THE era of civil litigation to recover the proceeds of fraud continues and, with it, the elaboration of principles of judge-made law as to when defrauded persons may recover. These principles reflect a truth of human nature. Apart from the silly, reckless and brazen, those who commit fraud wish their takings to be undetectable. A crucial development in the history of English law, therefore, was to accept that, for some purposes, persons may make juridical claims not only to an asset originally taken from them, but also to other assets representing the original: Maitland, “Trust and Corporation”, in Selected Essays (Cambridge, 1936), 171–72. The concept of assets “representing” the original assets is at the heart of the equitable and common law doctrines of tracing and civil recovery for fraud.

At issue in Federal Republic of Brazil v Durant International Corp. [2015] UKPC 35; [2016] 1 A.C. 297 was how ready a court will be to conclude that assets in a defendant's name, possession or control represent the plaintiff's original assets. The nominal plaintiff was the Federal Republic of Brazil, the active plaintiff being the Municipality of São Paulo. The Municipality had suffered fraud. Its then mayor, Mr. Maluf, secretly received bribes in person and through agents. The bribe moneys were converted into US dollars and paid into an account, with a New York bank, controlled by Mr. Maluf's son (“the C account”). Ultimately, the plaintiffs' claims for recovery were directed elsewhere. Messrs Maluf senior and junior were not the defendants, nor did the plaintiffs claim moneys or assets from them. The plaintiffs instead claimed that Durant, a company registered in the British Virgin Islands, received US$13,500,000 into its bank account in Jersey (“the D account”) as the proceeds of the bribes. They further alleged that US$10,500,055.35 of that money now stood in the Jersey bank account of another company registered in the British Virgin Islands, Kildare (“the K account”) and that the US$10,500,055.35 were traceable to the fraud practised on the Municipality.

The import of the case lies less in the plaintiffs' victory and more in their success in repelling two arguments which the Privy Council found “[c]onceptually … coherent and … supported by a good deal of authority” (at [18]). Both arguments are familiar from the textbooks. First, the defendants invoked so-called “backward” tracing. The defendants said the plaintiffs could only recover US$7.7 m because the last of the identified proceeds of bribery came into the D account after the last of the amounts making up the US$10,500,055.35 in the K account had been paid from the D account into the K account. How could Kildare hold the difference of US$2.8 m on constructive trust for the plaintiffs when Kildare was known to have held that sum prior to Durant's receipt of a corresponding sum of the identifiable proceeds of bribery? Secondly, the defendants invoked the so-called “lowest intermediate balance rule”. Only US$7.7 m could be traced, they said. For, after the proceeds of the bribes were paid from the C account into the D account where they were “mixed” with moneys derived from elsewhere, US$2.8 m were transferred back to the C account; the traceable proceeds of the bribes were diminished to that extent; thus, the plaintiffs could not recover above the lowest intermediate balance of US$7.7 m. Though these arguments were conceptually coherent and supported by authority, the Privy Council did not hesitate to reject them on the facts and to agree with the Jersey courts below (at [38]):

The plaintiffs could accordingly trace and recover the full US$10,500,055.35 in the K account. It can be expected that courts of England and Wales will follow this decision, though not strictly bound by it: see Brazil v Durant, at [18]–[27], [41]; Willers v Joyce [2016] UKSC 44; [2016] 3 W.L.R. 534, at [16].

To conclude that backward tracing is now available generally and that the lowest intermediate rule has been bidden goodbye would be easy – and wrong. The Privy Council limited its decision. Negatively, it denied “the argument that there can never be backward tracing, or that the court can never trace the value of an asset whose proceeds are paid into an overdrawn account” (at [40], emphasis added). By implication, their Lordships also denied that the lower intermediate balance rule necessarily limits claimants' entitlements to trace.

The Board also affirmatively limited the cases in which a claimant may trace backward – or into an overdrawn account or despite the lowest intermediate balance rule. Where a claimant seeks to trace backward, the claimant must establish “a close causal and transactional link between the incurring of a debt and the use of trust funds to discharge it” (at [34]). This must be judged by “look[ing] at the transaction overall”: if the “true overall purpose and effect” of, say, a series of interconnected transactions is that an asset or money should pass from one person to another, “the strict order in which associated events occur” will not, in law, bar a finding that the transferee holds the traceable proceeds of property transferred by the transferor (at [37]–[38]). Claimants seeking to trace into an overdrawn account despite the lowest intermediate balance rule will face a similar enquiry. In each case, the intentions of the parties to the apparent – and real – transaction will be central: Relfo Ltd. (in liq.) v Varsani [2014] EWCA Civ 360; [2015] 1 B.C.L.C. 14, at [57], [63]; R.C. Nolan, “The Administration and Maladministration of Funds in Equity”, in P.G. Turner (ed.), Equity and Administration (2016), ch. 4, 85–87; Cutts (2016) 78 M.L.R. 381, 397–404. However, the principles against tracing backward, or into overdrawn accounts or despite the lowest intermediate balance rule, will evidently continue to operate in all cases outside these affirmative limits.

When common law judges innovate, their method encourages them to explain the innovations in familiar terms and concepts, thus dimming the light of change. Nothing would be gained by pretending that Brazil v Durant makes no new law: two legal rules previously thought to limit tracing claims must henceforth be understood as limits which may or may not apply, depending on the facts. But this development is nonetheless consistent with the fundamental principles of tracing doctrine, and the concept at the heart of tracing: that an asset in the defendant's name, possession or control “represents” an asset to which the claimant was originally entitled.

How can the Privy Council have accepted the possibility of backward tracing without overthrowing fundamental principle or cultivating anomalies? Before Brazil v Durant, the weight of English authority supported the view that a claimant may only trace “forward”: that is, into assets acquired (1) subsequent to the misapplication of the claimant's assets and (2) by misapplying those assets. The causal requirement in (2) is undisturbed by Brazil v Durant. Not so the requirement in (1). However, too much should not be made of that fact. The Board declined a wider submission that a victim of fraud ought to be able to trace into “whatever” the defendant acquired “in exchange for incurring the debt”: cf. Smith [1995] C.L.J. 290, 292–95. Were that the principle, a claimant could trace into assets that a defendant bought on credit without then having intended to discharge that debt with the claimant's money. For instance, a claimant could trace into a car previously bought by a defendant on credit before the defendant formed an intention to pay off the debt with the claimant's money. That might be thought of as “true backward tracing”. Whatever the defendant's intentions in misappropriating another's money, the intention could not be to acquire the ownership of the car: the defendant already owned it: Nolan, ibid., at p. 89. Buying the car and taking the claimant's money are not part of a single co-ordinated transaction.

The “backward tracing” permitted by Brazil v Durant is consonant with fundamental principle because it would not permit tracing into the car in that example – and because it is scarcely backward tracing at all. Under Brazil v Durant, only assets to be acquired in the execution of a transaction may be traced into. Since transactions are by definition intended, it is a tautology – yet true – that a claimant may only trace into an asset acquired before it is paid for where the asset was intended to be acquired in the performance of a transaction: Relfo, at [63]; Nolan, ibid., at pp. 85–92; Cutts, ibid., at pp. 397–404. The utility of describing that as backward tracing is questionable. It is more accurate to describe Brazil v Durant as deciding that the assets into which a claimant may trace because they “represent” an original asset – or its traceable proceeds – are defined by the scope of the transaction, unlimited by accidents of the order and timing of the events by which the transaction is performed.