5^th Circuit Ruling Raises Concerns About Confidentiality in Federal Agency ADR

By Charles Pou, Jr.^*

When Congress approved the 1996 amendments to the Administrative Dispute Resolution Act (ADRA), most observers believed the difficult issues involved in protecting confidentiality in federal alternative dispute resolution had been resolved. After all, the act=s approach to confidentiality questions appeared clear and precise B indeed, its confidentiality section is the most detailed of any federal or state ADR statute.

As the Senate report for the original act stated, the act=s confidentiality A...protections are created to enable parties to ADR proceedings to be forthcoming and candid, without fear that frank statements may be used later against them. Thus, documents produced during an ADR proceeding, such as proposals to resolve the dispute, are immune to discovery unless certain specific conditions are met.@¹

In legislating, Congress gave parties in federally-related ADR proceedings an assurance that their dispute resolution communications would generally be Aimmune from discovery,@ and defined these protections in detail. The act forbids neutrals from disclosing such communications, and also states that they shall not Abe compelled to disclose@ such communications.

The confidentiality provisions of ADRA B the basic law governing ADR activities in federal administrative settings B reflected research sponsored by the Administrative Conference of the United States, and were the product of extensive negotiations among a diverse array of entities interested in balancing open government and dispute resolution needs, including the American Bar Association and Public Citizen. Most observers have found them to reflect an appropriate balance between the openness needed for legitimacy and oversight and the confidentiality necessary for many sensitive negotiations.

The act has expanded agency use of an array of processes that encourage informal resolution of disputes. However, recent related investigations by the U.S. Department of Agriculture=s Inspector General and a federal grand jury in Texas leading to a July 1998 decision by the 5^th U.S. Circuit Court of Appeals, have given rise to new uncertainty about the act=s scope and effect.

In 1994, the Inspector General of the U.S. Department of Agriculture (USDA) began inquiring into possible fraudulent misrepresentations allegedly made by debtors in the course of mediations regarding restructuring of farm loans. These took place under a state agricultural mediation program then operated by Texas Tech University under a grant from USDA=s Farm Service Agency (FSA) under the Agricultural Credit Act of 1987 (ACA).

The inspector general moved forward on two fronts. It announced that it would conduct a financial audit of the Texas Tech Mediation Program=s expenditure of its USDA grant funds, along with similar audits of state-level mediation programs in several other states. Its reports on these programs= operations recommended that FSA withhold all future grants to state mediation programs in Texas and three other states Auntil the agencies provide USDA representatives access to mediation records, unless a formal opinion or ruling is obtained that grant contract provisions dealing with access to records is unenforceable.@

The Texas Tech program subsequently ceased operating. The inspector general has conducted similar audits of additional states= agricultural mediation programs, and criticized those that did not provide access. Moreover, it has reportedly insisted that the standard grant agreement be amended to provide even more specific authority for its access to mediation files, and has refused to negotiate changes in that language with objecting programs.

In addition to auditing these programs, the USDA inspector general=s office issued administrative subpoenas for the information previously requested. These were resisted by the Texas Attorney General. In October 1996 a federal grand jury subpoena was issued to Texas Tech in connection with an investigation of possible criminal wrongdoing by parties to mediations. After several mediation participants moved to quash the subpoena B citing mediation confidentiality provisions in ADRA, the ACA and the Texas Alternative Dispute Resolution Procedures Act B the federal district court entered an order to quash.

A 5^th Circuit panel reversed the district court in July 1998, In re Grand Jury Proceedings, 148 F. 3rd 487 (5th Cir. 1998), determining that neither the federal ADR Act nor the Texas statute applied to the case. It stated that, while ADRA allows an agency to use a dispute resolution proceeding for the resolution of an issue in controversy that relates to an administrative program, there was no such Aissue in controversy@ involved in the case at hand.²

It appears that the 5^th Circuit erred in concluding the federal ADRA to be inapplicable in the Texas Tech agricultural mediation program. Under the act=s terms, it applies if an ADR process is employed to address any Aissue in controversy@ relating to any Aadministrative program.@ Moreover, the act defines these terms quite broadly.

The Texas Tech disputes would appear raise such an Aissue in controversy@ for at least three reasons. First, FSA was a party to nearly all of the cases mediated. Indeed, the USDA inspector general=s interest in obtaining relevant communications was precisely because of the federal agency=s involvement. Second, the issues in controversy relate directly to decisions and activities by FSA and other interested persons under the Agricultural Credit Act. Finally, the mediations were all conducted with the aid of funding from a grant program operated by a federal agency.

Of course, it is conceivable that another, inconsistent federal statute might control in a particular instance. In this case, the need for confidentiality protection to promote the use of ADR may need to be reconciled with the provisions of the Inspector General Act of 1978, which provides generally that each inspector general is authorized to Ahave access to all records, reports, audits, reviews, documents, papers, recommendations or other material@ relating to the program under investigation.

Certainly, the 5^th Circuit=s recent decision did not discuss, much less strike a balance between, the important, but somewhat divergent, goals of furthering accountability and promoting flexible, efficient decision making. That court=s decision simply concluded, without significant analysis or relevant citation, that ADRA does not apply and that, if it did, it would not prevent disclosure in that case.

As a result, the rationale for the court=s conclusions is unclear. It may have reached its conclusions by accepting one or more of several dubious arguments advanced by the U.S. attorney:

$ that ADRA was intended to apply only to administrative programs created after the act=s passage, and not to preexisting ones. This bald assertion finds no support in the act=s text or legislative history, and, if accepted, would render the act the only federal cross-cutting procedural statute on the books to be so limited.

$ that there is a distinction between Aconfidential@ and Aprivileged@ treatment, suggesting that the act=s Aconfidentiality@ protections are somehow inferior to those of a Aprivilege.@ In fact, this is an irrelevant distinction. The real question is: What does the applicable law require? Here, creation of any Anew@ privilege is not an issue; the act has already established a statutory scheme for defining and protecting dispute resolution communications. A core rationale for the act=s confidentiality provision was a perception that existing common law and statutory protection for settlement discussions was incomplete (or uncertain in application) and thus inadequate. Rather than rely on existing, presumably inferior, common law Aprivileges,@ Congress created a comprehensive statutory scheme that defines the extent of protection to be afforded dispute resolution communications in agency-related ADR.

$ that providing confidential data to a grand jury, whose records and deliberations are supposed to be secret, would not constitute Apublic disclosure@ because any use of the information would be limited. Certainly, the act=s prohibition on disclosure makes no distinction based on whether disclosure is Apublic.@ Given recent experience with such ostensibly Asecret@ prosecutorial and grand jury investigations, any assurances of prosecutorial or grand jury self-restraint offer little comfort to their targets. They certainly provide little encouragement for future parties in agency mediations to be Aforthcoming and candid.@

$ that the act is limited to civil matters and was never intended to apply to data sought in criminal investigations. There is no indication in the act or its legislative history of any such limitations. Indeed, one of the exceptions to the act=s general prohibition on disclosure deals explicitly with criminal matters. It states that a person may obtain disclosure of confidential communications if he or she can persuade a court to find that the need to Ahelp establish a violation of law@ is of sufficient magnitude in the particular case to outweigh the integrity of dispute resolution proceedings in general by reducing the confidence of parties in future cases that their communications will remain confidential.

In the ADRA=s careful balance between open government and confidentiality, Congress made clear the standards and procedures that should govern whenever issues of confidentiality arise in agency-related ADR. ADRA is the most detailed, and only specifically applicable, authority, because the act=s procedures explicitly provide for the possibility of disclosure whenever a compelling case can be made to a judge, and because courts have held that an inspector general=s Aofficial curiosity@ is not an adequate grounds to obtain access to protected documents.³ Many courts, confronted for the first time with this issue, would apply the Act whole cloth, as did the district court in Texas.

To safeguard against this, the act puts disclosure decisions in the hands of the courts. It should be apparent that serious harm could result if decisions as to what should, or should not, be confidential are left in the hands of the person seeking the information B whether to inspectors general, other law enforcement authorities, or any other person.

While no one would maintain that parties who fraudulently manipulate Amediation@ to cheat the government should be immune from investigation, an inspector general or U.S. attorney has the opportunity to convince a judge to grant access to specified information that would otherwise be protected by statute and ethical codes.

A judge who is asked to order disclosure of confidential communications would be expected to assure that parties have been given a chance to object to disclosure and then undertake a careful balancing using the Act=s criteria. This would require some meaningful consideration of the evidence sought and an explicit weighing of competing needs (including those of future parties and mediators). Inevitably, this case-specific inquiry would require considerably more than the conclusory (and premature) assertion, like the 5^th Circuit=s, that A[i]f an indictment is returned, any interest the [defendants] have in the confidentiality of their mediation sessions will have to give way to the public interest in the administration of justice.@

The harm of the 5^th Circuit=s dismissive approach could be substantial. It could create doubts and concerns that would extend far beyond any single mediation program, or even agricultural mediation programs across the nation. Prospective parties in ADR involving any federal agency subject to inspector general oversight would begin to worry that their communications could indeed be used against them later and decide to avoid mediating with the government altogether.

Such a scenario is hardly farfetched B in fact, in 1996, when Congress permanently reauthorized the act, it heeded concerns that some data=s availability under FOIA was inhibiting potential ADR use in agency programs, and provided that the act is a statute protecting against disclosure under FOIA.

Under ADRA, President Clinton has designated the attorney general as the authority responsible for convening an interagency working group to facilitate and encourage agency use of alternative means of dispute resolution under the act.⁴ It remains to be seen whether the Department of Justice can simultaneously encourage federal ADR use and also take positions in court B as it did before the 5^th Circuit B that would cut back on statutory protections intended to promote ADR use.

What is apparent, though, is that the briefs submitted by the U.S. attorney, and relied on by the 5^th Circuit, betray both a Aresults-oriented@ bias and a lack of understanding of the act=s intent and operation. This is troubling, and could prove detrimental to future agency ADR use. At a minimum, the Justice Department should assure that, in the future, its representatives possess a greatly improved understanding of, and appreciation for, a statute whose effective implementation it is now charged with promoting.

Without reaching any conclusions as to any individual audit=s accuracy or criminal investigation=s procedural fairness, and without assessing legal or moral culpability, it is still clear that every citizen has a right to fair procedure under law. The act=s coverage is broad. Its language precluding voluntary and compulsory disclosure is explicit. Its exceptions are narrowly drawn. Its procedures are clearly spelled out. And its stated intent is clear: to provide parties with the assurance that communications they make in ADR proceedings will not later be used against them. It is not too much to expect that an agency inspector general, a grand jury, or a U.S. attorney, should be prepared to meet those standards if it wishes to gain access to information denominated confidential by statute.

At the Section of Administrative Law & Regulatory Practice we are always looking for new and better ways to serve our members, the bar and the public. If you have any comments, ideas or features you would like us to incorporate, or if you have difficulties with any of the links in these pages, please contact the Section's Webmaster.