The President Will See Whom Now? Presidential Engagement with Organized Interests

Abstract

Though presidents often criticize organized interests, presidents also expend considerable effort engaging them. Using original elite interviews, a survey of lobbyists, and administrative data, I consider how this engagement manifests, why presidents engage interests, and with which interests presidents engage. Unlike in other institutions, presidents exercise substantial control over engagement with interests, and they engage to mobilize interests’ institutional resources in service of their goals. To optimize mobilization, presidents focus engagement on well-resourced interests and interests who share presidents’ preferences. Pairing over seven million White House visitor log entries from two administrations with lobbying and campaign finance records, I demonstrate that presidential engagement is informed by interests’ electoral and policy resources and partisan alignment, though these characteristics’ substantive effects are modest. My findings highlight coalition building with interests as an underappreciated source of presidential power and elucidate the degree to which presidents amplify the political voice of well-resourced and copartisan interests.

Though polar opposites in most respects, both Presidents Barack Obama and Donald Trump routinely assailed organized interests. ¹ Early in his presidency, Obama declared at a bill signing, “When I ran for President, I did so because I believed that despite … the influence of special interests, it was possible for us to bring change to Washington.” ² Also early in his term, Trump proclaimed he “want[ed] to get the special interests out of politics for good.” ³ Both also took action to limit the power of interests in their administrations, such as issuing executive orders restricting former lobbyists’ ability to serve in government. ⁴ Scholars suggest Trump, Obama, and other presidents express antipathy toward and distance themselves from interests not because of personal inclinations but as a consequence of the office’s isolation from the organized interest universe. Howell and Moe posit this isolation emerges from institutional design, as presidents’ “position of national leadership gives them far more freedom from special interest pressure … than their legislative counterparts” (2016, 102; see also Quirk and Nesmith 2005; Truman 1971). Furthermore, Light suggests this isolation stems from presidents’ deliberate choices, describing “a conscious effort [by the White House] to avoid interaction with most groups” (1999, 94). Consequently, scholars afford “far less attention” to presidents’ supposedly limited interactions with interests than to legislatures’ and executive agencies’ interactions with them (Loomis 2009, 403).

However, a closer look at the White House’s day-to-day activities reveals that presidents frequently engage with organized interests. ⁵ Although salient instances of engagement, such as President Bill Clinton providing donors nights in the Lincoln Bedroom, ⁶ often attract criticism, such anecdotes belie the White House’s commonplace engagement with interests in public and private settings. Recounting his early days as president, Barack Obama recalls “an endless flow of meetings with various constituency groups … to address their concerns and solicit their support” (Obama 2020, 285–6). Similarly, an aide to President Jimmy Carter reports the administration “tried to keep up very good relations with all the major interest group players in town … meeting with [them] … on a regular basis.” ⁷ The ubiquity of such engagement is exemplified by the longevity of the White House Office of Public Liaison, a unit maintained by every president since Gerald Ford responsible for cultivating relationships with interests (Peterson 1992; Pika 2009). ⁸ Though the White House may distance itself from interests in public-facing behavior, an aide to President John F. Kennedy admits, “[I]nterest groups are in the woodwork, under the floors, in the hallways, and in the rose garden” (Light 1999, 95). Highlighting this discrepancy between the prominence of interests in the White House and the “episodic and limited” attention scholars have paid to presidents’ interactions with interests, Loomis notes in the Oxford Handbook of the American Presidency that “such a gap is remarkable” (2009, 404–5).

This paper begins to address this gap by probing with which organized interests presidents engage. In doing so, this paper also considers two antecedent questions: to what extent do presidents or interests exercise control over engagement, and what motivations drive engagement? Although these questions remain underexplored in part because scholars have deemphasized the linkage between presidents and interests, they have also been overlooked because “the requisite data are difficult to amass” (Loomis 2009, 421). These questions require data on both the dynamics by which engagement between presidents and interests manifests and the occurrence of engagement; however, these phenomena are seldom observable.

I investigate these questions using original interviews, survey responses, and administrative data shedding light on presidents’ engagement with organized interests. First, I consider the degree to which presidents and interests exert influence over engagement with insights from 15 interviews with former White House officials and interest representatives and a survey of over 700 lobbyists. Unlike the traditional “outside-in” lobbying model, where interests pursue access to policy makers, my interviews and survey responses demonstrate that the relative prominence of an “inside-out” model, where presidents take a leading role in providing interests access, affords presidents substantial control over engagement (Shaiko 1998; Tenpas 2005). Second, I describe presidents’ motivations for engaging with interests and provide expectations for which interests they are likely to engage. In brief, interests’ institutional resources, such as lobbying capabilities and campaign contributions, make them important targets of presidents’ coalition-building efforts; through engagement, presidents hope to mobilize interest support or discourage opposition. To maximize the interest resources they marshal, presidents focus engagement on interests offering the highest expected rates of return—well-resourced interests and those sharing presidents’ preferences. Third, I evaluate these expectations using over seven million White House visitor log entries from the Clinton and Obama administrations and contemporary lobbying records to identify instances where presidents engaged with interests and model engagement as a function of interests’ electoral and policy resources and partisan alignment. My findings indicate presidents are more likely to engage with interests with larger resource endowments and who are copartisans, though the magnitudes of these effects are modest, as even low-resource and noncopartisan interests experience engagement at substantial rates.

My findings highlight the importance of presidents’ interactions with organized interests and contribute to key themes in the study of the presidency, organized interests, and representation. Although recent studies of presidential power focus on tools presidents wield independently, such as unilateral action (Lowande and Rogowski 2021), disbursement of selective benefits (Kriner and Reeves 2015), and appointments outside the advise and consent process (Kinane 2021), this study reemphasizes that presidents’ toolkits also include their ability to build and mobilize coalitions. Furthermore, my findings address the debate concerning presidential representation by considering representational priorities presidents exhibit through engaging interests, who are both independent actors and manifestations of preferences in the mass public. By engaging more with well-resourced and copartisan interests, my results suggest presidents provide more representation to subgroups that help advance their goals (Druckman and Jacobs 2015; Kriner and Reeves 2015) rather than serving as national representatives (Howell and Moe 2016). Finally, my findings illuminate presidents’ role in fostering or tempering the outsized voice upper-class interests enjoy in American politics. Although presidents, contrary to their role as national representatives (Bentley 1908; Quirk and Nesmith 2005; Truman 1971), engage more often with high-resource interests aligned with the upper class, this preference is modest, especially when compared with that of other institutions like Congress.

Presidents as Engagers-in-Chief

Presidents and organized interests interact in many well-documented ways, such as interests contributing to presidents’ campaigns and altering their activities in response to presidents’ priorities (Baumgartner et al. 2011). I focus on a specific type of interaction I term “presidential engagement,” or reciprocal communication and coordination between presidents and interests concerning electoral or policy goals. A key distinguishing trait of presidential engagement is that it requires active participation from both actors; for instance, although interests can expend campaign resources in support of presidents unilaterally, engagement requires dialogue between presidents and interests. Though presidential engagement can take place through many mediums of direct contact, such as phone calls and emails, my theoretical exposition and empirical analysis focus on in-person White House meetings because they are a valued form of interaction for both parties and because the significant time and effort the White House expends to facilitate them provides a strong signal of its engagement priorities. ⁹

Framing these interactions as presidential engagement presupposes presidents exert critical influence over their manifestation and conduct. This assertion diverges from the commonly understood “outside-in” lobbying dynamic characterizing other institutions, such as Congress, where interests’ motives for and strategic behavior to gain access and influence take precedence and policy makers are mere targets of interests’ pursuits (Austen-Smith and Wright 1994; Hall and Deardorff 2006; Miller 2022a). However, several descriptive accounts assert presidents’ motivations and strategic choices feature prominently in interactions with interests, particularly when compared with those of other policy makers, fostering an “inside-out” lobbying dynamic where presidents exercise substantial control (Peterson 1992; Shaiko 1998; Tenpas 2005). Unfortunately, limited systematic evidence exists demonstrating presidents exert more control over interactions with interests than other political elites. Substantiating this dynamic is important for my theoretical argument because presidents must wield considerable power over engagement for their incentives to influence with which interests they engage.

In this section, I augment extant descriptive studies with original interviews and survey responses from organized interest representatives and White House officials to demonstrate presidents wield important control over engagement (Miller 2022b). My interviews, conducted with 15 interest representatives and former White House officials between 2018 and 2019, provide detailed insights on the dynamics governing engagement from actors on both sides of the relationship. ¹⁰ My survey responses, collected from over 700 lobbyists in 2018, illustrate how these insights generalize to the broader population of interests. ¹¹ These novel data provide a window into typically unobservable mechanisms of elite decision-making surrounding the White House’s interactions with interests that provide presidents important control over engagement.

Conditions Empowering Presidents

Explanations for presidents’ relatively strong control over engagement fall into three categories. First, physical and logistical barriers make outside-in lobbying difficult for organized interests while providing the White House control over access (Peterson 1992). One important distinction between the White House and other policy-making venues is restrictions on physical access; whereas interests can easily enter Congress and many federal agencies to attend events or visit unannounced, White House access requires prior clearance. These barriers not only create distance between interests and the White House but also impose transaction costs on White House staff, who must exert time and effort to collect and submit visitors’ personal information to the Secret Service in advance; thus, the White House is disinclined to provide access. As one former White House official describes, “Going to a meeting at the White House is not trivial. You have to go through security, you have to get cleared in. There’s … care given to who is coming into the building” (Interviewee A).

Second, the White House faces stronger time and resource constraints on its ability to interact with interests than other institutions. Although all policy makers shoulder important responsibilities, presidents, as head of the executive branch and the sole nationally elected official, face an unceasing flow of demands for action that exceeds their powers and capabilities (Cronin 1980). Furthermore, the small size of the White House’s staff relative to those of the 535 members of Congress and hundreds of federal agencies leaves presidents little capacity for interacting with the thousands of organizations in the organized interest universe (Schlozman, Verba, and Brady 2012). ¹² Given a choice between fighting for a sliver of the White House’s attention or pursuing more accessible policy makers, interests tend to forsake the White House and focus outside-in lobbying on congresspersons and bureaucrats (Loomis 2009). As an aide to President Lyndon Baines Johnson explains, “There are 535 opportunities in Congress and only one in the White House. You get an hour to present your case before each representative; you get fifteen minutes once a year with the president. Where would you put your effort?” (Light 1999, 94). Because interests shift their outside-in lobbying effort away from the White House, presidents inherit considerable discretion over initiating engagement.

Third, the White House has responded to the challenges of interacting with interests with institutional innovations that increase its authority over engagement. The most prominent such innovation is the Office of Public Liaison (OPL), a White House unit created during the Ford administration to manage presidents’ relationships with interests (Pika 2009). An Obama administration staffer who worked in this unit described it as “the door to the White House… our job was to work with organized interests all the time, every day. Our role was to represent the President for these interests” (Interviewee A). Importantly, OPL functions as the White House’s primary point of contact with interests, centralizing engagement and serving as a clearinghouse for information flowing to and from interests. Although this helps the White House engage efficiently, it stymies interests’ outside-in lobbying to reach officials beyond OPL, leading them to lobby other venues where policy makers are more reachable (Interviewee B). Through institutional innovations like OPL, the White House has accumulated control over engagement (Kumar and Grossman 1984, 308).

Evidence of Presidential Control over Engagement

Because presidential engagement is the product of behind-closed-doors actions by myriad actors inside and outside the White House, marshaling evidence illustrating every dimension of the process is difficult. However, by collecting insights from political elites on both sides of the relationship, I can evaluate the veracity of several claims central to the argument that, relative to other institutions, the White House exercises substantial control over engagement. Specifically, I use my interviews and survey responses to show that interests focus more lobbying attention on Congress and the bureaucracy than the White House, the White House takes an active role in initiating engagement, and interests are more responsive to the White House than vice versa. Although this evidence is not definitive proof of White House control over engagement, its alignment with extant descriptive accounts affirms the argument that presidents’ motivations are critical to engagement.

First, I assess whether organized interests allocate less lobbying attention to the White House, thereby enabling presidents to exert more influence over engagement. In Figure 1, I present the distributions of lobbyists’ responses to questions asking how important direct contacts with officials in Congress, the federal bureaucracy, and the White House are to their lobbying strategies. The distributions for Congress and the bureaucracy accord with traditional outside-in lobbying accounts, with most respondents indicating direct contacts are “very” or “extremely” important (90.9% and 82.0%, respectively). However, only about a third of respondents (38.8%) assigned similar importance to direct contacts with the White House. Because interests are rational actors who allocate resources given expectations of success, this disconnect implies they devote more effort to obtaining direct contacts with Congress and the bureaucracy than the White House. Consequently, presidents must exercise greater initiative to engage interests focusing their access-seeking behavior elsewhere.

Figure 1. Importance of Direct Contacts for Organized Interests across Institutions

Note: Bar graphs indicate how important respondents reported having direct contacts with officials in Congress (left), the federal bureaucracy (center), and the White House (right) is for their overall lobbying strategies. Responses are weighted to reflect the characteristics of the sampling frame. Responses may not sum to 100% due to rounding; $ N $ between 709 and 714 per question.

Second, I explore the extent to which presidents initiate engagement with interests. My survey asked respondents who reported attending White House meetings during the Obama or Trump administrations the degree to which meetings were initiated by their organization or the White House. The distributions of responses (top row of Figure 2) indicate the White House often takes an active role in initiating meetings; the majority of respondents (54.1%) reported the Obama administration tended to take the lead or share responsibility for initiating meetings, and a smaller but still sizable proportion of respondents (42.2%) reported similar experiences in the Trump administration. ¹³ Former White House officials echoed that both presidents and interests initiate interactions but stressed that they and their colleagues exerted significant effort to cultivate engagement; one official explained, they “tried to be affirmative in engaging groups, companies, businesses, you name it … [and] it was also fairly regular for groups, voices, corporate actors, constituency groups, etc., to ask and request meetings or engagements or conversations. So it’s definitely both” (Interviewee D). However, several interest representatives indicated engagement was more commonly initiated by the White House; one lobbyist recalled, “The majority [of meetings arose from] the White House bringing people in” (Interviewee E), and another reported, “More often than not … [the White House] called stakeholders in” (Interviewee F). Furthermore, even when interests request meetings, the White House’s unique institutional characteristics afford it ultimate discretion in disposing of requests. Together, the survey responses and interviews indicate that although some engagement emerges from interests’ outside-in lobbying, the White House plays a more active role in manifesting engagement than traditional lobbying accounts anticipate.

Figure 2. Dynamics of Presidential Engagement with Organized Interests

Note: Bar graphs indicate respondents’ perceptions of dynamics underlying presidential engagement during the Obama (left) and Trump (right) administrations. The top row indicates the degree to which respondents thought direct contacts arose from organizations’ requests versus White House invitations, and the center and bottom rows indicate the frequency with which the White House granted organizations’ requests and organizations accepted White House invitations, respectively. These questions were posed to respondents who reported attending White House meetings during each administration. Responses are weighted to reflect the characteristics of the sampling frame. Responses may not sum to 100% due to rounding; $ N $ between 231 and 248 per question.

Third, I investigate how the White House and organized interests respond to each others’ engagement entreaties. The center and bottom rows of Figure 2 present the distributions of responses to questions concerning the frequency with which the White House acceded to interests’ requests and interests accepted White House invitations, respectively. These results demonstrate interests are more responsive to the White House than vice versa; although approximately three-quarters of respondents indicate both administrations “sometimes” or “usually” granted their organizations’ requests, over 90% replied their organizations “usually” or “always” accepted White House invitations. My interviewees unanimously agreed White House invitations are heeded more often than interest requests. One former White House official reported he rejected most requests because “I just didn’t have a whole lot of time to sit around and talk to people about whatever” (Interviewee C), and another relayed, “When the White House calls, people tend to take the call. People were generally always willing to meet” (Interviewee A). A lobbyist mirrored these sentiments, explaining, “I have not ever turned down a meeting. I always think of meeting as productive. In terms of getting the meetings, yes, it’s very challenging” (Interviewee F).

Taken together, this evidence illustrates that the White House’s interactions with organized interests differ fundamentally from those characterizing other institutions. Although Congress and federal agencies are besieged by interests’ outside-in lobbying, presidents enjoy first- and second-mover advantages providing significant discretion over engagement. As a first mover, the White House conducts inside-out lobbying through institutionalized structures that enable it to manage engagement with interests. As a second mover, the White House’s physical isolation, sparse staff resources, and institutional mechanisms provide unique opportunities to exercise discretion in fielding requests. To be sure, interests’ outside-in lobbying sometimes motivates presidential engagement; however, as illustrated in the middle and bottom rows of Figure 2, the White House’s second-mover advantage allows it to fulfill these requests less regularly than interests accede to its invitations. As one lobbyist explained the relative influence of the White House and his organization in manifesting engagement, “It works both ways, and it’s predominantly us reaching out to them… . [But] they’re the drivers. They determine whether they want to be responsive or not” (Interviewee G).

Presidents’ Motivations for Engagement

Because presidential engagement with organized interests is presidency-driven, presidents’ motivations for conducting engagement are critical to understanding with which interests they engage. Presidents, as goal-oriented actors, pursue electoral success for themselves and their copartisans, favorable policy, and positive legacies (Light 1999). Although they enjoy extensive unilateral powers to pursue their goals, the American constitutional system requires presidents to secure cooperation or assent from others to achieve many important aspirations. For example, presidents’ electoral success rests in voters’ hands. Again, policy initiatives necessitating legislation require congressional approval. Moreover, faced with myriad responsibilities that stretch the limits of their office (Cronin 1980), presidents often rely on other actors, like congressional leaders (Beckmann 2010) and party organizations (Galvin 2009), to provide resources to subsidize presidents’ pursuits.

Thus, presidents’ ultimate success hinges on their ability to build coalitions that can provide support at the ballot box and in Congress and other institutions on whose consent presidents rely and can subsidize presidents’ undertakings. In building coalitions, presidents seek not only to bolster support but also to temper opposition, as both absolute and relative gains improve their standing (Beckmann 2010). Though previous studies of presidential coalition building focus on Congress and the public (Edwards 2000), they largely overlook another class of actors whose support or opposition can affect presidents’ success: organized interests (but see Holyoke 2004; Kumar and Grossman 1984; Milkis and Tichenor 2019; Peterson 1992; 2008). Although interests lack formal powers, such as votes in elections or in Congress, they can be valuable targets of presidents’ coalition building because their institutional resources, like campaign funds and lobbying capabilities, can help or hinder presidents’ objectives. Through engagement, presidents can mobilize interests’ resources in directions favorable to them.

Organized interests’ institutional resources attract presidents’ attention for two reasons. First, interest resources can subsidize presidents directly as they pursue electoral, policy, and legacy goals (Hall and Deardorff 2006). For instance, presidents can use interests’ policy expertise and political intelligence to strengthen proposals and elucidate stakeholders’ preferences on them. In the context of regulatory review, Haeder and Yackee (2015) illustrate that the president’s Office of Management and Budget (OMB) is particularly responsive to comments from interests with policy expertise. Relatedly, Peterson (2008) describes how the second Bush administration’s energy task force relied heavily on the expertise of energy interests—in some cases using the very text of their proposals in regulations. Additionally, presidents can promote their electoral success by stimulating interests to support their campaigns through contributions and voter mobilization efforts (Kumar and Grossman 1984; Peterson 1992).

Second, organized interests can indirectly aid presidents by exerting pressure on actors on whose support presidents rely. For example, because presidents’ legislative success hinges largely on their party’s strength in Congress (Bond and Fleisher 1990), presidents can direct interests’ electoral resources to copartisan congressional candidates. Additionally, to assemble winning coalitions in Congress, presidents can leverage interests’ lobbying capacity to pressure members to support their initiatives (Beckmann 2010; Peterson 1992; Sullivan 1988). Furthermore, presidents can co-opt interests’ grassroots lobbying apparatuses to promote their initiatives among interests’ memberships and the public (Cohen 2012). Accounts of several presidential legislative priorities in recent decades, including Ronald Reagan’s Tax Reform Act (Milkis and Tichenor 2019, 273–4), George W. Bush’s Energy Policy Act (Peterson 2008, 305), and Obama’s Affordable Care Act (Jacobs and Skocpol 2012, 74–5), suggest success depended partly on presidents’ mobilization of interests to lobby Congress and the public.

Sometimes, presidents’ mere engagement sufficiently motivates organized interests to cooperate; when asked about his client’s willingness to comply with White House requests, one lobbyist reported, “If the White House asks, then you figure out a way to make that work for you … it’s very hard to say no to a White House” (Interviewee H). However, should rhetoric alone fail to secure cooperation, presidents’ unilateral powers provide ways to induce compliance. For example, presidents can help interests realize desired outcomes by employing executive orders, appointing preferred officials to executive and judicial positions, and exercising influence over the bureaucracy. Again, presidents can use their bully pulpit to generate attention and support for interests’ priorities; as Baumgartner et al. find, one of the strongest determinants for interests’ success in achieving or stifling policy change is presidential support through taking “an advocacy role in a public-policy debate” (2009, 238). Presidents can also help interests perform institutional maintenance by enabling them to advertise connections with high-level officials (Brown 2014). Thus, presidents have both motivation and means to obtain interests’ cooperation through engagement.

Engagement Strategy as Constrained Optimization

Although presidents’ limited time and resources and the vastness of the organized interest universe helps consolidate their control over engagement, these features also limit the number of interests with which presidents can engage; as one Obama administration official explains, “We wanted [to engage with] everybody, but, at the end of the day, there are a million organizations out there and every single organization isn’t going to be a part of everything” (Interviewee I). Presidents face a constrained optimization problem wherein they must determine which subset of interests, if engaged, provides the highest rate of return. General models of coalition building (Cox and McCubbins 1986; Dixit and Londregan 1996) and studies of presidential coalition building in other contexts (Beckmann 2010; Kriner and Reeves 2015) highlight two characteristics of potential coalition members that should inform presidents’ decisions: interests’ resource endowments and alignment with presidents’ preferences.

First, presidents focus engagement on interests with larger resource endowments to efficiently bolster their coalition’s resources or diminish those of their opponents. Because the costs of engagement (e.g., staff effort required to convene meetings) are fixed, presidents can more dramatically shift the balance of interest resources arrayed for and against them by targeting well-resourced interests (Cox and McCubbins 1986; Dixit and Londregan 1996). Presidents’ focus on well-resourced potential coalition members manifests in other contexts, such as Congress and the public. For instance, rather than lobbying each congressperson, presidents focus on congressional leaders who wield institutional power and can pressure members for presidents (Beckmann 2010). Again, in attracting public support, presidents tailor rhetoric and policies to appeal to citizens with more political resources, such as those in higher socioeconomic strata (Druckman and Jacobs 2015; Griffin and Newman 2016), and whose votes and voices are more valuable, such as swing-state residents (Kriner and Reeves 2015). Turning to interests, Milkis and Tichenor (2019) argue presidents are more likely to collaborate with social movements and associated interests when they possess institutional resources presidents can harness to achieve policy change, and Haeder and Yackee (2015) find OMB is more responsive to comments from well-resourced interests. Thus, presidents take a similar approach by focusing engagement on well-resourced interests.

Second, presidents engage more with interests aligned with their partisan or ideological preferences. This does not mean presidents never engage with interests with divergent preferences; indeed, engagement with such interests can persuade or demobilize opponents (Austen-Smith and Wright 1994; Milkis and Tichenor 2019). As a former White House official expressed, presidents engage with interests to both “increase the likelihood that you … maximize a positive reaction … [and] avoid a negative reaction” (Interviewee B; see also Beckmann 2010). Indeed, accounts of the Affordable Care Act indicate Obama’s engagement with interests naturally opposed to reform, like pharmaceutical companies, was critical to its enactment (Jacobs and Skocpol 2012, 69–75). However, because coalition builders’ entreaties are typically better received by those with similar preferences (Cox and McCubbins 1986), presidents allocate more effort to targets sharing their preferences. For instance, presidents’ rhetoric (Druckman and Jacobs 2015; Eshbaugh-Soha and Rottinghaus 2013) and policy decisions (Griffin and Newman 2016; Kriner and Reeves 2015) often appeal to partisan and ideological allies in the public. Similarly, presidents center congressional outreach on copartisans (Beckmann 2010). Likewise, presidents focus engagement on interests whose preference similarity predisposes them to react favorably.

Research Design

Presidents engage with organized interests through a variety of mediums including phone calls, emails, and White House meetings. Although the same strategic considerations confront presidents across mediums, my empirical analysis focuses on White House meetings for two reasons. First, across institutions, policy makers and interests alike perceive in-person direct contacts, or access, as the most effective medium for gaining counterparts’ attention, making preferences salient, and building relationships (Baumgartner et al. 2009). Second, because the White House must expend more time and effort to conduct in-person meetings relative to other mediums, presidents’ strategic considerations should manifest most strongly in this context. ¹⁴

Although all White House meetings bestow the benefits and exact the costs of engagement on interests and the White House, the circumstances of meetings can provide additional signals about presidents’ engagement priorities. One detail my interviewees highlighted as a marker of the importance the White House assigns to meetings is with whom in the White House interests meet, or the visitee. Although all White House personnel’s time is limited, that of the president and senior advisers is especially scarce because they oversee broader policy and managerial portfolios; thus, from an organizational perspective, engagement using high-level personnel is more costly to the White House. Interests, cognizant of these disparities in power and influence, place more value on engagement with high-level personnel; as one lobbyist expressed, “Relationships that are closer to the president are the most valuable” (Interviewee G). Consequently, the president or senior advisers may be more involved in engagement when the targeted interests are more central to presidents’ coalition-building efforts (Kumar and Grossman 1984, 293–4). Indeed, one staffer shared that although the White House tried to meet with all interests, it “might assign a more junior person to take the meeting or … elevate the meeting to a senior official depending on the significance of the issue and, frankly, the significance of the individual” (Interviewee A). Because presidents’ engagement priorities may be expressed through not only with which interests they meet but also with whom in the White House those interests meet, I consider both White House meetings writ large and as classified by visitees’ ranks.

Empirical studies of direct contacts between policy makers and organized interests typically confront data inavailability (Miller 2021; Schlozman, Verba, and Brady 2012, 295–304). In most contexts, contacts take place behind closed doors and records of them are rarely disclosed. However, recent releases of over seven million White House visitor logs records from the Clinton and Obama administrations provide a rare opportunity to observe White House meetings attended by interest representatives. By pairing these records with lobbying and campaign finance reports, we can assess how interests’ resource endowments and preferences inform presidents’ engagement decisions.

White House Visitor Logs

One of the world’s most secure workplaces, the Secret Service closely monitors all individuals entering and exiting the White House complex and records the comings and goings of individuals without permanent passes in the Worker and Visitor Entry System (WAVES)—more commonly known as the White House visitor logs. Each WAVES record contains information including the visitor’s name and the date and time of the visit. ¹⁵

Until recently, no sitting president had disclosed the White House visitor logs. However, in September 2009, President Obama announced his administration would voluntarily release its visitor logs every quarter. When the Obama administration left office, it had released nearly six million records spanning September 15, 2009 to September 30, 2016. Additionally, in 2015, the Clinton Presidential Library began fulfilling Freedom of Information Act requests by releasing the visitor logs for the final five years of the Clinton administration. As the Lobbying Disclosure Act data I leverage to identify organized interest representatives begins in 1998 (described below), I use the 1.3 million records from the final three years of Clinton’s presidency. In total, I use over seven million records from these administrations. ¹⁶

Because the only information the visitor logs provide about visitors is their names, we cannot discern which visitors were representatives of which organized interests from the visitor logs alone. Rather, to detect instances of presidential engagement with specific interests, we need data providing the names of persons representing each interest for time periods contemporaneous with the visitor logs. I obtain this information from reports interests filed under the Lobbying Disclosure Act of 1995 (LDA), which requires interests to provide the names of their registered lobbyists or those responsible for interests’ political advocacy. ¹⁷ Using these rosters of interests’ representatives, I identify instances of engagement by matching the names of lobbyists provided in each interest’s semiannual (through 2007) or quarterly (2008 and after) LDA reports with the names of White House visitors in corresponding time periods. ¹⁸ $ {}^{,} $ ¹⁹ Then, I aggregate all instances of engagement associated with each interest for each time period in which they filed an LDA report—semiannual for the Clinton administration and quarterly for the Obama administration—to construct measures of engagement. This level of observation—the organized interest-time period—constitutes my unit of analysis.

I use two measures of presidential engagement. First, I create a binary indicator for whether each organized interest-time period observation experienced engagement at least once. ²⁰ Second, acknowledging the White House may be more likely to engage with interests central to its coalition-building efforts with high-level personnel, I use the identity of each meeting’s visitee to create binary indicators for whether each interest-time period observation experienced engagement with (1) the president, vice-president, first lady, or an EOP staffer whose salary falls in the top quartile and (2) any other person at the White House complex. I refer to these types of engagement as “high-quality” and “low-quality,” respectively. ²¹

Organized Interests’ Resources and Preferences

I obtain measures of organized interests’ resource endowments and preferences using data from the Center for Responsive Politics (CRP). Because most interests specialize in either electoral politics or policy making (Tripathi, Ansolabehere, and Snyder 2002), I incorporate measures of both electoral and policy resources. First, I measure interests’ electoral resources as the aggregate amount of campaign contributions they make to candidates for federal office. I focus on aggregate contributions rather than only contributions to the president because aggregate contributions reflect interests’ electoral resource endowments, which presidents can channel in support of themselves and copartisans. Using CRP’s compilation of the Federal Election Commission’s reports for election cycles immediately preceding and including periods for which I have visitor logs—1995–1996 to 1999–2000 and 2007–2008 to 2015–2016 for the Clinton and Obama administrations, respectively—I generate for each interest-time period observation a binary indicator for whether the interest made any contributions in the preceding two years (i.e., the length of an electoral cycle) and a continuous measure of its total contributions in those two years. Second, I measure interests’ policy resource endowments using the lobbying expenditures indicated on their LDA reports in the previous time period. ²² Under LDA, activities for which expenditures must be reported include not only lobbying contacts themselves but also “any efforts in support of such contacts, including preparation or planning activities, research, and any other background work.” ²³ Thus, lobbying expenditures capture a wide range of interests’ policy resources, including staff salaries and expertise. Given the right-skewedness of both resource measures, I apply log transformations.

Measuring interests’ partisan or ideological preferences is less straightforward. Few measures of interests’ preferences exist, and those that do offer coverage for few interests filing LDA reports. For example, Bonica (2013) and Crosson, Furnas, and Lorenz (2020), who estimate interests’ ideology using campaign contributions and positions on legislation, provide scores for only 1,410 (4.8%) and 1,565 (5.3%) of the 29,405 interests in my analyses, respectively. Facing this limitation, I use qualitative information from CRP to construct a trichotomous measure of the partisan alignment of each of the 92 industries into which CRP sorts interests. ²⁴ For each industry, CRP provides a summary of its political activity and policy preferences, often assessing the partisan leanings of interests in the industry. ²⁵ If an industry’s summary expresses a preference for the Democratic or Republican Party, I code that industry as aligned with that party. If the summary does not express clear partisan inclinations, I code it as Independent, or aligned with neither party. Although this classification scheme may misattribute the preferences of some individual interests, studies using interest-level preference measures find that the distributions of preferences within industries typically exhibit clear central tendencies (Bonica 2013; Crosson, Furnas, and Lorenz 2020). Thus, industry-level partisan alignment is an imperfect but informative measure of interests’ preferences providing coverage for all observations. ²⁶

Estimation Strategy

The data structure poses nontrivial challenges for analysis. One set of challenges stems from inconsistencies in temporal units. Although some time-varying components of the data are observed daily, such as visitor log entries, others are observed less frequently, such as interests’ lobbying expenditures. Furthermore, because the frequency of LDA filings changed from semiannual to quarterly in 2008, the temporal units associated with information from LDA reports vary across the Clinton and Obama administrations. I address these disparities in two ways. First, to make temporal units of measurement consistent for all variables, I aggregate all data up to the temporal unit associated with contemporary LDA filing requirements—semiannual for the Clinton administration and quarterly for the Obama administration. Second, given the changes to LDA filing requirements, I estimate separate models for each presidency.

A second set of challenges arises from the data’s time-series cross-sectional structure. Each observation corresponds to an organized interest in a given time period, with each interest nested in one of CRP’s 92 industries. Consequently, observations are non-independent, as multiple observations correspond to the same interests, industries, and time periods. In order to recover unbiased parameter estimates and uncertainty measures, the analysis requires techniques that accommodate this non-independence. Another issue associated with the data structure is that one key covariate—interests’ partisan alignment—is measured at the industry level and thus repeated across observations of the same interests and industries. Unfortunately, conventional modeling approaches cannot estimate parameters for interest- and industry-level effects and partisan alignment due to collinearity.

To account for non-independence and collinearity, I use Bayesian multilevel models. Multilevel models account for non-independence in nested data structures by allowing parameter estimates to vary for each grouping, such as repeated observations of interests, industries, and time periods (Shor et al. 2007). Additionally, multilevel models can estimate parameters for covariates measured at the group level, such as industry-level partisan alignment. Each of my logistic multilevel models estimates a binary indicator of whether an organized interest-time period observation experienced presidential engagement as a function of its lobbying expenditures and campaign contributions; the partisanship of its industry; a series of organized interest-time period-level control variables drawn from LDA filings, such as on which issues it lobbied; and varying intercepts for each interest, industry, and time period. All models report no divergent transitions during sampling and indicate convergence with $ \hat{R} $ statistics of ≤ 1.10 for all parameters. ²⁷

Results

First, I discuss my analyses that measure presidential engagement treating all White House visits equally. Interpreting multilevel models can be difficult because of the large number of parameters estimated. To ease presentation, I provide predicted probabilities of engagement using an observed-value approach (Hanmer and Kalkan 2013), fixing the covariate values of all observations to their observed values while independently varying each covariate of interest—lobbying expenditures, campaign contributions, and partisan alignment—to demonstrate its marginal effect. This illustrates how the probability of engagement changes across levels of each covariate.

The left panes of Figures 3 and 4 present the predicted probabilities of engagement for interests with the levels of resources and partisan alignment specified by the far-left labels during the Clinton and Obama administrations, respectively. ²⁸ Before considering how variation in resources and partisan alignment affect engagement, it is worth noting that even interests who are not primary targets of presidential coalition building experience engagement at non-negligible rates. For instance, interests not making campaign contributions experienced engagement with a 0.50 probability every semester of the Clinton administration and a 0.29 probability every quarter of the Obama administration. With thousands of interests active in each time period, these substantively large probabilities underscore that presidential engagement is commonplace. Additionally, these probabilities reinforce that presidents do not engage only with well-resourced and copartisan interests, as engaging other interests can be important for coalition building (e.g., noncopartisans to persuade or demobilize).

Figure 3. Relationship between Organized Interests’ Resources and Partisan Alignment on Presidential Engagement (Clinton)

Note: In the left pane, points and lines represent the means and 95% credible intervals of the distributions of predicted probabilities calculated using an observed-value approach with parameter estimates from the Clinton administration model (see Table SI.6). For each of the three interest characteristics, I calculate the probabilities of engagement for all observations for each of the specified values while fixing other covariates to each observation’s observed values. In the right pane, I plot the means and 95% credible intervals for the differences between specific pairs of these distributions of predicted probabilities as indicated by the labels on the left side of the pane. For example, the three topmost points and lines in the left pane indicate the probabilities of engagement when lobbying expenditures are set to the first, second, and third quartile values. In the right pane, the two topmost points and lines indicate the differences in the distributions of probabilities when lobbying expenditures are set to their first and second and their second and third quartile values, respectively. Discrepancies between differences indicated in the right pane and differences between corresponding values in the left pane are attributable to rounding.

Figure 4. Relationship between Organized Interests’ Resources and Partisan Alignment on Presidential Engagement (Obama)

Note: In the left pane, points and lines represent the means and 95% credible intervals of the distributions of predicted probabilities calculated using an observed-value approach with parameter estimates from the Obama administration model (see Table SI.6). For each of the three interest characteristics, I calculate the probabilities of engagement for all observations for each of the specified values while fixing other covariates to each observation’s observed values. In the right pane, I plot the means and 95% credible intervals for the differences between specific pairs of these distributions of predicted probabilities as indicated by the labels on the left side of the pane. For example, the three topmost points and lines in the left pane indicate the probabilities of engagement when lobbying expenditures are set to the first, second, and third quartile values. In the right pane, the two topmost points and lines indicate the differences in the distributions of probabilities when lobbying expenditures are set to their first- and second- and their second- and third-quartile values, respectively. Discrepancies between differences indicated in the right pane and differences between corresponding values in the left pane are attributable to rounding.

I use the predicted probabilities in the left panes of Figures 3 and 4 to evaluate my expectations by assessing whether the differences between them across levels of lobbying expenditures, campaign contributions, and partisan alignment are statistically distinguishable. Importantly, we cannot rely on visual comparisons among these probabilities but instead must assess the differences between the distributions from which they are calculated. The right panes of these figures present the differences for specific pairs of probabilities, as indicated by the labels on the left side of those panes. I refer to differences whose 95% credible intervals do not include zero as statistically distinguishable. ²⁹

Considering first the predicted probabilities for varying levels of lobbying expenditures and campaign contributions, we observe consistent increases in the probability of engagement as resources increase. For example, in the left pane of Figure 3, whereas the probability of the Clinton administration engaging an interest with the first quartile of lobbying expenditures ( $ <\$\mathrm{10,000} $ ) is 0.40, this probability increases by 0.13 (33%) to 0.53 when expenditures increase to the median ($20,000). Further increasing expenditures from the median to the third quartile ($80,000) induces a smaller increase in this probability of 0.03 (6%) to 0.56. Similarly, in the left pane of Figure 4, the probability of the Obama administration engaging with an interest not making campaign contributions is 0.29, but this probability increases by 0.07 (24%) to 0.36 when contributions increase to the first quartile ($28,891). Further increases in contributions to the median ($97,088) and third quartile ($337,000) are associated with smaller additional increases in the probability of engagement. The right panes of both figures indicate each stepwise increase in lobbying expenditures and campaign contributions is statistically distinguishable, as the 95% credible intervals for the differences between the distributions of predicted probabilities exclude zero. Therefore, although presidents engage with interests lacking resources at not-insignificant rates, they are moderately more likely to engage with interests with larger electoral and policy resource endowments.

Turning to the predicted probabilities associated with interests’ partisan alignment, we see presidents are more likely to engage with interests in industries aligned with the Democratic Party—the party of both Presidents Clinton and Obama. For the Clinton administration, the probability of engagement for an interest in an industry aligned with the Democratic Party is 0.54, but shifting its industry’s alignment to the Republican Party or neither party decreases this probability by 0.04 (7%) and 0.05 (9%) to 0.50 and 0.49, respectively. Similar effects manifest for the Obama administration; the probability of engagement for interests in industries aligned with the Democratic Party is 0.33, but this probability decreases by 0.04 (12%) to 0.29 when industry alignment shifts to the Republican Party or neither party. The right panes of both figures indicate the differences in probabilities for interests in Democratic versus Republican and Independent industries for both administrations are statistically distinguishable. Thus, although presidents engage with interests of all partisan persuasions, they engage modestly more often with copartisans.

Engagement Quality

Next, I turn to my analyses accounting for engagement quality. Because the time and effort of presidents and senior advisers are more scarce and valuable to the White House, the effects of interests’ partisan alignment and resources (as well as the qualities resources afford, such as policy expertise and lobbying capacity) should be magnified for high-quality engagement with these personnel relative to low-quality engagement with others. To explore this expectation, I fit bivariate multilevel logistic regression models for each administration with high- and low-quality engagement as separate outcomes and compare the coefficient estimates for lobbying expenditures, campaign contributions, and partisan alignment for each type of engagement. ³⁰ Figure 5 presents the differences in the distributions of these estimates for high- and low-quality engagement. ³¹

Figure 5. Relative Effect of Organized Interests’ Resources and Partisan Alignment on High- and Low-Quality Presidential Engagement

Note: Points and lines represent the means and 95% credible intervals of differences in distributions of the coefficient estimates corresponding with interests’ lobbying expenditures, campaign contributions, and partisan alignment for high- and low-quality presidential engagement during the Clinton and Obama administrations (see Table SI.11). For example, the two topmost points and lines represent the difference between the distribution of coefficient estimates for lobbying expenditures when the outcome measure is high-quality engagement and the corresponding distribution of estimates when the outcome measure is low-quality engagement for each administration. Credible intervals for some differences are too small to be visible.

Considering first the relative effects of electoral and policy resources, I expect high-quality engagement coefficient estimates for lobbying expenditures and campaign contributions to be larger than those for low-quality engagement. However, the differences plotted in the two topmost pairs of Figure 5 indicate the estimates for these measures are not distinguishably different across engagement qualities; the differences in the lobbying expenditures coefficients are nearly zero, with narrow 95% credible intervals, and the differences in the campaign contributions coefficients are not distinguishable, with much wider 95% credible intervals. Moving to partisan alignment, where alignment with the Democratic Party is the omitted category, I expect high-quality engagement coefficient estimates for alignment with the Republican Party or neither party to be smaller than those for low-quality engagement. Three of the four differences for both types of alignment are in the expected direction, but all are substantively small with 95% credible intervals including zero. Thus, although the White House prioritizes general engagement with well-resourced and copartisan interests, there is no evidence representatives of interests with those characteristics are systematically more likely to meet with the president or his senior advisers when visiting the White House than with a lower-level staffer. ³²

Conclusion

Presidents often cast organized interests as villains who exercise “undue influence” ³³ and have “made a living bleeding our country dry.” ³⁴ However, as presidents publicly cast interests as foils, they also cultivate cooperation through engagement. Data inavailability long concealed this inconsistency. Using original interviews, survey responses, and administrative data, this paper demonstrates not only that presidents frequently interact with interests but also that presidents do so largely of their own volition. Like presidents build coalitions in Congress and the public, they court members of the organized-interest universe to provide support or withhold opposition, with some emphasis on interests with large resource endowments and who share presidents’ preferences.

These findings have several implications for our understandings of the presidency, organized interests, and representation in American politics. First, although recent discussions of the presidential toolkit focus on powers presidents exercise independently (Kinane 2021; Kriner and Reeves 2015; Lowande and Rogowski 2021), this study spotlights engagement with interests and the institutional mechanisms which facilitate it as an important instrument of power (Peterson 1992). Because presidents can benefit from mobilizing interests’ institutional resources in all contexts, including campaigning, policy making, and managing public opinion, incorporating interests into presidency studies may provide new insights on how presidents advance their goals. For instance, although some argue presidents seldom influence congressional votes (Bond and Fleisher 1990), accounting for presidents’ marshaling of interests to lobby congresspersons may reveal presidential influence in Congress. Again, though presidents’ ability to “go public” has waned in recent years, presidents may persuade and mobilize the public through interests’ outreach (Cohen 2012). The linkage between presidents and interests offers fertile ground to explore open and contested questions about the presidency.

Second, though recent studies of presidential representation stress that presidents emphasize representation of copartisans (Eshbaugh-Soha and Rottinghaus 2013; Kriner and Reeves 2015), my finding that engagement is also informed by interests’ resources suggests presidents’ representational emphases may vary across contexts. Most studies of presidential representation focus on the mass public, where each person possesses roughly equal resources; although some may be more willing or able to participate politically, each can only provide presidents a single vote in an election or nod of approval in a poll. However, when interacting with elites, such as congresspersons or interests, presidents recognize some actors wield more influence and focus representational effort accordingly. For interests, my results indicate engagement is informed by not only interests’ preferences but also their resources. Similarly, in Congress, Beckmann (2010) demonstrates presidents’ coalition building focuses on leaders from both parties rather than only copartisan leaders. Although copartisan leaders are more responsive to presidents, noncopartisan leaders hold important institutional powers and sway over their members and presidents communicate with them to bargain or demobilize opposition. Future studies of presidential representation should explore how presidents consider characteristics beyond preferences, such as resources, when allocating representational attention.

Third, this study offers a rare opportunity to assess how presidents contribute to organized interests’ ability to bias political outcomes. Interests representing businesses and the upper class, who tend to enjoy large resource endowments, pervade and exercise disproportionate influence in Congress and the federal bureaucracy (Schlozman, Verba, and Brady 2012), but some speculate presidents’ impulse for national representation dampens the influence of these privileged interests and provides a friendlier venue for interests “not effectively represented in the existing government” (Bentley 1908, 345; see also Howell and Moe 2016; Quirk and Nesmith 2005; Truman 1971). My results neither fully endorse nor fully refute this conjecture. In one respect, presidents’ preference for engaging with high-resource interests perpetuates disparities in political voice that benefit business and upper-class interests and marginalize other constituencies. However, presidents engage with even low-resource interests at not-insubstantial rates, and high-resource interests’ odds of experiencing engagement are far from guaranteed; for instance, during the Clinton presidency, even interests making no campaign contributions have a 0.50 probability of experiencing engagement, and increasing contributions to the third-quartile value (≈$200,000) only increases this probability to 0.63 (see Figure 3). Furthermore, whereas larger resource endowments increase interests’ probability of experiencing some form of engagement, I find no evidence resources exert stronger effects on access to the president and senior advisers than to lower-level staffers; thus, although higher resource levels induce modest increases in presidential engagement, they are not indicative of higher quality engagement.

An ideal assessment of how presidential engagement contributes to upper-class bias relative to policy makers’ provision of access in other institutions, like Congress, requires data from those institutions comparable to that from the White House visitor logs. Lacking such data, comparisons to Kalla and Broockman (2016) suggest presidents privilege high-resource interests to a similar or lesser extent than do congresspersons in providing access. In their field experiment, Kalla and Broockman (2016) find interests are 5 percentage points (12%) more likely to secure a meeting with a congressperson’s office and 13 percentage points (240%) more likely to secure a meeting with the congressperson or a senior staffer (i.e., high-quality access), when they reveal electoral resources by mentioning their members donate to campaigns. Although my effects of contributions on the probability of engagement conducted by any member of the White House are similar to those of Kalla and Broockman (2016), ³⁵ my effects of contributions on high-quality engagement are much smaller. ³⁶ Furthermore, unlike Kalla and Broockman (2016), I find no difference in the effect of contributions on interests’ access to high- and low-quality engagement. Thus, presidents may not counterbalance the advantages resources afford interests in securing access to other institutions, but they may contribute less to upper-class bias in political voice—particularly at the highest levels of government.

This paper not only illuminates important facets of presidents’ interactions with organized interests but also highlights lines of future inquiry. As only two recent Democratic administrations have disclosed their visitor logs, this paper cannot directly address how some forms of contextual variation, such as presidents’ partisanship and institutional changes affecting presidents’ engagement abilities, may condition engagement. In the former case, presidents of both parties face the same engagement incentives, but Republicans’ closer alignment with well-resourced business interests could augment the effect of resources on engagement for Republican presidents (Grossmann, Mahmood, and Isaac 2021). In the latter instance, the creation of OPL and its ability to coordinate engagement may enable modern presidents to more efficiently align interactions with interests with their goals than in previous periods. Both paths of research would benefit from additional records of engagement such as visitor logs from Republican administrations or more circumscribed information about presidents’ own engagement with interests from Presidential Daily Diaries or Oval Office tapes. Future work could also consider the interplay between presidential engagement and organized interest coalitions. Although the current analysis focuses on presidents’ direct engagement with interests, engagement with interests in coalitions may enable presidents to indirectly engage other coalition members. Additionally, presidents may use engagement to forge coalitions by mobilizing and coordinating the activity of interests sharing common goals. Finally, whereas this paper argues engagement mobilizes interests in service of presidents’ goals, it does not explicitly test whether presidents’ efforts are successful. Subsequent work should evaluate the efficacy of engagement by linking interactions between presidents and interests to interest behavior, such as congressional and grassroots lobbying on presidential initiatives.