liberal Democrats, averaged across the 2011, 2012 and 2014 local elections.


This suggests the Liberal Democrats still have a commanding position. But are there any signs of a decline in support more recently?

The graph below shows the percentage local election results, aggregated for the five wards. Labour’s best result against the Liberal Democrats was back in 2011, and their support locally has fallen since. The Liberal Democrat level of support has actually been pretty static across the three local elections at 38-39 percent of the vote, although this is a drop from the 52% they achieved in the local elections held on the same day as Clegg’s 2010 general election victory.

The most significant drop in support has been for the Conservatives: their support has halved since 2011. At the same time, UKIP’s support has doubled between 2012 and 2014, although this may reflect the fact that the latter election took place on the same day as the European elections.


Those polls

These local results seem to be at odds with the polls predicting Labour victory. The Oakeshott poll took place only a few weeks before the 2014 local election results. So, could the results be wrong?

After the 1992 general election result, red-faced pollsters had to change their methodology due to their missing of “shy Tories”. When I posted a few days agoweek about Scottish polls, I suggested that the polls might be having some difficulty in correctly identifying “shy Labour” voters there, just as they had in identifying shy No voters in the referendum.

There is some evidence that this may also be the case with Liberal Democrat support: that is one of the reasons why most polls of late (particularly YouGov) show them on single figures but ICM polls always put them in the low double digits. This is because most polls exclude “don’t knows” from their figures but ICM’s methodology show these as supporting the party they voted for last time around.

There is of course a slight problem with this coming home to the roost assumption – some Lib Dem voters in 2010 may now be undecided on who they will vote for in three months, but some of those will already have decided that it won’t be them again. But for the rest, it is probably a fairly reasonable assumption. The real Liberal Democrat level of support is probably somewhere between the YouGov and ICM positions.

All of the three polls showing a Labour lead exclude undecided voters (even the Lord Oakeshott one carried out by ICM themselves). But the data behind the poll does show how those “don’t knows” voted in 2010. So what happens if we take an ICM-style approach to these Hallam voters?

The methodology below is not precise: all the figures used below remove the pollsters’ weightings, although those weightings in fact have a relatively minor impact on the headline figures.

So, let us have a look first at that Oakeshott poll. The graph below shows the numbers of identified voters for each party in the darker colour at the bottom of each column. Above that, in the lighter colour, are the undecided voters shown as returning to the party that they voted for in 2010.


Those figures are certainly worse for Clegg than the Liberal Democrat councillors achieved three weeks later, but not enough to remove him from Parliament. A similar picture emerges looking at the Ashcroft results:


Note: the unweighted identified support varies from the headline figures due to the weighting issue. It might therefore be that the Liberal Democrat lead is greater once the undecideds are added in if the same weighting was applied.

Finally, looking at the Survation/ Unite poll, there are again more undecided voters that were previously Liberal Democrat than for other parties. This shows the parties neck and neck, rather than a ten-point Labour lead.



The local election results show that, whilst Liberal Democrat support has fallen since 2010, they are still the dominant party at a council level. Whilst the polls suggest that Clegg may be less popular in his constituency than his party, they may not really be as disasterous as the headline figures show. Given the affluence of the seat and the reduced (if still important) influence of the student vote, there does not appear to be a particularly strong narrative for why voters who still pick the Lib Dems locally would abandon their party leader in the massive numbers needed for a Labour victory.

So, with some regret, my prediction is that Nick Clegg looks likely to continue to be Hallam’s MP.