Friday, 29 August 2014

Clarity and consistency

Clarity is a rare commodity. I rarely achieve it. At the very least the odd paragraph of writing does what I wanted it to do. The best I can hope for is that I make fewer mistakes than the next blogger. Consistency is simply unrealistic. I could post about Labours problems with UKIP in a semi-coherent way and then be told it runs contrary to a tweet I made thirteen weeks ago. Damn! So do I disown the former in favour of the latter? No, I put it down to the human condition. We're not robots, we don't work in a clear and consistent manner. At least I don't, and I'd be suspicious of anybody who claimed they did!

Which brings me to UKIP. Imagine a game of tug-of-war. In the middle of the rope is a knot. At one end of the rope are the voters UKIP are drawing from the Conservatives. At the other end are those voters UKIP are drawing from Labour. The problem the party faces is that in trying to appeal to both, which is of course imperative from their perspective, they risk only making the knot tighter and tighter. It's a very difficult balancing act, and calls for a good deal of sensitivity to both sides. Give too much to the one end of the rope and they pull the party in its direction, and in so doing the other end simply abandons the game.

Of course this is not a tug-of-war, nor are the opinions of UKIP-leaning groups at the opposing ends of a right-left continuum. Which brings me back to consistency and clarity. Is it really that important? Clearly not if you look at a party like UKIP. It seems perfectly at ease drawing support from both Labour and the Conservatives and, frankly, doesn't care much for internal consistency. So long as it can continue to tempt voters to either end of the rope it's happy.

However there is a problem coming. Because this can only go on for so long. In a few weeks the party will gather in Doncaster with the avowed aim of stealing Labours thunder in its own back yard. Yet it will need to do so whilst preparing for a by-election in a seat which bears little or no resemblance to a Labour heartland. So it will move to the left for conference, and to the right for a by-election. Is this achievable? In the short term the answer is probably yes.

However if, as I expect, there are further defections from the Conservatives to UKIP, and no Labour defections, the party is going to find it increasingly difficult to portray itself as the party of the working man. Defections are good for UKIP right now, but if they're only from one party then you begin to resemble that party. Voters notice this, just as they noticed UKIP way ahead of anyone in the Westminster village.

So, what does it do? Well, it has made its bed now. Attracted by the idea of a broad church it has accepted the findings of academics like Matt Goodwin and Rob Ford, that it appeals to blue-collar 'left-behind' voters as well as social conservatives. The party hopes it can keep the fires burning between now and election time and get there without any major snafu's. It hopes the tug-of-war continues without a victor. All the time however, as Khruschev once said, the knot gets tighter and tighter and tighter.

Tuesday, 26 August 2014

The Conservatives and ethnic minorities

This morning I have written a piece for The Times in which I summarize the problem the Conservatives have with ethnic minority voters. As I say in the piece, "For the Conservatives the message is clear: adapt, or face a smaller pool of seats they can win or even defend, a process which will lead to a long period on the opposition benches." For over a century the Conservatives have been heralded as the 'natural party of government'. However the demographic shifts taking place in our constituencies will consign them to the natural party of opposition unless they make an effort to appeal to ethnic minorities.

But first the data.....

I have compiled age and ethnic group tables for every ethnic category from the 2011 census, and then aggregated them into parliamentary constituency files. I am therefore able to say how many of each ethnic group and age group are in each constituency. So, for example, I can say that 64.9% of 0 to 4 yr olds in Bethnal Green & Bow constituency are Bangladeshi; or that 16.4% of 8 to 9 yr olds in Batley & Spen are Pakistani; or that 78.4% of 15 yr olds in Birmingham Northfield are white, and so and so on....

When analysing the nature of ethnic minority populations it is important to take note of geographical dynamics. Ethnic minority populations are concentrated in mostly urban populations, meaning they are clustered geographically in London, the Midlands, north west and many of our major towns and cities. This is a consequence of original settlement patterns. As Ceri Peach* and others have pointed out the original settlement of immigrants to this country from the 1950s onwards has not been spatially random. Notwithstanding more recent diffusion and assimilation by ethnic minorities the pattern has remained largely unchanged. However this is a problem for the Conservatives because in order to win a majority of seats at the next election it will need to capture seats which are either in urban or suburban areas. It's simple numbers; there aren't enough rural seats with high proportions of white voters to win anything like a majority.

The picture which emerges from my data is of a population which will become increasingly diverse. This is because the large proportions of older white populations will eventually pass away and be replenished by younger age groups which already have higher levels of ethnic diversity than ever before. At the constituency level it means that many seats will also become more diverse. Which is bad news for the Conservatives. Why? Because the Conservatives perform extremely poorly among ethnic minority groups. Take a look at this analysis of the 2010 general election: Ethnic Minority Voting on the 2010 General Election. The implications of this research are catastrophic for the Conservatives. Not only will they lose a good proportion of older white voters, one of their key major voting blocks, but they will be replaced by higher proportions of voters who distrust and despise them.

For example, according to UK Polling Report the seat of Hampstead & Kilburn is the Conservatives number 1 target seat. Not only that, but it is also on the list of 40/40 seats published by Mark Wallace at Conservative Home; a list of seats the party has apparently identified "because of the examination of factors including demographic trends and local issues." In Hampstead & Kilburn 66% of the electorate is white. However in 2015 only 45.4% of first-time voters will be white. In addition less than half of 10 to 14 yr olds in the seat are white. This latter group will be the first-time voters in 2020. As older white voters pass away the constituency will resemble the 50:50 white:non-white split it currently has amongst younger voters. This is a seat with a 0.1% majority for Labour over the Conservatives. Demographic change alone will give Labour an in-built advantage in 2015 and 2020 and perhaps beyond. The seat will inexorably move away from the grasp of the Conservatives.

I could name dozens of seats where this self-same process will be repeated. Seats like Thurrock in the south, Wolverhampton South West in the midlands, Dewsbury in the north. Seats like Bedford, Halifax, Derby North, Dudley North, the list goes on and on..... For the Conservatives the hope must be that their period of opposition does not go on and on and on.....

*Everyone interested in this area should find articles by Ceri Peach. Having listened to him speak on more than one occasion, and read his work, he's something of an academic hero of mine. I also recommend the work of The Centre on Dynamics of Ethnicity at my alma mater, the University of Manchester. Their website is here: and their twitter handle is @EthnicityUK.

Saturday, 23 August 2014

First-time voters in 2015

I have pooled together the data for first-time voters in 2015 from the 2011 UK Census. To do this I have summed those aged 15, 16 or 17 in April 2011 as they will be 19,20 or 21 in May 2015, and will be voting for the first time. I have disaggregated the data by ethnic group as well, meaning I am able to show the ethnic breakdown of first-time voters in every constituency in England. I can take requests for this data through the blog contact form, but there will be a small charge for providing such data, and I reserve the right to refuse requests.

In 2015 around 4.5% of the electorate will be voting for the first time. Of this number 82.7% will be white, 3.9% will be mixed race, 2.1% Indian, 2.5% Pakistani, 1.1% Bangladeshi, 0.8% Chinese and 4.2% black.

However the distribution of first-time voters is non-random, as the two maps below show. On the left is the simple distribution map showing the proportion of first-time voters in each of the constituencies. On the right is a cluster map showing showing those parts of England where there is statistically significant clustering, both high and low.

The constituencies with significant high clusters of first-time voters to the east of London are South Basildon & East Thurrock, Gravesham, Basildon and Billericay, Rochester and Strood, Gillingham & Rainham, Sittingbourne & Sheppey, Chatham & Aylesford, and Tonbridge & Malling. Each of those seats has a particular ethnic breakdown in terms of the proportion of first-time voters. However they are almost exclusively white, with the exception of Gravesham where 21.3% of first-time voters are non-white.

The cluster map also shows a large part of the west Midlands having statistically significant proportions of first-time voters. However the first-time voters in the West Midlands are a much more ethnically diverse group, which reflects the particular demographics of the populations themselves. For example in Birmingham Ladywood 12.3% of first time voters are white. Birmingham Hodge Hill has the highest raw number of first-time voters with 6,131, 43% of which are Pakistani and 25.5% white. The pattern is similar in Perry Barr, Hall Green and Yardley constituencies in Birmingham (high proportions of first-time ethnic minority voters).

Below is an interactive map showing the number of first-time voters for each constituency in England. Click on the individual constituencies to reveal a pop-up showing the number of first-time voters and the proportion of the electorate they comprise. Ethnic breakdowns are not part of this map, and are available on request.

Friday, 22 August 2014

West Midlands Police & Crime Commissioner election

Here are the vote share and turnout maps for today's PCC by-election in the West Midlands:

And here are the % increases/decreases in each parties vote share between 2012 and 2014. There were three Independent candidates in 2012, meaning that the numbers are slightly misleading but still interesting to note how the vote changed:

Wednesday, 20 August 2014

Ashcroft marginals, comment 1

The first point to make regards the 18 to 24 yr olds. These polls were conducted during the summer vacation for University campuses. This is not so important in a seat like Bolton West, which has virtually no student households (well, 0.2% of households actually!). However it matters more in Plymouth Sutton & Devonport where 11.7% of households are student households. There is a large weighting for 18 to 24 yr olds in Plymouth Sutton & Devonport (from 25 in the original to 99 in the weighted base). The pattern is repeated in virtually all of Lord Ashcroft's polls here; the younger populations are weighted up whilst the older populations are weighted down.

It's difficult to determine whether students will choose to vote in their 'home' constituency in 2015 or stay on campus and vote where they are. They are entitled to register in either or both if they like. This is an important point in a seat like Southampton Itchen where the margins are so tight. The seat has student populations in the west of the seat across the Itchen bridge. Voter registration is going to be absolutely crucial. Not only that but they will need to make sure the students choose to vote in the seat rather than returning home to safer Labour seats elsewhere.

Saturday, 16 August 2014

Uxbridge; a quick note on the Ashcroft polls

Many thanks [again] to Lord Ashcroft for his poll of Uxbridge & South Ruislip, published this morning. This is the seat Boris Johnson is allegedly considering very strongly. Lord Ashcroft's poll is intended as a snapshot of opinion. I don't think he would want large inferences to be drawn from crosstabs with relatively small numbers. However it is often the case that the immediate response to the poll simply looks at the headline numbers and draws incorrect and unreliable inferences. This is not Lord Ashcroft's problem, but a problem of interpretation.

For example: Labour and the Conservatives are tied at 34% among 18-24 yr olds. This may surprise some but not me. There are particular 'types' of 18-24 yr olds, and in Uxbridge they come from an entirely different demographic base than they would in, say, Manchester Withington. The 18-24 yr olds in Uxbridge are predominantly, though not exlusively, young singles and sharers in private rented accommodation. They command good salaries but, for reasons of housing affordability, and because they cannot be sure where they are likely to be employed on a long term basis, prefer to rent. This group does not align strongly with any particular party, as opposed to 18-24 yr olds in different circumstances around the country, which have aligned strongly with Labour or the Lib Dems in 2010. When Boris' name is added to the Ashcroft poll the Conservatives lead Labour by 45% to 28% among this group, again reflecting Boris' popularity amongst this particular group of 18-24 yr olds.

It is very important, when interpreting constituency-level polls, to be aware of the very particular demographic composition of each constituency before drawing inferences. Which is why nerds like me spend an awful lot of time pouring over the crosstabs of Lord Ashcroft's polls. Not because I think they are useful in determining who will win the seat but because they are extremely useful in aligning what I already understand about particular constituencies in terms of demographics and previous voting behaviour with what is reported in the poll. In some cases Lord Ashcroft's polls have simply confirmed what I already suspected and on other occasions they have conflicted with it. Neither eventuality should be overplayed. Even a confirmation should be treated with suspicion.

Thursday, 14 August 2014

We need to talk about the Green Party

There has been understandable attention paid to the emergence of UKIP as a major factor in 2015. However comparatively little attention has been paid to the Green Party. Typically when I mention the Green Party to politico's there's a palpable shrug of the shoulders and a sigh, as if it's somehow beneath them. I'm a cold fish, so a party taking 6% in national polling, in my view, is at least worthy of analysis. I want to make it clear that I am relatively new to Green politics so if some of the points raised here seem naive it's because they probably are! Go easy on me.

I will start with the polls. Depending on which pollster you believe the Greens are capturing around 5% in voting intention polling. Looking through the crosstabs you find that around 10% of LibDems are intending to vote for the Greens in 2015. Their support is strongest among 18 to 24 year olds, around 12% of whom say they are intending to vote Green.

However national voting intention polling tells us comparatively little about the differential impact the party will have in 2015. Clearly there are some seats where the party either has a strong political base and good demographics, seats where they should expect to poll higher than the 5% recorded in VI polls. Seats like Brighton Pavilion or Manchester Withington, for example. Lord Ashcroft's poll of LibDem-Lab marginals had the Greens at 20% in Norwich South and 32% in Brighton Pavilion, the latter of which surprised me a little given the apparent unpopularity of the council in Brighton.

In terms of areas of the country where the party has a strong political base and a relatively good recent record in local and European elections there is a comparatively long list of successes for a party which claims so little attention in the national media spotlight. The maps below shows their vote share record in the 2014 European elections (I am including the Scottish Greens although I know they are a separate entity to the Green Party of England & Wales):

The party performed strongly across the south east, London, and south west and up into East Anglia. In terms of vote share the party achieved 6.9% of the popular vote and won three MEPs as a result (one each in London, the south west and south east). In local authorities like Brighton & Hove, Norwich, Oxford, Cambridge, Bristol, South Hackney, Hams, Stroud, Bath, York, Haringey and Lambeth the party performed much better than its national vote share.

As Patrick O'Flynn has pointed out, the Greens could eat into Lib-Lab contests in places like Cambridge:

Of course Mr O'Flynn would say that, given that he has announced his intention to put his name forward as the UKIP candidate in the seat. However the recent local elections in the seat point to a pronounced uptick in support for Labour from 2010. Labour's vote share in the Cambridge local elections was up 16% on 2010. In fact even the Greens saw a slight vote share dip of 1.3% from 2010 to 2014. The two biggest losers in Cambridge were the coalition partners. The LibDems saw their vote share drop from 37.8% in 2010 to a still respectable 29%, whilst the Conservatives saw their vote share fall from 22.5% in 2010 and 14.4% in 2014. UKIP stood just three candidates in the Cambridge local elections, polling 0.4% in total.

Making projections based on local and European elections is dangerous given the relative turnout in those elections compared to a general election. Rather it would be more appropriate to consider which demographic groups voted for the Greens in the May 2014 local elections and to then identify seats where such demographic groups account for a higher proportion of such groups.

My analysis has shown that the following demographic groups voted for the Greens in 2014:

Well educated singles living in purpose built flats
City dwellers owning houses in older neighbourhoods
Singles and sharers occupying converted Victorian houses
Young professional families settling in better quality older terraces
Diverse communities of well-educated singles living in smart, small flats
Owners in smart purpose built flats in prestige locations, many newly built
Students and other transient singles in multi-let houses
Young renters in flats with a cosmopolitan mix

However the Greens are in a fight with the other parties for these voters come the general election. The key for the party is to identify those seats where these groups account for very large proportions of total households. A list of the top ten seats with such demographics is:

  • Cardiff Central (41.1% of households) - 13% LibDem majority

  • Sheffield Central (39.7% of households) - Lab maj. of 0.4% over LibDems

  • Bristol West (37.6% of households) - 21% LibDem maj over Lab

  • Brighton Pavilion (36.7% of households) - Green-held

  • Manchester Withington (33.9% of households) - LibDem maj of 4.1% over Lab

  • Wimbledon (33.6% of households) - 24% Con maj over LibDems

  • Newcastle-upon-Tyne East (30.6% of households) - 12% Lab maj over LibDems

  • Tooting (30.6% of households) - 5% Lab maj over Con

  • York Central (29.5% of households) - 14% Lab maj over LD/Con

  • Hove (28.7% of households) - 3.7% Con maj over Lab

  • In many of these seats the Green Party can reasonably expect to increase its vote share, with the possible exception of Brighton Pavilion where one can plausibly make the case that every percentage point of Green support was maximised to the full in 2010. The assumption at this stage must be that the LibDems will lose many voters from 2010, and a good proportion of them have shown in the local and European elections that they are not averse to voting for the Greens. The challenge for the Greens is to retain those voters in 2015. If the polls are to be believed then 10% of 2010 LibDems have already decided to vote for the Greens in 2015. If these intentions do convert to votes it will hardly shake the political landscape, but it is at least worthy of some attention.