Shopping for a better deal? Party switching among grassroots members in Britain

ABSTRACT People who join political parties are motivated primarily, although not exclusively, by ideological impulses. So, given the often considerable ideological differences between parties, one might presume that very few those who later leave one party would be keen to join another. However, using a comprehensive 2019 survey of British party members, we not only identify several factors associated with switching parties (being especially socially liberal or socially authoritarian; being a Brexiteer; being a campaign activist; working in non-manual occupation; having a university degree; being a man rather than a woman; being older rather than younger; and being a current member of one of the country’s smaller political parties), but we also show, first, the flows between parties and, second, that party switching at the grassroots is far more common than many imagine. Even so, and notwithstanding the fact that switching impacts on different parties in different ways, it is not sufficient to support oft-voiced claims of widespread entryism into either of Britain’s two main parties – at least on the sort of scale that might account for Labour’s shift to the liberal-left or the Conservatives’ shift to supporting a hard Brexit.


Introduction
After decades of survey research, we have a fairly good idea of what it is that leads people to join political parties in the first place (see van Haute and Gauja 2015, 6-9, 193-194). And, while much more workand certainly much more work than can be done in a short research note like this oneis needed on the topic, we are beginning to get some idea of why they decide to leave them (Bale, Webb, and Poletti 2019, 146-164). Indeed, the reasons for both enrolling and quitting are similar in the sense that, although it may be not the be-all and end-all, ideology matters: most of us join a particular party primarily because we agree with what it wants to do (and disagree with what its rivals want to do); most of us who then leave do so mainly because we lose faith in the direction in which it seems to be headingsomething symbolized perhaps by particular policies or by, say, a change of leader (see Barnfield and Bale 2020). At the elite level, there is some research on "crossing the floor", "party switching" or "party hopping" by legislators (see for example, Heller and Mershon 2009;O'Brien and Shomer 2013;Morris 2018;Volpi 2019;and Keaveney 2019)research that shows that, although inter-party defection is actually quite rare, it is often ideologically motivated. But what we don't know much (if anything) about is the extent to which ordinary grassroots members who abandon one party go on to become a member of another. Moreover, why anyone would choose to do so could be seen as something of a mystery: after all, if someone is enough of an ideologue to join a party, wouldn't they, by definition, be relatively immune to the charms of its competitorsespecially since (unless perhaps they are a local councillor) they can hardly be tempted, like some of their parliamentary representatives, to jump ship purely to try and save their political career?
Yet according to many of those involved in British politics in recent years, be they active participants or observers, switching parties at the grassroots is by no means that uncommon. The massive increase in Labour Party membership which accompanied Jeremy Corbyn's elevation to the leadership in 2015 was often anecdotally associated, at least in the minds of his enemies (internal as well as external) with "entryism" on the part of people who had previously belonged to parties on the far-left fringe of the country's politics (BBC 2016; but see also Crick 2015 andPine 2016) or (more charitably, perhaps, and possibly less anecdotally) to the Greens. 1 Meanwhile, the Conservative Party's adoption of an ever harder position on Brexita position that eventually led to the replacement of Theresa May by Boris Johnson and the defection or ejection of "pro-European" Tory MPswas blamed by some (not least by some of those MPs) on entryism on the part of former members of UKIP (Murphy and Jacobson 2019; Thompson and Sylvester 2019; see also Loucaides 2019 and, for a dollop of scepticism, Wallace 2019).
In view of this, we seek here to make use of a survey of party members conducted immediately after the UK's 2019 general election to investigate the following questions: How many current party members have previously been members of other parties? Between which parties have flows of members occurred, and when did these flows of membership take place?
1 In a write-in section of a survey of party leavers in Britain conducted in 2017 by the Party Members Project, some 41% of 152 former Green Party members said they had left to join Jeremy Corbyn's Labour Party. At best this is only indicative but it is interesting to note that very, very few of the former members of the Labour Party, the Conservative Party or UKIP who responded to the write-in question asking them for their reason for leaving explicitly mentioned wanting to join another party. See Bale, Webb, and Poletti (2019, 158-161).
What factors seem to motivate such movements? And does the data lend credence to the various claims of "entryism" that were made in 2019?

Data
In order to answer this question, we surveyed 6823 members of six British parties (including registered Brexit Party supporters) within two weeks of the 2019 general election. Some 1478 respondents declared that they had previously belonged to a party other than the one to which they currently belonged. The (online) survey was conducted for us by YouGov and funded by the ESRC. Survey respondents were recruited from a panel of around several hundred thousand volunteers who are offered a small reward for completing a survey. Upon joining the panel volunteers are asked a broad range of demographic questions which are subsequently used to recruit respondents matching desired demographic quotas for surveys. Potential respondents for the party members survey were identified from questions asking respondents if they were members of any of a list of large membership organizations, including the political parties. Results reported in this article are not weighted in any way since there are no known official population parameters for the various party memberships. However, previous YouGov party membership surveys using unweighted data have generated predictions for party leadership contests that came very close to (that is, within just 1% of) the final official outcome, which provides considerable confidence in the quality of the data. Table 1 addresses the first of these questions by revealing that a remarkably high proportion of our sample claim to have previously been, or currently are, members of a different political party than the main one to which they are presently affiliated. Overall, nearly a quarter (23%) fall into this category. As a robustness check, we examined patterns of members' movements between parties using a small British Election Study panel sample from 2015 to 2019, and found that 9% of these respondents moved parties during this period (see Appendix, Table A.1). Given that this was a tenth of people switched party memberships in just four (admittedly turbulent) years, it is not surprising that double that proportion would have switched over the course of their entire political life-times, as our data suggests. When one weights the amount of switching by the relative sizes of these parties in 2019, it suggests that as many as 19% of party members at that time would have been "switchers"that is, people who had changed party membership at some point in their lives (see Appendix, Table A.2). In longterm perspective, this might be a relatively high figure given the extraordinary political tumult of the years surrounding Brexit. Not surprisingly, perhaps, the Brexit Party's registered supporters (the party had no "members" as such) and UKIP members, when considered as a single entity, were particularly likely to have been members of other parties, with more than a third claiming that to be the case. In fact, all UKIP members in our sample also registered as Brexit Party supporters for the 2019 election, although just 3.7% of BP supporters were also UKIP members. Thus, by late 2019 the overwhelming majority of BP supporters were no longer UKIP members. Indeed, there are very few UKIP members in our sample at alltoo few to be able to conduct reliable analysis; henceforward, therefore, in so far as this group of respondents is concerned, we shall focus our analysis on current BP supporters rather than UKIP members. The Greens also report a high proportion of other-party memberships (30%); as we shall soon see, Labour is the most prominent "other party" in question. By contrast, the two major parties (both of which do their very best to discourage any of their members belonging to other parties) have the lowest proportions of other-party memberships amongst their current members. Table 2 provides detail about the flows of members between all parties. As suspected, Brexit supporters who had other party memberships chiefly feature UKIP and Tory members amongst their ranks, some 41% falling into the former category and 32% into the latter. As a proportion of the total number of Brexit Party supporters in our sample (n = 1112), 24% had been affiliated with UKIP and/or the Conservatives, the remainder being new to party political involvement. Just as interesting is the reverse flow of members from BP and UKIP into the Conservative Party, since this bears directly on some of the claims of entryism that were heard in the months leading up to the 2019 general election. Over half of all Tory members with previous party memberships joined after 2015, and a third quit their former parties after 2017 and quickly moved on to the Conservatives in 2018 and 2019. Table 2 reveals that 29% of all Tory members who admit to having been members of other parties claim to have been UKIP members; interestingly, though, even more of them (39%) were former Labour members. Although not shown in Table 2, a further 29% of current Conservative Party members who admitted another affiliation also claimed to have been registered Brexit Party supporters at the time of our survey.

Results and discussion
As a proportion of all Conservative Party grassroots members, these figures amount to 3.2% who were former members of UKIP, 4.5% who were current Brexit Party supporters, and 4.3% who were ex-Labour members. This puts into perspective the scale of the entryist phenomenon; on these figures, around 8% of all Tory members in 2019 had a history of connections with UKIP or the Brexit Party. This is not to say that their impact may not have been significant in certain constituencies when it came to selecting party candidates. Nor is it to deny that the Conservative Party grassroots have increasingly come to favour "hard" forms of Brexit over the course of the past few years (Bale, Webb, and Poletti 2019, 66-67). But it would appear that in the vast majority of cases this will largely have been down to the changing views of those members who have no formal associations with UKIP or the Brexit Party.
As in the case of the Conservatives, many Labour members (39%) with previous party memberships joined the party after 2015. Much of this must be attributed to the effect of Jeremy Corbyn becoming leader of the party in 2015; indeed, it may also have contributed to it (Whiteley et al. 2019). Those Labour members who have had past lives in other organizations come mainly from the Green Party or Liberal Democratsor, intriguingly, from the amorphous "other parties" category. Our survey did not probe further into the precise provenance of this latter group, but it is possible that many of them have had previous connections with far-left organizations quite another type of "entryism" to that which the Conservatives have experienced. It was widely believed that the Green Party haemhorrhaged support to Labour after the election of Jeremy Corbyn as leader in 2015, so it is no surprise to find that nearly two-fifths of all Labour members with other party memberships came from the Greens. This amounts to 6% of the total Labour membership sample. A quarter of those with other party affiliations came from the Liberal Democrats, which is to say 4% of all Labour members, and 22% of Labour's influx from elsewhere (3% of the total membership) came from the miscelleaneous but probably leftist "other" organizations category.
The Liberal Democrats experienced a very significant influx of new members after the general election reversal of May 2015, and clearly, a substantial proportion of these individuals had histories of affiliation to other parties prior to that time. Indeed, around 70% of all Liberal Democrats who had previously been members of other parties joined after 2015, and twothirds of these did so almost immediately after quitting their former party. Virtually a half of all Liberal Democrats with previous party affiliations came from Labour, while a quarter came from the Tories. These figures amount to respectively 13% and 6% of the total Liberal Democrat membership.
The smaller parties have experienced high levels of cross-party flows,proportionately speaking. For the Greens, 34% of its members who had quit another party did so after 2017. Three-fifths of their recruits from other parties came from Labour and one-fifth from the Liberal Democrats, equal to 19% and 7% of the total party membership. The SNP is rather less dependent on recent quitters-and-quick-joiners, only 17% of all those with previous party affiliations (or 3% of the total membership) having left their former party after 2017 and then joined the Scottish Nationalists. A high proportion claim to have left their previous party and then joined the SNP prior to 2015 (61.6% of all those with previous party memberships, or 12% of the total current membership). This most likely reflects the extraordinary mobilizing impact of the 2014 independence referendum; it is striking that 43% of all those who had previously been members of other parties joined the SNP in the 2011-2015 period (ie, 8% of the total current membership). Table 2 reveals that some 57% of all SNP members with previous party affiliations came from the Labour Partyamounting to 11% of its total membership. This underlines the extraordinary realignment of Scottish politics over the course of the past decade, with Labour being replaced by the SNP as the dominant party north of the border.
So, what drives some people to quit one party and join anotheroften very quickly? Although space constraints prevent us from offering a comprehensive theory, previous research strongly points to a combination of two separate theoriesone covering why people leave parties and another why they join parties. People quit largely because of ideological differences with their parties and negative evaluations of their party leaders (Bale, Webb, and Poletti 2019, Ch. 8) and they join parties because of a whole range of factors included in the "general incentives model" of party membership, but especially ideological and expressive ones (Bale, Webb, and Poletti 2019, 76-84). The nature of our data makes it impossible to incorporate these factors into a single model because we lack contemporaneous measures of respondents' evaluations of their former party leaders and ideological proximities at the time they quit the first party. However, we can run a more limited model with our data, with the intention of identifying factors that might be relevant across time and party, per the logistic regression model reported in Table 3.
The dependent variable in the model is binary, distinguishing between respondents who have never been members of another party (coded 1), and those who have previously been or currently are members of another party (coded 2). What factors incline individuals to the former rather than the latter category? We test a number of possibilities here.
The first thing we test for is the degree of ideological radicalism. The supposition here is that the sort of people most likely to feel strongly enough to ditch membership of one party for another would have to be motivated by particularly intense political attitudes. Consequently, we enter two variables which distinguish ideological radicals from moderates, one relating to leftright ideology and the other to libertarian-authoritarian values. Empirically, "radicals" on each of these dimensions are defined as anyone lying more than one standard deviation from the mean; the overall sample mean on the Left-Right scale lies at 2.89 (where 0 = far left and 10 = far right), while the standard deviation is 2.48. Thus, radicals on this dimension are deemed to be those scoring less than 0.41 or more than 5.37, and they constitute exactly 33.5% of the sample. The sample mean on the Liberty-Authority scale is 4.78 (0 = very liberal, 10 = very authoritarian), and the standard deviation is 2.65. Therefore, radicals on this dimension are defined as people scoring less than 2.13 or more than 7.43, and constitute 43% of the sample. Simple bivariate analysis suggests, contra expectations, that it is moderates rather than radicals on the Left-Right scale who are more likely to claim other party memberships (by 24.5% to 21.5%, p = .006). However, on the Liberal-Authoritarian scale the reverse is the case, as radicals are more likely to have or have had other party memberships (by 27% to 21%, p = .000). When we examine the logistic regression model, we find that the latter effect is significant and signed as expected, after controlling for all of the other terms in the model: those who are especially liberal or authoritarian are indeed more likely to be party-switchers. However, left-right radicalism has the opposite sign and is not significant (unless one accepts p < .10).
We also test the effects of two other major values divides might have had on party-switchingpost-materialism and Brexit. Both of these are important attitudinal dimensions of contemporary British politics. While attitude towards Brexit is known to be related to the liberty-authority dimension (Sobolewska and Ford 2020), it is not perfectly correlated with it. Post-materialism is, of course, a long-recognized attitudinal dimension which is also not simply reducible to either left-right or liberty-authority positions (Hooghe, Marks, and Wilson 2002;Webb and Bale 2021, Tables 6.1 and 6.2). It therefore seems important to account for their possible impact as possible ideological drivers in their own right. As it happens, however, post-materialism is not significant in the model, even though simple bivariate analysis does suggest a modest difference between ideological materialists (19.1% of whom are party-switchers) and post-materialists (24.8% of whom are switchers). Even though this difference is statistically significant in a crosstabulation (p = .012), the effect loses impact in the context of the multivariate model. Brexit, on the other hand, does prove to be a significant factor, which is hardly surprising given all that it has meant for British politics over the past few years. It is widely recognized that this has become an important dividing line in the politics of the UK since 2016, with the capacity to realign patterns of political support (Fieldhouse et al. 2019, ch.9;Sobolewska and Ford 2020). A simple crosstabulation shows that those who voted Leave in the 2016 referendum on EU membership were more likely than those who favoured Remain to have switched party membership at some point (by 27.1% to 21.9%, p = .000). This effect remains significant (p < .005) in the multivariate model. This may, of course, turn out to be a period effectsomething that could be tested in a replication study in a few years' time.
The model also takes into account the level of activism of party members. In previous work, we have found that those members most engaged in activity on behalf of their parties are generally less likely to quit (Bale, Webb, and Poletti 2019, 162). Here, however, we uncover a rather different, though not mutually incompatible, finding: even though more active members are less likely to leave a party in the first place, if they do so, it seems that they are then more likely to go on and join another party. We test this with a variable which registers the mean number of activities they undertook on behalf of their current party in the 2019 election campaign: this scale runs from 0 to 9 (mean = 1.75). In the model, the higher a respondent's score on this scale, the more likely they are to have or have had other party memberships (p = .001).
Finally, we proceed to test for the impact of a number of demographic and party affiliation variables, most of which prove to be significant in the model. The inclusion of these standard control variables is for exploratory rather than a priori theoretical reasons. Non-manual ABC1 respondents prove to be more likely to switch parties than manual C2DEs: in a simple bivariate cross-tabulation the respective shares who have switched are 24.3% and 20.3% (p = .001), and this remains significant (p = .009) in the multivariate model. The difference between graduates and non-graduates is also significant in the model (p = .004), with the latter being less likely to have had more than one party membership. Bivariate cross-tabulation suggests that men are more likely to be party-switchers than women (by 25.6% to 20.0%), and this effect is confirmed in the multivariate model; compared to women, men are significantly more likely to be switchers (p = .000). Age is also significant: the older a respondent is, the more likely he or she is to have had other party affiliations (p = .016); presumably, the longer one has been around, the more opportunity one has had to experience the vicissitudes of political life and to have reconsidered one's own position and allegiance.
Lastly, we have to take account of and control for party affiliations. These are mostly significant. The reference category is SNP membership, and the picture that emerges from this is that members of the two largest parties are significantly less likely than SNP members to have switched parties, while the current members of the smaller parties are all more likely to include switchers. The exact bivariate picture is evident in Table 1 above.

Conclusion
Our research suggests that there are several factors which make members, at least in contemporary Britain, more likely to switch parties: being especially socially liberal or socially authoritarian; being a Brexiteer; being a campaign activist; working in non-manual occupation; having a university degree; being a man rather than a woman; being older rather than younger; and being a current member of one of the country's smaller political parties. Our research also reveals that party switching at the grassroots occurs far more frequently among rank and file members than it does among MPsnot surprisingly, perhaps, given the much higher sunk costs and reputational and career risks faced by the latter. That said, it is not sufficiently widespread to support claims of entryism into either of Britain's two main partiesat least on the sort of scale that might account for their recent positional shifts.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This work was supported by the Economic and Social Research Council under grant number ES/T015632/1.

Supplementary material
Supplemental data for this article can be accessed online at https://doi.org/ 10. 1080/17457289.2021.1941062.