Excerpted From: Jens Frankenreiter and Michael A. Livermore, Are Lawyers' Case Selection Decisions Biased? A Field Experiment on Access to Justice, 52 Journal of Legal Studies 273 (June, 2023) (23 Footnotes/References) (Full Document)


FrankenreiterLivermoreIn the United States, access to justice often requires access to lawyers. Although, in principle, litigants can represent themselves in court, as a practical matter pro se litigants face a variety of challenges that make it difficult to vindicate their rights. In many criminal proceedings, the US Supreme Court has found that access to counsel is so basic that the Constitution requires that representation be provided to indigent defendants. The existence of a robust market for legal services also provides evidence of the belief among consumers that competent legal counsel is valuable. Where it has been studied, legal representation is often associated with better outcomes (Eagly and Shafer 2015; Greiner, Pattanayak, and Hennessy 2013), although this effect is not universal (Greiner and Pattanayak 2012).

The role and importance of counsel vary by litigation context, but access to lawyers may be particularly consequential in legal fields in which fees are paid on a contingent basis. For these matters, lawyers provide not only counsel and advocacy but also financing for litigation costs. For many civil cases, whether a plaintiff is able to pursue litigation at all often depends on an attorney's decision about taking the case. This paper presents results from a large-scale field experiment exploring how demographic information (encoded in potential clients' names) affects elements of attorneys' intake decisions in private injury claims. In line with studies in other fields such as employment and housing, we find evidence suggesting that attorneys may disfavor members of racial or ethnic minorities in this context.

A variety of factors may influence attorneys' intake decisions. A utility-maximizing attorney will seek out cases that provide the largest expected benefit at the lowest expected cost. Typically, “benefit” can be understood in terms of monetary remuneration--the likelihood of a settlement or judgment times the anticipated amount, discounted by the contingency fee percentage. There may also be reputational benefits associated with taking on certain high-profile cases and nonpecuniary benefits akin to the psychological warm glow that accompanies charitable giving (Andreoni 1990). Costs are often similarly pecuniary but may also have reputational or psychological dimensions.

The decision whether to take a case is sequential in nature, as prospective attorneys and clients gather and reveal information during their preliminary interactions. There are important information asymmetries on both sides, as clients have access to privately held information about the case, and attorneys have experience in pursuing similar matters. Once a case is taken, an attorney-client relationship attaches, with all of the associated responsibilities and liabilities on the part of the attorney. Most important, the attorney-client relationship makes it difficult to abandon a client midstream in the course of litigation, even if information is revealed that indicates that a case has a lower expected payout than initially anticipated. This lock-in effect means that information collected prior to creating an attorney-client relationship has particularly high value.

Even though intake decisions can be consequential, they nevertheless must be made on the basis of quite limited information. For valuable claims, attorneys are in potential competition with each other, and delaying the decision to take a case with the goal of gathering more information may result in losing the client altogether. Protracted initial deliberations, especially with unsophisticated clients, also raise the possibility of creating an informal attorney-client relationship, with the attendant duties and liabilities. In this initial stage, attorneys must carefully balance the costs and benefits of delay. Because attorneys often have to make an intake decision on the basis of a preliminary understanding of the facts of a case, they may be influenced by guesses and hunches as much as reasoned analysis.

This study investigates the very beginning of attorneys' intake decisions--an initial request for information from a prospective client. Such a request represents the first gatekeeping opportunity for a lawyer. By failing to respond or responding negatively to such a request, an attorney can foreclose any dialogue before it begins. This initial contact also represents the lowest information point for the attorney, as a prospective client is reaching out cold, and the attorney has access to only the information that the requestee has chosen to reveal.

We are interested in how basic demographic factors revealed in a prospective client's name (primarily a sender's membership in a racial or ethnic group) might affect attorneys' choices about pursuing additional information at this initial stage. To investigate this question, we conducted a large-scale, email-based field experiment. Our work builds on a robust social and behavioral science literature that examines how various economic and political interactions are affected by race and gender, utilizing a name-based experimental design. Leveraging the same basic methodology developed in this literature (Crabtree 2018; Gaddis 2018), we sent initial information requests to more than 24,000 attorneys in jurisdictions across the United States, randomly assigning inquirer names that signal different demographic characteristics. Our research question is whether the rates of responses are sensitive to these demographic signals and what geographic and attorneys' characteristics predict response rates.

Our primary finding is that perceived race or ethnicity is a factor in attorneys' responses to initial inquiries concerning personal injury cases, with ostensibly Black or Hispanic inquirers less likely to receive a response than ostensibly White senders. The existence of this effect is validated by the split-sample approach employed. We constructed the models used to test our hypotheses on the basis of an analysis of roughly half the available data. After preregistering the resulting research plan, we used the other half of the data to confirm that our results can be replicated.

In apparent contrast to similar studies in other fields, the preferential treatment of White senders is mostly driven by higher response rates for senders with names suggestive of persons who are White and female. While this effect can be observed in both rounds of the study, it is markedly more pronounced in the first round. In particular, the results from the first round provide little indication that White male inquirers received more responses than male inquirers in other racial or ethnic categories. By contrast, in the second round, attorneys were more likely to respond to White male inquirers than to other male inquirers, although this difference is not statistically significant.

We furthermore find that the treatment effect may vary depending on the attorney's location and personal characteristics. Most important, we find some evidence that the preferential treatment of White inquirers is driven by certain subgroups of attorneys, among them attorneys with names that are more common among the White population in the United States. This result persists when we apply a matching procedure that seeks to eliminate potential confounding effects of the environments in which attorneys operate. We note that we were unable to replicate this result using the statistical test included in our preregistration statement. However, an analysis of the combined data suggests the existence of such in-group versus out-group effects, and we leave it to future research to confirm whether these results can be replicated. If supported in future studies, the finding suggests that this differential treatment cannot be explained solely as a response by attorneys to jurisdiction-level factors that affect expected payoffs.

The remainder of this paper is structured as follows. Section 2 provides more details about the theory and an overview of the existing literature. Section 3 describes the design and execution of the field experiment, Section 4 discusses the data and provides descriptive statistics. Section 5 presents the results. Section 6 discusses the implications of these results, and Section 7 concludes.

[. . .]

This paper reports results from a nationwide study of the effects of demographic information (as encoded in names) on attorneys' behavior at the initial inquiry stage of client intake for personal injury claims. The study implements an email audit of more than 24,000 attorneys from across the United States and documents whether response rates vary with inquirers' perceived racial or ethnic group. We find a discriminatory effect: members of the majority race or ethnicity are preferred. We also find tentative evidence that these effects vary depending on attorneys' personal characteristics. The latter finding might suggest that the observed differences in the treatment of different groups are not merely a reaction to anticipated differences in the expected payoffs from lawsuits involving different types of plaintiffs but that attorneys' selection decisions are at least partly a function of their personal characteristics. Mechanisms that could explain this latter effect include taste-based discrimination and ill-informed beliefs about members of other groups.

jens frankenreiter is Associate Professor of Law at Washington University in St. Louis.

michael a. livermore is Professor of Law at the University of Virginia.