Party Identification,the Policy Space and Business Donations to Political Parties

Business, Preferences and Political Finance

Power and preferences have been the two main concerns of scholars of business and politics. For a long time, the literature emphasised structural power (Lindblom, 1977; Offe, 1985: 170–220). The dependence of democratic politics on the capitalist economy gave business a structural advantage. According to this theory, the threat, or even possibility, of disinvestment was seen as enough to ensure that politicians maintain business-friendly policies. Structural power made intentional power, principally lobbying and financial donations, both more effective and less necessary. However, there is lots of evidence that democratic politicians defy businesses. Ultimately, office-seeking politicians are elected by voters, not firms, and policy-seeking politicians often want to constrain, rather than facilitate, businesses. The theory of quiet politics asserts that ‘[B]usiness power goes down as political salience goes up’ (Culpepper, 2011: 177). When voters and the media are paying attention, politicians are sceptical of business arguments. By contrast, the public have little interest in most of the arcana of public policy that impact business. In these low-salience areas, politicians tend to defer to the expertise of business. Business has a particular ‘informational privilege’ (Bernhagen, 2007: 135) that allows it to convince politicians to adopt policies that fully informed policymakers would reject as detrimental to their voters’ interests.

Power has also been a concern of the more specific literature on business financing of politics. The nature and the extent of the power of business donations depend on their motivation. Business donors have pragmatic or ideological motivations (Harrigan, 2017). ‘The pragmatic motivation seeks private goods … pragmatic money is interested money’ (McMenamin 2013b: 8). It is rare for business to directly buy policy in established democracies. Many of the more convincing accounts of the relationship between donations and policy can also be thought of in terms of salience and information. Policymakers are time-poor, as well as information-poor. They cannot give a full hearing to all of the informationally privileged business actors clamouring to lobby them. Moreover, politicians want to minimise the electoral costs of being perceived to make decisions for the good of businesses instead of voters. Money can improve access to policymakers in order to leverage informational advantages in low-salience areas. In these situations, there is a reciprocal exchange between business donors and politicians. In reciprocal exchanges, each actor’s part of the exchange is separately performed and terms are unstated and uncertain (Molm, 2000: 261–262). These reciprocal exchanges protect politicians against accusations that they are selling decisions, as there is no direct connection between the donation and public policy (McMenamin, 2013a: 22, 2013b: 12). This is one reason that statistical studies find it so hard to demonstrate a relationship between political finance and policy (Stratmann, 2017: 12–18). The other is that many donations are ideologically motivated (Ansolabehere et al., 2003; McMenamin, 2012: 2). Ideological payments promote a public good. ‘They express a preference for government based on a particular set of values and assumptions’ (McMenamin, 2013b: 8). Ideological donations are not strictly rational for a profit-seeking firm because there is little chance that such donations will decide an election. Thus, ideological donations are like voting or joining a social movement.

The preferences of business can be studied individually or systemically. The Varieties-of-Capitalism school studies how national, political and economic institutions shape business preferences over public policy. These institutions were born out of negotiations around the time of the birth of mass democracy and labour unionism. In multi-party systems, right-wing party leaders had incentives to delegate decision-making to business associations and labour unions (Martin and Swank, 2012: 3). They believed that a corporatist framework was more beneficial to business than fighting a losing battle in a legislature where farmers’ and workers’ parties could outvote pro-business parties. By contrast, in two-party systems, the right concentrated on the prize of a majority in the legislature and had much weaker incentives to delegate policy to representative associations. These different origins explain much about the extent to which business acts collectively and its preferences over economic and social policy (Martin and Swank, 2012: 227).

The ‘Open Economy Approach’ has also focused on the interaction of economic interests and political institutions and is divided into the sector and factor approaches. A factor of production is an input in the production of goods and services, such as land, labour, capital and entrepreneurship. The factor model assumes that business preferences will be driven by the relative prices of capital and labour. The sector model argues that business preferences will be shared with employees and other businesses in the same sector. For example, the sectoral approach might see workers and owners in manufacturing pitted against workers and owners in services, whereas, in the factor approach, the owners of manufacturing and services firms would be opposed to the workers in both sectors. Of course, ultimately, these economic interests have to be processed by institutions within the firm (Martin, 1995). Majoritarian electoral systems elect politicians in small territorial units, thereby emphasising sectoral preferences, whereas proportional electoral systems elect politicians in large constituencies, thereby facilitating the unity of capitalists.

These theories concentrate on policy preferences, not party preferences. They assume that party preferences follow from policy positions. In the Varieties-of-Capitalism school, the importance of party political differences depends on the national institutional comparative advantage. In co-ordinated economies, business associations enjoy deep relations with policymakers regardless of the composition of government (Hall and Soskice, 2001). In liberal economies, government–business relations are more distant and also subject to greater change when there is a governmental turnover. Liberal economies are associated with a more intense business preference for pro-business parties. The factor approach should be associated with a clear identification with the political right, but the sector approach allows for more nuanced partisan preferences, depending on the nature of the sector and common cause with workers. Both assume that businesses make a rational choice on the basis of a very, or somewhat, stable political economy. Neither pays much attention to a radical change in the positions of political parties.

The literature on business and political finance has neglected the preferences of businesses. In relation to ideological contributions, it has taken preferences over policy and between parties for granted. For over a century, this has been a safe assumption for many nations because there has been only one relevant dimension of competition and the order of parties on that dimension has not changed. This dimension has been the classic left-right continuum of state intervention versus the free market. The ideological distance between parties changes as does the location of the centre of the policy space. Nonetheless, parties virtually never exchange positions in two-party systems and very rarely do so in multi-party systems. Therefore, there has been an obvious long-term centre-right party, which can receive business funding. This stability has meant that ideological distance, issue salience and party identification have not been as important in the relationship between business and parties as they have been in the electoral market.

Changes in the political preferences of business have ramifications for many of the questions studied by political finance scholars. Regulation affects party competition differently if one party can count on donations from business. If this is so, public funding can reduce the financial advantage of such parties (Katz Mair, 1995, 2009; Potter and Tavits, 2013; Van Biezen and Rashkova, 2014). Bans, limits and disclosure requirements can transform a financial advantage for pro-business parties into a financial disadvantage (McMenamin, 2015). If some parties receive substantial funding from business, it is hard to resist the conclusion that this will affect public policy when they are in government (Thacker and Pierson, 2010). If parties are unequally dependent on business funding, this will influence how they construct their interests in political finance reform (Koß, 2011; Scarrow, 2004). The way in which broader public policy affects funders will depend on whether those funders represent the business sector (McMenamin, 2012: 9–13; Tomashevskiy, 2015). If some parties are more dependent on business than others, this will affect public opinion on political finance, especially preferences on the regulation of business donations and the political construction of scandals (Fisher, 2015a; McMenamin, 2015). Since attitudes to, and relationships with, business are so strongly associated with the left-right dimension, they are integral to most party systems. When business donations are legal, this means party positions on business tend to be also important for political finance. Indeed, if enforcement is weak, the same can be said for countries where business donations are illegal.

The literature on business donations to political parties has paid little attention to the dynamism of parties. This is for the good reasons that only one dimension has been relevant and the order of parties along this dimension has been so stable. Therefore, it is reasonable to think that businesses, unlike voters, have not been faced with a choice. In the absence of changes in the policy space big enough to change the preferred party of business, it has not been possible to assess whether enduring funding relationships are due to party identification on the part of business or the closeness of parties to business in the policy space. A major change to the policy space could reveal much about the policy preferences of donor businesses and the extent to which their donations reflect party identification or a rational choice based on the party’s policies. This is not the first study of donations to take inspiration from the voting literature. Fisher (2000: 187–188) looked at donors in terms of retrospective voting on the valence issue of economic performance, while this article looks at prospective voting. Ironically, there are some respects in which theories of voting apply better to political finance than they do to voting. Voting is a rare event, the occasion for which the voter does not choose. In some electoral systems, voters are offered a categorical choice and in others they are presented with an ordinal choice. Since donors control when, how often, and how much they donate, their donations can provide much more information about their preferences than do the votes of citizens. The next section argues that Great Britain is a particularly apposite case to study this question.

Business and Political Parties in Great Britain

Business donations to political parties in the Great Britain are largely ideological rather than pragmatic. Businesses do not switch between the two major parties depending on which is in government; they do not hedge by contributing to both and few contribute to the Labour party at all. Donations to the Conservative Party are dynamic and it is possible that some of this represents an interaction of ideological and pragmatic motivations, such that businesses only donate to their ideologically preferred partner the Conservative Party, but are more likely to do so when there are pragmatic reasons for donating. In other words, businesses contribute more to the Conservatives when they are in government and when they are popular, and therefore likely to retain, or regain, control of government. Nevertheless, my focus will remain on the ideological motivation, as this is where party political preferences are relevant.

The Conservative and Unionist Party has been the party of business in Great Britain for a century (Fisher, 1994). Business donors to the Conservative party are connected to the party by old social networks centred on elite education and social clubs (Bond, 2007). The social bases of business donations point towards party identification. The Conservatives moved to the right under Margaret Thatcher in the 1980s and then towards the centre under David Cameron from 2005. Being pro-market has not always meant that the party’s policies were in the interests of the business establishment. For example, financial deregulation under Thatcher undermined many banks in the City of London. Nonetheless, businesses have generally had to think about how much they like Conservative policy, not whether they prefer Labour or the Conservatives. The Labour Party has also travelled along the left-right dimension. It adopted a radical leftist manifesto in 1983 and was dragged to the centre by a succession of leaders. In 1998, Prime Minister Tony Blair even declared that Labour wanted to be the ‘natural party of business’. This does appear to have generated some new support, especially from very rich individuals but also from the corporate sector (Fisher, 1997: 240–241). Nonetheless, business continued to provide substantial funding to the Conservatives, while contributing little to Labour (McMenamin, 2011). Individual donations do not seem to have been made on behalf of businesses (McMenamin, 2011: 1037–1038). Thus, very few businesses considered a choice between Labour and the Conservatives. It might be argued that this suggests party identification among Conservative donors, who would not switch even to a competent and economically orthodox Labour party. However, New Labour remained committed to a range of centre-left social policies, including substantial public spending on health and education. Overall, party competition has not allowed for a distinction between party identification and rational policy choice.

In recent years, the Conservatives have reported a lower income than their Labour rivals (Fisher, 2015b: 134, 2018: 177), although this largely consists of small donations associated with the party leadership elections of 2015 and 2016 and did not have to be individually declared. In terms of declared donations, from 2001 to 2017, the Conservatives raised 15% more than Labour. If their income from business had been the same as Labour’s, they would have raised 12% less in declared donations. ¹ Since British parties are among the poorest in Europe (Poguntke et al., 2016: 665), it is likely that financial shortfalls have greater consequences than in countries where parties are not so lean.

Like other actors, businesses look to the leadership for an indicator of the policies of a political party. Therefore, I concentrate on the party leaderships, not the members or the parliamentary parties. For 100 years, business was able to assess politics along one dimension, the left-right continuum. Since the Conservatives were always to the right to Labour, business did not have much difficulty in choosing its favourite. The emergence of Jeremy Corbyn has re-emphasised the importance of this dimension. However, Brexit has introduced a new nationalism-globalism dimension. The Conservative party policy is still centre-right on left-right issues. However, its ‘Hard Brexit’ policy is a nationalist position that emphasises national sovereignty and local control and is hostile to the pooling of sovereignty with other European countries and any notion of foreign influence over trade policy. Many Brexiteers advocate ‘a global Britain’, champion free trade and de-regulation, and want to pivot away from declining Europe to the dynamism of Asia and Britain’s former colonies. They do not propose protectionism or self-sufficiency, but all such trade must happen in the context of national sovereignty. Indeed, national sovereignty, ‘taking back control’, not a new economic dynamism, often appears to be the ultimate aim (Shrimsley, 2018). For the first time in a century, there is no British party which is clearly pro-business. There is a Conservative Party that is nationalist in foreign economic policy and a Labour Party that is socialist in domestic economic policy. The ideological position of the parties as a whole has moved in the same direction, but much less dramatically than that of their leaderships (Schmitt Loughran, 2017). The Conservative Party has many MPs and members who are concerned about economic consequences of a ‘Hard Brexit’. Many Labour MPs still regard Mr Corbyn’s ideas as extreme and impractical. The opinions of these politicians will be of little comfort to British business should the United Kingdom leave the European Union (EU) without a customs union arrangement, or industries be renationalised (Gordon and Pickard, 2017).

British business largely supported staying in the EU. However, a small number of variables predicted whether a business was ‘remain’ or ‘leave’: where the headquarters is located, ownership structure, the location of profits or trade and calculations on foreign direct investment (McKay, 2016: 3). The business leaders in favour of leaving the EU traded mostly in the United Kingdom and saw a ‘cost advantage or the opportunity to influence the political process after separation’ (McKay, 2016: 3). After the vote in favour of leaving the EU, the preference for remain effectively became a preference for a ‘soft Brexit’ over a ‘hard Brexit’. A soft Brexit implies minimal disruption, such as leaving the EU institutions but retaining access to the European single market and therefore remaining subject to the European law. A hard Brexit implies leaving the single market and usually the customs union too. The United Kingdom would have full control over immigration and trade, but would face barriers to trade with the EU.

A hard Brexit was the policy of the Conservative government from the Party Conference in October 2016 to at least until the Chequers meeting of the cabinet in July 2018. The government maintained that divergence in regulation is consistent with continuing frictionless trade with Europe. Critics, including EU negotiators, said this was impossible. So, the implications of the government’s position seemed to point towards a harder Brexit than the government’s explicit preferences. Labour has been unclear and Mr Corbyn ambivalent for much of the period since the referendum. In February 2018, Jeremy Corbyn gave a speech committing Labour to negotiating a new customs union with the EU. It was both unsurprising and amazing that business leaders lined up behind socialist Corbyn. For example, Carolyn Fairbairn, director general of the Confederation of British Industry, said, ‘The Labour leader’s commitment to a customs union will put jobs and living standards first by remaining in a close economic relationship with the EU’ (Kentish, 2018).

Obviously, Corbyn and Brexit are unique and contingent phenomena. Corbyn had to be persuaded to run for party leader. He aimed to keep the flag flying for socialism in the Labour Party and did not initially regard himself as serious competitor for the leadership. The Brexit referendum resulted from a promise that David Cameron made to quell a rebellious faction in his party. He was not a Brexit advocate and Brexit was a minority position in the parliament that approved the referendum. Of course, the Brexit vote itself was only won by only 3.8%. Nonetheless, both are particular cases of more general phenomena. The seeming abandonment of the median voter for a more ambitious and inspiring ideological vision is commonplace in politics, even in largely two-party systems like Britain’s. The populist revolt against globalisation is the political theme of the 2010s. Centre-right parties in many countries, the United States, Germany, the Netherlands, Denmark and Austria, to name a few, tacked towards populist positions on immigration and economic globalisation. The observational equivalence of rational choice and party identification donations is typical of the history of many established democracies. The potential for the emergence of a second dimension in their contemporary politics is also evident in many established democracies. The tumult of recent British politics offers an unprecedented opportunity to study which matters most to the business sector: the left-right dimension or the nationalism-globalism dimension? It also allows us to probe whether donations are motivated by a rational choice or party identification. Next, I introduce the data I use to study these questions.

Data from the UK Electoral Commission

The data on British political donations, like that generated by perhaps all disclosure regimes other the United States, is under-utilised. I study all 18,963 donations reported by all units of the Conservative Party between 12 February 2001 and 31 December 2017. This represents 6166 days and approximately 6225 donors. The data were cleaned to reduce over-counting of donors due to typographical errors, spelling, punctuation, honorific variations and donations by different sections or brands of the same business. A prominent example of the latter is JC Bamford Excavators, JCB Sales, JCB Research and so on. Nonetheless, there is still some small over-counting of the number of distinct donors. The purpose of the analysis is to assess how this highly politicised section of the business community, the Conservative donors, reacted to the election of Corbyn and the decision to the leave the EU.

Much good work on business donations is organised as a panel. Since these data contain the identity of donors and time of donation, it appears to meet the criteria for a panel analysis. However, there are several limitations. While the data are daily, no business donated on a daily or weekly basis, very few more often than annually, with many contributing only once a parliament or, indeed, only once in the whole 17 years. There is no information about the donors other than the amount and timing of the donation. Therefore, they cannot enter the dataset until they make a donation, as it is not known if the business even existed prior to that. Since there is no information about the businesses, it is not possible to explore various economic arguments about why donation rates and strategies might vary by type of business. Previous panel analyses in this literature draw their sample from economic databases, prior to checking whether the businesses have donated or not. In many cases, including that of the United Kingdom, this ends up with very sparse number of donations. In this article, I am able to study a very large number of donations and donors. This research investigates how politicised businesses react to changes in the political system, especially the policy positions of political parties. To this end, the observations are daily aggregates. The dependent variable is the number of distinct donors on a given day. This is intended to gauge donor support for the Conservatives.

The Electoral Commission classifies donations by the legal status of the donor. I exclude all such categories other than individual, company and limited liability partnership. This means three categories representing substantial funds are not included. These are public funds, trusts and unincorporated associations. The latter two types of donors tend to be fundraising arms of the Conservative Party itself, such as Holborn and St Pancras Conservative Association or the Lord Woolton Luncheon Club. ² The other categories are only responsible for small amounts. I combine the legal categories of company and limited liability partnership and label them as business, to indicate the overwhelming probability that they are profit-seeking organisations. The dependent variable is computed for individual and business donors. Due to the regulatory changes of 2010, models are restricted to donations of at least £7500 to achieve consistency over time.

To account for variations in the pragmatic and ideological motivations over time, I include a number of control variables. First, there is an indicator for whenever the Conservatives are in government. In the British political system, the government is very powerful and the legislature traditionally weak. If government is associated with more donations, this could indicate a pragmatic motivation. Second, there is an indicator for electoral campaigns, defined as the period from the dissolution of parliament to election day. Ideological donors help their preferred party when its need is greatest; pragmatic donors may also calculate that the reciprocation is more likely when money is provided at a time of great need. Third, there is a count of the days since the election. Pragmatic donations should be negatively associated with this variable because the more time a party has in power the more likely it is to be able to deliver a policy benefit. Ideological donations should be positively associated with it, as donors will wish to help their party prepare for a forthcoming election. Fourth is the election countdown variable, with a maximum of 365 if there is 1 year to a mandatory election. This captures the traditional build-up of the parties’ war chests in the last year of a parliament (Fisher, 2000: 198–200). Fifth is the Conservatives’ lead over Labour as measured by ICM’s voting intention polls. Pragmatic donors will know that a popular party is more likely to be able to deliver benefits than an unpopular one. Popularity may also be positively associated with ideological donations, as businesses share in general approval of the party’s positioning in the policy space. Sixth and last is a time trend.

The independent variables are indicators for the periods since the election of Corbyn and the Brexit vote. Corbyn represents a major policy shift on the part of the Labour party on the left-right dimension. Brexit represents a major policy shift on the part of the Conservative Party on the globalisation dimension. If business donors did not react to these shifts, they seem to be driven by party identification, not rational choice. While the Brexit and Corbyn shifts were both very large, it is not possible to say which was larger. Nonetheless, a big difference between the reactions to the two shifts would appear to suggest that one dimension is more salient than the other.

Figure 1 shows donations by distinct individual and business donors, as well as the Conservatives’ lead in polls over Labour, across the whole time period. There is evidently a strong correlation between individual and business donations. Moreover, the importance of the control variables is obvious. The number of donations climbs steeply around each of the five elections in the dataset. There also seems to be a strong association between the popularity of the Conservatives and donations. The time trend is subtler, but it appears to increase over time.

OPEN IN VIEWER

Figure 1.

Donations by Distinct Donors (Lowess Smoothed).

Comparing business and individual donations allows a probe of the validity of inferences about business donations. There are very strong theoretical reasons to expect contrasts in the behaviour of businesses and individuals. Businesses are rational by design. Directors are obliged to promote the financial interests of shareholders. This means that businesses are more likely to make pragmatic donations, and less likely to make ideological donations, than individuals (Baker, 2018: 578; Burris, 2001; McMenamin, 2009). If control variables linked to pragmatism are more strongly associated with business donations, inferences about motivation are that much stronger. Relative business rationality is also relevant to preferences. Business preferences should be more driven by the policy space, as opposed to party identification, than the preferences of individual donors. To the extent that business reactions to policy shifts are stronger than those of individuals, inferences about the source of preferences are stronger.

Table 1 presents regressions evaluating the impact of Corbyn and Brexit on donations to the Conservative party by businesses and individuals. Since the dependent variable is an overdispersed count, I use a negative binomial model. The standard errors are HAC Newey–West to correct for heteroscedasticity and autocorrelation.

OPEN IN VIEWER

Table 1.

Negative Binomial Model of Donations to the Conservative Party.

	Model 1 (Business)	Model 2 (Individual)
Government	0.41*** (0.11)	0.14 (0.15)
Campaign	0.79*** (0.15)	1.23*** (0.12)
Days since election	0.0004*** (0.00008)	0.0004*** (0.00009)
Election countdown	0.0015*** (0.0004)	0.0011** (0.0004)
Conservative lead	3.27*** (0.41)	3.02*** (0.42)
Corbyn	0.41*** (0.05)	0.094 (0.09)
Brexit	−0.14** (0.056)	−0.254*** (0.038)
Trend	−0.00003 (0.00005)	0.0001* (0.00006)
Intercept	−1.56*** (0.128)	−1.29*** (0.16)
Log likelihood	4845	−6763
N	166	6166

HAC Newey–West standard errors in parentheses.

*

p ⩽ 0.1; **p ⩽ 0.05; ***p ⩽ 0.01.

The control variables show the expected contrasts between distinct business and individual donations. Government is associated with a large increase in business donations, but the increase for individual donations is just over a third of that for business and is statistically insignificant. The campaign is associated with more donations, whether business or individual. However, the increase is about 50% bigger for individuals, which makes sense, given that this period would appear to be more likely to attract ideological than pragmatic donations. The number of days since the election is positively associated with both business and individual donations, suggesting an ideological motivation, and the coefficients are virtually the same size. The election countdown is also positively associated with distinct business and individual donations, although the effect is bigger for businesses. Finally, both are again positively associated with the Conservatives’ lead over Labour. The coefficient is slightly bigger for business, perhaps suggesting a more pragmatic motivation, but the difference between it and the coefficient for individuals is statistically insignificant. These comparisons suggest the reassuring conclusion that business donations are more pragmatic than individual donations. Next come the two policy shifts. Corbyn’s election provided a very large stimulus to business donations, but did not have a statistically significant effect on individual donations. Brexit had a large negative effect on both, but the coefficient in the individuals’ equation is over 80% bigger than that in the business equation. These contrasts suggest that business reacts more to policy shifts than individuals, probably because individuals are more driven by identification with the Conservative Party instead of a rational choice linking party policy to financial gain. They might also imply that the businesses and individuals attribute different saliences to the two dimensions represented by Corbyn and Brexit. The left-right dimension may not matter so much to individual donors, compared to the globalisation dimension. Donation patterns suggest that both the left-right and globalisation dimensions are important to the Conservative business donors.

Figure 2 plots key coefficients and confidence intervals from the business equation. Its purpose is to illustrate the magnitude of the effects of the policy changes, compared to the other variables which capture ever-present aspects of political competition and the electoral clock. They suggest that both policy shifts had very large effects on donor behaviour. Corbyn’s election generated a bonus in business donations almost exactly the same as being in government. Brexit produced a reduction in donations that was over one-third of the gain from being in government. Both the policy shifts have effects that are much smaller than the influx of donations associated with a parliamentary election campaign. However, this is a very unusual and short period of political life. Corbyn’s effect was greater than a standard deviation increase in the number of days since an election (515 days), while the effect of Brexit was a little bit smaller than this standard deviation increase. Corbyn had an effect double the size of a one standard deviation (6%) increase in the Conservatives’ lead over Labour, while Brexit’s magnitude was 75% of this increase in the Conservatives’ lead. The coefficients from a negative binomial model are difficult to interpret directly. The incidence rate ratios for Model 1 imply that donations under Brexit were 87% of what they otherwise would have been and that donations after Corbyn were 151% of what they otherwise would have been. This analysis shows that both policy shifts had the potential to transform the relationship between the Conservative Party and its donors. Without Jeremy Corbyn, Brexit would have exerted financial pressure on the Conservatives. Due to the impact of Corbyn, the Conservative treasury has been insulated from the effect of the party’s turn towards economic nationalism.

OPEN IN VIEWER

Figure 2.

Selected Coefficients from Model 1.

These figures are consistent with different donation patterns. For example, it could be that party identification was unaffected if loyal donors contributed at the same rate before and after Brexit, but the number of new donors fell substantially. In order to explore this and related possibilities, I studied the re-donation and new donation rates before and after Brexit. I split each into two periods, so that I can distinguish re-donation and new donation for a given period of identical length. Given the finding above that Corbyn’s election provided a strong stimulus to business donations, I end the first period for Brexit prior to that event. This reduces the likelihood of finding a negative impact of Brexit on either re-donation and/or new donation.

Figure 3 shows the set-up for Brexit and pre-Brexit. The second periods for the Brexit and pre-Brexit calculations are quite similar in terms of the regular political variables studied above, as Table 2 shows. The Conservatives were in government and there was an election in both periods. The Conservatives had a four-point lead in the Brexit period, compared to virtually a dead heat in the pre-Brexit period. The average number of days until an election was also over three times greater in the pre-Brexit period. Again, this suggests a reduced likelihood of a fall in re-donation and new donation in the Brexit era. However, the Brexit period shows a reduction in both re-donations and new donations. In the pre-Brexit period, there was a loyalty rate of 55%, meaning that over half of those donating in the comparison period donated again. By contrast, in the Brexit period, this was only 33%. Similarly, the new donor rate was much higher in the pre-Brexit period, in which 78% were new donors, compared to only 58% in the Brexit period. Regardless of donor strategy, there was a large negative Brexit effect. ³ This is further evidence of the dominance of rational choice, rather than party identification.

OPEN IN VIEWER

Figure 3.

Research Design for Strategy Analysis.

OPEN IN VIEWER

Table 2.

Donor Strategies.

	Pre-Brexit	Brexit
Donors	453	356
Loyal donors	96	150
Dropouts	77	303
New donors	357	206
Loyalty rate	0.55	0.33
New donor rate	0.78	0.58
Government	1	1
Campaign	0.067	0.066
Conservative lead	−0.0006	0.042
Days to election	1264	409
Election countdown	121	0

These figures are calculated by comparing the first and second period donations for pre-Brexit and Brexit, respectively.

The interplay of the two types of donors is potentially interesting. If policy preferences are more important than party identification, a shift away from the preferred positions of business should make a party more dependent on party identifiers. We can see this already in Table 2. The drop in new donors means that the proportion of donors in the loyal category has increased, even as the loyalty rate has decreased. These donors may be hard-core party identifiers who will stick with the Conservative Party, irrespective of its economic policy.

The preceding analysis has tested the hypotheses that Brexit and Corbyn constitute structural breaks in business donations to the Conservatives. It is also possible to think of them as temporary disturbances, the influence of which should wane over time. I estimated models in which the Brexit and Corbyn effects decay by the exponent of 0.001 for each day that passes. This means that Corbyn has decayed to 0.43 by the end of the dataset and Brexit to 0.57. This procedure has little effect on the Corbyn coefficient, but reduces the size of the Brexit coefficient by 41%. If the decaying version of Brexit is combined with the binary Corbyn, or even a growing Corbyn, it increases massively. So, these robustness tests show that Corbyn is relatively impervious to specification, but that Brexit is sensitive. This is more evidence that the Corbyn effect is stronger than the Brexit one. However, there are good substantive and technical reasons to prefer the simpler binary coefficients reported in Table 1. Technically, there are 49% more observations of the Corbyn era and, as noted above, almost one-third more variation in the decay variable. Substantively, the nature of Brexit was highly uncertain during the period under study; therefore, it would be odd if its impact steadily decayed. Since the debate was ongoing, I next explore the relationship between donations to the Conservatives and different Brexit options.

The period since the referendum on leaving the EU is clearly a new era in British politics. Nonetheless, not all Brexits are equally damaging for business. British politics has been dominated by a debate about what sort of Brexit the United Kingdom should seek. A ‘hard Brexit’ aspires to a substantial disengagement from the EU, freeing the United Kingdom from the Court of Justice of the EU and allowing the United Kingdom to negotiate its own trade deals. A ‘soft Brexit’ means a modest separation with the United Kingdom remaining in the single market and subject to the EU’s jurisdiction, but withdrawing from its decision-making institutions. In Table 3, I present assessments of the impact of the debate on the nature of Brexit on Conservative donations. Since it is restricted to the period since the referendum, there are no Brexit, Corbyn, or Government variables. Neither is there an Election Countdown variable as at no point did British politics enter within a year to a mandatory general election. I add two new variables, Hard and Soft, which measure the number of articles every fortnight in the British press mentioning ‘Hard Brexit’ and ‘Soft Brexit’, respectively. ⁴ Figure 4 shows that the Hard and Soft variables are highly dynamic and strongly correlated with each other. However, Hard seems much more prone to spiking. The two tallest spikes relate directly to Prime Minister Theresa May’s two clearest speeches on Brexit. They indicate a high level of validity. Since the Hard and Soft variables are non-stationary, I use the first difference of both.

OPEN IN VIEWER

Table 3.

Negative Binomial Model of Donations to the Conservative Party.

	Model 3 (Business)	Model 4 (Individual)
Campaign	0.8437*** (0.0717)	1.5528*** (0.0664)
Days since election	0.0005** (0.0002)	0.0012*** (0.0002)
Conservative lead	7.9720*** (0.6938)	3.6657*** (0.8797)
Hard Brexit (1st difference)	−0.0052** (0.0023)	−0.0039** (0.0018)
Soft Brexit (1st difference)	0.0336*** (0.0048)	0.0167** (0.0069)
Trend	0.0022*** (0.0003)	0.0014*** (0.0003)
Intercept	−14.1868*** (1.9703)	−9.3489*** (1.7030)
Log likelihood	−507.8167	−669.8952
N	547	547

HAC Newey–West standard errors in parentheses.

*

p ⩽ 0.1; **p ⩽ 0.05; ***p ⩽ 0.01.

OPEN IN VIEWER

Figure 4.

Articles Mentioning Hard and Soft Brexit.

Like the previous equations, these exhibit the theoretically expected contrasts between business and individual donations. As before, the campaign displays a larger increase in individual donations than business donations, consistent with the idea that the campaign is associated with ideological donations and individuals are more likely to be ideologically motivated. Effectively, of course, this means the 2017 election campaign. Days since the election has virtually the same effect on business donations, as it has in the much longer time series. However, the effect on individual donations is over twice its previous size. Again, this fits the hypothesis that individual donations are more ideological, and ideological donations are more likely to be made as an election approaches. The Hard and Soft variables have a greater effect on business donations than on individual donations, perhaps because businesses are paying more attention to the Brexit news. The coefficient for Hard in the business equation is about one-third larger than in the individual equation. Soft has a much bigger effect in both equations, and the coefficient in the business equations is almost twice the size of its equivalent in the individual model. The metric of the Hard and Soft variables produces small coefficients. Table 4 uses incidence rate ratios to calculate the impact of one standard deviation changes in Hard and Soft on business donations.

OPEN IN VIEWER

Table 4.

Substantive Implications of Model 3.

Variable	Counterfactual	Change in annual donations	Percentage change in annual donations
D Hard	26 fewer articles	33	16
D Soft	9 more articles	68	33
Conservative lead	5.9% reduction	−91	−44

Counterfactuals based on one standard deviation changes in Brexit sample. Donations refer to distinct business donors giving £7500 or above.

Different versions of Brexit have substantial effects on business donations. A one standard deviation decrease in articles mentioning ‘Hard Brexit’ implies a 16% increase in donations. A one standard deviation increase in ‘Soft Brexit’ is predicted to increase donations by 33%. For comparison, a one standard deviation reduction (5.9%) in the Conservative lead is associated with a 44% reduction in annual business donations. Obviously, this analysis is restricted to a much shorter period than earlier models. Furthermore, this is a particularly volatile period in British politics. For example, Prime Minister Theresa May called an election in 2017 to exploit her commanding lead in opinion polls, but her popularity plummeted in the few weeks of the campaign. Also, the volume of mentions of Soft Brexit is a little bit thin for the sort of inferences I am hoping to make. Nonetheless, the figures clearly point towards business sensitivity to subtler moves on the globalisation dimension, as part of the Brexit negotiations within the Conservative Party and between the United Kingdom and the EU. Moreover, it is consistent with anecdotal evidence (Coates, 2018).

Conclusion

Many business donors to the Conservative party are not party identifiers. If they were, they would have maintained donations, irrespective of the policy shift represented by Brexit. There were very large reactions to the major events of Corbyn’s election, and the Brexit vote and donations were sensitive to news about the specific form Brexit might take. Instead of party identification, many business donors make calculations on the basis of the location of parties in the policy space and pay close attention to actual, and likely, moves in that policy space. This article has not shown, however, that donors switch party because of policy. Voting theory is a well-developed set of ideas that can reveal much about donor behaviour. It can also help integrate political finance with more general work on party competition. Perhaps best of all, voting theory can bring political finance into contemporary debates about dramatic changes in political ideology and public policy in established democracies.

The partisan preferences of business donors can no longer be assumed because of the demise of the consensus on globalisation. The role of the state in the economy and the management of inequality continue to be contested, as they have been since the first mass-suffrage elections. This dimension has recently been separated from contestation of the extent to which the state will engage in frameworks of multilateral economic governance. If centre-right parties move away from their previous pro-globalisation position and consequently lose the financial backing of business, this may have positive democratic consequences in the area of political finance. Where parties are dependent on business for funding, they have incentives to resist transparency, limits on donations and public funding. Consequently, if centre-right parties decided they do not need business funding, they may allow reforms that ensure that citizens, not firms, fund party competition.

The electoral consequences depend on the direct effect of policy changes on voters (voters gained less voters lost) and the indirect effect of voters lost due to more meagre financial resources. Vote-seeking and office-seeking are not always equivalent in multi-party systems. Shifts away from globalisation may increase the chances of new coalitions with the populist or radical right but reduce the likelihood of more traditional coalitions with liberal parties. Coalitions with parties that are more hostile to globalisation, and may previously have been pariahs, are also likely to reduce business funding. The British case offers both reassurance and a warning to other business-funded parties considering a foray along the globalisation dimension. It is reassuring that the financial flood from Corbyn’s election was greater than the financial drought attributable to Brexit. This suggests that the left-right dimension remains more salient than the globalisation dimension. A pro-business position on the left-right dimension may allow centre-right parties to limit the financial damage from anti-globalisation moves. It is a warning that Brexit could have been a severe financial constraint for the Conservatives if it were not for Labour’s lurch to the left. Other centre-right parties cannot rely on their competitors to be so obliging.