OAKES, Circuit Judge.
Tlie Social Security system provides primary benefits, based on one’s own work, to retired workers who meet specified conditions. It also provides secondary benefits to the surviving dependents, often spouses, of deceased workers. The spouse benefit system, which is at the heart of this case, has been said to involve “essential arbitrariness,” in that the provisions which make it up “incorporate no coherent rationale for entitlement.”1 One set of principles determines contribution and entitlement to primary benefits, while a completely different set determines entitlement to' spousal benefits, making, in the words of one commentator, “serious arbitrariness ... inescapable.”2
The arbitrariness that is implicated in this case involves the determination of which of successive surviving spouses (here widows) is entitled to secondary benefits based on the work record of the deceased. The basic test looks to the marital law of the state of the covered worker’s domicile at the time of his death to ascertain which claimant is the legal widow, and thus entitled to benefits. Unfortunately, this approach does not fully promote either uniformity or nonarbitrariness. Some states, for example, recognize de facto marriages,3 while others do not, a situation which would be likely to result in substantial injustice were entitlement to turn entirely on state law resolution of the question. In 1960 Congress responded to this problem by establishing a federal statutory test to provide benefits to persons who married in good faith unaware of legal impediments, such as a prior, undissolved marriage, to their marriage’s validity.4 [1069]*1069This statute, which grants entitlements to such “good faith” spouses, also contains a proviso that no “legal” spouse qualify for benefits.5 Thus, to sketch the basic statutory scheme, a second wife, such as appellant here, unaware that her husband was not divorced from his first wife, would be entitled to a widow’s benefit upon her husband’s death only if his first wife did not lay claim or was not entitled to those benefits.
This already complex portrait is further complicated when the surviving spouse is herself also entitled to primary, i.e., retired workers’, benefits. If those primary benefits are greater than the spousal benefits, the statute provides that there is no entitlement to the secondary benefits.6 In such a case, a second good faith or “deemed” wife is therefore entitled to surviving spouse’s benefits. The truly hard case — such as the one here before us — involves the situation where the secondary spousal benefits to which the legal widow is entitled are greater than the primary benefits which she may claim. Does the entitlement of the first wife to even a minute portion of the available surviving spouse’s benefits disentitle the second good faith spouse, or so-called deemed widow, to any and all such benefits?
Faced with that problem, this court in Rosenberg v. Richardson, 538 F.2d 487 (2d Cir.1976), construed the operable statute to mean that the “second Mrs. Rosenberg” (whose husband had previously obtained a Mexican divorce not recognized by New York law) was nevertheless entitled to her deemed widow’s benefits. We also held in Rosenberg that the deemed widow’s secondary benefits would be reduced by the amount payable to the legal widow, and that in no event could the total payments to widows exceed what could be paid to a single widow.
[1070]*1070Rosenberg has been criticized by other courts,7 has been said by a commentator to reach an “insupportable conclusion,”8 and we are now asked by the Secretary to overrule it.9 For reasons that we state below we decline to do so. Indeed, we believe the result reached in Rosenberg to be not only Solomonic but practical and we also believe its reasoning supportable by reference to the statutory structure and language, which, though by no means crystal clear, nevertheless evince a purpose served by our decision.
Facts
Our case is similar to but somewhat more complicated factually than Rosenberg. It is complicated further by an administrative muddle that is somewhat analogous to the muddled marital life that Samuel Capitano lived.
Sam Capitano evidently married Betty, nee Anna Jokubynas, on December 2, 1931, in Scranton, Pennsylvania. We say “evidently” because the name on her marriage certificate was “Petronella Jackaben.” Betty furnished, however, an explanation for the name discrepancies that satisfied one of the administrative law judges who dealt with this matter in its rather lengthy traverse through the administrative process, so that it was “reasonable to infer” that it was Betty Capitano who actually married Sam on that day.
Sarah Sparacio married Sam Capitano in Pittston, Pennsylvania, on April 7, 1951. Whether that marriage was “in good faith,” as she claims, or not, as the ALJ may have found (contrary to the Appeals Council’s subsequent decision), is an issue both of fact and of law which we discuss below. For the moment suffice it to say that Sarah was known as Mrs. Capitano, was called “wife” by Sam (who called himself “your loving husband Cap”), paid for Sam’s funeral, was “Wife & Beneficiary” of his life insurance, annually remembered the date of his death with rather moving newspaper notices, and bought a burial lot for the two of them. After Sam’s death in January, 1970, Sarah quite naturally applied for and received widow’s Social Security benefits. These benefits commenced in July, 1975.
Betty maintained during the administrative proceedings that she “never got a divorce and never even applied for a divorce” because she “did not want to give [Sam] the satisfaction of granting him a divorce so that he would be free to remarry.” She also claimed that from the time Sam left her in the 1940s until after he died she was “unaware of his whereabouts.” Betty claimed that she concluded ultimately that he must have moved to Buffalo, New York, where he had three sisters. Betty, whose account of her life with and without Sam occasionally stretches one’s imagination and patience, nevertheless also stated that one of Sam’s sisters with whom he was living had phoned her and told her that he was dying, that “I immediately went to Buffalo to the hospital to visit him” and that he “died while I was there and I attended the funeral.” Somewhat incredibly, Betty also stated in writing that “the name Sarah never came up while I was at the funeral, I never heard of this ‘Sarah.’ ”
Be that as it may, on December 7, 1977, Betty applied on the basis of her own earnings for primary retirement insurance benefits. Reluctantly, because she “did not [1071]*1071like the idea of going on” Sam’s work record, she also, some three weeks later, applied for widow’s benefits, apparently at the behest of the Social Security people.
Free access — add to your briefcase to read the full text and ask questions with AI
OAKES, Circuit Judge.
Tlie Social Security system provides primary benefits, based on one’s own work, to retired workers who meet specified conditions. It also provides secondary benefits to the surviving dependents, often spouses, of deceased workers. The spouse benefit system, which is at the heart of this case, has been said to involve “essential arbitrariness,” in that the provisions which make it up “incorporate no coherent rationale for entitlement.”1 One set of principles determines contribution and entitlement to primary benefits, while a completely different set determines entitlement to' spousal benefits, making, in the words of one commentator, “serious arbitrariness ... inescapable.”2
The arbitrariness that is implicated in this case involves the determination of which of successive surviving spouses (here widows) is entitled to secondary benefits based on the work record of the deceased. The basic test looks to the marital law of the state of the covered worker’s domicile at the time of his death to ascertain which claimant is the legal widow, and thus entitled to benefits. Unfortunately, this approach does not fully promote either uniformity or nonarbitrariness. Some states, for example, recognize de facto marriages,3 while others do not, a situation which would be likely to result in substantial injustice were entitlement to turn entirely on state law resolution of the question. In 1960 Congress responded to this problem by establishing a federal statutory test to provide benefits to persons who married in good faith unaware of legal impediments, such as a prior, undissolved marriage, to their marriage’s validity.4 [1069]*1069This statute, which grants entitlements to such “good faith” spouses, also contains a proviso that no “legal” spouse qualify for benefits.5 Thus, to sketch the basic statutory scheme, a second wife, such as appellant here, unaware that her husband was not divorced from his first wife, would be entitled to a widow’s benefit upon her husband’s death only if his first wife did not lay claim or was not entitled to those benefits.
This already complex portrait is further complicated when the surviving spouse is herself also entitled to primary, i.e., retired workers’, benefits. If those primary benefits are greater than the spousal benefits, the statute provides that there is no entitlement to the secondary benefits.6 In such a case, a second good faith or “deemed” wife is therefore entitled to surviving spouse’s benefits. The truly hard case — such as the one here before us — involves the situation where the secondary spousal benefits to which the legal widow is entitled are greater than the primary benefits which she may claim. Does the entitlement of the first wife to even a minute portion of the available surviving spouse’s benefits disentitle the second good faith spouse, or so-called deemed widow, to any and all such benefits?
Faced with that problem, this court in Rosenberg v. Richardson, 538 F.2d 487 (2d Cir.1976), construed the operable statute to mean that the “second Mrs. Rosenberg” (whose husband had previously obtained a Mexican divorce not recognized by New York law) was nevertheless entitled to her deemed widow’s benefits. We also held in Rosenberg that the deemed widow’s secondary benefits would be reduced by the amount payable to the legal widow, and that in no event could the total payments to widows exceed what could be paid to a single widow.
[1070]*1070Rosenberg has been criticized by other courts,7 has been said by a commentator to reach an “insupportable conclusion,”8 and we are now asked by the Secretary to overrule it.9 For reasons that we state below we decline to do so. Indeed, we believe the result reached in Rosenberg to be not only Solomonic but practical and we also believe its reasoning supportable by reference to the statutory structure and language, which, though by no means crystal clear, nevertheless evince a purpose served by our decision.
Facts
Our case is similar to but somewhat more complicated factually than Rosenberg. It is complicated further by an administrative muddle that is somewhat analogous to the muddled marital life that Samuel Capitano lived.
Sam Capitano evidently married Betty, nee Anna Jokubynas, on December 2, 1931, in Scranton, Pennsylvania. We say “evidently” because the name on her marriage certificate was “Petronella Jackaben.” Betty furnished, however, an explanation for the name discrepancies that satisfied one of the administrative law judges who dealt with this matter in its rather lengthy traverse through the administrative process, so that it was “reasonable to infer” that it was Betty Capitano who actually married Sam on that day.
Sarah Sparacio married Sam Capitano in Pittston, Pennsylvania, on April 7, 1951. Whether that marriage was “in good faith,” as she claims, or not, as the ALJ may have found (contrary to the Appeals Council’s subsequent decision), is an issue both of fact and of law which we discuss below. For the moment suffice it to say that Sarah was known as Mrs. Capitano, was called “wife” by Sam (who called himself “your loving husband Cap”), paid for Sam’s funeral, was “Wife & Beneficiary” of his life insurance, annually remembered the date of his death with rather moving newspaper notices, and bought a burial lot for the two of them. After Sam’s death in January, 1970, Sarah quite naturally applied for and received widow’s Social Security benefits. These benefits commenced in July, 1975.
Betty maintained during the administrative proceedings that she “never got a divorce and never even applied for a divorce” because she “did not want to give [Sam] the satisfaction of granting him a divorce so that he would be free to remarry.” She also claimed that from the time Sam left her in the 1940s until after he died she was “unaware of his whereabouts.” Betty claimed that she concluded ultimately that he must have moved to Buffalo, New York, where he had three sisters. Betty, whose account of her life with and without Sam occasionally stretches one’s imagination and patience, nevertheless also stated that one of Sam’s sisters with whom he was living had phoned her and told her that he was dying, that “I immediately went to Buffalo to the hospital to visit him” and that he “died while I was there and I attended the funeral.” Somewhat incredibly, Betty also stated in writing that “the name Sarah never came up while I was at the funeral, I never heard of this ‘Sarah.’ ”
Be that as it may, on December 7, 1977, Betty applied on the basis of her own earnings for primary retirement insurance benefits. Reluctantly, because she “did not [1071]*1071like the idea of going on” Sam’s work record, she also, some three weeks later, applied for widow’s benefits, apparently at the behest of the Social Security people. It does not appear in the record exactly what prompted this application, since her own application for retirement benefits had already been approved (as of July, 1981, she was receiving $415 per month in primary benefits) and since it is not certain whether Betty receives any more money by virtue of the widow’s benefit than she would receive by virtue of her own retirement benefits. Appellant’s counsel believes that at most Betty receives an additional $17.40 per month. But regardless of how much or even whether or not Betty’s fortunes improved, Sarah’s suddenly plummeted. As a result of the approval of Betty’s (i.e., the legal widow’s) application for benefits, Sarah was held disentitled to receive spousal benefits as of November 1, 1978. Her counsel advises us that since July, 1981, she has received some $233.90 monthly by virtue of her own primary benefits; if she received a spouse’s benefit she would, it is asserted, get an additional $198.50.
The facts pertaining to Sam’s two marriages, so far as they appear in the administrative record, are, as noted above, sketchy. Apparently Sam filed for a divorce from Betty in Luzerne County, Pennsylvania, on October 25, 1945, but the action stalled when Betty sought alimony pendente lite and when her attorney obtained an order requiring Sam to pay $150 attorney’s fees and $65 costs. While Sam moved around the country quite a bit thereafter, a search of the court records where he lived (including Las Vegas, Nevada) revealed no divorce.
Sarah testified that Sam had mentioned to her the fact that he had previously married and that he had children from this marriage, but that he did not talk much about his former wife because it was something that he “did not care to bring up” and that it was a matter about which she never asked “anything.” Sarah never saw Sam’s daughter and saw his son only once. Sarah also testified that Sam did tell her that he was divorced “but it was a delicate subject and I never tried to bring it up too much,” although other members of his family told her that he had been divorced. Throughout repetitive cross-examination from the AU, Sarah reiterated that Sam told her from time to time that he had secured a divorce but that he never told her where he had done so and that she had never asked him where or demanded to see a divorce decree.
Ruling against Sarah,10 the AU went to some lengths to distinguish the case from Rosenberg on the basis that in respect to the element of “good faith” Sarah “never made even the most cursory inquiry about [Sam’s] marital status” and that she “never asked” Sam about the divorce or where it occurred. The AU found an inconsistency in Sarah’s testimony because she explained that she and Sam were married in Pittston because his family was there, but in a statement made to the Social Security Administration in May, 1978, had stated that the only witnesses to the ceremony were the justice of the peace and his wife. The AU was not satisfied that a misstatement made by Sarah in February, 1970, to the effect that Sam had told her that he had been divorced in Pittston was attributable to the fact that her husband had died one month earlier and that when she filled in the form with the misstatement she did not have a clear mind and was still in a state of shock. The AU’s conclusions were not gently put; Sarah’s statements were “inconsistent,” her lapses of memory “convenient,” so that “the most charitable context in which her behavior [could] be placed ... is that ... she put on blinders and forged recklessly ahead ... without an inquiry that could have been settled by the production of court papers____” His ultimate finding was that “her good faith was not in the validity of her marriage, but in Sam, and there it was clearly misplaced because she was not acting as a reasonably prudent individual should have or would have.”
The AU also indicated that “[i]t may well be that Betty will get little if anything in a net calculation.” He did not, however, make this calculation, despite the fact that if Betty were to receive no additional money, she would technically not be receiving widow’s benefits. See supra note 6. Additionally, though Sarah’s attorney stated that Sarah received $200.70 from her own [1072]*1072retirement benefits, so that she would only be-entitled to receive an additional $161.80 out of the total of $362.50 available as widow’s benefits to Sam’s widow or widows, the AU found in the record no application by Sarah for retirement insurance benefits, or even her Social Security earnings record, so that he simply could not make the Rosenberg comparison of the net benefits of Betty and Sarah, even “if this factor was relevant and material to the disposition of this case.” The AU went on to find the “facts and equities” in the Capitano case different from those of Rosenberg and “closer” to those of Davis v. Califano, 603 F.2d 618 (7th Cir.1979), and Dwyer v. Califano, 636 F.2d 908 (3d Cir.1980), cases in which benefits were denied to “good faith” widows.
On request for review, the Appeals Council, on July 13, 1981, informed Sarah that while “it was previously determined that you did marry the wage earner in good faith,” in view of the fact that there was a “legal” widow, Sarah’s benefits had been properly terminated. The United States District Court for the Western District of New York, John T. Curtin, Judge, nevertheless held that the threshold requisite for applying Rosenberg had not been met since “the opinion of the Secretary that Sarah did not marry in good faith is supported by substantial evidence.” It is from Judge Curtin’s opinion and order that this appeal is taken.
Discussion
While the Government argues as Point I in its brief that the AU’s “finding that [Sarah] did not enter into marriage in good faith is supported by substantial evidence,” the AU made no such finding in so many words. He did find that Sarah was not entitled to widow’s insurance benefits as a deemed widow under section 216(h) of the Act “as she is precluded from such status because there is a legal widow entitled to such benefits.” He did not conclude that Sarah knew her marriage was invalid. The AU concluded only that she did not meet a standard which required that she not have placed trust in her husband without first having obtained concrete proof in the form of a search of the divorce records that the trust was merited. He purported to distinguish Rosenberg saying that the facts of Rosenberg “are so far removed from those of the instant case ... that it does not apply,” but the most that can be said is that, without finding that Sarah acted in bad faith, he found that she did not behave the way a reasonable and prudent bride-to-be should behave. Faced with those findings, the Appeals Council, while declining to review the decision and thereby upholding it, stated that
it was previously determined that you did marry the wage earner in good faith. The termination of your widow’s insurance benefits was in accordance with section 216(h)(1)(B) of the Act as interpreted by SSR 80-9c(C)(E) (1980).
We interpret the Appeals Council action as clarifying the determination of the Secretary, since the AU did not make an explicit finding as to whether or not Sarah acted “in good faith” under the statute. In the light of the ultimate finding by the Appeals Council that Sarah “did marry the wage earner in good faith,” we think that the district judge’s conclusion that “the opinion of the Secretary that Sarah did not marry in good faith is supported by substantial evidence” is simply wrong.
Nor would a finding that Sarah lacked good faith be supported by substantial evidence or be in accordance with law. Much of the AU’s opinion dealt with whether Sarah and Sam were legally married, a question totally irrelevant to the issue of Sarah’s good faith in going through with the marriage ceremony. While the AU asserted that Sarah “never made even the most cursory inquiry about [Sam’s] marital status,” there was no reason for her to have done so. At the outset, Sarah testified that Sam told her several times that he was divorced, and we think it unreasonable to demand a newlywed to conduct searching inquiries or independent checks into her spouse’s veracity. Indeed the AU admitted that Sarah had good faith “in Sam.” In our view that is enough to require a holding that she had good faith in thinking that he was divorced. The statute’s words “good faith” do not mean a belief in a set of facts that are not unreasonable to assume, but rather whether the spouse believes her spouse-to-be is free to marry. Indeed, our case is in this regard a fortiori to Kirkland v. Railroad Retirement Board, 706 F.2d 99 (2d Cir.1983), where the purported wife two years after her marriage signed a suretyship contract to guarantee her husband’s payments to his first wife and child, and also learned of the first wife’s knife attack on her husband. There this court ruled that those events had .“little if any bearing” on the [1073]*1073purported wife’s knowledge of her husband’s marital status two years earlier. Id. at 103.
The AU also placed some emphasis on the reply to a Social Security Administration interrogation of Sam’s brother Anthony to question Sarah’s good faith. But the question put to Anthony was, “Are you aware of. a divorce between Samuel and Betty Capitano?” The answer he gave in 1978 was “No divorce to my knowledge.” How Anthony’s knowledge in 1978 of a divorce has any bearing on Sarah’s state of mind in 1951 is totally unexplained by the AU. Nor are Sarah’s responses to Social Security questionnaires in 1970, 1975 and 1978 relevant to show her good faith or lack thereof in 1951, especially since she did not fill out the forms but merely signed her name to them after they were filled out by a Social Security worker. Moreover, Sarah explained that as to the 1970 misstatement she was still in a state of shock after her husband’s death. The AU does not explain why it was not reasonable for her to be in shock or why he believed her state of shock should not have put her in a confused state. See generally Small v. Califano, 565 F.2d 797 (1st Cir.1977) (Secretary’s failure to explain findings). The AU did not, in other words, weigh all the probative evidence.
As to the circumstances at the time of Sam’s and Sarah’s marriage ceremony, Sarah testified that she and Sam were married in Pittston in 1951 and this testimony is borne out by a certificate signed by Aider-man Harold Morrissey in which he certifies that “on the 7th day of April, at Pittston, Penna, Samuel Capitano and Sarah Sporacio were by me united in marriage.” The fact that an assistant clerk in Luzerne County could not find Samuel and Sarah’s names in her records does not permit the inference, drawn by the AU, that the marriage license used by Sam and Sarah was a “falsified document.” Even were the license proven to be forged, it would not in itself show any complicity on Sarah’s part.
The Social Security regulations themselves define “good faith” as meaning that “at the time of the ceremony you did not know that a legal impediment existed, or if you did know, you thought that it would not prevent a valid marriage.” 20 C.F.R. § 404.346(a) (1983). To the extent that the AU defined good faith as meaning a belief in a set of facts that are not unreasonable to assume, he was imposing a higher standard than required by the governing regulations. And to the extent that the district judge relied on the AU’s findings, his .conclusion was erroneous.
Rosenberg
Once it is seen that there was not a finding of bad faith on the part of Sarah and that, even if there were, there was no substantial evidence to support such a finding, Rosenberg becomes applicable. The distinctions drawn by the AU between this case and Rosenberg are totally insubstantial. The only conceivable basis for such a distinction would be that the monthly secondary benefit to which the first Mrs. Rosenberg became entitled was $1.40, while the first Mrs. Capitano may have been entitled to up to $17.40. But this is a difference without materiality where, as here, the total secondary entitlement which a surviving spouse of Sam Capitano would have received amounted to $362.00. Accordingly, we hold that Rosenberg applies, and reverse and remand to the Secretary for calculations.11
We have noted the criticism of Rosenberg. Perhaps the clearest statement of the reasons for disagreement is that of the Third Circuit in Dwyer v. Califano, 636 F.2d at 911-12. The first such reason there mentioned was that the plain meaning of the statutory term “benefit” in the section which states that the deemed widow’s entitlement “shall end with the [time] ... that another person is entitled to a benefit,” must mean any benefit. Rosenberg held that the term was ambiguous and should be read to mean a “full benefit.” But we are not as persuaded of the plain meaning of the statute as the Dwyer panel. We do agree with the Third Circuit that the legislative history of section 216(h)(1)(B) in large part only restates the basic language [1074]*1074of the statute. However, while the statute speaks in the singular “benefit,” the legislative history refers in the plural to “benefits.” S.Rep. No. 1856, 86th Cong., 2d Sess., reprinted in 1960 U.S.Code Cong. & Ad.News 3608, 3629, 3684. The only significance of this is, as we pointed out in Rosenberg, 538 F.2d at 591 n. 4, that Congress has in other contexts used “benefit” in the singular to mean “full benefit.” See 42 U.S.C. § 402(e)(2)(A) (“such widow’s insurance benefit for each month shall be equal to the primary insurance amount of such deceased individual”).
Second, the Dwyer panel argued that Rosenberg “ignores other language in section 216(h)(1)(B),” specifically the clause “has been entitled to a benefit.” It argued that if a legal widow previously, but no longer, received her spouse’s benefit and hence “has been entitled to a benefit” the deemed widow could not receive any widow’s benefits for that husband. Dwyer, 636 F.2d at 911. Dwyer cited Woodson v. Califano, 455 F.Supp. 457 (S.D.Tex.1978), for that harsh holding and properly noted that the appeal had been docketed in the Fifth Circuit. On appeal the decision in Woodson was reversed by the Fifth Circuit which declined to rehear the matter en banc, 656 F.2d 1169 (1981), rehearing en banc denied, 671 F.2d 118 (5th Cir.1982). While the facts and statutory provision implicated in Woodson differ somewhat from those of Rosenberg or the case at bar, the cases were sufficiently analogous for the Fifth Circuit to note that its interpretation “had the company of the Second Circuit” in Rosenberg, 656 F.2d at 1173. Woodson in other words takes the Rosenberg point of view.
Third, Dwyer argued that Rosenberg “based its result on the wage earner’s ‘fund’,” 636 F.2d at 911, and pointed out properly that Social Security taxes are not set aside in a separate account or fund so that any concept of a “fund” being “exhausted” has no significance. While Rosenberg did not use the word “fund,” it did refer to permitting a person in the second Mrs. Rosenberg’s position “to receive a widow’s benefit unless her husband’s work account is already exhausted in payments to a ‘legal’ widow.” 538 F.2d at 491. But not only was Rosenberg not “based” on the existence of a wage earner’s fund, it also clearly held that if for any reason, such as a beneficiary’s receiving her own primary benefits based on her own earnings, the amount available to the deemed widow is less than the sum represented by the surviving spouse's benefit, that benefit simply goes unused. There is thus no reduction in the social security trust reserves, and Rosenberg does not inject any kind of vesting concept into the statutory scheme. The Rosenberg panel was as aware as is this one that there is no “fund” for individual payments.
The fourth reason given in Dwyer for not following Rosenberg was that no opinion that court was able to find outside the Second Circuit followed Rosenberg. We are aware of other situations where this court has been alone — for a time — in taking a given view of a statute. See, e.g., North Haven Board of Education v. Bell, 456 U.S. 512, 520 n. 9, 102 S.Ct. 1912, 1917, n. 9, 72 L.Ed.2d 299 (1982). Be that as it may, since the decision in Dwyer, Judge McKay’s dissent in Martin v. Harris, 653 F.2d at 433, and the Fifth Circuit’s decision in Woodson v. Califano, have come along so as to make the Third Circuit’s statement no longer accurate or applicable. Thus, even if it is a reason for not following Rosenberg that no other court had done so, that reason no longer exists. It is worth noting that another panel in our own circuit has declined to distinguish Rosenberg. Kirkland, 706 F.2d at 104.
The words “is or has been entitled to benefit” in section 416(h)(1)(B) simply cannot be read as mechanically as the Dwyer, Davis, and Martin majority opinions would read it. It is uneontested that until the legal widow actually comes forward and applies for a benefit and is certified to receive it, the “deemed widow” may continue to receive the benefit. Thus, even if a legal widow existed and the Secretary knew her to exist and knew her to be the legal widow, the deemed widow would nevertheless be entitled to receive the benefit until the legal widow made a successful claim. As Judge McKay pointed out in his dissent in Martin, 653 F.2d at 435, the only adequate legal principle relative to this fact is “the policy against double dipping.” As he stated, and as the Woodson court was subsequently to hold, if a legal widow becomes disqualified to receive the benefit the deemed widow should be entitled to receive it. So too for the situation where the benefit is only partially utilized by the legal widow, and together the two widows’ claims do not exceed the secondary benefits which would be available based on the [1075]*1075work record of the deceased. In this sense, there is no “double-dipping.”
As Justice Rehnquist has written, the social insurance and spouse benefit statutory scheme will “have been expanded by amendment over a period of years so that it is virtually impossible to say that a particular amendment fits with mathematical nicety into a carefully conceived overall plan for payment of benefits.” Califano v. Goldfarb, 430 U.S. 199, 225, 97 S.Ct. 1021, 1036, 51 L.Ed.2d 270 (1977) (dissenting opinion). The same commentator who criticized this court's Rosenberg decision as being “unsupportable” nevertheless concedes that the “inescapable dilemma is how to allocate spouse benefits among [spouses],” and he adds that the abandonment of the principle expressed in the original eligibility test which required “a legal spouse living with or supported by the worker” “leaves no compelling basis for choosing who among several ‘spouses’ will receive benefits — assuming Congress retains an essentially all-or-nothing system.”12 But that is, of course, to assume what we believe is the incorrect answer to the critical question at issue here and in Rosenberg. Rosenberg took the view that it was not such an all-or-nothing system, and having in mind “that courts should enforce a statute in such a manner that its overriding purpose will be achieved, even if the words used leave room for a contrary interpretation,” Haberman v. Finch, 418 F.2d 664, 666 (2d Cir.1969), we see no reason to disturb that judgment.
Relying also on the principle that the Social Security Act is a “broad program of social insurance on which working people [can] rely to provide for themselves and their dependents in old age” and that the system is “one based on contributions, in which claimants are not recipients of ‘handouts,’ ” the Rosenberg court, 538 F.2d at 490, construed “benefit” as used in section 416(h)(1)(B) to mean a “full” benefit so that a person in Frieda Rosenberg’s position or in Sarah Capitano’s is entitled to this secondary benefit after subtracting the legal widow’s share and the primary insurance benefits due the “deemed widow.”
Until better advised by Higher Authority we will continue to follow Rosenberg and Kirkland.
On remand, calculations as to amounts payable to appellant.