In the following guest post, Stephen J. Choi, Jessica M. Erikson, and Adam C. Pritchard take a look at the plaintiffs’ attorney fee awards in “mega-settlements” in securities class action lawsuits. The authors ask the question whether the lawyers who lead these cases and negotiate the settlements are appropriately rewarded for their efforts. Choi is the Murray and Kathleen Bring Professor of Law at New York University School of Law. Erickson is Professor of Law & Associate Dean for Faculty Development at University of Richmond School of Law. Pritchard is the Frances and George Skestos Professor of Law at University of Michigan Law School. My thanks to the authors for allowing me to publish their article as a guest post on this site. I welcome guest post submissions from responsible authors on topics of interest to this blog’s readers. Please contact me directly if you would like to submit a guest post. Here is the authors’ article.


In our paper, Working Hard or Making Work? Plaintiffs’ Attorneys Fees in Securities Fraud Class Actions, we examine “mega-settlements” in securities fraud class actions. The cases that settle for hundreds of millions of dollars look like the success stories of securities fraud litigation. And indeed, lawyers fight fiercely for the right to represent the class in the cases that lead to these enormous recoveries. The substantial recoveries in these cases seems to have discouraged scholars from scrutinizing them. We attempt to fill this gap, focusing specifically on the fees awarded to plaintiffs’ attorneys in cases that result in mega-settlements. Are the lawyers who win the contest to represent the class in the highest stakes lawsuits being appropriately rewarded for their efforts? And do judges tailor fee awards to reflect the risks taken on by plaintiffs’ attorneys working on a contingent fee basis?

Mega-settlements stand apart as a distinct category of settlements in securities class actions. The bottom 90% of settlements—i.e., the settlements in the first nine deciles—average $11 million, with settlements in the ninth decile averaging $41.8 million. The settlements in the top decile, by contrast, average $295.5 million, more than seven times larger than the settlements in the decile below. Unsurprisingly, these settlements lead to significant fees for the plaintiffs’ attorneys—a mean of $39.5 million compared to a mean of $2.7 in the other nine deciles.

We examine how the high stakes in these mega-settlements influence plaintiffs’ attorneys in requesting attorneys’ fees and judges in awarding them. To support our analysis, we collected data on every securities class action with a disclosure claim filed in federal court between 2005 and 2016, a total of 1,719 cases. We gathered information on the contest for lead plaintiff, the allegations in the final consolidated complaint, potentially dispositive motions, and the resolution of each case. In every case that ended with a settlement, we collected data regarding the settlement terms, the fees requested by lead counsel, and the hours reported by the plaintiffs’ attorneys lodestar data. We also collected the awards made by courts.  We supplemented the litigation data with the defendant corporations’ market capitalization, which we measured on the last day of the class period.

Judges appear to give greater scrutiny to fee requests in the largest cases. The figure below shows how often judges reject or modify the fee requested by plaintiffs’ attorneys, broken down by the size of the company. (Settlement amounts positively correlate with market capitalization; suits against big companies settle for more, on average.)



We see a clear pattern of closer scrutiny as the size of the company increases, with a significant jump in the rejection rate for cases against the largest companies, which are most likely to result in mega-settlement. We conjecture that this more exacting scrutiny from courts gives attorneys an incentive to report more hours in the largest cases to bolster the argument for a fee award. Our regression analysis supports this conjecture. We find that larger fee requests are more likely to be rejected, but that a greater lodestar reduces the likelihood of rejection.

How do plaintiffs’ attorneys respond to the incentive structure that this creates for high-stakes cases? Our analysis of this question is shaped by the fact that judge almost never award plaintiffs’ attorneys more than one-third of the total settlement amount, regardless of the lodestar amount presented in the fee application. In smaller cases, this de facto 33% cap may constrain the hours that plaintiffs’ attorneys invest in the litigation. In higher-stakes litigation, however, plaintiffs’ attorneys are unlikely to run up against this cap because the expected settlement value is so large that any credible lodestar amount will still be well below a third of the settlement. As a result, plaintiffs’ attorneys can invest more hours into building the cases—for example, by researching possible claims, pouring through discovery, and filing and responding to motions—because they know that they are likely to be paid for this work if the case settles. We call this possibility the “working hard” hypothesis. At the same time, however, plaintiffs’ attorneys in higher-stakes litigation may also be more likely to inflate their hours by doing work that is duplicative or unnecessary, a possibility that we call the “making work” hypothesis.

We perform empirical tests to determine whether plaintiffs’ attorneys are working hard or making work in high-stakes securities class actions. We use the presence of multiple lead counsel as a proxy for a greater incentive for plaintiffs’ attorneys to generate duplicative or unnecessary hours. We conjecture that the presence of multiple lead counsel is driven more by relationships among attorney firms and between attorney firms and specific institutional investors rather than the merits of a specific case. (This possibility is the subject of a parallel empirical research project we are conducting). Using a multivariate ordinary least squares model, we find that having multiple lead counsel corresponds to 4,100 more attorney hours, even controlling for case characteristics and other factors. In the high-stakes cases, however, having multiple lead counsel increased the total hours reported by plaintiffs’ attorneys by 28,900 hours.

We find similar results after the U.S. Supreme Court’s decision in Halliburton II. We posit that Halliburton II’s price impact defense at the class certification stage provided a “shock,” increasing the need for more work on the part of plaintiffs’ attorneys in securities class actions, while at the same time providing cover for attorneys interested in inflating hours to do so (i.e., duplicative work on the part of multiple lead counsel). Using a difference-in-difference analysis, we show that hours significantly increased in cases with multiple lead counsel after Halliburton II, consistent with more “make work.”

We examine whether judges serve as a meaningful check on these agency costs. In determining what percentage of the settlement to award in fees, judges often consider the plaintiffs’ attorneys’ lodestar, as well as a multiplier. The conventional explanation for multipliers is that they reward plaintiffs’ attorneys in contingency fee litigation for the risk that they may expend resources in litigation but not receive any compensation. The higher the risk in a particular case, the higher the multiplier, at least in theory.

Our empirical tests, however, do not support the conclusion that courts calibrate the multiplier to the actual risk that plaintiffs’ attorneys face in litigation. Instead, we find that courts award higher multipliers in cases with pre-litigation observable characteristics that indicate a lower risk of dismissal – and a correspondingly higher probability of settlement – particularly against larger companies. These cases presumably pose less risk to plaintiffs’ attorney because of their higher ex ante probability of settlement. Our findings suggest that judges reward attorneys simply for winning the lead counsel spot in the largest cases. The figure below shows the relation between company size and the awarded multiplier.



Moreover, the results of our regression analysis show that the multiplier increases with the likelihood of settlement. To make matters worse, plaintiffs’ attorneys appear to exploit a low likelihood of dismissal by investing more hours in the litigation, especially in the largest cases for which the de facto 33% cap on fees unlikely to be binding. The figure below shows how cases that reach the class certification stage have larger reported hours, which is hardly surprising, but that the largest bump in fees is for cases against the largest companies.



Among the Large Market Cap actions, the stage that the litigation reaches before settlement corresponds to plaintiffs’ attorney hours. For Large Market Cap actions that settle, the mean hours for actions with a Class Certification Filing is 72.8 thousand hours. That figure is nearly six times the 13.0 thousand hours for actions without a Class Certification Filing. This overall pattern is confirmed in our regression analysis. This result is consistent with plaintiffs’ attorneys who are not limited by the de facto 33% plaintiffs’ attorneys fee cap investing more time once the risk of non-settlement diminishes. Judges reviewing fee requests appear to be poorly attuned to the amount of effort needed to prosecute such cases or the actual risk faced by the plaintiffs’ attorneys.

Overall, our findings indicate that plaintiffs’ attorneys may be receiving windfall fee awards in mega-settlement cases at shareholders’ expense. We note, however, that our research has limitations. We cannot directly observe whether firms are inflating their hours by doing unnecessary work. Without this direct evidence, we have to rely on proxies, such as the presence of multiple lead counsel or increased hours at certain stages of the litigation. We also recognize that the larger-stakes cases may differ from other securities class actions in ways that our data does not capture.

The complete paper is available for download here.