The increased unpredictability of UK politics.
Note this article was first published in October 2019, but has been later updated to include information regarding the 2019 election.
The last decade witnessed many shock results. The 2015 election produced a clear majority for the Conservative Party, defying many polling estimates. On top of this, the UK left the European Union, despite many commentators at the time confidently predicting a slightly win for Remain. The 2017 election produced rapid changing in polling and the outcome produced a hung parliament, further making political outcomes harder to predict. Finally, the 2019 general election gave the Tories a much higher than anticipated majority, with many former safe Labour seats surprisingly electing a Tory MP, often for the first time in their history. One reason behind this increased unpredictability is due to the rise in voter-volatility. This means that a higher proportion of the electorate develops the willingness to change parties at a more rapid pace than previously had been the case. This blog post outlines how trends in UK voter volatility increased across the last decade and by the time of the critical 2017 and 2019 general elections can be argued to be at very high, if not historic, levels. The article argues this development is the main reason why the UK political system has become increasingly unpredictable in recent years.
When using the oldest volatility measure, the Butler swing, it can be said that over the last decade elections have been increasingly volatile, and therefore less predictable. Moreover, the level of net volatility under this measure puts the 2019 election as the fourth most volatile election in UK post-war history. The level of volatility is also similar to that of the period where the 1974-79 parliament, another time that produced great change and unpredictable electoral behaviour. Therefore, the voters were creating historically high levels of change from one election to another.
The Pedersen Index takes into account gains and losses that occur under smaller parties. When using this method to calculate volatility between each electoral cycle in post-war UK history it can be said the past decade has been very unpredictable. The electoral cycles between 2010-15 and 2015-17 were historically the most volatile elections in post-war times. This importantly captures the increased levels of volatility created by the rise of smaller parties, such as UKIP. Additionally, this measure shows that amount of change from on election to another during the 2010s was higher than any other time-period. This all again shows the 2010s can indeed be said to be an unpredictable one and a decade of seismic change.
Gross volatility tracks the amount of voters who change support from one election to the next. Using BES panel data to track this change the last decade of British politics can be said to be a highly volatile period in British political history. Comparing stable votes, ballots where the same voter supported the same party two elections in a row, it can be said the parties’ voting base was as its most unstable level since the 1970s. The last decade was the first decade where vote retention was below the 60% in all elections. Moreover, vote retention for the two main parties was at its lowest level since the latter-half of the 1970s, the last time a realignment occurred.
Furthermore, voters switching between the two main parties has been consistently at its highest level since BES records began. This indicates great potential for the two main parties’ voter base to have changed. With four elections within the last decade, and each one producing large levels of voter switching between the main parties, there is a possibility that higher net levels of net volatility have led to a gradual shift in voting bases. Significantly, this may have been why elections became increasingly difficult to predict during the 2010s.
Table 2 - By-election volatility (in terms of losses and gains from the previous general election) from 1945 - 2020.
By-election results can be another indication of how volatile voters are feeling during the points between both local and national elections. Previous studies on volatility revealed that third and smaller parties performing well in by-elections can indicate desire for political change. This is because parties that lose control of seats in by-elections, especially those thought to be previously safe seats, can indicate a change occurring within British politics before a general election occurs.
Focusing on by-election results for incumbent parties having to unexpectedly defend their seat, it can be said that since 2005 by-elections have become increasingly difficult for the two main parties, regardless who is in government. Historically patterns have shown that the main opposition party has overall been the beneficiary of by-elections. However, since the parliamentary cycle of 2005-10 both parties have been losing their share of the vote in by-elections where they are defending. This means that volatility has increased in a way that appears to have caused increased instability for both the main parties regardless of who is in power. This increased level of volatility therefore appears to benefit smaller parties, especially those offering political change, rather than parties seen to be the establishment parties of Westminster. Therefore, as more parties share of the vote has rapidly changed from general election to by-elections, meaning even unpredictable by-elections can said to have become harder to forecast.
Volatility between elections: Opinion polls
Figure 4 below displays the change in average ratings for all parties in the opinion polls from 1945 to just before the election in 2019. It compares the average difference in support a party registered in one given quarter to the previous one. The average change is the mean of all the parties change in support, consequently meaning that this graph takes into account the rise of new parties. The closer the figure is to zero the less volatility has occurred from one quarter to the next, with higher numbers representing greater volatility. The graph shows that the decade experienced quite significant volatility, more so than most periods British politics has produced. Moreover, there are more opinion poll organisations conducting regular polls than in the past, resulting in greater confidence that this volatility is not due to sampling errors. On average parties change up to four percentage points in very short timespans towards the end of the last decade, resulting in leads being created, or cut, in shorter periods than British politics has historically allowed. Moreover, as the UK approached the 2019 election the variation in monthly and quarterly opinions polls increased to historic levels. Moreover, there are large fluctuations in volatility approaching, and during, general elections. This indicates that volatility was not just at historic high, but also was occurring during critical points in UK political history, making it hard to predict when decisive events would occur. Therefore, increased voter unpredictability led to elections becoming harder to estimate as the decade progressed.
Two main party volatility:
Focusing on the two main parties’ shares of the opinion polls, records have been quite volatile across the last decade. The polls show that there have been several exchanges of the lead, more so than other decades. There has also been many sharp increases and decreases of the two main parties’ shares of the vote, indicating a destabilisation of the vote for both parties. Crucially, towards the end of the last decade both the Labour and the Conservative levels of support appear to decline together at the same rate. Historically in British politics the patterns of their share of the vote have been mirror images, as one increases the other decreases at roughly the same rate. This is because these parties have typically fought over a small number of voters in the centre-ground and an exchange of these voters would result in a direct transfer of votes, with one party gaining and the other losing out. However, this pattern stopped as the 2019 election approached, indicating that these parties were now losing votes to smaller parties and were taking from each other’s traditionally solid support bases. This indicates that not only has volatility increased, but a particular type of volatility may have emerged, one which makes a predicting main party support harder. Moreover, 3rd and 4th party support also became harder to predict.
Third and Fourth Party Volatility:
Volatility from smaller parties, especially those that can be described as niche parties, has been quite high during the last decade of British politics. It has seen the rise of the SNP in Scotland and also the rise and fall of UKIP in England and Wales. Alongside this, the 2019 EU election witnessed the emergence of a new party, the Brexit Party, increasing volatility from smaller parties to its highest level in UK polling history. This level of volatility is quite extreme, with rapid gains and losses for these parties. This could indicate that these parties are being used as vehicles of protest with an aim to force the main parties to adopt certain positions. It could also indicate that these parties might have been used as a mechanism for voters to switch between the two main parties gradually, as occurred with the Liberal Party in the 1970s. Therefore, quick changes in fortunes for these smaller parties could indicate that rapid voter changing has been part of a process where voters realign their allegiances. For example, the dramatic fall in the share of the vote the Liberal Democrats once commanded during the first year of the coalition may have resulted in a transfer of votes of former Conservative voters to Labour. Meanwhile, the rise of UKIP and the Brexit party may have resulted in the gradual transfer of former Labour votes to towards the Conservative Party. If such changing of support did occur with rapid succession this could also help explain why British politics seemingly has become harder to predict.
Polls fluctuating in UK elections:
The above image displays how polls have also started to fluctuate to a much greater extent during election campaigns. Going into more detail of polling fluctuations during election campaigns, it can be said the polls show the UK to be at its most volatile level in electoral post-war history. Taking the average of the polls for the Labour, Conservative, Liberal and pro-Brexit parties at the start of a campaign, and then comparing this to the average poll during the last week of the campaign, the data shows more voters switching during an election. Historically election campaigns have done little to change the poll ratings of the parties, indicating that election campaigns might not have changed party fortunes in UK post-war history. However, since 2010 every election has seen higher levels of volatility during an election than existed pre-2010. The 2010, 2017 and 2019 elections showed the parties on average moving four points away from their starting position. Moreover, both in the 2015 and 2017 elections gains were underestimated, meaning the true amount of volatility was probably even higher than the graph suggests. This all indicates that voters now are more likely to consider changing their vote during an election campaign, and therefore such behaviour is making UK elections harder to project.
The last decade in British politics has become much more unpredictable due to the rise of voter unpredictability, also known as voter-volatility. It would appear old allegiances seem to be less meaningful, with voters having gained an increased appetite to switch parties, especially during general election campaigns. This increased unpredictability has coincided with the rise of smaller parties, and in England and Wales particularly the rise of UKIP. This trend accelerated post-Brexit and brought great uncertainty to UK elections. With the increased propensity for voters to switch their loyalties this could mean further uncertain times lay ahead for the UK.