Appeal from the United States District Court for the Northern District of Illinois, Eastern Division. No. 04 C 8211-Joan B. Gottschall, Judge.
The opinion of the court was delivered by: Easterbrook, Chief Judge
Before EASTERBROOK, Chief Judge, and MANION and KANNE, Circuit Judges.
Paul M. Daugerdas, a tax lawyer whose opinion letters while at Jenkins & Gilchrist led to the firm's demise (it had to pay more than $75 million in penalties on account of his work), designed a tax shelter for himself, with one client owning a 37% share. Like many tax shelters it was complex in detail but simple in principle, and to facilitate exposition we cover only its basics, rounding all figures.
Four entities played a role: Cemco Investors, a limited liability company that for tax purposes is treated as a partnership (so that its profits and losses are passed through to its owners); Cemco Investment Partners (the Partnership), Cemco Investors Trust (the Trust), and Deutsche Bank. (Forest Chartered Holdings, which Daugerdas controls, is Cemco's tax-matters partner. It handles Cemco's tax paperwork and has managed this litigation but did not participate in the underlying transactions. We do not mention it again.)
In December 2000 the Trust purchased from Deutsche Bank two options in euros, one long and one short. The long position had a premium of $3.6 million (nominally paid to Deutsche Bank) and entitled the Trust to $7.2 million if, two weeks in the future, the euro was trading at $.8652 or lower. (Back in 2000 the dollar was worth more than the euro.) The short position had roughly the same premium (nominally paid by Deutsche Bank to the Trust) and required the Trust to pay Deutsche Bank $7.2 million if, on the same date, the euro exchange rate was $.8650 or lower. In other words, if on the exercise date the euro was worth $.8652 or more, or $.8650 or less, the long and short positions would cancel out; but if the euro was trading for $.8651 on December 19, 2000, then Deutsche Bank would pay $7.2 million to the Trust. Like the options, the premiums matched almost exactly. The only money that changed hands was $36,000, which the Trust paid Deutsche Bank as the difference between the long and short premiums. And Deutsche Bank promised to refund $30,000 of this $36,000 if the options offset on the exercise date.
The Trust assigned its rights in these options to the Partnership, contributing some cash as well. The partnership spent about $50,000 to buy euros at an exchange rate of .89 dollars per euro. The next day Deutsche Bank remitted $30,000, because the options had offset. Soon the Partnership liquidated and distributed its assets (both dollars and euros) to the Trust. Next the Trust transferred to Cemco the i56,000 it had received from the Partnership. Finally, on the year's last business day, Cemco sold the euros for $50,000.
One would think from this description that the Trust (and the Partnership as its assignee) suffered a loss of $6,000-the net premium paid for options that cancelled each other out and hence lacked value on the exercise date-while Cemco had neither profit nor loss. Yet Cemco filed a tax return for 2000 showing a loss of $3.6 million! Its theory was that the Partnership's euros (later contributed to Cemco) had the same $3.6 million basis as the long option, and that a loss could be recognized once the euros had been sold. Cemco passed the loss to Daugerdas and his client, who reported it on their tax returns. What of the $3.6 million that Deutsche Bank "paid" the Trust for the short option, which almost exactly offset the long option? Well, Cemco took the view that this stayed with the Trust-and would never be taxed to the Trust, which, after all, had a net loss of $6,000-or if assumed by Cemco may be ignored because the options offset in the end. The purchase and sale of euros was the device used to transfer the basis of one option to Cemco while consigning the other option to oblivion.
A transaction with an out-of-pocket cost of $6,000 and no risk beyond that expense, while generating a tax loss of $3.6 million, is the sort of thing that the Internal Revenue Service frowns on. The deal as a whole seems to lack economic substance; if it has any substance (a few thousand dollars paid to purchase a slight chance of a big payoff) then the $3.6 million "gain" on one premium should be paired with the $3.6 million "loss" on the other; and at all events the deal's nature ($36,000 paid for a slim chance to receive $7.2 million) is not accurately reflected by treating i56,000 as having a basis of $3.6 million. The IRS sent Cemco a Notice of Final Partnership Administrative Adjustment disallowing the loss and assessing a 40% penalty for Cemco's grossly incorrect return. In the ensuing litigation, the district court sided with the IRS. 2007 U.S. Dist. LEXIS 22246 (N.D. Ill. Mar. 27, 2007).
Cemco says that in treating $50,000 of euros as having a $3.6 million basis, which turned into a loss when the euros were sold for exactly what they had been worth all along, it was just relying on Helmer v. CIR, 34 T.C.M. 727 (1975), and a few similar decisions. That may or may not be the right way to understand Helmer; we need not decide, for it is not controlling in this court-or anywhere else. The Commissioner has a statutory power to disregard transactions that lack economic substance. Compare Gregory v. Helvering, 293 U.S. 465 (1935), with Frank Lyon Co. v. United States, 435 U.S. 561 (1978). And the IRS has considerable latitude in issuing regulations that specify sorts of transactions that may be looked through. (These regulations avoid the need to litigate, one tax shelter at a time, whether any real economic transaction is inside the box.) A few months before Daugerdas set up this tax shelter, the IRS had issued Notice 2000--44, 2000--2 C.B. 255, alerting taxpayers to its view that Helmer could not be relied on, that purported losses from transactions that assigned artificially high basis to partnership assets would be disallowed, and that formal regulations to this effect were on their way. One of the transactions covered in Notice 2000--44 is the offsetting-option device.
Getting from a warning to a regulation often takes years, however, and did so here. Treasury Regulation §1.752--6, 26 C.F.R. §1.752--6, was issued in temporary form in 2003. T.D. 9062, 2003--2 C.B. 46. Two more years passed before the temporary regulation was made permanent. T.D. 9207, 2005--1 C.B. 1344, 70 Fed. Reg. 30334 (May 26, 2005). This sets the stage for Cemco's principal argument. It concedes that Notice 2000--44 and Treas. Reg. 1.752--6 scupper the entire class of offsetting-option tax shelters.
The regulation does this by subtracting, from the partner-ship's basis in an asset, the value of any corresponding liability. Thus if Cemco's basis in the euros comes from the premium for one option, then the premium for the offsetting option must be subtracted. That gets the basis back to roughly $50,000 (the value of the euros) $6,000 (the difference in the premiums). But as Cemco sees things the notice lacks legal effect, while the regulation cannot be applied retroactively. One district court has held that Treas. Reg. §1.752--6 does not affect transactions that predate it. Klamath Strategic Investment Fund, LLC v. United States, 440 F. Supp. 2d 608 (E.D. Tex. 2006). We disagree with that conclusion.
The regulation could not be more explicit: "This section applies to assumptions of liabilities occurring after October 18, 1999, and before June 24, 2003." Treas. Reg. §1.752--6(d)(1). (Transactions after June 23, 2003, are governed by the more elaborate 26 C.F.R. §1.752--7.) Why October 18, 1999? Because, although regulations generally do not apply to transactions that occur before the initial publication date of a draft regulation, see 26 U.S.C. §7805(b)(1)(C), the norm of prospective application "may be superseded by a legislative grant from Congress authorizing the Secretary to prescribe the effective date with respect to any regulation." 26 U.S.C. §7805(b)(6). Section 309 of the Community Renewal Tax Relief Act of 2000, Pub. L. 106--554, 114 Stat. 2763A--587, 638 (2000), enacts basis-reduction rules for many transactions and authorizes the IRS to adopt regulations prescribing similar rules for partnerships and S corporations. Section 309(d)(2) of the 2000 Act adds that these regulations may be retroactive to October 18, 1999. That's the power the Commissioner used when promulgating Treas. Reg. §1.752--6.
The district court in Klamath did not doubt that retroactivity could rest on the 2000 Act; Treas. Reg. §1.752--6 applies to partnerships (and LLCs treated as partnerships) a rule "similar" to the approach that Congress adopted for other business entities. Klamath held, however, that when promulgating Treas. Reg. §1.752--6 the IRS had not availed itself of that power. But if the IRS was not using that authority, why in the world does the regulation reach back to October 18, 1999? Retroactivity requires justification; to make a rule retroactive is to invoke one of the available justifications; and the choice of date tells us that the justification is the one supplied by the 2000 Act (in conjunction with §7805(b)(6)). A regulation's legal effect does not depend on reiterating the obvious. So Treas. Reg. §1.752--6 applies to this deal and prevents Cemco's investors from claiming a loss. Cemco is scarcely in a position to complain-not only because this tax shelter was constructed after the ...