After a very late night following the UK general election results, it’s clear that YouGov’s final projection of our MRP model properly told the story of the 2024 election. We called 92% of the seats correctly, and our model correctly projected: - Labour’s landslide victory - The extent of the Conservative vote collapse - The nearly perfect projection of Liberal Democrat seats - The Green party and Reform UK as growing electoral forces This was not a straightforward election. The electorate has shifted significantly over the last decade in a number of ways and various voter groups that were previously closely aligned to particular parties have become unmoored. A related point is that the 2019 Tory election-winning coalition – a unique blend of north and south cultural and economic conservatives and liberals – has fragmented beyond recognition in the past five years, with each of Labour, Reform UK, the Liberal Democrats, and the Greens benefiting. All of this made political modelling for the 2024 election especially difficult. However, a number of factors helped us. One of the most significant was a market-leading and completely new innovation; our “unwinding” algorithm which helped us deal with the unique, high-change situation which the data was pointing toward and for which we had no precedent or prior data to help guide us. Another was the quality of our data. Machine learning models such as MRP work best when high quality data is fed into the model. Because YouGov owns our own panel, we have a unique relationship with the people taking our surveys and have a deep understanding of who our panellists are. This helps us build and execute market-leading models at scale which can also be used for our clients to answer their own business challenges. There are things in the results that we will need to build into our future work. As a data and analytics business, we constantly look for ways to improve our methods and we will be doing a thorough review of any areas where we can take learnings from for next time to make our MRP even more accurate and how we can apply our success into other areas of YouGov’s work. Read more about YouGov’s projections compared to the election results and learn more about what made our MRP model work so well: https://lnkd.in/eDrgK5tv
YouGov’s Post
More Relevant Posts
-
Well done to YouGov's Political Team and all that support them in the background! I love seeing your passion and energy in the office. It makes me feel proud to be in your wider Public Team. YouGov called 92% of the seats correctly, and our model correctly projected: - Labour’s landslide victory - The extent of the Conservative vote collapse - The nearly perfect projection of Liberal Democrat seats - The Green party and Reform UK as growing electoral forces
After a very late night following the UK general election results, it’s clear that YouGov’s final projection of our MRP model properly told the story of the 2024 election. We called 92% of the seats correctly, and our model correctly projected: - Labour’s landslide victory - The extent of the Conservative vote collapse - The nearly perfect projection of Liberal Democrat seats - The Green party and Reform UK as growing electoral forces This was not a straightforward election. The electorate has shifted significantly over the last decade in a number of ways and various voter groups that were previously closely aligned to particular parties have become unmoored. A related point is that the 2019 Tory election-winning coalition – a unique blend of north and south cultural and economic conservatives and liberals – has fragmented beyond recognition in the past five years, with each of Labour, Reform UK, the Liberal Democrats, and the Greens benefiting. All of this made political modelling for the 2024 election especially difficult. However, a number of factors helped us. One of the most significant was a market-leading and completely new innovation; our “unwinding” algorithm which helped us deal with the unique, high-change situation which the data was pointing toward and for which we had no precedent or prior data to help guide us. Another was the quality of our data. Machine learning models such as MRP work best when high quality data is fed into the model. Because YouGov owns our own panel, we have a unique relationship with the people taking our surveys and have a deep understanding of who our panellists are. This helps us build and execute market-leading models at scale which can also be used for our clients to answer their own business challenges. There are things in the results that we will need to build into our future work. As a data and analytics business, we constantly look for ways to improve our methods and we will be doing a thorough review of any areas where we can take learnings from for next time to make our MRP even more accurate and how we can apply our success into other areas of YouGov’s work. Read more about YouGov’s projections compared to the election results and learn more about what made our MRP model work so well: https://lnkd.in/eDrgK5tv
How YouGov's seat and vote projections fared at the 2024 UK general election | YouGov
yougov.co.uk
To view or add a comment, sign in
-
Election Notes 6 – What Can Reform UK Achieve In This Election? (part 2) The polling data also shows is that Reform are the number one destination for 2019 Tory voters switching away this time. However, surveys have shown that only a minority of those switchers would go back to the Tories if Reform were not available to vote for. This is not a party whose vote looks likely to be squeezed. In my previous article, I mentioned the role of the Canadian Reform Party in the electoral earthquake that saw the governing Progressive Conservatives crash to just two seats in 1993. What was notable was that Reform’s support was geographically focused on western Canada, enabling them to win 52 seats, while the more widespread vote of the Progressive Conservatives led to a near wipe-out. This example has led various commentators to speculate that Mr Farage might attempt to emulate what happened in Canada, where Reform remained the larger of the two parties and ended up managing a reverse takeover of the older party, forming the present-day Canadian Conservative party and enjoying several years in office before the Justin Trudeau years. From Andrew Neill to a Guardian article, this notion has had a lot of coverage this last week. However, the analogy falls down precisely because the Canadian situation represents the reverse of the British scenario. Here, it is the Conservatives have the more geographically-rooted support and, even in the most apocalyptic worse-case scenarios, will win dozens of seats, Reform is still projected to win none, although this calculation may be upset by individual circumstances such as those of Mr Farage standing in Clapton and ex-Tory defector Lee Anderson, defending his Ashfield seat he won under a different party label. Such an outcome could leave Reform with voices in Parliament, but fewer than the Tories even if they get a similar number of votes. This quirk of the first-past-the-post system similarly impacted UKIP, where Douglas Carswell, like Anderson a Tory defector and like Farage standing in Clapton, won their sole victory in 2015. This situation might also be compared to the 1983 general election, when the Liberal-SDP Alliance vowed to replace Labour as the main left-of centre party, but won just 23 seats for 25.4 per cent of the vote, while Labour secured 209 with 27.6 per cent. In the end, the Alliance morphed into the Liberal Democrats and Labour recovered to eventually become a party of government again. Mr Farage may have shrugged off being attacked with a banana milkshake this week, but he may find the electoral system provides a banana skin for his party’s lofty ambitions.
To view or add a comment, sign in
-
So, the election results were...interesting. But I think that the next general election has the potential to be even more disruptive than this one. I'm not going to be providing a completely original analysis, but there might be a few nuances in there that I'd be keen to discuss with others. Firstly, Labour winning the election wasn't much of a surprise to most, but like most commentators have said, did they win because people wanted Labour or because people wanted the Tories out? Hard to say, especially with Labour winning seats in Scotland off of the SNP. To me, some of the most surprising results were in the things that have been barely reported on. Specifically, the rise in Independent MPs who aren't affiliated with a political party - like Jeremy Corbin for Islington North (whatever you think about him). We didn't have any Independent MPs after the 2019 election, and now we have 6, and they consisted of 2% of the vote. Those aren't big numbers, but could it mean the beginning of a shift away from political parties? I think that more independent MPs in parliament that don't have a horse in the RED vs BLUE race should mean that they'll only be on the side of their constituents. Which could be positive change. We can also recognise changes in voting patterns by the votes for Lib Dem, Green and Reform. I could see all of these parties gaining even more votes in the next election if Labour doesn't manage to make positive economic changes to centre their next campaign on. Now, I have to talk briefly about the First Past The Post vs the proportional representation system argument that's been ignited post-election. From what I've come across, more people want to switch to the PR system. I've also seen (and spoken to) people who have been historically in favour of it but are suddenly on the fence because that would mean that Reform would have a larger say in parliament this time around - with 14% of the vote. That's not how democracy works, and it is a failing of the system that they got 4,117,221 votes, and those voters aren't being represented. No matter what you think of Reform (I don't personally like them, but that's beside the point). Democracy can be a bitter pill to swallow, but ignoring these voters will probably only increase their presence in the next election. Just look at Trump and the disenfranchised South to see what will happen if you ignore a voting base. Therefore, in the next election, we could see even more splitting of the vote between the third parties (Lib Dem, Green etc.), a likely rise in Reform votes and hopefully a rise in Independent votes as well. But this will all hinge on whether Labour can make the changes they suggest and if the Conservatives can re-brand in 5 years. Either way, I think the next election will be entertaining, at least. And will probably be something that is taught in Politics Degrees in the future as a decisive shift in early 21st century UK politics. What do you think?
To view or add a comment, sign in
-
Election Notes 8 – Can The Liberal Democrats crack first-past-the post? (Part 1) One of the most salient moments of the first seven-party TV debate was when Liberal Democrat Daisy Miller was reminded of the party’s climbdown on tuition fees during its days in the coalition. The pain was all the greater when the party crashed from 57 seats to eight in the 2015 election, with their share of the vote down from 23 per cent to 8.1 per cent. Despite trying hard to rally the remainer vote in the Brexit-dominated 2017 and 2019 elections, seat tallies of 12 and 11 respectively represented little of a recovery, especially as the share of vote barely crept into double figures and the party continued to be fourth largest behind the SNP. However, while polls still show the party hovering around ten per cent in voter intentions, the projections suggest the party could win upwards of 50 seats, maybe as high as the 62 seats won in 2005. Some polls have even suggested they might overtake the Tories to become the second largest party, making Ed Davey leader of the opposition - as long as he doesn’t drown himself in one of his watery campaign stunts. Normally, the party is held back by one longstanding problem - the electoral system. In it’s manifesto, there is the usual promise to introduce proportional representation and it is easy to see why they want this; the Liberal Democrats and, before them, the Alliance and then the old Liberal Party have for decades had substantial vote shares without this translating into many seats. The outstanding instance of this came in the 1983 election, when the Liberals had formed an alliance with the SDP after their breakaway from the Labour Party. The Alliance gained 25.4 per cent of the vote, not far behind the 27.6 per cent recorded by a floundering Labour Party, Yet Labour still won 209 seats, the Alliance just 23. To understand why, one only has to look at the contrasting distribution of votes. Labour came first or second in 341 seats, the Alliance in 336. The key difference was that Labour still focused enough support on its core areas of support like inner cities and coalfields. The Alliance vote, by contrast, was far too evenly spread, with 546 of its 633 candidates polling within ten per cent of their national figure. By contrast, Labour wasted few votes on hopeless causes, losing 119 deposits, compared with the Alliance’s 11. Finally, there was a problem that has bedevilled Liberals and their successors at almost every election since the 1920s: Most of their target seats and best performances were in areas where they are fighting with the Conservatives, who were on their way to a landslide with just under 43 per cent of the vote. The situation came close to replicating itself in 1987, when the Alliance gained 22 per cent of the vote and just 22 seats, while the first election after the Liberals and SDP merged saw the Liberal Democrats pick up 20 seats for 18 per cent of the vote. Cont in part 2..
To view or add a comment, sign in
-
Twelve months out from the next election, a minority federal government is a live possibility According to analysis by Accent Research and the RedBridge Group, a hung parliament and a Labor majority are almost equally likely outcomes. There is essentially zero probability at this stage the Liberal-National Coalition will win more seats than Labor. Topline results indicate little movement in overall vote share since the last election. However, this hides localised movements. There are always significant regional patterns in voter behaviour. Labor is losing primary votes, particularly in the outer suburbs and regional centres, while the Coalition has gained small primary vote swings everywhere, except rural electorates. Australian elections are not necessarily won by the party, or parties, that win a majority of the popular vote. Rather, it is who can command a majority of support in the House of Representatives that forms government. Current vote patterns do not appear to be resulting in major seat gains for the Coalition, nor major losses for the Labor Party. While the Coalition is within striking distance of some outer suburban and regional seats held by Labor, such as Robertson, Gilmore and Lyons, they do not appear to be winning back urban seats they lost at the last election. Additionally, Labor is competitive in some Liberal-held seats, such as Menzies and Deakin in suburban Melbourne. This electoral geography makes it very difficult for the Coalition to regain government, or even look competitive. Using current electorate boundaries, a hung parliament or a Labor majority are almost equally likely outcomes, according to this model. If a minority government is the outcome, though, Labor is still more likely to still be the largest party in parliament. Of course our estimates won’t be 100 per cent accurate. These results are estimates of public opinion at a point in time. These were derived from a model-based approach called MRP, using a survey of 4,040 Australian voters conducted between February and May 2024. Electorate-level results have average 95 per cent confidence intervals of 6.7 per cent for the Coalition vote share, 4.6 per cent for Labor, 3.7 per cent for the Greens and 6 per cent for other parties and candidates. This shares information across seats, with voters assumed to behave in a related way to other voters with shared characteristics in similar divisions. Voter behaviour will also shift between now and the election, and these estimates are based on current electoral boundaries. A redistribution is underway and those boundaries will change in NSW, Vic and WA. We will also update our estimates once this is finalised. The full results and details on the methodology are available here: https://lnkd.in/g2iX4EST
To view or add a comment, sign in
-
-
Election Notes 4 (part 2) – Why The Polls Are Sometimes Wrong (and often right) What had been overlooked was the answers given to a secondary question beyond voting intention, which was how likely people were to actually go and vote. This indicated that fewer of those inclined to support Labour put themselves in the ‘very likely’ or ‘certain’ categories when it came to turning out compared to those leaning Conservative. This made a critical small difference. This, of course, was not an issue for the exit poll, because by definition all the people being polled in it have turned out. With this factor in mind, polling companies sought to avoid the same happening in 2017. One way to do that would be to apply some weighting based on assumed intent to turn out. However, 2017 brought an anomaly. After years of polls apparently skewing projected support towards Labour, the hung parliament defied the majority of polling organisations (Ipsos Mori, NOP and so on), which had pointed to an increased Tory majority. By contrast, Survation (which had not conducted a national poll in 2015) accurately predicted a much smaller lead, while YouGov used a new methodology, interviewing more people than normal (50,000) and asking a set of questions about different values and attitudes to create a ‘big data’ model to predict how people would vote. This caused a major stir three weeks out from the election when it (accurately) predicted a hung parliament. So, why were some polls right and others wrong? The answer may lie in the fact that those organisations that projected a larger Tory lead made the assumption that voters would behave as they had done in 2015. However, because Brexit was the prime issue in 2017 and this cut across party lines, voters switched between the parties in larger numbers than ever before. This was reflected in the fact that swings went in different directions, with the overall swing to Labour stronger in more pro-remain seats, while strong ‘leave’ seats went the other way. Among the low-lead pollsters, Survation had not tried any methodological tweaks and simply published its raw data. At the opposite end of the scale, the sophisticated YouGov method proved justified. True, YouGov said it was subject to a high margin of error, but even that was reflected in the fact that the outcome was on a knife edge; no fewer than 30 seats were won by majorities of under one per cent, compared with just 12 in 2015. It is easy to be critical of polling organisations, but their successes are often overlooked. Even in 2015, for example, they accurately forecast the SNP landslide in Scotland. Therefore, unless things change drastically, the polls foretell calamity for Rishi Sunak. What he cannot do is count on the pollsters getting it badly wrong.
To view or add a comment, sign in
-
Managing Director - Commercial Institutional Investment Sales, Finance and Research at Brookwood-Starboard Commercial
The Fed’s Decisions Now Could Alter the 2024 Elections The state of the economy will affect voting next November, and the Federal Reserve may find itself in a delicate position, our columnist says. What’s happening in the economy now will have a big effect — perhaps, a decisive one — on the presidential election and control of Congress in 2024. To a remarkable extent, the economy is what matters to voters, so much so that one long-running election model relies on economic data to produce accurate predictions without even considering the identities, personalities, popularity or policies of candidates, or the strategies, messaging or dirty tricks of their campaigns. Right now, that model, created and run by Ray Fair, a Yale economist, shows that the 2024 national elections are very much up for grabs. The economy is strong enough for the incumbent Democrats to win the popular vote for the presidency and Congress next year, Professor Fair’s projections find. But it’s not a slam dunk. Persistent — though declining — inflation also gives the Republicans a reasonable chance of victory, the model shows. Both outcomes are within the model’s margin for error. https://lnkd.in/gK7nUCzW
The Fed’s Decisions Now Could Alter the 2024 Elections
https://www.nytimes.com
To view or add a comment, sign in
-
Election time, finally. In the run up to the 4th of July general election, let’s remind ourselves of several important things. 1. Whether you vote Labour or Conservative, Reform or Green, your vote is essential to the process of democracy. If you don’t exercise your political right to vote, you aren’t being represented. 2. Your vote does matter, regardless of what you’re told. The term ‘safe seat’ gets thrown around at election time, but this year is different - constituencies such as Windsor are polled to vote Labour after, in this example’s case, more than a century in Conservative hands. Even if your seat does vote en masse for a specific candidate, this is a healthy part of democracy - the larger the majority a candidate wins, the more valid their claim to be a representative of their electorate. 3. Legitimacy is conferred on a given government through winning the popular vote, and the more people who vote in an election, the more democratic the result. In 2005, the Labour Party won 55% of seats in the House of Commons with 35.5% of the popular vote. On top of this, less than two thirds of registered voters cast their ballots that year, meaning that the Labour Party’s share of the total potential vote was less than 25%. If voters don’t engage, we run the risk of a result not dissimilar to this as a result of apathy and the predicted results of the election, which leads me to my next point. 3. Disregard opinion polling - there is no guarantee of accuracy, as indicated in the past. 1992 saw predictions of either a hung Parliament or a small Labour majority in the general election of that year; in actuality, the Conservative Party retained power. Allowing predictions of a large Labour majority this year to discourage you from voting not only damages the quality of the national result, but also upsets the balance in the House of Commons - no political party should hold a majority larger than New Labour did in 1997, as it paves the way for legislation to cruise through Parliament without proper scrutiny. That’s one of the ways in which terror laws which violated the Human Rights Act came about in the early 2000s. 4. Most important of all, a general election is better considered 650 separate elections, each taking place in a Parliamentary constituency. You’re not voting for the government - you’re voting for your local MP. Find out who your local candidate is; it’s them who you need to agree with and buy into the policies of. While it’s important to agree with the party you’re voting for, too, voting based on a personal inclination towards or against its leaders would be to do yourself a disservice.
To view or add a comment, sign in
-
Helping everyday people and their families take control of their financial futures by providing personal, face-to-face financial advice. Financial Planning | Financial Education | Local Advice
How many of you stayed up to watch the election results last night? I tried but found myself falling asleep by 1am! What I found interesting over the course of the night was to see how the vote share has landed versus the seats won. Labour with 33.8% of the vote share and a huge amount of seats at 410 whereas the Conservatives have 23.7% of the vote share but only a small amount of seats at 119. Interestingly enough Reform UK had the third largest vote share with 14.3% but only 4 seats versus the Liberal Democrats with 12.2% of the vote share with 71 seats. All of this has resulted in the worst election defeat in history for the Conservatives and one of the best election results for the Liberal Democrats. What does all of this tell us? I think it’s quite clear that the public have voted for change, and it could be said that many voters voted in protest yesterday because they demanded change. It has been commonplace in the last couple of elections that Nigel Farage’s party has gained popularity with voters but has struggled to win seats and so it is speculated that a lot of Conservative voters have voted Reform in this election. What does this mean for investments? At time of writing the FTSE 100 is up by 0.36% with the FTSE 250 up by 1.6%. The FTSE 250 is the more important index here as it is widely considered a better metric of the UK economy. The markets were fully expecting a Labour government and I don’t think it comes as a surprise to anyone. In terms of investment portfolios, the UK election results won’t make a huge difference right now but where the UK election results will make a difference is what Labour choose to do from here, I wouldn’t expect any form of ‘budget’ before October/November time and this will likely be our first indication as to any economic change. For further insights from yesterday's election results, check out our latest article on our website detailing the impact on investment now Labour has won the general election (link in comments) ⬇️ #GeneralElection #Markets
To view or add a comment, sign in
-
-
Election Notes 5 – Electoral wipeout scenarios (Part 2) The 1931 trouncing of Labour was, therefore, not the result of the party seeing a drastic drop in votes; they had a higher share than they won in 1923, which brought 191 seats and the first Labour government. Nor had the parliamentary party split in half like the Liberals had before the similar circumstances of the 1918 general election; MacDonald’s faction was a splinter group that, unlike Lloyd George’s Liberals in 1918, won far fewer seats than their erstwhile colleagues. The absence of a formal split is the key way Labour’s 1931 experience differs from that of any equivalent result for the Tories now. The similarity comes in the electorate’s willingness to punish a party it feels has failed to handle times of crisis. If a defeat of such magnitude should hit the Conservatives, they may take heart from the fact that two elections later in 1945, Labour were back in power with a landslide victory, as this time a new seismic global event played in their favour. In between, the 1935 election was referenced by commentators in 2019 as being the last time Labour had done worse than their 203 seat tally under Jeremy Corbyn, but any comparison is inapt, as at the time the 154 seats won represented a substantial recovery. The 1993 Canadian situation is the ultimate fantasy on the left. In that election, the oxymoronically-named Progressive Conservatives crashed from 156 seats (out of 308 in Canada’s House of Commons) to a mere two with just 16% of the vote. A key reason was that they, like the British Conservatives today, faced a new right-wing populist challenger called Reform. The difference, however, was that Reform’s support (nearly 19%) was heavily concentrated in western Canada, where they won 51 of their eventual 52 seats. This would be the equivalent of Reform UK winning 110 seats here. After further heavy electoral defeats for the divided Canadian right, they united to form the current Conservative Party of Canada, returning to office in the 2000s. What these scenarios tell us is that if the worst poll readings promise something that comes close to rewriting the record books, they still do not represent extinction events. Even if this election will bring the worst Tory defeat ever, it does not spell the end, despite the fact Britain’s Reform Party, now with the added spice of Nigel Farage taking charge, has ambitions to replace them.
To view or add a comment, sign in
Legal Counsel | Digital Law Professor @UANL | Legal Ops | Legal Tech | Legal Project Management | Pro Bono
3w👏🏻 👏🏻 👏🏻