Legislative Oversight and Control of Independent Portfolios: Government and Opposition Dynamics

Abstract

Research on legislative control dynamics has extensively examined how political parties use legislative tools to control portfolios and their respective heads in coalition governments. However, research has focused on partisan-run portfolios and has overlooked how control dynamics are affected when portfolio heads are independent, thus not affiliated with any party. This article addresses this gap by analysing parliamentary questions from 28 German city councils to determine how independent portfolios are controlled relative to partisan portfolios. The results show that all parties control independent portfolios more intensely than partisan portfolios. This is the case for both governing parties and opposition parties. However, while government parties control independent portfolios more than partisan portfolios, they still do so to a lesser extent than opposition parties.

Legislative oversight of the executive as a check and balance is one of the cornerstones of modern democracies and ensures that the executive branch's actions, decisions and performances are in line with the laws and constitution (Lupia and McCubbins 1994; Martin et al. 2014; Rockman 1984). As such, legislatures have a diverse set of oversight tools at their disposal. Parties use various control instruments such as parliamentary committees and hearings (McGrath 2013), budgetary oversight (Stapenhurst 2008) and parliamentary questions (Martin 2011). Although political science has extensively studied the mechanisms and dynamics by which legislative actors monitor and hold executive actors accountable, researchers have so far studied only cases where all portfolios are held by a minister who is a party member of one of the governing parties (cf. Höhmann and Sieberer 2020; Otjes et al. 2023; Raunio 1996). However, in many legislatures, it is quite common for ministries to be led by independents – for example, in Italy (Verzichelli and Cotta 2018), France (Bruère and Gaxie 2018), Sweden (Bäck and Persson 2018), Central and Eastern European countries, and the Baltic countries (Semenova 2018).

Since independents are not members of a party, they are free from the constraints of party loyalty. Thus, it can be assumed that they have more freedom to do what they believe is right than partisan ministers who are dependent on their party and must participate in the ‘party game’ (Müller 2000: 312). As a result, partisan ministers are more constrained by the additional need to respect their party's internal rules and hierarchies and, thus, less likely to follow their own beliefs if they are not in line with their party.

Moreover, independents are characterized by having special expertise for ministerial work, either in the form of knowledge of a specific policy area or in the form of high-level administrative and technical skills (Costa Pinto et al. 2018). Thus, independent portfolio heads differ in terms of their specific expertise and how parties can pressure a minister, while both hold prominent positions in the government. Consequently, there is a gap in our understanding of how parties exert control over these two distinct types of ministers and their respective portfolios. This research aims to address this gap by empirically examining the question, do parties in parliament control independent ministers and their portfolios differently from the way they control partisan ministers? This article seeks to contribute to a more nuanced understanding of whether the political alignment of a portfolio head affects party control behaviour.

To this end, I analyse a novel dataset of parliamentary questions (PQs) from 28 German city councils in cities with more than 100,000 inhabitants between 2011 and 2020. PQs are one of the most prominent and widely studied legislative control tools used to hold governments accountable, at both the national and subnational levels (Martin 2011; Otjes and Louwerse 2018; Otjes et al. 2023). Moreover, research on intra-coalitional control has shown that PQs are a useful control tool for monitoring individual ministers (Block 2024; Höhmann and Sieberer 2020; Martin and Whitaker 2019). Thus, PQs are also a suitable control tool for monitoring independent portfolios, as parties can monitor independent portfolios by asking specific questions directly to them.

In the empirical analysis, I focus on the German local level where independent portfolio heads are quite common (Ecker and Gross 2023). Choosing the local level allows me to analyse multiple cases from roughly the same period, a similar political culture and a similar institutional setting. The results show that independent portfolios are more tightly controlled: all parties address significantly more PQs to independent portfolios than to partisan-led portfolios. This is the case for both opposition and governing parties. However, even though all parties tend to control independent portfolios more than partisan portfolios, I still find differences between opposition and governing parties in their control behaviour towards independent portfolios. Opposition parties impose significantly more PQs on independent portfolios than governing parties.

Independent portfolio heads: who they are and why they get into office

Independent ministers are common in parliamentary regimes, such as Sweden (Bäck and Persson 2018) or Italy (Verzichelli and Cotta 2018), as well as in presidential regimes, such as Brazil or Mexico (Amorim Neto and Samuels 2010), or mixed regimes, such as France (Bruère and Gaxie 2018) or Portugal (Costa Pinto and de Almeida 2018). In a longitudinal analysis of 14 countries, Maurizio Cotta (2018) finds that the average share of non-partisan ministers among all ministers per country ranges from 16% (France) to 58% (Turkey).

Whether the political regime is a driving factor in explaining whether and how many independents become portfolio holders is an ongoing debate in current research. So far, scholars have found mixed empirical evidence that the political regime plays a central role in explaining the emergence and number of independent ministers. Ottavio Amorim Neto and Kaare Strøm (2006) find that independent ministers are more likely and thus more numerous in mixed regimes than in parliamentary regimes. Ottavio Amorim Neto and David Samuels (2010) compare all three regime types, and their results show that independent ministers are particularly likely in presidential regimes. They also find that independents are more likely in mixed regimes than in parliamentary regimes, but this effect is much smaller than the effect of presidential regimes. Cotta (2018) compares countries with mixed or parliamentary regimes and finds no clear result, as some mixed regime countries have the highest number of independent ministers, and other mixed regime countries have the fewest independent ministers. In between those two are the countries with parliamentary regimes.

Moreover, recently the results of Jean-Benoit Pilet et al. (2023) showed for 31 European countries that the share of populist parties within a cabinet and higher levels of perceived corruption positively affect the likelihood that a technocrat is appointed to the national government.

Why do governments appoint non-partisan rather than partisan portfolio heads? Especially when, from a strict party government perspective (Blondel 1991; Katz 1986; Strøm 2002), it seems much more advantageous for a party to stick with party officers who share the party's programmatic identity and are also skilled election campaigners for the party (Costa Pinto et al. 2018). The literature shows that there are several reasons why parties appoint independent ministers.

Most independent ministers can be considered technocrats due to their high level of education and qualifications – most commonly business executives, high-ranking civil servants and university professors (Rojon et al. 2023). Often, independent ministers have specific expertise in their ministerial portfolio due to their background, but sometimes they are chosen more for their general technical and managerial skills. Thus, compared to partisan ministers, independent ministers can be characterized as high-level experts. The results of Sebastien Rojon et al. (2023) show that citizens tend to prefer such non-traditional political actors over partisan MPs for 15 European countries. Moreover, Sophie Panel et al. (2023) find evidence that independents with an academic background and relevant professional experience have a higher degree of popularity than career politicians. So, overall, citizens seem to prefer independent technocrats to partisan ministers who are experts in the field of their portfolio (also called technopols) (Panel et al. 2023; Vittori et al. 2024).

From a party perspective, independents are chosen not only for their skills but also for several other reasons. (1) In times of crisis, independent ministers are often appointed as an alternative solution for crisis resolution (Alexiadou and Gunaydin 2015; Costa Pinto et al. 2018). (2) Independents are chosen as portfolio heads as a compromise between several parties in the event of a stalemate and are therefore more common in fragmented party systems (Amorim Neto and Samuels 2010; Costa Pinto et al. 2018; Shain et al. 1995). (3) Similarly, independents become the head of a portfolio because, in certain political situations, such as to resolve a coalition dispute, it may be advantageous to appoint someone without party ties (Costa Pinto et al. 2018). For example, to resolve a coalition dispute between several parties that all want a particular portfolio, appointing an independent minister may be a solution that each coalition party can accept without losing face. (4) Another reason why independents become ministers is that the ruling parties want to signal to the electorate that the government values policy expertise and is willing to implement reforms, thus giving portfolios to neutral technocrats (Alexiadou and Gunaydin 2015). (5) Governments may also choose independent ministers to push through unpopular reforms, such as cuts to the welfare state, and shift public blame to the independent minister to avoid high political costs for themselves (Alexiadou and Gunaydin 2015; Costa Pinto et al. 2018). (6) In presidential or mixed regimes, independents are often appointed by the directly elected head of the executive because the head of the executive wants to have reliable ministers around them (Amorim Neto and Samuels 2010; Amorim Neto and Strøm 2006; Martínez-Gallardo and Schleiter 2015).¹

Parliamentary questions as a legislative control instrument

PQs are widely used as a legislative control instrument (Russo and Wiberg 2010). Opposition parties use PQs to monitor the government in various legislatures (Martin 2011; Otjes and Louwerse 2018) – including parliamentary regimes (Russo and Wiberg 2010), mixed regimes (Borghetto et al. 2020; Hayward 2004) and presidential regimes (Mimica et al. 2023). Simon Otjes et al. (2023) analysed 53 Dutch municipalities, and their results show that PQs are used as a control instrument at the local level in the same way as at higher political levels.

However, PQs are not only used as a control tool by opposition parties. Government parties also use PQs to highlight their successes and to emphasize the interests of their constituencies (Zittel et al. 2019). In addition, government parties use PQs as an intra-coalition control tool to obtain information from government ministries under the control of a coalition partner (Block 2024; Höhmann and Sieberer 2020; Martin and Whitaker 2019). In intra-coalition control, PQs are suitable to bridge the information gap between the principal (coalition government) and the agent (individual minister) (Strøm et al. 2010).

The fact that PQs are an appropriate tool to address a specific portfolio makes them predestined to be used to monitor independent ministers and their portfolios. Furthermore, according to Shane Martin and Richard Whitaker (2019), written PQs must be answered by ministers, are resource-efficient for parties, can be used to extract specific information, and ministers are sensitive to the content of PQs.

How do parties use parliamentary questions to control independent portfolios compared to partisan portfolios?

From a principal–agent perspective, conflict arises due to differing interests and asymmetric information between agents and principals (Laffont and Martimort 2002; Lane 2008). When an agent possesses more information than the principal and exploits this advantage, conflicts can escalate as the agent pursues their interests. This pursuit of personal interest by the agent, when in contrast to the principal's preferences, results in an agency loss for the principal. However, the principal can monitor the agent, addressing the informational imbalances and reducing the potential for agency loss.

For partisan ministers, research on intra-coalition control has shown that parties tend to exert control, especially over ministers from coalition partners (Martin and Whitaker 2019). Adopting a principal–agent perspective (Müller 2000), they do so to mitigate the risk of an agent (a minister) deviating from their principal's (the cabinet) preferences and the coalition agreement to pursue their policy preferences (Laver and Shepsle 1990, 1994; Strøm et al. 2010; Thies 2001).

In contrast to partisan ministers, who are expected to adhere to the party line (Müller 2000), independent ministers have greater autonomy because they are not under the constraint of party loyalty (Rojon et al. 2023). Consequently, they may be more inclined to follow their agenda rather than strictly adhere to the government parties' manifesto or coalition agreement, confronting parties with a heightened risk of potential agency loss. Thus, it is highly likely that parties control independent portfolio holders differently compared to partisan ones.

In the following, I will first carve out in more detail why independent portfolios may be controlled differently from partisan portfolios and why both governing as well as opposition parties are more likely to control independent portfolios more closely. Subsequently, I will delve into possible explanations for why ruling parties may exercise less control over independent ministers than opposition parties do.

Previous research has identified several reasons why governing parties give portfolios to independents, even though this is counterintuitive from the perspective of party government because the government is appointing a minister who is not constrained by party discipline. Since independent portfolio holders are often chosen for their specialized expertise, either in specific policy areas or high-level administrative and technical skills (Costa Pinto et al. 2018), this expertise makes them valuable resources for their respective ministries and enables them to make informed decisions. However, the greater freedom to make decisions based on their expertise rather than party ideology makes independent ministers less predictable and more prone to agency drift (Laffont and Martimort 2002; Lane 2008). As a result, parties in both government and opposition may exercise greater control to ensure that the independent minister's actions remain within acceptable bounds with their respective party goals.

Additionally, independent ministers are often seen as neutral technocrats, chosen to bring professionalism and an impartial perspective to government (Alexiadou and Gunaydin 2015). While this neutrality can be an asset, it can also raise concerns for both opposition and governing parties about the independent's alignment with their respective party goals. Moreover, independent ministers are often selected to increase transparency and accountability in government (Alexiadou and Gunaydin 2015; Costa Pinto et al. 2018). Given their non-partisan status, their decisions may be subject to greater public scrutiny to ensure that they are acting in the best interests of the public rather than serving any hidden agendas. Parties, especially those in opposition, would likely capitalize on this public expectation by increasing their control efforts.

Moreover, information asymmetries may be an additional reason why parties may control independent ministers more closely. As stated in the literature, control can be used to close an information gap (Höhmann and Sieberer 2020; Martin and Whitaker 2019) and can thus be seen in these cases as a mechanism to obtain needed information. This mechanism is relevant for both opposition and government parties, albeit for different reasons. On the one hand, government parties can use oversight tools to obtain information without having to openly rebuke an independent minister and risk signalling cabinet disunity to the public. On the other hand, opposition parties can use oversight tools to gain information because they have less information and fewer clues about where an independent minister stands compared to a partisan minister since an independent does not have to toe the party line of the governing parties or a possible coalition agreement compared to a partisan minister.

Lastly, opposition parties in particular, but also to a lesser extent governing parties, have an incentive to monitor independent portfolios more closely, since signalling to an independent minister not only that they are being watched but also suggesting specific actions or policy positions may influence an independent minister to act more according to the respective party's interests.

In summary, I expect that all parties will control independent ministers more closely than partisan ones due to information asymmetries, the fact that an independent is not subject to party discipline, the high level of expertise of an independent and because independents are seen as neutral experts. Since parties need to ensure that a minister's actions remain within an acceptable range for them, they should therefore control independent portfolios more intensively than partisan ones. Thus, my first hypothesis is:

Hypothesis 1: Independent portfolios are subject to more intense control by all parties than partisan portfolios.

While both government and opposition parties have reasons to monitor independent portfolios more closely than partisan portfolios, there are also reasons why government and opposition parties may differ in the extent to which they monitor independent portfolios relative to each other. I first outline the reasons why government parties may be less inclined than opposition parties to control independent ministers and then focus on the reasons why opposition parties may be more inclined than government parties to control independent portfolios. Thus, the commonly found control pattern that opposition parties control more than government parties (Martin et al. 2014; Strøm et al. 2010) should also be found for independent portfolios when comparing the control behaviour of government and opposition specifically for independent portfolios.

Under a majority government, the governing parties decide who becomes the head of a portfolio (Bäck et al. 2009b; Cutler et al. 2016; Ecker et al. 2015), and the opposition parties do not. Thus, when governing parties decide to give a portfolio to an independent, they choose a person they trust at least to some degree and with whom they share a general vision of what the portfolio should do and on which policies it should focus (Kaltenegger and Ennser-Jedenastik 2022). Thus, independent ministers may be more closely aligned with government priorities and policies than opposition parties. In addition, government parties are more likely to have established working relationships and a degree of mutual trust with independent ministers than opposition parties. This trust may lead to a perception that independent ministers are aligned with the goals of the governing parties, reducing the perceived need for intense oversight by the governing parties. In addition, independent ministers are part of the government, and government parties may be less inclined to control them too aggressively, in order to avoid public disputes or controversies with independent ministers. Instead, they may prefer private discussions with the independent ministers rather than publicly addressing PQs to them which could escalate tensions and hinder smooth cooperation between portfolios.

Opposition parties, on the other hand, face multiple reasons why they should monitor independents very closely. For opposition parties, it is less clear than for governing parties what an independent minister stands for and what arrangements exist between the government parties and the independent minister. This information disadvantage is particularly important because an independent minister is not bound by the policy goals outlined in the governing parties' manifestos or by the coalition contract between the coalition parties, and thus, the opposition parties cannot be sure that the independent minister has the same agenda as a partisan minister from the governing parties would have. Thus, opposition parties are at an information disadvantage compared to governing parties and should, therefore, monitor independents more closely to close this information gap.

Since opposition parties need to differentiate themselves from the government in the eyes of the electorate (Strøm 1990; Tuttnauer and Wegmann 2022), actively monitoring independent portfolios can be used to demonstrate to voters their commitment to holding the government accountable and their diligence in addressing public concerns.

The appointment of independent ministers can sometimes be a response to crises or conflicts within a coalition government (Costa Pinto et al. 2018). Opposition parties may view these independent ministers as potential weak links in the government that could be exploited to expose government failures and inconsistencies.

In summary, government parties should control independent portfolios to a lesser extent than opposition parties do because government parties appoint independent ministers, are closer to independent ministers than opposition parties, trust independents more and are more inclined to avoid public disputes due to too vigorous monitoring. On the other hand, opposition parties should be more inclined to control independent portfolios than government parties to distinguish themselves from the government and to signal to voters that they are addressing public concerns and pointing out government weaknesses and inconsistencies. They should also be more likely to control independent ministers because they cannot be sure of the arrangements between the governing parties and an independent minister. Thus, my second hypothesis is:

Hypothesis 2: Parties with governmental responsibilities will control independent portfolios to a lesser extent than opposition parties will.

Data and methods

To empirically test my hypotheses and to analyse how portfolios headed by an independent are controlled, I focus on PQs from local parliaments – also called city councils – in major cities in Germany with more than 100,000 inhabitants. Focusing on PQs is a sensible choice because the literature review showed that PQs are an appropriate control tool to address portfolio holders directly, to close the information gap between the principal and the agent, and to mitigate the risk of an agent with divergent preferences pursuing their interest, leading to an agency loss for the principal. Since previous research has shown that independents are typical for mixed regimes (Cotta 2018), I chose to focus on the German local level, where independent portfolio heads are quite common (cf. Ecker and Gross 2023). In mixed regimes, the governing responsibility is shared between a directly elected head of the executive, such as a president or mayor, and a parliamentary majority, often in the form of a coalition cabinet. The mixed regime at the local level in Germany consists of a directly elected mayor – the head of the executive – and a coalition cabinet supported by the majority of the directly elected parliamentary actors in the form of the legislative councillors (Gross and Debus 2018).

I chose the German local level for multiple reasons. First, the case selection allows me to compare multiple local parliaments influenced by the same political cultures, under similar institutional settings and during nearly the same period. Second, PQs are an available legislative tool in all councils, and PQs are commonly used by local parties in Germany. Third, independent portfolio heads are common at the local German level (Ecker and Gross 2023). Fourth, analysing German major cities has the advantage that the city councils are highly professional and work similarly to the parliaments at higher levels (Egner 2015; Gross 2023). Furthermore, although local parties in Germany are integrated into national party organizations, they possess the independence to draft their own election manifestos (Gross and Jankowski 2020) and directly impact the policy outcomes in their respective cities through decision-making (Goerres and Tepe 2013).

At the German local level, ministries are called departments, and the head of a department is a deputy mayor. Department heads have to be elected by the local council and thus rely on the support of the parliamentary majority (Bogumil and Holtkamp 2013; Ecker and Gross 2023; Egner et al. 2013). Despite the legal requirement for public advertisement, portfolio allocation in German major cities is based on partisan considerations and follows a government–opposition logic (Ecker and Gross 2023). The results of Alejandro Ecker and Martin Gross (2023) show for 77 German cities that the distribution of seats per party in the local council influences the selection of department heads. Thus, parties tend to choose allocations that reflect their electoral strength. Furthermore, coalition agreements have a significant impact on the allocation decision. Coalition parties are much more likely to receive portfolio allocations (Ecker and Gross 2023). Overall, portfolio allocation at the local level in Germany follows a pattern similar to portfolio allocation at higher political levels, prioritizing proportionality and following a partisan logic.

I use an innovative dataset from the German local level, encompassing all PQs raised during one legislative period within German city councils of cities with a population exceeding 100,000 inhabitants between 2011 and 2020 (for a detailed description, see Martin Gross et al. (2024)). PQs were collected using web scraping and underwent manual editing as required. The dataset includes the dominant policy issue for each PQ coded based on the Comparative Agenda Project (CAP) coding scheme (Baumgartner et al. 2018; Bevan 2019). The labelling process was conducted using a semi-automated supervised classification approach.² I use a subset of the PQ dataset of 28 cities consisting of all councils with at least one independent portfolio (see Supplementary Material, Appendix A for details). I focus on all major parties of the German party system. The six major parties are: the Alternative for Germany (AfD), the Christian Democratic Union together with the Christian Social Union (CDU/CSU), the Free Democratic Party (FDP), the Greens (Bündnis 90/Die Grünen), the Social Democratic Party (SPD) and the Left (Die Linke). The dataset consists of 13,670 written parliamentary questions.³

Furthermore, I created a portfolio dataset consisting of all portfolios that were in place during the legislative period of the 28 cities covered in the PQ dataset. It is possible that during a legislative period, a portfolio's area of responsibility is rearranged or a new portfolio is created. It is also possible that a portfolio's head was replaced during a legislative period. I consider each rearranged, newly created or each change of a portfolio head as an individual case. Since, at the local level, PQs are not addressed directly to a portfolio, I determined, based on the portfolio's area of responsibility, which CAP policy issues are under the jurisdiction of the respective portfolio. The dataset consists of the name of the portfolio, the head of the portfolio, the party affiliation of the head, the date the portfolio started working, the date the portfolio stopped being in place, the duration of the portfolio being active and its areas of responsibility based on the CAP policy coding scheme. Overall, the portfolio dataset consists of 218 unique portfolios (see Supplementary Material, Appendix B for details). Of the total portfolios, 157 (72%) are partisan portfolios headed by a member of a party, and 61 (28%) are headed by an independent without a party membership.

The main dataset of this article combines the PQ dataset and the portfolio dataset. Since the first hypothesis tests whether parties control independent portfolios more intensely than partisan portfolios, I use all party–portfolio combinations (including independent and partisan portfolios) for the models dealing with the first hypothesis. Thus, the data structure is very similar to that commonly used in research on intra-coalition control, where the researchers examine the number of PQs a coalition party addresses to the coalition's portfolios (cf. Höhmann and Sieberer 2020; Martin and Whitaker 2019). It is important to note that in some cities, the mayor changed during the legislative period of a city council. Since the second hypothesis tests whether governing parties control independent portfolios to a lesser extent than opposition parties do, the party affiliation of a mayor is a key factor, and thus I take mayor changes into account and use as my unit of interest party–portfolio combinations during a mayoral term in a legislative period of a city council. The models focusing on the second hypothesis use all party–portfolio combinations where a party addresses PQs to an independent portfolio. I exclude partisan portfolios in this analysis because I specifically want to test whether the classic pattern that government parties control to a lesser extent than opposition parties holds for independent portfolios too. Since some portfolios are only in place for a short time, I only include cases where the portfolio has been in place for at least 30 days. Overall, the main dataset consists of 1,615 observations (parties direct PQs to a portfolio during a mayoral term), 445 are party–portfolio combinations with an independent head and 1,170 are such combinations with a partisan head.

To test my hypotheses, I use multilevel zero-inflated negative binomial models. I opted for negative binomial models since the models' dependent variables are over-dispersed.⁴ I chose zero-inflated models since the dataset consists of more zeros than would be expected by a negative binomial distribution.⁵ I include random intercepts for parties nested in cities to control for city differences in PQ usage and for the possibility that some parties generally ask more parliamentary questions than others.

Dependent variable and explanatory variables

My dependent variable is the number of written PQs during a mayoral term in a legislative period by a party that addresses a policy issue under the jurisdiction of a certain portfolio. I retrieved this data by combining the PQ dataset and the portfolio dataset. I created a count variable capturing the sum of PQs a party addressed on a policy issue under the jurisdiction of a certain portfolio.

My first explanatory variable captures whether the head of the portfolio addressed by the PQs of a party is an independent or a partisan head using a dummy variable Independent portfolio (0 = portfolio head is partisan, 1 = portfolio head is an independent). A portfolio is to be considered independent if the head has no party membership and thus is not an official member of any party (see Appendix B for details of the party affiliation of each portfolio head).

My second explanatory variable captures whether the party raising the question has governmental responsibilities or not using the dummy variable Government party (0 = party is in opposition; 1 = party has governmental responsibilities). Due to the dual executive structure of mixed regimes, such a party may be aligned with the mayor, or part of the majority coalition of the council, or may be aligned with the mayor and part of the majority coalition of the council (Gross and Debus 2018). For this variable, I use data from Marc Debus and Martin Gross (2016) and add all mayoral changes during the covered legislative periods based on information from the respective city websites.

The first explanatory variable is used to test Hypothesis 1, and the second is used to test Hypothesis 2. For the models testing the first hypothesis, the variable Government party functions as a control variable to capture government and opposition dynamics in the regression model (Otjes and Louwerse 2018; Otjes et al. 2023).

Control variables

Previous research has shown that several factors may play a role in explaining the number of written PQs a party addresses to a portfolio. Thus, I use multiple control variables to control for those factors. As found by Martin and Whitaker (2019), I expect parties to ask more questions about policy areas that are salient to them. Therefore, I control for the Issue salience that topics of the addressed portfolio have for the asking party. I use data from Gross et al. (2024) to capture issue salience. The authors automatically determined the 200 most important keywords per policy issue based on their labelled PQ dataset for the 19 policy issues coded according to the CAP coding scheme. They used the keywords per topic as separate dictionaries and assessed the frequency of keywords used in a manifesto. The salience of a particular policy issue in a party's manifesto was calculated by dividing the number of keyword occurrences by the total number of words in the manifesto and z-transforming the variable. I use this data to generate an issue salience score for each party–portfolio combination. Since it is common in the portfolio dataset for a portfolio to be responsible for more than one of the 19 CAP topic areas, I calculate the average salience value of all topics that a portfolio is responsible for. For example, if a portfolio is responsible for the environment and energy, I add together the salience values for a party of these two topics and then divide it by two. This measure is particularly appropriate for this study because it is a relational measure and, therefore, captures well the extent to which the salience differs between parties across policy issues.

The size of a party affects how many resources a party has to ask questions (Otjes and Louwerse 2018). For parties with more MPs, it is easier to raise more PQs compared to parties with only a few MPs. To control for this, I use the variable Seats in council (log). The variable is measured as the log of the number of local council seats held. The data used for this variable come from Martin Gross and Marc Debus (2018). Furthermore, I use the variable Total no. of PQs asking party (in 100) which takes into account the total number of PQs submitted by a party during the respective period measured in units of 100, to control for the effect that parties that generally submit more PQs are also more likely to address independent portfolios.

To capture issue salience differences between portfolios (Green-Pedersen and Mortensen 2010; Otjes and Louwerse 2018; Vliegenthart and Walgrave 2011), I also control for the total number of questions a portfolio received, using the variable Total no. of PQs to portfolio measured in units of 100. Finally, since portfolio heads and portfolio policy allocation have changed for some portfolios and the duration a portfolio is in place during a mayoral term differs, I use the variable Duration portfolio measured in units of 100 days to control for duration effects.

Empirical findings

In this section, I present the results of my analysis. First, I show descriptively whether the number of PQs parties addressed to partisan or independent portfolios differs and whether governing parties diverge from opposition parties in their control behaviour of independent portfolios. I then test my hypotheses using multivariate analysis.

Descriptive analysis

At the local level in Germany, portfolios can be led by a partisan or an independent portfolio head. Of the 218 portfolios, 61 are independent portfolios, and 157 are partisan portfolios. In the main dataset, 445 party–portfolio combinations during a mayoral term consist of the number of PQs a party addresses to an independently led portfolio, and 1,170 party–portfolio combinations consist of the number of questions a party addresses to a partisan portfolio.

Figure 1 shows that, on average, a party addresses 5.9 PQs to a partisan portfolio and 8.5 PQs to a portfolio headed by an independent. So, on average, parties address 3.6 PQs more to independent portfolios relative to partisan-led portfolios – this difference is significant (p < 0.01) in a one-sided t-test. The results suggest that parties tend to control an independent portfolio more strictly than partisan portfolios. Thus, these findings provide descriptive support for the first hypothesis.

Figure 1. The Average Number of PQs a Party Addresses to a Portfolio Led by Either a Partisan or an Independent Head

Note: Error bars show the confidence interval (p < 0.05).

Figure 2 takes the governing responsibility of parties into account and shows the average number of PQs parties addressed to partisan or independent portfolios comparing governing parties and opposition parties. Due to the mixed regime at the German local level, parties with governing responsibility are parties that are affiliated with the mayor and/or parties that are members of the council's majority coalition. Figure 2 shows that, on average, governing parties generally tend to raise fewer PQs than opposition parties. Figure 2 displays that both parties with governing responsibilities and opposition parties tend to address more PQs to independent portfolios than partisan portfolios. Furthermore, the descriptive results also support the second hypothesis since they show that parties with governing responsibilities tend to control independent portfolios less intensely than opposition parties do.

Figure 2. The Average Number of PQs a Party Addresses to Independent or Partisan Portfolios whether the Party Has the Mayor Is Part of the Council Majority or Is in the Opposition

Note: Error bars show the confidence interval (alpha = 0.05).

In summary, I find tentative evidence for my two hypotheses based on the descriptive results, but those findings need to be empirically substantiated in a multivariate analysis. Thus, in the following main analysis, I focus on factors explaining the number of PQs addressed to a portfolio and assess whether parties address more PQs to independent portfolios while controlling for those other factors. In addition, I test whether parties with governing responsibilities differ in their control behaviour of independent portfolios compared to how opposition parties behave.

Multivariate analysis

I test each of the two hypotheses with a zero-inflated negative binomial model. Model 1 (see Table 1) is used to test the first hypothesis using all independent and partisan party–portfolio cases, and Model 2 (see Table 2) tests the second hypothesis using all party–portfolio cases focusing on independent portfolios.⁶ Both models use in the binary part of each model an intercept and the variables Duration portfolio (in 100) and Total no. of PQs asking party (in 100). The duration variable is suitable to account for excessive zeros since the longer a portfolio has been in place, the less likely it is that a party addresses no question at all to this portfolio. In addition, the variable that captures the total number of questions a party asks in a council is also appropriate to account for excessive zeros because the more PQs a party asks in total, the less likely it is that a party will not ask at least one PQ to a portfolio.

Table 1. Zero-Inflated Negative Binomial Regression Model of Parliamentary Questions Used to Control the Portfolios

Notes: Multilevel zero-inflated negative binomial regression models with random intercepts (parties are nested in cities). Dependent variable: total number of PQs per portfolio asked by a party in local city councils. Standard errors of log-mean coefficients are in parentheses. IRR = Incidence Rate Ratio. Total no. of PQs to a portfolio, Total no. of PQs asking party and the Duration portfolio are measured in increments of 100. Significance levels: *p < 0.05; **p < 0.01; ***p < 0.001.

Table 2. Zero-Inflated Negative Binomial Regression Model of Parliamentary Questions Used to Control the Portfolios (Independent Portfolio-Only Models)

Notes: Multilevel zero-inflated negative binomial regression models with random intercepts (parties are nested in cities). Dependent variable: total number of PQs per portfolio asked by a party in local city councils. Standard errors of log-mean coefficients are in parentheses. IRR = Incidence Rate Ratio. Total no. of PQs to a portfolio, Total no. of PQs asking party and Duration portfolio are measured in increments of 100. Significance levels: *p < 0.05; **p < 0.01; ***p < 0.001.

Model 1 is used to test the hypothesis that independent portfolios are generally controlled more by all parties compared to partisan portfolios. The model uses the full dataset and all of the control variables in the count model section, thus accounting for multiple factors that impact how many PQs a party addresses to a portfolio. Model 2 uses a subsample consisting of the 445 party–portfolio combinations with independent portfolios to test whether government parties differ from opposition parties in their control behaviour of independent portfolios, holding other factors of influence constant. I use the same control variables as in Model 1 in the count part of the model. However, this time, I use the variable Government party to test my hypothesis.

Table 1 shows the results of the model testing the hypothesis of whether parties address more PQs to independent portfolios compared to partisan ones (Model 1). I find empirical support for the first hypothesis. The results show that parties ask independent portfolios significantly more questions compared to partisan portfolios. The respective Incidence Rate Ratio (IRR) shows that if the asked portfolio is independent, the number of PQs increases compared to a partisan portfolio by a factor of 1.15. Thus, the average number of PQs a portfolio receives is 15% higher when it is an independent one compared to being a partisan portfolio. In other words, holding all other variables constant, a 15% increase would mean that a party submitting 50 PQs to a given portfolio would submit approximately 58 PQs to the same portfolio if it were led by an independent. The control variables in the count part of the models show the expected significant effects except the variable capturing the number of seats a party has. In addition, in the binary section of both models, the variables Duration portfolio (in 100) and Total no. of PQs asking party (in 100) show a significant negative effect and thus account, as expected, for the presence of excessive zeros in the model.

Table 2 shows the results of the model testing the second hypothesis of whether parties with governing responsibilities control independent portfolios to a lesser extent than opposition parties do. Model 2 shows a negative significant effect for the variable Government party: if a party is part of the government, the party addresses significantly fewer PQs to an independent portfolio compared to an opposition party. The respective IRR displays that being a governing party reduces the number of PQs addressed to an independent portfolio by a factor of 0.73. Consequently, on average, a governing party addresses 27% fewer PQs to a portfolio led by an independent compared to an opposition party. In other words, holding all other variables constant, a 27% decrease would mean that a party submitting 50 PQs to a given portfolio would only submit approximately 37 PQs to the same portfolio if the party had governmental responsibilities. The control variables show the expected significant effects except for the variables issue salience and the number of seats a party has. Overall, these findings are in line with the second hypothesis and show that parties with governing responsibilities control independent portfolios less strictly than opposition parties.

Conclusion and discussion

Research on legislative control has extensively examined the role of political parties and coalition dynamics in how political parties use different legislative tools to control portfolios and their respective heads (Höhmann and Sieberer 2020; Otjes et al. 2023; Raunio 1996). However, research has focused on partisan-run portfolios and has overlooked how control dynamics are affected when portfolio heads are independent, thus not affiliated with any party. In this article, I provide one of the first analyses of how parliamentary oversight mechanisms are utilized to control independent portfolios compared to partisan portfolios.

Focusing on the local level in Germany, the results confirm the first hypothesis that all parties tend to exercise more intense control over independent portfolios, using PQs as a control instrument. This is consistent with the argument that independent ministers, due to their specialized expertise and non-partisan status, may require tighter monitoring to ensure their actions align with party objectives within acceptable boundaries. The empirical evidence suggests that parties recognize the need to manage the potential agency drift of independent ministers by actively engaging in oversight through PQs.

However, this finding also raises a critical question: does the observed discrepancy between government and opposition parties' behaviour towards independent ministers indicate a potential tension between effective oversight and harmonious governance? While robust oversight is essential for democratic accountability, overly stringent control measures could hinder cooperation and efficient policy implementation, especially within coalition governments. Thus, at its core, this question focuses on whether the appointment of technocrats as ministers has positive or negative implications for the functioning of legislatures. Shedding more light on this aspect in future research could help to better understand whether parliamentary oversight of the executive is particularly effective for independent ministers compared to partisan-led ministries, or whether stricter oversight tends to negatively affect the functioning of democratic systems.

The results of the analysis support the second hypothesis, demonstrating a distinction between parties with governing responsibilities and opposition parties in terms of their control behaviour towards independent portfolios. Government parties exercise less rigorous control over independent portfolios compared to opposition parties. This asymmetry can be attributed to the inherent trust and alignment that may exist between government parties and independent ministers, as well as the desire to maintain harmonious relations within the coalition and to present a united front to the public.

These findings contribute to the broader literature on legislative oversight and government accountability and provide valuable insights into how PQs are used as a control mechanism to monitor portfolios headed by an independent. In addition, the results of this study extend our understanding of how the local level functions and provide further evidence that local politicians use control instruments similarly to politicians from higher political levels (Otjes et al. 2023).

However, certain limitations should be acknowledged. The reliance on parliamentary questions as the primary data source, while informative, may not capture the full range of control mechanisms and interactions between parties and independent ministers. Future research could explore additional sources of data, such as committee hearings or media coverage, to provide a more comprehensive understanding of the control dynamics.

Moreover, although the German local mixed regime functions similarly to mixed regimes at higher political levels (Debus and Gross 2016; Gross and Debus 2018), and thus the findings should be generalizable to higher political levels, the local level still differs in important ways. One difference is that local issues rarely attract media attention. Thus, compared to national politicians who benefit from a national media spotlight, local politicians have a much harder time reaching the public. This makes it more difficult for them to use PQs to demonstrate their commitment to serving the public's interest, limiting their ability to connect with voters through this legislative tool. Another key difference is the policy areas that local politicians control. Unlike national politicians, who deal with broad issues like the economy, local politicians focus on things closer to home, like building roads, fixing schools and supporting local businesses. To ensure that these findings apply to other political levels, future research should test whether the findings on how parties control independents hold for other political levels. Moreover, by also focusing on cases from presidential or parliamentary regimes, it would be possible to test and ensure that independent control functions similarly across regime types.

For future research, it would be interesting to analyse whether the reasons behind the selection of an independent affect party control behaviour as independents are selected as ministers for different reasons (Alexiadou and Gunaydin 2015; Costa Pinto et al. 2018). For example, it may be the case that an independent who is known to have been selected to push through unpopular reforms is more tightly controlled by opposition parties that may not be in favour of those reforms than an independent who came to office because of their expertise and for being known to be a neutral technocrat.

Furthermore, independent ministers can be divided into two subtypes: those who are not members of a political party but have some loose affiliation and connections with a certain political party, and those who are not party members and have no form of association with any political party. So, the latter can be considered as ‘pure’ independents in a strict sense (Costa Pinto et al. 2018). In future research, it would be interesting to analyse whether pure independents are controlled differently from independents with a party alignment. It may be the case that the former are controlled more strictly due to a higher risk of agency drift and that the latter are controlled more similarly to partisan portfolio heads. In addition, to promote a deeper understanding and to advance the theoretical concept behind the control of independent portfolios, it would be beneficial if future research focused on studying the concrete motives for why parties decide to control independent ministers. In particular, expert interviews with politicians or their staffs could help enormously in gaining a deeper understanding of the processes behind the control behaviour and contribute to the theory of legislative control behaviour.

In conclusion, this article underscores the nuanced ways in which parliamentary oversight tools are employed to control independent portfolios. The evidence suggests that parties, driven by a combination of expertise, trust and accountability considerations, choose their control strategies based on their governing responsibilities or opposition status. By contributing to the literature on legislative control in this distinct setting, the article enriches the discourse on government oversight and the functioning of modern democracies.