Ponzi, Property, and Luck

“Ignorance is the best of law reformers. People are glad to discuss a question on general principles, when they have forgotten the special knowledge necessary for technical reasoning.” 1

I.     Introduction

Swindler is running a Ponzi scheme. Two or more victims invest with Swindler in the hope of turning a profit. The scheme eventually collapses. Some residue of the victims’ investments is found in the ruins, and the question is how to divide it between the victims. The basis of the division is a claim in restitution or unjust enrichment that is now asserted, explicitly or otherwise, by one fraud victim against another. Remedies in such cases are predominantly equitable, and the question of allocation between victims is sometimes described by comparing their “competing equities” or by asking “what equity requires.” Courts have recently been giving new answers to this old question. My topic is to ask why equity now seems to require something different from what it required in the past.

The traditional answers came from the property-rights end of the equity spectrum, the part that depends least on “equitable discretion.” But during the second half of the 20th century, for a variety of reasons, the law of equitable interests in property (as part of the law of restitution generally) suffered a dramatic decline in professional attention and awareness—a story that is old news by now.2 Propositions that were once part of what everyone knew became doctrine that could only be found in the library, and then only if one knew where to look. While this law was being forgotten, United States courts were encountering an extraordinary upsurge in Ponzi-type cases.3 As it happened, these cases fell to be litigated and decided by a new group of American lawyers and judges: the first professional generation—roughly speaking, of course—to attend law school after the courses dealing with equitable interests in property had been dropped from the standard curriculum.

Lack of awareness of the established rules has given courts an unusual freedom to decide for themselves and afresh what equity requires between victims of a common fraud. In three recurrent settings—all of them familiar locations within the landscape where exploded Ponzi schemes fall to Earth—recent decisions reach outcomes different from those that traditional authority would dictate. Ignorance has not produced a random redistribution, because the new results in victim v. victim cases have all been pushed in the same direction. Courts spread losses more widely by refusing to recognize the effects of fortuitous circumstances—luck—that by standard property rules would occasionally permit a few relatively fortunate victims to lose proportionately less than others. The move has been away from a view that takes victims as it finds them, toward one that seeks to rationalize and to equalize, so far as possible, the consequences of a casualty to which multiple victims are thought to have been uniformly subject.

This shift in outcomes has redrawn the established map of property rights by shortening the reach of legally protected ownership. By traditional rules, the owner of an asset lost to fraud can retake it so long as it can be identified—unless and until it comes into the hands of a bona fide purchaser. Protected ownership thus extends beyond a misappropriation, ending only when (1) the misappropriated asset can no longer be identified or (2) the rights of a protected purchaser intervene. Because they deny the owner’s usual right to restitution in such circumstances, recent decisions terminate ownership earlier: at the point where the asset first leaves the owner’s hands.

Most of the recent victim v. victim decisions do not see the problem in property terms. If they do, they assert—in the name of “equitable discretion”—an overriding authority to reallocate whatever entitlements might exist.4 More commonly, the courts—and the receivers whose rulings the courts approve—see themselves as writing on a blank slate, subject to a single guideline: a duty to divide contested property in whatever manner seems best under the circumstances.5

My intention is not to criticize the courts for ignoring the rules.6 My proposition is rather that the new outcomes in victim v. victim restitution cases make a striking instance of Holmes’ law reform by ignorance. If we observe that some “special knowledge” has evidently been forgotten, and that old questions are being decided instead on “general principles”—and coming out differently—it follows that the currently applicable “general principles” must be different from those that informed the older “special knowledge.” Different assumptions about what makes a desirable result have led new judges to new conclusions, and the question is what those different assumptions might be.

II.     Special Knowledge: Five Propositions

The rules of property law relevant to this discussion are those that determine the degree to which ownership of some asset is protected against involuntary dispossession. That the law gives some such protection is possibly the core idea of “property”—it would be hard to define “property” without it—but obviously such protection is not absolute. The question is how far this protection goes, against whom, and for what species of assets. Ordering competing interests in property is one of the law’s principal tasks, and the relevant law is sometimes complex. Still, most of the rules needed to understand victim v. victim restitution cases can be reduced to five fairly simple propositions.

1. Property is mine until I transfer it voluntarily—with some broad qualifications. Property claims are subject to statutes of limitations, so an involuntary dispossession is protected by lapse of time. Moreover, the law can protect my ownership only so long as I can identify the thing owned—though property law goes further than it might on behalf of dispossessed owners by incorporating an interesting “transitive” feature. It will sometimes give me ownership of a substitute asset to replace the one I have lost. If you misappropriate my X and exchange it for Y, I am the owner of Y so long as I can find it in your possession, and possibly beyond.7 A related instance appears in the law of “confusion of goods.” If I store 50 bales of cotton in the same warehouse that holds your 50 bales, my cotton is mine and yours is yours so long as we can tell them apart.8 If the warehouse were authorized to sell cotton from the common store that would change things, but each of us intends to withdraw his own bales at a later date. A fire at the warehouse destroys 40 bales. Of the 60 that remain, 20 still bear my labels and 30 bear your labels; the labels have been burned off the other 10. After I take my 20 and you take your 30 there is no more of either “my cotton” or “your cotton,” but there is a partial substitute for both in the commingled mass of 10 indistinguishable bales. You and I are tenants-in-common of this substitute asset in proportion to our presumptive contributions to it, so I take six and you take four.9 Other innocent victims—the warehouseman’s other customers, not to mention his general creditors—may well have lost more in the fire than we did, but that does not entitle them to share in our salvaged cotton.

2. Owners who can identify their misappropriated property prevail against competing creditors of the wrongdoer, but they may lose to subsequent purchasers. Protection against theft is the most robust, given the rule that “the thief gets no title and gives none.” Here, the case that tests the extent of protected ownership is not owner v. thief (too easy) but owner v. purchaser, where purchaser has acquired the asset from the thief, giving “value,” with no idea that his seller’s title was not good. Between these equally blameless parties the law protects the owner; protection lasts forever, against an infinite succession of purchasers, cut off only by the statute of limitations. Owner gets the same degree of protection when the misappropriation is achieved by egregious fraud or duress, making the transfer even more involuntary than usual. Conversely, ownership of money and certain money-substitutes receives less protection against theft than ownership of other kinds of property, because an innocent purchaser for value of money (a “bona fide payee”) gets to keep it, even if the money has been previously stolen and remains identifiable by the theft victim.10

3. Such cases aside, if A transfers X to B induced by B’s fraud, B obtains “voidable title” to X. B’s title is “voidable” in the crucial sense that it is ineffective against B’s transferor. If A discovers what happened and catches B with X still in B’s possession, A can rescind the transaction with B and obtain restitution of X. If B has meanwhile transferred X to C, the question is whether C qualifies as a bona fide purchaser. If so, A’s property in X was extinguished by the B-to-C transfer. Litigation of these competing claims of ownership (A v. C) puts a sharp focus on the conditions under which A might lose his property as a result of B’s fraud. A wins unless C acquired X from B (1) by a transaction of “purchase”; (2) without notice of A’s interest; and (3) in exchange for “value.”

4. Centuries of litigation have left a deep body of law on the meaning of these three requirements, with a particular focus on one class of C-type claimants who are not protected purchasers: namely, the lien creditors of B. The contest between A (who discovers his X in B’s possession) and C (B’s judgment creditor, who is looking hard for assets on which to levy) has been one of the most litigated in all of commercial law. A wins: because C has not acquired X by purchase, because C has not given value, or for both these reasons.11 In property terms, C has a lien on all of B’s assets, but that means a lien on “the actual estate of the debtor,” which does not include X. To put it another way: “the judgment creditor stands in the shoes of the judgment debtor,” with the result that C’s rights to X (vis-à-vis A) can rise no higher than B’s. The practical judgment underlying the distinction is not hard to see. Purchase transactions are so important, and they depend so significantly on the confidence of prospective purchasers, that the law will protect a purchaser’s expectations even when it means giving effect to an involuntary transfer. Claimants who do not qualify as bona fide purchasers have no comparable need for protection. B’s lien creditors would naturally prefer it if the X they have uncovered among B’s possessions were actually B’s property, but the fact that it is not will not have any appreciable effect on commerce; and there is no natural equity in robbing Peter to pay Paul.

5. Finally, victim v. victim restitution often requires application of the rules protecting purchasers to determine the ownership of stolen funds that have been repaid. Victim 1 (“V1”) is defrauded but later repaid by Swindler, who pays him with money obtained by a subsequent fraud on Victim 2 (“V2”). Is V1 exposed to V2’s restitution claim? V1’s instinctive response is that “I was only getting my own money back,” and the law has usually been prepared to see things that way; not because V1’s money is specifically identifiable, but because V1 is awarded ownership of the money repaid as a substitute for the money lost.

That result is reached by allowing V1 an affirmative defense to V2’s restitution claim, treating V1 as a “bona fide payee.”12 The conclusion is not altogether stable. When V1 received payment from Swindler he had no notice of the new fraud on V2—otherwise he would be out of luck—but has V1 given “value”? No one doubts that a seller who unknowingly accepts stolen money as payment in a current transaction gives “present value” for the money received. This is why Swindler’s supermarket and his wine merchant face no liability in restitution, even when it becomes clear that every dollar they ever received from Swindler had been stolen from Swindler’s victims. The initial question is whether any payee who receives money to reduce an antecedent debt should be regarded as giving “value.” If Swindler uses $1 million in stolen funds to pay down a preexisting loan from Bank, has Bank given value? What if Bank could give the stolen money back to the theft victims and reinstate the loan, all at the cost of a bookkeeping entry? The reason for the longstanding controversy on this point is that the discharge of an antecedent debt does not necessarily involve any detrimental change of position.

The equitable difficulty becomes more pointed in the victim v. victim scenario, when the antecedent debt being satisfied is essentially worthless. Assume that Swindler is judgment-proof; that he owes V1 $100 from an earlier swindle; that he acquires $100 by a new fraud on V2; and that V2’s $100 is paid over to V1. If V1 is allowed to retain $100 in real cash as a dollar-for-dollar replacement for the $100 previously stolen, the resulting substitution of assets is extraordinarily favorable to V1 and correspondingly harmful to V2.13

Prodded by statutes, modern U.S. law has moved steadily toward a more expansive definition of value, one in which satisfaction of antecedent debt usually counts just as much as groceries or wine.14 The usual explanation is that a concern with finality in commercial transactions—and in payment transactions above all—leaves no room for the laborious balancing of equities by which a more complex older law distinguished what might or might not constitute “value” in particular situations. Treating antecedent debt as value clearly promotes the finality of payments. Measuring that value at 100 cents on the dollar—even when the antecedent debt consists of an uncollectible claim against a crook—promotes finality even more decisively, though at a visible cost in fairness.

Even if we concede the overriding importance of finality in ordinary-course commercial transactions, things may look different out of the mainstream, in the essentially noncommercial setting of victim v. victim restitution. The statutory hegemony that treats antecedent debt as value, “no questions asked,” still encounters pockets of resistance where it creates windfall gains and losses between innocent victims. One such pocket has now appeared in the Ponzi cases. Until very recently, the orthodox and nearly unanimous rule was that a defrauded investor who withdrew funds before the collapse of the scheme became the owner of the funds repaid, up to the amount of the original investment. The victim’s claim against the scheme would be treated as value, dollar-for-dollar, for the money (stolen from subsequent victims) withdrawn from the scheme. This standard modern attitude toward antecedent debt as value has lately been drawn into question, and for a most old-fashioned reason: the protection it affords to finality of payment comes at too high a cost to equal treatment.

III.     Moving the Property Baseline

Recent decisions in Ponzi cases make fundamental adjustments to these traditional rules of ownership. The effect of the change in each case is to spread losses more widely in cases of victim v. victim restitution. The first and most radical alteration is the refusal to allow an owner to retake a misappropriated asset that remains identifiable in specie. At one remove is the refusal to let an owner retake what is undeniably the product of a misappropriated asset. This second change is no less a revision of existing entitlements: it is like requiring the owners of the commingled cotton to share it with people whose cotton formed no part of the mixture, or refusing to allow a secured party to follow the cash proceeds of collateral through the “lowest intermediate balance” of the debtor’s commingled bank account.15 A third departure from prevailing authority appears in the recent inclination to reject (or at least restrict) the affirmative defense of bona fide payee—the status of antecedent debt as value—in cases of victim v. victim restitution. Limiting the defense means that a defrauded investor who is repaid with money stolen from subsequent victims may or may not be treated as the owner of the funds repaid.

A.     Ownership of Identifiable Property

SEC v. Elliott grew out of “a massive Ponzi-type scheme” and a “web of false misrepresentations” which induced nearly two-thousand victims to part with money and securities.16 In one phase of the scheme,

Elliott talked investors into “loaning” him their securities. He convinced the investors that he could get them a return on their money far greater than they were currently earning in dividends from their securities. In exchange for the securities, Elliott gave them a promissory note, equal to the market value of the securities, promising to make monthly interest payments. The investors delivered the securities to Elliott with executed powers of attorney attached. Elliott could satisfy his obligation to pay either by returning the securities or by making a cash payment. The district court found, and we agree, that what in fact transpired was that the investors unwittingly transferred legal title in the securities to Elliott.17

Many of the securities “loaned” in this way had been sold, and the sale proceeds were impossible to trace. But some of the assets coming into the hands of the receiver consisted of securities that Elliott had left untouched, still registered in the names of the victims who had “loaned” them. The registered owners tried to get them back. The receiver, with the approval of the district court, chose instead to include these identifiable securities among the assets subject to ratable distribution. In the view of the district court:

To allow any individual to elevate his position over that of other investors similarly “victimized” by asserting claims for restitution and/‌or reclamation of specific assets based upon equitable theories of relief such as fraud, misrepresentation, theft, etc. would create inequitable results, in that certain investors would recoup 100% of their investment while others would receive substantially less. . . . [I]n the context of this receivership the remedy of restitution to various investors seeking to trace and reclaim specific assets as originating with them is disallowed as an inappropriate equitable remedy.18

The Court of Appeals agreed:

Legally, these investors occupy the same position as the other investors whose securities were sold. All investors were defrauded. All investors were cleverly persuaded to part with their securities. . . .

. . . .

    We cannot say that the district court abused its discretion by disallowing tracing. A district court has broad powers and wide discretion to determine the appropriate relief in an equity receivership.19

To reach this outcome, the court had to redraw an important baseline of property law: the one that protects specifically identifiable property against involuntary dispossession. The difficulties of doing so were neither acknowledged nor confronted, but the implicit reasoning of SEC v. Elliott and similar cases can be read between the lines. Remedies between fraud victims are equitable. This is because they are administered by a receiver, and receivership is equitable; or because the claimant is trying to “trace” something, and tracing is equitable; or because the claimant is asking for a remedy by constructive trust, and constructive trust is equitable; or because the claimant wants restitution, and restitution is equitable. But “equitable” means “in the discretion of the court.” This is understood to mean that a court exercising its equity powers is constrained only by its sense of justice in allocating ownership of disputed assets between competing claimants.

This is a travesty of equity jurisprudence, one in which everything has been forgotten, except the part about “discretion.” (There are equity problems that depend on the length of the Chancellor’s foot, but the basic rules validating and invalidating ownership of property are not among them.) By standard property rules, when A transfers X to B, induced by B’s fraud, A’s ownership of X is not extinguished by his initial dispossession. Instead A retains the right to rescind the transfer and recover X, if he can find it, either from B or from subsequent transferees (C, D, et al.), unless and until X comes into the hands of a bona fide purchaser. Resolution of such a dispute will often require nice equitable judgment on the part of the court—for example, whether the circumstances in which C acquired X were such that notice of A’s rights against B should be imputed to C. But the authority to decide such questions is not the authority to decide whether property rules assigning ownership between A, B, and C should be recognized or disregarded. The idea that A’s right to recover an identifiable X from either B or C might depend, not only on the facts of the A-B-C transfers, but on the “equitable discretion” of the court, made its first-ever appearance in the cases under review.

The assertion of a power to redraw property baselines in the exercise of equitable discretion starts with a piece of old-fashioned “technical reasoning” that is perfectly correct: the idea that A’s remedy against either B or C is “in equity” rather than “at law.”20 In any standard version of SEC v. Elliott, the fraud victims who were fortunate enough to discover their securities in Elliott’s hands would ask a court having equity powers for specific restitution by one means or another.21 Recent decisions point to the fact that such remedies are historically equitable as the first step in an inattentive non sequitur, enabling the conclusion that identifiable property may be restored or withheld in the court’s discretion. Apart from the fact that an impoverished understanding of “equity” is what gets this reasoning off the ground, the reference to equity is mostly a red herring. The implication is that these courts would have to respect established property baselines if the rights and remedies in question were legal rather than equitable. This is implausible on its face, given our dwindling capability to recognize the differences between historical jurisdictions, the lack of modern reasons to preserve the distinction, and the intolerably arbitrary results that would follow if we did. What the courts want and have assumed is the discretion to reassign property rights, and for this discretion “equity” is merely a fig leaf.22

The problem in SEC v. Elliott is to fix the point at which an owner loses his property in a particular asset, becoming instead the owner of a ratable share of something else. Law and equity draw the line in the same place: substitution occurs when the original asset is no longer identifiable. The answer rests on centuries of case law adjudicating the rights of owners or secured creditors whose assets or collateral had been combined, with or without authority, by thieves, bailees, or other persons in possession, then partially disposed of—so that the original property or its full value could no longer be restored to all the owners. But there is no “confusion,” and no substitution, so long as the separate contributions remain identifiable.23 Owners of gold and silver stored in the same warehouse do not contribute to make up each other’s losses, if the dishonest warehouseman steals more of one metal than the other.

That an outcome is fortuitous does not make it arbitrary, and a rule that puts the property baseline at the point of identifiability is the opposite of arbitrary: it is intensely practical. It pursues a simple, understandable objective—the protection of property against involuntary dispossession—exactly as far as circumstances permit. It makes outcomes turn on facts that are visible and incontrovertible. And because the rule coincides with our most elementary intuition about meum et tuum—about the circumstances that lead someone to say “that’s mine, not yours”—it enlists that intuition in support of its results. The arbitrary rule is the one that reassigns ownership in whatever manner the court considers appropriate.

B.     Ownership of an Identifiable Product

Defendants in United States v. Durham had persuaded 13 victims to invest a total of $806,750 in a nonexistent “advance fee loan financing business.”24 When the scheme was uncovered and defendants were arrested, “roughly $83,495.52 of the money was left.”25

    The charlatans used multiple accounts and company names to implement their scheme. However, by the end, they had only one account (Cypress, Ltd.) with assets (the $83,495.42) at one bank, Pavilion National. All but $8803.99 of the money in the Pavilion accounts could be traced to seven claimants, one of which was Claremont Properties. Only four of these seven filed claims. The Defendants had deposited and withdrawn almost all the money defrauded from the other claimants. A few days after Claremont funds were deposited, the Defendants were arrested.

    Over Claremont’s objection, the district court elected in the interest of equity to distribute the $83,000 pro rata rather than giving the bulk of it to Claremont and the other three victims whose funds had been traced. The court added the total claims, $806,750, and allocated the $83,000 by percentage against the total claim. Claremont’s total claim was $161,750—20% of the total claims. Therefore it would receive only $16,740,83. Uncontroverted evidence from the FBI showed that, if tracing were applied, Claremont was due $70,970.13 of the $83,000.26

The district court took a familiar tack, finding that:

[A]ll claimants stand equal in terms of being victimized by the defendant defrauders. The ability to trace the seized funds to Claremont and Northernaire is the result of the merely fortuitous fact that the defrauders spent the money of the other victims first. Allowing Claremont and Northernaire to recover from the funds seized to the exclusion of the other victims under the tracing principle would be to elevate the position of those two victims on the basis of the actions of the defrauders. The Court sees no justification in equity for this result.27

The Fifth Circuit approved this exercise of the district court’s “inherent equitable powers”:

    No one can dispute that tracing would have been permissible under the circumstances of this case. . . . The government in fact suggested that Claremont receive the traced funds. However, the court, in exercising its discretionary authority in equity, was not obliged to apply tracing. . . .

. . . .

    Sitting in equity, the district court is a “court of conscience.” Acting on that conscience, the lower court in the instant case rationally considered the positions of the victims and held that following the tracing principle would be inequitable. . . . Because the court used its discretion in a logical way to divide the money, the court committed no error requiring our intervention. For us to hold otherwise would be to chain the hands of the court in Equity to do what is right under the circumstances. We will not rob the lower court of the discretion essential to its function.28

So far this sounds a lot like SEC v. Elliott. But Durham introduces a new element, and a new source of confusion, in the reference to “tracing.” Decoding this term makes it easier to see that Elliott and Durham address the same property baseline, and that they move it in the same direction for the same reasons.

Restitution in Durham is slightly more “equitable” than in Elliott, because the relatively lucky victims can no longer identify their misappropriated property in specie. Instead they claim an entitlement to its direct product. Cases like Durham involve this sort of “following” or “tracing” because the misappropriated asset (money paid by Victim to Swindler) immediately undergoes a change of form, becoming bank credit. Next, Victim’s property is “confused” or “commingled,” because the state of Swindler’s bank accounts, and the sequence of deposits and withdrawals, will make it impossible to say whose money is still in the bank and whose has been spent. At this point Swindler’s bank account resembles a grain elevator from which someone has absconded with the combined wheat of different farmers—leaving no alternative to a ratable distribution of the remainder, in proportion to the farmers’ contributions.

The fact of confusion or commingling often means that ownership can be protected only by resort to “tracing rules” or “tracing fictions.” These are presumptions about whose otherwise indistinguishable property has been lost and whose remains, employed to shift losses where a ratable allocation would not be equitable. If dishonest Lawyer combines in a single account $1000 of Client A’s trust money and $1000 of his own, then withdraws $1500 for personal expenditures, A is entitled (as against Lawyer and Lawyer’s creditors) to all of the $500 that remains, not just half. The outcome is inexplicable except as the result of a legal presumption in favor of A: namely, that Lawyer spent his own money before invading client funds. Suppose instead that Lawyer makes one additional deposit to the same account and no further withdrawals: he adds $1000 received in trust for Client B, leaving a closing balance of $1500 when his defalcation is detected. B is entitled to $1000 of the balance (not just half), leaving only $500 for A. B recovers more than A, not because of a presumption in B’s favor—there is none on these facts—but because the sequence of events makes it impossible that any of B’s money has yet been spent. Full restitution to B is the standard result, though the fact that A loses $500 while B escapes loss is indeed “the result of the merely fortuitous fact that the defrauders spent the money of the other victims first.”

Durham and the cases that follow it exploit a confusion between the two different sorts of tracing in these last examples. In the original Ponzi case, Cunningham v. Brown, late-stage withdrawals by some of Ponzi’s desperate investors, following exposure of the fraud, had been condemned as avoidable preferences.29 These victims argued that they had not received the money as Ponzi’s creditors; rather they had rescinded their investments and recovered their own money. But their only means of identifying the funds paid out was the presumption that would have been available in a suit against Ponzi himself: that their funds were still present (and unspent) so long as they were within the lowest intermediate balance of Ponzi’s bank accounts.

Cunningham v. Brown held that a presumption or “fiction” that produced equitable results between some classes of claimants—favoring Client A over dishonest Lawyer, for example—would not be indulged to create priorities between claimants similarly situated. The court did not suggest that it would be inequitable to allow more fortunate victims to reclaim identifiable property or its product, should they be able to do so. On the contrary, Chief Justice Taft—in the course of explaining what equity would allow Ponzi’s victims to do—took as his principal example a nonfictional tracing claim very much like the one asserted in Durham:

But, even if we assume that . . . what the defendants here did was a rescission for fraud, we do not find them in any better case. They had one of two remedies to make them whole. They could have followed the money wherever they could trace it and have asserted possession of it on the ground that there was a resulting trust in their favor, or they could have established a lien for what was due them in any particular fund of which he had made it a part. These things they could do without violating any statutory rule against preference in bankruptcy, because they then would have been endeavoring to get their own money, and not money in the estate of the bankrupt. But to succeed they must trace the money, and therein they have failed. It is clear that all the money deposited by these defendants was withdrawn from deposit some days before they applied for and received payment of their unmatured notes. It is true that by the payment into the account of money coming from other banks and directly from other dupes the bank account as such was prevented from being exhausted; but it is impossible to trace into the Hanover deposit of Ponzi after August 1st, from which defendants’ checks were paid, the money which they paid him into that account before July 26th. There was, therefore, no money coming from them upon which a constructive trust, or an equitable lien could be fastened. In such a case, the defrauded lender becomes merely a creditor to the extent of his loss, and a payment to him by the bankrupt within the prescribed period of four months is a preference.30

The reason to read this much of Chief Justice Taft’s opinion is that Cunningham has come to be cited, in the cases following Elliott and Durham, for the proposition that property rights dependent on “tracing,” or on equitable remedies, or involving “restitution” in general, are to be recognized only in the discretion of the court—indeed, that recovery of property by any such means is only “ex gratia.”31 Properly understood, the authority of Cunningham is squarely to the contrary; but the matter-of-fact discussion of 1925 is a good example of the kind of “technical reasoning” that most lawyers no longer entirely follow. Difficulty in understanding what Taft was talking about has given today’s courts their freedom to rewrite the rules.

Durham moves the same property baseline that was shifted in Elliott, and in the same direction, though in the context of a claim to recover the product of an asset (e.g. money in a bank account) rather than the asset itself (e.g. registered securities). Either way, the point at which ownership of an asset is lost has been moved to an earlier stage of the transaction: from when the asset (or its product) can no longer be identified to when the victim was first induced to part with the asset.

C.     Ownership of Funds Withdrawn from the Scheme

Recent decisions make a comparable adjustment to a different aspect of the Ponzi aftermath. After money obtained from numerous victims has been untraceably commingled and mostly dissipated, some residue of their investments is marshaled and available for distribution. The problem of equitable division is exacerbated by the fact that some investors, but not others, have previously withdrawn money from the scheme. Any such payouts were necessarily made using money stolen from subsequent victims. Courts want to allocate available assets in proportion to losses, but the question—as in both previous examples—is how to measure each investor’s loss.

New methods of judicial accounting spread losses more widely by accelerating them: by recognizing, in those investors who withdrew funds before the scheme was exposed, only a limited ownership in the funds recovered. Making prior withdrawals subject to restitution equalizes the position of the victim who withdrew funds and the victim who did not. Refusing or qualifying the affirmative defense that would otherwise protect Ponzi withdrawals advances the point at which the fraud victim’s investment is irretrievably lost, thus moving a property baseline in the same manner, and toward the same point, as the decisions in Elliott and Durham.

Judges and receivers these days see the question of accounting for previous withdrawals as a matter of selecting the right “method of distribution.” An early decision reflecting this truncated conception, CFTC v. Franklin, put the problem this way:

It appears that the Receiver had four options in approaching these fake “profits” [i.e. the prior withdrawals from the scheme] and the investors who received them: (1) the investors could return the “profits” to the entire pool prior to the pro rata distribution; (2) the investors could both retain their “profits” and demand a full pro rata share of their initial investments; (3) the investors could retain their distributed profits but would be forced to subtract that profit after determining the pro rata share of their investments; or (4) the investors could retain their “profits” but would receive a pro rata share based on their initial investments minus the profit distribution, i.e., profits would be subtracted before determining the investor’s pro rata shares.32

Additional facts in particular cases have produced various qualifications of the four Franklin “options,” and a current list of “differing methodologies for claims allowance and distributions in Ponzi cases” is accordingly somewhat longer,33 but the analytical framework is the same. The modern approach starts at the point where a receiver with money in his hands is looking for a method of distribution. Put this way, the question has only one answer. The receiver is free to choose whichever method of distribution is most appropriate under the circumstances, so long as there is no “abuse of discretion”—which there never is. The objection to this approach is that it presents an old problem as if it were a new one. Even supposing that receivers make the right choice in every instance—and their new choices may indeed be superior—there is a significant cost to coherent legal analysis.

In choosing one method of distribution over another, a Ponzi receiver is necessarily adjudicating a common-law claim (and an affirmative defense) involving the fifth proposition of the “special knowledge” previously described.34 To repeat in two words: the money paid to Ponzi investors as “prior withdrawals” has been stolen from subsequent victims. Victims as a class have a claim in restitution to recover funds withdrawn from the scheme by other victims, except insofar as those recipients may be entitled to an affirmative defense. That defense, if there is one, is called “bona fide payee.” If A uses B’s money to discharge A’s debt to C, B will sue C in restitution; C will defend by asserting that C is, in effect, a protected purchaser of the money. The crux is whether discharge of A’s antecedent debt counts as “value” for this purpose. The question is difficult because discharge does not necessarily involve any change of position, and because antecedent debts are not all alike. In particular, if B and C are both Ponzi victims, C’s defense turns on the discharge of a debt that was probably worthless.

The receiver’s choice of methodology thus determines the extent to which satisfaction of (fraudulent) antecedent debt constitutes value in a Ponzi case. Reverting to the four possibilities canvassed by CFTC v. Franklin, in the passage just quoted: option (1) eliminates altogether the defense of bona fide payee as between Ponzi victims. Prior withdrawals must accordingly be restored to the receiver (i.e., to the victims as a class), permitting a ratable distribution. Option (2) denies any restitution of prior withdrawals, without regard to the extent of antecedent debt—thereby allowing fortunate recipients to retain ostensible “profits” exceeding their original investment. Option (3) is a compromise innovation, to be summarized presently, that has come to be known as the “rising tide method.”35 Option (4), usually known as the “net loss” or “net investment” method, is the orthodox approach. Withdrawals from a Ponzi scheme are considered to discharge an antecedent debt—and are therefore not subject to restitution—to the extent of the recipient’s original investment. A victim’s “net investment” yields the amount of the loss that is the basis of the receiver’s ratable allocation.

Franklin’s “option (1)” has not fared well, though it might seem to have a good deal to recommend it. If the point of the modern distribution exercise is to achieve a Durham-style ratable distribution—to eliminate the accidents of timing that allow one victim to recover his property where another cannot—it is unacceptable that a victim who happens to withdraw some or all of his investment should leave a greater loss to be suffered by those who have withdrawn less or nothing at all. It is even more anomalous to recognize value dollar-for-dollar in the satisfaction of an antecedent debt that was fraudulent in its inception and never collectible beyond some fraction of the funds invested. The straightforward way to deal with these objections would be to disallow the affirmative defense between fellow victims of a common fraud. A few Ponzi receivers have proposed just this—arguing that all previous withdrawals should be subject to restitution, permitting them to marshal the maximum receivership assets for ratable distribution.36 In this unqualified form, the argument for expanded restitution of Ponzi withdrawals is almost uniformly rejected. Most courts think it would be self-evidently unfair to ask fraud victims to surrender what they were fortunate enough to withdraw before the collapse of the scheme, and the SEC apparently agrees.37

No one—apart from a few successful Ponzi investors known as “net winners”—supports “option (2).” It would be inconceivable to let one victim walk away with an actual profit from a Ponzi scheme, when the result is necessarily to increase the losses suffered by fellow victims. It would be equally hard to explain the result in terms of the standard affirmative defense, since beyond the victim’s previous investment there is no antecedent debt to satisfy and no possible value being given. This leaves Franklin’s “option (4)”—commonly known as the “net loss method”—as the orthodox approach. It gives an innocent investor full ownership of any funds withdrawn from the scheme (in other words, an affirmative defense to his fellow-victims’ restitution claim), but only to the amount of his previous investment.38 “Net winners” remain liable in restitution to restore the excess of withdrawals over investments. Victims who have withdrawn part of their investments end up losing less than they would have—and less than the others, as a proportion of what they once had at risk. Whether this is fair or not depends on when we think a loss should be measured: when the victim first sends in the money, or only when that initial investment is net of withdrawals. The question is the same one addressed in Elliott and Durham, and the reasoning of those cases would measure the loss at the earliest possible time.

Making Ponzi victims liable to restore previous withdrawals—“option (1)”—would achieve this result directly. Most receivers do not ask for that much, and most courts would not give it to them. But it is against this background that the recent innovation in “distribution methodology”—the former “option (3),” now known as the “rising tide method”—takes its place. “Rising tide” moves the property baseline in the same direction as Durham and Elliott, and for the same reasons, but moves it only half a step.

Hypotheticals recently constructed by Judge Richard Posner are likely to anchor future discussion of this problem. The defendant in SEC v. Huber “operated a Ponzi scheme in which 118 investors lost a total of $22.6 million. . . . A receiver appointed to marshal and distribute the assets remaining in Huber’s funds was able to get his hands on some $7 million.”39 Eleven of the investors

had withdrawn portions of their investment from Huber’s funds before the scheme was exposed. With the approval of the district court the receiver counted the withdrawals as partial compensation for these investors’ losses. In doing so he was using what is called the “rising tide” method of allocating assets held by a receiver for distribution to creditors; the appellants argue that he should have used the “net loss” method (sometimes called the “net investment” method) instead.

    To understand the difference between the two methods, imagine that three investors lose money in a Ponzi scheme. A invested $150,000 and withdrew $60,000 before the scheme collapsed, so his net loss was $90,000. B invested $150,000 but withdrew only $30,000; his net loss was $120,000. C invested $150,000 and withdrew nothing, so lost $150,000. Suppose the receiver gets hold of $60,000 in assets of the Ponzi scheme—one-sixth of the total loss of $360,000 incurred by the three investors ($90,000 + $120,000 + $150,000). We’ll call these recovered assets “receivership assets.” Under the net loss method each investor would receive a sixth of his loss, so A would receive $15,000, B $20,000, and C $25,000 . . . .

    Under the rising tide method, withdrawals are considered part of the distribution received by an investor and so are subtracted from the amount of the receivership assets to which he would be entitled had there been no withdrawals. (When there are no withdrawals, rising tide yields the same distribution of receivership assets as net loss.) In our example, the total of withdrawn plus receivership assets is $150,000, but there is only the $60,000 in such assets to distribute. A, having been deemed (as a consequence of the rising tide approach) to have “recovered” $60,000 before the collapse of the Ponzi scheme, is entitled to nothing from the receiver, as otherwise the remaining sum of withdrawals and receivership assets—a total of $90,000 ($30,000 in withdrawals, all by B, and $60,000 in receivership assets)—would be insufficient to bring the remaining investors up to anywhere near A’s level. . . .

    For the “tide” to raise B and C as close to A as possible, B has to receive $15,000 in receiver assets, for a total “recovery” of $45,000, and C the remaining receiver assets, giving him $45,000 too. The division of withdrawals plus receiver assets is then 60–45–45 . . . versus 75–50–25 under the net loss method.40

This discussion is unnecessarily hard to follow, because it fails to identify the choices being made in terms of recognizable rules—rules applied in other contexts as well. Observe first that the reference to “the total loss of $360,000 incurred by the three investors ($90,000 + $120,000 + $150,000),” like the expression “net loss method” itself, begs the question of how each investor’s loss is to be measured. It assumes that the money previously withdrawn from the scheme by A or B belongs to them—resulting in a net loss for each that is less than their original $150,000 investment. If this is so, it is not because A or B has recovered identifiable property or its product, but because they are allowed to assert an affirmative defense to the restitution claim of the victims as a class.41 Whether that defense should be allowed is the question to be decided. Observe next that the object of the “rising tide method” is to equalize the ultimate consequences of the fraud for A, B, and C—“to raise B and C as close to A as possible”—but subject to a very significant constraint. The vital but unstated qualification is that no investor will be asked to put his hand in his pocket to repay any of the amounts withdrawn. Apart from that limitation, the “rising tide method” operates by denying the availability of the affirmative defense in a case of restitution between fraud victims.

The hybrid nature of this new result is easier to see in the traditional formulation of the problem. If the defense of bona fide payee were simply not available between fraud victims it would be a relatively simple matter to equalize the consequences of the fraud between investors who withdrew funds and those who did not. This was “option (1).” In Judge Posner’s example, A and B would be liable prima facie to restore their withdrawals of $60,000 and $30,000, yielding total “receivership assets” of $150,000. Imposing this liability in restitution means that each of A, B, and C has lost $150,000; ratable distribution leaves them all in the same position when the matter is concluded. On a net basis, A would pay $10,000 to the receiver, B would recover $20,000, and C would recover $50,000.

By contrast, the orthodox approach—in this context, the “net loss method”—takes victims as it finds them, without trying to calculate losses until the music has stopped. Everything is on a net basis: funds previously recovered from the scheme are funds that have not been lost, just as identifiable assets (or their traceable product) coming into the hands of the receiver still belong to their original owners. “Rising tide” imposes a rough compromise between the two alternatives. Investors will not be required to reimburse withdrawals, but they do not own them entirely, either. To the extent that restitution can be made by book entry, without repayment out of pocket (the case of B in Judge Posner’s hypothetical), previous withdrawals will be recovered for distribution to the other fraud victims.

This hybrid solution is not as arbitrary as it might seem at first glance. In standard restitution doctrine, the alternative to allowing the bona fide payee defense is not to make the payee automatically liable to restore all amounts received; it merely requires the payee to prove that he suffered a change of position on receipt of the funds, making it inequitable to require him to refund the money.42 This would doubtless be the case with many Ponzi victims who have spent the money obtained by withdrawals from the scheme, yet it is understandable that courts and receivers would not welcome the task of applying this equitable defense on a case-by-case basis. To replace this more laborious, individualized equity, “rising tide” substitutes a rough presumption: that a liability to make restitution in cash would constitute an inequitable hardship, while a reduction in one’s net distribution from the receivership would not.

IV.     General Principles

The shift from old to new outcomes is not the result of an enhanced concern with fairness. Every judge wants to treat competing claimants fairly, and fairness is assumed to mean equal treatment—“equality is equity.” Equal treatment of claimants similarly situated means ratable distribution of available assets. No court has ever proposed doing anything else. But identifying the assets available for ratable distribution begins with a measurement of each victim’s loss, and it is here that recent decisions have been rewriting the rules. By rejecting (or simply ignoring) doctrines that allowed particular victims to maintain or reacquire ownership of misappropriated assets, courts have moved up the point at which those assets are irretrievably lost. The effect of this readjustment is to narrow the set of circumstances potentially alleviating each victim’s ultimate loss, thereby increasing the loss of some individual victims and (in consequence) the aggregate assets available for ratable distribution.

Until fairly recently, victim v. victim cases such as Elliott, Durham, and Huber would have been decided by a fairly mechanical application of property rules. So long as these rules were widely recognized, they fixed the point at which an owner either lost or recovered an asset of which he had been involuntarily dispossessed—thereby answering the question of what belonged to whom. So long as property rules decided these cases, the only “general principle” required was one of the oldest and most basic: that of giving property to its owner, suum cuique tribuere.43 Nobody questions that principle, but it has no application where property rules have been effaced, leaving ownership indeterminate. The general principles underlying the new outcomes are necessarily something different.

A.     Incentives

A reflexive response on the part of some observers is to ask whether the new results in Ponzi cases might be explained in terms of instrumental goals unrelated to fairness, such as reducing the aggregate Ponzi-scheme loss. In other words, the general principle guiding the current Ponzi decisions might be that judges seek intuitively to maximize social utility. But this does not appear to be what is going on. For one thing, the judges and their receivers do not explain their decisions that way: instead they say they make the allocations they do because they seem fairer.44 In the second place, most of the changes being made—the lines that are being redrawn—are ill-suited to serve as incentives to useful behavior. This is almost certainly the case with regard to the first two of the rule changes taking place—denying to the fortunate fraud victim the right to retake identifiable property (as in Elliott) or its logically identifiable product (as in Durham). For a given investor in a Ponzi scheme, the choice between old and new rules on this point is easy to make after the fact but meaningless ahead of time—depending as it does on the combined odds that the investment opportunity is fraudulent, that either the investor or some rival will by blind luck be able to identify his contribution in the wreckage of the scheme, and that he (rather than the rival) will be the one who can do so. Unforeseeable consequences will not create incentives, and it is implausible that the location of the property baseline along this axis will have ex ante effects that might serve to reduce Ponzi losses.

This inquiry into incentives might be more profitably applied to the third of the “baseline” issues, the question whether (and to what extent) an investor who withdraws funds from the scheme will be allowed to keep the money withdrawn. Here the problem is that the postulated pros and cons do not explain the tendency of the recent decisions. As we have just seen, the novel method of distribution known as “rising tide” is a compromise by which (a) innocent investors are still protected (to the amount of their original investments) against having to repay prior withdrawals in cash, but (b) amounts withdrawn must be restored to fellow victims, to the extent restitution can be accomplished by book entry—in that previous withdrawals will be charged against each investor’s share of the ensuing distribution. Consistent with the other “baseline” changes, this departure from previous orthodoxy qualifies the repaid investor’s ownership of the funds withdrawn, treating some part of them as “already lost.” Repayments made using other victims’ money may or may not be yours to keep.

Recent comments on this issue by incentive-minded observers propose new explanations of the standard rule of “net loss,” alias “bona fide payee,” whereby the innocent investor becomes the owner of stolen funds paid out to him. They do not explain why the recent tendency of the decisions in Ponzi cases has been in the other direction. Thus, Saul Levmore points to a connection between the availability of the affirmative defense and the intuition “that Ponzi schemes collapse because investors exit and that law might best minimize waste and losses by encouraging exit.”45 An innocent but sophisticated investor who begins to have doubts that Ponzi is on the up-and-up will be encouraged to exit earlier if he knows that the funds he manages to withdraw will belong to him in the same way as the funds he never invested. If the rule were the opposite—so that funds invested in a Ponzi scheme were irretrievably lost at the moment of initial investment—the doubtful investor might as well leave his money where it is and hope for the best. Exit is thereby delayed and aggregate losses increased. This is a novel defense of the orthodox result, but it is a result from which the judges currently adopting “rising tide” are for some reason backing away.

In the Huber decision, already discussed, Richard Posner advances a more elaborate hypothesis to distinguish those investors who have made previous withdrawals from the scheme from those who have not. Investors who have made withdrawals and “have spent the money they withdrew . . . may find themselves with all or most of their savings still in the Ponzi scheme.”46 Investors whose sole remaining asset is their share of any distribution to be made by the receiver “would tend to place a high marginal utility” on that prospective dividend; yet their share will be less if prior withdrawals are counted against it. In short, “[t]he more investors in a Ponzi scheme who would receive nothing under rising tide and might therefore have difficulty paying their future expenses, the more likely the net loss method is to maximize the overall utility of the investors.”47 The reliability of the correlation between previous withdrawals and present-day “high marginal utility” might be questioned, but it is ultimately a fashionable paraphrase of the orthodox rule (namely, that innocent payees are protected) in one of its most traditional justifications. Restitution allows an affirmative defense to any defendant who can prove a detrimental change of position on receipt of the benefit in question.48 When antecedent debt is treated as “value”—and particularly when the “value” represented by fraudulent obligations is measured at 100 cents on the dollar—it is because repaid creditors (like bona fide purchasers generally) are presumed to have suffered a change of position, typically by spending the money on something else.49 A subsequent, unexpected liability to restore the money would therefore be an inequitable result.

The rule that treats antecedent debt as value, “no questions asked,” rests on an implicit calculation that it is not worth the judicial effort to investigate the facts of change of position (or “high marginal utility”) on an individualized basis. The new “rising tide” decisions suggest that this calculation is being reconsidered. Lawyers and judges in today’s Ponzi cases no longer recognize that when they debate the status of prior withdrawals they are addressing the age-old issue of antecedent debt as value, so their reconsideration is taking place beneath the surface. But if the compromise implicit in “rising tide” is spelled out, it is entirely plausible in traditional, equitable terms. The orthodox rule of value takes the fact of payment as a proxy for change of position by the payee. The modified rule rejects that equivalence, but only insofar as the payee’s liability in restitution can be satisfied without imposing an affirmative obligation to repay the money withdrawn. The fact the money is already in the hands of a receiver, so that the receiver is going to have to decide one way or another how to divide it up, makes it easy to justify this modest step in the direction of individualized equity.

B.     Loss Sharing

The common law lacks any principle permitting the sharing of casualty losses, even in circumstances where splitting the difference might seem the fair thing to do.50 If such a principle existed at all, its limits would be quickly reached. When A’s house is destroyed by a meteorite and B’s house next door is left untouched, no one (presumably) would put B under a legal obligation to contribute to A’s loss. When A invests $1000 with Ponzi and B intends to do the same but forgets to mail his check before the scheme is exposed, no one (yet) would want to equalize fraud losses between A and B. But when B’s check is deposited after Ponzi’s bank account has been frozen—so that withdrawals have ceased before B’s funds are added to the commingled fund—courts following Elliott and Durham will now require B to share A’s losses pro rata.51

Standard rules that determine the incidence of loss between multiple victims of a common fraud yield some of the best examples of an all-or-nothing tendency that some commentators have long wished to reform. The fraudulent architect in Vickery v. Ritchie deceived both Owner and Builder about the cost of the Turkish bathhouse that it was Owner’s lifelong ambition to construct, inducing them to sign counterparts of a purported agreement showing different prices for the work: Builder’s copy said $35,000, while Owner’s copy said $20,000. 52 The facts came to light after the bathhouse had been completed at a cost of $35,000, adding only $20,000 to the value of the property. Owner had already paid this much, so he was under no further liability to Builder. Many people think that the better outcome would be to split the difference.53 One of the most useful features of Vickery v. Ritchie is the way it illustrates the absence from our law of any obligation that would achieve this result.54

Cases involving the fraudulent agent of two principals present a streamlined version of the same situation. Agent embezzles from A, creating a shortage in his books at A’s establishment. Later, facing imminent exposure, Agent embezzles from B and uses B’s money to make up the shortfall with A. When these facts come to light B seeks restitution from A. Courts decide for A or for B based on technical distinctions—whether A was aware of the deficiency at the moment of repayment, or whether Agent’s guilty knowledge should be “imputed” to one of the parties (though not to the other) as principal—that betray the inadequacy of the available remedies.55 The most vivid examples lie closer to the problem at hand because they involve no agency, merely fraud. Swindler borrows money from Victim 1 by means of some confidence trick: for example, by giving a forged note secured by a mortgage on someone else’s property. When V1 begins to press for payment, Swindler repeats his successful trick to borrow from Victim 2, then uses V2’s money to repay V1 (or even persuades V2 to pay V1 directly). Discovering these facts, V2 sues V1. The decisions divide along the property baselines, because they reveal conflicting ideas of when the loss in successive-fraud cases is actually incurred. A court that decides to allow the affirmative defense (and deny restitution) might say that “[V2’s] loss was occasioned solely by the forgery in the second loan transaction. This loss would have been sustained regardless of whether the first loan was or was not in existence at the time and also regardless of whether the first loan transaction was forged or genuine.”56 A court that denies the defense (and grants restitution) can respond that “the defendant [V1] had already lost its money, although it did not realize the fact.”57

If the right thing to do in these cases is to split the difference, what is needed is a principle imposing contribution between victims of a common loss—as opposed to the usual contribution between parties jointly liable to a third person. Contribution between co-obligors is familiar from the examples of co-sureties and joint tortfeasors, though of course the principle is much broader.58 Contribution between losers is harder to envisage, though some marginal instances might be found. “General average” in admiralty works this way, but the analogy between cargo owners and fraud victims seems distant.59 A rule for splitting losses seems to be operating beneath the surface of some well-known cases of frustrated contracts, in which houses undergoing repairs (or being moved from one site to another) are destroyed by fire before the work is completed. Some decisions allow the performing party in such cases to recover in restitution for benefits conferred, notwithstanding the evident fact of a total loss resulting in no benefit to anyone. It is easy to conclude that the judges are only finding benefits in order to divide losses.60 But here, too, the contractual links between the parties—not to mention the availability of insurance—make it hard to see lessons for the victims of a common fraud.

We can assume for the moment that a strong impulse to impose contribution between “similarly situated” casualty victims lies somewhere behind the recent Ponzi decisions, and that such an impulse is not new. The difference between old and new outcomes might then be that the current setting of the problem allows courts to avoid the necessity of stating a principle and defining its limits. The Turkish bathhouse, the fraudulent agent of two principals, and the repeated confidence trick are fundamentally different from the Ponzi case because each is litigated directly as a matter of private obligations between the interested parties. Defendant has the money, plaintiff wants it back. Courts can decide the outcome on an all-or-nothing basis by manipulating liabilities and defenses, but a decision splitting the difference requires a principle no one has identified.

In the Ponzi cases, by contrast, the money is in the hands of a receiver. Contribution between losers no longer requires that a new legal theory be discovered—as it would in the case of Builder v. Owner, Principal B v. Principal A, or V2 v. V1—merely that some unfamiliar doctrine be ignored altogether. Doing the latter is obviously much easier. The extraordinarily difficult task of finding limits to a principle of contribution between losers—who is a loser, and how much has he lost?—is finessed by the age-old rule that possession is nine points of the law. The losers subject to contribution are the people whose assets have come into the hands of the receiver, while the losers who got their money out escape any such liability.

C.     Luck and Blame

If we take the judges and receivers at their word, the predominant reason for changing the rules of victim v. victim restitution in Ponzi cases has been to reduce the role of luck in determining the outcome. In refusing to allow recovery of specific property or its identifiable product, courts deprecate the idea that more fortunate victims’ losses should be affected by “the merely fortuitous fact that the defrauders spent the money of the other victims first.”61 They resist any distinction between claimants that would rest “on the basis of the actions of the defrauders”62 or on the “chance” that some victims can identify their assets within “the common pool.”63 In choosing “rising tide” over “net loss,” courts are responding to the same perception: that disproportionate distributions may have been fortuitous, and that an equitable recovery should not “penalize claimants based on the timing of their investments.”64

New decisions in Ponzi cases start from a simple premise: that it would be inequitable that victims’ recoveries (and hence losses) “be skewed based on fortuity even though everyone was defrauded in the same manner.”65 Courts that respected traditional property baselines did indeed permit recoveries to be “skewed based on fortuity” in this sense. What explains the difference? The idea that judges of an earlier era were less concerned with equitable outcomes is not even worth considering, nor is there anything novel in the idea of ratable distribution. What is different is that courts have chosen to draw some related property baselines in different places. The consistent result of the changes is that they advance the moment of irretrievable loss for the fraud victim: from the point at which a lost asset can no longer be retrieved, back to the point at which the victim first parted with it. Many of the things that might happen along the line connecting those two points—such as the ability to prove that one’s stolen money has not yet been spent—are purely the result of luck. The reason to redraw the baseline is to eliminate the effect of such fortuities on the claimants’ relative recoveries.

The unfairness of permitting victim’s recoveries to “be skewed based on fortuity” is never examined in these opinions, being treated as self-evident. But if such results are unfair, it is only in the trite sense that “life is unfair.” Every human outcome is “skewed based on fortuity,” and the law leaves most such outcomes as it finds them. According to the traditional baselines, the fortuity that permitted A to identify half of his property in Ponzi’s possession was no less significant in determining property rights than the fortuity that led Thief to steal less of A’s gold and more of B’s silver. Modern judges have been less willing to take the parties as it finds them.

The difference might be that judges in the recent Ponzi cases see themselves, not as returning property to its owners, but as allocating losses. Losses are like liabilities—to be allocated, so far as possible, on the basis of fault. If Ponzi victims are to blame at all, it is only for their negligence (or greed) in handing over their money. In this respect they stand on the same footing: “each was defrauded in the same way.” Their subsequent luck in recovering the money cannot alter their respective entitlements, if those entitlements are fixed as a matter of comparative responsibility. What was once a problem in property has become a problem in tort.

Most legal disputes are concerned with allocating a loss. The predominant idea of modern American tort law—that of allocating losses based on comparative responsibility—has become the least common denominator and the default rule of American private law generally. It is the lawyer’s reflexive response to any problem for which known rules provide no other resolution.

The shrinking extent of the “special knowledge” that once governed private-law relationships has naturally expanded the range of the topics left to be regulated by “general principles.” The one general principle with which modern American lawyers are immediately familiar is the allocation of loss on the basis of fault. Applying less-familiar rules at the periphery of Contract, Property, or Restitution—or Agency, Suretyship, Trusts, or any other topic in private law—has always involved either special knowledge or a trip to the library. Peripheral problems will occasionally arise, as they always did. But if an unfamiliar problem seems to yield at once to a universal solvent—producing results consistent with the way we have all been taught to think—few lawyers will look any further.

V.     Conclusion: A Thought Experiment

Lawyers increasingly expect to find their common law in statutes. Property baselines that used to establish rights in victim v. victim cases have been especially susceptible to neglect because they were never written down in this way. The erosion of these common-law property rules can be seen more clearly if we compare them to a version of the same rules that was written down: rules that fix rights between defrauded owners and bona fide purchasers of goods. When Swindler acquires goods by fraud from Owner and sells them to Purchaser, the resulting problem of loss allocation (Owner v. Purchaser) is a case of victim v. victim restitution, much like those that arise in the Ponzi aftermath.

A famous English decision of half a century ago, Ingram v. Little,66 hints at the way these cases might be decided in the U.S. today if they were not covered by the Uniform Commercial Code. Three ladies in Bournemouth placed an advertisement in the local paper offering to sell their used motorcar: “1957 Dauphine, 3800 miles, sky blue, £725 o.n.o.” A rogue calling himself Hutchinson came to look at the car. He talked at length about himself, offered £700, agreed on £717, and persuaded the ladies to take a check. He informed them that he was a reputable person named P.G.M. Hutchinson, of Stanstead House, Stanstead Road, Caterham, and when the sellers found just such a person listed in the telephone directory they were convinced their visitor was telling the truth. Three days later, “Hutchinson” (now calling himself “Hardy”) sold the car to Little, a used-car dealer in Blackpool, who gave him £605 in trade against the price of a Ford Consul. Little was acting in good faith. The check for £717 was returned unpaid. Sellers sued Little for conversion.

All parties were defrauded by “the rogue Hutchinson.” Which of them should bear the loss? In 1960, there was still room in English law to argue that a successful imposture (like a “fraud in the factum”) might so entirely vitiate the contractual intent of a defrauded seller that the purported sale was no more than a theft—leaving the loss with Little, notwithstanding Little’s good faith. Indeed, that is how the case was decided. The U.C.C. nails down a resolution the other way, in favor of the purchaser, and it would do so even if we adjusted the facts of Ingram v. Little to eliminate any hint of negligence on the part of the sellers.67

Now imagine for a moment that the U.C.C. had not addressed this problem, and that the rules about bona fide purchase of goods were only vaguely remembered today—like the rules permitting Ponzi victims to retake their misappropriated money and securities. Might we not expect to see the loss of the sky blue Dauphine allocated between the Bournemouth ladies and Little on the basis of their comparative negligence in dealing with Hutchinson, or—supposing them equally blameless—split 50/‌50? This was exactly the proposal put forward in a dissenting opinion by Devlin, L.J.:

The true spirit of the common law is to override theoretical distinctions when they stand in the way of doing practical justice. For the doing of justice, the relevant question in this sort of case is not whether the contract was void or voidable, but which of two innocent parties shall suffer for the fraud of a third. The plain answer is that the loss should be divided between them in such proportion as is just in all the circumstances. If it be pure misfortune, the loss should be borne equally; if the fault or imprudence of either party has caused or contributed to the loss, it should be borne by that party in the whole or in the greater part.68

Theoretical distinctions that have been forgotten need no longer be “overridden.” The result is that it no longer requires the boldness of a Lord Devlin to adopt a reform with which people already feel comfortable.

The process by which less familiar legal problems are reconceived as torts reveals an atrophy of legal learning, to be sure, but it carries with it a potential solution to at least one longstanding problem. There are victim v. victim cases in which nearly everyone’s instinct would be to split the difference, if they could. The problems in Elliott, Durham, and Huber are recognizable versions of the victim v. victim paradigms: Ingram v. Little; Vickery v. Ritchie; the lawyer who robs Client B to repay Client A; the forger who “borrows” by giving a forged note Victim 2, in order to redeem an earlier forged note given to Victim 1. The orthodox, all-or-nothing results in these cases appear to be explained—at least some of the time—by the fact that judges lacked the tools needed to split losses.

If the problem of victim v. victim restitution in each of these cases is perceived instead as a matter of comparative fault, the difficulty disappears. Lord Devlin’s tort-like resolution of Ingram v. Little was radical—unthinkable—because the case involved property, not tort. It would still be property law in the United States today, but only because the U.C.C. keeps it there. Out beyond the reach of statutes, where problems not elsewhere classified are all coming to look like torts, the problem of imposing a sharing of casualty losses may have found a partial solution; though as an initial step in this development, only to salvaged assets that have come into the possession of the court.

  1. [1]. Oliver Wendell Holmes, The Common Law 54 (1881).

  2. [2]. See, e.g., Douglas Laycock, How Remedies Became a Field: A History, 27 Rev. Litig. 161, 173–74 (2008) (“As the [20th] century wore on, equity casebooks focused more on equitable remedies and less on substantive equity.”). See generally John H. Langbein, The Later History of Restitution, in Restitution: Past, Present, and Future 57 (W.R. Cornish et al. eds., 1998); Andrew Kull, Rationalizing Restitution, 83 Calif. L. Rev. 1191 (1995).

  3. [3]. Anyone who reads the newspapers will share this impression, though I am not aware of actual statistics. As a crude proxy, Westlaw reports that the name “Ponzi” appeared in 535 state and federal decisions during the first 65 years of Charles Ponzi’s notoriety—or before January 1, 1985—and in approximately ten times as many since then. Decisions referring to “Ponzi” have recently been added to the list at a rate of more than 600 per year.

  4. [4]. See, e.g., SEC v. Credit Bancorp, Ltd., 290 F.3d 80, 88 (2d Cir. 2002) (finding that the existence of a constructive trust at state law “does not defeat the equitable authority of the District Court to treat all the fraud victims alike (in proportion to their investments) and order a pro rata distribution”); United States v. Vanguard Inv. Co., 6 F.3d 222, 226 (4th Cir. 1993) (finding entitlement to restitution at state law, assuming it exists, may be disregarded by a court exercising its “discretionary power” of equitable receivership). Invoked in this way, the asserted discretion to reallocate state-law entitlements in the context of a federal receivership strongly resembles the “free-wheeling equitable discretion to cut down entitlements when they are sought to be enforced in a bankruptcy proceeding”—a discretion that does not exist:

    Bankruptcy is an equitable procedure, and “equality is equity” (and vice versa), as numerous bankruptcy cases intone. These truisms have a particular appeal for those bankruptcy judges who would like to administer the bankruptcy laws in accordance with their personal notions of fairness. But it is now well settled that although the origins, procedures, and many of the remedies of bankruptcy are indeed equitable, a bankruptcy judge has no authority to cut down the entitlements that creditors seek to enforce in bankruptcy, except as provided by the Bankruptcy Code itself.

    In re Stoecker, 179 F.3d 546, 551 (7th Cir. 1999) (Posner, C.J.) (citations omitted). If the idea of the present discussion were to criticize the recent Ponzi cases on the ground that they were incorrectly decided, a good way to start would be to demonstrate that a federal equity receiver has no more power to cut down state-law property entitlements than does a bankruptcy judge. My present purpose is different, as I am about to explain.

  5. [5]. “There are no hard rules governing a district court’s decisions in matters like these. The standard is whether a distribution is equitable and fair in the eyes of a reasonable judge.” SEC v. Enter. Trust Co., No. 08 C 1260, 2008 WL 4534154, at *3 (N.D. Ill. Oct. 7, 2008).

  6. [6]. For criticism along these more traditional lines see, for example, Restatement (Third) of Restitution & Unjust Enrichment §59 & cmt. g and accompanying Reporter’s Note (2011); Chaim Saiman, Restitution and the Production of Legal Doctrine, 65 Wash. & Lee L. Rev. 993, 998 (2008).

  7. [7]. See Restatement (Third) of Restitution & Unjust Enrichment § 58 (“Following Property into Its Product and Against Transferees”).

  8. [8]. See generally 2 James Schouler, A Treatise on the Law of Personal Property §43 (3d ed. 1896) (“[S]o long as one can identify his own chattels, and take them away, the ownership of articles need suffer no change because all happened to be lumped in one lot.”); Lionel D. Smith, Law of Tracing 70 (1997) (“[D]ivisible mixture raises an issue for resolution only if the mixture is such that it is impracticable to extract from the mixture the exact material contributed.”).

  9. [9]. See, e.g., Intermingled Cotton Cases, 92 U.S. 651 (1875); Spence v. Union Marine Ins. Co., [1868] L.R. 3 C.P. 427. The law of confusion and commingling of goods, putting the change of ownership at the point where identification is lost, is of unusually long standing: illustrations propounded by the drafters of the Uniform Commercial Code would have been instantly recognizable to the Roman jurists. Compare Dig. (Ulpian, Ad Edictum 16) (commingled lead and silver do not become common property because—unlike gold and bronze—they can be separated again), with U.C.C. §9-336(f) & cmt. 4 (2013) (in which SP1’s collateral, eggs, and SP2’s collateral, flour, are combined by Debtor to make cakes).

  10. [10]. Miller v. Race, 97 Eng. Rep. 398 (K.B. 1758).

  11. [11]. The two requirements are closely related. “Value” is a shorthand rule that stands as a proxy for the purchaser’s change of position. Reaching into one’s pocket to pay for a particular asset—the usual setting of “purchase”—involves a change of position of the kind the law seeks to protect. Happening to discover A’s asset X among B’s possessions does not. See Restatement (Third) of Restitution & Unjust Enrichment §68 & cmt. b.

  12. [12]. Id. §67 (“Bona Fide Payee”).

  13. [13]. A claim by a trustee or receiver to recover Ponzi payouts is frequently characterized, though imprecisely, as an action to avoid a fraudulent transfer. (This is a makeshift analysis, devised by courts that struggled to locate a common-law claim they no longer recognized within statutes that do not directly accommodate it. See Andrew Kull, Common-Law Restitution and the Madoff Liquidation, 92 B.U. L. Rev. 939, 958–62 (2012).) If Ponzi payouts are viewed as fraudulent transfers, the problem of “value” is brought into sharper focus, because one of the determinants of transferee liability (and of the transferee’s affirmative defense) is whether the debtor has received, not just value, but “reasonably equivalent value.” Uniform Fraudulent Transfer Act §8(1) (1984); see also 11 U.S.C. §548(a)(1)(B)(i), (c) (2012). Underlying the test of “reasonably equivalent value” is the elementary idea that unpaid creditors are not injured by a transfer that reduces debtor’s assets and liabilities pari passu. That test is not satisfied if the claims of some lucky Ponzi victims (having a liquidation value of, say, 5¢ on the dollar) have been redeemed at par, leaving only 3¢ on the dollar for the victims who made no withdrawals.

  14. [14]. See, e.g., U.C.C. §§ 1-204(2), 3-303(a)(3) (explaining that “value” includes antecedent debt); see also 11 U.S.C. §548(d)(2)(A); Uniform Fraudulent Transfer Act §3(a). Prior to the adoption of the Uniform Commercial Code, a creditor who took goods in satisfaction of an antecedent debt would not have been protected as a bona fide purchaser. See R.F. Chase & W.C. Crais, Annotation,Right to Follow Chattel into Hands of Purchaser Who Took in Payment of Pre-existing Debt,11 A.L.R.3d 1028, 1031 (1967). The question is still an open one for transactions not governed by statute. See Restatement (Third) of Restitution & Unjust Enrichment §68 & cmt. f.

  15. [15]. See U.C.C. §9-315(b)(2) & cmt. 3; Restatement (Third) of Restitution & Unjust Enrichment §59(2)(c) & cmt. d.

  16. [16]. SEC v. Elliott, 953 F.2d 1560, 1565 (11th Cir. 1992).

  17. [17]. Id. at 1568–69.

  18. [18]. Id. at 1569 (quoting the district court’s Order Establishing Final Plan).

  19. [19]. Id. at 1569–70.

  20. [20]. As all lawyers were once instructed:

    [W]here property is obtained by fraud, unless (as seldom happens) the fraud be of such a nature as to prevent the legal title from passing, the only legal remedy will be an action for damages against the party committing the fraud; but equity, by creating an equitable obligation, can and will follow the property itself (until it comes into the hands of a purchaser for value and without notice), and compel a specific restoration of it.

    C.C. Langdell, A Brief Survey of Equity Jurisdiction, 1 Harv. L. Rev. 55, 67 (1887).

  21. [21]. “Tracing” and “following” are not involved if the claimant seeks recovery of his original property. All that is needed is that the court decree the invalidity of the A-B transfer, thereby restoring legal title to A. This might be accomplished using the word “rescission,” but “cancellation” is just as good when a deed has been delivered, and “avoidance” will do as an all-purpose synonym. “Constructive trust” as a metaphor for what happens next is widely understood, but it is probably logically superfluous: once the A-B transfer has been invalidated, B (or C) is in unauthorized possession of A’s property and will have to hand it over, whatever the form of words.

  22. [22]. Moving the facts in SEC v. Elliott by a hair’s breadth makes the problem of property rights strictly a legal one. Suppose Elliott’s victims gave him bearer securities of different issuers (thus leaving them identifiable), and suppose they were persuaded that the transaction was not a “loan” at all but merely a deposit for safekeeping. There is no transfer in such a case, any more than if Elliott had acquired the securities by theft. When the facts came to light, the victims whose securities were still found on deposit would ask for them back; their remedy (replevin) would have nothing equitable about it. Yet a court that accepted the Elliott reasoning—by which “all investors were defrauded, all investors were cleverly persuaded to part with their securities”—would find it intolerable to deny restitution in one case and allow it in another, depending on whether or not a particular victim had been tricked into surrendering “legal title.”

  23. [23]. See, e.g., Melville M. Bigelow, Law of Fraud and the Procedure Pertaining to Redress Thereof 97–98 (1877) (“If the goods can be easily distinguished and separated, as in the case of a mixture of articles of furniture, no change of property takes place.” (footnote omitted)).

    Confusion of goods takes place when, upon the proof, it appears that the property of each of the parties interested cannot be distinguished. If the goods can be separated, no change of property takes place, even though the act of mixing was fraudulent. The doctrine is applied in cases where chattels, such as corn or wheat, not capable of being identified, owned by different persons, have been fraudulently intermingled by the wrongdoer. It is not in any case to be carried further than necessity requires.

    Claflin v. Beaver, 55 F. 576, 577–78 (C.C.S.D. Ohio 1893).

  24. [24]. United States v. Durham, 86 F.3d 70, 71 (5th Cir. 1996).

  25. [25]. Id.

  26. [26]. Id. at 72. “An FBI Special Agent traced $70,000 of Claremont’s payments that were deposited and never withdrawn.” Id. at 72 n.3.

  27. [27]. Id. at 72 (citations omitted) (quoting the district court’s Order Overruling Objections of Claremont).

  28. [28]. Id. at 73 (citations omitted).

  29. [29]. Cunningham v. Brown, 265 U.S. 1, 7 (1924).

  30. [30]. Id. at 11–12 (citations omitted).

  31. [31]. Among numerous cases relying on Cunningham for this conclusion, see, for example, SEC v. Infinity Group Co., 226 F. App’x 217 (3d Cir. 2007). “[Cunningham] held that all innocent victims should share equally in the recovered funds because equity demands equal treatment.” Id. at 218; see also Liberte Capital Grp. v. Capwill, 229 F. Supp. 2d 799, 803 (N.D. Ohio 2002) (“[W]here tracing places one party in a superior position over another victim, equity dictates tracing rules be suspended.”); SEC v. Credit Bancorp, Ltd., No. 99 CIV. 11395 RWS, 2000 WL 1752979, at *13 (S.D.N.Y. Nov. 29, 2000) (“[D]istribution should be done equitably and fairly, with similarly-situated investors or customers treated alike.”).

  32. [32]. CFTC v. Franklin, 652 F. Supp. 163, 169 (W.D. Va. 1986), rev’d on other grounds sub nom. Anderson v. Stephens, 875 F.2d 76 (4th Cir. 1989).

  33. [33]. See Kathy Bazoian Phelps & Steven Rhodes, The Ponzi Book: A Legal Resource for Unraveling Ponzi Schemes §20.04(3)(a) (2012).

  34. [34]. See supra Part II (discussing the fifth proposition).

  35. [35]. The earliest judicial mention of a “rising tide method” appears in CFTC v. Equity Financial Group, LLC, No. CIV.04-1512RBK AMD, 2005 WL 2143975 (D.N.J. Sept. 2, 2005). Discussing a method of accounting for previous withdrawals, the court notes that the receiver’s proposal is “called the ‘rising tide’ method”—but without saying who calls it that or why. Id. at *6. The name “rising tide” appears to have been proposed by the receiver in Equity Financial, Mr. Stephen T. Bobo, as a more accessible label for what the Franklin court (and others) had previously called “option (3).” See Memorandum in Support of Motion of Equity Receiver for Authority to Make Interim Distribution on Account of Investor Claims at 15–17, Equity Fin., No. CIV.04-1512RBK AMD (D.N.J. Jan. 7, 2005), No. 100. “Rising tide” is now an accepted usage, though the expression is not particularly intuitive: the sense only emerges once the effects of the various distribution methods have already been understood. Anticipating the explanation about to be offered, we might say that “rising tide” lifts some claimants’ boats (because it eliminates some competing claims to the receivership assets available for distribution), but not all of them (because previous withdrawals will now be charged against a recipient’s distribution share). On the other hand, it does not cause any boats to sink (because previous withdrawals need not be reimbursed). See infra notes 40–42 and accompanying text.

  36. [36]. See Janvey v. Adams, 588 F.3d 831, 834 (5th Cir. 2009); SEC v. Forte, Nos. 09-63, 09-64, 2009 WL 4809804, at *3 (E.D. Pa. 2009).

  37. [37]. Both Janvey and Forte recount the SEC’s opposition to the receivers’ attempts to recover previous payouts from innocent recipients. Janvey, 588 F.3d at 834; Forte, 2009 WL 4809804 at *3–4. Because the problem is quintessentially a matter of law and equity between investors, not investor protection against wrongdoers, it is not clear why the SEC should have an official view at all, let alone why it should adopt this one. An amicus brief submitted in Janvey makes the reasonable point that in some (perhaps most) cases, the expense of litigating to recover payouts from some investors will outweigh the benefits to be gained for others; but it does not explain why this is not a decision for the receiver to make. Brief of the Securities & Exchange Commission, Amicus Curiae in Support of Appellees, Janvey, 588 F.3d 831, No. 84.

  38. [38]. See Restatement (Third) of Restitution & Unjust Enrichment §67 & cmts. f, i and accompanying Reporter’s Note (2011) (citing cases).

  39. [39]. SEC v. Huber, 702 F.3d 903, 904 (7th Cir. 2012).

  40. [40]. Id. at 904–05.

  41. [41]. Another passage of Judge Posner’s opinion makes this point aptly:

    Our appellants argue against rising tide on the ground that they shouldn’t be penalized for having withdrawn some of “their” money. But it was not their money; they withdrew portions of the commingled assets in the Ponzi schemer’s funds. Those were stolen moneys, albeit stolen in part from the eleven appellants. An investor has no entitlement to money stolen from other people. When investors’ funds are commingled, none being traceable to a particular investor, no part of whatever funds are recovered [by the receiver] is property of any investor. Instead each investor is a creditor, and has merely a claim to a share of the funds that is appropriate in light of the relative size of his investment and other relevant circumstances.

    Id. at 906–07.

  42. [42]. See Restatement (Third) of Restitution & Unjust Enrichment §§65, 67 & cmt. b.

  43. [43]. Dig. (Ulpian, Ad Edictum 1) (“Iuris praecepta sunt haec: honeste vivere, alterum non laedere, suum cuique tribuere.”) (“The precepts of the law are the following: to live honorably, to harm no one, to give to each his own.”).

  44. [44]. Judge Posner’s decision in Huber makes a limited exception to this generalization, as presently discussed.

  45. [45]. Saul Levmore, Rethinking Ponzi-Scheme Remedies in and out of Bankruptcy, 92 B.U. L. Rev. 969, 982 (2012).

  46. [46]. SEC v. Huber, 702 F.3d 903, 907 (7th Cir. 2012).

  47. [47]. Id. Judge Posner does not suggest that such factors decide the question one way or the other, but rather merely that they might be worth considering. Huber affirms a trial court decision to use “rising tide” because “[t]he cases treat the receiver’s choice among allocation schemes as one within the discretion of the district court to approve or disapprove, like other aspects of the administration of a receivership.” Id. at 908.

  48. [48]. See Restatement (Third) of Restitution & Unjust Enrichment §65 (2011).

  49. [49]. See id. §68 & cmt. b.

  50. [50]. On the theme of splitting the difference vs. all-or-nothing in the common law, see generally, John E. Coons, Approaches to Court Imposed Compromise—The Uses of Doubt and Reason, 58 Nw. U. L. Rev. 750 (1963); R. H. Helmholz, Equitable Division and the Law of Finders, 52 Fordham L. Rev. 313 (1983); Joseph Jaconelli, Solomonic Justice and the Common Law, 12 Oxford J. Leg. Stud. 480 (1992); Gideon Parchomsky et al., Of Equal Wrongs and Half Rights, 82 N.Y.U. L. Rev. 738 (2007); W.F. Young, Half Measures, 81 Colum. L. Rev. 19 (1981).

  51. [51]. For a pre-Elliott decision to this effect, see CFTC v. Franklin, 652 F. Supp. 163 (W.D. Va. 1986), rev’d sub nom. Anderson v. Stephens, 875 F.2d 76 (4th Cir. 1989). This is the case previously cited, supra note 32, for its list of “options” for dealing with the problem of prior withdrawals. It was on this different point—the ownership of traceable, post-freeze bank deposits—that the District Court decision in Franklin was reversed. In the Fourth Circuit Court of Appeals in 1989 it was still obvious that B’s ownership of identifiable funds could not be lost if Ponzi’s peculations had ceased before B’s funds came into Ponzi’s hands. But a loss-sharing result becomes easier to explain if it comes to be accepted that “ownership” is a matter of equitable discretion, or if the contest between A and B is no longer seen as a question of property at all. See, e.g., SEC v. Infiniti Grp. Co., 226 F. App’x 217 (3d Cir. 2007).

  52. [52]. Vickery v. Ritchie, 88 N.E. 835, 835 (Mass. 1909).

  53. [53]. See, e.g., Warren A. Seavey, Problems in Restitution, 7 Okla. L. Rev. 257, 264–66 (1954).

  54. [54]. There is no contract between the parties, no tort, and no unjust enrichment of Owner beyond the increased value of his property (or any greater amount he might have agreed to pay). See Restatement (Third) of Restitution & Unjust Enrichment §10 illus. 27 (“Mistaken Improvements”). The erroneous decision to the contrary by the Massachusetts judges merely underscores their frustration with the limited tools at their disposal.

  55. [55]. See id. §67 illus. 25–26 and accompanying Reporter’s Note; Warren A. Seavey, Comment, Embezzlement by Agent of Two Principals: Contribution?, 64 Harv. L. Rev. 431 passim (1950).

  56. [56]. Cal. Pac. Title & Trust Co. v. Bank of Am. Nat’l Trust & Sav. Ass’n, 55 P.2d 533, 537 (Cal. Dist. Ct. App. 1936).

  57. [57]. Nat’l Shawmut Bank of Bos. v. Fidelity Mut. Life Ins. Co., 61 N.E.2d 18, 22 (Mass. 1945). For a more extensive collection of the conflicting opinions in successive-fraud cases, see Kull, supra note 13, at 942–45.

  58. [58]. See Restatement (Third) of Restitution & Unjust Enrichment § 23 (“Performance of a Joint Obligation (Indemnity and Contribution)”).

  59. [59]. Cargo owners might easily be seen as joint venturers (along with the ship owner), facing known perils in common. If their implied contract for general average was unsuitable for some reason, they could write a different one.

  60. [60]. See Restatement (Third) of Restitution & Unjust Enrichment §34 & cmt. d and accompanying Reporter’s Note; Philip D. Weiss, Comment, Apportioning Loss After Discharge of a Burdensome Contract: A Statutory Solution, 69 Yale L.J. 1054, 1060–69 (1960) (reviewing U.S. decisions that award restitution for “Pickwickian benefit” in frustrated contract cases).

  61. [61]. United States v. Durham, 86 F.3d 70, 72 (5th Cir. 1996).

  62. [62]. Id. These phrases, quoted from Durham, have been endlessly repeated in subsequent decisions.

  63. [63]. United States v. Vanguard Inv. Co., 6 F.3d 222, 227 (4th Cir. 1993).

  64. [64]. In re Receiver, No. 3:10-3141-MBS, 2011 WL 2601849, at *2 (D.S.C. July 1, 2011).

  65. [65]. CFTC v. Rolando, No. 3:08-CV-0064(MRK), 2008 WL 5225851, at *4 (D. Conn. Dec. 10, 2008). A nearly universal theme of the newer decisions is the idea that, since the claimants are “similarly situated,” it would be “unjust to allow one defrauded claimant to recover at the expense of another, merely because the former has the good fortune of being able to trace his or her funds.” SEC v. Credit Bancorp, Ltd., 194 F.R.D. 457, 463–64 (S.D.N.Y. 2000). Both the starting premise about who is “similarly situated” and the idea of “one claimant recovering at the expense of another” beg the question of who owns what.

  66. [66]. Ingram v. Little, [1961] 1 Q.B. 31 (Ct. App. 1960).

  67. [67]. The relevant provisions of the U.C.C. might almost have been written specifically to preclude the lengthy treatment devoted to the question in Ingram v. Little:

    A person with voidable title has power to transfer a good title to a good faith purchaser for value. When goods have been delivered under a transaction of purchase the purchaser has such power even though

    (a) the transferor was deceived as to the identity of the purchaser, or

    (b) the delivery was in exchange for a check which is later dishonored, or

    . . . 

    (d) the delivery was procured through fraud punishable as larcenous under the criminal law.

    U.C.C. § 2-403(1) (2013).

  68. [68]. Ingram, [1961] 1 Q.B. at 73–74.


Distinguished Senior Lecturer, University of Texas School of Law; Reporter, Restatement Third, Restitution and Unjust Enrichment. 

The encouragement of Ralph Brubaker, Ward Farnsworth, Henry Smith, and Lionel Smith is gratefully acknowledged.