Donald R. Rollins was insured for $100,000 under a life insurance policy governed by the Federal Employees Group Life Insurance Act (FEGLIA). The insured failed, however, to designate a beneficiary. Upon his death, both his widowed second wife and his children by a previous marriage claimed the insurance proceeds. The second wife argued she was entitled to the proceeds under the terms of FEGLIA; the children contended they were entitled to the proceeds, either under the provisions of the statute or, alternatively, under the doctrine of constructive trust. The district court granted the second wife's motion for summary judgment, ruling that: 1) she was the beneficiary under the statutory scheme of FEGLIA, and 2) federal law preempted the imposition of a constructive trust upon FEGLIA insurance proceeds. The children appeal. We reverse and remand to the district court for a determination of the constructive trust claim under the laws of the state of Indiana.1I. FACTUAL BACKGROUND

Donald R. Rollins, the insured, married Kellen S. Rollins in 1968. Three children were born of that marriage: Donald R. Rollins, now approximately 16 years of age; Douglas R. Rollins, now approximately 15 years of age; and Misty D. Rollins, now approximately 13 years of age. During that marriage, the insured was privately employed. The marriage was dissolved in 1981. In its final decree granting the divorce, the Indiana Circuit Court provided, inter alia: "Court now orders the Petitioner [Donald Rollins] to maintain life insurance upon his life with the children named as beneficiaries thereon and further maintain life insurance policies on the lives [sic] of each child." In addition, the insured was ordered to pay child support. The insured, however, was not always faithful in making his child support payments and died leaving a deficiency.

In 1982 the insured married defendant Nancy J. Rollins. No children were born of this marriage. The defendant had a child by a previous marriage, but the insured did not adopt that child. On April 28, 1986, the defendant filed a petition for dissolution of the marriage. The petition was still pending on May 14, 1986, the date of the insured's death.

During his second marriage the insured became and remained a civilian employee of the federal government. As a federal employee, he was covered by a life insurance policy with a face amount of $100,000 issued by the defendant Metropolitan Life Insurance Company pursuant to the Federal Employees Group Life Insurance Act, 5 U.S.C. Secs. 8701-8716 (1982).2 FEGLIA allows the insured employee to designate the beneficiary of his insurance policy. In the absence of a designated beneficiary, FEGLIA provides that the insurance proceeds will be paid to the employee's spouse. If there is no surviving spouse, then the proceeds go to the employee's children.3 During the term of the policy at issue, the insured failed to designate a beneficiary. The defendant now claims that as the insured's widow she is entitled to the insurance proceeds.

The defendant and the insured had separated in April 1986 due to their marital difficulties and were living apart. In her verified petition to dissolve her marriage to the insured, the defendant alleged that there had "been an irretrievable breakdown of the marriage," and therefore she sought its dissolution and the restoration of the name she had used prior to her marriage. At the same time, she filed a verified motion for a temporary restraining order against the insured to prevent him from harassing or molesting her. As a basis for the motion, she alleged that the insured had "in the past caused or threatened her with bodily injury," putting her in fear for her personal safety. The restraining order was issued and was to remain in effect until the final hearing. The final hearing was set for July 14, 1986.

The insured did not contribute to the defendant's support during the separation. The couple had not yet entered into a written property settlement agreement, but had reached an oral agreement as to the disposition of all of their joint property, except for one gun, a reclining chair, and an end table. According to the defendant, the insured, notwithstanding the restraining order, visited her either very late on May 12 or very early on May 13, 1986. The insured was upset; the defendant and the insured had a quarrel. Later that same day, following the quarrel, the defendant discovered the insured's body at his trailer. The insured's death was ruled a suicide.

In Count I, the plaintiffs claim that the marriage between the defendant and the insured was over as a practical matter, with nothing but the formality of a hearing to technically dissolve it. The defendant claims that in spite of the bad state of the marriage, she and the insured were still married and were friends. The plaintiffs argue that the defendant was not the insured's widow under FEGLIA, but only his "friend," so that the defendant should not be eligible to receive the insurance proceeds in the absence of a specific beneficiary designation.

The defendant has a very simple view of this case. Her late husband was insured, he died, and as his widow, in the absence of some other designated beneficiary, she maintains that she is entitled to the insurance proceeds as a matter of law.

FEGLIA provides a rather simple priority of beneficiaries, without any reference to the common marital difficulties that often complicate real life, as in this case. We will not undertake the likely impossible task of trying to rewrite the statute to fairly provide for all eventualities. We agree that in spite of the total collapse of the marriage the defendant technically qualifies as the widow of the insured within the meaning of the statute.

The statute does not immediately solve this case however. The plaintiffs also raised the issue of constructive trust, which the district court failed to fully address. Therefore, even though the defendant qualifies as the insured's widow, we cannot simply conclude the matter by mechanically applying the statute.

In Count II of their complaint, the plaintiffs allege that the state court, in the dissolution of the insured's first marriage, imposed on the insured a fiduciary duty to maintain life insurance on his life for the benefit of his children, the plaintiffs. The plaintiffs contend that the insured's failure to name his children as beneficiaries of his government life insurance policy constituted a breach of trust and permitted the proceeds to go to the defendant, unjustly enriching her. The plaintiffs further allege that they are without a legal remedy to rightfully acquire the insurance proceeds for their support and education. Thus, the plaintiffs seek the imposition of a constructive trust.

The district court summarily rejected the plaintiffs' plea for the imposition of a constructive trust, on the theory that federal law preempted the application of equitable state law principles. For the reasons given below, we reverse the district court's decision on the issue of preemption and remand this case for consideration of the constructive trust claim on the merits.

Before addressing plaintiffs' constructive trust claim we note that there is some question about whether state or federal law governs this issue. The parties' briefs never clearly articulated which body of law determines the constructive trust issue. The district court appeared to believe that the issue was one of state law.

Federal courts must often interpret and apply federal statutes, frequently having to fashion rules of application for situations not envisioned by Congress and therefore not expressly provided for in a statute. When faced with these circumstances, a court must decide whether to fashion a uniform federal rule, or refer to state law for a rule of decision:

Controversies directly affecting the operations of federal programs, although governed by federal law, do not inevitably require resort to uniform federal rules.... Whether to adopt state law or to fashion a nationwide federal rule is a matter of judicial policy "dependent upon a variety of considerations always relevant to the nature of the specific governmental interests and to the effects upon them of applying state law." United States v. Standard Oil Co., 332 U.S. 301, 310 [67 S. Ct. 1604, 1609, 91 L. Ed. 2067] (1947).

Undoubtedly, federal programs that "by their nature are and must be uniform in character throughout the Nation" necessitate formulation of controlling federal rules. United State v. Yazell, 382 U.S. 341, 354 [86 S. Ct. 500, 507, 15 L. Ed. 2d 404] (1966).... Conversely, when there is little need for a nationally uniform body of law, state law may be incorporated as the federal rule of decision.

United States v. Kimbell Foods, Inc., 440 U.S. 715, 727-28, 99 S. Ct. 1448, 1458, 59 L. Ed. 2d 711 (1979) (some citations omitted). Application of these guiding principles leads us to hold that state law will provide the rule of decision in this case, as we see no need for national uniformity on the question of constructive trust under the circumstances of this case.

The doctrine of constructive trust is recognized under federal law. See Snepp v. United States, 444 U.S. 507, 515, 100 S. Ct. 763, 768, 62 L. Ed. 2d 704 (1980). Its uniform application in situations such as the present one could therefore be imposed. However, we think there is no need for national uniformity in this specific fact scenario, which touches so profoundly upon the law of domestic relations. By long tradition, family law is almost exclusively the domain of the individual states; there is, for all practical purposes, no federal domestic relations law. These facts militate in favor of looking to state law rather than fashioning a uniform federal law in the present situation. And Indiana law is the best source for a rule of decision in this case.

Indiana clearly has an interest in the resolution of this case. Donald and Kellen were divorced under the laws of Indiana; the divorce decree ordering him to maintain life insurance on himself for the benefit of the children emanated from an Indiana court. Indiana has a continuing interest in the maintenance of the children because they still reside in the state. Significant state policy interests thus favor the adoption of Indiana law on the issue of constructive trust, at least in the present circumstances.

We note that there are no federal interests compromised by our decision not to impose a uniform federal rule on the issue of constructive trust and FEGLIA insurance proceeds. As discussed more fully below, the FEGLIA was, from its inception, meant to place federal civilian employees on a par with private-sector workers, in terms of the benefits of group life insurance. We find no indication that Congress, in setting up such a program, intended government employees to be free of any legal demands that might be made upon their private counterparts. We therefore see no harm, and much good, in applying Indiana law to this question.

When the district court rejected the plaintiffs' constructive trust claim, it relied primarily upon Ridgway v. Ridgway, 454 U.S. 46, 102 S. Ct. 49, 70 L. Ed. 2d 39 (1981) and Metropolitan Life Ins. Co. v. McShan, 577 F. Supp. 165 (N.D.Cal.1983). Ridgway involved a fact pattern similar to the one before us now, with the significant exception that the Supreme Court was interpreting the Servicemen's Group Life Insurance Act of 1965 (SGLIA) rather than FEGLIA. When the insured, Sgt. Ridgway, divorced his first wife, a state court ordered the sergeant to keep his life insurance policy, or a replacement policy, in force for his three children who were then in the custody of his former wife. His former wife was named the beneficiary of the policy at that time. After a short time, however, the sergeant remarried and changed the beneficiary of his life insurance policy from his first wife to his second wife. The sergeant died a few months later. The first and second wives litigated the issue of entitlement to the insurance proceeds in state court. The Supreme Court of Maine imposed a constructive trust upon the proceeds on behalf of the children, but the United States Supreme Court reversed.

Ridgway detailed the statutory background of the sergeant's policy issued under SGLIA. In enacting SGLIA, Congress intended to make life insurance coverage available to active duty military personnel, particularly in combat zones during the Viet Nam war. Because of the war, commercial insurance offered only restricted coverage for service members.4 Ridgway, 454 U.S. at 50, 102 S.Ct. at 52. SGLIA established an order for policy beneficiaries. In the event that the service person did not designate a beneficiary, an unrestricted right, the proceeds would go to the surviving spouse, and if none, to the children. SGLIA also contained an anti-attachment provision that shielded the insurance proceeds from taxation and claims of creditors. Id. at 52, 102 S.Ct. at 53.

The majority in Ridgway found that the obvious purpose of SGLIA was to provide coverage to service personnel, no matter how hazardous the duty, and to protect their designated beneficiaries. Id. at 53, 102 S.Ct. at 54. As explained in Wissner v. Wissner, 338 U.S. 655, 70 S. Ct. 398, 94 L. Ed. 424 (1950), which the Ridgway majority found to be controlling, the possession of government insurance payable to the beneficiary of the service person's choice was designed to directly enhance morale, a purpose within congressional power pertaining to national defense. The Ridgway majority found a conflict between these federal interests and the state's interest in the welfare of the sergeant's children; of course, the federal law prevailed. Ridgway, 454 U.S. at 55-57, 102 S.Ct. at 55-56. The Ridgway majority also found that a constructive trust would be inappropriate because the complaint itself contained no allegations to support the constructive trust theory. Id. at 58-59, 102 S.Ct. at 56-57. In addition, the Ridgway majority noted that SGLIA's anti-attachment provision was an independent ground for rejecting the constructive trust argument. Id. at 60-61, 102 S.Ct. at 57-58.

The defendant suggests that imposition of a constructive trust in this case would require overruling Ridgway, but the defendant exaggerates the influence of Ridgway upon this case. The similarities between Ridgway and the present case are obvious, but there are significant distinctions. The insured in the present case was a civilian employee of the federal government, not a member of the armed forces. Therefore, there are no federal interests regarding military morale and national defense involved.

Further, the insurance program of the insured is under FEGLIA, not SGLIA. FEGLIA has the character of private insurance, in contrast with the military policy of SGLIA discussed in Ridgway. Where SGLIA is intended to bolster the morale of fighting forces, the impetus of FEGLIA is to provide low cost life insurance protection for the federal civilian employee's family. In 1954, when President Eisenhower proposed this insurance program to Congress, he likened it to the group insurance programs provided to employees by private employers, with the government substituting for the private employer. 1954 U.S.Code Cong. & Admin.News 3052, 3056.

FEGLIA is therefore best viewed as the public service analog to private group insurance, made available to federal government employees so that they may adequately provide for their families in case of death. Thus, for example, FEGLIA contains no anti-attachment provision, in contrast to SGLIA which does contain such a provision. Given the distinction between the purposes of FEGLIA and SGLIA, the absence of any anti-attachment provision in FEGLIA is understandable. And, just as the Court in Ridgway found the inclusion of an anti-attachment provision in SGLIA significant to its decision not to permit the imposition of a constructive trust under that statute, Ridgway, 454 U.S. at 60-61, 102 S. Ct. at 57-58, we find the omission of such a provision in FEGLIA supports our conclusion that nothing in FEGLIA dictates preemption of equitable state law principles. There is no reason to believe that Congress, when it passed FEGLIA, intended federal employees to enjoy protections against judicial decrees not enjoyed by holders of private group life insurance policies.

Finally, another crucial distinction is that the complaint in Ridgway failed to adequately allege a constructive trust theory. The present complaint, however, adequately puts the constructive trust theory in issue. The constructive trust issue was considered by the district court without the allegations of the complaint being challenged or found to be insufficient for that purpose.

The differences in personal circumstances between Sgt. Ridgway and the insured in this case are also striking. In the present case, the defendant had sued the insured for divorce and was seeking to enforce a restraining order to keep the insured from disturbing her. A late night quarrel between them preceded his suicide by only a few hours. The insured did not designate her as his beneficiary, unlike the situation in Ridgway. Instead, the defendant here seeks to take by default under the statute. Thus, Ridgway is readily distinguishable from the present case, on both law and facts, and thus is not controlling nor dispositive of the present case.5

In addition to Ridgway, the district court relied upon Metropolitan Life Ins. Co. v. McShan, 577 F. Supp. 165 (N.D.Cal.1983), in rejecting the plaintiffs' constructive trust argument. In McShan, the court held that a constructive trust was precluded by FEGLIA's provision pertaining to priority of beneficiaries. The court held that a constructive trust was precluded because SGLIA and FEGLIA have the same basic structure and therefore, as in Ridgway, the supremacy clause mandates that the federal scheme must prevail. McShan, 577 F. Supp. at 166-67. The court also found that the absence of an anti-attachment provision in FEGLIA was not significant in comparing the two insurance programs. Id. at 169. Our understanding of Ridgway forces us to disagree with the McShan court's interpretation of FEGLIA and the availability of the constructive trust remedy.

The McShan court believed that FEGLIA's provision allowing the insured to designate beneficiaries would be meaningless if a constructive trust could interfere. We do not view the invocation of equitable remedies in that light. In the absence of equitable intervention the statutory provisions establishing the priority of beneficiaries would continue to control. Thus they would not be rendered meaningless.

Further, there is nothing about the right to designate a beneficiary under FEGLIA that is uniquely federal, or bars the application of equitable state law principles. As Justice Stevens commented in dissent in Ridgway:

The right to designate the beneficiary of an insurance policy is a common feature in insurance contracts. It surely is not a right that can be characterized as uniquely federal in any sense. Moreover, the mere fact that the right has its source in a federal statute does not require that it be given a construction different from that given a comparable right created by state law or by private contract.

Ridgway, 454 U.S. at 79-80, 102 S.Ct. at 67 (Stevens, J., dissenting).

In addition, Justice Stevens noted that unless the application of the "well-established equitable doctrine [of constructive trust] does 'major damage' to 'clear and substantial federal interests,' we must respect it." 454 U.S. at 72, 102 S.Ct. at 63 (Stevens, J., dissenting) (footnote omitted) (citing United State v. Yazell, 382 U.S. 341, 352, 86 S. Ct. 500, 507, 15 L. Ed. 2d 404 (1966)).

Although the Ridgway majority did find a conflict between the provisions of SGLIA and Maine's law on constructive trusts, and therefore overturned the decision of the Supreme Court of Maine, we do not see such a conflict arising between FEGLIA and state law on constructive trusts. As noted earlier, SGLIA addresses important federal interests--national defense and the morale of service personnel--interests not at issue with FEGLIA. Thus while the Ridgway majority held that SGLIA's provisions preempted the application of equitable state law principles, we are persuaded by the differences between the two statutes that imposition of a constructive trust upon FEGLIA insurance proceeds would not do "major damage" to any "clear and substantial federal interests." Therefore, unlike the court in McShan, we hold that nothing in federal law conflicts with the imposition of a constructive trust upon these insurance proceeds if the district court finds that such relief is proper under Indiana law.

In addition to Ridgway and McShan, the defendant relies upon Knowles v. Metropolitan Life Ins. Co., 514 F. Supp. 515 (N.D.Ga.1981) and Huff v. Metropolitan Life Ins. Co., 675 F.2d 119 (6th Cir.1982). Neither of these cases influences our decision in this case. Knowles involved a contest between a widow and a first wife over FEGLIA insurance proceeds where the first wife held the supposed benefit of a divorce decree requiring the insured to maintain her as a life insurance beneficiary. The insured had instead designated his second wife, and now widow, as the beneficiary of the policy. The Knowles court held in favor of the widow, basing its decision upon the regulations governing the designation of beneficiaries which provide:

Sec. 870.902 Designation of beneficiary.

....

(d) Any person, firm, corporation, or legal entity ... may be named as beneficiary.

(e) A change of beneficiary may be made at any time and without the knowledge or consent of the previous beneficiary, and this right cannot be waived or restricted.

5 C.F.R. Sec. 870.902 (1988). By its terms, however, this provision applies only to the express designation of beneficiaries, not to designation by operation of the statutory default rule. Nothing in the purpose or language of FEGLIA argues for expanding the scope of this provision to protect beneficiaries designated by default rather than by the express choice of the policyholder. Knowles is therefore distinguishable from the present case where we are dealing with an insured who failed to designate any beneficiary. In the current situation, neither the statute nor the regulation imposes an anti-attachment provision. Furthermore, the Knowles court never addressed the issue of constructive trust which is the true issue in this appeal.

Nor does Huff influence our decision, due to key factual differences between it and the instant case. Huff involved a dispute between the decedent's first and second wives. The second wife argued, as does the defendant in the present case, that she was entitled to the proceeds as the insured's widow. The first wife contended that the decedent intended the insurance proceeds for the college education of his son by his first marriage. There was no court order that the insurance proceeds be so used, nor was there any pending divorce from the second wife. The issue of a constructive trust was not raised or decided. We therefore create no conflict by our holding in this case that a constructive trust may be imposed on FEGLIA insurance proceeds in these circumstances.

Cummings v. Briggs & Stratton Retirement Plan, 797 F.2d 383 (7th Cir.), cert. denied, 479 U.S. 1008, 107 S. Ct. 648, 93 L. Ed. 2d 703 (1986), foreshadowed our view of the present case. In Cummings we suggested that a constructive trust could be the proper remedy when a designated beneficiary of an insurance policy was other than the one contemplated by the divorce decree. Id. at 390 n. 5.6 Chief Judge Reynold's explanation in his district court opinion of the appropriateness of a constructive trust, though found by this court to be inapplicable to the pension benefit plan there at issue, is nevertheless applicable to the present case.7

There is, in fact, abundant authority (though not, it seems, in Indiana) for treating an insurance beneficiary as a constructive trustee for a prior beneficiary whose entitlement rested on a duty owed by the policyholder. See, e.g., Noll v. Brende, 318 N.W.2d 319 (S.D.1982); Hudspeth v. Stoker, 644 S.W.2d 92 (Tex.App.1982). Furthermore, a constructive trust may be invoked even where the unjustly enriched party is completely blameless. As Justice Stevens noted in his Ridgway dissent, the law of constructive trusts remains applicable in such situations:

The imposition of a constructive trust on these facts is common in the law, and has been recognized in cases in which no wrongdoing could be imputed to the designated beneficiary. Simonds v. Simonds, 45 N.Y.2d 233, 380 N.E.2d 189 [408 N.Y.S.2d 359] (1978); McKissick v. McKissick, 93 Nev. 139, 560 P.2d 1366 (1977); Richards v. Richards, 58 Wis. 2d 290, 206 N.W.2d 134 (1973); see also G. Bogert, Trusts and Trustees Sec. 475 (rev. 2d ed. 1978). As stated in Simonds v. Simonds, supra, at 242, 380 N.E.2d, at 194 [408 N.Y.S.2d at 364] (citations omitted):

"Unjust enrichment, however, does not require the performance of any wrongful act by the one enriched. Innocent parties may frequently be unjustly enriched. What is required, generally, is that a party hold property 'under such circumstances that in equity and good conscience he ought not to retain it.' A bona fide purchaser of property upon which a constructive trust would otherwise be imposed takes free of the constructive trust, but a gratuitous donee, however innocent, does not."

Ridgway, 454 U.S. at 71-72 n. 1, 102 S. Ct. at 63 n. 1 (Stevens, J., dissenting). The fact that the defendant in the present case would be unjustly enriched by the default provisions of the statute rather than by any wrongdoing on her part is therefore no bar to the plaintiffs' constructive trust claim.

And yet there remain circumstances which require additional consideration. The court order requiring the insured to maintain life insurance on himself with his children named as beneficiaries is not as carefully drawn as might be desired. It seems clear, however, that under the decree the children were intended to have the benefit of their father's insurance--regardless of the specific company, policy or amount. The decree seems adequate to help support the plaintiffs' constructive trust theory. The fact that the policy was not actually in place at the time of the marriage dissolution order, but rather acquired a number of years later, is also to be considered.

Some state courts have suggested that difficulties of this nature need not forestall the recognition of a constructive trust. See, e.g., Simonds v. Simonds, 45 N.Y.2d 233, 380 N.E.2d 189, 408 N.Y.S.2d 359 (1978) (first wife entitled to part of proceeds of policies substituted for policy required to be maintained under divorce decree); cf. McKissick v. McKissick, 93 Nev. 139, 560 P.2d 1366 (1977) (divorce decree's mistaken reference to a nonexistent insurance policy leads to imposition of constructive trust involving three existing policies). Other state courts, however, have adopted a more cautious approach in other circumstances. See, e.g., Rindels v. Prudential Life Ins. Co., 83 N.M. 181, 489 P.2d 1179 (1971) (no recovery for former wife and family of first marriage where policy named in divorce decree lapsed and later policy named second wife as beneficiary); Lock v. Lock, 8 Ariz.App. 138, 444 P.2d 163 (1968) (same); see also Carpenter v. Carpenter, 150 Ariz. 130, 722 P.2d 298 (App.1985) (surveying case law on equitable interests in after-acquired insurance policies), vacated in part on other grounds, 150 Ariz. 62, 722 P.2d 230 (1986).

In view of the paucity of Indiana caselaw on constructive trusts, we must remand this case to the district court for a determination of whether these insurance proceeds are subject to a constructive trust in favor of the children under Indiana law. We reiterate, however, that if a constructive trust exists under Indiana law in this case, nothing in federal law preempts an appropriate grant of relief.

In Ridgway, the majority referred to its case as "unpalatable," Ridgway, 454 U.S. at 62, 102 S.Ct. at 59, no doubt because concerns unique to that particular service personnel insurance program compelled the Court to divert proceeds from the children. This particular case involves a different insurance statute and the doctrine of federal preemption, which dictated the majority's holding in Ridgway, presents no bar to equitable relief in this factual situation.

Therefore, while we agree with the district court's holding that the defendant is the insured's widow within the meaning of FEGLIA's beneficiary-by-default provisions, that does not end the case; the district court must yet consider the merits of the plaintiffs' constructive trust claim. We therefore reverse the district court's grant of summary judgment, and remand this case to the district court for consideration of the plaintiffs' constructive trust claim under Indiana law.

REVERSED AND REMANDED for further proceedings in keeping with this opinion. Circuit Rule 36 shall not apply. Each party shall bear their respective costs.