Kenilworth and Southam Liberal Democrats

Summary

While national pundits are interested in the national picture, in individual constituencies electors often want to vote tactically to make a difference - in reality for the candidate who might be able to catch the front runner.

In Kenilworth & Southam the front runner is clearly the Tory, but because of the demographics only the Liberal Democrats could coalesce anti Tory votes to catch up and overtake.

Forecasters make a series of adjustments to the actual distribution of votes cast locally, to make the method look reasonable nationally, but it produces skewed results if the boundary changes are exceptional, as they are here.

We believe using old District elections is a poor basis for party political general election forecasting. There is much more likely to be unreliability because of personal candidate influence in District Council wards - less marked in the larger County Electoral Divisions. The 2005 County elections were held simultaneously with the general election in the same polling stations with the same turnout. This was when labour secured a large majority and is possibly a 'high water' mark.

Labour have not had any councillor elected anywhere in the constituency since 2002 (then it was one in Southam). Labour only polled 7% of the vote in 2009 and were either third or fourth in the seven divisions they contested, with the remaining three they did not even contest.

Objective

There are several national forecasting organisations, although they use subtly different methods between them. The primary object of these calculations is to give a national picture, so each takes to use a common approach across all constituencies within their own analysis, with the intention that it makes sense nationally. At a constituency level forecasting might be OK for modest boundary changes, but becomes unreliable for the 'all new' constituencies (of which there only 13 out of 650). Although they might not say so, the analysts can tolerate the greater error probable in these all new constituencies because of the greater validity of the 637 existing seats.

Within the 'all new' group there is a 'degree of change' with K&S being the 3^rd most changed in the UK. Even so, if the character of the surrounding seats had a 'smoothed out' representation of parties, the forecasting methods would be more reliable. The problem for the forecasters is where rural areas are split off adjoining conurbations; the Lab/Con/LD balance is often radically different between town and country. K&S is the most extreme example of this.

Forecasters' Assumptions

Forecasters apply a number of factors from local election results and come up with adjustments - all of which degrade the Liberal Democrat vote and inflate the Labour vote. They make several assumptions and use each one to apply adjustments to the numbers of actual votes cast.

The assumptions used by National forecasters are generally based on old district council or unitary council seats in an attempt to make it consistent across the country. Here we have both district and county councillors. Therefore the assumptions (some of which are questionable) made by national forecasters are that:

:

• Party political affiliation is the discriminator influencing choice of candidate at the district council level
• Wards are deliberately uncontested by parties even though there is significant electoral support within the Ward
• Uncontested wards can have party political percentages attributed to them by reference to nearby areas
• There are consistently 'weak' and 'strong' parties who are under and over represented at local level
• That tactical voting will be unaffected by boundary changes

Each of these assumptions is flawed to some degree and using a series of arithmetical and artificial "correction" factors distorts the results to a degree which is not credible.

1. A district councillor is broadly representative of 1800-2000 voters many of whom will know the candidate personally or by local reputation. Many voters make their selection on the personality of the candidate irrespective of party affiliation. This is not the case for county councillors who typically will represent 6000 electors, and will be less personally well-known than at district level. In this respect county elections are closer to general elections.

2. The main reason that District council seats are not contested is because of poor support for a political party in the area or simply a total lack of organisation. Not because they think they have representative support but then don't contest it anyway.

3. Allocating support pro rata from other wards, sometimes at many miles away and assuming voter behaviour in urban areas will be the same as rural areas is highly unreliable.

4. The assumption of weak and strong parties arises from the need to pro rata the different proportions of votes, cast across the country, contrasting between local and national elections. This is done on the basis of the national strength, or sometimes on regional strength. If the latter it relies on the constituency being in the correct region. (EC have it in the wrong region.)

5. Tactical voting tends to dominate in marginal constituencies, where voters are fully prepared to vote anti-Tory or anti-Labour, even if their sympathies are otherwise. Voters who find themselves 'moved' from one constituency to another are likely to vote differently.

Kenilworth and Southam

The pattern of voting considered for recent years is:

* BBC, ITV and SKY use published Univ of Plymouth figures, although use the method prescribed by the Univ of Plymouth gives different results to those published. The Univ is unable to explain this anomaly.

Weighting and adjustments

The forecasters have weighted real election results by assuming:

• That the strong Labour support in the centre of Rugby and Warwick/Leamington will carry over to rural areas.
• That turnout was different at local elections, which it wasn't in 2005 - it was identical.

The forcasters assume parliamentary voter intentions can be determined by looking back to various different District Council elections from 2003-06. This does not work well in K&S - some of the three Districts covered are "all up", some elect by thirds an in different years. To make matters worse some are multi member wards, some single member. Worse still, there are independent candidates and uncontested wards. And worse again, turnout tends to be low unless there is a general election at the same time.

Electoral Calculus has a particular assumption based on the national picture - that the Lib Dem vote is always diluted from local voting and Labour will always improve at a general election. Whilst it is true that a degradation of the Lib Dem vote does occur in response to the concentration on two party politics in the national media, (especially in Lab/Con marginals), this seems to be based in part on conviction that K&S is the East Midlands; whereas it is in fact the West Midlands.

Uncontested Seats

Basing the forecast on old District Council results gives a problem for uncontested Wards. University of Plymouth (Thresher and Rawlins) compensate for uncontested district council seats by assuming that they would have attracted a level of support equivalent to the minimum vote of an contested seat from the "donor" parliamentary area (Rugby 2004 in our case). Minimum Lib Dem vote in 2004 came from outside the new area - Lawford & Kings Newnham (11%); Labour from Dunchurch (9%). In their publication (Guide to the New Constituencies) the figures for K&S are different to that derived using this method, for reasons that the university cannot explain.

In Electoral Calculus, forecasted results the outcome is different depending on the party. If Lib Dem, the forecast results are predicted to still be zero (!); if Labour, EC add votes in the adjusted results. Whilst the Lib Dems may not see Lapworth as a strong area it beggars belief that not one single vote would cast.

Character of the area

The new Kenilworth and Southam constituency fundamentally has two forms of character. Kenilworth is a substantial south Warwickshire town. Labour has not held any council seats there at any level for many years. In 2009 they were in third or fourth place in electoral divisions that they did contest, with a maximum share of the vote of 15%. The Lib Dems were a substantial second place in every seat that they did not win.

The remainder of the constituency has a rural character very similar to the old Stratford on Avon constituency. The Lib Dems had been the principal opposition there since they ran the Council up until 2001.

No Labour councillor at town, district or county level has been elected since 2002 anywhere in the constituency.