The threat has two layers. The first is structural: AI may automate mid-skill tasks faster than labor markets can generate complementary roles. The result would be a polarized labor market— jobs growing at the high and low ends while the middle hollows out—with productivity gains flowing to capital and high-skill workers. The second, less discussed, is what happens to workers whose jobs persist in altered form: degraded job quality through algorithmic management, intensified surveillance, volatile pay, and opaque discipline and deactivation. The first layer is no longer hypothetical. Stanford’s 2026 AI Index reports that employment for software developers aged 22 to 25 has fallen nearly 20 percent since 2024, even as senior, judgment-heavy roles remain intact—a pattern the report calls “seniority-biased technological change.”¹ Any serious governance response must address both layers.

The Law Already Requires a Lot

Our starting point is that the international legal framework for governing AI at work largely exists. Articles 6 and 7 of the International Covenant on Economic, Social and Cultural Rights (ICESCR) guarantee the rights to freely chosen work and to just and favorable conditions of work,² and the Committee on Economic, Social and Cultural Rights has made clear—in General Comments No. 18 and No. 23—that these are binding duties requiring active state measures, not aspirational ones.³ Applied to AI, these duties mean that states cannot passively allow automation to displace workers. They must fund retraining, guidance, and placement, and they must legislate, inspect, and enforce to curb algorithmic scheduling that cuts rest time, monitoring that heightens stress, and opaque systems that trigger unfair dismissals.

A word of candor for U.S. readers: the United States signed the ICESCR in 1977 but never ratified it, so the Covenant does not bind it as treaty law. American companies are hardly exempt, though: they must comply wherever they operate, ratifying countries are converting the Covenant’s guarantees into enforceable law—the EU’s AI Act⁴ and Platform Work Directive⁵ among them—and domestic employment and discrimination law imposes parallel duties at home. That gap between signature and ratification is exactly why the coupling mechanisms we describe below matter most in the United States.

Corporations share these responsibilities. The UN Guiding Principles on Business and Human Rights require firms to conduct ongoing human rights due diligence and remediate harms.⁶ In the workplace-AI context, that means bias-testing hiring and evaluation systems, ensuring human review and appeal for significant automated decisions, and maintaining effective grievance mechanisms. Algorithmic bias is not inevitable. As the well-documented case of a U.S. health-management algorithm that underestimated the needs of Black patients demonstrated, bias is design-dependent—and design choices are auditable, correctable, and therefore governable.⁷

Complementary Institutional Roles

National responses so far are a patchwork: the EU classifies workplace AI as high-risk and presumes platform workers are employees; the United States relies on state and local rules and, where statutes fall short, union agreements; China regulates content more than worker protections; and Global South jurisdictions like Brazil and South Africa are legislating ambitiously but enforcing thinly. Inconsistency at this scale invites regulatory arbitrage—firms can route AI-intensive operations to the weakest jurisdictions, undercutting any national floor. No single jurisdiction can close that gap alone, which is where the international institutions come in.

At the international level, the institutions’ roles are complementary rather than duplicative. The UN Human Rights Council and the Office of the UN High Commissioner for Human Rights supply the normative baseline, linking workplace AI explicitly to ICESCR Articles 6 and 7 and issuing benchmarks for courts and inspectorates. The International Labour Organization (ILO) provides the standard-setting muscle: Its platform-work discussions, opened at the 2025 International Labour Conference, are on track for a Convention and Recommendation, slated for adoption in 2026, that would elevate protections such as presumed employment under algorithmic control and human oversight to global standards—and activate the ILO’s supervisory machinery.⁸ The WTO, without becoming a labor agency, can sequence digital-trade commitments with labor-impact assessments and transition finance so diffusion does not outpace states’ capacity to protect workers. And the International Trade Union Confederation and global unions supply “ground truth”: they collect evidence from workers—unpaid waiting time, accounts deactivated without explanation, safety incidents—and feed it to the international bodies that monitor compliance, while spreading model AI contract language from one country’s unions to another’s, which functions as regulation before any statute is passed.

Because even well-governed AI will displace and reorganize tasks, the governance mix must also include guaranteed minimums of income security and health care for all workers—what the ILO calls “social-protection floors”⁹ —and a “Just Transition for AI” facility: an idea borrowed from climate policy, funding wage-loss insurance, rapid re-employment services, and training for the jobs that complement AI rather than compete with it.

This layered architecture turns abstract guarantees into routines: notice to workers before an AI system is deployed, pace limits written into workplace safety plans, appeal rights backed by income protection, and funded retraining for those displaced. The lesson for lawyers and policymakers is that the AI transition is a design problem, not a legislative one: the law we need largely exists. Soft law supplies methods and metrics; hard law supplies duties, remedies, and sanctions. The task is to combine them by design rather than by chance so that productivity gains arrive with enforceable rights, worker voice, and social protection.

Labor-Rights Impact Assessment

Our central contribution is an architecture for coupling the two bodies of law that govern AI at work: hard law (the ICESCR, binding ILO conventions, national labor codes) and soft law (OECD and G20 AI principles, UN guidance, corporate codes). A system of many rule-makers can work, but only when concrete coupling mechanisms—what we call “knitting hooks”—keep soft law tied to enforceable duties rather than substituting for them.

The operational core is a Labor-Rights Impact Assessment (LRIA): a structured, recurring assessment—defining the baseline before rollout, reviewing at six months and then annually thereafter—that employers conduct before deploying AI in hiring, evaluation, scheduling, pay, discipline, safety, or termination. It tracks fixed, measurable indicators (disaggregated error rates for nondiscrimination, wage dispersion and scheduling volatility for job quality, pace-of-work and psychosocial metrics for occupational safety) and is tied to remedies, including individual rights to notice, explanation, and appeal.

Four hooks then bind the LRIA to hard law. First, incorporation by reference: Statutes recognize LRIA templates only when tied to enforceable remedies such as reinstatement, back pay, and penalties—so conducting the assessment becomes part of complying with the law, not a substitute for it. Second, comply-or-explain with a public registry: Employers publish LRIA summaries against a common standard or publicly justify deviations, putting a company’s workforce-AI practices where investors, regulators, and plaintiffs’ lawyers can read them. Third, contract hooks: Procurement rules, licenses, and collective bargaining agreements convert the LRIA’s metrics into binding obligations with audit rights—meaning a vendor that cannot document its assessment loses the contract. Fourth, enforcement presumptions: A missing or noncompliant LRIA creates a rebuttable presumption of unlawfulness in disputes over dismissal, pay, or scheduling—shifting the burden to the employer precisely where records are in its hands.

None of this requires a new institution or a new agency; every hook runs through levers that legislatures, regulators, general counsel, and unions already operate. And that is the practical point for corporate counsel and compliance officers: The LRIA is not a future regulatory burden but a present opportunity. Run before deployment and documented well, it is evidence of diligence—in litigation, before regulators, and in the procurement processes that increasingly demand it. Absent or perfunctory, it is becoming evidence of the opposite. The companies that build the assessment into their AI governance now will set the standard the rest are eventually measured against.

REFERENCES

1. Stanford Institute for Human-Centered Artificial Intelligence, The AI Index 2026 Annual Report, ch. 4 (Economy) (2026).
2. International Covenant on Economic, Social and Cultural Rights, arts. 6–7, Dec. 16, 1966, 993 U.N.T.S. 3.
3. U.N. Comm. on Econ., Soc. & Cultural Rights, General Comment No. 18: The Right to Work, U.N. Doc. E/C.12/GC/18 (2006); U.N. Comm. on Econ., Soc. & Cultural Rights, General Comment No. 23: The Right to Just and Favourable Conditions of Work, U.N. Doc. E/C.12/GC/23 (2016).
4. Regulation (EU) 2024/1689 of the European Parliament and of the Council (Artificial Intelligence Act), 2024 O.J. (L 1689).
5. Directive (EU) 2024/2831 of the European Parliament and of the Council on Improving Working Conditions in Platform Work, 2024 O.J. (L 2831).
6. U.N. Guiding Principles on Business and Human Rights, U.N. Doc. HR/PUB/11/04 (2011).
7. Ziad Obermeyer, Brian Powers, Christine Vogeli & Sendhil Mullainathan, Dissecting Racial Bias in an Algorithm Used to Manage the Health of Populations, 366 Science 447 (2019).
8. Int’l Labour Org., Standard-Setting Committee on Decent Work in the Platform Economy, Int’l Labour Conf., 113th Sess. (2025).
9. Int’l Labour Org., Recommendation No. 202: Social Protection Floors (2012).

Michael A. Santoro is a professor at Santa Clara University’s Leavey School of Business, and Brewer D. Stone is a graduate of the University of St Andrews (Scotland) and a fellow of the AI, Ethics, and Human Rights Lab at Santa Clara University’s Leavey School of Business. This post is based on their recent article, “Artificial Intelligence, Human Rights, and the Future of Work: Global Governance and International Organizations in the Age of AI,” published in the Journal of Human Rights and available here. An open-access version is available on ResearchGate.

Background

Under Rule 23 of the Federal Rules of Civil Procedure, plaintiffs seeking class certification must satisfy two categories of requirements. First, Rule 23(a) sets forth four prerequisites: numerosity, commonality, typicality, and adequacy of representation.[1] Second, plaintiffs must show that their action falls into one of three categories under Rule 23(b): (1) proceeding without a class would create inconsistencies or variance, be dispositive of non-class members’ interests, or impair or impede these interests; (2) injunctive or declaratory relief is appropriate on a class-wide basis; or (3) common questions of law and fact predominate over individual questions and class adjudication is superior to individual adjudication.[2]

In Comcast Corp. v. Behrend,[3] the Supreme Court addressed the third of these categories, Rule 23(b)(3). The Supreme Court held that a class action against Comcast on antitrust grounds had been improperly certified because plaintiffs had not satisfied the predominance requirement of Rule 23(b)(3). The Court emphasized that plaintiffs must satisfy “through evidentiary proof at least one of the provisions of Rule 23(b)”[4] and that district courts must subject this proof to “rigorous analysis.”[5] To satisfy Rule 23(b)(3), plaintiffs must present evidence of a model capable of measuring damages across the entire class that is consistent with their theory of liability.[6] Applying these principles, the Court found the Comcast plaintiffs’ model lacking because it failed to attribute damages to “the plaintiffs’ only viable theory” of liability.[7]

Despite the Supreme Court’s exacting language in Comcast, the decision has not historically been treated as a significant hurdle for plaintiffs seeking class certification, particularly in federal securities cases. The Fourth Circuit’s decision in Boeing marks a significant departure and development.

Facts

In 2018, the Federal Aviation Administration (FAA) grounded all Boeing 737 MAX airplanes following two crashes that killed everyone on board.[8] Boeing was subsequently subject to investigations by the FAA, Congress, the Department of Justice (DOJ), and the Securities Exchange Commission (SEC).[9] In response, Boeing committed to improving airplane safety, including eliminating “traveled work”—a process under which airplanes are moved down the assembly line even when work at a particular station is incomplete.[10] During this period, Boeing’s officers made numerous public statements about the company’s commitment to safety, workplace culture, the rate of production, and regulatory compliance.[11]

But on January 5, 2024, Alaska Airlines Flight 1282—a Boeing 737 MAX—made an emergency landing after a plug in the plane’s fuselage detached while the plane was in the air.[12] The following Monday, Boeing’s stock price dropped by 8%.[13] The issue was later tied to traveled work, which had allegedly continued despite Boeing’s statements to the contrary.[14]

Procedural History

In January 2024, plaintiffs brought a securities fraud suit under Sections 10(b) and 20(a) of the Securities Exchange Act, alleging Boeing deceived investors through dozens of false and misleading statements about its commitment to safety that artificially inflated or maintained Boeing’s stock price.[15] All of the alleged misstatements survived Boeing’s motion to dismiss.[16]

Plaintiffs moved to certify a class, supported by an expert report from economist Chad Coffman. Coffman’s report briefly addressed the issue of damages under Comcast in only a few pages, contending that plaintiffs would rely on the out-of-pocket methodology—measuring damages as the difference between artificial inflation per share at the time of purchase and artificial inflation per share at the time of sale.[17] As to the calculation of artificial inflation itself, Coffman suggested a number of potential approaches, but did not settle on a final method.[18] Boeing opposed class certification, arguing that Coffman had not provided a definitive methodology as required by Comcast.[19] Coffman’s rebuttal report, while longer, still failed to commit to a methodology.[20]

The district court granted class certification, noting the broad acceptance of the out-of-pocket methodology and reasoning that it fit plaintiffs’ theory of liability—that investors were damaged by purchasing Boeing stock at inflated prices.[21] Less than a week later, Coffman submitted a merits report settling on a particular methodology for calculating artificial inflation.[22]

The Fourth Circuit’s Decision

Boeing appealed under Rule 23(f), arguing that plaintiffs’ expert reports failed to set forth a class-wide, case-specific damages methodology consistent with their theory of liability as required by Comcast.[23] The Fourth Circuit agreed and reversed.

Lessons Drawn From Comcast

The Fourth Circuit began by outlining Comcast’s instructions for plaintiffs and district courts.

For plaintiffs, the court outlined five requirements:

• Plaintiffs must present a damages methodology explaining “how damages will be measured in a specific case.”[24] It is not enough for plaintiffs to present “[a] menu of options that the party will decide on later”; plaintiffs must “tell the district court what their actual methodology is and how it resolves the predominance inquiry.”[25]
• Plaintiffs must demonstrate how damages can be measured on a class-wide basis so the court can determine whether class-wide issues predominate.[26]
• The proposed methodology must be consistent with plaintiffs’ theory of liability. That means that plaintiffs must identify “their theory or theories of liability to compare against their identified damages methodology.”[27] It is not enough for plaintiffs to identify various theories of liability upon which they may rely.
• The methodology “must allow a just and reasonable inference of damages”[28] and cannot “be speculative.”[29]
• Plaintiffs must submit evidence “that their damages methodology satisfies all of the above requirements.”[30]

For district courts, the Fourth Circuit reaffirmed that they must “conduct a rigorous analysis to determine whether the plaintiffs’ damages methodology satisfies Rule 23” under Comcast.[31]

Considering the District Court’s Decision Under Rigid Comcast Framework

Applying this framework, the Fourth Circuit found plaintiffs’ class-certification motion and Coffman’s reports lacking. The court criticized plaintiffs for relying on the out-of-pocket methodology, which the court saw not as a damages methodology under Comcast but instead just “a legal description of damages.”[32] Comcast requires a methodology that explains “how to determine the artificial inflation embedded in Boeing’s stock price on any day of the class period.”[33] Instead, Coffman provided “a series of maybes, perhapses and what ifs.”[34] While Coffman recognized the need to disaggregate stock price declines to isolate the effects of Boeing’s alleged misrepresentations and to measure artificial inflation,[35] he never committed to specific means for doing so, instead simply listing possible methodologies in both his initial and rebuttal reports.[36]

The court also criticized plaintiffs’ failure to commit to a specific theory of liability, which deprived the court of the required comparison to ensure consistency between that theory and plaintiffs’ damages methodology. The Court observed that plaintiffs had posited different theories of liability at various points and, while it acknowledged that “litigants sometimes don’t want to tip their hand,” it reasoned that plaintiffs must do so at the class-certification stage, under Comcast.[37]

The court concluded that plaintiffs “necessarily prevented the district court from conducting the required rigorous analysis.”[38]By failing “to properly identify the two comparators—the damages methodology and the legal liability theory—the district court’s certification order necessarily failed to perform a rigorous consistency comparison.”[39] The lack of a robust methodology also made it impossible for the district court to address Comcast’s other requirements.[40]

Finally, the court addressed plaintiffs’ harmless-error argument—that Coffman had since filed a merits report with greater detail. Noting that plaintiffs had disclaimed a harmless-error argument at oral argument, the court in any event found it “inappropriate to let a class-certification decision stand based on a report which the district court didn’t even consider” and observed that the Supreme Court has emphasized that the rigorous analysis required must be performed before a class is certified under Rule 23.[41]The court also questioned whether the merits report was sufficient because Coffman’s methodology was premised on the argument that all challenged statements fundamentally concealed “the same underlying truth”—that Boeing was taking shortcuts with respect to safety and quality control.[42] In the court’s view, this could not be consistent with plaintiffs’ theory of liability, which relied on misstatements describing different types of information over a three-year period.[43] Having chosen to “base their claim” on such a variety of statements, “[plaintiffs] are stuck with those allegations when applying Rule 23(b)(3).”[44]

Key Takeaways

The Fourth Circuit’s decision provides public companies and their officers with an arsenal of powerful arguments to defeat class certification. Historically, Comcast has not been treated by district courts as a significant test for plaintiffs seeking class certification in securities class actions. Should the Fourth Circuit’s approach gain traction, it will join a growing chorus requiring greater analytical rigor at class certification, including recent rulings applying the Supreme Court’s 2021 decision in Goldman Sachs v. Arkansas Teacher Retirement Systems.[45]

This increased rigor poses several challenges for plaintiffs. First, plaintiffs must identify a detailed damages methodology at an early stage and cannot defer implementation issues to later phases. Second, plaintiffs must “tip their hand” and settle on a theory of liability,[46] making reliance on generic or aspirational misstatements with weaker links to the ultimate loss-causing event more difficult. Third, plaintiffs who plead numerous alleged misstatements, hoping some will survive a defendant’s motion to dismiss, face increased risks because they may struggle to fit a multitude of disparate statements within their proposed theory of liability at the class-certification stage.

Notably, the Fourth Circuit’s opinion is unlikely to be the final word on this issue. The court’s attempt to sidestep plaintiffs’ counterargument that its reasoning was inconsistent with the Second Circuit’s decision in Waggoner v. Barclays PLC[47] hints at a possible circuit split, particularly given the district court’s statement that its reasoning was consistent with Waggoner and that of other district courts.[48] Also, the Fourth Circuit is not alone in requiring greater rigor under Comcast—the Sixth Circuit recently reversed and remanded class certification partly on a similar basis.[49] But decisions such as Waggoner and a recent Ninth Circuit order denying leave to appeal on Comcast grounds[50] suggest that other circuits may favor a more flexible and less demanding approach. In time, therefore, the Supreme Court may be asked to return to its ruling in Comcast and clarify how significant a hurdle it presents to plaintiffs in securities class actions.

ENDNOTES

[1] Fed. R. Civ. P. 23(a)(1)–(4).

[2] Id. (b)(1)–(3).

[3] 569 U.S. 27 (2013).

[4] Id. at 33.

[5] Id. at 35.

[6] Id.

[7] Officer of Gen. Treasurer ex rel. of Emps Ret. Sys. v. Boeing Co., 2026 WL 2083048, at *10 (4th Cir. July 20, 2026).

[8] Id. at *3.

[9] Id.

[10] Id.

[11] Id. at *3–4.

[12] Id. at *4.

[13] Id.

[14] Id.

[15] Id. at *5.

[16] Id. at *6.

[17] Id.

[18] Id. at *7.

[19] Id.

[20] Id. at *8.

[21] Id.

[22] Id. at *8–9.

[23] Id. at *9.

[24] Id. at *11.

[25] Id.

[26] Id. at *12.

[27] Id.

[28] Id. (quoting Comcast, 569 U.S. at 35).

[29] Id. (quoting Comcast, 569 U.S. at 35).

[30] Id.

[31] Id.

[32] Id. at *13.

[33] Id.

[34] Id. (quoting Speerly v. Gen. Motors, LLC, 143 F.4th 306, 342 (6th Cir. 2025) (citation modified)).

[35] Id.

[36] Id.

[37] Id. at *14.

[38] Id. at *16.

[39] Id.

[40] Id. at *15.

[41] Id. at *17.

[42] Id.

[43] Id. at *18.

[44] Id.

[45] 594 U.S. 113 (2021).

[46] Boeing Co., 2026 WL 2083048, at *14.

[47] 875 F.3d 79 (2d Cir. 2017).

[48] Boeing Co., 2026 WL 2083048 at *17.

[49] In re FirstEnergy Sec. Litig., 149 F.4th 587 (6th Cir. 2025).

[50] SEB Inv. Mgmt. AB v. Wells Fargo & Co., 2025 WL 1243818, at *7 (N.D. Cal. Apr. 25, 2025), leave to appeal denied, No. 25-3021, 2025 WL 2028400 (9th Cir. July 17, 2025).

This post is based on a Cleary Gottlieb Steen & Hamilton LLP memorandum, “Fourth Circuit Reverses Class Certification in Boeing Litigation, Establishing a High Bar Under Comcast,” dated July 30, 2026, and available here.

Every MNC managing a global supply chain, every institutional investor operating under a social-impact mandate, and every executive-education program purporting to prepare leaders for the coming decade confronts an identical structural deficit. There is no objective, continuously updated signal of how the public actually expects corporate social commitments to unfold. Disclosure regimes have produced an extraordinary volume of data. What they have not produced, though, is a functioning market in beliefs about the future.

The deficiency is not simply informational. Sustainability reports, third-party ESG ratings, regulatory filings, and NGO scorecards are, without exception, retrospective; they can only describe what has already occurred. None assigns a probability to what has yet to occur. Yet it is precisely that forward-looking signal, the crowd’s best real-time estimate of whether a given social or environmental commitment will actually be met, that matters most for strategic decision-making, capital allocation, and any accountability regime.

Social impact prediction (SIP) markets are our proposed answer. A SIP platform runs on the same logic as any conventional prediction market, and the participants stake positions on the probability of a verifiable future event, with payouts calibrated to the difficulty and precision of that event. The menu is reoriented from tradeable events to outcomes tied to the sustainable development goals, climate commitments, biodiversity thresholds, supply-chain governance milestones, and related corporate accountability metrics.

Building such a market need not be complicated. Firms can begin simply by sponsoring a curated slate of prediction questions tied to verifiable outcomes within their own sector, supply chain, or public commitments. Each question then does triple duty (market signal, statement of belief, and data point in a continuously updated probability map of corporate social performance) and, because payouts are contingent on accuracy rather than sentiment, the architecture rewards participants for surfacing genuine beliefs rather than performative ones. The resulting information is distinct from (and in important respects superior to) conventional social-impact disclosure, precisely because it reflects what informed participants collectively expect to happen.

The value of a SIP platform is not confined to trading returns, nor to whatever share of proceeds is earmarked for social allocation. Aggregating thousands of informed lay predictions about social and environmental outcomes yields a dataset of genuine analytical power, with no real counterpart anywhere in the policy-research toolkit or the corporate intelligence function.

The distinction is worth dwelling on. A SIP platform is built to capture what the public expects to be true of the future, not what has already happened, and that is categorically different information. It captures collective anticipation, risk perception, and the weight lay stakeholders assign to competing causal narratives. For governance practitioners, that signal does something third-party ESG ratings cannot: It is forward-looking, continuously updated, and drawn from precisely the audience whose perceptions ultimately determine reputational and regulatory exposure.

For instance, our case study of the Amazon Rainforest (a measurable, satellite-monitored ecosystem whose future depends, in substantial part, on the governance decisions of corporations with supply-chain exposure) illustrates both the practical potential and the theoretical stakes of this design. A SIP platform operating on Amazon-linked questions would generate a publicly visible, continuously updated probability estimate of deforestation trajectories that ESG analysts, institutional investors, export creditors, and trade partners can read in real time, introducing a form of soft accountability that complements regulatory policy without displacing it. The private sector’s leverage over such outcomes is a function of exposure.

SIP offers something the conventional wisdom rarely can: real-time, stakes-bearing engagement with the causal complexity of social outcomes. An executive who wagers on whether a supply-chain deforestation commitment will actually be met is reasoning about that commitment in a fundamentally different register than one who merely reads about it in a quarterly sustainability report; the act of prediction compels genuine deliberation rather than performative assent. That same logic, we think, should extend to the regulators who will ultimately decide whether SIP markets are permitted to scale. The existing regulatory environment for prediction markets is in rapid flux, and it is far from obvious that categories built for commodity derivatives are the right frame for an instrument that is, in equal measure, financial contract, civic-participation mechanism, and information good. Within this sui generis regulatory framework of its own, SIP might be giving the CFTC the affirmative category it has been missing.

William S. Laufer is the Julian Aresty Endowed Professor of Legal Studies and Business Ethics, Sociology and Criminology, and co-director of the Zicklin Center for Governance and Business Ethics, at the Wharton School, University of Pennsylvania. Eduardo Saad-Diniz is a senior fellow at the Zicklin Center for Governance and Business Ethics, Wharton School, and a researcher at the ARC Foundation (Manaus, Brazil). This post is based on their forthcoming essay, “Going Long on Social Impact Prediction (SIP) Markets.”

The case arose from a challenge to a financing transaction approved by the board of MPower Financing, a Delaware PBC whose mission is to provide international students with access to financing for postsecondary education.  In 2025, MPower’s board approved a debt-for-equity conversion that would increase two of its existing lenders’ combined ownership stake from 25% to nearly 85%, substantially diluting the other stockholders.  The transaction was approved by a disinterested special committee, which, advised by independent legal and financial advisors, evaluated alternatives.  Stockholders sued, claiming that the committee breached its fiduciary duties by not pursuing alternatives that allegedly would have maximized stockholder value.

In rejecting the stockholders’ claim, the Court refused to extend the Revlon duty to maximize value in the corporate sale context to a PBC and confirmed the substantial protection afforded by the PBC statute’s safe harbor.  Under that safe harbor, a PBC director “will be deemed to satisfy such director’s fiduciary duties…if such director’s decision is both informed and disinterested and not such that no person of ordinary, sound judgment would approve.”

The Delaware PBC, while relatively new compared to the Delaware corporation, has become an important corporate form for private and public companies seeking to express their mission and stakeholder interests.  But choosing a PBC is not a free pass.  PBC directors should identify with clarity those interests, including the company’s public-benefit purpose.  Their decisions should include a clear record that the relevant interests were considered.

Today’s decision affirms that Delaware’s PBC statute provides substantial protections for directors leading today’s mission-driven companies, large and small.

This post is based on a Wachtell, Lipton, Rosen & Katz memorandum, “Delaware Court of Chancery Reaffirms Core
Principles of the Public Benefit Corporation,” dated July 29, 2026.

This marks a striking reversal. For more than three decades after the collapse of the Soviet Union in 1991, privatization, capital market liberalization, and shareholder primacy defined the prevailing model of economic reform. Governments were encouraged to retreat from ownership, while SOEs were expected to become more like private corporations. That consensus found its most influential expression in Henry Hansmann and Reinier Kraakman’s landmark essay, The End of History for Corporate Law, which argued that the shareholder-oriented model had effectively become the endpoint of corporate governance evolution.[1] The broader law-and-finance literature, particularly the work of Rafael La Porta, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert Vishny, likewise emphasized investor protection and private ownership as foundations of economic development.[2]

For much of the post-Cold War era, corporate governance ideas flowed from advanced Western economies to the rest of the world. Governments privatized SOEs, liberalized markets, strengthened investor protections, and sought to insulate business decisions from political influence. At the same time, Lucian Bebchuk and  Mark Roe argued that corporate governance systems remain shaped by political contexts and ownership structures, suggesting that convergence would never be complete.[3]

China is a prominent example of this institutional divergence. Rather than privatizing strategic enterprises, Beijing corporatized them, introduced market incentives, and sought to separate ownership from day-to-day management while retaining ultimate state control. The establishment of the State-owned Assets Supervision and Administration Commission (SASAC) in 2003 institutionalized the state’s role as a shareholder. As Curtis Milhaupt has observed, modern state capitalism relies less on central planning than on governments exercising influence through sophisticated corporate ownership structures.[4]

China’s model came with significant criticisms, raising legitimate concerns about political intervention, inefficient capital allocation, and distorted market competition. But it also challenged a longstanding assumption: that state ownership is necessarily incompatible with commercial performance.

More striking is that governments across different political systems and legal traditions are moving in a similar direction, rediscovering the strategic value of ownership.

The United States remains fundamentally organized around private enterprise. Yet Washington increasingly treats ownership and corporate control as matters of national security. The CHIPS and Science Act illustrates a broader industrial policy that combines public investment with strategic oversight of critical industries. More recently, the federal government has moved beyond traditional subsidies by taking a 10 percent equity stake in Intel[5] and imposing a 15 percent revenue-sharing condition on chip exports by NVIDIA and AMD.[6]

These developments do not mean the United States is becoming China. The two countries remain profoundly different in their political systems, legal institutions, and economic structures. Rather, the United States illustrates a broader transformation: Governments are increasingly willing to shape corporate ownership, investment decisions, and strategic control in the name of national security.

This transformation is visible across the Pacific. China’s SOEs remain the world’s most prominent example of strategic state ownership. Indonesia—a nation of more than 17,500 islands and Southeast Asia’s largest economy—has recently embarked on its own experiment in state ownership through Danantara, a new sovereign wealth fund designed to consolidate and manage more than 1,000 SOEs. Singapore’s Temasek Holdings, by contrast, represents one of the world’s most successful examples of professionally managed state ownership, demonstrating how institutional separation between political leadership and investment management can allow SOEs to operate with commercial discipline rather than bureaucratic direction.[7]

The challenge is whether governments can govern the corporations they own.[8] State ownership creates a distinctive governance challenge because governments simultaneously act as shareholder and regulator. Strategic ownership succeeds only when governments clearly separate those roles. International standards, including the OECD Guidelines on Corporate Governance of State-Owned Enterprises, emphasize that such separation is essential to prevent political objectives from overwhelming commercial judgment.[9] Institutional separation also promotes board independence, managerial autonomy, accountability, and commercial decision-making, which are the core objectives of modern corporate governance.

This governance challenge is increasingly universal. Whether in Washington, Beijing, Jakarta, or Singapore, governments are confronting the same fundamental question: How can public ownership advance long-term national interests without undermining competition, accountability, or managerial autonomy?

Recent geopolitical competition suggests that the question is no longer whether corporate governance systems will converge, but how governments increasingly employ ownership itself as an instrument of economic policy. Rather than abandoning markets, states are redefining their role within them through strategic ownership and corporate control.

Hansmann and Kraakman were largely correct that shareholder-oriented governance became the dominant model within corporations. What they could not have anticipated was the extent to which governments themselves would return as shareholders in strategically important industries.

The twentieth century’s corporate governance debate centered on how corporations should be governed. The twenty-first century is increasingly asking a different question: How should governments govern the corporations they own?

The future of capitalism will therefore not be defined by a choice between market capitalism and state capitalism. Rather, it will depend on whether governments can exercise ownership through institutions that preserve commercial discipline, political accountability, and competitive markets. As states increasingly return as strategic shareholders, designing those governance institutions is becoming one of the central corporate governance challenges of the twenty-first century.

ENDNOTES

[1] Henry Hansmann & Reinier Kraakman, The End of History for Corporate Law, 89 Geo. L.J. 439 (2001).

[2] Rafael La Porta, Florencio Lopez-de-Silanes, Andrei Shleifer & Robert W. Vishny, Law and Finance, 106 J. Pol. Econ. 1113 (1998).

[3] Lucian A. Bebchuk & Mark J. Roe, A Theory of Path Dependence in Corporate Ownership and Governance, 52 Stan. L. Rev. 127 (1999); Mark J. Roe, Political Determinants of Corporate Governance: Political Context, Corporate Impact (Oxford Univ. Press 2003). See also Umakanth Varottil, Proliferation of Corporate Governance Codes in the Backdrop of Divergent Ownership Structures, 24 Competition & Change 471 (2018).

[4] Curtis J. Milhaupt, Governance and Ownership: The State as Shareholder, in The Oxford Handbook of Corporate Law and Governance 589 (Jeffrey N. Gordon & Wolf-Georg Ringe eds., 2018).

[5] Rohan Goswami, Intel and U.S. Government Reach Deal for 10% Stake in Chipmaker, CNBC (Aug. 22, 2025), https://www.cnbc.com/2025/08/22/intel-goverment-equity-stake.html.

[6] Nvidia and AMD to Pay 15% of China Chip Sale Revenues to U.S. Government, FIN. TIMES (Aug. 10, 2025), https://www.ft.com/content/cd1a0729-a8ab-41e1-a4d2-8907f4c01cac.

[7] Luther Lie, Danantara: An SOE Superholding à la Temasek?, THE JAKARTA POST (Nov. 28, 2024), https://www.thejakartapost.com/opinion/2024/11/28/danantara-an-soe-superholding-la-temasek.html.

[8] Ernest Lim, Corporate Governance of State-Owned Enterprises: A Comparative Perspective, 41 Oxford J. Legal Stud. 663 (2021).

[9] Organisation for Economic Co-operation and Development, OECD Guidelines on Corporate Governance of State-Owned Enterprises (3d ed. 2024).

Luther Lie is a corporate lawyer with experience in New York, Indonesia, London, and Singapore. He writes frequently about corporate governance and state capitalism.

A coalition of tokenholders organized to direct treasury assets to ANT tokenholders through rapid governance votes, which the association characterized as a hostile extraction campaign. The association responded by moving treasury funds to a more restrictive multi-signature wallet structure, citing the need to protect the community’s long-term assets. Months of public dispute followed, including the departure of key contributors and a widening divide between the association’s leadership and the broader community it was constituted to serve. By late 2023, the association announced it would wind down entirely and distribute the treasury to tokenholders who chose to redeem their ANT tokens at a fixed exchange rate.

This case may seem arcane at first, but the underlying questions will be familiar to any conventional governance practitioner. Who owns the brand and the going-concern value when residual authority sits with one entity and operational capacity sits with another? Who controls the treasury when legal custody sits with one entity, but governance authority is distributed among tokenholders? When the parties disagree, by what process is the disagreement resolved, and how durable is the result? These are foundation-and-subsidiary questions, asked in an industry where the apex entity is a body of widely dispersed tokenholders, and where the answers are being worked out in public, across borders, at scale, on decentralized networks. The digital asset industry has accumulated more of these stress tests in the last several years than almost any other corner of the corporate world, and the experience is generating governance lessons.

One lesson has emerged repeatedly: Governance structures tend to perform under pressure only when three conditions are present: independence, skill, and will. The digital asset industry’s governance experiments provide unusually public examples of why each matters.

Zooming out, every frontier technology evolution eventually confronts many of the same governance questions, and these questions can in turn prove valuable to the standard corporate governance corpus. When the technology is novel, the operators are unconventional, the rules are unwritten, and the appetite for risk is high, structural choices made in those early years carry consequences well beyond their moment.

Founders and builders are rationally focused on immediate needs: shipping product, building networks, attracting capital. Governance architecture tends to wait. The tension between achieving high speed and governance rigor is predictable and, to a point, appropriate. It becomes a problem when the governance work that was deferred becomes a stress test the enterprise fails in a key moment.

In many instances, traditional governance frameworks provide tried-and-true answers for frontier tech companies. Some creativity may be needed to fit a new case, but existing practices have generally earned their standing. Occasionally, however, frontier operators do need innovative thinking on corporate governance to complement their technology and product innovations because of certain truly novel characteristics.

Both dynamics have played out across digital assets, blockchain, and decentralized networks over the last several years. The area merits close examination for governance lessons because the technology and the corporate structures genuinely depart from legacy frameworks in certain respects. Governance insights from these areas are broadly applicable to anyone designing accountability structures for frontier tech, including artificial intelligence or any industry where the pace of development outruns the institutional frameworks built for traditional systems.

The Organization of Digital-Asset Projects

For a corporate governance audience accustomed to the conventional Delaware C-corporation, the entities that govern major digital asset networks can look unfamiliar. Digital asset projects often operate through a two-tier structure. First, a foundation, organized either as a non-share-capital company or as a purpose trust, sits at the top. It owns the intellectual property of the underlying protocol, has custody of the treasury, and holds the ultimate governance authority of the network. Second, an operating company, usually a for-profit entity domiciled in a separate jurisdiction, employs the developers, ships the software, and runs the day-to-day business. There can be additional entities too, such as a separate structure for a software protocol intended to operate in a decentralized manner.

This bifurcation is the product of three converging pressures.

The first is regulatory. In many major jurisdictions, the lines between a security, a commodity, a money-transmission instrument, and an open-source software project remain unsettled. Separating the entity that develops the technology from the entity that stewards the protocol can reduce certain regulatory exposures, distribute liability across structures, and clarify the legal and tax position of each.

The second is philosophical. Decentralized protocols often aspire to be operated by their participants over time, not owned by a sponsor company indefinitely. A foundation provides the legal means through which that progressive decentralization can occur. Regardless of what happens with the operating company, the foundation persists as steward of the protocol’s intellectual and economic commons.

The third is jurisdictional. Most digital asset foundations are domiciled offshore, in places like the Cayman Islands, the United Arab Emirates, or Switzerland. These jurisdictions have purpose-built foundation regimes, well-developed corporate governance case law, and statutory infrastructure for non-shareholder fiduciary structures that most other jurisdictions do not. For projects with global stakeholders and significant operational complexity, access to established legal frameworks is often a significant advantage.

The result is that a major digital asset project may have, on paper, a board of directors that bears legal responsibility for tokens with billions of dollars in public market token value, a stakeholder community of tens or hundreds of thousands of tokenholders worldwide, and a regulatory exposure potentially spanning every major financial jurisdiction. The people sitting in those board seats are among the most consequential governance actors in the digital asset industry.

The Independent Director’s Role

The board of a digital asset foundation has duties that may differ from those common in traditional corporate or nonprofit governance. The foundation does not have shareholders in the conventional sense. Tokenholders carry economic interests and, in some networks, on-chain voting rights, but their formal contractual standing varies dramatically by network and is generally weaker than that of a public company shareholder. The foundation’s beneficiaries, in the legal sense, may be the community of network participants, the public interest in a functioning decentralized network, or some combination of both. There might also be equity investors in the operating company who received token allocations, creating crossover claims that span multiple entities. The fiduciary duties owed by foundation directors run, in most jurisdictions, to the foundation itself and to its stated purposes, although usually not to stakeholders such as tokenholders unless explicitly stated in the foundation’s constitutional documents.

The set of parties owed duties may be disorienting, but the protective function those duties perform is more familiar. The independent director’s job, in this domain as in any other, is to ask the questions that interested parties cannot reliably ask themselves. A founder evaluates a transaction from the perspective of the project’s growth. A tokenholder evaluates it from the perspective of price. An investor evaluates it through from the perspective of return. None of those perspectives is wrong, and a well-run organization needs all of them. None of them, separately or together, asks whether a transaction is appropriate, prudent, and consistent with the duties owed to the network and its broader stakeholder community. This is one example of the independent director’s job on a digital asset foundation.

This structural separation has a long history in conventional corporate governance. Listed companies are required by statute to maintain independent audit committees because the audit function cannot be delegated to people with an interest in the audit’s conclusions (Sarbanes-Oxley Act, Section 301). Stock-exchange listing rules require boards to have a majority of independent directors on the same logic (NYSE Listed Company Manual; Nasdaq Listing Rule 5605). The digital asset foundation model imports the principle into a new domain.

What Governance Pressure Looks Like

Most of the time, the role is unobtrusive. The directors maintain the legal, administrative, and procedural infrastructure that lets the operating team focus on building and operating. Routine board work rarely draws public excitement or interest. Done well, the routine itself builds the institutional credibility required for the board to act decisively when conditions change. Trust between board and operators is built in the routine.

When acute pressure arises, the structural fragility of the foundation governance model is tested. Conflict may emerge between the foundation and the operating company over the use of treasury assets, the disposition of intellectual property, or the strategic direction of the project. These conflicts are common in the second or third year of a project’s life, when the parties’ interests have begun to diverge from the unity that prevailed at launch.

Another source of conflict is regulatory pressure. The Securities and Exchange Commission, the Commodity Futures Trading Commission, the Department of Justice, and their counterparts in the European Union, the United Kingdom, and offshore jurisdictions have all moved from observation toward enforcement. Even where the regulatory direction has become more constructive, the volume and seriousness of oversight has materially increased. A foundation board’s documented governance process, the extent of its director independence, and the substance of its deliberations have begun to appear in licensing applications, settlement narratives, and enforcement records.

Then, there is counterparty pressure. Network participants, market makers, custodians, exchanges, and protocol partners present arrangements to the foundation that range from ordinary commercial dealings to structures whose terms benefit the counterparty at the foundation’s expense. Distinguishing between the two is, in practice, among the board’s most important tasks.

In our experience, the governance failures that create the most damage rarely stem from bad intentions. They emerge when authority, incentives, and accountability become misaligned under pressure.

Novel Challenges

One aspect making this role difficult in practice is the complexity of responsibilities and actors. The board has to be genuinely empowered to act independently of the founders and operating company it contracts with for services, and that empowerment has to be drafted into the constitutional documents, service agreements, and information-access provisions. Cayman foundation regimes and analogous BVI trust structures provide thinner statutory default protections for incumbent directors than onshore corporate law, leaving the foundation more reliant on its founding documents than a Delaware corporation.

The directors also need the experience to recognize familiar failure modes, which is harder than it sounds in an industry that moves at the speed of high-frequency trading. Failures in the digital assets sector include market-making contracts whose compensation is structured to pay off when the protocol’s token declines, vesting schedules engineered for insider exits, governance proposals that route treasury value to specific wallets, regulatory inquiries that read as routine but signal coordinated examination, validator concentrations that create silent single-points-of-failure, and counterparty contracts that exploit asymmetries in technical literacy.

Not only do directors have to be technically proficient enough to spot and understand these issues, they have to be willing to use what they recognize, in an environment where dissent against a project’s commercial ambitions can be mobilized against quickly and publicly, on social platforms and in governance forums, including by anonymous accounts with millions of followers in some cases.

None of these qualities substitutes for the others. A structurally empowered board without practiced judgment misidentifies the moments where the empowerment matters. A board with practiced judgment but without structural empowerment becomes commentary. A structurally empowered, judgment-rich board whose members are not willing to absorb the personal cost of dissent ratifies what the operating entity proposes. The architecture of frontier industry governance that experience has taught us to design for is one in which each of these is addressed at the constitutional stage of a project, rather than left to the personal qualities of whoever happens to occupy the seat.

The Frontier Industry Generalization

The digital asset industry has arguably produced more public stress tests of novel governance structures in a shorter period than any other sector. They have yielded lessons about how governance institutions perform under stress when the underlying technology, the participant stakeholders and communities, and the regulatory environment all confound standard governance practices.

Consider artificial intelligence. The leading laboratories have adopted novel governance structures of their own: nonprofit boards overseeing for-profit subsidiaries, capped-profit arrangements, and various forms of independent trust. These structures are being designed under competitive and commercial pressure, by institutions trying to balance growth imperatives against safety obligations to a public that has not yet fully articulated what it expects. The structural questions these institutions face are the same ones digital asset foundations have been working through: Who are the independent directors, what authority do they have, what happens when the operating entity wants to do something the oversight body considers contrary to its mission, and is the oversight body structurally empowered to stop it?

The 2023 governance episode at OpenAI offered early evidence of how these questions resolve under pressure. A board that appeared to hold the authority to remove the organization’s operating leadership found that authority contested in practice by the operating organization, by employees, by investors, and by the weight of public perception. The formal architecture survived, but the practical question of who governs whom was resolved largely by dynamics external to the governance design itself. Whatever one concludes about the merits of the underlying decision, the episode revealed a gap between the authority the board was designed to possess and the authority it was able to exercise when called upon. That gap is a design problem, and it is the same design problem digital asset foundation boards have been working through in less visible ways for years.

Anthropic has approached this design problem in its own way. Its Long-Term Benefit Trust appoints directors to the Anthropic PBC board, and as of 2026 Trust-appointed directors form a majority of that board. The arrangement is not beyond stockholder reach: a supermajority of stockholders can amend the Trust’s powers, and the threshold required to do so escalates over time. The mechanism keeps the oversight body partly accountable to the capital that underwrites the enterprise, while making that accountability progressively harder to invoke. Whether this calibration resolves the underlying tension between mission and investor return, or only defers it to a different moment, is one of the open structural questions of the next several years.

The lesson is straightforward. Novel governance structures only protect what they were designed to protect when independence, skill, and will are present together. Two of the three will not do. A skilled, independent board without the will to act becomes, in practice, a board that ratifies whatever the operating entity proposes. A willing, independent board without the experience to identify the moments where its will is needed fails at the critical juncture. A willing, experienced board without genuine structural independence finds its authority withdrawn precisely when it is most required. The three properties have to be present together. Those designing frontier-industry governance frameworks should be testing for all three at once.

Implications for Governance Designers

There are implications for corporate-governance researchers,  lawyers, and regulators who will evaluate these structures. The doctrinal vocabulary of independence is well-developed in conventional corporate governance, but what about governance designed for stress, where independence, skill, and will are tested simultaneously? The digital asset experience is generating the case material that the academic literature can build from.

For practitioners, the prescription is the same in every frontier domain. Treat board composition as a structural choice rather than an administrative one. Hire directors with the experience and the will to say no. Build in the structural independence that allows them to do so. Compensate them at rates consistent with the responsibility being delegated. Bring them into the work before the decisions that test structural integrity arrive on the table.

The governance infrastructure being built now, in digital assets and in artificial intelligence and in the industries that will follow, will set precedents for how the next generation of frontier enterprises is held to account. Done with care, it demonstrates that novel organizational forms can carry forward the accountability principles that conventional corporate governance developed over decades. Done without that care, the exposure falls on the founders themselves, on the communities these projects serve, and on the institutions that extended them trust at the moment when trust is called upon. 

Cobus Pietersen is a co-founder of Caliber, a provider of independent-director services, and he advises digital asset projects and investment structures on governance and fiduciary matters Marc Piano is a partner at Horizons Global, a provider of corporate-governance services, and he is an independent director and governance professional based in the Cayman Islands

Public companies were not persuaded of the wisdom of the reform. In comment letters to the FASB, Starbucks, Pfizer, General Motors, Boeing, CIGNA, Uber, Marathon Oil, and other firms argued that the required breakdown is (i) more detailed than investors need, (ii) poorly aligned with how they manage their businesses, and (iii) too costly to justify the benefit. While we cannot precisely estimate the cost argument, in a new paper, we test their objection that disaggregated labor-cost information is not useful enough to be worth disclosing.

In a recent post on this site, the authors showed that one dimension of disaggregation carries information: Firms that voluntarily separate selling costs from administrative costs reveal economically distinct activities. We examine a complementary dimension that the new standard targets: the labor content of expenses, and how finely it should be disaggregated before it helps investors.

Measuring labor costs is hard precisely because firms do not disclose them. We use data from Revelio Labs, which standardizes millions of public employment records to estimate firm-level compensation, to build wage measures for roughly 25,000 firm-years from 2009 to 2022. The data let us split labor into three functional categories, general and administrative (G&A), sales and marketing (S&M), and research and development (R&D), a detail that no mandated financial-statement data currently provide. We validate our wage estimates against three benchmarks: voluntarily disclosed staff expense, median employee pay from proxy statements, and an industry-imputed wage measure based on voluntary disclosures. We caution that the Revelio measures are estimates, not audited figures, and are subject to selection and measurement error. Both sources of noise bias our tests toward finding smaller effects, so our estimates should be read as a conservative floor on the usefulness of audited labor-cost information.

Our first finding is that detail matters, and the useful detail is functional. Separating SG&A into aggregate wage and nonwage components yields only modest gains in predicting future performance. The larger gains arise when wages are split by function. The three components behave in economically distinct and intuitive ways: S&M wages track near-term revenue growth, R&D wages are the most informative about long-horizon revenue growth and future SG&A intensity, and G&A wages are comparatively uninformative about future fundamentals. These patterns persist when we control for the matched nonwage expense within each function, so labor is carrying information that nonlabor costs do not. Additionally, the gains are largest among firms with smaller workforces, where hiring choices map more directly onto commercialization and innovation.

The information also has capital-market consequences. Periods of high wage volatility, particularly in G&A and R&D wages, are associated with larger analyst revenue-forecast errors and with greater market illiquidity. Firms that voluntarily disclose aggregate wages attenuate some of the G&A-related uncertainty, but the effect is imprecise and does not fully resolve the forecasting errors or liquidity costs tied to more forward-looking inputs such as R&D. Taken together, the evidence is consistent with labor-cost disaggregation providing information that analysts and investors do not already have, and the usefulness of both aggregate and disaggregated labor costs.

What does this mean for the standard? The companies’ empirical premise does not hold up: Labor-cost information is useful, and the market prices its absence. More important, our evidence cuts in a second direction that the debate has largely missed. The standard requires compensation to be disclosed within the functional captions firms already present, but it does not require firms to separate R&D, S&M, and G&A on the face of the income statement, meaning that the labor-cost disaggregation our results find most useful may not be disclosed. Because the largest predictive gains come precisely from splitting labor functionally, a rule that stops at compensation-within-existing-captions is a step forward, but it can be improved by requiring firms to report R&D, S&M, and G&A in the income statement.

We caution that we speak only to the benefits side of the ledger. We do not observe firms’ compliance and preparation costs, and a complete cost-benefit verdict will have to wait until the standard has been in force for several years. On the benefit side, however, the data provide support for more disaggregated labor-cost disclosure and a reason to think the FASB can go further in the future.

Yue Chen is an assistant professor at The Chinese University of Hong Kong, Kalash Jain is an assistant professor at Columbia Business School, Nan Li is an assistant professor at Rotman School of Management, and Shivaram Rajgopal is a professor at Columbia Business School This post is based on their recent paper, “Decision Usefulness of Disaggregated Labor Costs: Ex-Ante Evidence and Implications for the FASB’s Expense Disclosure Mandate,” available here.

The Upside Is Limited

The policy underlying the push to open private markets to retail investors rests on two claims: that private markets deliver superior risk-adjusted returns and private assets improve diversification. The Proposal takes both as given, citing isolated studies. But a careful look at high-quality studies shows that both claims are disputed and unresolved at best.

The timing also raises serious concerns. Private equity faces a multi-year shortfall in distributions, a global backlog of portfolio companies, depressed exits, and interest rates that have strained the business model, and traditional institutional investors have responded by scaling back their allocations to the asset class. As recently as two years ago the private equity industry was telling regulators that the success of private markets depended on serving institutional investors exclusively. That reversal reflects no change in the underlying facts, only a change in sponsors’ fortunes.

Even if we assume  that institutional investors outperform public markets in their private equity investments, retail investors cannot expect to replicate those returns. Private markets cannot be indexed, and wide variation in returns is the norm. Outcomes depend on manager selection, access, bargaining power, timing, and negotiated terms—precisely where retail capital is systematically disadvantaged. Meanwhile one feature is entirely certain: higher fees. Compounded across a lifetime against near-zero index fees, even a modest fee drag will result in an enormous transfer from savers to asset managers.

The Downside Is Underexamined

The Proposal says little about the risks, and there are three in particular worth emphasizing.

First, retail investors have a well-documented history of poor decision-making even in the world’s most transparent and liquid markets. They chase past performance, ignore fees, respond to marketing rather than fundamentals, and select –high-price options when identical, cheaper ones are available. If decades of regulatory effort to make public markets investor-friendly have not fixed this, there is little basis for optimism about private markets, which are far more opaque, illiquid, complex, and heterogeneous.

Second, the likely vehicle for delivering 401(k) access to private investments compounds the problem. The approach  most attractive to private asset managers will be to embed a percentage of private assets inside 401(k) target date funds—vehicles that hold an assortment of assets and serve as the default for tens of millions of participants. That structure makes costs and risks nearly invisible. Participants who cannot identify high fees when plainly disclosed will certainly not detect them inside target date funds or attribute underperformance to the private assets in those funds.

Third, opening private markets to retail investors would erode the very features that make private  markets valuable. Private equity has long created value at the portfolio company level through active ownership with a long-term horizon and freedom from the constraints on public companies. Each would be undermined by retail capital: Funds will grow larger and less responsive, retail liquidity needs will drag on any illiquidity premium in private markets, and more regulation and litigation will almost certainly follow a flood of retail capital. The result is likely to be a market that has private equity’s high fees and opacity while delivering no more than a public index.

The Limits of Plan Fiduciaries

For these reasons, the magnitude of any net benefit that retail investors could theoretically receive from private markets is much smaller than the Proposal acknowledges, even in the best-case scenario. Critically, that scenario requires that retail investors rely on intermediaries to adequately represent their interests in the unforgiving private markets. The Proposal assumes plan fiduciaries will be able to play this role. We are much less optimistic.

The historical record gives little reason for confidence. It has been orthodoxy among financial economists for nearly 50 years that retail investors are best served by low-cost index funds. For decades after the introduction of index funds, however, plan fiduciaries funneled savers into high-fee active products that systematically underperformed, often offering only expensive share classes in cases where an identical, cheaper one existed. An extensive literature shows a long history of 401(k) fiduciaries making conflicted decisions that failed to optimize the retirement outcomes of plan beneficiaries.

This pattern broke in the late 2000s, thanks in significant part to ERISA fiduciary litigation and more prescriptive oversight. The Proposal treats the resulting narrowing of plan menus toward low-cost index funds as a harm litigation caused, but it should be framed as one of the great welfare gains in the history of American retirement saving. The Department of Labor rightly notes that the costs of litigation are real and, as in every area of the law, that at least some litigation has been meritless. But the Proposal’s focus on only the cost side is not a balanced accounting.

We are not convinced that the processes set forth in the safe harbor will resolve these concerns. One important reason: Under the terms of the Proposal, a fiduciary lacking the skills to analyze private market investments is directed to seek assistance from a “qualified investment advice fiduciary.” The Proposal assumes that such market participants will be able to offer independent assessments, but the harsh reality is that no unconflicted advice will realistically be available in these circumstances.

While private asset managers and their affiliates obviously have the greatest expertise, it goes without saying that they have a vested interest in moving more 401(k) money into the system. Retail asset managers might appear neutral, but after two decades of fee compression due to the rise of passive products, a large-scale return to actively-managed products lacking objective benchmarks would provide a massive windfall. Similarly, while there are scores of private market consultants, their continued existence relies on clients allocating to private markets. All of these market players thus face profound conflicts, and because performance data and benchmarks are easily manipulable, it won’t be difficult for an adviser to assemble a record satisfying the safe harbor factors.

This conflict can be readily illustrated. BlackRock, the largest of the retail asset managers, has a white paper claiming that private-market exposure would allow 401(k) participants to retire with up to 15 percent more assets, based on certain assumptions buried in the endnotes. These assumptions include 11 percent annual private-equity returns against 8 percent for public equities, and 10 percent private credit returns against 5 percent for public fixed income, sustained for 40 years. These assumptions are unreasonable: When private credit beats public equities by 200 basis points a year for four decades, these projections are clearly inconsistent with basic principles of corporate finance.

Defined Benefit Plans Are No Precedent

Proponents of private assets in 401(k) plans often analogize defined contribution plans to defined benefit plans. Defined benefit plans also have retirement savers as their ultimate beneficiaries and have been investing in private markets at scale for decades, so why shouldn’t defined contribution plans do the same thing? The DOL invokes this logic throughout the Proposal when it treats the former’s performance as illustrative of the latter.

But these two forms of intermediation differ in fundamental ways. At the level of incentives, defined benefit sponsors bear investment outcomes on their own balance sheets and face political scrutiny and legislative oversight. 401(k) fiduciaries, on the other hand, share neither the upside nor the downside of participants’ investments and face little labor-market or reputational discipline, since employees do not choose jobs based on menu construction. Those incentive differences matter more in private market investments, where performance is harder to measure, terms are more complex, valuations are discretionary, and outcomes turn on access and sophistication.

Structural differences cut the same way. Defined benefit plans persist indefinitely and have predictable payouts, giving them time horizons ideally suited to illiquid assets. Defined contribution plans face job-change rollovers, hardship withdrawals, and liquidity needs that are hard to predict and that spike during downturns, creating fire-sale risk precisely when private assets are hardest to value and sell. Intermediation is also far more complete in a defined benefit plan, where the fiduciary makes every allocation decision and thereby shields beneficiaries from return chasing, panic selling, and naïve diversification. And because defined benefit plans rarely transact at interim valuations, they can wait for cash-flow realizations, while defined contribution participants will unavoidably need to transact at estimated valuations sometimes—exposing them to dilution, first-mover advantage, and rushes to redeem when interim marks turn out to be wrong.

Retail Returns

In public markets, intermediation quality hardly matters. This is the crowning achievement of the last two decades of 401(k) practice: A mediocre fiduciary and an excellent one produce nearly the same outcome, because the product does the work and requires zero monitoring. Private markets offer no such backstop. There is no index, no passive vehicle delivering the asset class return at low cost, and enormous dispersion between top and bottom managers. Retail investors will earn exactly what their intermediary is capable of earning for them, in an environment where private market returns will likely be continuing a downward trajectory.

How good will the DOL’s proposed intermediation approach be? When we see fiduciaries that internalize none of the outcome, that face negligible monitoring, that are advised by parties who profit from saying “yes,” that leave allocation and redemption decisions to beneficiaries, and that are insulated from the litigation that has been the primary source of discipline on their conduct, we see little reason to be optimistic.

William W. Clayton is a professor at BYU Law School, and Elisabeth de Fontenay is a professor at Duke University School of Law. This post is based on their recent comment letter, “Private Assets in Defined Contribution Plans: A Comment Letter on the 2026 DOL Proposal,” available here.

Earlier this month, the U.S. District Court for the Eastern District of New York answered all these questions in the affirmative.[2]  Relying principally on the plain language of the statute, the court held that the automatic stay applies to successive motions to dismiss, even when some claims survived a prior dismissal motion and even when the successive motion seeks only partial dismissal.  The court also held that the stay applies not just to fact discovery but also to expert discovery and class certification proceedings.  Patel appears to be the first decision to apply the PSLRA’s automatic stay after the close of fact discovery to expert discovery, and also the first in the Second Circuit to expressly apply the stay to class certification proceedings.  Given that the Second Circuit has consistently ranked among the nation’s busiest forums for securities litigation, including as the top forum in 2025,[3] Patel is likely to carry significant weight in future disputes over the scope of the stay.

The decision reinforces the breadth of the stay and confirms that the broad statutory language means what it says:  “all discovery and other proceedings” means all and “any motion to dismiss” means any.  The decision also confirms that unless plaintiffs can meet their burden of proving that one of the two, narrow statutory exceptions applies—i.e., that particularized discovery is needed either to (i) preserve evidence or (ii) prevent undue prejudice—the automatic stay will not be lifted.

Overview

The Patel securities litigation relates to a global recall of respiratory devices, including CPAPs, by Philips RS North America LLC, one of the U.S. subsidiaries of the Dutch company, Koninklijke Philips N.V.  As often occurs following a stock-price decline, plaintiffs promptly filed a securities fraud action against Philips and various of its executives.  The Philips defendants moved to dismiss the complaint, and in September 2024, the court granted in part and denied in part their motion.  Over the course of 2025, the case moved through fact discovery and class certification discovery and briefing.  Near the end of 2025, the parties completed class certification briefing, and fact discovery concluded.

As the case was heading toward expert discovery on the merits, plaintiffs sought and obtained leave to file an amended complaint.  The new pleading added new alleged misstatements, new alleged corrective disclosures, a new theory of scienter, and a new individual defendant (Philips’ current CEO).  The Philips defendants again moved to dismiss.  While Philips’ CEO sought dismissal of all claims asserted against him, the other Philips defendants sought partial dismissal in light of the prior motion to dismiss ruling.

With the filing of their motion to dismiss, the Philips defendants notified plaintiffs and the court that the PSLRA’s automatic stay had been triggered.  As a result, the stay paused any further expert discovery and the additional class certification discovery and briefing made necessary by the amended complaint.  Plaintiffs strenuously objected, arguing that because certain of their claims had already survived dismissal and the Philips defendants’ motion to dismiss only sought partial dismissal, the PSLRA’s protective function against potentially meritless claims had already been served.  Plaintiffs also argued that there was no authority that the PSLRA stay applied after fact discovery had closed and encompassed expert discovery or class certification proceedings.

Following briefing and argument, on May 15, 2026, Magistrate Judge Henry rejected plaintiffs’ arguments, ruling that the Philips defendants’ motion triggered the stay and that the stay encompassed both expert discovery and class certification proceedings.[4]  Plaintiffs appealed, and District Judge Korman affirmed Judge Henry’s ruling on July 15, 2026.[5]

Implications

Key Takeaway 1:  The PSLRA stay applies to successive motions to dismiss, even partial motions to dismiss, and even when some claims previously survived dismissal.

It is not uncommon for courts, in ruling on motions to dismiss, to decide that claims based on certain alleged misstatements were adequately pled while dismissing others with leave to replead.  When that happens, as a practical matter, defendants can typically only move to dismiss portions of the amended complaint.  The same is true late in a case, when scheduling orders let plaintiffs amend to match their pleading to the evidence—any dismissal motion at that stage can likely be only a partial one.

In either of these situations, defendants should invoke the automatic PSLRA stay, citing Patel.  Plaintiffs will likely urge the court, as the Patel plaintiffs did, to look beyond the statute’s text and focus on its purpose.  The Patelplaintiffs argued that once at least some claims had survived dismissal, a successive motion challenging only a subset of claims no longer threatened the viability of the case as a whole—and therefore no longer justified the PSLRA stay.  The court rejected that argument outright.  As Judge Korman explained, “the PSLRA states that the stay of discovery applies during ‘the pendency of any motion to dismiss,’ without listing an exception either for successive motions or for when some claims have already been deemed viable.”[6]  It does not matter that claims based on some alleged misstatements will be proceeding irrespective of the outcome of the motion; a motion to partially dismiss is still a motion to dismiss.  The only way around the stay is to satisfy one of the two statutory exceptions.

Other courts, both within and outside the Second Circuit, have confirmed the applicability of the automatic stay whenever any type of motion to dismiss is pending.  For example, a California district court held in In re Lantronix, Inc. Securities Litigation that the stay applied even though defendant’s initial motion to dismiss only sought partial dismissal, holding that “[e]ven with [defendant’s] concession on a portion of [plaintiff’s] claim, the pendency of the Motions to Dismiss . . . still trigger[s] the PSLRA’s discovery stay provisions.”[7]  While that case involved an initial motion to dismiss, other courts later extended the principle to successive motions to dismiss, most often where the successive motion sought full (not partial) dismissal of the amended complaint.[8]  Another court, in Sedona v.Ladenburg, went even further and applied the stay to a successive motion seeking partial dismissal, but it did so before much, if any, fact discovery had occurred.[9]

Patel takes the PSLRA’s text to its logical endpoint.  In Patel, the court enforced the stay during a successive motion seeking partial dismissal after fact discovery had closed and class certification briefing (on the prior complaint) had been completed.  In doing so, the court confirmed that application of the stay turns on the existence of a pending motion to dismiss, not the stage of discovery or whether the motion seeks only partial dismissal.  Patel thus provides the clearest articulation of the PSLRA’s mandatory stay, free from any of the procedural wrinkles that are ultimately irrelevant to the interpretation of the clear statutory text.

Key Takeaway 2:  The PSLRA stay applies to expert discovery, not just fact discovery.

As discussed above, toward the end of, or even after, fact discovery, plaintiffs may seek to amend their complaint based on evidence developed during discovery.  If defendants move to dismiss the amended complaint, they should then invoke the PSLRA stay to pause any remaining discovery, including expert discovery.  Patel confirms, for the first time, that the PSLRA stay applies just as much to expert discovery as fact discovery.

Again, the holding turned on the statute’s plain language.  Plaintiffs tried to carve expert discovery out of the stay by arguing that no court had expressly applied the stay to expert discovery.  The Patel court rejected that argument, explaining that the statutory language, which encompasses “all discovery,” is “very clear” and does not make “a distinction [] between the types of discovery that [are] stayed.”[10]  In affirming, Judge Korman underscored this point, criticizing plaintiffs for “invent[ing] a distinction not reflected in the text of the PSLRA or the case law.”[11]   The court’s reading also makes sense as a practical matter because otherwise, parties would have to undertake potentially needless expert discovery on issues, claims, and alleged misstatements that are pared away by the motion to dismiss.

Key Takeaway 3:  The PSLRA stay reaches class certification proceedings as well.

Motions to dismiss and class certification proceedings often run on parallel tracks.  Early in a case, plaintiffs may move for class certification while also seeking to amend their complaint.  Later in a case, as in Patel, plaintiffs may seek leave to amend while their class certification motion is already pending.

Patel is the first decision from a court in the Second Circuit to expressly apply the PSLRA stay to class certification proceedings.  Plaintiffs argued that the PSLRA’s reference to “other proceedings” was limited to “litigation activity relating to discovery,” and therefore did not cover class certification proceedings.  But the court agreed with the Philips defendants that the substantial changes introduced by the amended complaint called for supplemental class certification discovery, including additional expert reports and expert depositions, and such activity fell squarely within the statute’s stay of “all discovery.”  As a result, the court did not need to separately address the Philips defendants’ alternative argument that class certification proceedings fall squarely within the statute’s reference to “all . . . other proceedings,” as other courts outside the Second Circuit had ruled.[12]

The court’s decision is all the more logical in light of the recent price impact decisions by the Supreme Court and the Second Circuit in the Goldman Sachs litigation.[13]  Those decisions made clear that in evaluating class certification motions, courts must compare, on a statement-by-statement basis, the genericness of an alleged misstatement against the specificity of an alleged corrective disclosure to determine whether the fraud-on-the-market (Basic) presumption applies to each particular statement.  Courts also are required to shorten class periods where the alleged corrective disclosures “do not reveal additional falsity of actionable misstatements.”[14]  As a result, prior to briefing class certification, defendants need to know whether any of the alleged misstatements or corrective disclosures are being dismissed from the litigation.

The plaintiffs’ bar made similar observations in a recent Law360 article[15] discussing the PSLRA stay.  The authors explained that, although the PSLRA does not expressly state that it applies to class certification proceedings, it effectively does so because class certification requires discovery.  “Thus, in staying discovery, the PSLRA effectively prohibits class certification until after any motion to dismiss has been decided, regardless of how ‘other proceedings’ is construed.”  Defendants routinely rely on fact and expert discovery to rebut the Basic presumption, and courts must remain “open to all probative evidence” bearing on price impact.[16]  Until a motion to dismiss settles the operative alleged misstatements and corrective disclosures, class certification proceedings are likely to be inefficient, if not mooted entirely by subsequent developments in connection with the dismissal motion.  Patel recognizes that practical reality, and in doing so, it aligns the Second Circuit with courts in other jurisdictions that have reached the same result.

Conclusion

Amended complaints are a recurring feature of securities litigation.  While the changes may not always be as extensive as those in Patel, defendants confronted with an amended complaint should be aware of the procedural protections that accompany successive motions to dismiss, even if only in part.

The aforementioned Law360 article highlighted Leone v. ASP Isotopes Inc.,[17] a securities class action in which the court ordered the parties to engage in simultaneous motion to dismiss and class certification briefing.  It does not appear the court in that case was reminded of the mandatory, automatic stay provided by the PSLRA, leaving defendants to engage in expert discovery and class certification briefing before the viability of the claims had been fully tested.  Patel confirms that Leone is an outlier and that, at least in the nation’s busiest securities litigation courts, the stay extends to expert discovery and class certification proceedings, and also applies to partial motions to dismiss.  Defendants should therefore be mindful of these protections and raise them promptly.

That lesson rings true even when the litigation has reached a relatively advanced stage.  Although successive motions to dismiss may not always dispose of the entire case, their resolution may substantially reshape the litigation. Claims and theories may be narrowed, defendants may be dismissed, and alleged misstatements and corrective disclosures may also be dismissed.  Those changes can materially affect what work is ultimately required, including expert work and class certification work.  Patel recognizes that reality and gives effect to one of the PSLRA’s core objectives:  preventing parties from incurring unnecessary litigation costs before the contours of the case have settled.

Going forward, Patel leaves little room for securities litigants to narrow the PSLRA stay through procedural distinctions or policy-based arguments.  Its message is straightforward:  “any motion to dismiss” means any, and “all discovery and other proceedings” means all.

ENDNOTES

[1] 15 U.S.C. § 78u-4(b)(3)(B).

[2] Patel v. Koninklijke Philips N.V. et al., No. 1:21-cv-4606 (E.D.N.Y.).

[3] Cornerstone Research, Securities Class Action Filings:  2025 Year in Review, at 22 (2026).

[4] May 15, 2026 Order; see also Dkt. No. 128 (transcript of argument).

[5] July 15, 2026 Order.

[6] July 15, 2026 Order.  In other contexts, courts have stated that the PSLRA stay provision is “crystal clear,” and “there is nothing ambiguous about the meaning of ‘any’ in the stay provision.”  Altimeo Asset Mgmt. v. Qihoo 360 Tech., 2022 WL 1663560, at *1 (S.D.N.Y. 2022); In re Smith Barney Transfer Agent Lit., 2012 WL 1438241, at *2 (S.D.N.Y. 2012).

[7] 2003 WL 22462393, at *2 (C.D. Cal. 2003).

[8] See Moab Partners v. Macquarie Infrastructure Corp., 2025 WL 2080507 (S.D.N.Y. 2025); In re Smith Barney, 2012 WL 1438241 (S.D.N.Y. 2012); Lian v. Tuya Inc., 2024 WL 1932623 (S.D.N.Y. 2024); Fosbre v. Las Vegas Sands, 2012 WL 5879783 (D. Nev. 2012).

[9] 2005 WL 2647945 (S.D.N.Y. 2005).

[10] Dkt. No. 128 at 21-22.

[11] July 15, 2026 Order.

[12] See In re ValuJet, Inc., 984 F. Supp. 1472, 1481-82 (N.D. Ga. 1997); Vignola v. FAT Brands, Inc., 2019 WL 13038337, at *2 (C.D. Cal. 2019); Spears v. Metro. Life Ins., 2007 WL 1468697, at *5 n.2 (N.D. Ind. 2007); Winn v. Symons Int’l Grp., Inc., 2001 WL 278113, at *2 (S.D. Ind. 2001).

[13] Goldman Sachs v. Ark. Tchr. Ret. Sys., 594 U.S. 113 (2021); Ark. Tchr. Ret. Sys. v. Goldman Sachs, 77 F.4th 74, 81 (2d Cir. 2023).

[14]  E.g., In re Veon Sec. Litig., 2025 WL 66444, at *6 (S.D.N.Y. 2025).

[15] Jesse Jensen and Alexandra Forgione, NY Securities Class Action Ruling Holds Rare Timing Insights, Law360 (January 28, 2026).

[16] Goldman Sachs, 594 U.S. at 122.

[17] 2025 WL 3484821 (S.D.N.Y. 2025).

The authors wish to thank Ju Hee Ahn and Bridget E. Le Donne for their significant contributions to this article.

In a new paper, we analyze how corporate governance practices can undermine weak law enforcement policies. A good example is policies that provide leniency to corporations in exchange for cooperation and implementing compliance programs or commitments to report misconduct.

Standard approaches to corporate governance focus on ensuring that managers act in the interests of shareholders. In the ideal world, law enforcement would prevent the pursuit of profit from harming customers, employees, taxpayers, or the public. However, when laws are inadequate or enforcement is weak, internal governance mechanisms may push managers to engage in profitable misconduct.

Policies that determine how external rules are interpreted and enforced may reduce deterrence.  These policies can, perversely, create incentives for managers to engage in the higher levels of misconduct that maximize profits while taking into account the expected penalties. In our paper, we show that adjustments in managerial compensation often weaken and may entirely undo the potential deterrence of penalties. Insurance and indemnification can also undo the impact of enforcement actions against corporate leaders by shielding them from consequences.

Because detecting and investigating corporate misconduct is difficult and costly, enforcers often rely on corporations to establish compliance programs, conduct internal investigations, report misconduct, and provide information. The rewards for cooperation can include lower fines, reductions in other penalties, and decisions not to prosecute. In May 2025, for example, the U.S. Department of Justice revised its corporate enforcement policy so that corporations that voluntarily disclose, cooperate, remediate, and satisfy other criteria can receive a declination of prosecution. Speaking about the changes, the head of DOJ’s Criminal Division said that “never before have the benefits of self-reporting and cooperating been so clear.” More recently, the U.S. Attorney’s Office for the Southern District of New York announced a new Corporate Enforcement and Voluntary Self-Disclosure Program for Financial Crimes. The program applies even where misconduct has already been detected and reported in the media. In exchange for self-reporting and cooperation, there will be no financial penalties, no corporate monitor, and no prosecution of the corporation.

Prosecutors like leniency policies because they encourage corporations to reveal misconduct that may have gone undetected for many years. A well-functioning compliance program can increase the probability that misconduct will be detected earlier and shorten its duration. Self-reporting can reveal misconduct that may have otherwise remained hidden. If cooperation reduces enforcement costs and helps detection of misconduct, it can be beneficial.

However, corporations governed in the interest of shareholders will adopt a compliance program, self-report, or otherwise cooperate with authorities only if doing so benefits shareholders. Authorities must therefore offer benefits that make cooperation more profitable than non-cooperation. The prospect of lower penalties that makes cooperation attractive in those cases can also make the underlying misconduct more profitable and thus more attractive to corporations and their shareholders to pursue before it is revealed. We show that, even if cooperation may reduce the likely duration of misconduct, it can be profitable for corporations to engage in higher levels of misconduct before it is detected. The result may be more harm to society from misconduct.

By giving significant or even complete discounts on fines and other penalties for voluntary compliance or self-reporting, authorities give corporations the choice of how to respond (e.g., when to self-report). The corporation will choose the most profitable approach and may do more harm even as they seem to be complying. Mandatory monitoring, whistleblower programs, and properly audited reporting obligations, by contrast, can increase detection without making misconduct more attractive. Leniency will not be needed and will thus not make fines simply a cost of doing business.

The case for offering leniency in exchange for compliance programs is even weaker when such programs are ineffective or largely cosmetic, which is more likely when the misconduct is profitable. In our analysis, we assume that compliance programs genuinely increase the probability of detection. If corporations receive credit and leniency for programs that just create the appearance of compliance, then the leniency will not improve detection but instead make the misconduct more profitable and frequent.

Our analysis suggests that cooperation may not be the best approach to enforcement in the corporate context. Cooperation policies should be evaluated by asking whether they increase detection and reduce the underlying harm, not merely whether they reduce enforcement costs or induce corporations to help authorities.

Overall, if misconduct and law breaking is profitable, internal governance mechanisms that align managers with shareholders, which may be valuable when laws are well designed and effectively enforced, can exacerbate the conflict between corporations and society and the potential harm corporations may cause when enforcement is weak. Our analysis underscores the importance of a broader perspective that might be called societal corporate governance, one that ensures that the interactions between external and internal governance are constructive and lead corporations and the individuals acting on their behalf to produce genuine benefits without causing undue harm.

Anat R. Admati is the Joseph McDonald Professor of Finance and Economics at Stanford Graduate School of Business, Nathan Atkinson is an assistant professor at the University of Wisconsin Law School, and Paul Pfleiderer is the C.O.G. Miller Distinguished Professor of Finance, Emeritus at Stanford Graduate School of Business. This post is based on their recent paper, “Profitable Misconduct, Corporate Governance, and Law Enforcement,” available here.