What MRPs miss about local elections

“MRPs have become a mainstay of election coverage. They’re interesting and provide excellent insight into general trends and sentiment. They’re good at what they do. But what they do is model votes. At a General Election in the UK, this works well, but the story is more complicated for local elections, argues Jenevieve Treadwell . Enjoying this post? Then sign up to our newsletter and receive a weekly roundup of all our articles. Multi-Level Regression and Post-Stratification (MRPs) is a statistical modelling technique that combines large-scale polling with demographic and geographic data to estimate vote shares at a granular level. There were at least three MRPs just for London’s local elections this cycle — from YouGov , More in Common , and JL Partners . Coverage predicted a seismic shift: historic Labour losses, Reform breaking through, the Greens remaking the map. In the end, London was remarkable in how well the two largest parties held up. So, was it a polling miss? In fact, vote share projections were broadly accurate. The graphs below show the vote shares of the five largest parties in London against the projections from each pollster. The closer a dot is to the line, the closer it is to the actual result. Overall, the MRPs have a fairly similar level of accuracy. Each pollster’s “MAE” (mean absolute error), the average size of the error, is around 4pp. An MAE of 4pp means that, in a given borough, the pollster was off by about 4 percentage points on average. If the projections were broadly accurate, why did London’s political geography experience less severe shifts than many pundits predicted? Because MRPs model votes, not seats. In our electoral system, getting enough votes is not sufficient ; where you get them is also important. Winning a seat in a General Election is relatively straightforward. A party’s candidate needs to win more votes than the others across the whole constituency. What makes a constituency seat easier to win also makes it easier to predict. Local elections are neither. To control a borough, a party must win a majority of council seats, but the distribution of the vote means that the party in control may not necessarily be the most popular. A borough is not a single unit, as a constituency is, but is made up of multiple wards. Each of these wards has multiple ‘seats.’ Across London’s 32 boroughs, there are nearly 700 wards, with over 1,800 councillors. Each of these seats is its own competition, with dozens of candidates vying to win. Where the votes are won matters more than just the absolute number. These dynamics come into full effect at the ward level. In Bexley, before the election, many were predicting that Reform would be neck-and-neck with the Conservatives. And in vote-share terms, that was the case. Reform and Conservatives won 32.3 per cent and 36.5 per cent, respectively. But this did not translate into political power. Reform won seven council seats to the Conservatives’ 29 seats. In Wandsworth, Labour won the popular vote (33.6 per cent vs Con 31.2 per cent) but Conservatives converted this into more seats (29 of 58 seats). We can think of seat competitions as “won”, “competitive”, and “no-hope”. Winning results in the election of a councillor. Competitive means the candidate was within 10 per cent of the winning threshold. In these cases, differences in campaign, candidate, or strategy may have shifted the result. But for some candidates, there is no hope; they were more than 10 percentage points behind the winning candidate’s total vote. Even changes in campaign, candidate and strategy were unlikely to result in a victory. Reform’s experience in Bexley is almost entirely a story of inefficient votes. Before the election, many were predicting that Reform would be neck-and-neck with the Conservatives. And in vote-share terms, that was the case. Reform and Conservatives won 32.3 per cent and 36.5 per cent, respectively. This happened because Reform’s vote was spread too thinly geographically . The Conservatives enjoy extremely high levels of support in the south of the borough. And Labour’s relatively low borough-wide vote share (19.4 per cent) belies their extremely staunch support in the north of Bexley, often between a third and half of the vote. Labour’s concentration in the north allows it to win those seats, while the Conservatives’ concentration in the south allows them to do the same. But support for Reform in Bexley is less variable. Consistently winning around a third of the vote is rarely sufficient to beat the Conservatives or Labour in either of their strongholds. Vote inefficiencies are not only bad for the inefficient party; they can also damage other parties. For example, Wandsworth is dominated by Labour and the Conservatives, with the Greens still a minor party. They were competitive (within 10 per cent of the winning threshold) in just one ward, and the average Green candidate in Wandsworth was 887 votes short of the winning threshold. At this election, in efficiency terms, every vote for the Greens in Wandsworth was “lost”. But lost votes still affect the election. Labour and the Greens draw from the same broad pool of left-liberal voters. In any given ward, that pool is finite – and when these parties compete, those voters are split between them, sometimes leaving neither enough to win. Balham clearly demonstrates this. The best-losing Labour candidate got 2,364 votes, 94 votes fewer than the worst winner (a Conservative). And the best Green candidate only got 1,027 votes. If fewer than one-in-ten of those Green votes had gone to Labour instead, Labour would have gained an additional councillor. For insurgent parties, this is a lesson they must learn. A good election strategy can make a huge difference in a campaign. Targeting some areas (where you’re close second) at the expense of others, to redistribute resources and scrape a win. These choices over resources and priorities shape campaigns and are the lessons smaller parties need to learn to compete and grow. Until they get this right, we might see the continued dominance of Labour and the Conservatives in much of London. Enjoyed this post? Sign up to our newsletter and receive a weekly roundup of all our articles. All articles posted on this blog give the views of the author(s), and not the position of LSE British Politics and Policy, nor of the London School of Economics and Political Science. Image credit: Steve Travelguide on Shutterstock The post What MRPs miss about local elections first appeared on LSE British Politics .

Summary generated from the RSS feed of LSE British Politics and Policy. All article rights belong to the original publisher. Click through to read the full piece on blogs.lse.ac.uk/politicsandpolicy.