944 F.2d 334
7th Cir.
1991
POSNER, Circuit Judge.

This is a diversity suit by a prisoner against the two federal public defenders who represented him unsuccessfully in his effort to fend off the revocation of his parole. The suit alleges that the defenders committed legal malpractice in violation of the common law of Illinois and seeks damages. The district judge dismissed the suit on the ground that public defenders have absolute immunity from liability for damages.

The litigants and the district judge apparently were unaware that shortly after this suit was filed Congress amended the Federal Tort Claims Act, 28 U.S.C. §§ 1346(b), 2671-2680, to make that Act the exclusive remedy for persons alleging common law wrongs by federal employees committed within the scope of their employment, whichever branch of the federal government they are employed by (formerly the Act was limited to employees of the executive branch). 28 U.S.C. § 2671. If federal public defenders are federal employees, the amendment (section 3 of the Federal Employees Liability Reform and Tort Compensation Act of 1988 (the “West-fall Act,” as it is known), Pub.L. 100-694, 102 Stat. 4563) has extinguished the plaintiffs claim. (Section 8(b) of the Act makes it retroactive. Arbour v. Jenkins, 903 F.2d 416, 420 (6th Cir.1990).) We therefore asked the parties to brief this question.

Literally, federal public defenders are federal employees. They are employed by the Administrative Office of the U.S. Courts. But it is uncertain whether they were intended to come within the scope of the amendment. Congress had already taken care of the problem of legal malpractice suits against federal public defenders by authorizing the Administrative Office to buy them liability insurance. 18 U.S.C. § 3006A(g)(3). This statute was not repealed by the Westfall Act, not explicitly anyway.

The purpose of the Federal Tort Claims Act is germane to the scope of the new amendment. That purpose (as the Westfall Act states explicitly, § 2(a)(2)) is to lift federal sovereign immunity in order that victims of torts committed by federal employees in the scope of their employment can sue the federal government under the doctrine of respondeat superior, just as if the government were a private employer enjoying no immunities. The theory (or perhaps one theory) underlying the doctrine of respondeat superior is that it gives the employer an incentive, perhaps greater than that created by liability simply for negligence, to use his control over the employee to prevent tortious misconduct. Konradi v. United States, 919 F.2d 1207, 1210-11 (7th Cir.1990). That theory falters when the employer is the federal government and the employee is a lawyer representing a person whom the federal government is prosecuting. A litigant cannot direct his adversary’s lawyer. Polk County v. Dodson, 454 U.S. 312, 321-22, 102 S.Ct. 445, 451-52, 70 L.Ed.2d 509 (1981). It would be odd to make the federal government answerable for the legal malpractice of federal public defenders, when the acts constituting malpractice are beyond the federal government’s power to control. One might think the public defender a purely nominal government employee and indeed an involuntary one (involuntary on the part of the government, that is). The government hires him not because it is in the business of defending the people it prosecutes but because the Sixth Amendment has been held to require it to furnish counsel to indigent criminal defendants. One way to discharge this constitutional duty is to hire public defenders; another is to reimburse appointed counsel. In neither case is the government meaningfully an employer responsible for the care with which the lawyer does his work. In relation to the federal government a federal defender is functionally an independent contractor rather than an employee, and a principal is not liable for the torts of his independent contractors. Anderson ¶. Marathon Petroleum Co., 801 F.2d 936, 938-39 (7th Cir.1986).

Given the severely attenuated and indeed adversary “employment” relationship it might seem quite enough that the Administrative Office has been authorized to pay the federal public defenders’ liability insurance premiums. The insurance carrier can monitor the defenders’ activities without being accused of interfering with an opposing lawyer’s handling of his case.

We have stated the arguments against applying the Westfall Act as forcefully as possible but still they are far from conclusive. If the Act applies, the government can still reduce its exposure to liability for the malpractice of federal defenders by exercising greater care in deciding whom to hire and whom to fire. And there is more than a hint of fiction in deeming the Administrative Office of the U.S. Courts an adversary of the defenders’ clients, as if the federal government were really a unitary entity or the judiciary’s administrative arm a branch of the Justice Department. The language of the Westfall Act, certainly, tugs in favor of classifying defenders as federal employees for tort claims purposes; and finally it would simplify administration of the Act to read it literally.

But if the case against applying the Act to public defenders is not conclusive, nei ther is the case for application. If the canons of statutory construction are to be trotted out, and the “plain meaning” canon invoked in favor of application, we must not overlook the canon against repeals by implication. If the Westfall Act applies to public defenders, the statute authorizing the Administrative Office to buy liability insurance for the defenders is nullified; they don’t need the insurance if they don’t have the liability. As so often, the canons are an unruly team, pulling in opposite directions and therefore requiring that the judicial decision, if it is to be candid, be placed on other grounds, such as what in all likelihood Congress would have provided had it foreseen the interpretive question that has arisen.

We need not bite the bullet and determine the Westfall Act’s coverage in this case, however. For the Act to kick in, the employee must notify the Attorney General of the suit; the Attorney General must certify that the defendant employees were acting within the scope of their employment when the tort occurred; and, if he refuses to so certify, the employees themselves must petition the district court. 28 U.S.C. §§ 2679(c), (d). No certification or petition has been filed in this case and, so far as we are aware, the defendants have not even notified the Attorney General of the suit. Given the elaborate statutory procedures for notification, certification, and petition, we do not think that in their absence either the district court or this court is obliged to determine on its own initiative whether the suit is against a federal employee acting within the scope of his employment. The suit proceeds as an ordinary tort action against an individual, as it would if, for example, the defendant, although unquestionably a federal employee within the meaning of the Westfall Act, had been acting outside the scope of his employment when he committed the tort.

We conclude that the Westfall Act does not bar this suit — not yet anyway, for doubtless on remand (and there must be a remand, as we shall see shortly) the district court will be asked whether the defendants can still invoke the Act’s protection or have waived it, a question complicated by the Act’s having come into existence after the suit was filed. But there is another preliminary issue that we cannot duck: whether the case is within the diversity jurisdiction. It is not, if the prisoner’s domicile is the place of his incarceration, because that is Illinois, the same state of which the defendants are citizens.

We agree with the four other circuits that have considered the question (we had not until this case) that since domicile is a voluntary status, a forcible change in a person’s state of residence does not alter his domicile; hence the domicile of the prisoner before he was imprisoned is presumed to remain his domicile while he is in prison. Stifel v. Hopkins, 477 F.2d 1116 (6th Cir.1973) (per curiam); Housand v. Heiman, 594 F.2d 923, 925 n. 5 (2d Cir.1979) (per curiam); Jones v. Hadican, 552 F.2d 249 (8th Cir.1977); Polakoff v. Henderson, 370 F.Supp. 690, 693 (N.D.Ga.1973), aff’d per curiam, 488 F.2d 977 (5th Cir.1974). The presumption is rebuttable — a prisoner might for example decide he wanted to live in another state when he was released and the federal prison authorities might therefore assign him to a prison in that state, and that would be the state of his domicile. The plaintiff in this case lived in Virginia before he was imprisoned, and, so far as appears, intends to return there when he is released. That is his domicile, and the suit is therefore within the diversity jurisdiction.

We move on to the question of immunity. In granting the federal public defenders absolute immunity, the district judge relied on a decision by this court which indeed so held, as a matter of federal common law. Robinson v. Bergstrom, 579 F.2d 401, 410-11 (7th Cir.1978) (per curiam). But this decision was in all likelihood overruled by Ferri v. Ackerman, 444 U.S. 193, 100 S.Ct. 402, 62 L.Ed.2d 355 (1979), a case that differs from the present one only in that the defendant there was appointed from practice rather than being employed by the government as a public defender. Although some language in the Court’s opinion could be used to argue for a distinction between a private lawyer and a public defender, the defendants in our case have confessed error on the district judge by conceding that the only significance of Robinson v. Bergstrom is that it might be predictive of whether they have absolute immunity under Illinois state law, a possibility left open in Ferri: “when state law creates a cause of action, the State is free to define the defenses to that claim, including the defense of immunity.” Id. at 198, 100 S.Ct. at 406.

There are no cases in Illinois on the question but we expect that the Supreme Court of Illinois would agree with the basic reasoning of Ferri, which is that the differences between court-appointed counsel, on the one hand, and retained counsel, on the other, are too slight to justify granting absolute immunity to the one while exposing the other to malpractice liability to his client with no immunity at all. Moreover, Illinois statutes provide for indemnification of public defenders (state, of course, not federal as here), Ill.Rev.Stat. ch. 34, U 5-1003; ch. 127, H 1302, implying a judgment that adequate protection from legal harassment can be provided to them without granting them absolute immunity. Even if there were no provision for indemnifying federal public defenders, we think it likely that Illinois would take the position that there ought to be and that it is the business of the federal government to provide it, pending which reform the state will allow victims of malpractice by federal defenders to sue them. But as noted earlier the Criminal Justice Act contains a provision authorizing the Administrative Office of the U.S. Courts in effect to indemnify federal public defenders. We hold that they do not, at least as a matter of Illinois law, enjoy absolute immunity from being sued for malpractice.

The defenders have another string to their bow, however. They argue that Sullivan is collaterally estopped to complain of their alleged malpractice because in two habeas corpus proceedings in which he challenged the revocation of his parole the courts resolved all but two of the malpractice issues by ruling adversely to Sullivan on claims that he argues his defenders failed to raise. That’s fine, but it leaves the two issues: first, whether the Parole Commission incorrectly calculated the unexpired portion of his District of Columbia sentence and the defenders failed to point this out at the revocation hearing, and second whether they also failed to point out to the Commission that it should make that unexpired portion run concurrently with the sentence that Sullivan is now serving in Illinois.

The decision of the district court must be vacated and the case remanded to consider the merits of these two claims of malpractice (as well as the possibility that the defendants can still invoke the Westfall Act—but all that that will do, since the case is already in the federal courts, will be to cause a substitution of the United States as the party defendant for the two defenders; it will not abate the suit. 28 U.S.C. §§ 2679(d)(1), (2)). For the guidance of the court and parties on remand we point out that Sullivan is correct in arguing that the Parole Commission, in deciding the issues relating to the D.C. sentence, was required to apply the parole laws and guidelines of the District of Columbia Code. We so held in Johnson v. Williford, 821 F.2d 1279, 1282-88 (7th Cir.1987) (as have a number of other courts, illustrated by Brewer v. Swinson, 837 F.2d 802, 805 (8th Cir.1988), and Walker v. Luther, 830 F.2d 1208 (2d Cir.1987), both of which cite Johnson with approval), and mention the point only to note an arguable but we think spurious conflict with Pryor v. Brennan, 914 F.2d 921 (7th Cir.1990), which did not cite Johnson. Pryor, in common with other decisions, such as Moss v. Clark, 886 F.2d 686 (4th Cir.1989), holds that violators of the District of Columbia Code who are incarcerated in federal prisons earn good-time credits pursuant to federal law rather than the D.C.Code. Both lines of decision are correct. The provision of the Code relating to good-time credits is expressly limited to offenders housed within the District of Columbia, D.C.Code § 24-428(a), while the provisions relating to parole violations expressly grant the Parole Commission “the same power and authority over prisoners convicted in the District of Columbia ... as is vested in the District Board of Parole over prisoners confined in the penal institutions of the District of Columbia.” § 24-209.

Affirmed in Part, Vacated in Part, and Remanded.