A look back at the polls: March 2009

by Stephen Tall on April 5, 2009

We tend not to be too poll-obsessed here at LDV – of course we look at them, as do all other politico-geeks, but viewed in isolation no one poll will tell you very much beyond what you want to read into it. Looked at over a reasonable time-span and, if there are enough polls, you can see some trends.

Here, in chronological order, are the results of the eight polls published in March:

Tories 42%, Labour 30%, Lib Dems 19% – Populus/Times (9th March 2009)
Tories 41%, Labour 31%, Lib Dems 17% – YouGov/S. Times (15th March)
Tories 42%, Labour 32%, Lib Dems 14% – Mori/unpublished (17th March)
Tories 42%, Labour 30%, Lib Dems 20% – ICM/Guardian (18th March)
Tories 41%, Labour 30%, Lib Dems 17% – ComRes/Independent on S. (22nd March)
Tories 41%, Labour 31%, Lib Dems 17% – YouGov/Telegraph (27th March)
Tories 44%, Labour 31%, Lib Dems 18% – ICM/S. Telegraph (29th March)
Tories 40%, Labour 28%, Lib Dems 18% – ComRes/Independent (31st March)

Which gives us an average rating for the parties in March as follows, compared with February’s averages:

Tories 42% (-1%), Labour 30% (+1%), Lib Dems 18% (n/c)

March was a much more steady month for polling – in February, you may recall, polling was all over the place, with the Tories were scoring anywhere between 40-48%, Labour 25-31%, and the Lib Dems 14-22%. Currently we see a more steady situation, with the Tories polling in the 40-44% range, Labour 28-31% and the Lib Dems – still the outliers – 14-20%.

April may prove to be more ineresting and fluid, as we see if the aftermath of the G20 summit reaps polling dividends for Labour, and as the parties begin active campaigning for June’s European and local elections.

The biggest polling moves was in the assessment of the party leaders, with David Cameron receiving a significant boost, most likely associated with the understandable public sympathy following the death of his son, Ivan. In February, Mori’s satisfaction rating for Mr Cameron was +9%; YouGov’s was +12%. By March, this had more than doubled to +22% and +29% respectively.

Gordon Brown’s personal ratings have jumped about at least as much. In February, his satisfaction rating was well in the negatives at -34% (YouGov) and -38% (Mori); by March these had climbed from these depths to -25% for both YouGov and Mori. (Though a second YouGov survey, asking just about Brown and Cameron, showed his personal rating at -42%, the worst since the pre-Lehmen’s collapse leadership crisis of August/September).

Nick Clegg’s ratings have remained steady and moderately positive, with YouGov once again giving him a +4% rating, and Mori showing him at +12% (against +9% in February). The challenge for Nick remains moving the one-third of voters who currently have no opinion at all about him currently into the positive camp.

With the general election now almost certain – save exceptional circumstances – to be held in May/June 2010, it’s worth reminding ourselves how current Lib Dem ratings compare with other Parliamentary cycles 15 months from the national poll:

March 2009 LD average: 18%
February 2004: 23%
March 2000: 13%
February 1996: 16%
January 1991: 9%

In this historical context, then, Lib Dem ratings are holding up very well. Of course, the unknown factor – as those who are ‘bearish’ when it comes to Lib Dem poll ratings will note – is that the Tories’ ratings are considerably higher than has been the case at any other time post-’97, and most Lib Dem seats have a Conservative challenger.

Against this, a more ‘bearish’ viewpoint, is the known incumbency boost from which Lib Dem MPs disproportionately benefit, and which has been reflected in recent years’ local elections – how many people actually think, for example, that Chris Huhne in Eastleigh or Norman Baker in Lewes will be defeated, whatever the national polls might say?

Truthfully, and tritefully (I admit), I don’t think we have an earthly hope of knowing the answer to who is right, the ‘bears’ or the ‘bulls’ – or indeed whether Tory losses in the south will more or less be offset by gains from Labour in the north and elsewhere – until the results are actually in. Which is, of course, why poll analysis is such speculative fun.

Enjoy reading this? Please like and share:

No comments

If people haven’t seen it todays You Gov is Con 41 N/c Lab 34 +3 LD 16 -1. Though its worth noting that the field work was crammed into two days immeadiately after the G20.

As ever a very fair write up which takes the issue as far as it can reasonably go without massive amounts of detail.

I think there will be be a case for a public debate between the bears and the bulls on this subject however not until after the euros which in them selves will give hard data for comparison with the last cycle.

by David Morton on April 5, 2009 at 2:33 pm. Reply #

Since we have an FPTP voting system rather than a proportional system, percentage poll ratings can provide only a rough guide.

It’s seats that matter, and here a better guide is the spread betting market (which was virtually spot on in the 2005 general election).

I have just checked the Sporting Index site, which currently has the Tories at 345-350 seats, Labour at 230-235 and the Lib Dems at 43-46 (the ‘sell’ price is first and the ‘buy’ price second).

Extrabet’s current spreads are similar: Tories 344-350, Labour 228-234, and Lib Dems 45-48.

Meanwhile at Ladbrokes, the shortest odds (9/4) are on 40-49 Lib Dem seats. William Hill is offering 2/1 on 40-49 seats.

by Simon Titley on April 5, 2009 at 4:22 pm. Reply #

Mark – In 2005, the spread betting markets were spot on by about three or four months before polling day.

Of course, assuming an election in May or June 2010, we are further away than that. In the intervening period, circumstances can change.

If Lib Dem prospects improve, we can expect the spread betting markets to move upward from their current prediction of around 45 seats.

However, one should not assume that “on past form the party will do better than those numbers” because the next general election won’t be like the last three. The reason the spread betting markets predict a net loss of Lib Dem seats is that, unlike in the previous three elections, the Tories are doing much better than Labour.

My point is that the spread betting markets provide a better guide than the polls to the likely share-out of seats, assuming there are no significant changes in the meantime.

Remember, the value of your investment can go down as well as up.

by Simon Titley on April 5, 2009 at 10:18 pm. Reply #

Populus/The Times

Con 43 +1
Lab 30 n/c
LD 18 – 1

by David Morton on April 6, 2009 at 7:58 pm. Reply #

Mark – My basic point is not confusing.

It is that, if you want a prediction of the number of seats, then the spread betting markets are a better guide than opinion polls.

Of course, both the spread betting markets and the polls are snapshots of opinion now and are subject to change.

What ought to concern us that the spread betting markets currently project a net loss of 15 to 20 seats for the lib Dems when the poll ratings seem to be no worse than in previous parliaments.

The opinion I dispute is that, because Lib Dem poll ratings are similar to those at an equivalent stage in previous parliaments, we can confidently expect a similar or better result in terms of seats.

With a clear Tory lead over Labour, that is not an assumption one should make automatically.

by Simon Titley on April 8, 2009 at 7:57 am. Reply #

Seven out of eight of these polls show the Tories doing less well (albeit slightly) than in 1979, 1983, 1987 and 1992.

At the equivalent stage in the 1966-70 and 1974-79 Parliaments (when there were unpopular Labour governments), the Conservatives were doing much, much better.

Cameron’s strategists must be worried that during the course of the campaign they will slip into the 30s as the Lib Dems pull up.

In my judgment, an overall Tory majority is unlikely.

by Sesenco on April 8, 2009 at 8:50 am. Reply #

Simon – you seem to be implying the spread betting markets and the opinion polls are unrelated to each other. I’d have thought the former drives the latter – whether punters reacting to polls or trying to anticipate polling movements.

Before concluding spread betting markets are better predictors we would need to see significant eveidence over a period of time. Referring to one period of time in 2005 isn’t exactly conclusive.

We’d also need to know about the volumes of those markets, too – eg, has the existence of PoliticalBetting.com increased the level of punters betting on political outcomes? Is the political betting cognoscenti being diluted?

I’m not sure it’s possible to know the answers to these questions – but I’d want to before using spread betting as a more reliable predictor than polling trends.

by Stephen Tall on April 8, 2009 at 9:19 am. Reply #

Stephen – There is no doubt that PoliticalBetting.com has increased the interest in the spread betting markets.

But participants in that market remain more likely to be the cognoscenti, insiders with a relatively sophisticated knowledge of individual constituencies, because I can’t see how this market would appeal particularly to many punters outside this group.

Therefore punters in the spread betting markets are unlikely to be ‘driven’ by the polls; rather, the polls will be one element of a more sophisticated judgement.

My original point still stands: if you want to judge projections of seat totals (as opposed to the overall percentage vote), the spread betting markets are a better guide than nationwide opinion poll results. Not perfect but better.

by Simon Titley on April 8, 2009 at 7:25 pm. Reply #

Leave your comment


Required. Not published.

If you have one.