Which voters abandoned Labour in 2010?

Labour's would-be leader Ed Balls has launched an attack on what he sees as the 'traps' that could derail the party as it seeks to recover from electoral defeat. Writing in the Times, the former Schools Secretary says one such trap is potentially ceding the 'radical' centre ground of politics to the coalition.

Yet at the same time he warns that Labour may have become too zealous in its pursuit of the centre ground due to myths over the party's electoral fortunes.

The Claim

Assessing the groups of voters which failed to back Gordon Brown in May, Mr Balls argues that the party should not subscribe to "the myth that our biggest challenge is to "win back" middle-income voters. They largely stuck with us at the election while we lost the support of too many people on lower incomes who felt we were no longer on their side."

So are the statistics on his side?

Analysis

Full Fact contacted Mr Balls campaign team, who informed us that the source of his claim is the analysis of the 2005 and 2010 general elections produced by Ipsos-Mori.

The pollster actually analyses the voting patterns by social group, rather than income level, but nevertheless we confirmed with Mr Balls' office that he was in fact referring to social grouping rather than income level.

With this point in mind it seems the data does back up the point made in the Times this morning.

For social group DE, which comprises unskilled workers and the unemployed, there was a drop of 8 per cent in Labour's estimated share of the vote, while for those in category C2, namely skilled workers, there was an 11 per cent reduction in support.  

This is compared to a much smaller 4 per cent fall in support among the C1 (clerical/junior managerial) category, and a fall of 2 per cent among the AB professional and managerial group.

Such findings are backed up by research conducted by the company of US pollster Stan Greenberg, which again found that the biggest decline in support for Labour was among the DE group (down roughly 11 per cent) and the C2 group (down roughly 13 per cent).  

However this is not a phenonemon exclusive to the 2010 election. The trend for Labour to lose more voters from these two groups goes much further back.

This analysis of ICM polling figures shows the same trend from the 1997 election to the 2005 result. Despite electoral victory, Labour was losing support among the groups Ed Balls identifies.

Indeed when we spoke to Ipsos-Mori, we were provided with data suggesting the trend went back even further. We were told that in 1992 approximately two in five people who voted Labour were in the DE category, while analysis in 2009 this group represented one in four votes for the party.

Conclusion

Going on the data seen by Full Fact, Ed Balls claim seems largely justifiable, in the sense that support for Labour fell less among traditionally higher earning social groups, than for people such as unskilled workers.  

There is without a doubt more room for comprehensive analyses of the types of people who contributed to the electoral demise of the last Labour Government, but the use of the statistics available seems sound.

However while the numbers may pass the Full Fact test, we would quibble with Mr Balls' wording.

By talking in terms of low income groups, Mr Balls ignores the point that those in the C2 category may not necessarily be such low earners, as Anthony Wells of pollster YouGov explained.  

"I wouldn't classify C2 as a low income group though — I'd class AB as higher income, C1 C2 as middle income, DE as low income. C2 is classic white van man - people like plasterers, plumbers or electricians and so on — tending to be aspirational and not badly off.

"It would be wrong to assume there is necessarily a correlation of interests between C2 and DE voters," he said.

When we put this to Mr Balls campaign team they defended the phrasing, arguing 'low income groups' was 'slightly more comprehensible language' for most people reading the article.