Introduction
Modern societies are heavily regulated by laws that affect the daily lives of all of us. In parliamentary democracies, the process of initiating these laws is dominated by governments more than legislatures (Martin, Reference Martin2004; König et al., Reference König, Lin and Silva2023; Martin and Vanberg, Reference Martin and Vanberg.2004: 14; Laver and Shepsle, Reference Laver and Shepsle.1996: 41), and studies have shown that government bills here have a high passage rate (Cheibub et al., Reference Cheibub, Przeworski and Saiegh.2004; Brunner, Reference Brunner2013: 578). With this prominent legislative role of government in parliamentary democracies, research has focused on themes such as determinants of the content of governments’ policy agenda (Jennings et al., Reference Jennings, Bevan, Timmermans, Breeman, Brouard, Chaqués-Bonafont, Green-Pedersen, John, Bjerre Mortensen and Palau.2011; Green-Pedersen et al., Reference Green-Pedersen, Bjerre Mortensen and So.2018), factors that pose barriers for the government in pursuing its policy agenda (Bräuninger and König, Reference Bräuninger and König.1999; König et al., Reference König, Lin and Silva2023), and means to overcome these (Fortunato et al., Reference Fortunato, König and Proksch.2013). Many governments are minority governments, and much research has focused on ways for these to obtain a majority (see Field and Martin, Reference Field and Martin.2022: 335). The ability of minority governments and minority coalitions to form legislative coalitions or alliances with nongovernment parties can be a precondition for the government’s stability and policymaking effectiveness (Klüver and Zubek, Reference Klüver and Zubek.2018; Krauss and Thürk, Reference Krauss and Thürk.2022). Yet, according to a recent study, such research ‘has paid little attention to the principal-agent problem’ inherent in governance, where ‘the ministerial office-holder has informational advantages’, which can make ‘policy-making more complex and difficult’ (König and Lin, Reference König and Lin.2021: 694). Principal–agent problems are inherent in ministerial government when ministers are delegated authority within a given departmental jurisdiction (Strøm, Reference Strøm2000). PA problems are most likely to arise in multiparty cabinets. Complete unity is by no means given in single-party cabinets either. Yet, studies have to a large degree ‘bypassed the role of individual politicians and have assumed that political parties are unitary actors’ (Alexiadou, Reference Alexiadou2015: 1052). As Bäck et al. (Reference Bäck, Müller, Angelova and Strobl.2022: 255) note, ‘it remains unclear whose preferences eventually shape policymaking in coalition governments and to what extent’. I turn attention to these unexplored matters concerning cabinet PA problems, individual ministers, and political output in this article.
Control and oversight mechanisms have become apparent in contemporary governments, which can help mitigate potential agency loss (Bergman et al., Reference Bergman, Back and Hellström.2021a, Reference Bergman, Back and Hellström.2021b. See also Müller and Meyer, Reference Müller and Meyer.2010). One such mechanism is cabinet committees or inner cabinets (Andeweg and Timmermans, Reference Andeweg and Timmermans2008; Ie, Reference Ie2022). Yet, it has been argued that ‘despite their importance within the executive in many parliamentary systems’ cabinet committees ‘have largely been neglected as objects of study’, and consequently, future studies ought to focus on cabinet committees in relation to ‘policy coordination’ (Ie, Reference Ie2022: 230). In other words, the scope of cabinet committees’ impact on ministers’ legislative activity remains an open question (cf. also Bergman et al., Reference Bergman, Angelova, Bäck and Müller.2023: 23).
In this article, in addition to the suggested gaps in the literature, I am studying whether membership of central cabinet committees increases a minister’s legislative activity compared to nonmembers. I am investigating individual ministers’ number of laws proposed to parliament during parliamentary sessions as a measure of legislative activity. As mentioned, ministers are considered to be given a crucial role in the process of proposing legislation. But there may be reasons for a minister’s party to stray from the overall cabinet agenda in coalitions, and furthermore, there can also be reasons to expect that some ministers will pursue policy goals not in line with that of their own party. On the other hand, control mechanisms in cabinet are set in place not least to keep ministers from doing so. What happens in reality between these somewhat contradictory expectations is important for our understanding of policymaking in cabinets and parliamentary democracies in a broader sense. I argue that ministers appointed to certain central cabinet committees enjoy an advantage in the pursuit of realizing their policy goals. These committees are given such a prominent position in cabinet decision-making as coordination mechanisms to avoid agency loss upon ministerial appointments, which will arguably increase members’ legislative activity. To explore this somewhat unknown territory, I constructed a dataset consisting of individual ministers’ number of proposed laws in the Danish parliament per annual session. The dataset covers parliamentary sessions from 1975 to 2022, and approximately 8000 law proposals were counted. Denmark is a well-suited country for the study at hand, as the political system shares many similarities with other Western European democracies. Furthermore, cabinet governance here, on the surface, contains the paradox of having de jure autonomous ministers and yet also a long history of using cabinet committees in governments’ decision-making processes.
The analysis shows that membership of the so-called Economic Committee significantly increases ministers’ legislative activity, whereas membership of the Coordination Committee shows less certain results. The study offers several contributions that advance our understanding of intra-cabinet control mechanisms, cabinet agenda setting, policymaking in parliamentary democracies, political parties, and cabinet governance in general. Ministers and their legislative activity are being studied individually, as opposed to studying parties in cabinet as unitary actors, which provides new insights into the legislative role of ministers in parliamentary democracies. While it is a single case study focusing on Denmark, the implications are relevant for our knowledge of ministers as policymakers in all parliamentary democracies where central cabinet coordination bodies are installed and help us understand who actually shapes policy in these systems.
Ministers’ legislative activity – risks of ministerial drift and means to avoid it
While policymaking is a fundamentally important part of participating in government, it also comes with several difficulties. The vast and complex tasks faced by the national government are divided and shared between individual ministers, each responsible for separate parts of the government machinery. Ministers are granted a crucial informational advantage upon appointment due to the resources and expertise of their departments (Laver and Shepsle, Reference Laver and Shepsle.1990: 874; Martin and Vanberg, Reference Martin and Vanberg.2004). Some even consider them to have a near-monopoly position to formulate detailed legislative proposals within their respective policy areas (cf. Bäck et al., Reference Bäck, Müller, Angelova and Strobl.2022).
Upon appointment, a Prime Minister and/or a party leader may have had certain wishes regarding what each minister should do in office. But as Strøm (Reference Strøm2000: 270) emphasizes, ‘any delegation of authority creates the risk that the agent may not faithfully pursue the interests of the principal’. Such delegation problems occur when ministers are appointed. Adverse selection can occur when ministers with undesired characteristics are appointed, even when Prime Ministers want to appoint ministers in line with their own views. The availability of personnel, as well as party political and constitutional factors, may constrain the Prime Minister. Prime Ministers are also expected to consider other characteristics, such as gender, geography, and age, in the formation of cabinet. Prime Ministers may even feel pressured to keep ministers who are publicly popular, regardless of their potential flaws (Nielsen, Reference Nielsen2022a). Agency rent can occur if ministers turn out to lack sufficient skill or simply if ‘the minister has different policy preferences’ from those of the Prime Minister (Dowding and McLeay, Reference Dowding and McLeay.2011: 159; see also Dowding and Dumont, Reference Dowding and Dumont.2009: 2).
Delegation problems in cabinets are often studied with regard to ‘inter-party’ agency loss (Bäck et al., Reference Bäck, Müller, Angelova and Strobl.2022: 257). From this perspective, ministers may have incentives to use their informational resources to their own or their party’s advantage in the policymaking process (Martin and Vanberg, Reference Martin and Vanberg.2020; Bäck et al., Reference Bäck, Müller, Angelova and Strobl.2022; Bergman et al., Reference Bergman, Angelova, Bäck and Müller.2023; see also Andeweg, Reference Andeweg2000). A minister might try to shape certain policies or aspects of a legislative proposal (positive agenda setting), or even hesitate or temporarily refuse to introduce certain policies (negative agenda setting or gatekeeping) to cabinet (Bäck et al., Reference Bäck, Müller, Angelova and Strobl.2022: 257). Ministers may also attempt to shape the proposals coming from their own departments in a way that deviates from the cabinet agenda (Laver and Shepsle, Reference Laver and Shepsle.1996: 32–33) and present it as the best possible solution.
However, I assume in this study that intra-party agency issues can also occur in cabinets, similar to those mentioned above. Indeed, ‘governing creates opportunities for intra-party tensions to arise’ (Greene and Alexiadou, Reference Greene and Alexiadou.2023: 239), and one study even found party factions to affect policy (Schumacher, Reference Schumacher2012). When Prime Ministers are to appoint ministers, ‘different factions within a party must be satisfied’ because ‘even in single-party governments there are personal rivalries for the top jobs and ideological tension between wings and factions within the party’ (Dowding and Dumont, Reference Dowding and Dumont.2009: 2–3; see also Dowding and McLeay, Reference Dowding and McLeay.2011: 159). In a nutshell, Indriðason and Kristinsson (Reference Indriðason and Kristinsson.2013: 824) claim that ‘ministers are rarely perfect agents of their parties’ (see also Dewan and Myatt, Reference Dewan and Myatt.2005, Reference Dewan and Myatt.2010; Indriðason and Kam, Reference Indriðason and Kam.2008). This point is linked to a second assumption central to this study, namely that ministers are policy-seeking. Parties assuming government are often considered policy-seeking (Krauss, Reference Krauss2018: 1284). Holding ministerial office is an obvious way to shape public policy, for instance, to fulfill electoral pledges. In this view, individual ministers are agents of their parties. But, as described, ministers may not unambiguously be considered so. Therefore, I also assume that individual ministers are policy-seeking, which may in itself be a factor for intra-party divisions to arise when parties enter government. These are good reasons for looking beyond parties in cabinet as unitary actors, and therefore, ministers are studied individually in this article.
As Prime Ministers and party leaders would be aware of risks of agency loss – inter-party as well as intra-party-related – we have reason to expect that they will seek to ensure that the content of the ministers’ law proposals applies to the overall government agenda. This (along with other factors) encourages Prime Ministers to install monitoring devices and coordination and control mechanisms in cabinets (e.g., Müller and Meyer, Reference Müller and Meyer.2010; Bergman et al., Reference Bergman, Back and Hellström.2021b: 700ff; Thies, Reference Thies2001; Carroll and Cox, Reference Carroll and Cox.2012). Institutions, such as cabinet committees and ‘inner cabinets’, are indeed set up as such mechanisms (Andeweg and Timmermans, Reference Andeweg and Timmermans2008). Studies have also highlighted the drafting of coalition agreements in multiparty cabinets as an instrument to mitigate agency loss and control cabinet’s policy agenda (Klüver et al., Reference Klüver, Bäck and Krauss.2023; Indriðason and Kristinsson, Reference Indriðason and Kristinsson.2013; Moury, Reference Moury2011; Moury and Timmermans, Reference Moury and Timmermans.2013). In some cases, ministries with overlapping or bordering jurisdictions may have ministers from different parties appointed, and junior ministers can also be appointed as ‘watchdogs’ (Fernandes et al., Reference Fernandes, Meinfelder and Moury.2016; Lipsmeyer and Pierce, Reference Lipsmeyer and Pierce.2011). However, how such control mechanisms affect ministers’ legislative activity remains an open question (König and Lin, Reference König and Lin.2021). Bergman et al. (Reference Bergman, Angelova, Bäck and Müller.2023: 5) have argued that once a government is formed, ‘none is in a position similar to that of an acting minister when it comes to influencing policy content’, even despite the presence of an extensive legislative agreement. The authors go on to suggest that issues emerge one-by-one during the legislative term, which increases the uncertainty of ad hoc policy ‘without institutional mechanisms, which foster partisan commitment’ (Bergman et al., Reference Bergman, Angelova, Bäck and Müller.2023: 6). Certain cabinet committees should be considered such an institutional mechanism for containing ministerial drift and agency loss after government formation. In the past, full cabinet meetings may have had a similar function, but, as Peters and Helms (Reference Peters and Helms.2012: 31) has it, ‘the rise of cabinet committees […] have gradually transformed the full cabinet to a court of appeal, and effectively strengthened, rather than weakened, the prime minister’ (see also: Ie, Reference Ie2022: 116; Andeweg, Reference Andeweg2000).Footnote 1
While all ministers attend full cabinet meetings, cabinet committees comprise only of some of these ministers (see also Appendix D). Cabinet Committees are commonly mentioned as control mechanisms to avoid inter-party agency loss in multiparty governments (Martin and Vanberg, Reference Martin and Vanberg.2013; Oppermann et al., Reference Oppermann, Brummer and van Willigen.2017; Dragu and Laver, Reference Dragu and Laver.2019). In single-party governments, cabinet committees have also been mentioned as playing an important role in the centralization of intra-executive decision-making processes (Heffernan, Reference Heffernan2003; Ie, Reference Ie2022: 118; Dunleavy, Reference Dunleavy2003). As such, certain cabinet committees should be considered as means to avoid both drift of ministers from different parties in cabinet and of ministers within parties in cabinet. It has been suggested that future research should pay attention to consequences rather than causes of the type of organization of cabinet decision-making, and whether the degree of hierarchy and central control mechanisms in these is related to ‘the policy outcomes of the cabinet’ (Vercesi, Reference Vercesi2012: 22; see also: Ie, Reference Ie2019: 484; Ie, Reference Ie2022: 230). This article aims to contribute to fill this gap in the literature.
Cabinet committees and their members as advanced policy initiators
As certain cabinet committees are installed to coordinate cabinet policymaking and, in so doing, to contain agency loss ex post ministerial appointment, there are reasons to expect that committee members will enjoy more opportunities to propose laws and, perhaps more importantly, nonmembers will face obstacles in doing so. I focus on cabinet committees concerned with overall decision-making and coordination in cabinet. Such cabinet committees can be regarded as an institutionalization of hierarchy in cabinet decision-making processes. For them to work effectively to contain agency loss, they must take decisions – not only prepare decisions – and consequently function as ‘partial cabinets’ (Andeweg, Reference Andeweg2000: 387).Footnote 2 Bergman et al. (Reference Bergman, Back and Hellström.2021b: 701) show that in 13 of the 16 countries (81 percent) they analyze ‘cabinet committees are among the most common mechanisms used’ (Bergmann et al., Reference Bergman, Back and Hellström.2021b: 702) as coalition management mechanisms. They argue that these appear ‘to deal with the main challenge of reaching collective decisions in situations in which the individual incentives of […] the minister formally in charge of the policy area might differ from that of the majority in the cabinet’ (Bergman et al., Reference Bergman, Back and Hellström.2021b: 700). This highlights the prominent role that cabinet committee members enjoy in cabinet policymaking. Similarly, single-country case studies show how cabinet committees are crucial for cabinets’ policy coordination, also in single-party cabinets (for review, see: Ie, Reference Ie2022: 117; see also Kolltveit and Shaw, Reference Kolltveit and Shaw.2022: Ch. 4 and 5; Dunleavy, Reference Dunleavy1995, Reference Dunleavy2003; Ie, Reference Ie2019).
In Danish cabinets, two cabinet committees have been particularly important. These are the Coordination Committee and the Economic Committee. Their permanent members meet regularly, while nonmembers can be invited to join the meetings if they have a case on the agenda. One study based on 21 interviews with former and current permanent secretaries and ministers found it to be the Prime Minister and the Minister of Finance ‘and their committees, who control policy-making’ (Rhodes and Salomonsen, Reference Rhodes and Salomonsen.2021: 74) in Danish governments. ‘Their committees’ refer to the Coordination Committee, which the Prime Minister chairs, and the Economic Committee, which the Minister of Finance chairs today. Accordingly, these two committees constitute the ‘inner court’ of Danish cabinets (Rhodes and Salomonsen, Reference Rhodes and Salomonsen.2021: 77; see also Nielsen Reference Nielsen2020, Reference Nielsen2023).Footnote 3 This claimed role of controlling cabinet policymaking is crucial for the proposed causal mechanism under investigation in this article. As long as members agree with each other, they can propose new initiatives, while nonmembers at some point must have permission from a committee to do so. This should be expected to increase committee members’ legislative activity compared to nonmembers. A policy-seeking committee member would arguably try to assign himself with new initiatives when possible, while nonmembers don’t enjoy the same possibilities of doing so.
Law proposals are in Denmark a matter of discussion in these two cabinet committees (Jensen, Reference Jensen2018: 157–158. See also Nielsen, Reference Nielsen2020: 440–441). We would assume that members of the committees will try to affect the content of the proposals presented in the committees by nonmembers. Accordingly, this lies at the core of the committees’ role, and it is what distinguishes committee members from the rest of the ministers: The chance for them to intervene in policies under the jurisdictions of other ministers (Nielsen, Reference Nielsen2020). The permanent members can perform sometimes very strict control over nonmembers. Nonmembers know that they will be met with scrutiny when proposing initiatives to the committees, and this may even prevent them from proposing new initiatives, which would consequently decrease, or at least delay, nonmembers’ legislative activity. It requires very thorough work for nonmembers to prepare their cases for committee meetings (Dybvad-udvalget, 2023: 325; see also Rhodes and Salomonsen, Reference Rhodes and Salomonsen.2021: 76). It can be time-consuming even getting their initiatives on the agenda for a meeting, and this process can be prolonged if the senior bureaucrats preparing the committee meetings don’t find the case sufficiently prepared, or if other matters are more urgent for the committees. This preparation and the fact that nonmembers need to be invited to meetings in the committees, while the committee members meet regularly, will also delay the process of promoting law proposals for nonmembers (see Dybvad-udvalget, 2023: 93).
When nonmembers attend meetings, the committee members and ultimately the chairman will, in the end, draw the conclusion in case of disagreement. According to one study, if a minister doesn’t agree with the conclusion, the minister will have to accept it and follow it, or ultimately resign (Christiansen, Reference Christiansen2021: 150). This entails a bargaining advantage for members of the committees, as they know Danish ministers quite seldom resign due to policy disagreements (Nielsen, Reference Nielsen2020: 442; Reference Nielsen2022a: 367).
As mentioned, all ministers are assumed to have an informational advantage related to issues under their jurisdiction, which strengthens their role as policy initiators. However, according to a rare field study carried out at the Danish state-level bureaucracy, membership of a central cabinet committee provides the members’ ministries and their private secretaries with privileged all-round knowledge of what is going on in other ministries (Trangbæk, Reference Trangbæk2021: 251. See also Dybvad-udvalget, 2023: 95). Thus, the cabinet committees as an institution reduce nonmembers’ informational advantage, consequently strengthening members’ role in cabinet policymaking.
It is beyond the scope of this article to elaborate in depth on the process of appointing members to cabinet committees, and generally, we have limited systematic knowledge on this matter. Intra-party allocation of portfolios is seldom subject to studies (cf. Smith and Martin, Reference Smith and Martin.2017: 131). In many cases, it seems to be persons that Prime Ministers and party leaders want as close advisors who are appointed membersFootnote 4 , or in some cases, persons who by appointment could be expected to appease intra-party divisions.Footnote 5 Sometimes, newer and younger ministers are appointed, seemingly with consideration to the party’s line of succession.Footnote 6
A few further remarks should be made on appointment to committees. First, one could argue that some ministers are planned to have a large legislative activity and are thus appointed to a cabinet committee. However, I am not aware of any accounts that propose this. It could also be that some portfolios are always represented in committees due to a persistent large legislative activity, rather than committee membership being the decisive factor for legislative activity. Empirically, however, there seem to be no over-encompassing criteria related to the portfolio automatically conferring to permanent membership of a central cabinet committee. For instance, ministers with economy-related portfolios are often members of the Economic Committee, but it is not always the case. Instances have occurred of Ministers of Industry and Business Affairs, Taxation, and Foreign Trade not being members, while ministers such as those for Higher Education and Research, Cultural Affairs, and Education have been permanent members on several occasions. Similarly, there has been a large variation in the composition of the Coordination Committee. Second, it could also be that certain supposedly active or skilled ministers are appointed to the committees, which increases legislative activity more than the actual committee membership. Such qualifications would arguably be seniority or perhaps portfolio-related experience, which may perhaps have made way for a cabinet committee appointment in some cases (see online Appendix D). But on the contrary, such ministers are also just as likely (or more likely) to be appointed to a post that is presumably difficult to handle, which would be an argument for that minister to focus narrowly on their portfolio rather than for them to be involved in the coordination of policies in all other ministries in the central cabinet committees. Footnote 7 Finally, as in appointing ministers in general, other criteria might outweigh seniority, and in some parties, there will be few or no very experienced candidates to choose from.
All things considered, there is reason to believe that membership of a central cabinet committee entails an advantage for ministers as policy initiators. Thus, the following hypotheses are proposed:
Hypothesis H1a: Membership of the Coordination Committee will increase a minister’s number of laws proposed to parliament during a parliamentary session.
Hypothesis H1b: Membership of the Economic Committee will increase a minister’s number of laws proposed to parliament during a parliamentary session.
Data and Methods
This section presents the research design of the study. Data collection and measurements of the dependent and independent variables are also elaborated on.
Laws as a measure of ministers’ legislative activity
Ministers’ legislative activity is the dependent variable of this study and is measured as individual ministers’ number of laws proposed to parliament. Proposing laws to parliament should indeed be considered as a means for politicians to pursue their policy goals. The approach is, to some extent, inspired by a study on the legislative activity of female ministers in four Latin American countries, which measures legislative activity as ‘the number of bills and laws initiated by men and women’ (Escobar-Lemmon and Taylor-Robinson, Reference Escobar-Lemmon and Taylor-Robinson.2016: 233). Similarly, when the term ‘legislative activity’ is used in the subsequent analysis, it denotes the number of laws proposed to parliament.
Individual ministers’ policy outcomes in parliamentary systems have only rarely been subject to in-depth analyses. Only recently, some studies have focused on the conditions of policy output change, while not focusing on individual ministers (Martin and Vanberg, Reference Martin and Vanberg.2020; Bäck et al., Reference Bäck, Müller, Angelova and Strobl.2022; Bergman et al., Reference Bergman, Angelova, Bäck and Müller.2023; see also Klüser, Reference Klüser2023). One recent paper studied individual ministers’ involvement in cabinet decision-making, albeit with a narrow emphasis on crisis management during the first wave of COVID − 19, based on data from full cabinet meetings (Shpaizman, Reference Shpaizman2023). Alexiadou (Reference Alexiadou2015) focused on the impact of ministers’ personal backgrounds on social welfare policy output under their jurisdictions. The dependent variable is an index of welfare state decommodification, including insurance programs related to pension, unemployment, and sick pay. Becher (Reference Becher2010) investigated data on the influence of labor ministers on unemployment insurance entitlements and employment protection legislation, while Bojar (Reference Bojar2019) studied the effect of ministers’ party-political backgrounds on the annual allocation of budgetary resources.
These studies are most relevant to our understanding of ministers’ legislative activity. However, when attention is turned to specific policy areas, not all cabinet ministers can be included in the analysis (Becher, Reference Becher2010: 39–40; Alexiadou, Reference Alexiadou2015: 1064; Bäck et al., Reference Bäck, Müller, Angelova and Strobl.2022). When the dependent variable is measured as an economic value or a measure that presupposes larger spending (Bäck et al., Reference Bäck, Müller, Angelova and Strobl.2022: 269–270; Bojar, Reference Bojar2019; see also: Bergman et al., Reference Bergman, Angelova, Bäck and Müller.2023), much political output is also excluded from the analysis. Thus, we miss the wider scope of the potential impact of the committees’ interference. Laws could be considered important and certainly subject to cabinet coordination, even if they are (next to) cost-neutral. Therefore, studying ministers’ number of laws proposed to parliament offers a different perspective on ministers’ influence on policy that includes all line ministers and the potential variations in legislative activities between them.
This measure implies that laws are important. When including all laws, one could, from different perspectives, even normatively argue that some laws are more important than others and conversely that some laws are unimportant, which consequently would make this measure of legislative activity disputable. Yet, for analytical purposes, it is not straightforward to distinguish more or less important laws. For instance, even if a law addresses only a narrow or quite local audience, it means that others are not addressed by it, which could stir anger or frustration among other voters, causing indirect (nationwide) consequences. And even if a new law upholds the status quo, it could arguably be seen as an impact itself, although it would not be counted in a measure of policy output change (like Bäck et al., Reference Bäck, Müller, Angelova and Strobl.2022). The status quo might have been altered in the case of another cabinet, and a lack of change may also provoke frustration among voters. Arguably, any law will have at least minimal impact for some people and at least one politician or bureaucrat would assumably argue for its importance (see also Escobar-Lemmon and Taylor-Robinson, Reference Escobar-Lemmon and Taylor-Robinson.2016: 232–233).
Research design
Cases in the dataset are ministers in a given year. I have coded the number of laws proposed to parliament for each minister for the parliamentary sessions in the Danish Parliament from 1975–1976 to 2021–2022 including both. Ministers’ number of laws proposed to parliament is count data and includes an excess number of zero laws. As the conditional variance of the dependent variable is greater than the mean, I employ negative binomial regression models in the analysis.
1975 is a reasonable starting point for the analysis, as the legislative process has been institutionalized in such a way that makes for a comparative analysis for such a long period of time with regard to ministers’ legislative activity. Cabinets have dominated the legislative process by proposing nearly all passed laws, and the total number of laws proposed per session didn’t vary dramatically. Furthermore, during this period, cabinets in Denmark introduced annual law programs to parliament as a means of legislative planning. These are published each year at the annual opening of parliament on the first Tuesday of October.Footnote 8 To investigate the overall legislative culture of cabinet’s influence on the legislative process, I include only ordinary parliamentary sessions starting in October. This precludes sessions that start sometimes in between due to general elections, where the introduction of a law program has not consistently been the norm.Footnote 9
Cabinet committee membership status for each minister is the independent variable of main interest. I focus on the central cabinet committees that serve as de facto inner cabinets in cabinet decision-making processes: The Coordination Committee and the Economic Committee. The Coordination Committee existed from 1968–1971, 1978–1980, and then continuously from 1982 to the present. The Economic Committee dates back to at least 1947, and it has in many years functioned as a de facto inner cabinet in the absence of a coordination committee. Only in the years 1982–1988 was the Economic Committee not given a prominent role in cabinet decision-making and was eventually closed in 1990, but from January 1993 to the present, it has existed continuously. In the years that both committees have existed, the intention is for them to supplement each other with differing divisions of labor between them and with some ministers as members of only one of them. Whereas Danish cabinets typically include some 20 ministers, each committee typically comprises 4–8 ministers (see also online Appendix D). Ministers are coded according to membership of each of these two cabinet committees. Members are coded 1, whereas ministers are coded 0 if they weren’t members, or if the given committee didn’t function as a central cabinet committee during the given session.
I focus on legislative activity for line ministers, excluding Prime Ministers and Ministers of Finance. In the Danish context, these two are the coordinators, not subjects to coordination. The Prime Minister is the main principal in cabinet. The Minister of Finance has chaired the Economic Committee in most of the years studied, has been member of both committees consistently, and due to the limited capacity of the Danish Prime Minister’s Office, ‘the Ministry of Finance acts as the Cabinet Office’ for the Prime Minister (Rhodes and Salomonsen, Reference Rhodes and Salomonsen.2021: 79).
Comparing how many laws ministers propose to parliament during a session is complicated due to dismissals, resignations, and a considerable number of reshuffles, etc. Some sessions have also been interrupted because of calls for a general election. In the analysis, I focus on laws introduced by ministers holding office at the beginning of parliamentary sessions. Their time on the post is introduced as a control variable measuring their share of the total session holding the post. If ministers leave office or are reshuffled to another post, they are only included for the first post they hold. If a minister enters or leaves a cabinet committee during a session while holding the same portfolio, the minister is included as a new case and is coded according to the changed membership status.
When Prime Ministers and Ministers of Finance are excluded, the dataset includes 855 (unweighted) cases of ministers holding office from the beginning of a parliamentary session for shorter or longer periods. Table 1 provides an overview of ministers’ legislative activity per session, measured in two different ways (see also online Appendix E). Ministers have proposed between 0 and 60 laws during parliamentary sessions (the case of 60 certainly is an outlier, as there are only three cases above 50, 11 above 40, and 23 above 30). Ministers holding office at the beginning of a session have on average proposed 9.7 laws per annual session (weighted). When Prime Ministers and Ministers of Finance are excluded, the mean is slightly higher and even a little higher for members of the Coordination Committee. For members of the Economic Committee; however, this mean is 15.3 proposed law per session, which marks a noteworthy difference from the rest of the ministers.
Table 1 Descriptives. Ministers’ legislative activity
1 Weighted according to time of session on post.
2 Only ministers on same post for an entire session, who didn’t change cabinet committee membership status during session. PM = Prime Minister, MoF = Minister of Finance.
Control variables
Several variables are introduced that could also be expected to have an impact on ministers’ number of laws proposed to parliament. The number of members of each cabinet committee at the beginning of the session is included, as the scrutiny performed by a committee may be less effective when more ministers comprise it. A variable for ministers’ number of laws in the annual law program is also included (see Appendix A). Similar to coalition agreements (see Section 2), we have reason to expect that the number of laws ministers are assigned in the law program will affect the ministers’ number of laws proposed to parliament (cf. Klüver and Zubek, Reference Klüver and Zubek.2018: 725).
Variables for the ministers’ personal characteristics are also included. This includes political experience. As mentioned, a politically experienced minister may be more skilled in the ‘legislative craftmanship’, popularly speaking. Variables are included to control for whether ministers are elected to parliament (dummy), how long they have held their current ministerial posts at the beginning of the session, and whether they have held different ministerial posts earlier in their careers (dummy). A dummy variable is included for the rather few cases in which the minister held two (or more) portfolios simultaneously. Ministers are also coded for whether they are party leaders, as these may have decreased legislative activity due to obligations not related to legislative work. The minister’s gender is also included as a dummy variable (cf. Escobar-Lemmon and Taylor-Robinson, Reference Escobar-Lemmon and Taylor-Robinson.2016).
As a considerable body of literature is concerned with cooperation between parties in coalitions, some variables are included to control for factors related to this. Ministers are coded as to whether they are from the Prime Minister’s party or from a junior party in a coalition (dummy). This may affect a minister’s legislative behavior (see e. g. Green-Pedersen et al., Reference Green-Pedersen, Bjerre Mortensen and So.2018). A dummy variable is also included to control for whether the cabinet is a single-party or coalition cabinet.
For the years studied, Denmark has only had minority governments, except for the parliamentary session of 1993–1994 (arguably, this cabinet’s majority status could be disputed). Cabinets’ parliamentary strength has varied greatly, and as such, the cabinets’ and ministers’ party’s vote share in the latest election is controlled for. The number of ministers in cabinet is also controlled for. Having fewer members in cabinet to share the legislative work with would be expected to positively affect a minister’s legislative activity.
I include ministers’ time on the post during the session as a percentage share of the total session as a control variable. As a robustness check, I run similar models in which this variable is not included, but where cases are weighted according to it (see Appendix B). Dummy variables are also included to control for whether general elections were called during the session and whether one or more parties were included (e.g. 2016) or excluded (e.g. 1996) in/from cabinet during the session. Finally, a variable is included for the total number of cabinet-proposed laws in each session.
Arguably, some variable controlling for ministers’ portfolios would be of interest, as there may be something embedded in the ministry itself that affects the minister’s legislative activity. However, over-time comparisons of portfolios are inherently difficult, and it is beyond the scope of this study to do so in a reliable way. Ministries recurrently merge and split, and portfolios’ content change frequently (see Sieberer et al., Reference Sieberer, Meyer, Bäck, Ceron, Falcó-Gimeno, Guinaudeau, Ejnar Hansen, Kolltveit, Louwerse, Müller and Persson2021). This has also been the case in Denmark for the entire period studied in this article (Sieberer et al., Reference Sieberer, Meyer, Bäck, Ceron, Falcó-Gimeno, Guinaudeau, Ejnar Hansen, Kolltveit, Louwerse, Müller and Persson2021: 779). According to Klüser (Reference Klüser2022: 5), a ‘lacuna’ on data on what tasks ministries are actually assigned with has persisted, and consequently, ‘researchers often still seek to infer ministerial jurisdictions from the corresponding ministerial name tag’. Including a variable based solely on the names of ministries would not be a valid control for portfolio in this study. To mention only one example, there have been considerable differences in what tasks Danish ministers of justice have been assigned, although this particular ministry is one of the few to never change name. Most other ministries change names from time to time, and others again come and go. Further, in a cross-country comparative study including Denmark, Klüser (Reference Klüser2022: 13) found that ‘parties in coalition governments tend to share ministerial responsibility for contentious policy issues that are comparatively salient to them’. As Denmark has experienced many often quite different coalitions over the last forty years, this further complicates the issue of controlling for portfolio in a longitudinal study such as this. Finally, the included variable for ministers’ number of laws in the law program will, to some extent, also function as a proxy for the portfolio (see also Appendix A).
Findings
Table 2 shows negative binomial regression results for determinants of individual ministers’ number of laws proposed to parliament. In Model 1, both variables for cabinet committee membership have a significant effect. Interestingly, with a rate ratio of 0.9, membership of the Coordination Committee decreases a minister’s number of laws proposed to parliament. The 95% confidence interval of the rate ratio is 0.81–1.00, and we can reject hypothesis H1a that membership of this committee will increase ministers’ legislative activity, while the negative effect is somewhat statistically uncertain. There is more statistical certainty regarding the effect of membership of the Economic Committee. Holding all other predictors constant and with a significant rate ratio of 1.26 (95% CI 1,14, 1,41), membership of the Economic Committee is associated with an increase of 26% in the number of laws proposed to parliament. This confirms H1b. With a median of some eight proposed laws per minister per session (Table 1), this denotes a noteworthy increase. When cases are weighted according to ministers’ time on post during the session, similar effects are found (Appendix B). This suggests that committee members use their positions to assign more laws to themselves, or perhaps it is more likely that the members’ proposals are not turned down as often. Considering the rejection of H1a, one explanation could be that the Coordination Committee also aims to focus on broader strategic perspectives (see Rhodes and Salomonsen, Reference Rhodes and Salomonsen.2021: 78; see also Christiansen, Reference Christiansen2021: 150). It also serves as a court of appeal when an agreement can’t be reached in the Economic Committee. The Economic Committee may be given the role of more in-depth scrutiny of ministers’ proposals, not least also in cases of less salient and less party-political conflictual matters.
Table 2 Negative binomial regression. Determinants of ministers’ legislative activity
Negative binomial regression coefficients reported. Standard errors in parenthesis.
*p < 0.10, **p < 0.05, ***p < 0.01.
Reconsidering the descriptives presented in Table 1, the median for a minister’s number of proposed laws is lower than the mean. This suggests that ministers’ number of proposed laws – also for ministers within the cabinet committees – are quite unevenly distributed between the ministers in cabinet. This raises the question of whether the effect of cabinet committee membership applies (equally) to all ministers. To investigate this matter, in Models 2 and 3, respectively, ministers are divided into two groups according to whether they had equal to or less than eight laws in the law program (Model 2) or above (Model 3). This cutoff resembles the median for ministers’ number of laws in law programs (Appendix A) and divides the cases into two (almost) equally sized groups. In both models, the cabinet committee membership variables show similar effects to those in Model 1. The variable for Coordination Committee membership has no significant effect, whereas Economic Committee membership has a highly significant positive effect on ministers’ number of proposed laws. More noteworthy is it that the effect seems to be larger for ministers with fewer laws assigned to them in the law program, suggesting that membership is even more important for ministers with lower legislative activity. The results hold when ministers’ tenure is used as a weight instead of being included in the models as a control variable (Appendix B). Ministers with few laws planned for them may also be more eager to obtain new initiatives, as one surplus law makes a relatively larger difference for that minister than for one with an already extensive number of laws assigned.
Regarding the remaining variables, a few results should be noted.Footnote 10 In Model 1, a minister’s number of laws in the law program has a small but highly significant positive effect on the number of laws the minister proposes to parliament in the succeeding session. In other words, the number of laws a minister is assigned in law programs serves as a valid predictor for the number of laws the minister will propose to parliament in the session to come. Next, only in Model 2 has the cabinet vote share a significant effect on ministers’ number of proposed laws. In neither model has the total number of ministers in cabinet any effect. Therefore, even smaller cabinets with weaker parliamentary support don’t differ significantly in terms of individual ministers’ legislative activity. Considering the vast literature on minority governments, this is an interesting area for more in-depth studies. Finally, considering the extensive literature on coalition governance, the dummy variable for single-party cabinets also provides interesting results. For ministers with fewer laws in the law program, legislative activity is decreased in single-party cabinets, while there is no significant effect of this variable in Models 1 and 3. It may be that coalitions have to delegate some law proposals to smaller ministries in order to maintain a balance between ministers from different parties. Exploring this matter would also be an important venue for future studies.
Robustness check
In Appendix C, as a further robustness check, I propose two alternative measures to explore the effect of cabinet committee membership on ministers’ legislative activity. Logistic regression is applied, as both are categorical variables. First, I coded ministers according to whether they had ten or more percent of the cabinet’s total number of laws proposed to parliament per session. Second, I have coded ministers according to whether they are among the five ministers in cabinet with the most proposed laws. Overall, the logistic regression models support the results of Model 1 in Table 2. Coordination Committee membership has no significant effect in any of the models. Economic Committee membership increases the probability for a minister of being among the five ministers in cabinet who proposes most laws by approximately 300 percent (OR = 3.97, 95% CI = 1.98, 7.96). The odds of a minister proposing ten or more percent of the total cabinet-proposed laws increase by 263 percent (OR = 3.63, 95% CI = 1.65, 7.99) for members of the Economic Committee compared to the rest of the ministers. The coefficients for the Economic Committee variable have lower values when the ministers’ time of a session on the post is used as a weight instead of an independent variable, but the positive effect remains significant.
Conclusion and discussion
Governments comprised of individual ministers overwhelmingly dominate the process of initiating laws in parliamentary democracies. Previously, it has been argued that ministers are unassailable policy dictators, but also that line ministers are largely constrained by an inner core of ministers in the cabinet. Yet, it has remained an unanswered question whether such control mechanisms affect ministers in their work as policy initiators. Consequently, we have needed a better understanding of who dominates policymaking processes in parliamentary democracies. Following previous studies’ claims for future research, I have tried to shed light on rather unexplored themes related to ministers’ role as policymakers and cabinet coordination mechanisms. I used the number of laws proposed to parliament by individual ministers as a measure of ministers’ legislative activity, which should be regarded as a way for ministers to shape public policy, although such analyses have been absent. One advantage of including such a measure of ministers’ legislative activity is that it allows USA wider perspective than the few previous quantitative studies that have focused on (individual) ministers’ influence on policy.
Cabinet committees as cabinet control mechanisms, which are seldom included in empirical quantitative studies, were introduced in the statistical models as determinants of ministers’ legislative activity. The statistical models presented show that even when several other factors are controlled for, membership of the Economic Committee will have a significant positive effect on a minister’s legislative activity. This holds even for ministers who are not assigned a large number of laws in the law program at the beginning of a parliamentary session. This implies that members of the Economic Committee enjoy larger degrees of freedom – or are allowed to be more policy-seeking – than their fellow colleagues in cabinet with regard to proposing laws to parliament. The fact that a similar effect is not found for membership of the Coordination Committee could imply a division of labor between the two committees.
Although this is a single-country study, these results have important implications for our understanding of ministers’ role in parliamentary democracies. They show that hierarchy in cabinet decision-making can affect political outcomes, which has so far remained an open question – not least due to the absence of studies connecting cabinet committees to policy outcomes. This somewhat explorative nature of this study makes it difficult to predict the exact generalizability of the results. But many countries have developed hierarchical decision-making processes in cabinets, and we have no reason to believe that mechanisms set up to coordinate and control ministers’ policymaking in these should not have a similar effect on ministers’ legislative activity. Next, the results show that even though many ministers obtain ministerial offices to shape policy, ministers also from the same party don’t share the same possibility of being policy-seeking. This highlights that important nuances can be missed if we only assume parties as unitary actors in studies of legislative processes and in our attempt to understand who shapes political outcomes in parliamentary democracies. Finally, it should also be stressed that ministers’ often claimed informational advantage in cabinet policymaking doesn’t apply equally to all ministers. As argued, cabinet committees can limit this informational advantage for nonmembers. Giving cabinet committees a crucial coordinating role in cabinet policymaking can be a means to cope with the complexities implied in delegating this advantage to line ministers (cf. König and Lin, Reference König and Lin.2021). However, the presented results suggest that this also reduces nonmember ministers’ opportunities to propose laws.
Extensive cabinet coordination will effectively lead to acentralization of cabinet decision-making processes, thus limiting ministerial autonomy (Andeweg, Reference Andeweg2000: 381). This is not only important for our understanding of political institutions but also for society in general. Recently, an inquiry stated that increasing centralization of decision-making within Danish cabinets – indeed related to the role of cabinet committees – compromised the inclusion of professional and technical advice of the departments in the ministries in the policymaking process (Dybvad-udvalget, 2023: 95). Thus, centralized decision-making processes may negatively affect the quality of laws.
I propose four further avenues for future research. First, I focused only on ministers’ legislative activities as laws proposed to parliament. However, other legislative activities could also be subject to cabinet coordination, for instance, extra appropriations within the framework of existing laws. Exploring this would contribute with more nuance to the results presented in this article. Second, and related to this, a further exploration of the role of law programs, both qualitatively and quantitatively, in cabinet policymaking processes is important, as the results show that they affect ministers’ number of laws proposed to parliament. Such short-term legislative agreements seem to function as a cabinet control mechanism, but they haven’t been thoroughly investigated as such (see, however: Nielsen et al., Reference Nielsen, Juul Christiansen and Dagnis Jensen.2023; see also: Zubek and Klüver, Reference Zubek and Klüver.2015; Klüver and Zubek, Reference Klüver and Zubek.2018). Third, the results didn’t show any effect on ministers’ legislative activity of either party vote share or the dummy for ministers from Prime Ministers’ parties or junior coalition partners. These relationships are especially important for understanding the potential gains for smaller coalition partners in seeking to participate in government. Studies of these factors as determinants of legislative activity would expand our understanding of cabinet coalitions (and minority governments). Fourth, we would also be curious about the number of proposed laws that eventually get passed. Although a high rate of laws proposed by the government in Denmark are passed, there could be interesting differences between the individual ministers for laws passed and not passed (cf. Escobar-Lemmon and Taylor-Robinson, Reference Escobar-Lemmon and Taylor-Robinson.2016).
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S1755773923000310
Acknowledgments
The author wishes to thank Radoslaw Zubek (University of Oxford) for valuable feedback and comments.
Funding
The project is supported by the Carlsberg Foundation.
Competing interests
No competing interests to declare.
Open access
