MAYER BROWN LLP, Formerly Doing Business as Mayer, Brown, Rowe & Maw LLP, Appellant v. INTERNAL REVENUE SERVICE, Appellee, 562 F.3d 1190

Opinion for the Court filed by Circuit Judge BROWN. BROWN, Circuit Judge: In this appeal, the parties dispute whether disclosure of certain Internal Revenue Service (“IRS”) settlement information could risk “circumvention of the law” — a category exempted from disclosure under the Freedom of Information Act (“FOIA”), 5 U.S.C. § 552. Id. § 552(b)(7)(E). On cross-motions for

summary judgment, the district court held the settlement information is covered by FOIA exemption 7(E). Because the disclosure of such information could risk circumvention of the law, we find the information falls within the FOIA exemption and therefore affirm the ruling of the district court. I. Background The FOIA request in this case involves lease-in/lease-out (“LILO”) arrangements between tax-exempt entities and taxable entities.

LILO arrangements occur when a tax-exempt entity owns and uses property but shifts significant tax deductions by signing a long-term lease with a taxable entity, reserving the right to cancel the lease. The tax-exempt owner then “subleases” the property back from the taxable entity. Arguably, the only purpose of the LILO scheme is to reduce the tax burden of the taxable entity, who otherwise has no ownership interest in the property. In 1999, the IRS disallowed deductions based on LILO schemes. In 2004, Congress made LILOs illegal, but the statute had only prospective effect. See 26 U.S.C. § 470.

The IRS continues to audit taxpayers engaged in LILO transactions and disallows the reported deductions. Many of these cases are settled. In 2004, Mayer Brown LLP filed a FOIA request for various information relating to the IRS’s LILO settlement practices. Through the course of litigation, the IRS turned over numerous documents, and the parties resolved almost all other issues by agreement.

The only remaining issue concerns one small class of information, described by the district court as “settlement strategies and objectives, assessments of litigating hazards, [and] acceptable ranges of percentages for settlement.” Mayer Brown v. IRS, No. 04-2187, slip op. at 3 (D.D.C. Nov. 28, 2006) (order granting motion for clarification). On cross-motions for

summary judgment and after reviewing the information in camera, the district court held that the IRS does not have to turn over the remaining settlement information because FOIA exemption 7(

E) applies. II. Discussion The sole issue in this case is whether the information requested by Mayer Brown meets FOIA exemption 7(E). That exemption states, in relevant part: (

b) This

section [i.e., mandatory FOIA disclosure] does not apply to matters that are ... (7) records or information compiled for law enforcement purposes, but only to the extent that the production of such law enforcement records or information ... (E) ... would disclose guidelines for law enforcement investigations or prosecutions if such disclosure could reasonably be expected to risk circumvention of the law. 5 U.S.C. § 552(b) (emphasis added).

Of the several requirements under exemption 7(E), the only one disputed by the parties is whether disclosure of the information “could reasonably be expected to risk circumvention of the law.” Id. “Circumvent” is nowhere defined in the statute; Webster’s defines the word as “[t]o avoid by or as if by passing around.” Webster’s II New College Dictionary 209 (3d ed., 2005). The breadth of the phrase “circumvention of the law” encompasses more than direct violations.

While using information to violate the law is one example of circumvention, it is also circumvention of the law to evade punishment after committing a violation. Thus, for instance, circumvention of the law includes both the person who violates the law and the fugitive who escapes the consequences of a previous violation. See United States v. Arias, No. 94-3011, slip op. at 1 (D.C.Cir. Mar. 25, 1995) (affirming enhancement of sentence because the defendant was a fugitive for two years after the initial criminal violation).

Used in a general way, “the law” encompasses both prohibitions against certain behaviors as well as the legally prescribed consequences for violations. Indeed, many statutes both define a violation and describe the applicable punishment. See, e.g., 26 U.S.C. § 7201 (“Any person who willfully attempts in any manner to evade or defeat any tax ... shall, in addition to other penalties provided by law, be guilty of a felony and, upon conviction thereof, shall be fined not more than $100,000 ... or imprisoned not more than 5 years, or both----”).

What constitutes circumvention varies, depending on the kind of prohibition and the enforcement mechanisms involved. A. Exemption 7(

E) shields information if “disclosure could reasonably be expected to risk circumvention of the law.” 5 U.S.C. § 552(b)(7)(E). If the FOIA request here sought a checklist used by agents to detect fraudulent tax schemes or the words most likely to trigger increased surveillance during a wiretap, the applicability of the exemption would be obvious. But enforcement of the tax laws, a largely self-policed obligation, depends heavily on the personal probity of taxpayers and the deterrent effect of severe and certain sanctions.

And, as a slew of high profile cases have recently demonstrated, compliance will often be delayed until enforcement (or unfavorable exposure) is imminent. Presumably, the importance of deterrence explains why the exemption is written in broad and general terms. It does not simply apply when information will definitely lead to circumvention of the law. The IRS does not have to prove that circumvention is a necessary result; the statute exempts information that would “risk circumvention of the law.” 5 U.S.C. § 552(b)(7)(E) (emphasis added). Showing a risk, of course, is a lower standard than showing a certainty.

But the statute does not stop there. Rather than requiring the IRS to prove a risk of circumvention, the statute exempts information that would “be expected to risk circumvention of the law.” Id. (emphasis added). Risk of circumvention is not required — only an expectation of such a risk. Moreover, this expectation of a risk of circumvention need not be undeniable or universal; the risk need only be “reasonably” expected. Id. (emphasis added). Adding a final strike against a rigidly narrow

interpretation, the exemption does not force the IRS to show this reasonably expected risk with certainty— only that disclosure “could reasonably be expected” to create such a risk. Id. (emphasis added). In short, the exemption looks not just for circumvention of the law, but for a risk of circumvention; not just for an actual or certain risk of circumvention, but for an expected risk; not just for an undeniably or universally expected risk, but for a reasonably expected risk; and not just for certitude of a reasonably expected risk, but for the chance of a reasonably expected risk.

The settlement guidelines requested here certainly meet this threshold as information that “could reasonably be expected to risk circumvention of the law.” Id. Although the settlement guidelines requested are not “how to” manuals for lawbreakers, the exemption is broader than that. Exemption 7(

E) clearly protects information that would train potential violators to evade the law or instruct them how to break the law. But it goes further. It exempts from disclosure information that could increase the risks that a law will be violated or that past violators will escape legal consequences. Though the information here does not necessarily provide a blueprint for tax shelter schemes, it could encourage decisions to violate the law or evade punishment. B. Tax evasion (like many crimes, to varying degrees) involves a cost-benefit analysis on the part of the law-breaker.

Information about acceptable settlement ranges quite clearly affects the cost-benefit analysis of potential evaders because it informs their economic calculus. Some potential evaders, upon learning the range of settlement percentages, may decide that the range is low enough to make evasion an appealing gamble. In this way, disclosure of the information can create an incentive for increased evasion.

For example, suppose hypothetically the requested information revealed that the IRS’s acceptable settlement range goes as low as 35% of the amount due, and a potential evader expects to gain a greater sum through an illegal tax scheme. Equipped with this information, the potential evader might decide the risk of a 35% settlement is low enough to gamble and violate the tax laws. Similar reasoning applies to other categories of information at issue.

Litigation hazards may include types of illegal tax shelters the IRS does not have the resources to pursue, situations which make witnesses unsympathetic or hard to find, or practical complications for investigating certain types of schemes. A potential evader who is made aware of the IRS’s perceived litigation hazards will know how to best structure an evasion so as to avoid the maximum enforcement efforts of the IRS.

Knowing how to evade in a way the IRS deems more difficult to detect or prosecute also enters into the cost-benefit analysis of a potential evader; a person with such knowledge may feel emboldened because she believes she can execute a scheme the IRS will be loathe to prosecute. Mayer Brown points out that Congress prohibited LILO transactions in 2004. Because, they argue, the 2004 statute will effectively eliminate all LILO arrangements, disclosure presents no danger. Of course, this argument ignores the limitless ingenuity of lawyers and accountants when it comes to tax shelters.

Returning to the example above, if a potential evader knows that the IRS’s acceptable settlement range goes as low as 35% for LILO arrangements, she might infer that a similar range would apply to a slightly different tax model. The LILO settlement ranges could be used as a baseline or a reference point for analogous schemes. Indeed, the IRS has estimated the existence of “almost 400 SILO transactions” (an arrangement very similar to LILOs) representing “claimed tax deductions in excess of 35 billion dollars” as evidence of the potential use of LILO settlement information for other schemes. J.A. at 140.

The information about settlement ranges for LILO transactions may enter into the cost-benefit analysis for other tax shelters and, in some cases, could convince potential evaders that a questionable scheme is worth the risk. Mayer Brown urges this Court to adopt an extremely narrow

interpretation of the exemption, arguing the IRS has a high burden to specifically prove how the law will be circumvented. We are aware the language of FOIA’s exemptions “must be narrowly construed.” Dep’t of Air Force v. Rose, 425 U.S. 352, 361, 96 S.Ct. 1592, 48 L.Ed.2d 11 (1976). But broad language — even when construed narrowly — is still broad language. Although some FOIA exemptions set a high standard, see, e.g., 5 U.S.C. § 552(b)(6) (requiring the agency to show “a clearly unwarranted invasion of personal privacy”), the text of exemption 7(

E) is much broader. In matters of statutory construction, the text is our primary guide. See Sierra Club v. EPA 536 F.3d 673, 679 (D.C.Cir.2008). Rather than requiring a highly specific burden of showing how the law will be circumvented, exemption 7(

E) only requires that the IRS “demonstrate!] logically how the release of [the requested] information might create a risk of circumvention of the law.” PHE, Inc. v. DOJ, 983 F.2d 248, 251 (D.C.Cir.1993). The IRS has logically shown how a risk of circumvention might result; Mayer Brown’s desire for an even narrower reading is simply not supported by the text.

In addition to potential future violators who may be encouraged to evade, those who have already evaded the tax laws (either through questionable lease transactions or other conduct) may use the information when deciding whether to come forward — a different but equally rational cost-benefit analysis, factoring in the degree of risk, the likelihood of enforcement, and other factors. If a tax cheat knows the IRS’s acceptable settlement ranges, and if her liability is substantial, she might decide the ranges are so high that it is better to remain in hiding.

If the consequences of confession seem especially harsh, a past evader may feel every effort must be undertaken to cover up the violation. If her efforts actually impede investigation or prosecution, not only has she circumvented the law by avoiding the legally prescribed consequences of her evasion, but she has also violated an additional law by obstructing justice. See, e.g., 18 U.S.C. § 1503 (prohibiting attempts “to influence, obstruct, or impede, the due administration of justice”).

The information about settlement goals, litigation hazards, and settlement ranges is just another factor in an evader’s cost-benefit calculation about whether to cooperate. Mayer Brown insists the term “circumvention of the law” applies only to future conduct. Thus, a disclosure that merely inhibits expiation for past conduct — a circumvention which has already occurred— does not implicate exemption 7(E). But even if we accept Mayer Brown’s premise that exemption 7(

E) is solely forward-looking, there is a logical flaw in its argument. Information that encourages a past violator to remain in hiding affects that violator’s decisions in the future. The fact that a tax evader has circumvented the law in the past does not mean she cannot also circumvent the law in the future by avoiding the legally proscribed consequences of her actions. The decision to evade the legal consequences of the initial violation is not just made once; it is a decision that is made anew when there is additional information.

If the disclosed information reveals severe costs in coming forward, it may influence future conduct, satisfying even Mayer Brown’s assumption that the exemption is only forward-looking. C. The consequences of a contrary

interpretation of exemption 7(

E) highlight the problem. In this case, companies using LILO schemes would love to have information about the IRS’s objectives of settlement, assessment of litigation hazards, and acceptable ranges for settlement. Why? Because this information would inform their cost-benefit analysis about the advantages of evading the law. Constructing a phony tax shelter may only be worthwhile if the IRS’s acceptable settlement range is below 80% of the tax liability. Once armed with (hypothetical) information that the IRS’s acceptable settlements are between 60% and 75%, a questionable tax scheme becomes viable.

Even a failure may be a win. And, once also armed with information about which cases the IRS does not like to litigate, the illegal tax shelter can be designed to minimize the chances of litigation or the likelihood of sanctions. In other analogous contexts, similar information could be sought.

Everyone seeking to minimize tax liability — not just LILO users — would love to have all of the IRS’s guidelines for all other schemes to learn when the agency is likely to seek enforcement versus when the agency views the specific context as too difficult to litigate because of concerns about lack of agency resources for certain types of cases, unsympathetic witnesses, or complicated investigation requirements.

Once equipped with information about settlement guidelines for all tax evasion schemes, particularly crafty evaders could plan their behavior to maximize the litigation hazards for the IRS and evade tax laws when, based on the settlement ranges, it was economically beneficial for them to do so. Outside of the tax context, corporations would request settlement information from the Antitrust Division at the Department of Justice, so they could know exactly which kinds of illegal transactions are likely to draw the enforcement wrath of DOJ.

Knowledge of settlement ranges would foster an accurate calculation of the risks and benefits of potential illegal transactions; this knowledge would inspire illegal conduct when, based on settlement information, such conduct proved to be economically rational. Criminal defendants, too, would love to have access to the settlement guidelines from their local U.S. Attorney’s office.

Information about likely ranges of plea offers would not only provide leverage to push, for a more lenient deal (for instance, by giving the defendant knowledge about when a prosecutor may be bluffing), but it would also enter into the ex ante cost-benefit calculus of soon-to-be criminals when deciding whether to break the law in the first place. All of these examples deal with information that could increase the risk of circumvention of the law.

Although other exemptions may cover such information as well— see, e.g., 5 U.S.C. § 552(b)(5) (exempting from disclosure “inter-agency or intraagency memorandums or letters” which meet certain requirements) — we need not decide whether other exemptions also apply. As the above examples illustrate, under the most sensible

interpretation of exemption 7(E), disclosure of settlement guidelines can increase the risk of circumvention of the law; reading the statutory language otherwise stretches any reasonable meaning out of the text. While there may be some legitimate uses of the requested information, the potential for misuse amply supports the IRS’s argument for exemption. People prepared to do the right thing would simply negotiate in good faith. Disclosure in this case would clearly be of enormous benefit to potential evaders and past violators hoping to escape punishment. FOIA exemption 7(

E) covers these circumstances, exempting disclosure when the information “could reasonably be expected to risk circumvention of the law.” Settlement guidelines fall squarely within this category, and, therefore, the information requested in this case is not subject to mandatory disclosure. The ruling of the district court is affirmed. So ordered. . Of course, we recognize ''[mjaterial like [discussion and digests of useful caselaw] is precisely the type of information appropriate for release under the FOIA.” PHE, Inc. v. DOJ, 983 F.2d 248, 251-52 (D.C.Cir.1993). Where the information is nothing more than garden-variety legal analysis, it does not fall under exemption 7(E).