Him Too? The Corporate Consequences of Epstein Connections

Between September 2025 and January 2026, the US Department of Justice released more than three million pages relating to Jeffrey Epstein, a convicted sex offender whose network extended across business, academia and politics. The central harm was borne by Epstein’s victims. But the files also provide something governance researchers rarely observe: a detailed, dated record of how a compromised but well-connected individual interacted with public-company leaders over more than a decade. Our findings suggest that the principal corporate risk lay not in the access Epstein’s network provided, but in the norms transmitted through it.

In our recent paper, ‘Him Too? Analyzing the Effects of Epstein Connections’, we use these records to examine the corporate consequences of such ties. Directors are recruited partly for their networks, and well-connected boards are often viewed as a valuable asset. Epstein’s network challenges that assumption: did his ties primarily expand firms’ access to influential people, or did they expose firms to harmful norms?

What the files show

We searched for more than 92,000 CEOs and directors of US public companies across 1.3 million text-bearing documents. A large language model helped distinguish documented contact—correspondence or meetings—from incidental mentions. The analysis centres on 841 directors with frequent documented contact.

Three findings emerge. First, investors treated these associations as material. S&P 500 firms linked to Epstein in news coverage following the 30 January 2026 release lost about 3.7% of their value over three trading days—roughly $1.4 billion for the median firm. The firms themselves were generally not accused of wrongdoing; the revealed associations through executives and directors were enough to move prices. Yet press coverage captured only the surface: the documents identify more than four times as many connected S&P 500 firms as press reports did.

Second, Epstein was a genuine connector within corporate America, rather than a collector of isolated acquaintances. US companies are already linked through directors who sit on multiple boards. Adding Epstein’s documented ties makes this network nearly eight times as dense and brings the average pair of large firms within two steps of each other. The effects are strongest in finance and technology. These ties increase network density more than shared alumni affiliations with Yale or Princeton.

Third, proximity is associated with corporate misconduct. Firms with more connected directors experience significantly more adverse incidents reported by the media, regulators and NGOs, particularly governance failures and employee mistreatment. Within the same firm over time, one additional connected director is associated with about 1.6 additional incidents per year—slightly more than half the annual count for the average firm. The pattern also appears in independent legal data, where these firms are associated with more financial-regulatory enforcement, more securities and racketeering litigation, and larger employment-misconduct settlements.

Could this simply be because poorly governed firms were more likely to appoint Epstein-connected directors? The evidence suggests that this is unlikely to be the full story. The estimates are based on changes within the same firm over time and remain similar when firms are compared only with peers in the same industry and year. Incidents also decline after a connected director dies and leaves the board—an exit unlikely to have been driven by the firm’s governance.

Access or contagion?

The files allow us to distinguish two features of elite networks that usually move together. The first is network expansion: the improvement in a firm’s position when Epstein’s ties are added. The second is norm contagion: exposure to Epstein-connected firms through directors who sit on multiple boards.

When both are considered together, a firm’s improved network position is not associated with misconduct. Exposure to connected peer boards through shared directors is—even when the firm has no direct Epstein tie of its own.

Linguistic measures of the directors’ emails point in the same direction. Directors who gain network access write more formally over time, while those who become more exposed to connected peers write less formally. This pattern is consistent with closer and more permissive interactions. The evidence therefore points less to the access the network created than to the norms it transmitted.

Implications

Boards typically assess a director’s network by its size, centrality and access to influential people. Our findings suggest that this assessment is incomplete. The same ties that convey information and opportunity may also shape judgments about acceptable conduct and embed firms in social environments carrying governance, legal and reputational risks.

For boards, investors and regulators, the practical lesson is that a director’s network should be evaluated not only by whom it provides access to, but also by the norms circulating through it.

The authors’ paper is available here. 

Marina Gertsberg is an Assistant Professor (Senior Lecturer) in Finance at the University of Melbourne and a Research Affiliate at IZA@LISER.

Michaela Pagel is an Associate Professor, Washington University Olin Business School, an NBER Research Associate, CEPR Research Fellow, and Research Network Affiliate, Center for Economic Studies (CES-ifo).

Ekaterina Volkova is an Associate Professor in the Department of Finance at the Faculty of Business and Economics, University of Melbourne.