On September 23, U.S. District Judge William Pauley refused to approve a settlement between the CFTC and Deutsche Bank covering the bank’s failure to report swaps properly. Instead of rubber-stamping the settlement, the judge asked the parties for additional briefing. With that move, Judge Pauley followed in the footsteps of two of his colleagues in the Southern District of New York, judges Jed Rakoff and Victor Marrero, and several other district court judges around the country.

As I describe in more detail in an essay recently published in the Yale Law Journal Forum, “Securities Settlements in the Shadows,” and available here, judges Pauley, Rakoff, Marrero and their peers will have fewer similar opportunities in the future. After Dodd-Frank, securities settlements are increasingly filed in administrative proceedings, where they avoid judicial review.

Section 929P(a) of the Dodd-Frank Act authorizes the Securities and Exchange Commission to obtain a civil fine for a securities violation against “any person” in an administrative proceeding. Before Dodd-Frank, the SEC could secure civil fines against registered broker-dealers and investment advisers in administrative cease-and-desist proceedings, but had to sue in U.S. district court non-registered firms and individuals, including public companies and executives charged with accounting fraud, or traders charged with insider trading. After Dodd-Frank, except for a few remedies that can only be obtained in court, the SEC can choose the forum in which it prosecutes enforcement actions.

Since Dodd-Frank, the SEC has litigated more contested cases before administrative law judges (“ALJs”)—a move that has received considerable attention. At the same time (and without much notice), the SEC also shifted its settlement filings from district court to administrative proceedings, just as it shifted litigation in contested actions to ALJs. Until Dodd-Frank, the SEC filed most of its settlements in district court: Only one-third to one-half of settlements were filed in administrative proceedings. In 2013, the practice shifted sharply, and, by fiscal year 2015, the SEC filed five times as many settlements in administrative proceedings as it did in court (click here for figures and original data on SEC settlement practices).

The shift is significant because most SEC cases settle. When a settled enforcement action is filed in court, a federal judge reviews the settlement to ensure that it is fair, reasonable, and in the public interest (some circuits, including the D.C. Circuit, review for adequacy and appropriateness as well). Judicial review is usually a rubber stamp, but not always. In 2009 and 2011, Judge Rakoff refused to approve two settlements between the SEC and large financial institutions, the former for failing to penalize individual defendants and the latter for failing to include any proven or admitted facts. Although both settlement denials were controversial and one was ultimately reversed by the Second Circuit, the SEC vowed to change the enforcement practices that Judge Rakoff had rejected.

By contrast, settled actions filed in administrative proceedings receive no formal external scrutiny. Settlement negotiations remain confidential until after the Commission approves the settlement in a closed hearing.

Unlike the contested cases where defendants have filed dozens of lawsuits challenging the constitutionality of ALJs, defendants who settle and the SEC both prefer to file settlements in administrative proceedings. In fact, since October 2013, all SEC settlements with large Wall Street banks—the sorts of settlements that Judge Rakoff rejected—have been filed in administrative proceedings. By doing so, the SEC and defendants avoid the publicity, the uncertainty, and the risk of delay caused by judicial review.

A wholesale shift to in-house settlements may not be desirable, given the near-complete absence of transparency and oversight, at least until after the settlement has been finalized. Operating in the shadows, SEC settlement practices can evolve quickly and stealthily as a result of internal and external pressures, including budgetary constraints, regulatory capture, and short-term political interests. This is not an idle concern: There is some evidence that SEC settlement practices have drifted from announced priorities. Contrary to Chair Mary Jo White’s stated initiative to prioritize securities fraud cases, recent SEC cases are less likely to prosecute scienter-based violations. Contrary to an expressed commitment to prosecute more individuals, the SEC has recently targeted fewer. Despite a stated goal of factual admissions, few settlements include them, even those that accompany guilty pleas or convictions in parallel criminal proceedings.

I do not assert that only courts can police SEC settlements or that courts should be in the business of setting enforcement priorities. Agencies are generally better informed about both. But agencies and their staff are also subject to significant pressures than can cause enforcement practices to drift from what is optimal to what is more expedient. Judicial oversight of SEC settlements operated as an external constraint on drift, and this constraint is now gone. Whatever one thinks of Judge Rakoff’s interventions, he voiced the concerns of many that financial enforcement agencies were not doing a good enough job.

In the essay, I propose two approaches that could be implemented immediately without legal change and at low cost: Congressional oversight committees could exercise more probing review of enforcement policy and practices, or the SEC leadership could begin reporting on how it is meeting the goals that it sets for itself, including targeting individuals, seeking factual admissions and prosecuting fraud. At the least, the SEC should be as explicit about meeting the goals it sets for itself as it is about announcing them.

This post comes to us from Professor Urska Velikonja of Emory University School of Law. It is based on her recent paper,“Securities Settlements in the Shadows,” available here.