Monday 14 July 2014

New Elections etc site: ElectionsEtc.com

A former student of mine, Jonathan Jones (@JJ_159) has kindly set up a new combined general election forecast page and blog at ElectionsEtc.com with a twitter account @ElectionsEtc. Please follow those links for future updates and discussion of the forecast.  

I’m very grateful to Jonathan who suggested the new site, set it up and is managing it all for free in his spare time. I think it is great. I hope you like it.  

Although Jonathan is a Lib Dem, the forecast and commentary will remain a non-partisan data-driven statistical exercise. 

There is no change in the methodology associated with the site move. This blog will stay up for the record, but will no longer be updated.

Friday 4 July 2014

Why is the 2015 general election forecast trending?

A few people have asked me this question about my forecast at http://users.ox.ac.uk/~nuff0084/ge15forecast/. The trends are definitely there as you can see from the graphs below. But since they are well within the very broad prediction intervals, there is a danger of reading too much into them. Certainly we are far from having enough information to say the model isn't working well for this electoral cycle.

In fact, it is more remarkable that public opinion and the forecasts haven't drifted or fluctuated more over the last nine months than it is that the forecasts show a bit of a trend from the initial forecast.  The current forecast is very close to the original one given the width of the prediction interval.

That said, it is worth saying something about the reasons for the trends. Some of them were expected and some are due to deviations from the expected trajectories of the parties in the polls.  I'll start with the latter and then return to the former.

First, then, trends in the forecasts partly reflect trends in the polls relative to their expected trajectories which are based on the average historical patterns in electoral cycles for different parties according to whether they are in government or opposition and their results at the last general election.

The Conservatives and Liberal Democrats were expected to recover some of their mid-term losses over the last nine months. In fact the Conservatives have roughly been level pegging around 32% in the polls since October, so their forecast election share has declined slightly. They haven't made the recovery that history suggests they ought to have done since the autumn.

The Liberal Democrats were steady in the polls at 10% but have dropped a bit to 8% since the local and European elections. So their forecast general election share has declined from 16% to 12%.

Labour, meanwhile, has seen its poll share decline roughly by the average from previous electoral cycles, and so their forecast share of the vote has remained pretty steady around 32%. Arguably by matching expectations Labour have done better than the Tories who failed to meet them over the last nine months.

As a result of a declining Conservative forecast share and steady Labour forecast share the chances of a Conservative lead in votes or seats has drifted down, and conversely the probability of a Labour lead has drifted up in the graph below. It still remains more likely that the Tories will emerge ahead. But 57% probability is not much greater than a half.

The seats prediction without the prediction interval (uncertainty range) has always pointed to a hung parliament.  With the prediction interval there is a slightly better than 50% chance that one or other of the two main parties will emerge with a parliamentary majority. Since October the chances of a hung parliament have increased from 40% to nearly 50%.

This is firstly because the forecast number of Tory seats has dropped (from 315 to 305) but also because the uncertainty around this forecast has declined. There is now a greater chance that the result will be close to the main estimate than there was in October. This is partly because there is less time for change, and so polls 300 days from the election are more informative about the eventual result than polls 600 days before.

This is the sense in which trends in the forecast were expected. It was expected that the prediction intervals would narrow slightly with time and so (without big changes in the seats forecast) the chances of a hung parliament were also expected to increase as the chances of big changes in before the election decline.

While declining, the chances of big changes between now and the election remain substantial. The prediction intervals on the shares of the vote differ by party, but they are still now around plus or minus 7. Plenty of room for both a big Labour or big Conservative lead to emerge.




Monday 26 May 2014

What do the 2014 European and local election results mean for the opinion polls and next year’s general election?

Labour emerged narrowly ahead of the Conservatives in both the local and European Parliament elections. I discussed the implications of the local election results in a previous post on Friday. The results of the Euros only came through last night.  This post considers the lessons learnt from both elections for the general election, including the likely accuracy of the opinion polls.

The shares of the vote in Table 1 below are all comparable in the sense that they all regard voting across Britain as a whole. There are some methodological issues that explain differences (e.g. locals figures are projections from parts of England to the rest of GB; polls are polls and not votes) but the main differences are due to the kind of election.

People vote differently in locals, European and general elections. Not only is it wrong to suggest that either the local or Euro election vote share is a forecast (or even nowcast) of the general election, it was not even possible to tell what the European election result was going to be from the (earlier-released) local elections on the same day.

Table 1. Percentage shares of GB vote 2014

2010 GE
vote
Locals PNS
Locals NEV
Euros vote
GE polls
Con
37
29
30
23.9
32
Lab
30
31
31
25.4
35
LD
24
13
11
6.9
9
UKIP
3
17
18
27.5
14
Green
1


7.9
4
Other
5
10 (inc Grn)
10 (inc Grn)
8.4
6






Lab lead over Con
-7
+2
+1
+1.5
+3






PNS: BBC Projected National Share of the vote
NEV: Sunday Times, Rallings and Thrasher National Equivalent Vote
Polls: UK Polling Report Polling Average on election day (22nd May 2014)

The differences between locals, Euros and general election voting follow a familiar pattern. For most countries, smaller parties do better and governing parties suffer in European Parliament and local elections relative to national general elections. These long-established phenomena are also visible in mid-term general election polls, but much more dramatically in the local and Euro elections.

In Britain, since at least 1999, more Eurosceptic parties (especially UKIP but also the Conservatives) have tended to do rather better in the European elections than in the local elections on the same day. Conversely, as the most pro-European major party, the Liberal Democrats have typically done much worse in European than general elections or local elections where they do best. This pattern was repeated again, probably more strongly, this year.  It is remarkable that the UKIP share of the Euro vote was a full 10 points higher than the average of the estimates of the local election vote share (PNS and NEV).

Such and other differences between local, European and general election voting behaviour suggest general election vote intention polls are a better guide to the next general election vote share than either local or European election shares. But there are still some lessons from the elections both directly for the general election next year and for our interpretation of the opinion polls.

Protest voting and government recovery?

There is a tendency in Britain for governments to recover from mid-term setbacks. While in some respects 2009 provides a good example and basis for comparison with 2014 (both saw Euro and local elections a year before the general election) it may be an extreme case given the MPs’ expenses crisis broke during the campaign. 

Nonetheless, the pattern of change in Table 2 is indicative of a more general one from other election cycles. It shows that a government recovery can be substantial. Gordon Brown managed to claw back a full 7 points in the year before the general election.

Table 2. Percentage shares of GB vote 2009: Local and Euro elections, and General Election opinion polls (with 2005 and 2010 GE)

2005 GE
vote
Locals PNS
Locals NEV
Euros vote
GE polls (May 09)
2010 GE
vote
Con
33
35
35
27.8
39
37
Lab
36
20
22
15.7
23
30
LD
23
26
25
13.8
19
24
UKIP
2
10
10
16.5
9
3
Green
1


8.6

1
Other
5
 9
(inc Grn)
8
(inc Grn)
17.6
10
(inc Grn)
5







Lab lead over Con
+3
-15
-13
-12.1
-16
-7







PNS: BBC Projected National Share of the vote
NEV: Sunday Times, Rallings and Thrasher National Equivalent Vote
Polls: Rallings and Thrasher polling average for month prior to election day (4th June 2009)

Labour’s recovery halved the lead of the Conservatives between 2009 and 2010.  Although the levels of the Conservative and Labour vote shares are rather different in the polls, locals, and especially the Euros, the Conservative leads over Labour are more similar and tell the same story: mid term setback followed by government recovery, albeit not a big enough one to secure re-election.

Polling accuracy

One reason why some commentators look to local and Euro elections provide an insight into the prospects of the parties for the next election is that they are, so the cliché runs, real votes in real elections.

But opinion polls still provide the best guide.  The question then is whether local and Euro elections tell us anything about the likely accuracy of the polls at the general election?

For good reasons, there were no attempts (I know of) by pollsters to forecast a national local election share of the vote. The European elections, however, provide the only opportunity before the general election to assess the ability of the pollsters to forecast the share of the vote in a Britain-wide election.

There is an excellent polling post-mortem by Anthony Wells with the detailed figures here. Some of the newer internet pollsters hugely overestimated UKIP (by 3 or more points).  There were no telephone polls so we cannot say whether they would do any better. (They were not statistically significantly better or worse in 2010.) But clearly some of the higher estimates of the UKIP share in general election vote intention polls in recent years seem to be problematic in light of the Euro results.

Perhaps most importantly for the general election, all of the final Euro polls (apart from ICM) over estimated Labour and under estimated the Conservatives.  The average (including ICM) of the six polls put Labour 0.8 points too high and the Tories 1.8 points too low. All suggested Labour leads of at least 3 points. More commonly the lead was 4 points, with the highest up to 7 points.  The actual lead was just 1.5 points.

If this kind of bias occurs in the polls in the run up to general elections it could be seriously misleading as to the likely outcome. Even though the polls did not under-estimate Labour in 2010, it was one of just two elections since 1974 where they did not, perhaps because it was one in which Labour were at an historically low point.  The polls did overestimate the Tories in 2010, and in every other election since 1974 apart from 1992.

So the direction of the bias in this year’s Euro opinion polls fits a long-standing tendency. Table 1 above also provides some indication that there may currently be a pro-Labour bias in the Labour lead in general election vote intention in the polls. Whereas the Labour leads in the PNS, NEV and Euros were between 1 and 2 points, that in general election vote intention is 3 points.  Well within the bounds of statistical significance and the polls have varied, but the difference is in a direction that suggests a traditional bias might currently be in operation.

Putting the government recovery and polling bias phenomena together

My long-range general election forecasting model assumes that future bias in the polls is more likely to be what it was on average in the past than to completely disappear in 2015. As a result the forecast is for a larger forecast Tory lead in 2015 than one would expect just from the tendency for governments to recover and oppositions to fall back in the final months.

Based on the general election voting intention polling average in the table above, my forecasting model suggests the Conservatives are likely to emerge with a clear lead in the share of the vote and 60% chance of being the largest party in Parliament.

The performance of the Conservatives and Labour relative to each other in the opinion polls and elections last week provides more comfort than concern for quality of the model. Chris Prosser’s model forecasting general election results from local election results alone produces a remarkably similar forecast share of the vote.

The biggest risk for the model is not so much that UKIP might get a very large share of the vote in 2015: they are already forecast to get 12%.  Rather it is the (broadly) uniform national change assumption that looks like potentially being the weakest link.

One key question is which of the two main parties UKIP will take votes of more in the key marginals compared with overall.  As discussed here, it currently looks like the overall damage to the two main parties from UKIP is roughly equal. Even if it was not, the overall hit is reflected in the overall share of the vote and so in the uniform change projections. More importantly, there was no evidence in the local election results that the swing between the Conservatives and Labour was any different in marginal constituencies to elsewhere.

The second question is whether UKIP can mount successful campaigns to win particular constituencies. Even on the basis of their stunning results last week, it will be difficult for them to win more than a dozen, or even a handful, of seats in the Commons.  The party has proved its ability to organize, but the UKIP vote is geographically less flat, but even in this year’s elections their highest peaks were still lower than those of the main parties.


Acknowledgements: Thanks to the BBC, John Curtice, Rob Ford, Jon Mellon, Rosie Shorrocks, and Michael Thrasher for their help with both data and analysis. 

Friday 23 May 2014

Unraveling the 2014 local election changes in the share of the vote: who suffered most from UKIP?

A naïve glance at the overall changes in the share of the vote since 2010 in the table below suggests that UKIP’s big gains came as the expense of the biggest losers, the Liberal Democrats.


Change % vote since 2010
Change % vote since 2012
Con
-6
-0.2
Lab
+5
-8
LD
-14
-3
UKIP
+13
+10
Grn
+4
+2
BNP
-2
-0.2
Oth
+0
-0.6
(Overall changes in the share of the vote across the BBC’s 1008 Keywards.)

This is clearly one of those cases where the net changes in the share of the vote can be misleading.  The opinion polls have commonly told us that Labour have benefitted primarily from the collapse of the Liberal Democrats, whereas the UKIP surge has come mostly from the Tories. 

If all the Lib Dem losses since 2010 were to Labour then they should be up 14 not just 5. The Green 4 point gain is likely to have largely come from the LDs, but that would still leave Labour 5 points short of where they would have been if they were simply benefitting from the Lib Dem losses.

Conversely, on the right of the political spectrum, if, apart from 2 points from the BNP, all of UKIP’s 13 point rise since 2010 were from the Tories, then Conservative would have been down by 11 points, not 6 they were actually down by. This discrepancy of 5 points corresponds to the 5 points Labour seem to have been short by in the above paragraph.

On this rough reckoning then, UKIP took 5 points from Labour, and 6 points from the Conservatives.  So, by comparison with 2010, it looks like Farage hurt both major parties roughly equally, not disproportionately from the Tories.

The story about change since 2012 also in the table is also revealing here.  That shows much more clearly how Labour have fallen and UKIP risen in local election vote shares over the last two years.  What seems to have happened is that between 2010 and 2012 UKIP took votes mainly from the Conservatives, but between 2012 and 2014 they have had more success in attracting Labour voters. The net effect is that the UKIP rise from 2010 to 2014 has been at similar expense to Labour and the Conservatives.

This is certainly a story that fits the main pattern of change in the general election vote intention opinion polls. The narrowing of the Labour lead over the past two years has been mainly due to a Labour fall and UKIP rise.

Another way of looking at this issue is to consider who suffered most in the wards where the UKIP vote share went up most.  In BBC Keywards fought by all four parties, the correlation between the change in the UKIP share of the vote since 2010 and that for the Conservatives was very similar to that between UKIP and Labour: both close to -0.3.  This provides further evidence that UKIP hurt Labour by about the same amount as they hurt the Conservatives.

This pattern is also visible in the main source of regional variation.  Across wards fought by all three main Westminster parties, UKIP were up by just 6 points in London, but by 14 points elsewhere in England: an 8 point difference.  The corresponding (partially) compensating London/elsewhere differences in the performances of the Conservatives, Labour and Liberal Democrats were 2.0, 2.6 and 1.4 points respectively.  These figures not only reinforce the picture of Labour suffering similarly if not more than the Tories from UKIP, but they also show an (albeit smaller) impact on the Liberal Democrats.

So much for votes. What about seats? Among the 53 UKIP gains in the BBC Keywards, 22 came from the Conservatives, 19 from Labour and 12 from the Lib Dems.  So a similar pattern again. 


Acknowledgements: Thanks to the BBC, John Curtice, Rob Ford, Jon Mellon and Rosie Shorrocks for their help with both data and analysis.

What do the 2014 local election results mean for next year’s general election?

The Labour lead of 2 percentage points in the BBC’s projected national share of the local election vote (PNS) is too narrow a lead for Labour to suggest they will be ahead in a general election next year. 

Since it was first calculated (in 1983) oppositions have usually had bigger leads in the PNS in the year before a general election, but they still went on to lose the next election.  The changes of government (in 1997 and 2010) were preceded by very big leads for the opposition in the PNS (16 points for Labour in 1996 and a 15 point lead for the Tories in 2009).

The message from the PNS is in line with general election vote intention in the opinion polls, which have on average also shown a narrow lead for Labour recently.

Even though their PNS is lower than the 23% they scored last year, the UKIP challenge remains significant. But since they seem to be hitting both main parties roughly equally, it is perhaps more important to think about both the impact on the major parties and it’s distribution across the country.

A uniform swing projection of the House of Commons from the PNS suggests that UKIP would not win any seats. This is partly because in 2010 their share was very evenly distributed.

There were some signs last year of UKIP’s vote becoming more concentrated in areas with more older people, fewer graduates and fewer ethnic minorities. 

UKIP share became still more variable this year. They went up more where they started strongest. So they are building up strong support in particular places, which is what they need to do to win seats in Westminster.  Indeed UKIP won the share of the vote across all the wards in the Great Grimsby constituency. This was the seat that Robert Ford and Matthew Goodwin predicted would be the most fertile territory for a UKIP general election victory.

(Rotherham has also been mentioned as an effective UKIP constituency win, but it was not part of the BBC keywards data collection so I’m not in a position to be able to confirm.)

Moreover, there are several seats where, based on the results of this week’s local elections UKIP would not be far behind the winning party. These include Portsmouth North (5.9% behind Con), Portsmouth South (6.1% behind LD), Southport (9.3% behind LD), Bradford South (9.4% behind Lab). 

Just as in local elections, UKIP are more likely to win seats in the Commons where the share of the vote between the three main parties is more evenly divided. In a general election if it becomes apparent that UKIP are a serious challengers this will lead to the kind of attention that might generate strong efforts to encourage tactical voting both for and against UKIP. So it may be that UKIP will find it difficult to divide and conquer.


Acknowledgements: Thanks to the BBC, John Curtice, Rob Ford, Jon Mellon and Rosie Shorrocks for their help with both data and analysis.