Lawyers as Lobbyists: Regulatory Advocacy in American Finance

Abstract

Administrative agencies have undertaken an increasingly substantial role in policymaking. Yet the influence-seeking that targets these agencies remains poorly understood. Reporting exceptions under the Lobbying Disclosure Act allow many of the most powerful advocates to characterize their activity as lawyering, not lobbying, and thereby fly under the radar. Using agency-generated records on lobbying activity, financial reporting, and personnel databases specific to lawyers, as well as LinkedIn, we describe a vast subterranean world of regulatory influence-seeking that the social-science literature has (mostly) ignored. Regulatory lobbying is systematically different from legislative lobbying. It involves different kinds of people and different lobbying firms that bring specific forms of expertise and distinct networks. Our key findings about how regulatory lobbying differs include the following: (1) the regulatory lobbying sector is highly segregated from the reported lobbying sector, with many regulatory advocates failing to consistently register or report earnings commensurate with their activity level, (2) the number of unregistered regulatory advocates working on the implementation of the Dodd-Frank Wall Street Reform Act plausibly exceeds 150% of the registered lobbyists working on that law, (3) the most effective regulatory lobbyists and law firms involved with regulatory lobbying have incomes that dramatically outpace leading reported lobbying firms (which are also mostly law firms), and (4) back-of-the-envelope calculations and more sophisticated decomposition regressions imply that aggregate expenditure on lawyer-lobbying is several multiples of reported lobbying spending. We introduce the case of a particular lawyer-lobbyist and provide a theoretical discussion to situate and contextualize these findings. Collectively, this work opens a window into neglected domains of politics and reveals an important and understudied form of political inequality.

We learned very quickly that if you combine legal disciplines with lobbying, you have a much better chance of getting something accomplished than if you just lobby. … We would always try to combine legal disciplines with lobbying.That’s pretty much how lobbying is now done.

Thomas Hale Boggs Jr., founder of Squire Patton Boggs

Lawyers play a vast but underappreciated role in lobbying in the United States, especially regarding the many hard-to-observe regulatory issues facing executive agencies. In the years following passage of the Dodd-Frank Wall Street Reform Act, influence-seekers of many kinds, but especially lawyers, frequently met with the regulators charged with implementing more than 300 separate rulemaking mandates contained in the 2010 statute. Until these rules were completed, the law would not be fully effective. Indeed, some of the law’s most important provisions—for example, those relating to problematic executive pay practices—have still not been implemented more than a decade later, in large part due to the blistering lobbying campaigns waged against federal financial agencies charged with developing those statutory provisions into operative regulations. ¹ Regulatory lobbying campaigns, largely run by lawyers, are systematically different from legislative lobbying campaigns, which are also often run by lawyers. Even though lawyers are involved in both instances, each kind of lobbying aims to persuade fundamentally different policymakers (i.e., bureaucrats versus politicians) and, to be effective, must involve the specific actors, methods, expertise, and networks that cater to the respective audience. Even more than legislative lobbying, regulatory lobbying occurs beyond the bounds of disclosure laws. Moreover, even though legislative and regulatory forms of lobbying are systematically different, they are also related: the ability of regulatory lobbying to stymie the implementation of statutes shapes the incentives of interest groups in fighting over the content of legislation in the first place, opening a different venue in which lobbying can “succeed” or “fail” in influencing public policy (Ban and You 2019). A large literature on regulatory capture, iron triangles, and subsystem politics attests to the pervasiveness of regulated industries’ influence over regulatory processes (Gordon and Hafer 2005). Even so, this literature has often been faulted for lacking specificity about the mechanisms—the actors, motives, and methods—by which industry achieves these policy outcomes (Carpenter and Moss 2013).

Our article focuses on a set of actors—we call them lawyer-lobbyists—whose influence is seldom directly seen but nevertheless is often felt in the regulatory process. There is no better example of such a figure than Rodgin Cohen. For more than a decade, “Rodge” Cohen was the managing partner of the elite law firm Sullivan and Cromwell. Described as the “trauma surgeon of Wall Street” ² and “the preeminent go-to lawyer on Wall Street,” ³ Cohen owes his renown largely to what he has done in Washington, DC. Cohen spent much of the years from 2009 to 2012 shuttling between Washington and New York City. The New York Times described his work less as brokering or structuring deals between banks and more as capitalizing on his “trusted relationships with people in government.” One Treasury official who ran the federal bank bailout regarded Cohen as “one of my kitchen cabinet of advisers,” who was “always available as a sounding board.” That was in 2009, a year before Dodd-Frank was enacted. In the years after its enactment, Cohen met regularly with officials at the Federal Reserve, FDIC, and elsewhere (see table 1).

Table 1. Partial list of Cohen’s disclosed contacts

Notes: This table presents counts of occurrences of Rodgin Cohen’s name in meeting disclosures posted by regulatory policymakers. Sources for the Federal Reserve Board staff meetings are stakeholder meeting logs posted on the FRB website. All other counts come from agency chief executives’ public calendars. Periods indicate years where no meeting is recorded in these sources, although other meetings may have occurred. Note that the data establish minima; important and known meetings between Cohen and Treasury Secretary Geithner (Geithner 2014) are not included, for example.

At this point, however, a puzzle arises. At no point during the 2007–8 financial crisis or since has Rodgin Cohen registered formally as a lobbyist. But for at least six different years from 1998 to 2004, he did so register. ⁴ We presume that Cohen has complied with relevant lobbying regulations. Indeed, we do not believe that he has skirted either the minimum activity requirements ⁵ or the minimum expenditure requirements ⁶ under the Lobbying Disclosure Act. Instead, he has more likely relied on the exemptions in the Lobbying Disclosure Act for many of the most important categories of regulatory influence-seeking, including comment letters, “communications” more broadly made in response to notices in the Federal Register, and participation in other forms of agency proceedings (2 USC 1602(8)(B)). Cumulatively, these exemptions for activity that is often conceptualized as “lawyering” rather than “lobbying” allow an enormous amount of agency-focused influence-seeking without having it legally qualify as lobbying or generating any lobbying reports.

Our question here, however, is not a legal one about the effectiveness of some regulatory regime but rather a scholarly one about the activity that regime seeks to regulate: How good is our understanding of lobbying if it is premised on concepts and measures that fail to pick up what advocates like Rodgin Cohen are doing? That Cohen’s work plausibly involves lobbying is clear from two facts: (1) six times before 2005, he registered formally as a lobbyist, and (2) much of his work involves meeting with officials who make regulatory policy. Additional confirmation that Cohen’s work involves lobbying comes from firsthand accounts of lobbied officials. In his 2014 autobiography Stress Test, Tim Geithner describes multiple instances of the “omnipresent” and “inevitable” Rodge Cohen calling or meeting with the then-treasury secretary about various matters, including extraordinary financing and regulatory decisions intended to prop up the business of Cohen’s struggling clients such as Bear Stearns (p. 158), Lehman (p. 179), AIG (p. 209), and Wachovia (p. 218)—often in ways that would also provide significant aid to Cohen’s more successful clients, such as Goldman Sachs and JP Morgan Chase. That at least some of Cohen’s advice on these matters was probably persuasive is evident from Geithner’s admission that he later personally sought to have Cohen appointed as his deputy treasury secretary (p. 349).

Cohen is an extraordinary case admittedly, but he is nonetheless a cardinal example of a much broader pattern: the lawyer as regulatory lobbyist. Cohen has made many millions of dollars from his legal work, and the companies he represents have likely made billions of dollars of additional profits from his influence and that of people like him. To be sure, scholars have long known that lawyers like Cohen, especially in Washington, frequently work as lobbyists (e.g., Horsky 1952). Indeed, Heinz et al. (1993) open their well-known book The Hollow Core with a story of a lawyer whose lobbying on a proposed rule was delayed through the ham-handed influence-peddling of a “heavyweight” former cabinet secretary. Although scholars have not ignored the important tension between expertise and connections that Heinz and coauthors imply through this story (Bertrand, Bombardini, and Trebbi 2014; Egerod and McCrain 2023), studies have paid much less attention to another equally substantial Hollow Core theme: the professional differences between lawyers, influence-peddlers, and other kinds of advocates and experts involved with the policymaking process. This lack of attention is unfortunate because the differential visibility of the activity of various kinds of lobbyists is likely to profoundly influence scholarly understanding of the subject. In particular, given the reporting exceptions for lawyerly representation in administrative proceedings, the role of attorneys is likely to be underappreciated. De Figueiredo and Richter’s (2014) excellent review of the lobbying literature does not use the words “lawyer” or “attorney” once, nor does Bombardini and Trebbi’s more recent (2020) review of the same topic. Baumgartner and colleagues’ contemporary classic Lobbying and Policy Change (2009) does discuss lawyers at some length, particularly in connection with sampling procedures that focus on issues before Congress. Tellingly, they emphasize that their procedures likely undercount “the lawyers who file public interest lawsuits or design the legal defense strategies of trade associations…. Likewise, many lawyers, engineers, and other specialists who work on regulatory issues for companies do not consider themselves legislative lobbyists and are not required to register as such.” Although Baumgartner and coauthors acknowledge that their Congress-centered approach to lobbying will miss figures like Cohen, they also assume that the most important lawyer advocates work for trade associations or as litigators (as a transactional attorney, Cohen does neither). Even though their approach provides a meticulous effort to understand the typical issues being lobbied on in Congress—composing “the most representative sample that has ever been used in a study of lobbying”—we respectfully submit that any empirical strategy that ignores the Rodgin Cohen phenomenon will systematically miss critical, institutionalized, and expensive patterns that constitute modern policy lobbying.

Of course, it is easy to claim that our understanding of special-interest politics has been hamstrung by poor or missing data, which is widely acknowledged. What is harder to show is that consideration of other data leads to a different and more complete understanding. To that end, we focus on the analysis of lobbying records produced by financial regulatory agencies (Carpenter et al. 2020), rather than Lobbying Disclosure Act (LDA) data, and ask what these documents reveal about the lobbyists who conduct the advocacy in this forum. A key descriptive fact that readily emerges from these records is that vast legions of regulatory lobbyists engage with executive agencies. Crucially, these regulatory lobbyists leave few traces of their activity in public lobbying databases. We also note that frequently, although not invariably, these regulatory advocates are lawyers. The observation that regulatory lobbyists are often lawyers opens many possibilities for measurement, facilitating understanding of this hard-to-observe segment of the lobbying industry. These measures help us differentiate the regulatory lobbying industry from the “lobbying industry” as it has been conventionally understood through LDA reports. Top regulatory lobbying firms’ professional standing and earnings are, for example, substantially higher than those of top lobbying firms. Using databases of lawyers at large firms, we aggregate the number of lawyers specializing in government and public policy and determine that there are about as many of these individuals as there are registered lobbyists—which means that the plausible “number of lobbyists” on related issues is double the number commonly measured.

The understandings that emerge from this empirical effort are also theoretically and substantively important. The people who show up for regulatory lobbying are systematically different from those who show up for lobbying, as customarily observed. The predominance of “transactional lawyers” among these regulatory advocates raises important theoretical issues. Why do the people who show up to lobby an agency about the implementation of a statute differ so much from those who lobbied Congress on that same statute to begin with? Why do lawyers—and, at that, not litigators but transactional lawyers—dominate the practice of regulatory lobbying instead of another class of professionals, most obviously economists? Why does it matter that lawyers are the dominant professional class in regulatory lobbying or lobbying more generally? We only begin to sketch some answers to these and related questions.

Lobbying by Transactional Lawyers

Although we are not the first to make claims about the underappreciated importance of the regulatory advocacy conducted by lawyers, we do not believe that the significance of this activity has been as systematically and forcefully documented as it should be. Even though the lobbying literature is swiftly becoming familiar with the rulemaking context in which so much lobbying activity occurs (e.g., Dwidar 2022; Carpenter et al 2022; Libgober 2020a; Libgober 2020b; Potter 2019; Yackee 2019; McKay and Yackee 2007; Yackee and Yackee 2006), there remains in our view considerable unfamiliarity with what regulatory lobbying is and how a transactional attorney such as Rodgin Cohen could figure so prominently. This is crucial: Rodgin Cohen is the kind of lawyer who goes to boardrooms, not courtrooms, and who structures deals, rather than files lawsuits. Given the likelihood of confusion about how and why transactional attorneys would have such influence, we consider it appropriate to begin by presenting some theoretical and factual primitives about regulatory lobbying drawn from the social-science literature. Having done this, we consider in depth the work of regulatory lobbyist Rodgin Cohen, whom we choose to focus on because he is the most frequently appearing regulatory lobbyist in our data. Finally, we turn to the quantitative description of our data and present some key stylized facts about lawyerly lobbying and regulatory advocacy.

Although we emphasize patterns that we expect will have substantial generalizability, it is worth explaining at the outset why we focus on US financial regulation. First, the area has been particularly active in recent years as legislation has sought to dial-back previous decades’ efforts at deregulation (Barton 2022). Second, in various case studies, financial industry intervention into rulemaking under Dodd-Frank was observed to be tilted decisively toward large bank holding companies (Krawiec 2013; Carpenter and Libgober 2018; Libgober 2020a; SoRelle 2020). Third, finance has been identified as a plausible source of structural inequality (Jacobs and King 2021), both in the returns to capital (Piketty 2014) and the financialization of the US economy (Krippner 2011). Fourth, the sheer size of some of the companies involved in finance, when combined with the size of their portfolios, means that seemingly small changes in lobbying activity and success can aggregate to billions of dollars in value and hence billions of dollars in explicit and implicit redistribution. Finance is undoubtedly not representative of all policy areas, and we both admit as much and rely on that fact in our study design. However, if scholars in political science and allied fields study policy advocacy in finance without giving regulatory lawyers a central place, they will miss most of the action and, we believe, the site where inequality in resources becomes most dramatic. As an exemplary case, finance may illuminate dynamics of lawyer-lobbying in other fields where much of the most important policymaking is delegated to bureaucratic actors; for example, telecommunications, environmental policy, health and technology policy, and labor.

Regulatory Advocacy as Lawyerly Lobbying: Toward a Theory

Regulatory advocacy entails the statement and defense of an interested position in the regulatory policy domain. This domain is defined as generalized and prospective policymaking processes in regulatory administration. Formal or informal rulemaking is the paradigmatic example; however, standard setting via adjudicatory processes is also common and within the scope of regulatory patterns we have in mind (Breyer 1982). Excluded from our definition would be activities taking an interested position with respect to purely retrospective or exclusively individualized application of policy to a particular individual or case. Helping a corporation navigate issues around criminal liability at the hands of an impending governmental prosecution is not what we mean by regulatory advocacy. Nor does regulatory advocacy, as we understand it, cover nonprecedential private letter rulings from the IRS about some esoteric tax matter. So defined, regulatory advocacy is practiced by what Thomas and LaPira (2017) and LaPira (2015) call “unregistered policy advocates”; that is, individuals who are “paid to challenge or defend the policy status quo, to subsidize policymakers with information, or to closely monitor intricate policy and political development that are not readily available to the public—or those who offer expertise, knowledge, and access in support of these activities—yet who do not register as lobbyists” (LaPira 2015, 229).

Although regulatory policy is general and prospective, regulatory advocacy is typically conducted on behalf of a particular client; hence, for self-interested reasons specific to the entity footing the bill. Yet, the fact that a generally applicable policy is at stake creates important constraints on advocates. In arguing that a rule implementing a statute should be changed to reduce its cost, a regulatory advocate might argue on behalf of Company X (claiming that a stringent rule would damage Company X). While doing so, the lawyer-lobbyist does not argue for a waiver for Company X specifically but instead pushes for a general reduction in the stringency or cost of the rule. Although it is understood that Company X would benefit from this change in policy, and one may suspect probably would benefit relatively more than competitors who are pressing other issues, advocates typically must give regulators a reason and incentive for the policy change, often couching their concerns more broadly in some legitimate policy goal (Trumbull 2012; Libgober and Rashin 2023). To the extent that advocates can find compelling distinctions between their clients and firms within their sector, general standard setting can devolve into allocating firm-specific benefits. Indeed, Libgober (2020a) finds that a set of firms who lobby about rules obtain higher stock market returns than their most similar competitors on the announcement of those rules, evidence that regulators adopt policy wedges disproportionately benefiting the firms that engage in regulatory advocacy.

Although regulatory policymaking through rulemaking allows for distributive politics to some degree, other administrative policymaking (more rarely observed) allows advocates to appeal directly for discretionary distributive benefits. A particularly notable example is when Rodgin Cohen appealed to then-treasury secretary Geithner for bailout loans to Bear Stearns. For lawyers, this case of advocacy should be distinguished from the judicial defense (in court or through prelitigation negotiation and all prelitigation filings) of a company charged under enforcement of an existing rule or statute. In other words, what makes participation in rulemaking or Cohen’s appeal for a discretionary loan “regulatory advocacy,” and not traditional lawyering work, is the fact that the participants did not need to have either a law degree or admission to a bar to advocate in this way, nor was the expertise brought to bear particularly legal but rather was mostly related to knowledge of the policy domain and subject at issue. For legal defense of a company targeted in regulatory enforcement, some minima of inclusion and training in the legal profession are required to enter, and selection among those eligible is related to their expertise in legal procedure.

In scholarship and the popular imagination, former members of Congress and their staffs play an outsized role in legislative lobbying because of the human and relational capital that can only be acquired in those jobs (Hirsch et al. 2023). In reality, there are a large number of domain and policy experts armed with PhDs and other credentials that are also active in lobbying because legislators require the information they possess (Ban, Park, and You 2023; Hall and Deardorff 2006). A substantial theoretical question, therefore, is why lawyers are mobilized so extensively in regulatory advocacy, as opposed to some other profession or background. If there are no formal requirements granting lawyers a monopoly on participation, if legal training does not readily lend itself to providing deep insights into the costs and benefits of regulatory policy choices, and if lawyers do not typically engage in the activity being regulated, why are there so many lawyers there? Even though we pose these questions theoretically, the answers have important normative stakes. We strongly agree with Bonica and Sen (2020) that the dominance of lawyers in American government is emphatically a policy choice, although one so deeply entrenched that many consider it inevitable—despite comparative evidence from other national and subnational contexts showing that it is possible to make do with few, if any, attorneys. At the same time, there are usually reasons why a particular market equilibrium prevails, and understanding these reasons is also crucial for any normative assessment.

Six Reasons Why Lawyers Dominate Regulatory Advocacy

We propose several explanations for why lawyers dominate regulatory advocacy. First, regulatory agencies function through administrative procedures. These procedures do not typically bind Congress or other legislatures. At the national level in the United States, these procedures are structured by the Administrative Procedure Act of 1946 and a body of associated statutes and case law. These procedures are, in a sense, the way agencies think. They are critical because (a) they define a normal way of doing things for agencies as they consider new regulations, draft rules, receive and consider comments, and finalize regulations, and (b) they set up powerful veto players—courts—that can veto agency action if they deem the agency not to have followed the procedures. Although judicial review is theoretically separable from administrative procedures, they are often linked, as in the case of the Administrative Procedure Act. Administrative procedures magnify the voices of those who can speak to administrative procedural constraints; they especially enhance those same voices in cases where judicial review may lead to the imposition of huge costs for regulators who are perceived to have not followed procedures. The role of administrative procedures in making regulatory policy, especially backed by judicial review, encourages lawyers to occupy privileged positions not only on the government’s side of policymaking but also on the side of private influence-seekers, especially those lawyer-lobbyists who have experience doing detailed work on rulemaking or rule application.

Second, scientific and technical complexity are often expressed through law, especially complex contracts. Put differently, what is technical or specialized about certain policy areas is not merely the scientific basis of the policy but the role of law in expressing that science and even in defining scientific or technical concepts. In the case of finance, for example, it is one thing to understand the economics of aggregating many mortgages into a product that can be bought and sold as securities, while it is quite another to know what mortgage-backed securities agreements look like and how they anticipate issues such as foreclosure or nonperformance by intermediaries. Other financial activities such as bank acquisitions and mergers, branching, and new forms of loan generation also involve contracts and implicate governing legal regimes that importantly require nuanced ordinary or sophisticated usage. Similar developments exist in telecommunications and biotechnology, where assets are defined by intellectual property (itself a complex of contracts), by rights to a spectrum or a molecular model, by highly technical definitions and concepts, or by all these factors. Whether in finance, telecommunications, biotechnology, or other areas, the complexity of the contracts that define the field is sufficiently bewildering that lawyers—and not just any lawyers, but highly specialized ones—must be involved.

These two considerations produce a third: lawyers can speak to constraints of feasibility, given that whether an idea can be implemented depends on how well it translates into administrative text and its potential reception by the courts. Economists and other advisers with policy expertise might more credibly speak to an idea’s theoretical or normative desirability but are less well positioned to advise on its administrative and legal feasibility. Given individual processing limitations and the challenges of coordinated action, bureaucratic decision making tends to focus on what solutions can be adopted first, often relying heavily on precedent, while deemphasizing comprehensive value pursuit, a full search of alternatives and their trade-offs (Lindblom 1959). In this sense, being steeped in the precedent of what has been done is a critical advantage, and the ability to do an exhaustive welfare analysis of possible courses of action is less important than one might think.

Fourth, the industrial organization of law facilitates the legal dominance of regulatory advocacy. Companies generally do not hire individual lawyers but law firms, and this business has become increasingly dominated by “big law.” Bigger law firms can specialize in ways smaller firms cannot, using entire teams of lawyers. Larger law firms can cross-subsidize one area of work (a particular field of law such as food and drug law or a kind of activity such as writing notice-and-comment letters) by its profits in another. Such cross-subsidization can reflect a stable equilibrium in the legal market, either because the large law firms that engage in cross-subsidization enjoy rents or because cross-subsidizing these firms effectively seeds areas of future revenue growth.

Fifth, lawyer mobilization stems from informational credibility. Regulators need information from regulated firms, but this reliance creates a trust problem. Lawyers help grease the machinery by allowing the regulator to trust that the information they get from companies is legitimate. They also help communicate the intensity of preferences of companies credibly. Know-how and know-who are both essential elements of the craft, as is the fact that lawyerly advocates come from a relatively small world where dishonesty can produce outsized social and professional sanctions. The fact that regulators cannot sanction regulatory advocates for misconduct in the same way that courts can punish lawyers for malpractice makes these social punishment mechanisms more crucial. Although the revolving door’s influence may be overstated, there is little doubt that the revolving door can enhance the credibility of a regulatory advocate. If a former general counsel of an agency that is writing the rules is now working for affected Company X directly or with a law firm and says that a certain argument or interpretation is a reasonable policy view, who in Company X or even at the regulatory agency can dispute that?

Sixth, and finally, agencies depend on lawyers to generate compliance with rules and enforce agency mandates. ⁷ Law firms make substantial investments in interpreting and analyzing regulatory announcements. Various kinds of client advisories, memos, presentations, and reports are the concrete manifestations of these efforts. Some of this work product is made available exclusively to clients as “club goods.” Occasionally, it may even be offered up freely for general public consumption, one suspects as a form of advertisement for the club goods and even more bespoke advising one pays for. ⁸ The interpretations that counsel offer in public, semiprivate, and one-on-one advising play a significant role in determining what actions regulated entities will take or not take, what sort of procedures they will develop or not develop, and how the regulated field operates post-implementation. Given the role of lawyers as authoritative interpreters of agency action, and the reliance that regulated entities will place on them, regulators have significant incentive to engage these firms and consider their perspective in making regulations. Indeed, the incentive for regulators to be open-minded to input from these vicarious enforcers is all the greater if regulatory attention and budgets are limited or governmental priorities are conflicted: such constraints erode administrative enforcement capacity and make agencies more dependent on law firms as vicarious enforcers. Law firms and regulated industries are not ignorant of these dynamics, and strategies aimed at weakening an agency’s ability to enforce the rules may also aim to influence its incentives in making the rules in the first place.

In our conceptualization, lawyers who work for private law firms serve a function that trade and industry associations may have played more prominently in prior decades and still play in other comparative contexts. Indeed, we will provide correlational evidence to this effect. Classic studies in the interest-group literature from the twentieth century, ranging from general business mobilization (Bauer, Pool, and Dexter 1963) to the farm lobby (Hansen 1991), emphasized peak associations such as the National Association of Manufacturers or the American Farm Bureau Federation. Later in the twentieth century, lobbying began to change with the rise of established Washington-based lobbying outfits such as Akin Gump Strauss Hauer & Feld LLP and Squire Patton Boggs LLP. These organizations appear to focus predominantly on legislative lobbying. The migration of policymaking to the regulatory and administrative realm, however, means that many of these more traditional advocacy vehicles are weakened or need legal expertise to augment their influence.

To be sure, financial industry associations are still active in regulatory advocacy. However, as Krawiec (2013) showed in her careful study of the Volcker Rule, general industry associations met less often with agencies writing the Volcker Rule than with bank-allied law firms. Our point is not that trade associations have disappeared but rather that advancing the industry interest in practice often falls to lawyers and law firms with dense networks of client representations and not only to trade groups. This shift from trade groups to law firms has substantial normative stakes. Although one can certainly debate the extent to which trade associations are democratic, at the very least, they typically have few constraints on entry, and their formal procedures allow the preferences of their broad membership to translate into collective positions. Law firms have no such guarantees. Instead, they sell their services to the corporations offering the largest book of business, and bigger companies need more legal work than smaller ones. Considering these differences, we think the substitution of trade associations for lawyers and law firms is a notable and worrying development from the standpoint of political and economic inequality.

A Case of Exemplary Influence: H. Rodgin Cohen and Sullivan and Cromwell

Rodgin Cohen and his firm Sullivan and Cromwell (S&C) occupy a distinctive role at the nexus of Wall Street and Washington. Much of their work, we argue, should be interpreted as regulatory advocacy and as a kind of lobbying.

Cohen is deeply connected to the bank holding companies (BHCs) that dominate much of global systemic finance. These connections stem partly from the fact that, as a transactional attorney, he helped build these organizations in the first place. Much of Cohen’s renown within the legal industry came from his role in facilitating the wave of mergers of banks in the 1990s and early 2000s. While doing so, he helped form the very large BHCs that would later dominate financial rulemaking (Krawiec 2013; Libgober 2020a). Although much of this work occurred in the market rather than in the halls of government, the Wall Street Journal reports that even earlier in his career, far before the financial crisis, he was active in influencing policy by “helping draft the rules that led to the emergence of powerful national banks” (Pressler 2009).

Another important source of connections between Cohen and BHCs is his leadership role in a law firm well known for placing its alumni “in-house” with its clients. In December 2017, for example, Karen Patton-Seymour left her partnership at S&C to become a partner and co-general counsel of Goldman Sachs (Greene 2021). When Darrell Cafasso left his partnership at S&C to lead Goldman’s litigation and regulatory proceedings group under the title “managing director and associate general counsel,” the New York Law Journal remarked that he was “treading the well-worn path from the Wall Street law firm to the corporate legal department of a key longtime client, Goldman Sachs & Co.” ⁹ As an apex figure within an apex Wall Street law firm, Cohen has long occupied a central position within the New York transactional bar.

During the financial crisis of 2007–8, Cohen did much of his most important work by directly appealing for governmental aid on behalf of companies that he represented. These included (1) his personal request to then-treasury secretary Timothy Geithner to aid Bear Stearns, which was then acquired by the BHC, JP Morgan Chase (another accomplishment in his mergers and acquisitions [M&A] work); (2) his work on the closure of Lehman Brothers, whose brokerage unit was then acquired by Barclay’s (as Cohen switched his representation to Barclay’s); and (3) his appeal for an $85 billion bailout from the Federal Reserve for American Insurance Group (AIG). As with his M&A work, his “trauma surgery” work often reads as transactional; however, it has important linkages to policy battles fought in years past. The very powers that the Federal Reserve Board (FRB) was exercising in its 2007–10 bailouts were powers that Cohen had helped give it earlier. In a 2010 interview, Cohen acknowledged that he was personally involved in an amendment to the Federal Deposit Insurance Corporation Improvement Act (FDICIA) that transformed the Federal Reserve’s emergency lending powers (Cohen 2010). That amendment would serve as the statutory basis for $29 trillion in loans made from mid-2007 to 2010, $4.5 trillion of which were revolving loans to Morgan Stanley and Citigroup (Felkerson 2011; Martens and Martens 2019).

The BHCs that survived the financial crash of 2007–8 were confronted with a wave of new regulations following the Dodd-Frank Act of 2010. These companies turned to Cohen and S&C for many of their needs. Cohen met many times with federal agency officials from 2011 to 2020 (table 1), during which time he was not registered as a lobbyist and thus was never officially counted as a lobbyist. He also met repeatedly with FDIC leadership and with FRB officials. These meetings continued through the Trump presidency, as Cohen often met with Treasury Secretary Steven Mnuchin and, on at least one occasion, with Mnuchin and President Trump’s personal attorney at the time, Marc Kasowitz. ¹⁰ At least three features of Cohen’s meetings speak to possibly more general patterns. First, it was often the case— in two-thirds of the events we documented—that S&C was the only law firm with attorneys in the room at these meetings. Indeed, in these cases, Sullivan and Cromwell was the only nonbank, nongovernmental entity at the meetings. When Cohen and S&C were at the table from 2011 to 2020, in other words, other “big law” and “white-shoe” law firms generally were absent. Second, industry associations almost never appear in the meetings at which Cohen appeared. Third, in almost half (9 of 21) of Cohen’s meetings with FRB officials, Goldman Sachs representatives were also at the meeting. No other bank or BHC appeared more than twice at these meetings.

Although Cohen may have been the S&C attorney most actively engaged in post-crisis policymaking and regulatory advocacy, he was far from the only S&C attorney actively engaging financial regulators on matters of importance to the firm’s BHC clients. Wells Fargo turned to S&C for representation in negotiating a $1 billion settlement with financial regulatory agencies (Barber 2018). S&C partner Jay Clayton, later nominated to lead the SEC by President Donald Trump, reported $7.62 million in earnings from 2015 to March 2017 from his representation of Goldman and hedge-fund pioneer William Ackman’s Pershing Square Capital Management (Barber 2017). In early 2021, S&C announced two vice chairs of the firm from within its ranks, Robert J. Giuffra Jr. and Scott D. Miller, both of whom had been active in representing Goldman.

In addition to engaging behind closed doors with regulators in the post–Dodd-Frank era, Cohen took public regulatory stances similar to those normally associated with industry lobbyists. Specifically, he took the view that financial regulation after Dodd-Frank had become too stringent, too little informed by trust between regulator and regulated, and too confrontational. In 2015, he argued publicly that “the regulatory environment today is the most tension-filled, confrontational and skeptical of any time in my professional career.” ¹¹ He blamed this fact on a “supposition of regulatory capture [which] has become as pervasive as it is false.” His comments clearly indicate that Cohen did not view this tension as a good thing and that he believed that regulatory capture was not happening. Later that same year, when the Federal Reserve began to change methodologies for stress tests after the financial crisis, Cohen interpreted the methodological changes as substantive changes to rules and opined, “If we’re going to increase these capital requirements, shouldn’t there be notice and opportunity for comment?” ¹² Whether Cohen was correct on either count—he might well have been— is immaterial to our analysis. These statements clearly amount to a form of policy and industry advocacy.

General Patterns of Pre-Notice of Proposed Rulemaking (NPRM) Meetings

We have documented the regulatory work of Rodgin Cohen and S&C that, in our view, meets the substantive definition of lobbying activity but that almost entirely fails to appear in the LDA databases that social scientists tend to use, either because of intentional or unintentional nonreporting or, most likely in our view, through the plausible reliance on many reporting exceptions for regulatory advocacy. But how generalizable is their case? Are there others doing the same thing?

As mentioned earlier, although Cohen is the private individual who appears most frequently in the Federal Reserve’s records about whom it met with during rulemaking, he is very far from the only such individual. To these records, we now turn. We collected data about 905 meetings held with FRB officials related to the Dodd-Frank law between its enactment in 2010 and the conclusion of our data collection in 2018. These meetings are described in PDF logs prepared by the Federal Reserve staff and posted on the Federal Reserve website under a section titled “Regulatory Reform: Communications with the Public.” ¹³ These logs occasionally include richly descriptive materials, from PowerPoint presentations to journal articles or depositions. Typically, however, they are relatively spare and list only the attendees, their organizations, and a terse paragraph about the matters discussed. For our data collection, we scraped all available records at the time of data collection from the website, automatically extracted the names of the meeting participants and their organizations, and manually double-checked and standardized the names of individuals and organizations across logs; for example, we linked cases where an advocate’s name appears as “Rodge Cohen (SullCrom)” with those where the name appears as “H. Rodgin Cohen (Sullivan & Cromwell).”

The first thing to note about these meetings is that at least two studies have concluded they were venues of remarkable influence. Krawiec (2013) shows that industry interests dominated these meetings, with banks appearing particularly often, in contrast with the notice-and-comment process. As Krawiec and others (Carpenter and Libgober 2018) argue, it is well known that the Volcker Rule was watered down in different ways, both pre-NPRM and during the notice-and-comment process. Second, Libgober (2020b) shows that the attendance of BHCs at these meetings is associated with robust, abnormally positive, asset price movements as soon as 20 minutes after the rule is announced. Libgober gives these patterns a cautious interpretation, but they are consistent with third-party audiences (here, investors) detecting monetizable gains in the revenue and value of BHCs that participated in these meetings.

Across these 905 meetings, 6,155 people met with Federal Reserve officials about the Dodd-Frank Act. By comparison, 4,516 registered lobbyists reported lobbying activity on Dodd-Frank or a related predecessor bill. Although we do not conclude that the number of regulatory advocates is 50% larger than the number of registered lobbyists, it is safe to say that at least as many regulatory advocates met with the Federal Reserve (just one of more than a dozen implementing agencies) as participated in the officially measured lobbying process for Dodd-Frank.

With the names of these regulatory advocates in hand, we tracked down information on these individuals from scrapes of LinkedIn, ALM Legal Compass, the Center for Responsive Politics, and related public databases (Kim 2018). The first thing to note about these meetings is that the most senior, most powerful, and likely highest-paid individuals at these law firms are disproportionately represented. We found 319 exact name matches between those who met with FRB officials and the ALM Legal Compass; of these 319, fully 235 were partners, and 15 of these were not only partners but also firm leadership (e.g., “managing partner” or “executive partner”; Rodgin Cohen is just one of these 15). Clearly, law firms are not merely “showing up” to these meetings but regard them as matters of great importance, sending their most experienced, most powerful, and most expensive talent.

Table 2 shows information about the top firms by participation in Dodd-Frank meetings and several top firms by amount of annual reported lobbying expenditures. Sullivan and Cromwell tops the list of firms by meetings, with almost as many meetings as the nearest competitors Debevoise, Davis Polk, and Cleary Gottlieb combined. Notably, one can count on one hand the number of meetings attended by known lobbying heavyweights such as Patton Boggs and Akin Gump. Brownstein Hyatt, a top-five lobbying firm by reported expenditures, appears in no meetings. It is hard to overlook that all these firms are law firms. Yet in terms of professional prestige and status, it appears that the “white-shoe” set engages notably more in regulatory advocacy. ¹⁴ The American Lawyer magazine ranks law firms based primarily on the metric of profits per partner, and it typically places all but one of the top regulatory advocacy firms in the top 20 firms nationally. Even the least profitable of these firms (Debevoise) never falls out of the top 50. By contrast, law firms reporting substantial lobbying earnings have a much more tenuous position in the Am Law 200. Patton Boggs occasionally climbs to similar heights as the regulatory lobbying firms, but there are years it does not crack the Am Law 100. Akin Gump is consistently around the country’s thirtieth most profitable law firm. However, it is an outlier: most of the top reported lobbying firms are like Brownstein Hyatt, holding onto rankings in the bottom 100 of the top 200 US law firms.

Table 2. Top regulatory lobbying and reported lobbying firms are segmented

Source: Federal Reserve meeting logs. For reference, Federal Reserve governors Janet Yellen and Jerome Powell appear in 24 and 25 meetings, respectively. Daniel Tarullo, a governor particularly involved with Dodd-Frank implementation, appears in 52 meetings. Sean Campbell, deputy associate director of the division of research and statistics, is the most frequent meeting participant, having attended 94 stakeholder meetings.

Figure 1 makes this analysis more comprehensive through a scatterplot of all Am Law 200 firms between 2010 and 2017. It compares the profits per partner with reported lobbying earnings for all firm-years simultaneously. First, it is notable that many firms report zero dollars in lobbying revenue, and many others report hundreds of thousands of dollars, if not millions. Sometimes these are the same firm, just in different years. For example, the second-most prominent law firm in our meetings data (Debevoise) reports zero lobbying revenue in 2010 and a quarter-million dollars in lobbying earnings in 2011. Given the scale of reported lobbying earnings, these are more similar amounts: it is rare to see year-to-year reported lobbying earnings within a firm differ by orders of magnitude. Firm profitability is also quite persistent. Given that the position of each firm in this graph is relatively stable, it makes sense to consider how the top reported lobbying and top regulatory lobbying firms differ with respect to their positions in the field. Figure 1 confirms the segmentation and differences in profitability we alluded to earlier. Viewed as law firms, those with high revenue from reported lobbying are not especially profitable. Indeed, during this period, the average profits per partner in the five reported lobbying firms in figure 1 was $1.1 million, an enviable salary for most American workers. However, the average annual salary across the entire American Lawyer 200 law firms was more than $100,000 higher. In this sense, law firms heavily involved in reported lobbying are slightly “below average.” They are far inferior in terms of profits and professional status to the firms most involved in regulatory lobbying whose partners, on average, in this same period, earned $2.6 million, more than twice as much. Viewed as reported lobbying firms, however, firms such as S&C or Davis Polk hardly rate.

Figure 1. Lobbying revenues and profits per partner in the American Lawyer magazine’s Top 200 Firms (2010–17)

Notes: Each point represents a firm-year among the American Lawyer magazine’s annual survey of the top 200 US law firms (e.g., Sullivan and Cromwell in 2013). For presentational reasons the x-axis has been logged and reflects the amount of lobbying expenditures. The y-axis reflects the profits per partner on a linear scale. Orange triangles highlight the position of firms with consistently high levels of regulatory lobbying: Davis Polk, Sullivan & Cromwell, Debevoise, Cleary, and Morrison & Foerster. Blue-green squares reflect the position of law firms with consistently high levels of reported lobbying: Patton Boggs, Akin Gump, Brownstein Hyatt, Holland & Knight, and K&L Gates. Gray circles reflect law firms that may engage in both or neither, year after year or only occasionally. Whatever the case, they do not appear as the top firms on either dimension in our data. Lines show the mean annual profits per partner for each category of firm between 2010 and 2018.

Although the profits per partner metric enables a reasonable guess as to the difference in earning power of advocates at reported lobbying and regulatory lobbying firms, it is probably the case that the true differences are even larger because the regulatory advocates appear to be relatively senior within their firms. Put differently, Rodgin Cohen, as S&C managing partner, must earn more than the average partner and likely quite a bit more. ¹⁵ Table 2 highlights the identity of the top regulatory advocate for each firm and whether this person has registered as a lobbyist. In pertinent part, their story is like Rodgin Cohen’s: they are all relatively senior partners. Like Rodgin, all these lawyers have registered as lobbyists, although many with very limited consistency, especially in the post–Dodd-Frank era. Oliver Ireland, a former FRB general counsel and later partner at Morrison and Foerster, disclosed lobbying activities between 2001 and 2010 but not thereafter. Randall Guynn of Davis Polk disclosed lobbying only between 2009 and 2010 and not before or after. None of these titans of the legal profession and regulatory advocacy would appear to be significant influence-seeking actors based on lobbying reports.

Turning from the individual to the organizational advocates, we note that industry and law firm participants at FRB meetings appear more often than traditional industry associations. We measure the participation of industry associations such as the American Bankers Association (ABA), the Securities Industry and Financial Markets Association (SIFMA), other umbrella associations, and a range of state bankers’ associations. All these associations qualify as the natural kind of industrial lobbyist or advocacy organization according to classic accounts in political science (Bauer et al. 1963; Salisbury et al. 1987). Participants representing bankers’ associations attended 178 FRB meetings (19.7% of the sample). Law firms attended 209 meetings (23.7%), and large financial institutions (mainly BHCs) attended 360 such meetings (39.8%). Whereas S&C attended 5.6% of the FRB meetings, ABA representatives attended 3.9% and SIFMA representatives attended 2.8%. It is remarkable that a single law firm attended more FRB rulemaking meetings about Dodd-Frank than the top peak associations in US finance.

Finally, to support our earlier claims about substitution, we examined statistical predictors of the participation of law firms at FRB meetings, measured as the intensity of participation: number of participants from the organization or kind of organization present at the meeting. Table 3 displays the results from a regression of the number of law firm participants at a meeting, controlling for the time of the meeting, the number of total attendees (as a kind of denominator), the presence of large financial institutions, and the presence of any number of federal agencies (SEC, Treasury, Office of the Comptroller of the Currency, FDIC, etc.). Across specifications of this predictive regression, the presence of more industry association representatives is systematically associated with fewer participants from law firms. Using dichotomous indicators of whether any representative from that type of organization attended yields similar findings. Meetings at which one or more financial industry associations were present were systematically associated with a 19 (18.8%) percentage-point reduction in the probability of law firms attending the same meeting (z = -3.30, p < 0.001).

Table 3. When do more law firm participants appear in FRB meetings?

Notes: Robust standard errors in parentheses; estimate of alpha (distributional parameter in negative binomial regression) suppressed but statistically significant in each case.

Given the observational nature of our data, it is not possible to interpret the negative correlation between law firm participation and bankers’ association participation as causal. Law firms and bankers’ associations may simply appear at different meetings where different agendas are discussed or as different issues arise. Although the consistency of our findings with various fixed effects challenges simple selection stories based on time or topic, more complex selection stories are hard to rule out empirically. Nevertheless, these regressions buttress our argument that law firms can serve not only as a possible direct substitute for traditional industry associations but also as implicit or functional substitutes. If the effect is one of selection, it means there are important venues of regulatory advocacy at which selection pressures invite law firms and not industry associations. And as we know from the general statistics, these selection pressures lead to greater law firm participation compared to association participation generally.

Have We Undercounted Lobbyists by Excluding Regulatory Advocates?

We now turn the question asked by Thomas and LaPira (2017) and other scholars. How much of the lobbying universe is missed when we rely on legalistic definitions under the Lobbying Disclosure Act and exclude what these scholars call policy advocates or what we call “regulatory advocates”?

We note first that many registered lobbyists are in fact lawyers as well (see table 4). Examining the set of registered lobbyists between 1998 and 2020 (50,701 total) who have readily identifiable public LinkedIn profiles (5,435 lobbyists), ¹⁶ 4,371 list some kind of degree information in their profile; of these, 1,661 use either “jd,” “law,” “llm,” or “llb” to describe their degree. Among lobbyists whose expertise and training are observable via self-report, in other words, nearly four in ten (38%) list legal training or qualifications. Similarly, 4,516 lobbyists filed reports on Dodd-Frank or any of its predecessor bills, and of these 650 have public LinkedIn profiles with some degree information: 38.5% of lobbyists with degree information listed on LinkedIn indicate they are lawyers, a self-reported number that matches up well with the LinkedIn-based estimate of the broader registered lobbyist population.

Table 4. Rough estimates of lobbyist and regulatory advocate populations, from Dodd-Frank and in general.

Sources: Center for Responsive Politics (CRP), ALM Legal Compass, FRB meetings database.

Examining financial regulation patterns more broadly, we can begin to compare the number of people involved in lobbying on legislation and in regulatory advocacy. A total of 4,516 registered lobbyists reported lobbying activity about Dodd-Frank or a related predecessor bill. Yet looking at meetings with the Federal Reserve alone, a number almost 50% larger appeared at FRB meetings after Dodd-Frank was passed (6,155). Of these 6,155 individuals, some 953 were lobbyists according to CRP data. ¹⁷ That leaves more than 5,000 people who met with the Fed but do not have the same name as any individual and organization who ever registered as a lobbyist. What fraction of those who met are also “lobbying” in the true or fair sense of the term? Perhaps not all of them—but we believe, by any reasonable definition, that the number rests far above zero. Put one way, if just half of the never-registered regulatory advocates appearing at Fed meetings can be considered lobbyists of a sort, then the total aggregate of “lobbyists” active on this legislation represents an undercount by 64 percent, in large part because of the systematic undercounting of advocates who are active in rulemaking.

If we generalize this kind of exercise to consider what the total lobbyist population looks like and we compare it against the total population of lawyers that are eligible for inclusion in regulatory advocacy, we arrive at a similar general scale of potential error. The Center for Responsive Politics lists 50,701 lobbyists that it identified between 1998 and 2020 and lists 11,611 for 2017. Yet if we examine the population of lawyers in the ALM Legal Compass who work in the areas of Policy and Government, 29,655 attorneys appear, and fully 16,769 of these are at the partner level.

It is important to recognize that, just as official lobbying measures undercount what should count as policy advocacy, these measures of lawyer and law firm practice also undercount regulatory advocacy. For instance, the law group Milbank, as recently as two years ago, did not list a single lawyer in any regulatory or governmental practice group. Yet, it also reports on its website that “we frequently, and successfully, represent our clients before a variety of government agencies including the Federal Reserve, the Comptroller of the Currency, SEC, CFTC, NY DFS, and the Federal Deposit Insurance Corporation. Our work results in formal and informal approvals, staff opinions and no-action letters—including several that have expanded the law.” After a website update, Milbank began listing 21 lawyers in the related practice group with work experience justifying the representations its former website used to make. Our database of 29,655 lawyers working in Policy and Government does not mention “Milbank,” so indeed, these figures are also an undercount of the number of potential regulatory advocates. As we show in the following section, even quite conservative assumptions about the extent to which these lawyers are involved in regulatory advocacy and not merely litigation produces telling expansions of the estimated amount of actual “money in politics.”

How Much Money Is There in Regulatory Advocacy?

We have established that the regulatory advocacy population in finance is plausibly as large or larger than the registered lobbying population in finance. By examining the debate about the “number of lobbyists,” we address one issue about whether the neglect of regulatory advocacy leads us to miss the amount of lobbying (Thomas and LaPira 2017). Yet, what do available data suggest about the aggregate expenditure on these matters?

We begin with a simple fact: legal expenditures by large BHCs dwarfs their lobbying expenditures by orders of magnitude. Figure 2 displays the aggregate level of expenditure for legal expenses, lobbying, and PACs for five major BHCs over different periods of the last 20 years, with Goldman Sachs (Sullivan and Cromwell’s principal BHC client) displayed separately for illustration. These data are derived from quarterly consolidated financial statements of BHCs (Form FR Y-9C), often known as “call reports” because early versions of these reports were expected to be produced by banks in response to unpredictable “calls” of regulators. ¹⁸ BHC call reports are the most-used financial statement for banks. They are now reported quarterly for larger banks and semiannually for smaller banks, but all banks file these reports, and the bank’s CFO must sign it and acknowledge the possibility of jail time and million-dollar penalties for intentional misrepresentations and omissions (18 USC 1001, 18 USC 1007). Since 2002, BHCs have reported legal expenses above a percentage threshold of non-interest “overhead” expenses. The exact percentage threshold has shifted greatly over the years and, in some years, depends on the size of the bank as measured by assets under management. Using a hierarchical Bayesian model that considered these complex censoring patterns, we estimated the proportion of legal overhead expenses and imputed a range of possible legal spending values. For each of these firms, legal expenses generally are 100–1,000 times greater than their reported lobbying or PAC expenditures per year.

Figure 2. Legal, reported lobbying, and PAC spending for five leading BHCs.

Notes: Black lines indicate reported spending by these firms; gray areas convey the estimated legal spending derived from our Bayesian model. Changes to reporting thresholds cause black lines to disappear for some entities for some years. Note that because Goldman Sachs was not a BHC before the financial crisis, it did not complete the FR Y-9C, and so overhead expenses cannot be used to impute the extent of lawyer spending for Goldman at all prior to 2010. OpenSecrets data are used for the lobbying and PAC series.

To be sure, legal expenditures include numerous activities that are not regulatory advocacy and even do not have much to do with regulation, such as the writing of interorganizational and intraorganizational contracts, employment issues, and so forth. That said, some fraction of that legal spending is on regulatory advocacy and lawyer-lobbying. Even were that fraction less than 1%, it would still add up to substantial dollar amounts for the money in politics literature. Note that to visualize all three series simultaneously, the y-axis in figure 2 representing dollars in estimated spending is presented on a log scale.

To better understand how much of this legal spending is related to lobbying, we constructed a novel panel dataset of all BHCs and their legal expenses. Additional details on the database are available in the supplemental appendix. The dataset includes measures of BHC legal expenditure by year from 2002 to 2018 and a range of other measures, including (a) bank financials, including assets, liabilities, and interest-bearing and non-interest-bearing expenses; (b) civil and criminal actions and judgments as measured by federal court filings (measured by the Federal Judicial Center), (c) federal enforcement actions (actions and penalties) taken by three regulatory agencies (the Federal Reserve, the Office of the Comptroller of the Currency, and the Federal Deposit Insurance Corporation); and (d) the number of pre-NPRM meetings and NPRM comments made by the BHC on ongoing federal rulemaking at the Federal Reserve.

Because of mergers and bank failures, the resulting data yield an asymmetric panel but one that covers 798 BHCs that reported at some time from 2002 to 2018 (see supplemental appendix A). The total legal expenditure reported by BHCs in this data amounts to $113.9 billion dollars. This is different from popular media reporting on banks’ legal costs, which includes the monetary amounts of settlements that are not supposed to be included in legal costs according to FR Y-9C instructions (Griffin and Campbell 2013).

We analyzed these panel data using ANOVA and two-way fixed-effect regressions with lagged dependent variables (including Arellano-Bond estimation to account for problems that arise in using a lagged endogenous variable), using forms of the following estimating equation (see supplemental appendix A):

$$ {L}_{it}=\alpha +\delta {L}_{i,t-1}+{\beta}^{\prime }{\mathrm{X}}_{it}+\unicode{x03B3} {R}_{it}+{c}_i+{m}_t+{\varepsilon}_{it} $$

where L measures bank legal expenses, X is a set of control variables varying over bank and year, R measures regulatory advocacy factors (observed meetings and rulemaking comments), c specifies a set of bank-specific fixed effects, and m specifies a set of year-specific fixed effects. All Greek letters are parameters (vectors or scalars) to be estimated or represent unobservable error (ε). The vector X and the variable R can also include leads and lags of relevant variables, which we do not state here in the equation for reasons of simplicity and space, but which we report in relevant tables.

We conducted three sets of analyses with this model. The first retrieved the estimate of γ from the equation and examined the total expenditure attributable to a regulatory variable across the dataset. The second collapsed the dozens of independent variable measures we have to a handful of factors by means of principal components analysis and then examined a regulatory advocacy factor that combines meetings and comments, as well as the total legal expenditure associated with this factor across the dataset (controlling for the other factors, bank and year fixed effects, and the lagged dependent variable). We also conducted an analysis of variance in several forms: (1) an ANOVA with the factorized variables (continuous and coded as such); (2) an ANOVA with all the variables from the panel regression model (with two specifications, one of which removes the one-year lead of the meeting count variable); and (3) and as a supplement calculating the change in R-squared between models with and without the regulation factor.

Table 5 reports the results of these estimations, with the full models being available in supplemental appendix A. For each component of variance explained (first half of table 5) or legal spending per meeting (second half of table 5), the rightmost column reports the implied annual expenditure across firms per year. The basic ANOVA exercise produces an estimate explained by the regulatory advocacy factor of just under 1% (0.9%), which would imply annual spending by all BHCs of $137.5 million. Using ANCOVA with just the meetings variables produces larger estimates of variance explained. Including the lag and present-year value of regulatory advocacy meetings yields a variance explained of 2.3% (implying an annual total regulatory expenditure of nearly $358 million). Adding the lead of BHC regulatory meetings nearly doubles the variance explained, bringing it to 4.6% (with an implied annual expenditure of $696 million). Another glimpse comes from the added explanatory value of the regulation factor (3.05%) to the model, divided by the period in which the regulator factor varies most (2010–18). This produces a “spending explained” estimate of $3.7 billion, or close to a half-billion per year ($459M) from Dodd-Frank’s enactment to 2018.

Table 5. Estimates of BHC legal expenditure (in thousands of dollars) associated with regulatory advocacy

Source: Legal Expenditure from Board of Governors of Federal Reserve System, Consolidated Financial Statements for Holding Companies, Form FR-Y-9C (2002–18), deflated using FRB’s GDP Implicit Price Deflator; comments and meetings data from Libgober (2020a). Estimates from Arellano-Bond models exclude lead estimate (year before meeting), which are often statistically significant. See supplemental appendix A. Reported lobbying benchmarks come from the authors’ analysis of OpenSecrets bulk lobbying data.

In a saturated predictive model, each pre-NPRM meeting with regulators is associated with $14.4 million in legal expenses reported the year before by the same bank. There are 905 such meetings reported in our dataset. ¹⁹ Across the dataset the associated expenditure tied to regulatory advocacy would amount to $765 million dollars per year. A statistically preferable Arellano-Bond model that properly instruments for the lagged dependent variable (see the supplemental appendix) produces even larger estimates, with $34.4 million per meeting, implying almost $1.9 billion in annual expenditures. A one-unit increase in the factorized regulation variable that bundles meetings and comments is associated with $10.1 million in legal expenses. The total “amount” of factorized regulation reported is 1,966 “units,” for an aggregate associated expenditure of $1.17 billion per year from 2002 to 2018. For circumspection, we have mainly focused our discussion on the more conservative estimates, but either way, tens of millions of dollars in legal expenditure appear to be moving in synchronicity with pre-NPRM meetings in granular two-way fixed-effects models with lagged dependent variables and many controls.

To get a sense of how much more “lobbying” we might observe if we counted regulatory advocacy in finance alone, we note that total reported lobbying by BHCs and other commercial banks in 2019 was a “mere” $54 million. The panel estimates suggest that, controlling for temporal and bank-specific factors and for dozens of covariates measuring criminal and civil legal exposure, regulatory enforcement, mergers and acquisitions activity, and operating factors, 2.5 to 20 times the amount of reported BHC lobbying spending is moving in synchronicity with observable regulatory advocacy. ²⁰ A key feature of this estimate is that a large range of regulatory enforcement and adjudication factors are being controlled for, with bank-by-year variables measuring civil liability cases, judgments and appellate activity, regulatory enforcement cases by three federal regulators, and criminal enforcement activity as well.

These estimates seem large, and we emphasize that we are not aiming for precise causal estimates. Yet from interviews with observers of the industry, the estimates appear plausible. One reason is that banks bundle their advocacy with other expenditures. Although there certainly is a “spot market” of companies seeking and law firms providing one-off ad-hoc regulatory services, we do not believe the advocates who dominate these processes are available for one-shot ad-hoc representations. Rather, their advocacy is part of a larger set of client services, much of which involves purely transactional work. For this reason, practitioners like Rodgin Cohen only work for those firms that have a sufficient scale of nonpolitical work. Law firms make most of their money through billable hours and retainers that obligate the firm to be available for clients when they need it. To get a top regulatory advocate like Rodgin Cohen, a BHC must direct lots of billable hours to Sullivan and Cromwell and pay retainer fees. ²¹ If Goldman does not spend tens of millions on billable hours for Sullivan and Cromwell’s lawyers, then that bank does not get the services of Rodgin Cohen or other top partners in the regulatory advocacy space.

This means that not all the $780 million–$1.04 billion per year is directly spent on regulatory advocacy. Given the presence of bank-by-year–varying measures of enforcement and civil liability, however, it does mean that such money is likely inseparable from regulatory advocacy; thus, changes in regulatory advocacy result in BHCs spending billions of dollars on legal services and shift many billions of dollars across law firms.

Implications and Conclusion

The idea of lawyers, especially the transactional lawyers that our theoretical perspective has identified, engaging in policy advocacy would inform a range of literatures in political science, not only in American politics but also in comparative politics, as well as in economics, sociology, and law (Hacker et al. 2021). Although the activity of lobbyists and quasi-lobbyists in policy advocacy has long been studied, the activity of transactional lawyers as lobbyists has not. The implications of this realization are significant and would include, at minimum, the following considerations:

1. 1. The population of lobbyists, and the amount of lobbying activity. By generalizing our approach, a broader conception of lobbying could be revealed, with more expansive measurements. Thomas and LaPira (2017) and LaPira and Thomas (2017) have documented the amount of lobbying activity, and Cain and Drutman (2014) have shown that regulatory change shapes the work and influence of revolving-door lobbyists. If, however, legal practitioners are active in influencing policy through rulemaking, and if they are not coded as lobbyists, then these estimates, although accurate for “official” lobbying, undercount the population of those who influence public policy and do so by directly advocating for policy change within the organs of government.

2. 2. The amount of money in politics: Generalizing our approach would also help reconceptualize and differently measure the amount of “money in politics.” Ansolabehere, de Figueiredo, and Snyder (2003) and others have estimated the amount of money in politics, and a long follow-on literature has continued the endeavor of accounting. Substantial legal expenditures by firms (directly on legal representation, in-house on legal representation, and through associations) are excluded from this tally. Aggregate data suggest an immense, multibillion dollar per year increase in legal expenditures by BHCs immediately after the financial crisis and passage of the Dodd-Frank Act, an amount continuing for years and even decades after enactment. The research agenda should attempt to decompose the different kinds of regulatory advocacy spending that occur in response to regulatory enactments and rules.

3. 3. Political inequality: The first and second implications deserve more general study, but they can be linked to potentially immense inequalities in political resources. Bartels (2008), Gilens (2012), and Schlozman, Verba, and Brady (2012) have aggregated evidence of inequality in US national policy with wealthier populations regularly receiving better or more responsive policy outcomes in Congress, and wealthier interests having better, fuller, and more consistent representation (see also Schlozman 1984; Page, Bartels, and Seawright 2013). If political inequality is reinforced by the role of lobbyists in the rulemaking process, particularly in early, agenda-setting meetings (Carpenter 2023; Yackee 2012), these patterns may be accentuated. There is also the possibility that, in some policy areas, the influence of lawyers may ameliorate lobbying and representational biases at earlier stages of the policymaking process.

4. 4. Lobbying outcomes: Defining lobbying success or failure at the legislative stage—as in Baumgartner and Leech (1998) and Baumgartner and coauthors (2009)—is vital, but other forms of success may emerge only much later. Legislative lobbying success may be reinforced or reversed via regulatory advocacy, and these “next-generation outcomes” may happen not only through the text of rules but also in paths that were not taken or changes in the interpretations of rules published long ago, as when Cohen persuaded the Treasury Department to apply regulations in ways favorable to Bear Stearns. This point extends but is not reducible to the classic McCubbins, Noll, and Weingast (1987) argument on how winning/losing at the legislative stage (the enacting coalition) is connected to winning/losing in rulemaking (see also Balla 1998). The work of transactional lawyers in regulation should affect political scientists’ calculus of aggregate wins and losses over the long-run cycle of policy change, including considerations of paths that were possible but foreclosed by effective regulatory advocacy (Carpenter 2023; Yackee 2012).

5. 5. Strategies of influence: For companies engaging in “nonmarket strategy,” lobbying at the legislative stage and lobbying at the administrative stage may be complements or may be substitutes. Complementary relations may hold if firms wish to reinforce gains made at the legislative stage by ensuring that they are not diluted in implementation. Legislative and administrative lobbying may be substitutes if companies wish to avoid negative publicity from legislative lobbying (a more transparent process) or if companies see one venue as potentially risky in that lobbying will induce counter-lobbying (Austen-Smith and Wright 1994). We know of no study that examines the rationality of trade-offs between legal-lobbying influence over administrative policy and traditional lobbying influence over legislators.

The theoretical and empirical costs of not expanding our understanding of lobbying to include “lawyerly lobbying” in regulatory advocacy are substantial. Scholarly definitions of lobbying should not depend on the legislature’s decision to delegate policymaking, or else social scientists would be allowing legislative decisions to shape both the definition of politics and the venue and form of industry influence. Nor should scholarly definitions of lobbying be constructed in a way that, on their own, creates incentives to shift activity from one venue to another. To understand modern patterns of political influence, scholars need to examine developments at the nexus of two powerful forces: the rise of the administrative state (especially policymaking by rulemaking) and the emergence of corporate and regulatory law as a subprofession.