Partisan Influences in Dutch Politics

Abstract

We investigate whether and how partisan politics played a role in the Netherlands. To do so, we review existing literature and prepare descriptive statistics from recent datasets. We focus on two related questions: (i) Are there effects of partisan politics in the Netherlands in the past 60 years? (ii) Do developments in party and voter behavior influence the relevance of partisan influences for Dutch governance? As we find that there is only limited existing research on partisan politics in the Dutch context, we empirically explore the relationship between government ideology and typical, country-level partisan outcome variables. We find mixed evidence for partisan policymaking in the Netherlands. Some findings are in line with partisan theory, some are counter-indicative. Exploring the role of party and voter movements, we document a shifting political landscape in which the potential for traditional partisan influences has become smaller.

1. Introduction

It is often assumed that government ideology plays an important role in determining economic policymaking. A typical interpretation of partisan theory (for example, Hibbs 1977; Alesina 1987) is that the ideological ‘color’ of the government affects policymaking because political parties have an incentive to pursue policies that are in the best economic interests of their constituents. The baseline prediction of partisan theory is that left-wing governments favor more expansionary economic policies than right-wing governments (Potrafke 2017). Many scholars have examined whether this expectation can be confirmed.¹ For this special issue, we focus on the Netherlands and review the evidence for partisan influence in Dutch politics. We ask two main questions: (i) Is there any evidence for partisan politics in the Netherlands? and (ii) Did developments in the party and voter space influence the relevance of partisan politics for the Netherlands?

We start with an overview of the Dutch political landscape and discuss how the political and governance structure may affect the potential for the politicization of governments. We then continue with an overview of the role of partisan politics in the Netherlands. Unfortunately, we can only rely on little existing literature for this overview. To our knowledge, there are no country-level studies for the Netherlands. The few existing studies focus on local taxes (Allers et al. 2001; Allers and Rienks 2024), for which no partisan effects are found, and on the appointments of civil servants (Ennser-Jedenastik 2016), which seem to follow partisan logic.

Given the lack of empirical research on partisan politics on the macro level, we conduct a simple empirical analysis that describes the relationship between government ideology and typical partisan outcome variables in the Dutch context since 1960, such as GDP, government spending, and taxation. Besides sketching the relationship in various figures, we also perform a series of multivariate regressions to estimate conditional correlations between government ideology and our policy indicators. Our results are mixed. Some correlations are in line with partisan ideology, for example, for GDP growth, inflation, and government debt. Yet, we also find counterintuitive associations, for example, regarding total government expenditure, the primary balance, and labor income taxes. Moreover, the following caveats apply. First, we cannot draw causal conclusions from these analyses. Second, the Netherlands has not had a purely left-wing government in the past seven decades. Therefore, there is no counterfactual to compare to the policies under the predominantly right and center governments in that period. We see our analysis as a first step toward investigating partisan influence on country-level policies in the Netherlands and invite future research to confirm our results using more adequate identification strategies.

We then turn to our second main question and ask whether movements of Dutch parties and voters alike could have played a role in the relevance of partisanship for policymaking. Based on a descriptive analysis inspired by existing findings for Western democracies in general and the Netherlands in particular, we conclude that there may be little room for traditional partisan politics in recent times. On the one hand, political competition between parties seems to have shifted to different policy issues than economic ones. This has created opportunities for new parties on both the ‘New’ left and the populist radical right to enter the Dutch political stage. On the other hand, the traditional class cleavages, on which partisan theory is based, seem to have lost importance, while new cleavages appear. Given that these developments are not just unique to the Dutch context, a re-interpretation of partisan politics on other ideological dimensions could create fruitful avenues for future work.

2. The Dutch political context

In this section, we discuss the most important characteristics of the Dutch political system.² We start with a description of the characteristics of the Dutch party system. We also give an overview of important periods in the development of the Dutch political landscape as it stands now. In this context, we highlight two elements of the Dutch political system that have played a significant role in shaping the Dutch political context: pillarization and consociationalism. We then move from parties to governments and discuss the governance structure and the partisan ideology of Dutch governments over time.

2.1 The party system

The Netherlands has a multi-party system, which has always been quite fragmented. In the post-war period, the number of parties competing in elections ranged from 10 parties in 1946 and 1956 to a record high of 37 parties in 2021 (Andeweg et al. 2020; Van Holsteyn and Irwin 2022). As the entire country is one electoral district, there is formally no geographical representation of subnational regions in the parliament.³ This implies that there is a direct link between parties’ vote share and their seats in parliament and that the threshold to gain a seat is only determined by the number of valid votes cast relative to the number of available seats in parliament.⁴ As such, one of the characteristic features of the Dutch system is its high proportionality (Pellikaan et al. 2018; Andeweg et al. 2020). Another defining feature is the openness of the party system, which shows in the high inflow of new political parties and the many different and credible governing alternatives, including new political parties as coalition partners (Mair 2008; Pellikaan et al. 2018).

As a result of this proportionality and openness of the party system, the Dutch electorate is one of the most volatile among Western democracies (Mair 2008). Equally important in explaining this high volatility is the disappearance of pillarization, which structured the Dutch electorate along class and religious cleavages in the past (Aarts and Thomassen 2008; Mair 2008; see next section). While pillarization gave structure and stability to the Dutch party system and electorate until the mid-1960s, traditional cleavages eroded afterward. As a result, electoral volatility increased until the Dutch party and voter space were slowly dominated by an economic left-right dimension (Aarts and Thomassen 2008).⁵ However, this newly found structure was disrupted when Pim Fortuyn's party, the Pim Fortuyn List (LPF), entered the political stage during the 2002/2003 elections.⁶ According to existing literature, two main reasons can be given for this sudden change. First, mainstream parties had moved closer to the center on the traditional left-right dimension, which opened up space for parties on the extremes (Pennings and Keman 2003; Andeweg et al. 2020). Second, Fortuyn was able to politicize socio-cultural issues that had become increasingly salient in the electorate since the mid-1990s. As a result, a second socio-cultural dimension emerged as a relevant dividing factor alongside the traditional left-right dimension (Aarts and Thomassen 2008; Pellikaan et al. 2018). Consequently, electoral volatility rapidly increased, creating instability and unpredictability in the electorate (Aarts and Thomassen 2008; Mair 2008). Even though the success of the LPF was short-lived, lasting opportunities arose for the (populist) radical right (Pellikaan et al. 2018). We return to this in Section 4, where we also elaborate on the consequences for partisan influences.

Today, five dominant party families can be found in the Dutch political space (Andeweg et al. 2020).⁷ First, a Christian party family originated from the religious pillars. Currently, the dominant parties in this family are the Christian Democratic Appeal (CDA) and the Christian Union (CU). Second, a Socialist and Social Democratic party family developed from the socialist pillar. Dominant parties in this family are the Labor Party (PvdA) and the Socialist Party (SP). Third, a Liberal party family consisting of the People’s Party for Freedom and Democracy (VVD) and the Democrats ’66 originated from the liberal pillar. The remaining two party families are not direct descendants of any of the pillars: the Greens, with a dominant party, the GreenLeft (GL), and a Populist Radical Right family, with the Freedom Party (PVV) as the dominant party.⁸ We refer the reader to  Appendix 6.1 for other (smaller) political parties included in the party families. Given its considerable role in the formation of Dutch political parties and party families, as well as in the developments of the political space, we continue with a discussion of pillarization.

2.2 Pillarization and consociational democracy

Pillarization means that society and societal life are structured according to a few highly organized subcultures or pillars. In the high-time of pillarization in the Netherlands before the mid-1960s, one could identify four pillars: the protestant,⁹ catholic, socialist, and liberal pillars. Within each pillar, specific institutions and organizations associated with the pillar cater to the needs of their constituents. These institutions and organizations comprise political parties, trade unions, employers’ organizations, schools, healthcare, broadcasting, universities, newspapers, etc. (Andeweg et al. 2020; Figure 2.2). While strict social segmentation means that there is little interaction and cooperation between the people that constitute the pillars on the mass level, there is a strong tendency for cooperation among the elites of the pillars. For instance, as an absolute majority in parliament is generally not achieved (see Section 2.3), political leaders need to forge coalitions and consensus in the political arena. This form of political governance is called consociational democracy, and pillarization is one of its drivers (Andeweg et al. 2020). It is, therefore, not surprising that the highly pillarized period between 1945 and the elections of 1967 is referred to as the ‘heyday of consociational democracy and of political stability’ (Andeweg et al. 2020, p. 40). What followed were periods of increased political competition and polarization. Yet, consociational traditions and principles seem to persist, exemplified perhaps by the formation of the ‘purple’ governments (a coalition between socialist and liberal parties) between 1994 and 2002.

Following Lijphart (1975), Andeweg et al. (2020, p. 35) identify several governance principles that coincide with (are required by) consociational democracy: executive power sharing or grand coalition, a high degree of autonomy of the pillars, proportionality, and a minority veto. Regarding proportionality, for instance, distributional questions are often solved by awarding each pillar its fair share, contributing to relative political stability despite high social division. Yet, while the practice of proportionalism can be successful when, for example, appointing higher civil servants (see Section 3.1), one cannot easily solve questions regarding abortion, decolonization, the energy transition, or national identity by awarding pillars their fair share. This can lead to tensions between the pillarized groups and political instability even in countries like the Netherlands that are perceived to be very consensus-oriented.

2.3 Governance and policymaking

Against this background, we discuss how are policy decisions taken in the Netherlands and whether the Dutch governance structure is conducive to partisan behavior in policymaking. The Netherlands is a unitary state with a monarch as the Head of State. Given the more representative role of the monarch, the central government is responsible for determining and steering the policies in the country. This could provide space for partisan ideology to influence policymaking. Yet, policymaking in the Netherlands is functionally decentralized with a high degree of corporatism, a segmented bureaucracy, and an active role for courts as policymakers (see Andeweg et al. 2020). This governance structure, which stems from pillarization and consociational democracy, speaks against a high degree of politicization in Dutch policymaking. Moreover, local governments (ie, provinces and municipalities) have less discretion than comparable units in, say, Germany or the USA, despite having their own councils. Local policymaking also shows low levels of politicization as local governments are perceived as local administrations rather than places for political competition (Andeweg et al. 2020).

On the other hand, the Netherlands is a parliamentary democracy with changing political majorities, giving room for partisan influences on policymaking. The Dutch parliamentary system consists of two chambers: the First Chamber (Eerste Kamer, ‘Senate’), whose members are elected by the members of the provincial councils, and the Second Chamber (Tweede Kamer, ‘House of Representatives’), whose members are directly elected every four years by the electorate. The Second Chamber deals with day-to-day political decisions and is responsible for drafting and amending legislative proposals, while the First Chamber reviews bills as a last instance. In this paper, we refer to the Second Chamber as ‘the parliament’. Given that the political majorities in the Second Chamber are decisive for the formation of the government, there may be room for partisan influence on policymaking in the Netherlands. However, governments in the Netherlands are coalition governments, and historically, no single party has ever controlled a majority of the seats in the parliament. This short review highlights that there is potentially limited scope for partisan influence in Dutch politics—especially compared to other Western democracies.

2.4 Partisan ideology of Dutch governments

As parliamentary majorities change, also governments display different degrees of ideological identity. In this section, we document the ideological identity of Dutch governments over time.

As mentioned, in the Netherlands, no single party has ever occupied more than 50% of the seats in the second chamber. So, all governments are coalition governments. Examples of this are the two so-called ‘purple’ governments under Prime Minister Wim Kok (from the Labour Party; PvdA) between 1994 and 2002, which were composed of the Labor Party (PvdA; social democrats; left), the Peoples Party for Freedom and Democracy (VVD; conservative liberals; right) and Democrats 66 (D66; social liberals; center).

We illustrate the political ideology of Dutch governments in Figure 1. The data come from the Comparative Political Data Set (CPD) by Armingeon et al. (2023), which comprises annual political and institutional data for 36 countries (1960–2021). Using data for Dutch governments, Figure 1 shows the share of cabinet posts held by right, center, and left parties per year (weighted by the number of days in office in a given year). Left parties denote social democratic parties or parties to the left of them. Center parties denote Christian democratic parties or Catholic parties. Right parties denote liberal and conservative parties ( Appendix 6.1; Armingeon et al. 2023; Appendix A).¹⁰

In the Netherlands, the government cabinet comprises ministers and ministers without a portfolio. The latter are ministers without their own ministry but who are assigned specific tasks that fall within the domain of one or more ministries (for example, Minister of Development Cooperation within the Ministry of Foreign Affairs). Together, these ministers form the Council of Ministers, which determines the policies of the government. The share of cabinet posts in the Armingeon et al. (2023) dataset, which we use as an indicator of government ideology, is based on the Council of Ministers. For instance, the second Rutte cabinet (November 2012 to October 2017; coalition between VVD (conservative liberals) and PvdA (social democrats)) comprises eleven ministers and two ministers without portfolio, of which seven were right-wing (53.85%) and six were left-wing (46.15%).¹¹

In addition to the shares of cabinet posts, Figure 1 indicates each government’s ideological ‘color’ along the x-axis. For this, we use the so-called Schmidt index, included in the CPD. Accordingly, periods marked in dark gray denote a ‘hegemony of right/center parties’, which means that 100% of cabinet posts are held by right and/or center parties. Medium gray periods denote ‘dominance of right/center parties’ when more than two-thirds but less than 100% of cabinet posts are held by non-leftist parties. Light gray periods denote a ‘balance of left and right parties’ when both left and center/right parties hold cabinet post shares between one-third and two-thirds each. In principle, the Schmidt index also includes ‘dominance of left parties’ and ‘hegemony of left parties’. Yet, no Dutch government has ever reached the critical point of more than 60% of cabinet post shares of leftist parties.¹² An implication of this lack of fluctuation between right and genuinely left governments is that it may become harder to observe partisan effects in the Netherlands.

Moreover, it should be noted that, as a summary of underlying cabinet post shares, the Schmidt classification into hegemony, dominance and balanced governments necessarily omits information. For example, since the end of 2010, when the Dutch governments were led by a right-wing party (VVD) under Mark Rutte, 5 years have been marked as balanced between left and right. However, many would probably describe politics as largely right-wing. Similarly, the period of the purple governments under Wim Kok (1994–2002) saw some social democratic policies alongside strict consolidation policies typically associated with center/right governments (see Section 3.2). Hence, it is unclear which flavor of government ideology prevailed during these ‘balanced’ periods. One should, therefore, not mistake the ‘balanced’ periods as periods with a ‘social democratic’ government ideology.¹³

3. Effects of partisan politics in NL

3.1 Previous findings

Only a few empirical studies investigate the effects of partisan politics in the Netherlands. Recall that almost all policy fields in the Netherlands are serviced by the central government. This limits the role of local governments in the provinces and municipalities in determining socio-economic outcomes, leading to little variation between these administrative units being exploited by researchers. The few existing studies focus on local property taxes (Allers et al. 2001; Allers and Rienks 2024), one of the few policy areas that are under the discretion of local governments, and the appointment of civil servants (Ennser-Jedenastik 2016). To our knowledge, there are no time-series studies (such as Erlandsson 2004; Potrafke 2012; Blinder and Watson 2016; and Voia and Ferris 2024) that investigate the effect of government ideology on macroeconomic variables, public finance, or other policy areas in the Netherlands.

3.1.1 Local property taxes

Studying the effect of local government ideology on outcomes in Dutch municipalities, Allers et al. (2001) estimate how the ideological composition of municipality councils affects the local tax burden in the Netherlands. The expectation is that municipalities governed by left-wing parties have higher property taxes than municipalities governed by right-wing parties. The authors use the property tax per capita in 1996 in 593 Dutch municipalities as the dependent variable. As independent variables, they use the political composition of the council (0 = governed by right-wing parties; 1 = governed by left-wing parties) and the difference between the political composition of the executive board and that of the council. After including a number of control variables, they find that municipalities with a higher share of council seats occupied by left-wing politicians impose higher property taxes. This finding is according to expectations and suggests evidence for partisan influence in Dutch local policymaking.

More recently, however, a follow-up study by Allers and Rienks (2024) addresses some of the shortcomings of Allers et al. (2001). This study shows little empirical support for partisan influence on local property taxes in the Netherlands. In particular, Allers and Rienks (2024) estimate the same effect of council composition on property taxes in a panel of Dutch municipalities in 1998–2021. Their study design includes time-varying municipality-level controls, as well as, municipality-fixed effects and period-fixed effects. Their results show no robust significant effects of individual parties on property tax revenues and rates. The effects that are significant are small: for example, a one percentage point increase in the seat share of social liberals leads to a 0.6 Euro decrease in property tax revenue. Although Allers and Rienks (2024) do not report results excluding the municipality-fixed effects, it is likely that their inconclusive results, compared to the strong positive effects in Allers et al. (2001), are driven by the inclusion of these fixed effects. In additional analyses, using indicators for left and right majorities in the council¹⁴ instead of individual party shares, Allers and Rienks (2024) find results contrary to expectations. For instance, they find a (marginally) significant positive effect of having a right-wing majority in the council on property tax revenue. As mentioned above, one would rather expect that left governments have a positive effect on property tax revenue than right ones. So, overall, the authors find no robust evidence for partisan effects on the local level in the Netherlands.

3.1.2 Appointment patterns of civil servants

A second area of partisan influence on Dutch public policy examined in the literature is the appointment of senior civil servants or secretaries-general (Ennser-Jedenastik 2016). Secretaries-general sit at the top of each ministry and function as the link between the minister and the ministerial bureaucracy. They are appointed by the cabinet following the proposal by the respective departmental minister and the Minister of the Interior. Secretaries-general often carry party affiliation. The data presented by Ennser-Jedenastik (2016) suggests that around 70% of all secretaries-general have a party affiliation—in the late 1980s and early 1990s, this number was even as high as 100%.

Ennser-Jedenastik (2016) investigates four different motivations for a partisan appointment: (1) ideological agreement, where the party affiliation of the secretaries-general is the same as that of the minister; (2) coalition governance, where the coalition government appoints secretaries-general to a ministry whose party affiliation is different than the minister’s to maintain checks and balances; (3) party family issue priorities, where secretaries-generals are assigned to ministries depending on the perceived policy-field competences of their party; (4) consociational representation, where secretaries-general are appointed such that their party affiliation is in proportion to the parties in parliament. The empirical results lend support to (1) ideological agreement and (4) consociational representation as motives for partisan appointment patterns of Dutch senior civil servants. Overall, this suggests that partisan appointment plays a role in the Dutch elite bureaucracy. However, as Ennser-Jedenastik (2016) notes, it remains unclear whether such party-dependent appointments have an effect on policymaking.

Given the lack of empirical research on the effects of partisan politics in the Netherlands, we now present a simple empirical analysis of the relationship between government ideology and typical country-level partisan politics variables in the Dutch context.

3.2 Empirical analysis

3.2.1 Macroeconomic variables: GDP, unemployment, inflation

According to partisan theory, governments implement policies that serve the economic purposes of their constituency (for example, Hibbs 1977; Alesina 1987; Potrafke 2017). Hence, left-wing governments, whose constituents are assumed to be largely dependent on labor income rather than capital income, would favor expansionary policies that boost production and increase employment in the country. In contrast, right-wing governments, whose constituents are assumed to hold capital and do not depend as much on labor income, would favor policies that keep inflation low.

To gauge whether there is evidence for such partisan policymaking in the Netherlands, we plot GDP per capita growth, inflation, and the unemployment rate between 1960 and 2021 alongside government ideology in Figure 2. The shaded bars in the background indicate government ideology: the darkest shade indicates a period of ‘hegemony of right/center parties’; the medium shade indicates a period of ‘dominance of right/center parties’; and the lightest shade indicates periods of ‘balance of left and right parties’ (following Armingeon et al. 2023). These data on government ideology are taken from the Comparative Political Dataset 1960–2021, which also contains GDP growth, inflation (both from OECD), and the unemployment rate (from AMECO).

Figure 2 shows the development of the three variables in a small open economy that strongly depends on worldwide developments. GDP growth clearly dipped during worldwide recessions in the mid-1970s and early 1980s, as well as more recent ones associated with the 2008/2009 financial crisis and the COVID-19 pandemic. A striking and more ‘domestic’ feature is the spike in unemployment in 1981/1982, which can be traced back to the overheating economy in the 1960s, leading to rapidly increasing wages throughout the 1970s. The steep economic downturn in 1981/1982 forced firms to lay off workers as they could no longer afford the high wages (for details, see, for example, Bakker 1999). However, in the following 20 years, governments (of all colors) managed to get the unemployment rate below 5% again.

As a small open economy, one may expect that worldwide economic conditions strongly influence the Dutch economy and that domestic policymaking may not be as consequential as in other, bigger countries. When combined with the observation that government color does not vary much over the period we observe, one may expect only a limited effect of domestic government ideology on GDP growth, inflation, and unemployment. In line with this, the graphs in Figure 2 neither suggest an obvious link between more left-wing governments and low unemployment nor more right-wing governments and low inflation, as suggested by partisan theory. To exemplify, although the unemployment rate decreased substantially in the period of higher left-wing party influence between 1990 and 2002, there were also periods where the same balance of left and right parties in the government coincided with increasing unemployment.

To further investigate, we run a set of multivariate regressions ( Appendix 6.2) in which we regress the three macroeconomic variables on shares of cabinet posts of left-wing (Table A2a) and right-wing (Table A2b) parties while controlling for domestic and international economic conditions and the lag of the respective dependent variable. Each year between 1960 and 2021 represents one observation. Coefficients can be interpreted as contemporaneous conditional correlations.

Contrary to observations from Figure 2, which did not suggest partisan effects, results regarding the influence of left-wing parties on GDP growth and inflation are in line with the predictions of partisan theory. Both coefficients are positive and statistically significant at the 5% level, meaning that a higher share of left-wing parties is associated with higher (short-term) production and increasing inflation. For example, one more left-wing minister in a 13-minister cabinet, such as the Rutte II cabinet (see Section 2.4), would be associated with a 7.7 × 0.014 ≈ 0.11 percentage point increase in GDP growth and a 7.7 × 0.015 ≈ 0.12 percentage point increase in inflation. In comparison, a one percentage point increase in OECD GDP growth is associated with a 0.73 percentage point increase in Dutch GDP. Finally, a higher share of left-wing parties is not significantly related to unemployment. We do not find significant associations for the influence of right-wing parties. All in all, these results suggest some partisan influence on typical macroeconomic indicators, but only when considering the influence of left-wing parties.

3.2.2 Public finances

The next topic of our analysis concerns public finances. Following partisan theory, left-wing governments are more likely to engage in deficit spending, increasing public debt and budget deficits than right-wing governments (for example, Alesina et al. 1993; Cusack 1999; Potrafke 2017; Volkerink and De Haan 2001). To gauge whether there are discernible effects of partisan influence on public finances in the Netherlands, we plot government revenue, expenditure, gross debt, and primary balance as percent of GDP alongside government ideology in Figure 3. The data on public finances come from the IMF’s Public Finances in Modern History database (Mauro et al. 2015).

In line with empirical evidence for OECD countries (see Potrafke 2017 and references therein), there is no clear indication that more left-oriented governments in the Netherlands increase public debt or run negative primary balances compared to right-wing governments. In fact, in the 1973–1978 period, during the ‘Purple Governments’ 1990–2002, and in the 2013–2017 period, public debt seems to have fallen when more left-right-balanced governments were in power. There is also no clear indication that the influence of left-wing parties in the balanced governments led to excessive government expenditure. In fact, toward the end of the aforementioned periods, expenditure fell below revenue, leading to positive public balances.

As above, we supplement this visual analysis by running several multivariate regressions. This time, we use the four indicators of public finances as the dependent variable, see Tables A3a and A3b in  Appendix 6.2. We control for the domestic economic climate (GDP growth, inflation, and unemployment) and important demographic indicators (total population, urbanization rate, and age dependency ratio).

The results show that, in line with partisan theory, the share of left-wing government posts is positively related to gross government debt (significant at the 1% level). However, not in line with partisan theory, it relates positively to the government balance (sig. at 10% level) and negatively to government expenditure (sig. at 5% level). There are no significant effects of the influence of left-wing parties on revenue. Having one more left-wing minister in a 13-minister cabinet is associated with a 7.7×0.15 ≈ 1.16 percentage point increase in government debt (as a share of GDP). To place this in context, government debt ranges between 35% and 77%. The share of right-wing cabinet posts is not significantly related to the four public finance indicators. So, while we find some evidence for a typical partisan influence on the public debt (again, only for the increasing influence of left-wing parties), most of the conditional correlations are not in line with typical partisan predictions.

3.2.3 Taxation

Our next field of interest concerns taxation or how governments generate their revenues. Here, we focus on tax rates rather than tax revenues, as governments have less influence on the latter. Following the same partisan logic we used before, left-wing parties should favor higher taxation of capital and corporate income and lower taxation of labor income because their constituents depend more on labor income than on capital income (for example, Hibbs 1977; Alesina 1987; Osterloh and Debus 2012; Potrafke 2017). Moreover, as the constituencies of left-wing parties are often from the lower end of the income distribution, they should prefer lower consumption taxes, which are usually regressive (for example, Beramendi and Rueda 2007; Angelopoulos et al. 2012).

To inspect whether such partisan patterns exist in the Dutch tax system, we plot effective tax rates together with the color of the government (Figure 4). The data on effective tax rates is taken from Kostarakos and Varthalitis (2020), who compiled a database of effective tax rates on labor, capital, and corporate income and consumption for 31 OECD countries between 1970 and 2017. First, comparing labor income and capital income taxes, we note that labor income taxes are higher than capital income taxes in the entire period. Both increase by about 8 or 9 percentage points between 1970 and the early 1990s. These hikes seem independent of government ideology. Then, during the two purple governments (1994–2002), we see reductions in labor and capital income taxes, where taxes on labor income decreased slightly more than those on capital income. Yet, while the reduction of capital tax rates continues until about 2012 through several governments (mostly right-wing and center-right), labor income taxes increase again from 2006 onward. For instance, the difference between the two tax rates is 7.1 percentage points in 2001, while it is 18.4 percentage points in 2012. During the period of balanced governments since 2012, this difference has reduced again to 9.8 percentage points as a consequence of sharp increases in capital income taxes. Taxes on consumption have been on a slight upward trend since 1970, with no obvious pattern of partisan influence. Yet, we note that corporate taxes increased during the purple governments in the 1990s. The latter increase has consequently been reversed again during the following right-wing and center-right governments. Similar to taxes on capital income, the centrist (or balanced) governments from 2012 onwards have increased taxes on corporate income.

Overall, this visual inspection does not indicate much partisan influence on effective tax rates until the mid-1990s. However, there are more signs of the influence of government ideology during the period that followed, which started with the purple governments (a coalition between social democrats, social liberals, and conservative liberals).

As for earlier topics, we supplement this graphical analysis with multivariate regressions. In these regressions, we use the above five effective tax rates as the dependent variable and the share of left-wing (right-wing) party cabinet posts as the main independent variable. We add Dutch GDP growth, total population, the urbanization rate, and the age dependency ratio as control variables. On the one hand, and in contrast to expectations based on partisan theory, we observe a positive and significant association between the share of cabinet posts held by left-wing parties and effective taxes on labor income. On the other hand, there is a positive association between left-wing parties’ influence and capital income tax. This would be in line with partisan effects. All other coefficients are statistically insignificant, including those on the share of right-wing parties.

3.2.4 Composition of government expenditure

Finally, we study how Dutch governments distribute their expenditure over various government functions. To investigate, we plot government expenditure as a share of GDP in a selection of policy areas alongside the familiar data on government ideology in Figure 5.¹⁵ The data on expenditure come from Statistics Netherlands and has been available since 1995. Social protection is perhaps the most obvious area of government expenditure that can be influenced by partisan politics. Following partisan logic, social expenditure is expected to be higher under more left-wing governments because left-wing parties favor redistribution from the rich to the poor (for example, Kittel and Obinger 2003; Herwartz and Theilen, 2017; Potrafke 2017). As Figure 5 shows, social protection is the largest item on the government expenditure sheet in the Netherlands, with an average of 16.1% of GDP. It has come down from nearly 19% since reforms under the purple governments in the mid-1990s. The graph does not suggest an obvious partisan effect on Dutch social protection expenditure.

Expenditure on the health sector is usually not a divisive partisan politics issue. Parties from both ends of the spectrum would have an interest to increase spending in this sector (for example, Potrafke 2010; Jensen 2012). For example, left-wing parties seek to expand health care coverage to the poor, making it accessible to as many as possible, while right-wing parties cater to the needs of the aging population, which tends to vote more conservatively. Both strategies entail higher health-related expenditure. Health expenditures were stable until 2006, after which the Dutch health insurance system was reformed as a response to the unpopularity of the system implemented by the Purple Governments in the 1990s. The aim was to introduce more competition into the system and to make a choice of insurance plans available to all insurance takers (Kroneman et al. 2016). While executed by private insurance firms who compete for clients, the system is deemed very social as it is compulsory for everyone. The increase in health expenditure from 5.4% to 6.5% of GDP in 2006 is likely attributable to this reform. Yet, the increase in 2009 is likely attributable to the global financial crisis and the corresponding drop in GDP. 2012 saw further privatization in the hospital sector of the health system.

Spending on education is associated with partisan politics as left-wing parties favor public schooling and right-wing parties favor more private alternatives (for example, Busemeyer 2009; Potrafke 2017). So, periods with a stronger involvement of left-wing parties should see increased public expenditure on education. Figure 5 does not suggest such a relationship. Spending on public services (for example, Potrafke 2011) has been declining since the mid-1990s, which does not suggest a strong association between left-wing parties and a big state. Furthermore, although spending on economic affairs (for example, Potrafke 2011) seems to be mostly tied to macro developments (financial crisis, COVID-19), there may be some hint that spending decreases during times of high shares of right-wing parties in the government (2002–2007 and 2009–2019).¹⁶ Finally, spending on military defense (for example, Bove et al. 2017), often associated with right-wing politics, has been fairly stable during the period covered by our data, suggesting little partisan influence.

As for the other areas of partisan influence, we extend our graphical analysis by running multivariate regressions. As dependent variables, we use expenditure in each of the categories of government spending. We control for GDP growth, inflation, unemployment, and various demographic factors (total population, female labor force participation, the urbanization rate, and the age dependency ratio). Our results (Tables A5a and A5b in  Appendix 6.2) do not yield statistically significant coefficients for either left-wing or right-wing cabinet post shares, except for a negative correlation between right-wing parties and defense and military spending. This goes against partisan logic. The overall lack of association suggests little partisan influence on the composition of government spending in the Netherlands. It should be noted, however, that the limited amount of data for government spending in categories (available only since 1995) reduces our sample size, making our estimates less precise and excluding periods that are potentially relevant for partisan effects.

When looking at partisan influence on the composition of government spending, it may be instructive to also consider the composition of spending as a share of total spending instead of as a share of GDP. Therefore,  Appendix 6.2.4 includes a figure in which we plot the composition of government expenditure as a share of the total expenditure of the government. This keeps the effect of total spending constant and allows us to investigate whether governments of different color put different emphasis on the allocation of their total budget on different policy areas.

Overall, the trends displayed by the categories of government spending are similar to the ones in Figure 5, except that spending on social protection appears to be more constant (between 34% and 38% of total spending) and that the negative slope of public service expenditure is less steep. In Figure 5, we observed a fairly strong reduction of spending on social protection and public services at the beginning of the observed period. That we do not see the same reduction in  Appendix 6.2.4 may indicate that cuts in total government spending (see Figure 3) were financed by reducing spending on social protection and public services. Despite this, the graph does not suggest a strong influence of partisan ideology on government spending as a share of total expenditure either. We also ran the same regressions as above but using the individual categories of government spending as a share of total expenditure as the dependent variable. Results do not differ qualitatively.¹⁷

It is difficult to draw definitive conclusions about the state of partisan influence in Dutch politics from the descriptive analysis and conditional correlations. For one, our analysis should be supplemented with more sophisticated empirical methods. Yet, some of our observations could be helpful for hypothesis-building in future work. Second, the lack of true left-wing governments in the past 70 years does not allow us to identify periods that could serve as proper counterfactuals for periods dominated by right-wing parties. The periods of left/right-balanced governments cannot pose as counterfactuals as they convey ambiguous signals about the ideological color of the policies implemented in that period. In addition, as balanced governments are coalitions of left and right parties, we may see leftist influence in some policy fields and right-wing influence in others. For instance, the juxtaposition of labor-friendly tax policies in the 1990s and the cutting of social protection expenditure in the same period could be interpreted as an (implicit) agreement between the left and right government parties in their coalition accord.

Given our analyses above, we conclude that there is no evidence for a consistent relationship between government ideology and policymaking in the Netherlands. For each statistically significant association suggesting a relationship in line with traditional partisan theory, for example, for GDP growth and inflation, we find another that is opposite to what one would expect, for example, for the primary balance and labor income taxes.¹⁸

4. Movements of parties and voters

In this section, we present two developments in the political landscape that have played a role in the relevance of partisan politics for Dutch politics: one on the supply side and one on the demand side. As with any supply and demand setting, these developments are endogenous and should be considered as such. Additionally, they are not unique to the Netherlands and can be found to different degrees in other Western democracies too.

4.1 Previous findings

On the supply side, using a ‘top-down’ approach (Evans and De Graaf 2013), we look at how political parties have positioned themselves along the traditional, that is, economic, left-right dimension. However, according to most political scientists (for example, Marks et al. 2006; Albright 2010; Bakker et al. 2012; Bakker and Hobolt 2013; Bornschier 2018), the political space in Western (European) democracies is best represented not by one but by two (or more) dimensions.¹⁹ We, therefore, also introduce a second cultural, or socio-cultural, dimension.²⁰ If the party space is multi-dimensional, it depends on what dimension(s) and to what extent party positions are polarized, as the choices that voters have are restricted by what political parties offer (Evans and De Graaf 2013; Evans and Tilley 2012). In other words, when parties converge on one dimension, voters are less able to convey their preferences along this dimension. When parties at the same time diverge on the other dimension, voting is more likely to be a demonstration of preferences along this polarized dimension. Hence, if political parties do not compete on the economic dimension and, thus, do not offer clear economic alternatives to voters, traditional (economic) interpretations of partisan politics may lose their relevance. Indeed, it has been shown that differences between parties in this area may have stabilized or even decreased, while those on the social dimension have increased (Pennings and Keman 2003; Andeweg et al. 2020; Gethin et al. 2021). Moreover, increased polarization of parties on socio-cultural issues and decreased polarization on economic issues created opportunities for ‘New’ left parties and, even more so, populist radical right parties to gain traction (Bornschier 2010, 2018; Pellikaan et al. 2018; Spies and Franzmann 2011).²¹ While often not clear about their economic policy position (Mudde 2009), these parties have aligned along the social dimension.

On the demand side, using a ‘bottom-up’ approach (Evans and De Graaf, 2013), we look at how changes in the constituents of political parties contribute to political cleavages. Traditionally, the constituents of the left consisted of working class (low income; low education) voters and those of the right of the upper class (high income; high education). The class was, thus, a prominent determinant of the political left-right cleavage. As seen above, the predictions from partisan theories follow this cleavage in that left-wing parties are expected to promote economic policy that benefits their electoral base, ie, laborers, and right-wing parties to a different electoral base, ie, capital owners (for example, Alesina 1987; Hibbs 1977). However, over the last decades, class cleavages seem to have declined as a result of a reduction in traditional industries (Ford and Jennings, 2020) or to have lost relevance as a result of the electorate increasingly voting along the cultural dimension (Van der Waal et al. 2007; Van der Meer et al. 2012). On the one hand, it is suggested that this has led to a more volatile and potentially dealigned electorate (Mair 2008; Ford and Jennings 2020). On the other hand, new socio-structural cleavages have been put forward along which the electorate is aligned. A prominent cleavage is one along educational lines, resulting from the increase in the level of education throughout society (Bornschier 2018; Ford and Jennings 2020).²² The resulting divide saw higher educated voters moving to the (‘New’) left and lower educated moving to the (populist radical) right (Bornschier 2010, 2018; Gethin et al. 2021). Given this evidence of changes in the electoral base of political parties, the predictions from (traditional) partisan theory may not be applicable anymore. Against this background, we discuss trends in party and voter movements in the Netherlands.

4.2 Party movements in The Netherlands

In order to show trends in the supply side of the Dutch political system, we use the Manifesto Project Dataset (MPD; Budge et al. 2001; Klingemann et al. 2006; Lehmann et al. 2023). The MPD is a database created by the Manifesto Project, in which the content of election manifestos of political parties in more than 50 countries from 1945 onwards is analyzed. The MPD therefore allows scholars to study changes in party positions over time. More specifically, the MPD contains information on the fraction of sentences in a manifesto (relative to the total number of coded sentences) that fall into different policy-relevant categories. Combining these categories gives measures of party positions on dimensions of ideology.²³ We rely on a two-dimensional representation of party ideology based on Bakker and Hobolt (2013). The two dimensions capture party ideology on a traditional economic left-right dimension and a second dimension that captures the divide between what the authors refer to as libertarian and authoritarian values. The party scores on the economic left-right dimension are based on issues such as economic incentives, productivity, social services, and nationalization; scores on the libertarian-authoritarian dimension are based on issues like multiculturalism, traditional morality, law and order, and environmental protection. We refer to this dimension as the cultural left-right dimension, where we consider libertarian values to be on the left and authoritarian values on the right. The measures of party positions are created by deducting the sum of left-wing statements from the sum of right-wing statements. Therefore, a higher score indicates a more right-wing position.²⁴

Although the MPD covers a longer time frame, we cover party positions for all elections in the period from 1971 to 2021 for most of the political parties that were active during this period. We impose this restriction in order to make the data on party positions comparable to our voter data, which we will introduce in the next section. Given the fragmentation of the Dutch party landscape (see Section 2.1) and for ease of comparison, we report the positions of party families rather than those of individual parties.²⁵ To show movements in the positions of party families over time, we calculate the average score for the economic and cultural dimensions for every combination of party family and election year. By calculating the difference between the position of the party family and that of the average party, we control for general trends, i.e., movements of all party families in the same direction.

Figure 6 shows the movements of party families on the economic left-right dimension relative to the position of the average party. After a rightward move by the Christian parties and a leftward move of socialist parties during the 1970s, the figure shows a relatively stable divide on economic issues until the start of the 21st century. For the rest of the observation period, Christian parties remain closer to the liberals than the socialists. The arrival of the greens, or the ‘New’ left, did not considerably affect the positioning of the other parties. From the early 2000s onwards, however, there are noteworthy movements. First, during the last two decades, the greens have positioned themselves in between the socialists and the three other party families, taking a slightly more left-wing position than the average voter. Second, the radical right has established itself on the economic right, in some elections even taking a more economically right-wing position than the liberals, most prominently during the two most recent elections in our data (2017 and 2021). Third, the difference between party families has not changed much after the arrival of the radical right. An exception to this is the most recent election in 2021. Between the 2017 and 2021 elections, we see a rightward move on economic policy issues of both the Christian parties and a leftward move of the socialist parties, increasing the divide on economic issues during the 2021 elections. This suggests that for most of the observation period, the ‘traditional’ economic divide between the party families remained relatively stable, also after the 2002/2003 elections. Except for the most recent elections in our sample, differences in economic policy issues between party groups seem not to have become more pronounced over time.

Figure 7 shows party family movements on the cultural left-right dimension relative to the position of the average party. During the seventies, the liberals, socialist, and Christian parties differed little on socio-cultural issues. Yet, from the 80s onwards, we can see a clear rightward move of the Christian parties on this dimension. This created a difference in party positions between them on the right and the liberals and socialists on the left, relatively speaking. A difference that remained relatively stable over the entire period. With the arrival of the greens, an alternative on the cultural left of the liberals and socialists was created. Unlike on the traditional economic dimension, where the radical right's position does not differ much from the liberal parties, the position of the radical right does differ substantially on the cultural dimension, especially from 2006 onwards when the PVV entered the political supply. The rise of the radical right posed a novel, more right-wing alternative to this dimension. This increased the political divide on social policy issues. Two additional changes are noteworthy. First, the elections in 2003 seem to have caused a rightward spike in the position of the liberal parties and, especially, the greens. On closer inspection, this spike is largely due to an increase in sentences coded in the right-wing social category ‘Political Authority’, which covers references in the manifesto that relate to competence to govern and government stability (Lehmann et al. 2023). This is not surprising given the turbulent times during the 2002 and 2003 elections.²⁶ Second, it is important to note that the difference between the liberals and socialists on socio-cultural policy issues are very small and do not differ much from the position of the average party. Both party families have moved in sync throughout the observation period, with the exception of the 2003 election, which suggests that they do not present alternatives to voters on the social left-right dimension.

All in all, these descriptives suggest that the divide, that is the difference between parties, on economic issues has remained relatively stable over time, at least until the 2021 elections. Neither the greens nor the radical right influenced the divide on the economic dimension substantially. This is generally in line with the conclusions of Pennings and Keman (2003) and Andeweg et al. (2020). Things are different in the cultural dimension, where the entrance of the greens and, later (and mostly), the radical right created a divide on socio-cultural policy positions, especially after the 2002/2003 elections. This is in line with the conclusions of Aarts and Thomassen (2008) and Pellikaan et al. (2018). Moreover, the combination of a stable divide on economic issues and the lack of a clear difference on social issues for the socialists and the liberals, specifically after this divide became increasingly salient, suggests that there are no clear policy differences between these traditional left-right party families. In line with, for example, Spies and Franzmann (2011), Bornschier (2018) and Pellikaan et al. (2018), this seems to have created opportunities for the ‘New’ left (that is, the greens) and, especially, the radical right in the Netherlands. At the same time, traditional partisan politics may play little to no role, given that political competition occurs on different policy issues than economic ones. As such, economic policy issues may not be relevant in explaining party choice (Evans and De Graaf 2013).

4.3 Voter movements in The Netherlands

In order to show trends in the electorate over time, we rely on the Dutch Parliamentary Elections Studies (DPES), a voter survey conducted around each election from 1971 onwards.²⁷ Even though each election study covers some election-specific topics, many questions are part of the core variables of the DPES and are therefore asked in each wave. Next to turnout and (self-reported) voting behavior, the DPES core covers variables such as political efficacy, satisfaction with the government, and voters’ views on a set of political issues. It also includes data on the socio-economic background of respondents. As we are interested in changes in political cleavages in the Dutch electorate over time, we use data on self-reported social class, education, income, and voting behavior. Our main sample, therefore, consists of those who voted. Summary statistics are provided in  Appendix 6.5.²⁸

Before using the data to describe the average voter of each party family, we like to note several stylized facts about the entire electorate over time, assuming that the DEPS sample is a good representation of the electorate. Figure 8 shows the ways in which Dutch voters have changed since 1971. First, we see that the share of respondents that are highly educated has increased substantially.²⁹ Second, in line with depillarization, the share of people that follow a religion has decreased over time.³⁰Figure 8 also includes trends in self-reported social class.³¹ We see that while the share of people identifying as the working class has decreased, the share of people who identify as part of the upper class has increased. Fourth, the share of women in our sample and the share of people living in urban areas has remained relatively stable.³² We consider these overall changes in the Dutch electorate in our further analysis below (details follow).

We now turn to the characteristics of the average voter of each party family. Given traditional class cleavages and in line with the assumptions underlying traditional partisan politics, we would expect green and socialist voters to be part of the working class and voters of the other party families to be part of the upper class (Lipset and Rokkan 1967; Hibbs 1977; Alesina 1987; Ford and Jennings 2020). In order to examine whether the Dutch electorate (still) follows such traditional class cleavages, we show two figures. The first, Figure 9, is on self-reported social class; the second, Figure 10, is on education. Following Gethin et al. (2021) and Bornschier (2018), we would expect traditional class cleavages to have decreased while a new cleavage on education is arising.

To obtain Figures 9 and 10, we use information on (reported) voting behavior and party families and calculate party family averages for our variables of interest for each election since 1971. This shows how the average voter of each party's family changed over time. There are two potential confounders in this picture. First, the general trends in the Dutch electorate that we discussed above may distort averages per party family. To adjust for these general trends, we calculate the distance between the average voter of a party family and the average voter in our full sample. Second, we also observe that, for example, the socialist and social democratic parties lost vote shares since 2002/2003, while liberal and radical right parties gained.³³ This shift of the vote base from left to right could confound the image if the voters that moved are not a random sample but have a specific socio-economic background. Unfortunately, we are not able to track and reconstruct such selected voter migrations between parties here. The conclusions from our descriptive analysis are, therefore, subject to this caveat.

Regarding Figure 9 on social class, we see that before 1990 a larger share of constituents of the liberals considered themselves upper-class relative to the average in the Dutch electorate, while the share of Christian and socialist upper-class voters was lower than the average. For the liberals and socialists, this follows traditional class cleavages, but given that Christian parties can be considered center-right (see Section 4.1 and  Appendix 6.1), this share would be expected to be higher. Over time, however, the share of upper-class voters in the Christian parties’ electoral base increased to the average. Interestingly, the electoral base of the greens considers themselves more upper class than the average voter. From the mid-2000s, this share decreased toward the average. Moreover, the share of constituents of the radical right that considers themselves part of the upper class (relative to the average voter) is even smaller than constituents of the socialist parties.  Appendix 6.5 presents a figure that shows the share of constituents that identify as working class; similar trends are visible. The share of working class voters among socialist and radical right parties is substantially larger than average, while it is lower than average for voters of the liberals and greens. The constituents of the Christian parties are somewhere in between; but below the average voter. Note that the trends we see for the constituents of the greens and the radical right are not in line with class cleavages; there even seems to be a reversal of this cleavage for these groups. Note also that the constituents of the socialists and liberals still align along class.³⁴

What about an educational cleavage? Figure 10 shows trends over time in the share of highly educated among the constituents of the five-party families. Over time, we see that the share of highly educated socialist voters lies close to the average. The education of the constituents of Christian parties fluctuates a bit more but also lies close to or slightly below the average. When we consider the liberals, greens and radical right, a cleavage on education seems to have been emerging since the 90s, when the green parties entered the political landscape. While the greens have attracted voters with relatively high education (that is, higher than the average) from the start, the constituents of the liberals have become more highly educated over time. Since the arrival of the radical right, this party group has attracted voters with education levels substantially lower than average.³⁵

Combining these developments along class and educational lines, the following picture emerges. On the one hand and in line with Gethin et al. (2021), class cleavages have seen a reversal for the constituents of the greens and the radical right. However, constituents of the socialist and liberal parties largely follow class lines indicating that there is still some role for class-based voting in the Netherlands (Van der Waal et al. 2007; Van der Meer et al. 2012). At the same time and in line with Gethin et al. (2021) and Bornschier (2018), a new cleavage along educational lines has emerged. This new cleavage is specifically pronounced for green (or ‘New’ left) and liberal voters on the one side versus radical right voters on the other side. Our 'bottom-up' descriptive analysis thus indicates that the amount of voters with (only) class-based characteristics on the left and on the right is declining (Evans and De Graaf 2013). Consequently, the predictions from traditional partisan politics may not be applicable to the materialization of new cleavage lines in the Netherlands. This is in agreement with the conclusions from our ‘top-down’ investigation.

5. Conclusion

In this paper, we have reviewed the evidence for two related questions: first, what are the policy effects of partisan politics in the Netherlands, and second, did developments in the Dutch party and voter space influence the relevance of partisan politics in the country? Regarding the first question about partisan effects, we found no previous literature on the national level and only a few studies on the subnational and local levels. These studies show that there is no robust evidence for partisan politics in municipal councils on local property taxes. At the same time, there is some evidence of partisan influence on the appointment of higher public servants.

As there is no previous literature on partisan effects on economic policy at the national level, we presented graphics that juxtapose the ideological ‘color’ of Dutch governments with several typical policy variables, such as GDP per capita, unemployment, inflation, public finances, effective tax rates and areas of public expenditure. We supplemented this with contemporaneous conditional correlations based on multivariate regressions. Although this exercise is not sufficient to identify more than simple associations, these descriptives suggest no consistent role for partisan politics. We find some relations in line with partisan expectations; however, we find as many counterintuitive relations contrasting it. Our interpretation is that (the legacy of) the pillarization of Dutch society and associated consociationalism left little room for consistent politicization of policymaking, at least before the 2002/2003 elections. The following period of increased political competition may have provided more fertile ground for the expression of partisan interests, but not according to the traditional economic left-right dimension (cf Andeweg et al. 2020).

What the Netherlands has in common with other countries is the rise of populist right parties since the turn of the century. This phenomenon is associated with the second question about developments in the party and voter space. Previous literature, not only for the Netherlands but also for other Western democracies, suggests that movements of parties along dimensions of political ideology and simultaneous movements of voters along social and cultural cleavages have freed up a voter base for right-wing populist parties (for example, Spies and Franzmann 2011; Bornschier 2018; Pellikaan et al. 2018). Using the Manifesto Project Database and the Dutch Parliamentary Election Studies, we presented descriptive evidence that aligns with these findings. Perhaps in contrast to other national contexts, however, the established party families in the Netherlands did not so much converge on the traditional economic left-right dimension, but maintained a relatively stable distance from each other.³⁶ Simultaneously, the entries of the greens and radical right parties have led to a new political divide on socio-cultural policy issues. Moreover, regarding voters, we observed that, while the constituencies of some party families (liberal, Christian) remained stable with respect to social class identification, others have seen a reversal (green, social democratic) after the entry of the radical right parties. In line with previous literature (for example, Bornschier 2018; Gethin et al. 2021), we also detected a growing divide among voters along educational lines.

The process of depillarization and resulting increased political competition in the Netherlands could be conducive to observing partisan influence on policymaking. Yet, there may be little to no role for traditional partisan politics, given that political competition has occurred mainly on different policy issues rather than economic ones. We, therefore, call for future research into ‘new’ partisan effects, which considers political divides along the socio-cultural policy issues. We also see potential in researching the determinants and consequences of voter migrations from one party (family) to another, conditional on selection along new and old cleavage lines. Furthermore, the Dutch context examined in this chapter stipulates specific questions. For instance, the study by Ennser-Jedenastik (2016) on civil service appointments raises the follow-up question of whether partisan patterns in these appointments translate into policy effects on the national level. The Netherlands could also be an interesting study ground for how partisan effects depend on coalition agreements between parties that usually have diverging policy objectives. Finally, one could study to which extent (the legacy of) pillarization and consociationalism suppressed the effects of partisan policymaking and how the demise of these two characteristics of the Dutch political system may have fostered partisan politics.

Footnotes

References

6 Appendix

6.1 Overview of Dutch political parties

 

6.2 Multivariate regression analysis of government ideology on partisan variables

6.2.1 Macroeconomic variables

 

6.2.2 Public finances

 

6.2.3 Taxation

 

6.2.4 Composition of government expenditure

 

6.3 Details on Bakker and Hobolt (2013)

 

6.4 Additional statistics and figures: party movements

 

6.5 Additional statistics and figures: voter movements