Information escrows.

• Author: Ayres, Ian
• Position: Introduction into II. Using Allegation Escrows to Mitigate Initial Claim Aversion B. Additional Costs and Benefits of an Allegation Escrow Regime, p. 145-172

A variety of information escrows--including allegation escrows, suspicion escrows, and shared-interest escrows--hold the promise of reducing the first-mover disadvantage that can deter people with socially valuable private information from disclosing that information to others. Information escrows allow people to transmit sensitive information to a trusted intermediary, an escrow agent, who only forwards the information under prespecified conditions. For example, an allegation escrow for sexual harassment might allow a victim to place a private complaint into escrow with instructions that the complaint be lodged with the proper authorities only if the escrow agent receives at least one additional allegation against the same individual. We assess the benefits and costs of allegation escrows and discuss how they might be applied to a variety of claims, including sexual harassment, date rape, adultery, and corporate and public whistleblowing. We also show how analogous shared-interest escrows might be used in workplace dating and adoption contexts to facilitate the discovery, of parties' mutual interest when unintermediated expressions of interest might themselves be harassing.

TABLE OF CONTENTS INTRODUCTION I. INFORMATION ESCROWS IN ACTION A. Commitment and Posthumous Escrows B. Shared-Interest Escrows II. USING ALLEGATION ESCROWS TO MITIGATE INITIAL CLAIM AVERSION A. Modeling the Impact of Escrows on the Communication Equilibrium B. Additional Costs and Benefits of an Allegation Escrow Regime 1. Legal Issues 2. Managing Salience 3. The Matching Algorithm III. APPLICATIONS A. Applications for Allegation Escrows 1. Sexual Harassment, Date Rape, and Other Sexual Assaults 2. Whistle-blowing and Allegations of Nonsexual Wrongdoing 3. Suspicion Escrows 4. Insecurity Escrows CONCLUSION INTRODUCTION

A familiar narrative of sexual harassment begins with a reluctant initial allegation of abuse that is quickly followed by other accusers stepping forward with similar allegations. A victim of abuse is reluctant to bring the first claim, in part because the accused routinely responds by trying to impeach the credibility of the accuser, characterizing the accuser as "a nut or a slut." (1) This initial claim aversion is a rational concern. In a "he said / she said" credibility contest, an uncorroborated accusation of harassment is unlikely to prevail. Though initial accusations often inspire additional allegations from other victims that can serve to corroborate the initial claim, isolated claimants deciding whether to make the first accusation often cannot be sure whether other victims exist and whether those other victims will have the courage to make a supporting allegation. Victim reluctance to take on the risk of "going it alone" gives rise to the well-known concern that there might be a substantial underreporting of harassment. (2) Even recidivist harassers may go unchallenged because, among isolated victims, there can be a first-mover disadvantage to making the initial accusation. A challenge for public policy is to seek out ways to encourage victims of sexual harassment to take on the risks associated with making initial allegations. (3)

Of course, not all accusations are true. The standard narrative can also be read as suggesting that there may be too many accusations. Once one or two accusations of harassment are lodged, it may become too easy for putative victims to make false, copycat accusations based on unsubstantiated accounts of alleged, long-ago events. While it may be comforting to infer guilt from the multiplicity of accusations, policymakers should also be concerned about whether cascades of potential copycat complaints substantially enhance the likelihood of an ill-supported guilty finding.

In this Article, we redeploy the game-theoretic "information escrow" technique to make progress on the twin concerns of underreporting of initial truthful allegations and overreporting of false copycat allegations. We propose the use of an allegation escrow to allow victims to transmit claims information to a trusted intermediary, a centralized escrow agent, who forwards the information to proper authorities if (and only if) certain prespecified conditions are met. Specifically, the escrow agent would keep harassment allegations confidential, unutilized, and unforwarded until the agent has received a prespecified number of complementary harassment allegations concerning the same accused harasser. For example, if the escrow agreement specified the accumulation of two additional allegations as a triggering event, then the agent would wait until the escrow had received three separate allegations concerning a particular alleged harasser before forwarding the information to specified authorities and initiating a complaint on behalf of the three alleging parties.

An allegation escrow holds the promise of mitigating the first-mover disadvantage in making a complaint. A victim can place the first allegation into escrow with diminished fear that she will bear the sole brunt of the adversarial reaction, and with confidence that her escrowed allegation will be released only if accompanied by at least one other allegation against the same individual. (4) Information escrows might thus secure more initial allegations because the alleging victim can rest assured that her initial allegation will not be seen unless it is part of a larger pattern of alleged misconduct.

More precisely and more subtly, an allegation escrow helpfully creates uncertainty for the victim regarding whether she is making the initial allegation. At the time of placing an allegation into escrow, an alleging victim will not know whether any prior allegations against the same offender have already been placed into the escrow. This means that an accuser will not know at the time of making an escrowed accusation whether she is the first or the second (and triggering) accusation. This incomplete information also helps respond to the copycat concern. Absent collusion among the allegers, (5) investigating authorities need not be worried that later-in-time allegations placed in escrow are false copycats of an initial escrowed allegation, for the simple reason that the subsequent allegers would not know that an earlier allegation had been made.

Indeed, the possibility that different victims would independently--behind the veil of the escrow--allege similar details of harassment could enhance the credibility of each allegation. If multiple students independently escrowed allegations that a particular professor's harassment included inappropriate bugging, the very similarity of the professor's modus operandi would be strong evidence that the allegations were true. In contrast, if it is widely known that a student accused a professor of sexually harassing her by inappropriately hugging her, then investigating authorities would need to consider whether subsequent complaints of inappropriate hugging by the same professor were triggered by reports of the first complaint. Besides the possibility of false follow-on allegations, there is also the possibility that subsequent allegers were inappropriately primed by their knowledge of the first accusation into reinterpreting the behavior of the accused as harassing. (6)

But allegation escrows are not a panacea. While escrows hold the potential for mitigating the twin concerns of initial underreporting of truthful allegations and subsequent overreporting of false allegations, this Article will also discuss a variety of ways in which placing an intermediating escrow mechanism between allegers and investigating authorities might be counterproductive. Most importantly, we will consider circumstances where allegation escrows may reduce the sheer quantity of actionable complaints. In a world with escrows, some victims' complaints will never exit the escrow mechanism. Therefore, we envision a system in which victims retain the right to go it alone by directly lodging a complaint even if they have previously placed an allegation in escrow. Indeed, we will discuss ways in which escrows can be designed to encourage and facilitate subsequent conversions of escrowed allegations into independent go-it-alone complaints. But there remains the possibility that an escrow system would leave some harassment uninvestigated and some harassers undeterred. Even if, on the whole, escrows would increase the quantity and quality of deterrence, some unpaired escrow allegations could remain forever impounded. An important goal of the subsequent analysis is to determine the circumstances under which the benefits of escrows outweigh their costs.

Allegation escrows come in many shapes and sizes. In the following Sections, we will discuss design issues (including more detail on triggering events, interim reporting, and matching criteria) and legal issues (including the legal relationship between the escrow agent and the escrow depositors, and what duties, if any, are owed to the accused). We will also show that information escrows might be applied to many different types of information. At least conceptually, allegation escrows can be applied to almost any context in which victims experience claim aversion because of a reluctance to go it alone. (7) As we'll see, allegation escrows might be used in the workplace not only to respond to instances of sexual and racial harassment, but also as a complementary tool to protect whistle-blowers in making allegations concerning corporate or government misconduct. Outside of the workplace, escrows might be put in place to respond to allegations of date rape, an area in which claim aversion similarly leads to well-recognized problems of underreporting. (8) Indeed, we'll see that it might even be possible to create "suspicion escrows," where mere suspicions of adultery or other misconduct could, if matched, be disclosed.

Ultimately, we will conclude that the case for deploying information escrows is stronger in...