The Chairman of an investigative committee of the Pennsylvania House of Representatives and other leaders of the House ("the Legislators") have refused to comply with a federal grand jury subpoena seeking certain records of the committee's investigation into alleged contractual fraud.1 They assert a common law speech or debate privilege for state legislators cognizable in federal courts under Federal Rule of Evidence 501. On the basis of this asserted privilege, they petitioned the district court to quash the subpoena. Applying a balancing test to the competing interests, the district court granted much of the requested relief. The United States, acting for the grand jury, now appeals.

In United States v. Gillock, 445 U.S. 360, 100 S. Ct. 1185, 63 L. Ed. 2d 454 (1980), the Supreme Court rejected a claim that a speech or debate privilege prevents a prosecutor from using evidence of a state legislator's legislative acts in a prosecution against him. The Legislators and the district court believe that Gillock is distinguishable because it involved a more compelling federal interest than that at stake in this case; essentially they claim that Gillock rejected only an absolute privilege and does not preclude a qualified speech or debate privilege that protects the documents requested here.

Gillock, however, was not based on a general balancing of competing interests that may differ from case to case. Instead, the Supreme Court engaged in a particularized analysis of whether the interests protected by the Constitution's Speech or Debate Clause justify extension of that protection to state legislators. An analysis along the lines suggested by Gillock leads also to rejection of a qualified privilege. Neither the threat of harassment by the federal executive or judiciary, nor the dangers of distraction, nor the potential disruption of confidential communications justifies a qualified privilege for the full range of legislative activities normally protected by the Speech or Debate Clause.

Although we do not foreclose the possibility of a qualified privilege limited only to confidential deliberative communications, that issue is not squarely before us on the present record. The district court erred in concluding that the records at issue were insulated from production by a qualified privilege. Its order limiting the subpoena will therefore be reversed.

In early 1985, a select committee of the Commonwealth of Pennsylvania's House of Representatives (the "Committee") began an investigation into alleged improprieties, including payoffs, in the procurement of granite for an expansion of the state capitol in Harrisburg. The Committee reviewed numerous documents. It also interrogated witnesses in both open and closed session and through its investigators, in the process generating transcripts, summaries, notes and memoranda of its interviews.

In January, 1986, at the suggestion of the Pennsylvania Department of General Services, the United States Attorney for the Middle District of Pennsylvania convened a grand jury to investigate the same occurrences. Shortly thereafter, agents of the FBI met with Representative Nicholas Colafella, Chairman of the Committee, and with Committee counsel Reizdan Moore. These men advised the agents generally of the progress of the investigation but confined themselves essentially to matters of public record. Although Moore stated in an affidavit submitted to the district court that he had disclosed his impressions about potential sources of evidence and the Committee's intentions regarding certain witnesses, Colafella and Moore denied the agents access to documents, transcripts and notes not yet made public. Colafella and Moore also refused to reveal the names of many witnesses, the substance of what they said or whether they may have committed any federal crimes.

On January 15, 1986, following this interview, the U.S. Attorney caused a grand jury subpoena duces tecum to be directed to Representative Colafella or the records custodian of the Committee, demanding production of the following documents:

All records in your care, custody and control relating to an investigation of the purchase of granite use [sic] in the Capitol addition project as follows:

1. Memos, reports, summaries, interview notes, record review notes, correspondence and any other documents of an investigative nature compiled, adopted, prepared or sent and received by any committee investigator including, but not limited to, the listed individuals and companies named on the attached list [omitted].

2. Agenda, minutes, summaries and all transcripts of public and private hearings, along with copies of all documents or exhibits referred to in said transcripts summaries, agenda or minutes during testimony or in comments by committee members or staff.

The subpoena also stated that in lieu of personal appearance, the documents could be mailed to the U.S. Attorney's office.

Colafella responded by moving to quash the subpoena, asserting that he need not produce the records because of a federal common law speech or debate privilege for state legislators under Federal Rule of Evidence 501. With the consent of the United States Attorney, several legislators were permitted to intervene in alignment with Colafella, including the Speaker of the Pennsylvania House of Representatives, the majority and minority leaders and the Bipartisan Committee for the Administration of the House. The district court held an immediate hearing, and the parties stipulated to many facts including the fact that no legislator was then a subject or target of the investigation.

Much of the hearing focused on the nature of the documents sought by the subpoena and whether they would reveal the thought processes of legislators or their aides. The U.S. Attorney repeatedly stated that he sought only "hard" evidence:

... I am basically asking for ... investigative reports of interviews conducted with yet other individuals and the sworn testimony of other individuals whether taken publicly or in executive session. I want no attorney-client privilege nor do I ask that the grand jury be tainted by documents that are privileged in any other manner including legislative thought processes or work processes.

The Legislators informed the court that the Committee had already made many of the transcripts and documents public, but the U.S. Attorney rejoined: "It is my belief that there are additional documents and interviews which have not been made public which are vital and relevant. If it could be represented that there are not, I would withdraw my subpoena and abide the event."

The Legislators responded that the documents sought would in fact reveal the thought processes of committee members and their aides. The Legislators argued that the revelation of thought processes significantly impinged on legislative independence and the values of federalism. Moreover, the Legislators claimed that the subpoena did not further significant federal interests in law enforcement because: (1) most of the subpoenaed information was obtainable by the Grand Jury from other sources, and (2) the fact that no legislator was under investigation indicated that the Grand Jury sought to use the Committee merely as its investigative tool.

Following oral argument and full briefing, the district court agreed substantially with the Legislators. The court adopted a general balancing approach to the privilege question and essentially accepted the Legislators' evaluation of the competing interests. The court therefore quashed the subpoena in most particulars. The Legislators remained obligated only to disclose the identity of committee witnesses and to produce documents which the Committee had not itself authored if the Grand Jury could demonstrate that they were unobtainable through other means.

The district court entered its order on January 31, 1986. On February 6, the Committee released numerous documents to the public and sent copies of these documents directly to the U.S. Attorney. The U.S. Attorney did not believe that these documents included all those within the reach of the subpoena and therefore appealed to this court. When the Legislators filed their brief, however, they appended an affidavit by Committee Counsel Moore which stated that "all relevant or remotely relevant information regarding payoffs which was obtained on or before February 6, 1986, has been released to the public and turned over to the United States Government...."

The case was listed for oral argument. In view of the Moore affidavit, the panel was concerned whether a viable case or controversy existed and inquired whether any documents sought by the subpoena remained unproduced. The ensuing colloquy did not resolve the question. At the conclusion of oral argument, the panel suggested to counsel that they attempt to resolve their differences and report to the Court in writing. However, the parties remained unable to settle their dispute or even to agree whether documents subject to the subpoena remained unproduced. Uncertain whether the case was or was not moot, the panel retained jurisdiction but remanded the matter to the district court for a determination of mootness. 802 F2d 446.

The district court received supplemental submissions and determined that "a case or controversy still exists." The court found that the Committee had not released several specific items "which fall within the scope of the subpoena:"

1) memorandums, summaries of interviews, interview notes, reports of interviews with Carl Lunderstadt, John Scott, Robert Billingsly, Mr. Bernardo.

2) memorandums, summaries, notes, reports of interviews of potential witnesses interviewed by the committee's investigator, Stanley Gochenour and the committee's attorney, Reizdan Moore.

The district court also found that the Legislators did not have to produce these items under the terms of its previous order quashing most of the subpoena's request.

The case has now returned to us from the district court, and we have received supplemental submissions. Despite the district court's findings, the Legislators continue to press a claim of mootness. They argue first that the U.S. Attorney effectively modified the subpoena by stating to the district court that he would withdraw the subpoena if the Legislators represented that all vital and relevant documents and interviews had been made public. The Legislators contend that they have complied with this modified subpoena through Moore's representation that "all relevant or remotely relevant information" has been turned over to the government. The Legislators additionally claim mootness because the Committee ceased functioning on November 26, 1986 when the legislature expired to be replaced by a newly elected legislature. As of that day, the Legislators claim, Representative Colafella no longer had authority to produce the documents requested by the subpoena.

In addition to these mootness claims, the parties have reasserted their previous arguments on the merits.

The Legislators' suggestion that they have complied with the subpoena by providing the statement requested by the U.S. Attorney amounts to a kind of compliance by estoppel argument. Even assuming the viability of such a theory, however, one requirement of any estoppel claim is a showing of detrimental reliance. The Legislators have not explained how they relied on the U.S. Attorney's statements to their detriment.

Furthermore, the U.S. Attorney stated that he would be satisfied only if the Legislators represented that no relevant "documents or interviews" remained unavailable. Moore's affidavit referred only to "relevant information." The district court has specifically found that the Legislators have not turned over many relevant documents. This finding indicates at the least that Moore's view of "relevant information" does not conform to the U.S. Attorney's understanding of all relevant "documents." It further indicates that documents sought by the subpoena remain unproduced. Because the Legislators do not challenge that finding as unsupported, we must agree with the district court that the Legislators have not complied with the subpoena; hence the case is not moot on the ground of compliance.

The Legislators' second claim of mootness, lack of possession or legal control, is equally without merit. A party's lack of possession or legal control over documents requested by a subpoena is normally a valid defense to a subpoena and justification for a motion to quash. But such a defense presents a question on the merits not one of mootness. See In the Matter of Grand Jury Impaneled January 21, 1975 (Freedman & Cortese), 541 F.2d 373, 377 (3d Cir.1976) (treating claim of lack of legal authority to produce records as question of the merits). Moreover, it may involve a number of factual disputes, and to succeed on such a defense, Representative Colafella has the burden of producing at least some evidence that he does not have possession or control of the documents sought. See McPhaul v. United States, 364 U.S. 372, 378-79, 81 S. Ct. 138, 142-43, 5 L. Ed. 2d 136 (1960) (party claiming lack of control as defense to congressional subpoena duces tecum must testify that he does not possess or control records); United States v. O'Henry's Film Works, Inc., 598 F.2d 313, 318 (2d Cir.1979) ("[W]hen the agent fails to produce documents that are the subject of a valid summons or subpoena, if called before a court, he must give sworn testimony that he does not possess them."). The Legislators have offered no evidence whatsoever, and we will not postpone our consideration of the district court's order because of unsubstantiated allegations that might support a separate, independent, ground for quashing the subpoena.2 We therefore turn to the merits.3III. THE MERITS

A grand jury exercises broad investigative powers and generally has both the right and the duty to procure every man's evidence. United States v. Dionisio, 410 U.S. 1, 9-10, 93 S. Ct. 764, 769-70, 35 L. Ed. 2d 67 (1973). This power is limited only by constitutional, common law or statutory privilege, or by the exercise of the district court's supervisory authority to "quash or modify the subpoena if compliance would be unreasonable or oppressive." Fed.R.Crim.P. 17(c); see Branzburg v. Hayes, 408 U.S. 665, 668, 92 S. Ct. 2646, 2650, 33 L. Ed. 2d 626 (1972). While appellate courts will normally defer to the district court's exercise of its supervisory authority, see In re Grand Jury Proceedings (Schofield II), 507 F.2d 963, 967 (3d Cir.1975), cert. denied, 421 U.S. 1015, 95 S. Ct. 2424, 44 L. Ed. 2d 685 (1975); In re Grand Jury Matters (Young), 751 F.2d 13, 16 & n. 3 (1st Cir.1984), "[j]udges may not, in the guise of exercising supervisory power, create new privileges or enlarge or distort existing ones," Young, 751 F.2d at 18. The existence and scope of a privilege are questions of law which Congress has directed courts to consider "in the light of reason and experience." Fed.R.Evid. 501.

Because the district court relied on a theory of a common law speech or debate privilege for state legislators, we must reverse its order if the district court's understanding of the privilege was in error. To analyze that question, we must first discuss the contours of the speech or debate privilege for members of Congress, for the Legislators' contend that the same or similar interests motivating that privilege justify a comparable, albeit qualified, privilege for state legislators. We then analyze the Supreme Court's decision in United States v. Gillock, 445 U.S. 360, 100 S. Ct. 1185, 63 L. Ed. 2d 454 (1980), and the district court's reasons for distinguishing it. Finally, we apply the lessons of Gillock to decide whether we should recognize the qualified privilege claimed in this case.

B. The Federal Congressional Privilege and Gillock

The speech or debate privilege for members of the United States Congress has its source in the Speech or Debate Clause of the Constitution, Art. I., Sec. 6, cl. 1. Its " 'central role' ... is to 'prevent intimidation of legislators by the Executive and accountability before a possibly hostile judiciary.' " Eastland v. United States Servicemen's Fund, 421 U.S. 491, 502, 95 S. Ct. 1813, 1821, 44 L. Ed. 2d 324 (1975) (citing Gravel v. United States, 408 U.S. 606, 617, 92 S. Ct. 2614, 2623, 33 L. Ed. 2d 583 (1972)). It also prevents disruption of Congressional operations by preventing distractions or interference with ongoing activity. See id. 421 U.S. at 510 n. 16, 95 S. Ct. at 1824-25 n. 16 ("Where we are presented with an attempt to interfere with an ongoing activity by Congress ... [t]he speech or debate protection provides an absolute immunity from judicial interference.") The Speech or Debate Clause accomplishes these goals primarily by providing Congressmen and their aides with absolute immunity from criminal or civil suit, see Gravel, 408 U.S. at 615-16, 92 S.Ct. at 2622 (criminal); Eastland, 421 U.S. at 503, 95 S.Ct. at 1820 (civil), for "acts that occur in the regular course of the legislative process," United States v. Brewster, 408 U.S. 501, 525, 92 S. Ct. 2531, 2544, 33 L. Ed. 2d 507 (1972).

In order to insure that Congressmen are not held liable for legislative acts, the Speech or Debate Clause mandates an evidentiary privilege that prohibits evidence of a Congressman's acts from being used in a proceeding against him. See United States v. Helstoski, 442 U.S. 477, 488-89, 99 S. Ct. 2432, 2439-40, 61 L. Ed. 2d 12 (1979). Because "hostile questioning by the executive branch before a possibly hostile judiciary," may also impinge on legislative independence, In re Grand Jury Investigation ("Eilberg"), 587 F.2d 589, 597 (3d Cir.1978), the speech or debate privilege is also testimonial and protects a Congressman or his aides from judicial questioning about his legislative acts, Gravel, 408 U.S. at 615-16, 92 S.Ct. at 2622; Eilberg, 587 F.2d at 596.

Both prongs of this evidentiary privilege have two crucial features. First, in keeping with the broad scope of Congressional immunity, the evidentiary privilege applies broadly to evidence or testimony about all "acts that occur in the regular course of the legislative process." Brewster, 408 U.S. at 525, 92 S.Ct. at 2544. It therefore applies to activity taken in the course of legislative factfinding. Eastland, 421 U.S. at 504-05, 95 S.Ct. at 1822 (1975); see also Government of the Virgin Islands v. Lee, 775 F.2d 514, 521 (3d Cir.1985).4 Second, the speech or debate privilege is "absolute;" hence, it cannot be overcome by any countervailing interest no matter how strong. Eastland, 421 U.S. at 509-510 n. 16, 95 S. Ct. at 1824 n. 16.

For many years, courts of appeals differed over the recognition of a speech or debate privilege for state legislators. Some courts held that because the evidentiary privilege for Congressmen exists primarily to protect Congressional immunity, a comparable evidentiary privilege for state legislators, who lack this immunity, made little sense. See, e.g., United States v. Craig, 537 F.2d 957 (7th Cir.) (en banc), cert. denied, 429 U.S. 999 (1976); United States v. DiCarlo, 565 F.2d 802, 807 (1st Cir.1977), cert. denied, 435 U.S. 924, 98 S. Ct. 1487, 55 L. Ed. 2d 517 (1978). In contrast, a panel of this court recognized the privilege because of the historically accepted view that "a legislature will function better in the public interest with the protection than without it." In re Grand Jury Proceedings (Cianfrani), 563 F.2d 577, 583 (3d Cir.1977). This conflict between the circuits was resolved by the Supreme Court in United States v. Gillock, 445 U.S. 360, 100 S. Ct. 1185, 63 L. Ed. 2d 454 (1980).

In Gillock, the Supreme Court considered whether a common law speech or debate privilege barred federal prosecutors from introducing evidence of a Tennessee state representative's support for particular legislation in a prosecution against the representative for bribery.5 Chief Justice Burger's decision for the court separated the purposes of the speech or debate privilege into "[t]wo interrelated rationales ...: first, the need to avoid intrusion by the Executive or Judiciary into the affairs of a coequal branch, and second, the desire to protect legislative independence." 445 U.S. at 369, 100 S.Ct. at 1191.

The interest in preventing intrusion, the Court held, provides no support for a privilege for state legislators in federal cases because the Federal Government has supremacy over state legislatures in areas where the Constitution grants its power to act. Although the Court noted that interests of comity and federalism "command careful consideration," 445 U.S. at 373, 100 S.Ct. at 1193-94 it held that "federal interference in the state legislative process is not on the same constitutional footing with the interference of one branch of the Federal Government in the affairs of a coequal branch," 445 U.S. at 370, 100 S.Ct. at 1192 (citations omitted), and could not justify the privilege requested in that case.

Turning to the second rationale, the interests in legislative independence, the Court recognized that "denial of a privilege to a state legislator may have some minimal impact on the exercise of his legislative function" and that interests of comity commanded some concern for this impact. 445 U.S. at 373, 100 S. Ct. at 1193-94. But the court noted that it had traditionally drawn the line at protection of state legislators in civil suits. The courts have no power, it reasoned, through the creation of an evidentiary privilege "to immunize criminal conduct proscribed by an Act of Congress." 445 U.S. at 372, 100 S.Ct. at 1193 (emphasis omitted) (citing Gravel, 408 U.S. at 627, 92 S.Ct. at 2628). Accordingly, the Court held that no common law speech or debate privilege bars federal authorities from using evidence of legislative acts in prosecutions against state representatives.

The district court distinguished Gillock on the grounds that the need to use evidence of legislative acts against the Tennessee representative provided a more compelling federal interest than the need in this case to obtain evidence from a state legislator that is independently obtainable. In doing so, the district court essentially read Gillock as denying state legislators only an absolute speech or debate privilege. In the implicit reasoning of the district court, Gillock did not preclude a qualified speech or debate privilege for state legislators, i.e., a privilege applicable to the full range of legislative activities normally protected by the Speech or Debate Clause for members of Congress but, when claimed by state legislators, surmountable by a federal prosecutor's particularized showing of need.

Having distinguished Gillock to its satisfaction, the district court proceeded to balance the competing interests of federal law enforcement and state legislative independence in this particular case to decide whether it should uphold the privilege claim. On the federal side, the court noted that the grand jury could obtain virtually all the information it sought from an independent investigation and that no state legislator was a target of the investigation. On the side favoring the privilege, the district court placed "the integrity of the legislative process." Balancing these interests, the court held "that the policy underlying the asserted privilege outweighs that favoring disclosure" and thus justifies the protection requested by the Legislators. Only if the Grand Jury could demonstrate a particularized need for a specific document by showing that it could not obtain the document through another source would the district court be prepared to command compliance with the subpoena.

We agree with the district court that Gillock did not by its terms foreclose all possible forms of privilege for state legislative activity. Gillock made clear the importance the Supreme Court attached to the criminal context in which the case arose. Furthermore, the federal interest was obviously vital in that case because, in the Court's own terms, the privilege sought by Gillock threatened to immunize conduct proscribed by Congress. In contrast, the interest in enforcing the subpoena in this case is less crucial. The subpoena will assure the Grand Jury all the hard evidence obtained by the Committee, and the statements obtained by the Committee when the trail was fresh may provide insights not available later on or may provide important corroboration. Recognition of the privilege, however, is unlikely to thwart the investigation entirely. In light of this difference, the district court was correct not merely to cite Gillock and to forego all further analysis but instead to evaluate the request for a qualified privilege with the additional use of its "reason and experience." Fed.R.Evid. 501, see typescript supra at 14.

Notwithstanding differences between this case and Gillock, we disagree with the district court's method of analysis in three important respects. First, the district court analyzed the Legislators' privilege request entirely in the context of this case. Such a narrow focus was in error because a court must first decide whether a qualified privilege exists or should exist before deciding how to apply it to a particular case.

By insisting on a two-step process, courts guide their discretion with rules developed from accumulated wisdom about the situations that justify a privilege. The requirement of general privilege rules also enables private actors to rely on those rules and thereby to engage more freely in the conduct privileges aim to encourage. Thus, although the concept of the qualified privilege permits courts to uphold or reject privilege claims in light of the particulars of an individual case, the decision to recognize a qualified privilege must still follow from a more broad-based view of how the privilege will work in general.

Second, although all privilege analysis involves some balancing, we disagree with the abstract nature of the district court's balancing approach. Common law privileges should be "accepted 'only to the very limited extent that permitting a refusal to testify or excluding relevant evidence has a public good transcending the normally prudent principle of utilizing all rational means for ascertaining truth.' " Trammell v. United States, 445 U.S. 40, 50, 100 S. Ct. 906, 912, 63 L. Ed. 2d 186 (1980) (citations omitted). The justification for a privilege must be specific, for courts should apply privileges "only to the extent necessary to achieve their purpose." Freedman & Cortese, 541 F.2d at 382 (citing Fisher v. United States, 425 U.S. 391, 403, 96 S. Ct. 1569, 1577, 48 L. Ed. 2d 39 (1976)).

In Trammel for example, the Supreme Court rejected the historically accepted privilege that permitted one spouse to exclude the testimony of the other despite the claim that it fostered "the harmony and sanctity of the marriage relationship." Trammell, 445 U.S. at 44, 100 S.Ct. at 909.6 The court found that the asserted benefit rested on a generalization and not on the required, sound, practical explanation of how the privilege would protect the relationship, id. at 51-53, 100 S.Ct. at 912-14. Similarly here, the district court should not have terminated its analysis with the assertion of the need for "legislative independence" but should have examined more specifically how a common law speech or debate privilege would realistically assist legislators to perform their function.

Finally, we do not believe the district court's efforts merely to distinguish Gillock paid proper respect to its significance. Gillock's reasoning undercuts the weight we may accord interests of comity and federalism, and Gillock's practical impact undermines the extent to which recognition of a speech or debate privilege for state legislators may realistically further the interests that the privilege has traditionally safeguarded. To analyze the Legislators' claim, we must evaluate the specific reasons for and against a qualified speech or debate privilege in light of the theoretical and practical limitations imposed by Gillock.

D. Analysis of Qualified Speech or Debate Claim in Light of Gillock

As we have discussed, the core rationale for the speech or debate privilege has traditionally been the protection of legislative independence through a "prophylactic" rule assuring members of Congress that their legislative activities cannot lead to harassment by the executive or the judiciary. United States v. Johnson, 383 U.S. 169, 182, 86 S. Ct. 749, 756, 15 L. Ed. 2d 681 (1966); see also Eastland, 421 U.S. at 502, 95 S.Ct. at 1821. The need to prevent this harassment has overridden the danger of legislative misconduct inadvertently protected by the privilege. See Brewster, 408 U.S. at 516, 92 S.Ct. at 2539. Gillock made clear, however, that the concern with potential harassment rests on an underlying concern for maintenance of the separation of powers. 445 U.S. at 370, 100 S. Ct. at 1192. Federalism and comity do not raise the same concerns because the Supremacy Clause sanctions "interference" with state legislatures by the federal executive and judiciary so long as they pursue legitimate federal aims. Accordingly, Gillock 's reasoning suggests that prevention of intimidation is simply not a legitimate rationale for extension of the privilege to state legislators.

As a practical matter, Gillock 's holding also makes the prevention of intimidation and harassment impossible. The greatest potential for intimidation, of course, lies in the power of the federal government actually to prosecute legislators for their actions. The prophylactic function of the privilege is therefore undermined once federal officials may prosecute legislators for their legislative acts and obtain and use evidence of legislative acts for that purpose. Although the federal government may have less interest in intruding on state legislative functions when no misconduct is alleged, a qualified privilege that gives way when misconduct is alleged cannot insulate state legislators from intimidation. A qualified privilege therefore cannot serve a purpose for state legislators similar to that which lies at the core of the privilege for members of Congress.7

In addition to this core rationale, Gillock also recognized that legislative independence may in some senses be unrelated to the threat of intimidation. Even apart from the potential impact federal intimidation may have on a legislator's activity, a subpoena may burden a state legislator with onerous production requirements or may force a state legislator to disclose confidential communications. Having found that the core rationale is inapplicable here, we must still examine whether these threats to the functioning of the legislature justify creation of a qualified privilege that may apply to cases in which the federal interest is less compelling than it was in Gillock.

We certainly agree that a subpoena directed at the records of an ongoing legislative investigation could be unduly burdensome and may intrude on important legislative interests. We take this problem seriously because the citizenry not only has a strong interest in the success of federal criminal prosecutions to punish and to deter misconduct, but also a strong interest in the success of state legislative investigations that may lead to changes preventing future misconduct. Furthermore, the burden on legislators is a speech or debate concern. The Supreme Court has held analogously that the Speech or Debate Clause shields Congressmen from suit to block a Congressional subpoena because making the legislators defendants "creates a distraction and forces Members [of Congress] to divert their time, energy, and attention from their legislative tasks to defend the litigation." Eastland, 421 U.S. at 503, 95 S.Ct. at 1821.

We do not believe, however, that a state legislator's interest in avoiding the burdens of compliance with a subpoena is alone sufficient to justify creation of a speech or debate privilege. Gillock instructs us that any such privilege must be qualified, not absolute, and must therefore depend on a balancing of the legitimate interests on both sides. Yet Rule 17 of the Federal Rules of Criminal Procedure already authorizes a district court to quash or modify a subpoena that is unreasonable or oppressive. That rule permits the district court to protect state legislators from undue disruption of their activities while still accommodating the need for information by the party issuing the subpoena. Interests of comity and federalism require that a district court respect the needs of state legislators, but Rule 17 continues to be the proper vehicle for accommodating the competing interests.

The state legislator's need for confidentiality is more troublesome. As a panel of the District of Columbia Circuit has noted, the legislator's need for confidentiality is similar to the need for confidentiality in communications between judges, between executive officials, and between a President and his aides. See Nixon v. Sirica, 487 F.2d 700, 717 (D.C.Cir.1973). The need for a full, frank exchange of ideas has led courts to recognize qualified privileges for each of these governmental decisionmakers. See, e.g., United States v. Nixon, 418 U.S. 683, 708, 94 S. Ct. 3090, 3107-08, 41 L. Ed. 2d 1039 (1974) (recognizing privilege for Presidential communications because "[a] President and those who assist him must be free to explore alternatives in the process of shaping policies and making decisions and to do so in a way many would be unwilling to express except privately"); United States v. Weber Aircraft, 465 U.S. 792, 802, 104 S. Ct. 1488, 1494, 79 L. Ed. 2d 814 (1984) (recognizing privilege for executive officials because history of governmental privileges "recognizes a need for claims of privilege when confidentiality is necessary to ensure frank and open discussion and hence efficient governmental operations") (citations omitted); In Certain Complaints Under Investigation by an Investigating Committee of the Judicial Council of the Eleventh Circuit ("Hastings"), 783 F.2d 1488, 1519 (11th Cir.), cert. denied, --- U.S. ----, 106 S. Ct. 3273, 91 L. Ed. 2d 562 (1986) (recognizing judicial privilege because "[j]udges, like Presidents, depend upon open and candid discourse with their colleagues and staff to promote the effective discharge of their duties.").

Even while granting governmental officials a qualified, confidentiality privilege, however, the courts have taken pains to insure that the privilege applies only "to the very limited extent" that the "public good" in confidentiality transcends the value of "utilizing all rational means for ascertaining truth." Trammell, 445 U.S. at 50, 100 S.Ct. at 912 (quotation omitted). In Nixon, for example, the enormous importance and sensitivity of the President's personal communications motivated the Supreme Court to hold that these communications are "presumptively privileged," requiring the prosecutor to make a threshold showing of need and permitting the prosecutor access only to relevant, admissible evidence. Nixon, 418 U.S. at 713-16, 94 S.Ct. at 3110-11. However, the Court held that the President's "generalized interest in confidentiality," id. at 713, 94 S.Ct. at 3110, could not prevail over the need for "evidence that is demonstrably relevant in a criminal trial," id. at 712, 94 S.Ct. at 3110.

When dealing with lesser governmental officials, courts have treated claims of privilege even more strictly. They have required a threshold showing that communications of executive or judicial officials involve confidential deliberations about the decisions entrusted to them before those communications become presumptively privileged. See, e.g., Hastings, 783 F.2d at 1520-22 (11th Cir.1986) (judicial privilege applies only upon showing that matters under inquiry implicate communications among a judge and his staff concerning performance of judicial business "such as ... the framing and researching of opinions, orders and rulings"); N.L.R.B. v. Sears, Roebuck & Co., 421 U.S. 132, 150-54, 95 S. Ct. 1504, 1516-18, 44 L. Ed. 2d 29 (1975) (executive privilege for agency officials applies only to documents "reflecting advisory opinions, recommendations and deliberations comprising part of a process by which governmental decisions and policies are formulated") (quoting Carl Zeiss Stiftung v. V.E.B. Carl Zeiss, Jena, 40 F.R.D. 318, 324 (D.D.C.1966).

These precedents reflect the judicial reasoning and experience required by Rule of Evidence 501. In accordance therewith, we do not believe that the needs of state legislators for confidentiality justify the creation of a qualified privilege for the full range of legislative activities normally protected by the Speech or Debate Clause. Not all acts "that occur in the regular course of the legislative process," Brewster, 408 U.S. at 525, 92 S.Ct. at 2544; will involve confidential communications, and many communications actually made in confidence have only limited need for confidentiality. For this reason, our prior decisions have recognized that confidentiality does not lie at the root of the concerns motivating a privilege for all legislative speech or debate. The speech or debate privilege is at its core a "use privilege" not a privilege of nondisclosure. See Government of Virgin Islands v. Lee, 775 F.2d 514, 523 (3d Cir.1985); In re Grand Jury (Eilberg), 587 F.2d 589, 595-96 (3d Cir.1978).

Our survey of the suggested rationales for the speech or debate privilege for members of Congress convinces us that after Gillock even a qualified speech or debate privilege for state legislators would not realistically serve the purposes that the Speech or Debate Clause is intended to advance. Neither the threat of harassment, the dangers of distraction, nor the potential disruption of confidential communications justifies a qualified privilege for the full range of legislative activities normally protected by the Speech or Debate Clause. We therefore refuse to recognize such a privilege; hence we reject the underlying rationale for the district court's decision.

Although we reject a qualified speech or debate privilege, our discussion does not preclude the possibility of a more narrowly tailored privilege for confidential deliberative communications. Notwithstanding their advocacy of a broad speech or debate privilege, the Legislators have relied in oral argument on confidentiality concerns. The district court apparently assumed that state legislative documents are privileged merely because they may somehow reveal the "thought processes" of state legislators. Although we are reluctant to go so far as to recognize a confidentiality privilege and to delineate its contours, we believe we must analyze the district court's assumption to provide guidance for remand. We note in this regard that this case has already been before us twice and that the Grand Jury's investigation has been delayed by more than a year.

To analyze the Legislators' confidentiality concerns, we look for guidance to the "deliberative process privilege" for executive officials. Of all confidentiality privileges for government officials, that privilege has probably received the most intricate explication by the courts. See generally M. Larkin, Federal Testimonial Privileges Sec. 5.02 & 5.07, at 5-13 to 5-30, 5-87 to 5-90 (1986); Project, Developments in the Law Privileged Communications, 98 Harv.L.Rev. 1450, 1592, 1620-23 (1985). We also believe that it provides a useful analogy for a confidentiality-based privilege for state legislators because executive agencies, like state legislators, engage in a wide variety of activities, including factual investigations for quasi-legislative rulemaking.8

The "deliberative process privilege" applies only to confidential deliberations of law or policymaking, reflecting opinions, recommendations or advice. See NLRB v. Sears Roebuck & Co., 421 U.S. at 150-54, 95 S.Ct. at 1516-18; EPA v. Mink, 410 U.S. 73, 89 & n. 16, 93 S. Ct. 827, 837 & n. 16, 35 L. Ed. 2d 119 (1973). "[I]n most situations, factual summaries prepared for informational purposes" will not be protected even if they are part of an investigative record. Paisley v. C.I.A., 712 F.2d 686, 699 (D.C.Cir.1983) (citing cases). Even if documents contain advisory opinions, factual material which is severable is not protected. See Mink, 410 U.S. at 88-91, 93 S.Ct. at 836-38. Furthermore, the party seeking disclosure may overcome the claim of privilege by showing a sufficient need for the material in the context of the facts or the nature of the case, see Sears Roebuck & Co., 421 U.S. at 149 n. 16, 95 S. Ct. at 1516 n. 16; FTC v. Warner Communications, Inc., 742 F.2d 1156, 1161 (9th Cir.1984); or by making a prima facie showing of misconduct, see KFC Nat'l Management Corp. v. NLRB, 497 F.2d 298, 305 (D.C.Cir.1974), cert. denied, 423 U.S. 1087, 96 S. Ct. 879, 47 L. Ed. 2d 98 (1976); Singer Sewing Machine Co. v. NLRB, 329 F.2d 200, 208 (4th Cir.1964).

These limitations reflect the careful tailoring of the privilege to achieve its purpose of protecting confidentiality without unduly inhibiting the truth-finding process of litigation. Without holding that a deliberative process privilege does protect state legislators, we therefore reject the district court's assumption that a privilege protects all state legislative documents which may somehow reveal the "thought processes" of state legislators. Confidentiality concerns do not sweep so broadly. If a deliberative process privilege does protect state legislators, it is limited to communications involving opinions, recommendations or advice about legislative decisions. In addition, severable, factual material cannot be privileged.

In this case, the U.S. Attorney has repeatedly disclaimed interest in confidential deliberations and has stated that he seeks only hard evidence. Most if not all of the documents sought in this case are notes, summaries and transcripts of interviews that are unlikely to contain deliberative communications. We therefore suspect that the district court will be able to resolve any dispute about confidential portions of subpoenaed material without depriving the Grand Jury of the materials it seeks merely by exempting or excising any confidential deliberations that the documents contain. If necessary, the court may order in camera inspection of disputed documents. See Kerr v. United States District Court, 426 U.S. 394, 406, 96 S. Ct. 2119, 2125, 48 L. Ed. 2d 725 (1976) ("in camera review is a highly appropriate and useful means of dealing with claims of governmental privilege").

For the foregoing reasons, the order of the district court limiting the subpoena will be reversed and the case remanded for proceedings consistent with this opinion.