UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT
June 27, 1984
JOSEPH A. SULLIVAN, SHERIFF, DEPUTY SHERIFF SAULL, DEPUTY SHERIFF RODRIZGAS, DAVID OWHES (DIRECTOR OF INMATE SERVICES), THOMAS KELLY (WARDEN HOLMESBURG PRISON), GUARD BERSANE, GUARD HANNAH, GUARD OCEHSLE, DEPUTY SHERIFF JOHN DOE, DEPUTY SHERIFF JOHN DOE, DEPUTY SHERIFF JOHN DOE, DEPUTY SHERIFF JOHN DOE, DEPUTY SHERIFF JOHN DOE UNITED STATES MARSHALS SERVICE, APPELLANT
On Appeal from the United States District Court for the Eastern District of Pennsylvania.
Gibbons and Becker, Circuit Judges and Atkins, District Judge.*fn*
Opinion ANNOUNCING THE JUDGMENT OF THE COURT
GIBBONS, Circuit Judge:
The United States Marshals Service, an intervenor, appeals from an order of a United States Magistrate in a pending civil action, denying the intervenor's motion for reconsideration of an order directing the United States Marshal for the Eastern District of Pennsylvania to transport a state prisoner witness from the nearest county jail to the United States Courthouse in Philadelphia and to maintain custody of that prisoner while he is in the federal courthouse. The underlying civil action is within the subject matter jurisdiction of the district court, and is assigned for trial before a United States Magistrate by consent of the parties. 28 U.S.C. § 636(c)(1). The Marshals Service contends that the district court, and hence the United States Magistrate, lacked statutory authority to enter the order appealed from. The Bureau of Corrections, Commonwealth of Pennsylvania, the Sheriff of Philadelphia County, and the City of Philadelphia, which have also intervened, defend the order. We hold that the district court, and thus the magistrate, has statutory authority to direct the United States Marshal to maintain custody of a state prisoner witness in the federal courthouse while he is there in connection with judicial proceedings, but lacks such authority to direct the Marshal to transport that prisoner between the county jail and the federal courthouse.
This dispute, between the United States Marshals Service and the Pennsylvania corrections officials as to who must bear the expense of transporting and maintaining secure custody over state prisoners required as witnesses in civil proceedings in the federal courts, is entirely collateral to the underlying action. The Magistrate's order disposes of the dispute fully and finally. Review of final orders in cases tried by consent of the parties before United States Magistrates lies directly to this court.*fn1 28 U.S.C. § 636(c)(3). Thus we have appellate jurisdiction. Story v. Robinson, 689 F.2d 1176, 1177 (3d Cir. 1982).
At issue in this appeal is the authority of a United States District Court to shift the burden of expense in responding to unquestionably lawful federal civil process from the Commonwealth of Pennsylvania to the United States Treasury. The Commonwealth understandably prefers that federal taxpayers bear that burden, and since the United States is often seen as the ultimate deep pocket target, that preference has a plausible appeal. But the United States has not conferred on the federal judiciary any authority to lay hands on its Treasury. To the contrary, the Constitution provides that "no Money shall be drawn from the Treasury, but in Consequence of appropriations made by Law. . . ." Art. I, Sec. 9, cl. 7. Absent statutory authorization, what the Commonwealth proposes is a raid in its interest on the United States Treasury.
This court has dealt with the contention advanced on behalf of the Commonwealth, the County and the City in the recent past. In Story v. Robinson, 689 F.2d 1176 (3d Cir. 1982), the Commonwealth and Allegheny County appealed, as intervenors, from a writ of habeas corpus ad testificandum directing that state prisoner witnesses be transported from state prisons to a county jail near the federal courthouse in Pittsburgh. The Commonwealth objected to transporting the prisoners without reimbursement by the United States, and urged on us several statutes which might authorize the district court to qualify the writ by a reimbursement condition. The first statute relied upon was 28 U.S.C. § 569(b), which instructs the Marshals to "execute all lawful writs, process and orders executed under the authority of the United States. . . ." We held that this statute was not a separate source of judicial authority to issue writs, process and orders, but only a source of the Marshals' duty to execute them when they have been issued pursuant to some other source of authority. 689 F.2d at 1179. A second statute was 28 U.S.C. § 571, which authorizes the United States Marshal to act as disbursing officer for the United States Attorney and for federal court personnel. We held that this statute conferred no authority to relieve state custodians of the financial burden of complying with federal process. 689 F.2d at 1179. A third statute relied on by the Commonwelath was 28 U.S.C. § 567(2), which provides that each United States Marshal shall be allowed "the expense of transporting prisoners. . . ." We held that this statute was no authority, at least absent some intergovernmental agreement, see 28 C.F.R. § 0.111(j) (1983), for relieving the Commonwealth of such expense. Id. Finally, the Commonwealth relied on 28 U.S.C. § 1651. We rejected the Commonwealth's contention that this statute required the grant of the motion to impose the entire cost of compliance with the writs of habeas corpus ad testificandum on someone other than the state custodian. 689 F.2d at 1180.
In the instant case, neither in the district court nor before this court did the Commonwealth and its subdivisions rely upon section 1651 as authority for an order directed to the United States Marshal.*fn2 Technically, the question is open, for while Story v. Robinson rejected the contention that section 1651 required an order that the cost of producing prisoner witnesses must be assumed by the Marshals Service, it did not decide whether the section nevertheless authorized such an order as a matter of discretion. No party has briefed that question, but since such a construction of section 1651 might arguably justify the order appealed from, we address it.
A construction of the All Writs Act as a grant of authority to transfer to the United States Treasury the expense of compliance with witness process is unsupportable. None of the authorities construing section 1651 interpret it as a grant of authority for the issuance of an order compelling a non-party Executive branch officer to satisfy the obligation of state custodians to respond to witness process. United States v. New York Telephone Co., 434 U.S. 159, 54 L. Ed. 2d 376, 98 S. Ct. 364 (1977), on which the Commonwealth relied in Story v. Robinson, supra, involves the authority of the district court to direct process to non-parties who are so situated that their cooperation is essential to the exercise of its jurisdiction otherwise attaching. The Marshals clearly is not so situated. The only cooperation necessary to the exercise of the district court's jurisdiction over pending civil or criminal cases is that of the custodian of needed witnesses.*fn3
In McIntire v. Wood, 11 U.S. [7 Cranch] 504, 3 L. Ed. 420 (1813), the Court held that section 14 of the Judiciary Act of 1789, from which section 1651(a) is derived, was not a source of substantive authority to mandamus federal executive branch officers. Cf. McClung v. Silliman, 19 U.S. [6 Wheat.] 598, 5 L. Ed. 340 (1821) (state court cannot mandamus federal officer). Compare Pub. L. No. 87-748, § 1(a), 76 Stat. 744 (1972), codified at 28 U.S.C. § 1361 (1976) ("Action to compel an officer of the United States to perform his duty"). Since the Marshals are not custodians, they are not appropriate respondents to writs of habeas corpus. Orders directing them to assume the burden of compliance are therefore in the nature of mandamus. Thus the definitive construction of the All Writs Act in McIntire v. Wood precludes reliance on that statute.
Any other interpretation of section 1651 would permit district courts to impose the expenses of litigation upon the Treasury of the United States whenever that seemed to the court to be a good idea. Both common sense and appropriate concern for separation of powers caution against such an arrogation of judicial power. Certainly section 1651 authorizes writs in aid of the district court's jurisdiction otherwise obtaining. Thus, as we held in Story v. Robinson, it authorizes writs of habeas corpus ad testificandum directed to custodians. But it is a long leap from authority to compel custodians to produce witnesses necessary for the exercise of jurisdiction, to authority to compel non-custodians to bear the expense of that production simply because they have access to a deeper pocket.
Custody of state prisoners in the United States Courthouse presents a separate issue, because courthouse security is involved. A federal statute provides that "the United States Marshal of each district is the marshal of the district court . . . and may, in the discretion of the respective courts, be required to attend any session of court." 28 U.S.C. § 569(a). Addressing that provision, we observed in Story v. Robinson:
This statute quite clearly authorizes the district court to require the attendance of the Marshal at any judicial proceeding, including any proceeding at which the testimony of a state prisoner may be required. It seems plain, then, that the court has authority to direct the marshal to take prisoners into custody during the time their attendance at court is required.
689 F.2d at 1180. The Marshals Service acknowledges that this section of the statute makes the Marshals responsible for courthouse security, and that funds are appropriated for that purpose, but urges that absent some finding of a security risk, it should not be construed to relieve the state custodians of the duty to produce their prisoner witnesses and maintain them in secure custody.
We hold that section 569(a) authorizes federal trial courts,*fn4 in the interest of security and flexibility, to require the Marshal of the district to take custody of state prisoner witnesses while they are in a federal courthouse for purposes of testimony. Requiring an individualized "special security risk" finding would be completely impractical. The development of a record for such findings is beyond the capabilities of the litigants for whose lawsuits the prisoner witnesses are needed. Because of the critical need for courtroom security this case presents one of those rare instances in which the trial courts must be recognized as having virtually unreviewable discretion. Nor is the alternative of depending on state corrections officers for the maintenance of security within federal courthouses a viable alternative. Since 1789 the United States has depended upon its own marshals for the security of its own court facilities, and few if any federal judges would be comfortable with any other arrangement.
The provision of the order requiring that the Marshal of the District Court transport prisoners between the county jail and the federal courthouse, however, presents a distinctly different issue. In Story v. Robinson we reserved the question "whether section 569(a) provides authority for a direction, in the interest of courthouse security, that prisoner witnesses be taken into custody at a point outside the courthouse." 689 F.2d at 1180. We did not intend to suggest that federal trial courts had unlimited discretion to enter such orders. Obviously the security problem outside the courthouse is of an entirely different order of magnitude than that within. The presumption should be that there is no federal court security interest until state prisoners are delivered to the custody of the Marshal of the district at the federal courthouse. Situations obviously can be imagined in which that presumption may be overcome: particularly dangerous or threatening prisoners, for example, or perhaps prisoner witnesses whose own safety may be at risk. But those situations are likely to be the exception. Thus we hold that unless a federal trial court makes a specific finding that special security circumstances exist which require an order that state prisoner witnesses be taken into the Marshal's custody at a situs away from the federal courthouse, section 569(a) does not authorize such an order. There is no other authority for such an order.
The order appealed from will be affirmed insofar as it requires the Marshal of the District Court to take into custody while in the federal courthouse state prisoner witnesses whose presence is required for testimony in civil proceedings in the federal court, but reversed insofar as, without a finding of special security circumstances, it required the Marshal to transport those witnesses between the county jail and the federal courthouse.
BECKER, Circuit Judge, concurring in part*fn1 and concurring in the judgment.
The opinion announcing the judgment of the court proceeds along the following syllogism. A writ of habeas corpus ad testificandum requires the custodian to produce the prisoner at a trial. Since the magistrate in this case ordered federal marshals to deliver a prisoner of Pennsylvania to the federal courthouse, the order appealed from cannot be a writ of habeas corpus ad testificandum but must instead be a writ in the nature of mandamus directed at the Marshal Service. But, continues the syllogism, under McIntire v. Wood, 11 U.S. (7 Cranch) 504 (1813), and McClung v. Silliman, 19 U.S. (6 Wheat.) 598 (1821), the All Writs Act does not authorize federal court to issue writs of mandamus against federal executive branch, and since neither 28 U.S.C. § 569(a) nor 28 U.S.C. § 569(b) (nor any other arguable provision) provides an alternative source of authorization, the magistrate committed an error of law when he directed that the Marshals Service share responsibility for transporting state prisoners to federal court pursuant to a writ of habeas corpus ad testificandum.
With one exception,*fn2 I find each of the critical steps in this syllogism to be questionable. Moreover, having in mind the role of the Marshal as the executive arm of the court and the need to accord flexibility to trial judges if the judicial system is to function fairly, I believe that the consequence of the historic approach taken by Judge Gibbons' opinion is unfortunate. Ultimately, I cannot deny the analytic force of his opinion, and I am thus constrained to concur in the judgment. Given the institutional importance of today's decision and the circuit split it creates, however, I believe it important to probe the weaknesses of the syllogism implicitly relied upon, and to explain the problems of judicial administration likely to result from the failure to adapt the precedents of antiquity to the needs of a modern society, hence this separate opinion.
I. Analytic Problems
A. The Rigid View of Habeas Corpus Ad Testificandum
It is hornbook law that "when served with a writ of habeas corpus ad testificandum calling for the production as a witness of one lawfully restrained of his liberty, the sheriff, jailer, or other custodian of such person is bound to bring him into court to give his testimony. . . ." 97 C.J.S. Witnesses § 30c (1957). It is similarly the ancient learning that "A prisoner produced in response to a writ of habeas corpus ad testificandum is still in custody and the court or commissioner issuing the writ has no power to authorize the custodian to suffer his prisoner to go at large pending the proceedings in which he is called to testify, or to take him from the custody of one officer and put him in another." Id. There is no provision anywhere in traditional authority that permits modifications of this writ to allow custody to be transferred to federal officials before the courthouse is reached. Thus, this segment of Judge Gibbons' syllogism is probably correct.
The rigidity with which habeas corpus ad testificandum is viewed, however, rests in some tension with the flexible character of the All Writs Act, which has been held in three post-war cases to authorize the issuance of writs not precisely comporting than those of tradition. In Price v. Johnston, 334 U.S. 266, 282, 92 L. Ed. 1356, 68 S. Ct. 1049 (1948), Harris v. Nelson, 394 U.S. 286, 299, 22 L. Ed. 2d 281, 89 S. Ct. 1082 (1969), and United States v. New York Telephone Co., 434 U.S. 159, 171, 54 L. Ed. 2d 376, 98 S. Ct. 364 (1977), the Court has repeatedly interpreted the All Writs Act to be a "legislatively approved source of procedural instruments designed to achieve the rational ends of the law." Moreover, as the Supreme Court stated in Price, "since law' is not a static concept, but expands and develops as new problems arise, . . . the forms of the habeas corpus writ authorized by [section 1651] are not only those recognized in this country in 1789, when the original Judiciary Act containing the substance of this section came into existence. . . . Rather, it is a legislatively approved source of procedural instruments designed to achieve the rational ends of law.'" 334 U.S. at 282. This expansive language must give pause to any court seeking to circumscribe the discretion of a district court that seeks to fashion equitable procedures for assuring the presence of witnesses in proceedings before it.
More than the mere language of the cited cases might suggest the All Writs Act to provide a source for the directive issued by the magistrate in this case. In Price the issue was whether a federal appeals court had power to permit a state prisoner to come forward to argue his own appeal. Despite the absence of any traditional writ or express statutory provision authorizing release of a prisoner on these grounds, the Supreme Court held that section 1651 gave the appeals court such power. In Harris, a case much like this one in that the district court perceived a need for greater fact finding than other statutes would have allowed, the Supreme Court held that, despite the absence of specific statutory authority, the district court could issue a discovery order in connection with the habeas corpus proceeding before it. In New York Telephone, the Supreme Court affirmed an order of the district court directing a telephone operating company to assist the FBI in installing pen registers. After summarizing the case law noted above, the Supreme Court stated there that the All Writs Act empowers federal courts to issue orders to "persons who, though not parties to the original action or engaged in wrongdoing, are in a position to frustrate the implementation of a court order or the proper administration of justice . . . and even encompasses those who have not taken any affirmative action to hinder justice." 434 U.S. at 174.
I suppose, however, that there is a difference between the creation of new writs to satisfy new needs and the expansion or modification of existing writs. With the latter, the traditional forms and doctrines may be said to express the limits of the writ. To rephrase the matter in statutory parlance, 28 U.S.C. § 2241(c)(5) (1982), which gives federal courts authority to issue writs of habeas corpus ad testificandum, may preclude reliance on 28 U.S.C. § 1651 (1982) (the All Writs Act in modern garb) to create processes that mimic section 2241 writs in most of their features but which speak with their own voice on several critical issues. Thus, I cannot quarrel too much with a conclusion that the magistrate did not (and statutorily likely could not) issue a writ under 28 U.S.C. § 1651 directing the Marshals Service to share responsibility for transporting a state prisoner to federal court so that he might testify in civil proceedings. Instead, the magistrate issued a traditional writ of habeas corpus ad testificandum, and tagged on a writ, arguably in the nature of mandamus, directing the marshals to share responsibility for transport. The question then becomes, of course, whether the magistrate had authority to issue this supplemental writ.
B. The Ways of Denying Authority
The opinion announcing the judgment of the court relies on the hoary cases of McIntire v. Wood, 11 U.S. (7 Cranch) 504, 3 L. Ed. 420 (1813), and McClung v. Silliman, 19 U.S. (6 Wheat.) 598 (1821), to reach the conclusion that the All Writs Act, 28 U.S.C. § 1651 (1976), does not authorize a writ of mandamus against a federal executive officer. While I ultimately agree with the opinion of Judge Gibbons on this point, I am not without serious doubts.
Standing alone, I do not think the cryptic opinion in McIntire v. Wood could support the result reached today. In that case, a person claiming to own land by virtue of a federal statute brought suit in federal court in order that the court might direct a land officer to perform various acts apparently needed to perfect title (or to bring suit to perfect title). There being no "arising under" jurisdiction as there is in modern federal practice, the Supreme Court held that the 11th Section of the Judiciary Act of 1789 did not permit the suit to be brought in federal court. Generally put, that section vested the then-circuit courts with jurisdiction only over certain types of diversity and alienage cases, and cases in which the United States was the plaintiff or petitioner. Were we writing on a clean slate, we would say that the only comment McIntire made with respect to the All Writs Act was that it did not in and of itself create federal jurisdiction; though Congress might have granted jurisdiction to the federal courts in cases wherein performance by some federal officer was a precondition to enforcement of some federal entitlement, it had not done so in the All Writs Act.
It is only because of McClung v. Silliman and because of that case's interpretation of McIntire that I feel constrained to adhere to the judgment announced today by Judge Gibbons. In that case, the same litigant who had not been permitted to sue in federal court in McIntire sued in state court and asked it to issue a writ of mandamus to the recalcitrant federal land officer. The state court believed it had the power to issue such a writ, but, on the merits, declined to do so. Because of this factual background, it would seem that the issue before the Supreme Court was essentially one of federalism: whether the state court had the power to issue orders to a federal executive official.
For reasons that are lost to antiquity the Supreme Court in McClung determined whether a federal court could issue such a writ in cases (such as the one before it) where diversity of citizenship existed. Unaided by consultation of McClung, I would have thought that a federal court could issue such a writ under those circumstances; as I read McIntire, the problem was an absence of statutory jurisdiction, not some limitation on remedial powers. In McClung, however, the Supreme Court spoke to the contrary and stated (so far as I can discern) that, because there was no jurisdiction, a federal court of original jurisdiction could not issue a writ of mandamus against a federal official, even in cases where the parties were of diverse citizenship. Why there would be no jurisdiction in such a case is a mystery that perhaps only those deeply schooled in the mindset of that era could solve.*fn3
My bewilderment does not permit me to disregard what the Supreme Court has written, however. Thus, so long as I accept the assertion that the magistrate issued a writ of mandamus against the Marshals Service, and the notion that a marshal, like the land officer in McClung, is a member of the federal executive branch, then I must resign myself to the decision reached by Judge Gibbons.
C. The Status of the Marshals
Generally speaking, the United States Marshals serve as the executive arm of the federal judiciary and, since the Judiciary Act of 1789, have been obligated to execute all lawful "precepts" directed to them under the authority of the United States. As the Fifth and Eleventh Circuits have observed, "The raison d'etre of the Marshal Service is to service the federal forum in civil as well as criminal litigation." Ford v. Allen, 728 F.2d 1369, 1370 (11th Cir. 1984) (per curiam); Ballard v. Spradley, 557 F.2d 476, 481 (5th Cir. 1977). And from what I have observed during my years on the district bench, the Marshals Service performs numerous tasks for members of the judiciary. Finally, this court has declared that "in discharging his duties as a Marshal he acts concurrently as an officer of the federal judiciary and as an executive officer." United States v. Krapf, 285 F.2d 647, 649 (3d Cir. 1961).
I conclude, however, that these factors do not take the Marshals Service out of the executive branch of our government. The President appoints the United States Marshal for each judicial district, 28 U.S.C. § 561(a); and the Attorney General supervises and directs the United States marshals in their performance of public duties, 28 U.S.C. § 569(c) (1976). Thus, while there is plainly a special relationship between the Marshals Service and the federal judiciary, the marshals formally remain part of the executive branch of our government. While I would feel considerably more comfortable with this result, were I able to discern the reasoning of McIntire and McClung in their apparent restrictions on federal mandamus power and thus to consider whether those reasons justified a limit on mandamus against federal marshals, on balance I agree with Judge Gibbons' conclusion that federal marshals are the beneficiaries of whatever limits there may be under 28 U.S.C. § 1651 on the issuance of writs of mandamus against federal executive officials.
II. The Need For Flexibility
My acceptance of the rigid historicism of the opinion announcing the judgment of the court is also hindered by my belief that, from a perspective of judicial administration, the flexibility and creativity exhibited by the magistrate in mediating the vexing financial problems of Pennsylvania and the United States is precisely what we should be encouraging. If the dead hand of precedent did not cabin my freedom of action, I would applaud the actions of the magistrate here.
It is important to explain why today's holding may hamper the ability of the federal courts to find facts accurately, particularly in prisoners' rights suits under 42 U.S.C. § 1983, which, for obvious reasons, frequently require invocation of the otherwise unusual writ of habeas corpus ad testificandum. Cf. Harris v. Nelson, 394 U.S. 286, 22 L. Ed. 2d 281, 89 S. Ct. 1082 (1969) (expanding fact finding capacity of federal courts via section 1651 in habeas corpus proceeding brought by prisoner). Many such suits involve baseless claims and in many cases "ad testificandums" will be unnecessary. But where a district court or magistrate determines that the testimony of a state prisoner actually would assist a court or jury in its fact finding capacity, it should not be deterred from obtaining the prisoner because the state is unable to pay the full costs of witness production and because the similarly strapped Marshals Service is itself hard pressed to shoulder any such burden.
It may be said, of course, that the state assumes the costs of producing those it has chosen to incarcerate when federal authorities determine that temporary release of the prisoner is required in order to vindicate federal rights. This court has recognized, however, that because states have compelling interests in incarcerating dangerous persons, the federal courts should consider the costs of production on a custodian before issuing writs of habeas corpus ad testificandum. See Jerry v. Francisco, 632 F.2d 252 (3d Cir. 1980) (per curiam). Yet, given (1) the requirement of Jerry v. Francisco that courts consider the costs to the state before issuing an ad testificandum; (2) the principle that issuance of an ad testificandum writ lies in the discretion of the trial court; and (3) financial realities, we fear that, faced with a holding that the court cannot exercise its discretion to order the Marshals and the state to share the costs of transportation, there may be cases in which the district courts may be impelled to exercise their discretion to deny ad testificandum writs when a state is short of funds. Such denials will, at least in some instances, impair the vindication of civil rights.
Several methods exist whereby the concerns expressed in parts I and II of this opinion can be treated and the difficulties I help create today can be ameliorated. The first would be for me to change my mind. I will confess to have wavered many times in the course of reviewing this appeal. Ultimately, however, I cannot escape McClung, a decision of the Supreme Court, which no matter how antiquated, I am not free to overrule. The Supreme Court, by contrast, has far greater freedom and greater authority in interpreting its cases than do courts of appeal; it also may find that, because statutes can adapt to fit the needs of changing times, the All Writs Act now permits what, in the time of McIntire and McClung it did not.*fn4 With a split in the circuits now existing, there is now all the more reason for Supreme Court review in this area.*fn5 Finally, of course, legislative action may provide for greater flexibility than exists by virtue of our decision today. Congress may conclude that our interpretations of its prior commands has been erroneous or decide to issue new commands in light of changed circumstances. While I do not presume to tell Congress what result to reach, I do suggest that it is a problem, among many, worthy of its consideration.
ATKINS, District Judge, concurring in part and dissenting in part.
I concur in Part I and Part II.B. of Judge Gibbons' opinion. However, because I agree with the decisions of the Fifth and Eleventh Circuits in Ballard v. Spradley, 557 F.2d 476 (5th Cir. 1977) and Ford v. Allen, 728, F.2d 1369 (11th Cir. 1984) (per curiam), and with the Ninth Circuit in Wiggins v. County of Alameda, 717 F.2d 466 (9th Cir. 1983), cert. denied, 465 U.S. 1070, 104 S. Ct. 1425, 79 L. Ed. 2d 749 (1984), I respectfully dissent from the judgment.