MICHAEL J. KAPLAN, Chief Judge.

Here the Court is asked to decide whether the Second Circuit's decision in In re Riddervold is still good law after the United States Supreme Court decision in Barnhill v. Johnson. Implicit in the request is a deeper question of whether that determination is appropriate for this Court to make, or whether this Court instead must abide by Riddervold and leave the question of Riddervold's continuing vitality to higher authority. Is a trial judge obliged to obey a seventeen year old decision of the Circuit Court of Appeals after the United States Supreme Court announces a rule that casts doubt upon, but does not clearly undermine, the Circuit Court's analysis?

After laboring without success to find clear guidance, this writer finds that a theory of vertical precedence that some find improper has application in this limited circumstance. The theory is the "predictive" (or "proxy") model of vertical precedent by which this judge must predict what the Riddervold Court itself (the Second Circuit Court of Appeals) would now rule in a Riddervold type of case, in light of the Supreme Court's decision in Barnhill and other intervening factors.

Once determining that that model applies, the role of trial judge is not to choose how to apply Riddervold and Barnhill as precedent (a task which frustratingly leads this writer to conclude that he does not know what the law currently "is" in light of Barnhill), but requires instead an analysis (or "prediction") of how the Second Circuit Court of Appeals would itself reconcile the "vertical precedent" of Barnhill with the "horizontal precedent" of Riddervold.

This writer believes that the Second Circuit itself would find that this Court would not be bound to follow the ruling of the prior panel in Riddervold, and that consequently he may decide the matter anew, but in light of Barnhill, as controlling authority and in light of Riddervold as persuasive authority.

Doing so, I conclude that wage garnishments deducted during the 90 days before bankruptcy are recoverable by the Debtor as voidable preferences under 11 U.S.C. § 547 and 522(h) and (i).

BACKGROUND 1. Riddervold

In 1981, the Court of Appeals for the Second Circuit decided in Riddervold v. Saratoga Hospital, that if a judgment creditor duly levied a wage execution under N.Y.Civ. Prac. L. R. § 5231 (McKinney 1995) before the beginning of the ninety-day preference period contemplated in 11 U.S.C. § 547(b)(4)(A), then deductions taken from the judgment debtor's wages pursuant thereto, and within the ninety-day period, are not voidable preferences under 11 U.S.C. § 547. See 647 F.2d 342, 346 (2d Cir. 1981). Without discussing the question of whether the issue was to be decided under federal or state law, and without mentioning 11 U.S.C. § 547(e) which expressly defines the time at which "a transfer is made" for purpose of § 547, the court examined the language of § 547 and the consequences of levy of a wage execution under state law, and found "nothing in the language or the policy of the 1978 Code" to obviate the state-law consequences of such a levy. Riddervold, 647 F.2d at 347. The court therefore adopted the pre-Code view that a levy executed outside the ninety-day preference period is a "continuing levy" the effect of which is to leave the debtor "no property or interest in property subject to the levy which can be transferred" after the sheriff has served the income execution on the employer. Id. at 346.

2. Barnhill

In 1992, the United States Supreme Court was asked a different question under the same statute: whether delivery of a check is deemed to be a "transfer" for the purpose of § 547, or does the "transfer" occur when the creditor received the check, or when the check is honored by the bank. See Barnhill v. Johnson, 503 U.S. 393, 394-95, 112 S.Ct. 1386, 118 L.Ed.2d 39 (1992). In Barnhill, the Supreme Court stated:

`What constitutes a transfer and when it is complete' is a matter of federal law. This is unsurprising since 11 U.S.C. § 547(e) itself provides a definition of `transfer' but that definition in turn includes references to parting with `property' and `interest in property.' In the absence of any controlling federal law, `property' and `interests in property' are creatures of state law.

Id. at 397-98, 112 S.Ct. 1386 (citations omitted).

The court then examined the rights and duties enjoyed under state law by each party to a check transaction, and the more general definition of "transfer" contained in 11 U.S.C. § 101(54). Id. at 398-401, 112 S.Ct. 1386. The court concluded that by virtue of 11 U.S.C. § 101(54) and 11 U.S.C. § 547(e)(2)(A), the "transfer" occurs when the check is honored by the bank. See Id. at 407, 112 S.Ct. 1386. Consequently, where the check was delivered before the beginning of the ninety-day period, but was honored within that period, the transfer may be voidable under 11 U.S.C. § 547.

3. The Issue

Courts outside the Second Circuit have questioned whether the Riddervold decision survives the Barnhill admonition to look to federal law to resolve the question of when a "transfer" has occurred. See, e.g., Chiasson v. First Tennessee Bank Nat'l. Assoc. (In re Kaufman), 187 B.R. 167, 171 (Bankr.E.D.La. 1995).

Now it is argued that this Court, otherwise bound to apply the Riddervold decision to this case in which the facts addressed are identical to those addressed in Riddervold, must instead view itself to be bound by the Supreme Court's command in Barnhill, even though the court in Barnhill addressed a different type of transaction (a payment by check) and never addressed Riddervold or the matter of wage executions.

There does not appear to be any clear authority on this question. There may be several reasons for this: many times a higher court will cite lower court decisions that are not directly on point, but which were based on reasoning of which the higher court disapproves; a lower court would not publish on the fact that it will no longer apply the disapproved analysis. Also, in instances like this (where the court confronted with the problem is inferior to the court whose reasoning seems to have been undermined by the High Court determination), an inferior court will not publish its election to apply the intermediate court's decision and leave it to that court, on subsequent appeal, to overrule itself if that intermediate court believes it must.

This writer believes it important to find the governing rule. Though there is only $1,153.85 in dispute, it is the assetless Debtor, not a creditor or bankruptcy estate, who would be entitled to these monies (under 11 U.S.C. § 522(h) and (i)) if they are recoverable under 11 U.S.C. § 547. For this Court to defer to the Second Circuit Court of Appeals for an answer to this question would require that the Debtor must appeal first to the District Court, and then, if necessary, to the Circuit Court, to simply find out whether the U.S. Supreme Court's decision commanded the present Court to award him judgment for $1153.85 at the outset. Of course, decisions are not to be made on the basis of who may least afford the burden of appeal. Rather, this burden is addressed here only to underscore the importance of this Court determining as a matter of law where its duty lies, rather than simply presuming that its duty is to defer to the reviewing courts.

(a) Stare Decisis, Generally.

As noted at the outset, a lower court's acceptance of a higher court's ruling as precedent that "binds" the lower court is sometimes known as "vertical stare decisis" or as "vertical or hierarchical precedent." It has been said that without that binding effect, there would be anarchy. No inferior court may ignore clear higher authority that itself has not been foreclosed by yet higher authority. Thus, there can be no doubt that this Court is bound to apply the teaching of both Riddervold and Barnhill if they are not inconsistent. But if they are inconsistent, what is the test? Must Riddervold apply unless there is no way to reconcile it with Barnhill? Or must Barnhill apply if it casts significant doubt on Riddervold? What of the fact that Riddervold's facts were directly on point and Barnhill's were not?

There is an additional consideration. The state statute interpreted by the Riddervold court has been amended since the Riddervold decision. The amendments are of a nature that may or may not have led the Riddervold court to a different decision (had the amended version of the statute been in effect at the point of time that was material in the Riddervold case). How is this circumstance presently to be accommodated?

(b) The "predictive model," and its prediction.

As confessed at the outset of this decision, the present writer is simply at a loss to know what the present law "is" in this case. That being so, this writer considers it appropriate to accept the guidance advanced by Professor Caminker that there are times when an inferior court may properly apply the "proxy model" (also called the "predictive model") rather than the "precedent model" of adjudication, and that one of those times is when the law is unclear. See Evan H. Caminker, Precedent and Prediction; the Forward-Looking Aspects of Inferior Court Decisionmaking, 73 Tex. L.Rev. 1 (1994).

Specifically, there are ambiguous precedents here. One is more recent and from the highest authority, but not on "all fours" with the present case. The other is older and from only intermediate authority and interprets a state statute that has been amended in a respect that may or may not be material; but this precedent is on "all fours." I am bound by both. Under such circumstances, some judges seeking to avoid reversal might be found to "interpret [the] ambiguous precedents or fill the gaps therein . . . in the manner she believes her reviewing court would do. [But unless] such a judge candidly acknowledged her intention, this motivation and methodology would not be discernible from her written [opinion]. . . . [Some] degree of predictive behavior likely lurks behind the lines of some and perhaps many inferior court opinions." Id. at 78. As Professor Caminker might observe, I am motivated simply by candor to use the discourse of prediction rather than the discourse of precedent.

Offering my best prediction of what the Court of Appeals for the Second Circuit will say about Riddervold when now asked, in light of Barnhill and the amendments to CPLR § 5231, I believe that the Second Circuit would first determine the extent to which horizontal stare decisis commands adherence to its own earlier decision in Riddervold. Fortunately, there is a very clear rule in that regard. It is well-settled in before the Second Circuit Court of Appeals that one panel of that court is "bound by a decision of a prior panel unless and until its rationale is overruled, implicitly or expressly, by the Supreme Court or [by the Circuit Court] en banc." United States v. Ianniello, 808 F.2d 184, 190 (1986) (citing In re Jaylaw Drug, Inc., 621 F.2d 524, 527 (2nd Cir. 1980)). See also City of Timber Lake v. Cheyenne River Sioux Tribe, 10 F.3d 554, 557 (8th Cir. 1993); Finkel v. Stratton Corporation, 962 F.2d 169 (2nd Cir. 1992) (the prior panel rule "does not apply where an intervening Supreme Court decision casts doubt on the prior ruling.") Thus the Circuit would focus less on the factual distinctions between Riddervold and Barnhill, and more on the effect of Barnhill on the rationale or reasoning of the Riddervold decision.

Examining the rationale and reasoning of the Riddervold decision carefully, and observing that no mention is made of 11 U.S.C. § 547(e), one cannot be certain whether that provision was never pointed out to the Riddervold panel, or whether the panel considered the provision but rejected § 547(e) as having any mentionable significance. Whichever, the Riddervold panel's emphasis on interpretation of the state garnishment statute raised a clear implication that state law, and state law alone, was dispositive of the question of the time of the transfer of the judgment debtor's interest in the garnishable portion of wages yet to be earned.

That approach may seem at first blush to be irreconcilable with the statement in Barnhill, that "what constitutes a transfer and when it is complete is a matter of federal law." But the Barnhill Court went on to say that while the Bankruptcy Code itself provides a definition of "transfer," "that definition in turn includes references to parting with `property' and `interests in property.' In the absence of any controlling federal law, `property' and `interests in property' are creatures of state law." See Barnhill, 503 U.S. at 398, 112 S.Ct. 1386. The Barnhill Court proclaimed it "helpful" to "sketch briefly the rights and duties enjoyed under state law by each party to a check transaction." Id. After doing so, the Supreme Court returned to the definition of transfer contained in 11 U.S.C. § 101(54), reached the conclusion under both state law and § 101(54) that the "transfer" occurred at the time that the check was honored, and then finally noted that conclusion to be "consistent with § 547(e)(2)(A)." See id. at 400-01, 112 S.Ct. 1386.

Riddervold is not necessarily irreconcilable with Barnhill. It is conceivable that if the Barnhill decision had already been rendered at the time the Court of Appeals was deciding Riddervold, the court might still have found that the answer ultimately lay exclusively in state law and would have reached the same decision.

But that is not the test. A new panel of the Second Circuit would not be bound by Riddervold if it found that Barnhill "cast doubt" on the reasoning or rationale of Riddervold. And when one looks most carefully at the Riddervold reasoning, one finds that indeed Barnhill casts doubt on the continuing vitality of Riddervold. This is because the Riddervold rationale was that "[s]ervice of income execution on the employer in effect works a novation whereby the employer owes ten percent of the employee's salary not to the employee but to the sheriff for the benefit of the judgment creditor." Riddervold, 647 F.2d at 346. This was based upon the fact that although the court found a "dearth of authority" on the subject, a state-law principle announced by Judge Learned Hand in connection with a case under the prior bankruptcy law in 1910, characterized an income execution as a "continuing levy." The court carefully looked to see if there was a reason to find that the "continuing levy" principle was no longer applicable.

The court considered two factors. It first looked at the fact that the state garnishment statute made it clear that the employer is under no liability to pay the sheriff until the wages are earned, and the court observed that if "employment is terminated by resignation or dismissal, . . . the levy shall thereafter be ineffective." Id. at 346. In view of this, the court stated that "this does not require us to hold that the portion of the salary subject to the income execution vests in the employee for a fleeting second after it has been earned, when in fact the employer becomes bound at that very time to pay it to the sheriff." Id.

Secondly, the court, recognizing that the authorities upon which it was relying were decided under the Act of 1898, returned to the 1978 Code and stated, "we find nothing in the language or the policy of the 1978 Code that forbids their continuing application." Riddervold, 647 F.2d at 342.

The Barnhill Court too recognized the importance of state law in deciding when a transfer has occurred for purposes of the preference statute, but that court ultimately tested its conclusion against subsection (e) of 11 U.S.C. § 547. This was entirely consistent with its declaration that "what constitutes a transfer and when it is complete is a matter of federal law." Barnhill, 503 U.S. at 398, 112 S.Ct. 1386 (internal quotations omitted). Had the Riddervold court similarly tested its decision, it would have had to explain how the holding that the transfer of the deductible portion of the judgment debtor's wages could have occurred when the employer was served, given the fact that § 547(e)(3) expressly states that "a transfer is not made until the debtor has acquired rights in the property transferred." It is respectfully submitted that a debtor can have no rights in wages that are not yet earned, and that the Riddervold court's emphasis on the rights of the levying creditor vis a vis the employer, rather than vis a vis the debtor, became problematic once Barnhill was decided.

To lay the matter to rest, we find further clear direction from the Supreme Court. After laying out the background of the matter before it, it defined its "task" as "determin[ing] whether, under the definition of transfer provided by § 101(54), and supplemented by § 547(e), the transfer that the trustee seeks to avoid can be said to have occurred before [the beginning of the 90 day period]." Barnhill, 503 U.S. at 397, 112 S.Ct. 1386. If that was the "task" for the Supreme Court in the Barnhill case, then it would seem clear that the "task" for the Riddervold court was to determine "whether, under the definition of transfer provided by § 101, . . . and supplemented by § 547(e), the transfer" of the wage deductions occurred before or within the ninety day period.

Barnhill casts doubt on the Riddervold decision because Riddervold was based almost exclusively on state law and was not tested against the provisions of § 547(e)(3). Testing the state-law conclusion against § 547(e)(3) was, in my view, commanded by the Supreme Court in Barnhill.

CONCLUSION

Finding, by means of the "predictive model," that Riddervold is no longer binding, does not mean that Riddervold is necessarily wrong. It is still of strong persuasive value, particularly where, as here, the High Court decision cast doubt only on the rationale or reasoning of Riddervold, and not on the ultimate holding (which addressed a different set of facts).

After giving great deference to Riddervold, and noting that there are no factual disputes here, I find for the Debtor. There are three reasons:

(1) As profoundly instructive of state law as Riddervold was, I do not believe that the Riddervold holding can be reconciled with the emphasis upon 11 U.S.C. § 547(e)(3) that was commanded by Barnhill.

(2) There were amendments in 1987 to the state's wage garnishment statute, giving a judgment debtor, for the first time, a right to return to state court to seek modification or release of wage garnishments. When that fact is coupled with the language regarding loss of employment that was examined by the Riddervold court, the two combine to constitute sufficient reason (in my view) to hold, in the language of the Riddervold court, "that the portion of the salary subject to the income execution [does vest] in the employee for a fleeting second after it has been earned."

(3) I am persuaded by the decisions from outside the Second Circuit that considered, but rejected, the authority of Riddervold. See, e.g., Taylor v. Mississippi Learning Inst. (In re Taylor), 151 B.R. 772, 776 (Bankr.N.D.Miss. 1993); Malone v. Fidelity Nat'l. Bank (In re Dunn), 56 B.R. 275, 277 (Bankr.M.D.La. 1985).

The Clerk shall enter judgment for the Debtor and against the defendant in the amount of $1,153.85 plus costs, and interest from the date of the filing of the Complaint, once the plaintiff's attorney files an affidavit of amount due and the defendant has had an opportunity to object raise any objection to the computation.

SO ORDERED.