Financing party branches: Evidence from the German federal system

ABSTRACT

Adequately financed branches contribute to the integration of regional interests into statewide parties. Yet, we have limited knowledge about the determinants of branches’ varying income levels in federal contexts. To address this shortage, this article elucidates why branches receive donations from citizens and businesses to different degrees. I hypothesise that party competition at the state level, the difference in regional economic performance and parties’ historical legacies can account for the level of branches’ donation revenue. Analysing German statewide party branches’ income from 2009 to 2017, this study finds support for the facilitating impact of state and federal electoral contests on donation levels. Regional economic disparities, by contrast, only marginally affect donation revenues. At the same time, parties’ path-dependent developments help explain asymmetries in average revenue levels between western and eastern branches. The study’s findings suggest that intense regional party competition contributes to branches’ financial independence within the statewide party organisation.

KEYWORDS:

• Statewide parties
• donations
• party financing
• Germany
• party competition

Introduction

Federal theory suggests that political parties running in multiple states integrate regional interests into federal politics (Bednar 2009; Filippov, Shvetsova, and Ordeshook 2004). By running party offices at the state and federal level, statewide parties incorporate regional interests into their agenda and settle conflicts between state regions internally (Detterbeck 2016b; Petersohn, Behnke, and Rhode 2015, 631; Verge and Gómez 2012). In this process, branches serve as an important channel through which regional interests are represented in the statewide party. However, for this integrative channel to work, regional branches require sufficient resources to run regional party offices properly and express regional interests in the decision-making process within the statewide party. And yet, while research has already investigated the role of regional parties in federal systems (Aarebrot and Saglie 2013; Deschouwer 2003; Detterbeck 2012), we still know relatively little about the determinants of branches’ income in federal contexts. This article aims to address this understudied aspect of party organisations in federal systems. The German federal system provides a good study environment in which to investigate the determinants of branch revenues. Because of Germany’s symmetric federal design, party branches operate in similar institutional contexts, which is why institutional factors can be mostly neglected when seeking to explain differences in donation levels (cf. Freitag and Vatter 2008, 17).

At the same time, German party branches can generate their income independently of the statewide party office. In contrast to membership dues and a large share of state subsidies, all donations made by individuals and business remain with the receiving branch and are not redistributed within the federal party organisation. Consequently, donations represent a revenue type that provides statewide party branches with additional means to run a regional infrastructure. Indeed, as Figure 1 shows, donations make up a considerable portion of many branches’ total revenues, yet to various degrees. First, branches of some statewide parties (AfD and FDP) appear to be more dependent on donations than others. Second, party branches’ donation incomes are, in many cases, relatively unstable over time. In light of the relative importance of donations for a large share of statewide party branches and longitudinal variation, I ask: What explains the differences in statewide party branches’ donation revenues?

Figure 1. Share of donations in branches’ total income, 2009–2017.

To identify potential explanatory factors of branches’ donation levels, I draw on approaches in federal theory and research into party organisations. Specifically, branch donation levels may be affected by political competition both at the state and federal levels, the socio-economic environment in the respective regions, and branches’ path-dependent trajectories, meaning that historical legacies account for varying incomes. To test these accounts empirically, based on an original dataset compiling finance data from 2009 to 2017,¹ I compare donation levels within and across party branches and apply multilevel negative binomial regression models to donation levels. Lastly, I discuss the empirical findings in light of the presented theoretical approaches derived from scholarship on federal systems and party organisation.

The results indicate that competition-related factors boost branches’ donation income, while economic conditions only marginally play into revenue levels. Both state and federal elections increase branches’ donation level, where higher fragmentation of state party systems reinforces the concentration of donations in electoral years. Moreover, parties’ path-dependent trajectories account for partially asymmetrically integrated statewide organisations in western and eastern states. Overall, the results imply that, despite the role of regional economic interests in German federal politics (Jeffery 1999, 2003), branches in states showing comparatively poor economic performance do not generally generate higher donation revenue than their counterparts in more affluent environments. Instead, political competition raises branches’ donation incomes, expanding their budget for running party offices and financing electoral campaigns.

Party financing in federal environments

Multilevel party competition and the financing of statewide party branches

Research into party politics at the regional level has argued that the form and extent of regional competition is an important determinant of how regional parties organise themselves and compete with each other (Deschouwer 2003). Since citizens have increasingly given up their long-term ties to a single political party (Flanagan and Dalton 1984) and instead became more likely to vote for different parties between elections, it can be assumed that competition has become an important driver of regional party politics. As Detterbeck and Renzsch (2004, 88) state, the weakening of enduring citizen-voter linkages has been replaced by conflicts over territorial interests and corresponding institutional arrangements. Because of this process, territorial interests, rather than political colours binding state actors together, might increasingly shape politics in the states, which could eventually nurture a more competitive setting at the subnational level. Party branches might exploit these local environments by running campaigns tailored to territorial issues and investing in organisational infrastructure to establish linkages with the regional electorate. In the long term, this could pay off for branches in gaining citizen support and advantages over competing branches.

From this perspective, competition should exert a significant influence on statewide party branches’ donation revenues. Specifically, there are three aspects of party competition that likely account for varying donation levels: (1) the intensity of political competition in the states, (2) whether branches hold executive office at the state and federal level, and (3) electoral campaigns.

First, the nature of party competition at the state level may influence branch revenue levels. The effect of fierce competition on the amount of funding is well known from first-past-the-post (FPTP) systems, in which candidates, commonly running on a party ticket, seek to generate as much revenue as possible for their electoral campaign. For instance, in the United Kingdom, marginal seats attract more donations to candidates than seats that will be most likely won by a candidate of the dominating party (Johnston and Pattie 2007; Pattie, Johnston, and Fieldhouse 1995). In competitive constituencies, candidates’ investment in the campaign oftentimes is a decisive factor for the election outcome. Applying this mechanism to mixed-member proportional (MPP) representation, which is widely used in German state elections, particularly fragmented state party systems may incentivise partisans to donate for their party to run an effective and elaborated campaign. As fragmented party systems are characterised by citizens voting for various smaller parties (Coleman 1995, 141), party supporters could be more willing to support their party with donations in order for it to gain strength vis-à-vis competing party branches and to run capital-intensive electoral campaigns that appeal to undecided voters.

There are other forms of party competition besides fragmentation. For one thing, the closeness of the previous electoral race might have implications for future elections, yet, as research into elections at the German state level shows (Müller 2018; Schniewind 2008, 199–267), there is considerable voter volatility between state elections. Moreover, the usual period between elections is five years, making it rather difficult to anticipate the closeness of the upcoming electoral race.

What is more, the coalition size in the federal parliament could also be seen as an indicator of party competition. However, a slim majority of a party coalition in a state parliament does not necessarily indicate a higher degree of competition in the next state electoral contest. For instance, two branches forming a coalition that holds a supermajority in a state parliament could aim to achieve a higher share of the vote in the next election to be no longer dependent on the coalition partner. Similarly, a minimum-winning coalition (i.e. a coalition narrowly passing the majority threshold) may consider the upcoming election highly competitive, as the incumbent parties could prefer to continue their coalition after the next election. Therefore, depending on the various state branches’ coalition preferences, the coalition majority size of state governments can lead to diverging degrees of competition among parties in the German states.

Given these considerations about the potential influence of party competition, I theorise that a low degree of party fragmentation, marked by a voter concentration to a low number of political parties, is likely to result in lower donation revenues. Vice versa, fiercer party fragmentation should lead to higher donation levels. Hence,

Hypothesis 1: The higher state party fragmentation is, the higher branches’ donation level in that state.

Besides the degree of party fragmentation in state party systems, it could matter whether parties are in state government or not. While some business actors donate for ideological reasons because they want to back a party sharing their political ideology, others have pragmatic intentions to donate; that is, they try to influence the policymaking process to enforce specific policies conducive to their businesses (McMenamin 2013). Pragmatic donations, however, are only effective when they are given to branches holding executive authority and governmental responsibilities. From a theoretical perspective, pragmatic donations to governmental parties present in two ways. On the one hand, discrete donations are characterised by a relatively high and direct financial contribution by the business donor to the party, aiming to influence policymaking processes. Reciprocal donations, on the other hand, are characterised by relatively low sums of money given to the party on a continuous basis. By doing this, the business actor aims to establish long-term ties to the party organisation (McMenamin 2013, 12–13). Whatever type of pragmatic donations a business actor chooses, governmental parties should receive higher donations than opposition parties.

This theoretical claim has found support in empirical studies, showing that governing parties tend to enjoy higher incomes than opposition parties (Fink 2017; Lösche 1993). Furthermore, business donations have been found to be influenced by power status at the state level: Fink (2017), analysing private and business donors’ spending patterns in the German states, showed that only corporate donors tend to give more to governing parties than to opposition parties.

Besides attracting pragmatic donations, governing state branches are likely more salient in the public and media, encouraging citizens and partisans to donate money to a branch in state government. It has been shown that rerunning candidates have a higher propensity to be reelected in elections, as officeholders usually enjoy higher media coverage and public attention than their challengers (Egner and Stoiber 2008; Stoiber and Egner 2008; Träger, Pollex, and Jacob 2020).

Adapting the incumbency effect to party branches’ income in the German states, one may argue that parties in government are granted more public attention, thereby becoming more relevant to business actors’ and citizens’ concerns, resulting in a higher willingness to donate money to a governing branch than to the state opposition (Lösche 1993, 222). Based on these two mechanisms, I derive the following hypothesis:

Hypothesis 2: If a party branch is in government at the state level, its donation level increases.

The governmental status at the federal level may also affect branches’ income levels. Detterbeck (2016a, 274) states that when a German statewide party is in opposition at the federal level, politicians at the state level become more influential in the party organisation. This effect may also apply to branches’ income. Donors could be more willing to support branches financially when their party is sitting on the opposition benches in the federal parliament, as supporters may wish to see policy changes implemented at least at the state level. Put differently, partisans adhering to a statewide party currently forming the federal opposition could identify a window of opportunity for their state branch to win the upcoming state election. Aiming to take advantage of that window, one would assume sympathisers to be more willing to equip the branch with financial means.

Alternatively, branches benefit from their umbrella party’s governmental status at the federal level. Like the incumbency effect at the state level, branches whose statewide party is in federal government would gain benefits from pragmatic donations and increased public attention. In this line of argument, the incumbency effect at the state level not only pays off for the federal party but also for the branches at the subnational level. This produces two competing hypotheses:

Hypothesis 3a: If a statewide party is in opposition at the federal level, its branches’ donation level increases.

Hypothesis 3b: If a statewide party is in opposition at the federal level, its branches’ donation level decreases.

It is well known that parties raise more money in election years than in regular periods (Fisher 2018). In the context of federal (and some unitary) systems, elections are held at multiple levels, where both state and federal elections can incentivise donors to give money to statewide party branches. While it is very likely that state elections contribute to branches’ donation revenues, the extent to which statewide parties’ federal offices and branches are intertwined (cf. Fabre and Swenden 2013) could moderate federal elections’ facilitating effect on branches’ income. If statewide parties are organised dualistically – the federal office and branches are only weakly meshed – one would expect federal elections to have little effect on branches’ donation revenue, as the federal office seeks to receive donations directly in federal elections rather than via their regional branches.

Conversely, in a firmly integrated statewide party organisation, regional branches and the federal office collaborate closely in federal elections, resulting in higher donation income for branches in national electoral contests. Here, branches are heavily involved in national election campaigns and seek to appeal to donors in federal races too. To test the effect of both state and federal elections, I will put the following hypotheses under empirical scrutiny:

Hypothesis 4a: In state election years, branches’ donation level increases in that state.

Hypothesis 4b: In federal election years, branches’ donation level increases.

In any case, the nature of political competition in the respective state may interact with elections at the state level. When fragmentation between parties is high, partisans could be even more eager to provide the branches with financial means. Party sympathisers may believe that branches require more funds to become visible in rather plural and challenging state party systems, while in less fragmented settings, the modest number of competitors does not intensify donation behaviour. Hence,

Hypothesis 4c: The more fragmented a state party system, the higher branches’ donation levels in state election years.

In addition to factors related to political competition in multilevel systems, a second – yet not necessarily conflicting – approach advances that parties’ behaviour is influenced by the surroundings and conditions in the various regions (Aarebrot and Saglie 2013). As Detterbeck (2012, 23–25) puts it, ‘[i]n order to integrate a strongly diversified society, there is a need to develop political styles and mechanisms which allow for the expression of territorial diversity. Political structures are thus seen as products of their societal context’.

In Germany, economic disparities are the most prevalent cleavage in federal politics (Jeffery 1999, 160). Considerable differences in states’ economic performances could result in diverging income levels. In states where the economy is comparatively strong, party branches are likely to benefit from this situation because individuals and corporations tend to harvest more financial means. As a result, more affluent citizens and enterprises may result in higher branch revenues. This produces:

H5: The higher a state’s economic performance, the higher branches’ donation revenue in that state.

Historical legacies and statewide party branches’ donation levels

Apart from competition-related and contextual approaches to statewide party branches’ financial behaviour, party branches’ trajectories may furthermore account for varying income levels. In the case of Germany, parties exhibit a firmly institutionalised structure, which partially originates in strict organisational regulations stipulated by the German Parteiengesetz (Party Law). Parties must organise in accord with the federal order, which requires well-equipped party branches employing a considerable amount of staff (Nassmacher 2009, 383). Even though the number of generous donations to parties has decreased over the past (Saalfeld 2000, 99), this funding source makes up a relevant share of parties’ revenues and is especially critical for the party on the ground, as a mediocre amount of money may contribute significantly toward the local unit’s budget (Lösche 1984, 39). Furthermore, since it is the local branch that is closest to members on the ground (Detterbeck and Jeffery 2009, 74), state branches enjoy the advantage of having the most direct access to donations.

At the same time, parties deal with institutional environments in different ways. Most importantly, the party’s historical legacy, or, in other words, “sticky” party traditions’ (Hepburn and Detterbeck 2013, 88), influence the development of intra-party organisational structures. In this line of argument, the institutionalisation of organisational structures during the party’s formative years shapes the organisation’s development. With respect to statewide parties, it may matter whether the organisation was initially established at the federal or state level; parties founded from top-down may remain relatively centralised, while bottom-up party institutionalisation may result in a rather decentralised organisation in the long term.

Additionally, more regionally integrated branches should appeal to individual and businesses donors considerably more than branches that do not enjoy much public support in their regional political environment. Parties legacies are thus likely to affect branches’ donation levels. There is a rich literature on the trajectory of German statewide parties, which allows formulating empirical expectations regarding diverging donation levels that can be traced back to path-dependent developments of party organisations since the beginning of the Federal Republic.

As the two dominant parties in German party politics, the Social Democratic Party (SPD) and the Christian Democratic Party (CDU) are likely to generate a relatively high number of donations across German regions. Moreover, besides their prominent role in politics at the federal level, many branches have been part of state governments, illustrating their principal role in state politics. It can thus be expected that the two parties’ donation levels should not vary too enormously between branches.

Yet, the different sequence of establishing the two parties is likely to have affected the branches’ composition of revenue. The SPD has been categorised as a ‘unitary organisation’ (McMenamin 2013, 29), as the party was established first at the federal level and then founded branches in all German states. By contrast, the Christian Democrats (CDU), with its Bavarian collaboration partner the Christian Social Union (CSU), first organised in the various states and established regional organisations in the states. Contrasting with the SPD’s formative years, the CDU’s federal umbrella party was founded after establishing regional organisations in 1950. Because of this sequence, the federal party was comparatively week in the first decades of its existence. CDU chairman Helmut Kohl later strengthened the federal party’s office and its resources in the 1970s, yet the Union’s state branches have preserved their influential role in federal politics (Detterbeck 2012, 136).

Based on these two different trajectories at the first stages of the parties’ organisational evolution, one may expect donations to make up a higher portion of CDU branches’ income because of its regional units’ regional independence. Besides these differences, the unification process in the 1990s should have affected both parties to similar degrees. In the aftermath of integrating the eastern states into the Federal Republic, CDU and SPD were coping with the structural heterogeneity between ‘old’ and ‘new’ states.² As Jeffery (2005, 83) points out, traditional western German parties have not fully integrated into all eastern states. Nonetheless, though confronted with higher electoral volatility in the east (Mannewitz 2017, 231), SPD and CDU were able to gain strength in the eastern states and built up organisational structures, such as well-equipped branch offices. However, advancing a more reluctant assessment, Detterbeck (2012, 136) notes that all eastern state branches are still ‘organizationally weak’. Empirically, we may thus expect the CDU’s and SPD’s eastern branches to receive fewer donations than their western counterparts.

The Free Democratic Party (FDP) has been traditionally more dependent on donations than other German parties (Lösche 1984, 60), both at the federal and state level. However, its trajectory shares some similarities with that of the CDU, as the organisation was first founded at the state level and later established at the federal level (Filippov, Shvetsova, and Ordeshook 2004, 243). Likewise, the Green Party, founded in 1980 in the (western) Federal Republic, implemented a decentralised rather than a centralised internal structure. In so doing, the party sought to establish grass-roots-oriented funding patterns to avoid power centralisation (Lösche 1984, 95–96). Thus, it is likely that the FDP and Green branches generate substantial donation revenue across states. Similarly to the CDU and SPD, however, western branches should exhibit higher donation levels than their eastern counterparts.

Lastly, The Left and the Alternative for Germany (AfD) are likely to exhibit significant differences in donation income levels between western and eastern branches. As Abedi (2017, 458) notes, ‘[d]ifferences in western and eastern voting behaviour have their roots in the discrete identities’, which can mainly be explained by citizens’ distinct experiences in different political cultures prior to unification and lasting socio-economic disparities (Abedi 2017, 475). The socialist Left Party’s eastern branches continue to be more successful in electoral contests and have considerably more members than their western counterparts. In a similar vein, the far-right AfD is most successful in eastern state elections (Abedi 2017, 457). Therefore, the two parties’ higher electoral support may suggest that The Left and AfD generate more donations in the east, leading to financial asymmetries between western and eastern branches.

Overall, statewide parties’ historical legacies are likely to explain variance in income levels that cannot be captured by the level of competition and economic performance in the various states. Table 1 summarizes the expectations on branches’ donation revenues derived from parties’ trajectories from the outset of the Federal Republic. Empirically, path-dependencies are challenging to study with conventional regression analyses, which is why the following empirical study builds on both descriptive and inferential quantitative techniques. While the effects of competition and state economy can be investigated in a regression framework, path dependencies may become apparent in systematic deviation in generated income along German parties’ distinct trajectories outlined in this section. In addition to estimating the effect sizes of competition- and economy-related factors, I shall investigate to what extent branches within statewide parties differ internally with regard to their donation levels.

Table 1. Expectations on branch income differences and empirical findings between and within German statewide parties.

Party	Expectations from party trajectory	Empirical findings
CDU	Eastern branches receive slightly fewer donations than their western counterparts; higher share of donations in overall income than SPD	Relatively balanced individual and corporate donations between eastern and western branches; higher share of corporate (but not individual) donations in overall income than SPD
SPD	Eastern branches receive slightly fewer donations than their western counterparts; lower share of donations in overall income than CDU	Relatively balanced individual and corporate donations between eastern and western branches; lower share of corporate (but not individual) donations in overall income than CDU
FDP	Eastern branches receive slightly fewer donations than their western counterparts; highest share of donations in overall income among smaller parties	Relatively balanced individual donations and moderately higher corporate donations in eastern than western states; substantially higher share of donations in branch income among smaller parties
The Left	Western branches receive substantially fewer donations than their eastern counterparts; lowest share of corporate donations in overall income among smaller parties; most branches do not accept donations from businesses	Substantially higher individual and corporate donations in eastern than western states; marginal corporate donations
Greens	Eastern branches receive slightly fewer donations than their western counterparts	Slightly lower individual donations in eastern than western states; higher corporative donations in northern/southern than other states
AfD	Western branches receive fewer donations than their eastern counterparts	Relatively balanced individual donations between eastern and western branches; substantially higher corporate donations in eastern than western states

Research design

Subnational research has become a popular approach to hold contextual factors relatively constant while investigating multiple subunits of a political system (Freitag and Vatter 2008, 17; Jacob and Pollex 2021). In Germany, the Party Law, including parties’ obligation to publish their incomes and expenditure, applies to all branches, facilitating the comparative analysis of income patterns.

The dependent variable of the analysis is donations received by statewide party branches per annum measured in euro. To test whether the various explanatory factors affect types of donations in different ways, I differentiate between donation revenue from citizens (i.e. individual donations) and legal bodies (i.e. corporate donations). In Germany, donations are the only income type that remains entirely with the receiving branch within the statewide party. In contrast, other sources, such as fees, mandate holder fees (‘party taxes’) and state subsidies, are at least partially redistributed based on intra-party arrangements that vary from one statewide party to the next. While it is important to note that donations to branches do not necessarily originate from donors residing in the respective state, it can be assumed that the lion’s share of donations comes from local individuals and businesses. The data is taken from the annual reports on parties’ incomes, expenditure and assets, which the parties must submit to the President of the Bundestag (Deutscher Bundestag n.d.).

In the reports, state branches’ income is listed separately for revenues of the branches’ central offices (Landesgeschäftsstellen) and local units of the branch (nachgeordnete Gliederungen). As both build the party branch, the two levels’ revenue is summed up to a merged value for each state branch. The data encompasses the period from 2009 to 2017, whereby two complete electoral cycles at the federal level can be examined.³ Therefore, this period allows for testing the impact of state and federal elections and the incumbency effect in at least two government terms per state. Overall, the dataset (N = 791) encompasses the annual donation revenue of the six main statewide parties’ branches over nine years. Since the AfD was founded in 2013, data on its branches are only available for four years. To receive an overview of the income levels within the Union ‘party family’, the CSU is considered in the descriptive part of the analysis but not in the regression models because it is not a statewide party.

The empirical analysis proceeds in two steps. Firstly, to systematically shed light on differences in branches’ income levels that may point to path-dependent developments, I calculate the average income from individual and corporate donations of party branches per 1000 inhabitants between 2009 and 2017. Next, I report the extent to which each branch deviates from the average income among their statewide party’s branches. More formally, I calculate $\text{\%}\mathrm{devianc}{\mathrm{e}}_{pb}=(1-{\displaystyle \frac{\mathrm{incom}{\mathrm{e}}_{pb}}{{\sum }_{b=1}^N(\mathrm{incom}{\mathrm{e}}_{pb})/N}})\times 100,$where p is the statewide party, b a branch of a statewide party, N the number of branches per statewide party (i.e. 16 for each statewide party), and $\mathrm{incom}{\mathrm{e}}_{pb}$ the branch’s income divided by 1000 state inhabitants. This measure allows comparing the variance across parties as it reports the percentage deviation of each branch from the average branch income of the respective statewide party. Substantially, negative values would indicate lower, and positive values higher, income than the average branch of the statewide party. By contrast, values close to zero per cent suggest that the branch approximates the average branch income of its statewide party. I display the deviations in a heat map of German states by party to facilitate the evaluation of deviances between and within parties.

In a second step, I implement regression models to identify statistical predictors of donation levels. The independent variables of these model are operationalised with the following indicators. Firstly, the effective number of electoral parties (ENEP) measures the level of party fragmentation in the states based on all parties’ vote shares in state elections (Laakso and Taagepera 1979).⁴ I choose this measure over the elective number of parliamentary parties (ENPP), as smaller statewide parties regularly fail the 5 per cent electoral threshold but nevertheless operate in states with relatively weak electoral support, including running branch offices and electoral campaigns.

Secondly, governmental status is coded as a binary variable, indicating whether a party is in state government (1) or not (0). Another variable measures whether a party is in the opposition at the federal level (1) or not (0). Moreover, a binary variable indicates whether a state election (and thereby an election campaign) was held in the year of the observation (1) or not (0). Similarly, another binary variable is created that indicates if a federal election took place (1) or not (0). Lastly, a continuous variable takes account of the respective state’s GDP economic performance.⁵

I implement negative binomial regression models to estimate the association between the independent and dependent variables. This generalised linear model is preferable to linear models because of the outcome variable’s specific characteristics (i.e. donation revenue in euro). Specifically, branch revenues deviate substantially between states and parties, leading to a highly skewed distribution of the outcome variable. However, as linear models should be implemented with reasonably normally distributed data, this model specification may lead to spurious results with donation data. By contrast, negative binomial models can accommodate heavily positively skewed and overdispersed count data in its parameterisation (Hardin and Hilbe 2007, 200–203), which applies to the distribution of donations, a variable on a currency scale. Therefore, in line with previous analyses using financial data as an outcome variable (McMenamin 2019), I employ negative binomial models.

Further, the regression model takes account of the multilevel structure of the data (Gelman and Hill 2006). For instance, while GDP is measured at the state level, the incumbency variable refers to the branches in the states. To take account of the data’s nested structure, the model includes random effects for (1) state and (2) parties nested in states.

Results

Donation levels of statewide party branches

Figure 2 displays the deviation of each party branch from the average level of individual donations per 1000 inhabitants. Within the Union party family, the Bavarian CSU and the Rhineland-Palatinate CDU branch generate about 75 per cent more individual donations than the average CDU branch.⁶ Although the highest donation levels within the CDU occur in the western German states, eastern branches do not receive remarkably less income from individual donations. For instance, the eastern CDU branch of Mecklenburg-Western Pomerania exhibits a similar relative donation level as the western branch of North Rhine-Westphalia.

Figure 2. Deviation of average individual donations to branches of (a) CDU, (b) SPD, (c) FDP, (d) The Left, (e) Greens and (f) AfD.

Like the CDU/CSU’s donation revenue, the SPD branches in the east receive donations levels comparable to western counterparts. The only exception is the Saxony branch, which receives almost 70 per cent less than an average SPD branch. In contrast, The Left’s individual donations mirror its asymmetric integration: All eastern branches generate remarkably more income from that source than Western branches. Even though the AfD has been the most successful in eastern state elections, western party branches, most notably in Baden-Württemberg, obtain on average the same amount of donations as their eastern counterparts. As for the share of donations in the overall income of the two German principal parties, CDU branches do not appear more reliant on donation revenue than SPD branches across states (see Figure 1 in Online Appendix).

In line with the Greens’ and FDP’s more muscular organisational strength in western states, their party branches tend to receive more individual donations in the west than in the east. However, the asymmetry tends to be more prevalent within the Green Party, which generates most of its individual donation income in southern and northern states. In contrast, the eastern Saxony-Anhalt branch receives about 70 per cent less than an average Green branch. Substantially, this branch generates only 10 euro per 1000 inhabitants, one of the lowest figures among all branches’ individual donation income. The FDP branches in the east, particularly the Saxony branch, appear to enjoy higher financial support from individual donors, suggesting that this party has appealed to private donors in the east to a greater extent than the Greens. Like AfD branches, FDP branches are the most dependent on individual donations in their overall income (see Figure 1 in Online Appendix).

Turning to business donations, Figure 3 shows the deviation of each branch from the average corporate donation by statewide party. The CDU/CSU’s income from that source does not vary substantially between east and west, with the Bavarian CSU having the most corporate donations per inhabitant at its disposal. Yet, the Mecklenburg-Western Pomerania branch in eastern Germany also receives a comparatively high level of business donations per capita. This finding is somewhat surprising, as economic performance is considerably higher in the western states, most notably in Baden-Württemberg, but business actors do not support the CDU to a higher degree in that state than in others.

Figure 3. Deviation of average corporate donations to branches of (a) CDU, (b) SPD, (c) FDP, (d) The Left, (e) Greens and (f) AfD.

As the CDU has been traditionally closer to employers’ interests than the SPD, it is not surprising that SPD branches receive on average less from business donations than the CDU (see Figure 2 in Online Appendix).⁷ While income of this type is comparatively evenly distributed across states, the Saarland branch turns out to generate the highest income from business donations. As for The Left, once more, the eastern branches receive more than their western counterparts. However, since The Left accepts business donations only under exceptional circumstances, the level of that revenue type is close to zero. In contrast to individual donations within the statewide party, most of AfD’s eastern branches appear to receive the highest relative amount of corporate donations, where business donations of AfD branches tend to be generally low. The Greens’ and FDP’s revenue levels from corporate donations do not differ substantially between east and west.

Summarised in Table 1 along with the expectations about income differences derived from party trajectories, the four ‘traditional’ parties CDU/CSU, SPD, Greens and FDP do not generally receive a higher amount of donations in the west. Instead, eastern branches of these statewide parties have been particularly successful in generating business donation and, in a few cases, even exceed their western counterparts in corporate donations per state capita. While The Left is asymmetrically integrated across the states in terms of both individual and corporate donations, the AfD, despite its better electoral performance in the east, has also been able to acquire donations from citizens in the west.

Determinants of donation levels

Table 2 reports the results of the multilevel regression models. To analyse the impact of the dependent variables on overall donation levels, Models 1–3 estimate the effects of the explanatory variables on the sum of individual and corporate donations. In Model 1 (without interaction effect) and Model 2 (with interaction effect), state government participation, state elections and federal elections turn out to be statistically significant. To see whether this pattern holds for (1) a smaller sample without AfD observations and (2) individual and corporate donations separately, I refine the aggregated regression model.

Table 2. Multilevel negative binomial regression on total, individual, and corporative donations.

	Model 1 (all donations)	Model 2 (all donations)	Model 3 (all donations)	Model 4 (individual donations)	Model 5 (corporate donations)
	std. Beta	std. Beta	std. Beta	std. Beta	std. Beta
Intercept	12.062 *** (0.400)	12.058 *** (0.399)	12.208 *** (0.436)	11.911 *** (0.349)	9.305 *** (0.840)
ENEP × 10	0.006 (0.032)	−0.055 (0.032)	−0.079 * (0.031)	−0.062 * (0.031)	−0.113 (0.124)
State government	−0.001 (0.051)	−0.007 (0.050)	−0.011 (0.046)	−0.005 (0.048)	0.205 (0.184)
Federal opposition	0.034 (0.051)	0.036 (0.050)	0.037 (0.046)	0.033 (0.048)	0.073 (0.204)
State election	0.459 *** (0.037)	0.461 *** (0.036)	0.470 *** (0.035)	0.399 *** (0.035)	0.967 *** (0.152)
Federal election	0.305 *** (0.031)	0.326 *** (0.030)	0.340 *** (0.030)	0.283 *** (0.029)	0.905 *** (0.132)
State GDP	0.001 (0.146)	0.016 (0.143)	−0.068 (0.134)	0.014 (0.139)	−0.443 (0.539)
State inhabitants	0.752 *** (0.200)	0.727 *** (0.196)	0.788 *** (0.183)	0.737 *** (0.193)	1.094 (0.571)
(ENEP × 10) × State election		0.205 *** (0.034)	0.228 *** (0.033)	0.194 *** (0.033)	0.348 * (0.143)
Random effects state	✓	✓	✓	✓	✓
Random effects parties nested in states	✓	✓	✓	✓	✓
AfD included	✓	✓		✓	✓
Observations	791	791	711	791	791
Log-Likelihood	−10283.892	−10266.384	−9238.646	−10102.647	−8536.582

Standard errors in parenthesis. Random effects for states and parties nested in states. GDP and State inhabitants rescaled from 0 to 100 and rounded to the nearest integer. ENEP multiplied by 10 and rounded to the nearest integer. Quadratic parameterisation of the negative binomial distribution (Hardin and Hilbe 2007).

* = p < 0.05; **=p < 0.01; *** = p < 0.001.

First, it may be that the AfD’s income patterns (that is, a considerably high share of donations in the overall income, see Figure 1) are too different from the established parties’ revenue structures. I thus test whether the results for all parties hold when dropping all AfD branches from the sample (Model 3). Besides participating in state government, all estimates retain their statistical significance, providing support for the effects identified for elections at the state and federal levels, population size and the interaction term between party fragmentation and state elections.

Second, the independent variables could exert different effects on individual and business donations. To test this proposition, Model 4 shows the effects on individual donations only. In this specification, state population increases individual donations the most, followed by state elections (β = 0.399) and federal elections (β = 0.283). This finding suggests that citizens tend to donate more to branches in state election years and, to a lower extent, in federal electoral contests. Compared to the other effect sizes, the negative impact of party fragmentation (ENEP) is negligible.

As Model 5 shows, a similar pattern emerges for corporate donations. Again, both state elections (β = 0.967) and federal elections (β = 0.905) exert a considerable positive effect on business donations. In contrast, recognising that GDP does not yield a significant coefficient, economic performance is negatively associated with corporate donations (β = −0.443). This somewhat surprising finding suggests that branches in states with lower economic performance tend to receive more corporate donations. While governmental status is not associated with individual donations, branches participating in state government tend to receive more business donations than branches sitting on the state legislature’s opposition benches. In contrast, if the umbrella party is in the federal opposition, that branch will generate less revenue from business donations. However, both coefficients of both federal and state government participation fail the significance test.

As for the hypothesised facilitating effect of a fragmented state party system and state elections, the interaction effect of ENEP and state election yields a positive association with both individual (β = 0.194) and corporate donations (β = 0.348). This result suggests that, as hypothesised in H4c, the higher party fragmentation, the higher branches’ donation revenues in state election years. Figure 4 illustrates this effect: From the state with the lowest degree of fragmentation to that with the highest, the models predict a rise for individual and corporate donations in state election years. While the confidence intervals are relatively large, branch observations in the state of Berlin and a few in Brandenburg, Bremen, Mecklenburg-Vorpommern, Schleswig-Holstein, Saxony, Saxony-Anhalt and Thuringia yield higher ENEP values (see Online Appendix Figure 4). This group of states is of different population sizes and located in different regions, even though most of them in eastern Germany. This variance indicates that the interaction coefficient is likely not driven by outlying ENEP in one or a few states.

Figure 4. Predicted counts of a) individual and b) corporate donations with 95% confidence intervals (see Online Appendix for predicted counts of all donations).

Overall, the analysis suggests that competition-related variables affect the sum of individual and corporate donations. Moreover, this pattern holds for the sample without the more recently established AfD and individual and corporative donations separately.

Conclusions

What accounts for varying donation levels of German statewide party branches? This article addresses this question by deriving potential determinants of statewide party branches’ donation levels from scholarship on party organisations in federal systems. From a competition perspective, regional party competition, governmental status and electoral campaigns may influence statewide branches’ donation revenues (Hepburn and Detterbeck 2013; Hildebrandt and Wolf 2016). From an economic angle, states’ varying economic performance may similarly shape donations levels (Detterbeck and Jeffery 2009; Jeffery et al. 2014). Lastly, parties’ different historical legacies could account for persisting asymmetries in branch revenues since parties are organisations that have evolved practices and institutions over time (Hepburn and Detterbeck 2013).

To test these approaches empirically, this article examines German statewide party branch revenues from donations between 2009 and 2017. The empirical analysis suggests that branches’ donation levels are decisively affected by regional political competition. Elections at the state and, to a lower extent, federal level politicise both citizens and businesses to support state branches financially, providing them with additional means for running party offices and electoral campaigns. Moreover, this mobilisation effect appears to become stronger in more fragmented party systems, in which branches generate more donations in the electoral season but less in non-election years.

By contrast, states’ economic performance only matters to a low degree for branch revenue levels, indicating that the economic environment only marginally influences branch incomes. Interestingly, corporate donations to state branches tend to be higher in states with lower economic performance, even though the effect failed to be statistically significant. At first sight, this finding seems to be counterintuitive as one would expect more business donations in states with higher economic performance. McMenamin (2012) explains this puzzle, arguing that businesses already participate in the policymaking process in coordinated market economies like Germany. From this perspective, businesses operating in states with comparatively weak economic performance may want to expand their influence at the state level to influence policies conducive to their enterprise. Firms in more affluent states, however, may not see the necessity to exert influence on party branches because they already consider the state economy to be competitive. Overall, the hypothesis that high economic performance increases party branches’ income in the states substantially could not be confirmed.

Therefore, the divide between poor and prosperous states in German federal politics is not reflected by branches’ donation levels. Nonetheless, as the descriptive inquiry of donation levels in the states has shown, most statewide parties exhibit asymmetrical patterns that mirror their historical trajectories. While the largest statewide parties, namely CDU and SPD, are integrated into both east and west, the Green Party and FDP tend to generate a higher income in the latter. By contrast, The Left is the most established in eastern states, which can be explained by the party’s historical strength in that region (Detterbeck 2016a). The far-right AfD has been remarkably capable of appealing to individual donors in all states, yet only to business donors in the east. Thus, similarly to differences in voting behaviour and political attitudes between eastern and western Germany (Abedi 2017; Mannewitz 2017), the distinct political environments also manifest themselves in their branches’ donation revenue.

Overall, the findings of this study imply that party competition at the state level is crucial for branches to generate individual and business donations. Without a competitive nature of subnational politics, branches would likely be in a financially weaker position. Simultaneously, statewide parties’ historical legacies account for asymmetric donation levels within statewide parties, most importantly, for their more robust integration in either western or eastern Germany. In sum, intense party competition at the state level appears to strengthen branches’ financial independence, which can help them formulate regional interests more confidently in decision-making processes within the statewide party.

As this article has sought to show, empirical analyses of branch revenue provide fertile ground for research into multilevel party politics. Even though the current availability of financial data on regional parties sets limits to the number of potential cases other than Germany, the trend to adopt more transparent party finance regulation in established democracies may ultimately offer additional promising data sources in the future.

Replication material

Replication material can be found at https://doi.org/10.7910/DVN/97EDVS.

Acknowledgements

I am indebted to Chris Jardine, Sabine Kropp, Christoph Nguyen, Ugur Ozdemir, Greta Schenke, Noel Treffinger, Editor Louise Tillin and two anonymous reviewers for helpful comments on previous versions of the manuscript. All remaining errors are my own.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1 The full dataset can be accessed at: https://doi.org/10.7910/DVN/FSCDPI.

2 However, as Rueschemeyer (1998) remarks, it is important to note that the SPD was required to build up branches from scratch after unification, while CDU and FDP could integrate existing branches of former ‘block parties’ into their statewide party.

3 Federal elections were held in 2009, 2013 and 2017.

4 $\mathrm{ENEP}=1/{\sum }_{i=1}^n{p}_i^2$. Parties below the 5 per cent threshold were also considered for calculating the ENEP based on state election results. The ENEP values were assigned to all state branch observations in the respective legislative period.

5 The data for state GDP and inhabitants was taken from the Federal Statistical Office. A longitudinal overview of state GDP and inhabitants can be found at www.statistik-bw.de.

6 In absolute terms, both branches generate about 400 euros per 1000 inhabitants on average.

7 The average SPD branch corporate donation level is 32 euro per 1000 inhabitants, while CDU branches generate on average 120 euro per 1000 inhabitants.