Lobbying the executive branch: Unpacking access to political heads, political advisers, and civil servants

Abstract This article systematically examines how access of business groups and nongovernmental organizations (NGOs) to the executive branch of the European Union varies across political heads, civil servants, and an understudied yet critical intermediary figure of the executive branch: political advisers. Building upon exchange theory, we argue that the occurrence of a meeting between public officials and interest groups depends on information and legitimacy sought and offered by both types of actors, the public officials’ public exposure, and the interest groups’ lobbying strategies. The empirical analysis is focused on the executive body of the European Union (i.e., the European Commission). Our results show that, while political advisers and civil servants are more likely to meet with business groups than with NGOs, political heads are not biased in favor of any of these two groups.


Introduction
Public officials in the executive branch rely on interest groups' policy input to design, develop, and implement sound and legitimate public policies. Constrained by time and resources, public officials demand high-quality and timely information about the content of public policies. As a consequence, they are expected to interact with a wide set of interest groups to ensure that they are aware of the different perspectives on a policy issue. In addition, the interaction with interest groups may foster the legitimacy of policy processes as it signals to society that all the voices with a stake in the issue at hand have at least the chance to express their concerns to public officials. 1 A central theme for public policy scholars is to understand with whom public officials are more likely to meet in their quest to obtain all the necessary and relevant policy information while enhancing the legitimacy of the process. In that regard, this article extends accumulated knowledge in the field by examining access of interest groups to the different layers within the executive branch. While lobbying research has focused on how interest groups seek to shape policy initiatives with political heads and bureaucrats, this study seeks to disaggregate the executive branch further by also studying the mobilization of interest groups around political advisers.
Public officials, particularly in the administrative side of the executive branch, are more likely to meet with business groups (i.e., business associations and individual corporations) when compared to citizen groups or nongovernmental organizations (NGOs). 2 This is problematic as it presents us with governments that mostly rely on input from business groups, which might cause biased and less legitimate public policies. If public officials are mostly listening to one side of the interest group community and do not facilitate equal access to all groups, the democratic quality of governance systems might be in jeopardy. However, previous research on interest groups either looks at the executive as a whole and compares it to other policy venues, 3 focuses on one of the multiple entry points to this venue, such as expert groups 4 and open consultations, 5 examines the interaction between interest groups and regulatory agencies, 6 or looks at business firms 7 and their characteristics when assessing access, that is, without comparing with other types of interest groups. 8 To deepen our understanding of public officials' interaction with interest groups, we argue that it is necessary to have a more fine-grained distinction of the different types of public officials within the executive branch. More specifically, we pay attention at the internal organizational structure of the executive by not only distinguishing the administrative and political layers, 9 but going beyond it. Whereas political heads-that is, (prime) ministers in most European countries, commissioners in the European Union, or the president and the department secretaries in the United States-set the course of action of public policies, public officials in the administrative side-that is, top civil servants-execute the strategies put forward by political heads and must develop the detailed small print. We go beyond this distinction and make an additional important contribution by including a key intermediary actor in the political-administrative divide: the political advisers. 10 Political advisers are cabinet or personal office members of European Commissioners, members of the Office of US Secretaries, or private office members of UK Secretaries. Positioned in-between political heads and civil servants, political advisers have been mostly overlooked in the interest group and public management literatures. Ultimately, the article addresses the following research question: How does access of business groups and NGOs to the executive branch vary across political heads, political advisers, and top civil servants?
In this article we seek to show that these three actors of the executive branch have different policymaking demands, and thus they will prioritize meeting with different types of interest groups. Whereas civil servants often need technical information to develop policy proposals and ensure its output legitimacy, political heads require political information that reinforce the input legitimacy that enables them to signal to the public their unbiased nature. 11 Based on the incipient literature on political advisers, 12 we expect the latter to demand both political and technical information. In developing the resource-exchange explanation, 13 we argue that public officials' public exposure has an important role in explaining access across the layers of the executive branch. 14 Civil servants outside the public light are more concerned about output legitimacy; in contrast, political heads, which are often in the spotlight and prone to public scrutiny 15 pay more attention to input legitimacy. 16 While political advisers are similar to civil servants in that they are out of the spotlight, they differ from civil servants in that we expect them to seek both types of legitimacy.
Empirically, we build a database consisting of the meetings of 3,359 interest groups with the executive branch of the European Union-European Commission (EC)-between 2014 and 2021. Hence, we analyze the effects of being a business group or an NGO on the likelihood of gaining access to political heads (i.e., president, vice president, and commissioners), political advisers (i.e., appointed by the political head as members of their personal office), and top civil servants (i.e., directors general). We find that whereas political heads are not significantly biased in favor of business groups or NGOs, political advisers and top civil servants are significantly more likely to meet with business groups than with NGOs.
Interest group access to public officials of the executive branch The executive branch is the arena where policy decisions are prepared and implemented. To develop viable policies, public officials in the executive branch need information-both political and technical information-aimed at reinforcing input and output legitimacy. 18 Both business groups and NGOs target the executive to shape the content of public policies. Yet, according to previous research, business groups dominate access to the executive branch in the United States, 19 the European Union 20 and in European countries such as Belgium, 21 Switzerland, 22 and Denmark. 23 However, previous research often considers the executive branch at the aggregate level and compares it with other policy venues such as the media or the legislative arena. 24 Other scholars have studied one of the multiple entry points to the executive, namely, expert groups, 25 online consultations, 26 officials in regulatory agencies, 27 or the political heads of the executive. 28 Consequently, we still need to unpack the executive branch as a whole and systematically assess how access of business groups and NGOs varies across the different types of public officials. 29 We conceptualize the interaction between different public officials in the executive branch with business groups and NGOs as an exchange of information for access. 30 We argue that information and legitimacy needs vary among public officials; as will the information provided and the legitimacy promoted by different interest groups. Furthermore, we further develop exchange theory by considering the public exposure of officials and the strategies of groups, which ultimately are related to whether access occurs.
First, the function of political heads is to lead public policy, set the agenda, determine the strategy, define the policies that need to be passed, and oversee the civil servants. Political heads are appointed to lead the executive branch or one ministry. In this vein, political heads require political information defined as information about the support of core actors, the strength of opposition or support for a specific policy proposal, the framing of specific provisions found in proposed legislation, and the political problems related to implementation processes. 32 Because of their prominent position within the executive, political heads are more visible, and thus frequently subject to public scrutiny and to the media. Whether political heads are affected by reelection or not (in many systems political heads cannot be reelected or do not partake in electoral politics), their reputation and credibility is likely to be vulnerable to the attention of the broad public and are "therefore more exposed to institutional pressure in the social system." 33 As succinctly noted by Holyoke 34 when describing presidential figures, political heads' lives "are subject to a high degree of public scrutiny, [and thus] have a powerful incentive not to become closely identified with organized interests." To avoid being perceived as biased by the public, political heads prioritize input legitimacy through inclusive approaches when interacting with interest groups and thus are more likely to grant equal access to different types of interest groups. In other words, due to their executive tasks, their public exposure, and their input legitimacy needs, political heads are more likely to pay attention to the different voices within the wider public debates. 35 Secondly, the function of civil servants is to provide competence and unbiased implementation in the governance process. 36 The principles and values of top civil servants are different from those entailed by political heads and include political neutrality, hierarchy, competence and expertise. 37 Civil servants' primary responsibilities are to draft regulations based on the guidelines provided by political heads and to enable formulated policies into concrete implementation in conformity with legislative intention. 38 Civil servants should not determine the range or object of public policy 39 but must competently propose policy solutions that are technically sound. As a consequence, top civil servants need technical and scientific knowledge and policy expertise to draft and implement regulations that are only broadly sketched by the political heads of the executive branch. 40 In relation to this, the administrative work of civil servants is normally conducted out of the public sight and is thus less exposed to public scrutiny and media attention. As such, the legitimacy of civil servants rests on the quality of their proposals associated with notions of effectiveness and efficiency, in other words, it is focused on output legitimacy. 41 A third type of public official in the executive branch has drawn attention from scholars as of late: political advisers. 42 By political advisers we refer to individuals in the personal/private office of political heads. While executive structures vary importantly across "Western" systems, political advisers are found (among others) in the private offices of UK senior ministers, the office of US secretaries, and the cabinets (personal offices) of EU Commissioners.
Core executive studies are starting to explore how this "third element" operates. 43 Political advisers are appointed to serve an individual political head, their position is temporary, and they are recruited on the basis of political criteria 44 and, at least in part, because of their loyalty to the political head and 32 Yackee (2012). 33 Goodstein (1994); Oliver (1991 Demir and Nyhan (2008). 39 Demir (2009). 40 Yackee (2012, 377). 41 Coen and Katsaitis (2019); Schmidt and Wood (2019). 42 Craft (2015). 43 Shaw and Eichbaum (2018). 44 Hustedt et al. (2017). commitment to its political agenda. 45 Importantly, political heads increasingly surround themselves with advisers. 46 A typical minister in the 1950s had perhaps three advisers, while today the numbers is around 10. 47 Moreover, scholars argue that ministerial political advisers are gaining influence in governments around the world. 48 One key function of political advisers is to help bridge the political head with civil servants and other external stakeholders. 49 Political advisers of the EC communicate with external interests and with civil servants on behalf of the political head, while minding the political opportunities and threats to the political heads. 50 In that regard, political advisers act as institutional buffers, protecting civil servants against political demands from political heads and enabling them "to concentrate on the provision of neutral competence." 51 Because of their intermediate position, 52 political advisers are expected to need both political information and policy expertise. In other words, political advisers should be able to interpret technical knowledge and communicate this to their political heads while simultaneously paying attention to political development and political information that might shape the nature and the technicalities related to the legislation. In relation to their legitimacy concerns, political advisers are less subject to public exposure and thus they interact with different types of groups without having to worry about the input legitimacy implications of these meetings. Yet, a key function of political advisers is to act as bodyguard of their political heads, which might lead them toward a more balanced interaction with both business groups and NGOs.

Supply side: Access goods of business groups and NGOs
On the other side of this relationship, we find interest groups that pursue different types of strategies and possess different types of information that affect their ability to reinforce the legitimacies of the policy process. The literature has developed and tested theoretical arguments that consider group type as an empirically and normatively important determinant of access to political arenas. 53 More specifically, the distinction normally boils down to business groups versus NGOs. The business groups category includes firms, professional associations, and associations of firms organized at a national or supranational level (e.g., Google, the European Bicycle Manufacturers Association, the Irish Cattle and Sheep Farmers' Association, or Business Europe). NGOs, in contrast, represent the interests or concerns of citizens, such as environmental or consumer concerns (e.g., Human Rights Watch, Greenpeace Europe, the Federation of Swedish Family Forest Owners, or the Norwegian Consumer Council).
Business groups represent specific interests and a well-defined constituency, and thus face fewer hurdles related to their ability to mobilize and lobby. 54 Business groups have been traditionally associated with having technical expertise and specialized information that they obtain in their everyday activities. 55 At the aggregate level, business groups (either individual firms or associations of companies) are expected to possess valuable expertise that is critical for the success and effectiveness of public policies. 56 As highlighted by Yackee and Yackee, 57 business interests are able to provide public officials with high-quality input for three reasons: They are more capable of understanding industrial data, have their own scientifically sound technical data, and are more likely to have lawyers, lobbyists, or experts 45 Haeder and Yackee (2015); Lewis (2008). 46 Shaw and Eichbaum (2018). 47 Robson (2015). 48 Hustedt et al. (2017). 49 Craft (2013); Robson (2015). 50 Gouglas et al. (2017). 51 Shaw and Eichbaum (2020). who are trained in drafting convincing arguments. As a consequence, and as noted by Kanol 58 "research in Europe shows that administrative actors need technical expertise more than political support, so they grant access to groups which possess such expertise. Sectional groups, and especially business groups, possess such expertise more than public interest groups." In terms of legitimacy, and mostly due to their presumed possession of technical expertise, business groups can provide public officials with output legitimacy. 59 Another important characteristic of business groups' lobbying activities is that they tend to avoid conflict and politicization of policy issues. 60 Thus, they are expected to use inconspicuous mechanisms that do not attract media attention and that do not create or expand overt conflict. 61 In summary, business groups are more likely to access the technical rule-writing civil service, who seek expertise and technical knowledge and operate out of the public spotlight. 62 We expect this bias to be less pronounced with political advisers because of their political nature and reliance on their political heads-irrespective of their low public exposure. Last, because political heads are the most exposed figures within the executive, business groups will temper their access to these public officials. 63 NGOs, in contrast, often represent interests that are part of the public domain and incentivized by public support expressed through their membership-base. 64 At the aggregate level, NGOs are more likely to specialize in collecting and providing political information that is relevant to their cause. In doing so, NGOs tend to be better positioned to reinforce the input legitimacy of policy processes because they claim to represent the public interest as opposed to a business's particular economic interests. As empirically shown by Flöthe,65 NGOs offer information on public preferences more often than business groups. NGOs, therefore, add to the breadth and participatory character of the decision-making process: the legitimacy they bolster is input oriented. 66 Moreover, NGOs often rely on outside lobbying strategies that entail the mobilization of their membership base and seek conflict expansion. Importantly, NGOs struggle to maintain their members, which are crucial to ensure organizational survival. 67 So, seeking conflict expansion though outside lobbying strategies "provides publicity to citizen groups among a broader audience, which may ensure the survival of the organization." 68 Consequently, NGOs are expected to target public officials at the top of the hierarchy within the executive branch-the most visible ones. By doing so, they are expected to effectively publicize their actions among their membership base and expand the conflict. In contrast, NGOs will be less interested in reaching out to political advisers and civil servants due to their low public exposure.
Based on the discussion in the previous sections, we draw the following hypotheses: H1: Top civil servants within the executive branch are more likely to meet with business groups than with NGOs H2: Political advisers within the executive branch will equally meet with business groups and NGOs In this article we aggregate individual firms and business associations under the same label: business groups. However, we acknowledge that some of the factors presented in the preceding text might vary across firms and business associations. More specifically, business associations might be more willing than firms to engage in political battles and provide political information about their membership base. Nonetheless, for the sake of parsimony, this article exclusively addresses the distinction between NGOs and business groups. 62 Binderkrantz and Christiansen (2015); Fraussen et al. (2015); Haeder and Yackee (2015); Yackee and Yackee (2006). 63 Keller (2016). 64 Coen and Katsaitis (2019). H3: Political heads within the executive branch are more likely to meet with NGOs than with business groups

Research design
In this study we tackle the question how does access of business groups and NGOs to the executive branch vary across political heads, political advisers, and top civil servants? by quantitatively examining the meetings between interest groups and public officials of the EC.

The executive context: European Commission
To study how interest groups access varies across levels of the executive branch, we examine the meetings of interest groups with EC officials. The EC is the main executive power of the European Union and provides the empirical setting to test our hypothesis. It proposes legislation, manages the European Union's day-to-day business and budget, enforces the rules, and negotiates international trade agreements on behalf of the European Union. The EC drafts, enforces and monitors EU laws, but it does not pass laws (Klüver, 2013). As noted by Broscheid and Coen,69 the key to successful lobbying in the EC is not political patronage but the provision of information and legitimacy. The EC, with "its executive instruments and directives, acts as the focal point in the early stages of the lobbying process." 70 In general, the EC needs both technical and political information and legitimacy, in particular due to its indirect connection with citizens. The EC explicitly states that it seeks to have open, transparent, and equitable consultation and engagement processes that facilitate policy interaction with any group with a stake in the issue under discussion. 71 More importantly for the purpose of this article, the EC has a clear political-administrative division. First, at the top of the hierarchy we find the president and vice presidents of the EC as well as the commissioners (i.e., political heads). These political heads "take decisions on the Commission's political and strategic direction." 72 Although commissioners do not have a direct reelection motive with the electorate, these political figures still need to protect their reputation and credibility vis-à-vis their national governments and the EU population in general. At the administrative level we find the Directorates General (DGs), which are formed by employees of the European Civil Service and are responsible for technical policy preparation and implementation oversight. DGs prepare proposals for their commissioners, which can then be put forward for adoption in the college of commissioners. The DGs are headed by a top civil servant in charge of these administrative departments. Lastly, there is an intermediate position called the cabinet of the commissioner, which is staffed with political advisers. The cabinet should be understood as a personal support office of the political head. Each commissioner has six cabinet members (i.e., political advisers): two of them must be women, no more than three can have the same nationality as the Commissioners, three members need to be civil servants of the EC, and all of them must reflect the European Union's regional diversity. Together, they have the responsibility of giving political guidance to the commissioner. 73 The EC is an excellent case study because of the neat distinction between the three types of public officials we are interested in. Unlike, for example, the US executive (and to a lesser extent the UK government), the EC has no intermediate political appointees (beyond the personal office) between the commissioner and the civil service. The United States has many political appointees between the secretary and civil service-in addition to the personnel in the Office of the Secretary-such as undersecretaries and deputy secretaries. This is also the case in the United Kingdom-with secretaries, and senior and junior ministers-and in many other national governments.

The data
Our empirical analysis relies on data from the Transparency Register. 74 More specifically, since 2014 public officials of the EC are requested to publish their direct meetings with external stakeholders. 75 This data is made available through Transparency Register, together with a set of additional interest group-level variables. The dataset includes 3,359 interest groups (either business groups or NGOs) that have had a meeting with one EC official between November 12, 2014, and February 25, 2021. This six-year data-frame enables us to overcome potential bias due to policy and issue cycles that might be inherent in cross-sectional data.
The dataset is limited to top public officials (i.e., president, vice presidents, commissioners, commissioner cabinet advisers, and directors general), thus our analysis covers the most visible access patterns in direct meetings. Regarding the quality and reliability of Transparency Register data, Greenwood and Dreger 76 showed that the Transparency Register had reached a satisfactory level given the current high standards of data entry among the different interest groups. We do not rule out the possibility that EC officials also meet with organizations outside the commission and thus do not report this in the register. In that regard, both our theoretical assumptions and our empirical analyses are focused on the public side of access, which may vary if we examine the "nondisclosed" or off-the-record interaction between interest groups and public officials.

Dependent, explanatory, and control variables
Our main dependent variable is access to the different officials of the EC. Access is present when an interest group has "entered a political arena (parliament, administration or media) passing a threshold controlled by relevant gatekeepers (politicians, civil servants or journalists)." 77 In other words, not all groups have access to a political arena as it is something that must be won or granted. 78 In this article we measure access to three different types of public officials: one to capture access to top civil servants, another one focused on access to political advisers, and a third one measuring access to political heads. In all three instances, access is a binary variable indicating whether groups had access to top civil servants (i.e., directors general), political advisers (i.e., cabinet members), and political heads (i.e., president, vice president, and commissioners). 79 Regarding our explanatory variable, we operationalize group type following the broad terminology of the Transparency Register. More specifically, we refer to business groups when the group is categorized as "[corporate] in-house lobbyists or trade/professional associations" by the Transparency Register, and we code the group as an NGO when it categorizes the group as "non-governmental organizations." 80 We include six control variables: volume of resources, scope, whether or not the group has supranational interests, whether organizations have their headquarters in Brussels (where the EC is housed), the years in the Transparency Register, and whether the group has been part of an expert group of the EC in the last years. First, mobilizing, surviving, and gaining access to public officials is neither easy nor cheap as it requires the generation of valuable information. In this vein, we expect that groups 74 http://ec.europa.eu/transparency/regexpert/index.cfm. The data of the meetings with different public officials of the EC has been structured by Integrity Watch (see http://intergritywatch.eu).  Binderkrantz and Pedersen (2017, 307). 78 Halpin and Fraussen (2017, 726). This article focuses on "access" rather than "level of access" because the intention is to examine the extent to which different public officials meet with a more or less diverse set of interest groups, rather than capturing which are the groups that obtained a privileged position among public officials (Maloney et al., 1994). 79 Whereas top civil servants and commissioners are one single person, political advisers are group 4-6 officials. Although this may affect the "accessibility" of groups to this layer of the executive. It is still relevant to examine and compare this to the other layers because our focus is on whether or not there is bias in favor of business groups. 80 A calculation of Krippendorff Alpha for 50 randomly selected observations, coded by one of the authors in terms of business and citizen group, gives a value of 0.98, showing a high level of reliability. endowed with more resources are more likely to gain access to public officials across the executive branch. 81 Second, scope refers to the number of fields of interest on which the groups are active. This relates to the idea of generalist versus niche organizations. Whereas generalists are expected to be more relevant for political heads, niche-oriented groups (with detailed expertise on specific policy issues) are more likely to gain access among civil servants and political advisers. Third, supranationality is a binary variable where groups self-report whether their interests are restricted to subnational/ national issues or also cover EU/supranational ones. Given the EU nature of the policies developed by the EC, we expect that supranational groups are more likely to access public officials across the executive layers. Fourth, having a government affairs headquarters in Brussels is operationalized as a dichotomous variable. Being close to the decision-making center is expected to facilitate access to public officials at the different levels of the executive. 82 Fifth, we control for the number of years an organization has been registered in the Transparency Register. Our argument here is that those organizations that have been registered for longer, are more likely to be known by public officials, and thus might experience less hurdles to gain access. Lastly, we also control for whether the organizations have participated in one of the expert groups of the EC. Groups that have done so are expected to be particularly relevant for civil servants in need of technical information. 83 Tables A1 and A2 in the Appendix summarize the key descriptive statistics of all the variables introduced and how they correlate among themselves.

Results
In this section we first present the main descriptive statistics and a bivariate analysis of our dependent variables with the explanatory factors to then move to a multivariate analysis. Before moving to our main analyses, we explore which groups are more likely to gain access to the commission. That is, if we consider the full population of interest groups registered in the Transparency Register (n = 10,243), we observe a clear bias in favor of business groups, both at the commission level, and at each of the layers of the executive branch. Nonetheless, the standardized coefficient of the three logistic regressions already points toward a trend, namely that business groups are more likely to gain access to civil servants, followed by cabinet members and, the lowest coefficient is for access to political heads (see Table A3 in the Appendix). This finding is not surprising considering the portfolio of the European Union and its regulatory nature. However, many groups in the Transparency Register do not seek access to the Commission and focus on other EU institutions such as the parliament or the council. 84 In that regard, our analyses focus on groups that have accessed the commission, enabling us to study the dynamics more neatly between different types of groups and the three layers of the executive branch.
At a general level, we observe that the EC is biased in favor of business groups since 2,305 out of the 3,359 groups with access (i.e., 68.62%) are business groups. Yet, this distribution varies across the layers of the executive-note that a given interest group can appear in more than one of the layers within the EC. Firstly, out of the 1,017 groups with access to top civil servants 76.3% were business groups (X 2 (1, N = 3,359) = 39.968, p = 0.000). Secondly, 70.55% of the 2,659 groups with access to political heads' advisers were business groups (X 2 (1, N = 3,359) = 22.101, p = 0.000). Lastly, 69.66% of the 1,869 that met with political heads were business groups (X 2 (1, N = 3,359) = 2.122, p = 0.145). That is, aligned with our expectations, we observe that civil servants are the ones that more heavily rely on input from business groups, whereas political heads report a very similar distribution to the one found at the EC level. If we look at the significance levels of the chi-square analyses, we can only confirm our H1. H2 is not confirmed because we still observe a bias in favor of business at the political advisers' level. Regarding H3, whereas we expected a higher incidence of NGOs among political heads, we find that access of business groups and NGOs is not significantly different at this layer. 81 Grant (2001).  Table 1 presents the results of our multivariate analysis. Due to the dichotomous nature of our dependent variables, we opted for three logistic regression models, one for each type of public official. When controlling for several variables, business groups are more likely to access top civil servantsconfirming H1. Yet, H2 is not supported because business groups have significantly higher access to political advisers than do NGOs. However, the odds ratio of business groups gaining access is higher in Model 1 than in Model 2; this is at least aligned with the logic of our expectations, where a business group is more likely to meet with top civil servants than with political advisers. Last, H3 is not confirmed as political heads are neither significantly biased toward NGOs nor business groups. The nonsignificant p-value of Model 3 indicates that the likelihood of gaining access to political heads is very similar for business groups and NGOs. Figure 1 presents the predicted probability that business groups and NGOs have of gaining access across the three layers of the executive branch-relying on the models of Table 1. As shown, business groups have a higher predicted probability of gaining access to the three types of public officials. However, we observe important differences. Firstly, the highest change in the predicted probabilities between NGOs and business groups is found among civil servants (23.5% for NGOs and 33.3% for business groups). Secondly, the highest predicted probabilities are found at the political adviser level because of the higher number of groups with access to this venue. Although we also observe the bias in favor of business groups among political advisers, the difference between the predicted probabilities is not as acute as in the civil service level (i.e., 75.5% for NGOs and 80.9% for business groups). Lastly, although business groups have a higher predicted probability of gaining access to political heads, this is clearly not significant (i.e., 55.4% for NGOs and 55.8% for business groups).
Regarding the control variables, the resources of an interest group is significantly and positively related to access in each of the models, confirming previous studies that find that resources matter when it comes to gaining access. 85 As expected, and based on the value of the odds ratio, the scope is particularly relevant to gain access to political heads, although this variable is also positively and significantly related to both civil servant and political advisers. The supranationality of the groups is significantly and positively related to meeting with both civil servants and political advisers' level.  Dür and Mateo (2013).
Surprisingly, this variable is not significant among political heads, meaning that they are less attentive at whether groups have supranational interests or not. Having a Belgian headquarter is not related to top civil servants, yet it is positively related with having access to political advisers. Interestingly, we find an inversed relationship with political heads, that is, having the headquarters in Belgium is negatively related to the likelihood of gaining access to political heads. That is, the relational capital derived from having an office close to where the public officials of the executive branch work, is beneficial to gain access to political advisers, but it seems to penalize access to political heads. The number of years in the Transparency register seems to increase the likelihood of access to the different types of public officials. Lastly, having been member of an expert group of the EC is particularly relevant for gaining access to top civil servants and, to a lesser extent, to political heads-but it does not matter to interact with political advisers.

Discussion and conclusions
The interaction between public officials in the executive branch with interest groups has attracted scholarly attention in the last years. The general consensus so far is that public officials in the executive more frequently meet with business groups than with NGOs. However, there are good reasons to think that the executive is not a homogenous branch. The different organizational characteristics of public officials across the executives combined with the different preferences of interest groups lead to unequal government interactions with business groups and NGOs. In this article, we look in detail into the executive branch and extend supply-demand exchange theory of lobbying, 86 by considering interest groups' nature and public officials' public exposure. More specifically, the article makes three important contributions: First, we unpack the executive branch and conceptually distinguish three relevant layers within this venue that is often considered as a homogenous entity. In doing so, we examine the role of political advisers, an increasingly relevant figure that has been mostly neglected in the field of interest group studies. Second, we theorize on the relationship between public officials at each layer of the executive branch and different types of interest groups by relying on exchange theory, yet accounting for the public exposure of officials. Third, we empirically demonstrate that there is value in unpacking the executive branch because the levels of bias in favor of business groups vary significantly. Aligned with the general consensus, this article finds that, at the aggregate level, public officials of the EC meet with more business groups than with NGOs. However, a detailed analysis into the meetings with different actors within the executive branch of the European Union sheds a more nuanced story where the politics-administrative distinction and the political advisers play a crucial role. Supporting H1, the results indicate that civil servants meet more often with business. This is consistent with the former's need of technical information for rule writing and with the latter's preference for out-of-the-spotlight lobbying. Regarding H3, political heads meet equally with different types of groups. 87 This is aligned with the idea that political officials seek an allure of neutrality 88 to underscore their input legitimacy 89 and with NGOs' preference for outside strategies. 90 An important finding falsifying our H2 is that political advisers meet more with business groups than with NGOs. Importantly, we propose that exposure and visibility of political actors in the executive branch explains this difference. Compared to accessing the political head directly, the lower visibility of political advisers (as opposed to the high visibility of the political head) make them valuable targets for corporates who prefer operating away from the lights. 91 As noted by Haeder and Yackee 92 political advisers represent an attractive alternative option for group who want to voice their concerns to the higher levels but want to avoid public exposure and also the hurdles they may face when seeking access to top political figures. Yackee 93 found that US business interests and public officials value the lower visibility of ex parte (i.e., off-the-record) meetings that occur during the preproposal regulatory stage. In short, political advisers can be seen as a substitute for business groups that seek to shape political heads' view on the policy issue but that want to avoid public exposure. 94 From the demand-side perspective, one potential explanation of this findings is that political heads use their less visible political advisers to grant "behind the scenes" access to business groups. In doing so they can still present themselves as "neutral" agents that equally listen to different societal groups. That is, the null-finding related to political heads and the strong bias found for political advisers might be endogenous to the strategy of political heads.
We recognize that while exploring the EC is one of this study's strengths, adding a comparative element to the existing body of US literature and exploiting the EC's clear differentiation between political heads, political advisers, and civil servants, this context is also a limitation. Two idiosyncrasies of the EC limit the generalizability of our findings. First, the EC is a particular executive arena, that has a lower level of political partisanship, it is highly technical, and it is entrusted with many regulatory powers, 95 where civil servants are in their posts for long periods and political heads do not seek reelection. 96 Second, political heads have to comply with some conditions when appointing their political advisers, and this might change in national context where ministers often have full discretion when nominating their advisers. Another limitation that requires more attention in the future is the noncontextual data used to conduct the analyses. Previous research has shown that access and influence are shaped by the nature of the issue (i.e., complexity, salience, and conflict). 97 To further assess which interest groups gain access at each level of the commission, it is necessary to account for the nature of the policy issue under discussion, not only the policy area 98 but also the conflict, complexity, and salience of the issue at hand. In addition to this, our hypotheses are based on the assumption that business groups often have technical information, whereas NGOs, at the aggregate level, have more political information. Although this observation has been confirmed in the past, we acknowledge that there might be important variation within groups, which is why we call for research that disentangles how interest groups are internally organized 99 and link it to the type of information they possess. 100 Last, our theoretical framing proposes that both supply and demand explain the pattern in meetings observed-and indeed our research design cannot disentangle causality. Access is simultaneously gained and granted in our framing. This article argues that the administrative-political divide and more specifically the different types of public officials within the executive branch leads to unequal 88 Holyoke (2004). 89 Coen and Katsaitis (2019). 90 Dür and Mateo (2013).  Weiler and Brändli (2015). 95 Kassim et al. (2013). 96 Coen et al. (2021). 97 Klüver et al. (2015). 98 Coen and Katsaitis (2013); Vikberg (2020). 99 Albareda (2018, 2020). 100 Flöthe (2019). access of business groups and NGOs. In that regard, future research examining the relationship between interest groups and public officials could move beyond resource-exchange explanations by considering two additional elements: (structural) power of interest groups 101 and behavioral dynamics of public officials. 102 In conclusions, our study has important implications for public officials and interest groups. Firstly, particularly political advisers and top civil servants should consider the biases they are part of, and either remedy these or at a minimum be keenly aware of the overrepresentation of some stakeholders and the potential biases this might generate for final policies. Secondly, our study signals that business groups who stay away from public spotlight particularly target civil servants and (more importantly) political advisers. In that regard, NGOs could also expand their lobbying strategies and seek access to less visible venues such as political advisers that, as political guiders of their heads, may have an important say on final policy issues. Ultimately, our findings call for further studies examining the executive branch not as a monolithic entity, but as a multilayered organization that includes a diverse set of public officials with unequal agendas, needs, and priorities that affect their interactions with interest groups.