In the wake of the 2007-2008 financial crisis, Congress enacted the Dodd-Frank Wall Street Reform and Consumer Protection Act, Pub. L. No. 111-203, 124 Stat. 1376 (2010), including provisions aimed at redressing the "complexity and opacity" of securitizations that it saw as preventing investors from adequately assessing risks in a securitized portfolio. S. Rep. No. 111-176, at 128-29 (2010). In § 941 of the act, 15 U.S.C. § 78 o -11, Congress directed the defendant agencies (plus two other banking agencies 1 ) to prescribe regulations to require "any securitizer" of an asset-backed security to retain a portion of the credit risk for any asset that the securitizer "transfers, sells, or conveys" to a third party, specifically "not less than 5 percent of the credit risk for any asset." 15 U.S.C. § 78 o -11(c)(1)(B)(i). The reasoning was that "[w]hen securitizers retain a material amount of risk, they have 'skin in the game,' aligning their economic interests with those of investors in asset-backed securities." S. Rep. No. 111-176, at 129. The agencies responded with the Credit Risk Retention Rule, 79 Fed. Reg. 77,601 (Dec. 24, 2014).

The Loan Syndications and Trading Association (the "LSTA") represents firms that serve as investment managers of open-market collateralized loan obligations ("CLOs") (a type of security explained in some detail below). 2 It challenges the agencies' decision, embodied in the rule, to apply § 941's credit risk retention requirements to the managers of CLOs ("CLO managers"). See 12 C.F.R. § 244.9 ; 17 C.F.R. § 246.9 . The LSTA's primary contention is that, given the nature of the transactions performed by CLO managers, the language of the statute invoked by the agencies does not encompass their activities. We agree.

We note by way of background that the LSTA initially petitioned for review of the rule in this court. We held that we lacked jurisdiction and transferred the case to the district court. Loan Syndications & Trading Ass'n v. SEC , 818 F.3d 716 , 724 (D.C. Cir. 2016). The district court granted summary judgment in the agencies' favor, finding that they could reasonably read § 941 to treat CLO managers as "securitizers." Loan Syndications & Trading Ass'n v. SEC , 223 F.Supp.3d 37 , 54-59 (D.D.C. 2016). The district court also rejected the LSTA's argument that the rule's methods for determining credit risk were arbitrary and capricious. Id. at 59-66 . The case has now returned to us on appeal of both rulings. Because we agree with the CLO managers that they are not "securitizers" under § 941, the managers need not retain any credit risk; we therefore need not address the risk calculation issue.

* * *

We review the Credit Risk Retention Rule for reasonableness under the familiar standard of Chevron, USA, Inc. v. NRDC, Inc. , 467 U.S. 837 , 104 S.Ct. 2778 , 81 L.Ed.2d 694 (1984), "which ... means (within its domain) that a 'reasonable agency interpretation prevails.' " Northern Natural Gas Co. v. FERC , 700 F.3d 11 , 14 (D.C. Cir. 2012) (quoting Entergy Corp. v. Riverkeeper, Inc. , 556 U.S. 208 , 218 n.4, 129 S.Ct. 1498 , 173 L.Ed.2d 369 (2009) ). Of course, "if Congress has directly spoken to an issue then any agency interpretation contradicting what Congress has said would be unreasonable." Entergy , 556 U.S. at 218 n.4, 129 S.Ct. 1498 .

The LSTA, rightly, does not suggest that Chevron is inapplicable due to the multiplicity of agencies. As they were authorized only to act jointly, and have done so, there is no risk that Chevron deference would lead to conflicting mandates to regulated entities. See Collins v. NTSB , 351 F.3d 1246 , 1252-53 (D.C. Cir. 2003). And there is nothing special to undermine Chevron 's premise that the grant of authority reflected a congressional expectation that courts would defer to the agencies' reasonable statutory interpretations. See Smiley v. Citibank (S.D.), N.A. , 517 U.S. 735 , 740-41, 116 S.Ct. 1730 , 135 L.Ed.2d 25 (1996).

As we are reviewing the district court's grant of summary judgment to the agencies and denial of summary judgment to the LSTA, our review of the district court is de novo. District Hosp. Partners, L.P. v. Burwell , 786 F.3d 46 , 54 (D.C. Cir. 2015).

The statute directs the agencies to issue regulations

to require any securitizer to retain an economic interest in a portion of the credit risk for any asset that the securitizer, through the issuance of an asset-backed security, transfers , sells , or conveys to a third party.

15 U.S.C. § 78 o -11(b)(1) (emphasis added). And Congress defined a "securitizer" as:

(A) an issuer of an asset-backed security; or

(B) a person who organizes and initiates an asset-backed securities transaction by selling or transferring assets, either directly or indirectly, including through an affiliate, to the issuer ....

Id. § 78 o -11(a)(3) (emphasis added).

The two key words in determining whether § 941 can be reasonably read to encompass CLO managers are "transfer" and "retain." The two subsections quoted above have the effect of authorizing requirements that an entity which transfers assets to an issuer retain a portion of the credit risk from the underlying assets that it transfers. In their ordinary meaning, words directing that one who "transfers" an asset must "retain" some interest in the associated risk refer to an entity that at some point possesses or owns the assets it is securitizing and can therefore continue to hold some portion of those assets or the credit risk those assets represent-that is, the entity is in a position to limit the scope of a transaction so that it transfers away less than all of the asset's credit risk. See, e.g., FDIC v. Meyer , 510 U.S. 471 , 476, 114 S.Ct. 996 , 127 L.Ed.2d 308 (1994) ("In the absence of [a statutory] definition, we construe a statutory term in accordance with its ordinary or natural meaning."); 2A Sutherland Statutory Construction §§ 47:28 -29 (7th ed. 2014) (similar).

But CLO managers do not hold the securitized loans at any point. Instead of being a financial institution originating or acquiring assets and then securitizing them, a CLO manager meets with potential investors and agrees to the terms of its performance as well as the risk profiles and tranche structures the CLO will ultimately take. See Wells Fargo & Company, Comment on Credit Risk Retention Proposed Rules 27 (July 28, 2011) ("Wells Fargo Comment"). The manager then directs a Special Purpose Vehicle ("SPV") (a corporation operating as the manager's agent-although the documents often define the manager as the SPV's agent, Argument Tr. 32-33) to issue notes in exchange for capital from the investors, the various notes reflecting the terms of the agreement and the kind and size of the investments. Board of Governors of the Federal Reserve System, Report to the Congress on Risk Retention 22-23 (Oct. 2010) ("Board Report"). Only then does the SPV-using the investors' money and operating at the recommendation of the manager-purchase the assets to securitize them. Wells Fargo Comment 27.

The loans underlying CLOs are very large loans made to already highly leveraged companies, often in the retail or manufacturing sectors of the economy. Usually no single bank originates the entirety of a loan. Rather, multiple banks "syndicate" under a lead arranger, each holding only a portion of the loan. Syndicated loans are "actively traded amongst financial institutions in a secondary market place," and purchased on these markets by a range of investors, including institutional investors, hedge fund managers, and, of course, CLO vehicles. Wells Fargo Comment 27. The number of syndicated loans in a CLO pool is typically small relative to other asset-backed securitizations. Ordinarily, a pool is made up of 100 to 250 loans, usually all made to moderate or large companies that generate a wealth of risk profile data for review by CLO managers and investors. See Securities Industry and Financial Markets Association, Comment to Joint Regulators on Credit Risk Retention Proposed Rules 67-68 (June 10, 2011); JPMorgan Chase & Co., Comment to Joint Regulators on Credit Risk Retention Proposed Rules 58-59 (July 14, 2011). But CLO managers neither originate the loans nor hold them as assets at any point. Rather, like mutual fund or other asset managers, CLO managers only give directions to an SPV and receive compensation and management fees contingent on the performance of the asset pool over time. See American Securitization Forum, Comment to Joint Regulators on Credit Risk Retention Proposed Rules 133-34 (June 10, 2011); U.S. Treasury Department, Report: A Financial System that Creates Economic Opportunities 102 (Oct. 2017) ("Treasury Report"); Board Report 22. The agencies do not question these characterizations of the CLO securitization model. See Appellees Br. 4-7; Board Report 22-23.

The agencies' interpretation seems to stretch the statute beyond the natural meaning of what Congress wrote; it turns " retain " a credit risk into " obtain " a credit risk. Even under Chevron , after all, agencies only "possess whatever degree of discretion [an] ambiguity allows." City of Arlington v. FCC , 569 U.S. 290 , 296, 133 S.Ct. 1863 , 185 L.Ed.2d 941 (2013) (quoting Smiley , 517 U.S. at 740-41 , 116 S.Ct. 1730 ). But the agencies assert some defenses, which we'll now explore, taking the key words in order.

"Transfer" ordinarily implies that a person with control over an asset via possession or ownership gives it or conveys it by sale to a transferee. E.g., Black's Law Dictionary 1727 (10th ed. 2014) ("to pass or hand over from one to another, esp. to change over the possession or control of"); 18 The Oxford English Dictionary 395 (2d ed. 1989) ("To convey or take from one place, person, etc. to another"; " Law. To convey or make over (title, right, or property) by deed or legal process").

The agencies point to dictionaries they claim support the idea that a third party can be said to "transfer" something if it is somehow the cause of a transfer between two other parties. E.g., The American Heritage Dictionary 1832 (4th ed. 2000) ("To convey or cause to pass from one place, person, or thing to another"). We think the natural reading of "cause" in these definitions means nothing more than that the verb "transfer" means to engage in the act of making a "transfer" (as a noun). Merriam-Webster's Dictionary of Law 495 (2d ed. 2011) (defining the verb as "to cause a transfer" and the noun as "a conveyance of a right, title, or interest in real or personal property from one person or entity to another" or a "passing of something from one to another"). But such a meaning doesn't necessarily encompass a person's playing any causal role in any transfer. Moreover, the possible ambiguity and meaning of statutory language depend on "the specific context in which [the term] is used, and the broader context of the statute as a whole." Robinson v. Shell Oil Co. , 519 U.S. 337 , 341, 117 S.Ct. 843 , 136 L.Ed.2d 808 (1997). Here the specific context is § 941's imposition of a risk retention rule on entities that " sell or transfer" assets to the issuer, indicating that Congress had in mind the ordinary sense of a conveyance between two parties, whether the conveyor was acting for financial remuneration ("sell") or was merely shifting economic value between or among an entity and its subsidiaries or affiliates ("transfer"). See also Freeman v. Quicken Loans, Inc. , 566 U.S. 624 , 634-35, 132 S.Ct. 2034 , 182 L.Ed.2d 955 (2012) (invoking the "commonsense canon ... that a word is given more precise content by the neighboring words with which it is associated" (quoting United States v. Williams , 553 U.S. 285 , 294, 128 S.Ct. 1830 , 170 L.Ed.2d 650 (2008) ). More broadly, § 941 is designed to reach those entities that "organize[ ] and initiate[ ]" securitizations " by transferring" assets to issuers. 15 U.S.C. § 78 o -11(a)(3)(B) (emphasis added). The language does not seem to apply to a person or firm that causes an SPV, whose value belongs to the investors, to make an open-market purchase from wholly independent third parties.

Besides seeming to reverse the apparent flow of the "transfer," the agencies' disregard of context leads them to embrace a reading of "transfer" that would include any third party who exerts some causal influence over a transaction. It would thus sweep in brokers, lawyers, and non-CLO investment managers who, though they play a part in organizing securities and "causing" the transfer of securitized assets, are clearly not the initiators of securitizations that Congress intended to regulate. That the agencies' interpretation sweeps so far beyond any reasonable estimate of the congressional purpose confirms our view that the interpretation is beyond the statutory language.

The agencies argue that § 941's qualification of "transfer" with the phrase "directly or indirectly" sufficiently broadens the term to cover CLO managers' activities. See 15 U.S.C. § 78 o -11(a)(3)(B). But it would be odd for those adverbs to eradicate the ordinary boundaries around the word "transfer." As the Supreme Court recently said in rejecting a government claim that a provision holding defendants liable for property "obtained, directly or indirectly" from certain crimes could make a defendant liable for property he never obtained at all, "The adverbs 'directly' and 'indirectly' modify-but do not erase-the verb 'obtain.' " Honeycutt v. United States , --- U.S. ----, 137 S.Ct. 1626 , 1633, 198 L.Ed.2d 73 (2017). Further, the immediate context of "directly or indirectly" suggests a meaning far narrower than the agencies' reading. Section 941(a)(3)(B) refers to one who organizes an asset-backed securities transaction "by selling or transferring assets, either directly or indirectly, including through an affiliate, to the issuer." 15 U.S.C. § 78 o -11(a)(3)(B). Congress's use of "through an affiliate" as an example of "directly or indirectly" (indeed, perhaps the only example contemplated by Congress at all) suggests that the adverbs were meant to assure that, despite the often complicated array of affiliates, depositors and SPVs that financial institutions use to create and sell asset-backed securities, the credit retention rule would reach a transferor no matter how many intermediaries it used. But to be covered by clause B of § 941(a)(3), the party must actually be a transferor, relinquishing ownership or control of assets to an issuer.

The meaning of "retain" in this context makes this point clear, as it lacks even the theoretical ambiguity that a person with no possessory interest could effect a "transfer." One occasionally uses "retain" colloquially to mean "acquiring something for the first time," as when one first gains legal counsel by "retaining" a lawyer. But the agencies conspicuously fail to offer a single real-world example of anyone ever using "retain" to encompass a process like the activity that the rule would require of CLO managers: going out into the marketplace and buying an asset they never before held.

The agencies argue that "retain" doesn't presuppose any voluntary anterior possession. It is of course true that one can keep something that one acquires under duress. Thus, say the agencies, so long as the rule's "requirement is continuing, and it requires 'retention' of risk on an ongoing basis," the rule fits a reasonable reading of Congress's text. Appellee's Br. 31 (citing Black's Law Dictionary 1509 ("To hold in possession or under control; to keep and not lose, part with, or dismiss.") ). But it is an astonishing stretch of language to read a mandate to "retain" to apply to one who would never hold the item at all apart from the mandate, with no congressional text mandating the prior acquisition.

Indeed, the record indicates the potentially serious consequences of shifting the statute's meaning from "retain" to "obtain." Recall that before the rule CLOs would cause SPVs to issue notes to investors and then use investor capital to purchase loans on the open market. Under the rule, CLO managers must now acquire investments that may not be suitable for them, necessitating significant amounts of capital that they may neither have nor have access to. Treasury Report 102. The agencies themselves identified a subset of managers operating between 2009 and 2013 with relatively poor access to capital, and concluded that "it would be reasonable to estimate that the exit of [those managers ] could impact current levels of capital formation by CLOs by 37 percent." 79 Fed. Reg. 77,730 /2. Some commentators forecast much larger reductions. See id. (discussing Oliver Wyman study). The resulting recourse of borrowers and lenders to other devices would presumably somewhat neutralize the net effect on overall supply of capital to the leveraged loan market. Id. at 77 ,730/2-3. That the agencies' reading of the statute would require non-transferring parties to obtain large positions in the relevant securities seems too large a surprise to have been intended.

* * *

So far we cannot see how the language of § 941 sustains the rule's application to managers of open-market CLOs. But the agencies make a special argument that requires a somewhat detailed response. If CLO managers are not covered by § 941, they contend, our interpretation "would do violence to the statutory scheme" and "creat[e] a loophole that would allow securitizers of other types of transactions to structure around their risk retention obligation." Appellee's Br. 20. Policy concerns cannot, to be sure, turn a textually unreasonable interpretation into a reasonable one. See General Dynamics Land Systems, Inc. v. Cline , 540 U.S. 581 , 600, 124 S.Ct. 1236 , 157 L.Ed.2d 1094 (2004) ; cf. Spectrum Pharm., Inc. v. Burwell , 824 F.3d 1062 , 1068 (D.C. Cir. 2016). Regardless of that, however, we think that the agencies overstate the supposed loophole.

The argument pressed by the agencies in their final rulemaking asserts that if CLO managers are not covered by § 941, then the statute is open to "easy evasion." 79 Fed. Reg. 77,655 /1. The theory runs that any securitizer could "evade risk retention by hiring a third-party manager to 'select' assets for purchase by the issuing entity that have been pre-approved by the sponsor," eviscerating the rule. Id. At the very least, the agencies argue, CLOs would become a type of securitization "without meaningful risk retention," contrary to Congress's purposes. Appellee's Br. 29. At worst, all other securitizations might conform themselves to the CLO structure and thus defeat Congress's purposes entirely.

We understand the agencies' plight, as they must give force to a statute that links the organization and initiation of a securitization to the transference of ownership interests in the assets that are securitized. Occasionally cases may arise, such as this one, in which those "organizing and initiating" the securitization do not do so "by transferring" the securitized assets to the issuer, while those that do transfer the assets are not the entities who organize or initiate the securitization in any meaningful way. However, if that is a "loophole," it is one that the statute itself creates, and not one that the agencies may close with an unreasonable distortion of the text's ordinary meaning.

In any event, we do not find the supposed loophole as worrisome as the agencies do. First, the loophole is to a considerable extent a problem of the agencies' own making. To see why requires us to trace several definitions through the statutes and regulations.

Congress required the agencies to craft a risk retention rule for "securitizers," defined as: (A) "an issuer of an asset-backed security," or (B) "a person who organizes and initiates an asset-backed securities transaction by selling or transferring assets, either directly or indirectly, including through an affiliate, to the issuer." 15 U.S.C. § 78 o -11(a)(3). Although the agencies commonly refer to the "issuing entity" of a security and define "issuing entity" as "the trust or other entity created at the direction of the sponsor that own or holds the pool of assets to be securitized, and in whose name the [securities] are issued," Credit Risk Retention: Proposed Rule, 76 Fed. Reg. 24,098 n.37 (Apr. 29, 2011) ; 12 C.F.R. § 244.2 ; 17 C.F.R. § 246.2 , they did not use that commonsense definition here. In fact, the agencies define the "issuer" of clause B as the issuing entity, but in clause A they noted that in multiple other regulations "the term 'issuer' when used with respect to an asset-backed security ("ABS") transaction is defined to mean the entity-the depositor-that deposits the assets that collateralize the ABS with the issuing entity." 76 Fed. Reg. 24,099 /1. Accordingly, the agencies do not interpret "issuer" in clause A to mean "issuing entity," but rather to mean "depositor." Id. The agencies in turn define "depositor" as either: (1) "[t]he person that receives or purchases and transfers or sells the securitized assets to the issuing entity"; (2) "[t]he sponsor" if the sponsor transfers assets directly to the issuing entity; or (3) the person that receives or purchases and transfers or sells the securitized assets to the issuing entity through a trust. 12 C.F.R. § 244.2 ; 17 C.F.R. § 246.2 . That is, the clause A "issuer" under the rule is not what an ordinary reader would think is an issuer but is rather the "entity that transfers assets to the issuer." Indeed, the second definition of a "depositor"-"the sponsor"-makes this clear. A "sponsor" is "a person who organizes and initiates a securitization transaction by selling or transferring assets, either directly or indirectly, including through an affiliate, to the issuing entity"-almost exactly the language of clause B. Id. In sum, the agencies have interpreted "issuer" in clause A of Congress's definition of a "securitizer" to be the exact same thing as the substance of clause B-the entity transferring assets to an issuer. See 15 U.S.C. § 78 o -11(a)(3). The agencies have thus dropped the actual issuing entity out of the statutory definition altogether.

The LSTA does not challenge the agencies' definition of an "issuer," so we need not wrestle with the question whether the agencies' interpretation mistakenly turns a statutory clause into mere surplusage. 3 Cf. Duncan v. Walker , 533 U.S. 167 , 174, 121 S.Ct. 2120 , 150 L.Ed.2d 251 (2001). The agencies contend that the disjunctive "or" in Congress's definition of a securitizer gives them discretion to elect which entity will retain the credit risk and that they have "elect[ed] to impose the risk retention requirements on the sponsors of open market CLOs rather than on the issuing entities." Appellee's Br. 35. That may be so, but that choice does not justify the agencies' stretching the definition of a "sponsor" to cover those who do not, as required by clause B, have a relationship to the assets such that one can reasonably say that they "transfer" the assets and could be required to "retain" a portion of the assets' risk. The protest that no securitizer can then be found to retain the risk arises because the agencies have defined a whole class of securitizers-the issuing entities-out of Congress's statute. To be sure, bringing them back in might make no useful difference here; it may be that here the only serious candidate for "issuer" is the SPV and that it holds the assets for the life of the securitization for the benefit of the investors, who are the very parties Congress sought to protect. See Argument Tr. 26-29. But addressing the ramifications of this would take us into far deeper waters than the briefs or rulemaking record explore, or than resolution of the case requires.

A second answer is that the agencies' feared hypothetical loophole is unlikely to materialize, even under their rather circuitous definitions. The agencies' concern is that banks or other financial institutions who actually organize securitizations will employ putatively third-party "managers" to sponsor securitizations without themselves transferring assets, thus creating "a situation in which no party to a securitization can be found to be a 'securitizer' because the party that organizes the transaction and has the most influence over the quality of the securitized assets could avoid legally owning or possessing the assets." 79 Fed. Reg. 77,655 /1. The agencies raised two scenarios at argument: one where regulated entities might try to get "bad assets" off their balance sheets by appealing to third-party managers to offload them, and one where ABSs might trade on a market and get re-securitized into ever more risky yet opaque offerings. Argument Tr. 35-37. But in the first scenario, where the actual organizer is the institution that holds the assets and "pre-approve[s]" the selections of "a third-party" manager, then both the rule and the statute logically apply. 79 Fed. Reg. 77,655 /1. The bank or financial institution that is actually calling the shots is organizing the securitization " by transferring " its assets and can be required to retain credit risks that it is already holding. In fact, the agencies' first hypothetical has provided a textbook example of sponsoring a securitization by indirect transfer through agents and intermediaries, the type of indirect transfer that is absent in this case, where CLO managers act as independent contractors with investors rather than as agents of an originating financial institution.

In the agencies' second scenario, they fear that re-securitized ABSs (the so-called "CDO squared," or collateralized debt obligations made up of other collateralized debt obligations), which became one of the more toxic defaulters of the financial crisis, would escape regulation because they often take a structure similar to that of CLOs: trading on an open market, purchased by third parties on behalf of an SPV that re-securitizes the purchased securities. Argument Tr. 35-36. But this concern ignores that many if not most of the primary securitizations that produce ABSs traded on these markets are now subject to the agencies' risk retention rule and the improved underwriting standards the agencies expect to flow from it. Moreover, any CLO so constituted must, by definition, have acquired the ABSs in open-market purchases at the direction of the investors.

Finally, to the extent other asset-backed securitizations were to truly take the form of open-market CLOs, it is highly doubtful that their falling outside the reasonable coverage of the statute need be a cause for concern. CLO managers are not direct or indirect agents of the originators, but rather negotiate their contracts with investors, usually with compensation based on performance. Open-market CLOs thus mitigate the problems Congress identified with the originate-to-distribute model: (1) The organizers' compensation dependency already gives them "skin in the game." Cf. S. Rep. No. 111-176, at 128-29. (2) Because they purchase relatively small numbers of unsecuritized loans on the open market through arm's-length bargaining, the activities of CLO managers present a far weaker version of the opacity that Congress identified in other ABS markets. Cf. id. And (3) both the superior incentives and relative transparency reduce the likelihood that such financing will generate anything like the decline in underwriting standards that the more famous ABS market is thought to have brought about. Perhaps for these reasons, CLOs weathered the financial crises relatively well. In contrast to 435 ABS collateralized debt obligations that defaulted, no more than six CLOs defaulted during the crisis, and all six included features atypical for CLOs and were eventually cured. Joint Appendix 527, 875, 1132 (citing Moody's Report, "Recent Performance of Market Value Collateralized Loan Obligations" (May 5, 2010); S&P, "Cash Flow and Hybrid CDO Event of Default Notices Received as of Jan. 25, 2011"; Jeremy Gluck, "CLOs versus CDOs" It's the 'L' That Matters," Moody's CLO Interest (July 2010) ). Indeed, Congress in its statutory scheme seems to be trying to achieve through regulation the incentives and transparency that the CLO market achieved through its business model. To the extent other securitizations seek to conform to this model, so much the better for the investors Congress seeks to protect.

In any event, the agencies' policy concerns cannot compel us to redraft the statutory boundaries set by Congress. Our commentary on those concerns only reinforces the reasoning that the ordinary meaning of § 941 does not extend to CLO managers. See K Mart Corp. v. Cartier, Inc. , 486 U.S. 281 , 291, 108 S.Ct. 1811 , 100 L.Ed.2d 313 (1988) ("In ascertaining the plain meaning of the statute, the court must look to the particular statutory language at issue, as well as the language and design of the statute as a whole."). The agencies have gone beyond the statute to require managers to "retain" risk by acquiring it. Even if their concerns about a policy loophole had merit, the statutory language does not support this radical shift in meaning.

* * *

The judgment of the district court is reversed and the case is remanded with instructions to grant summary judgment to the LSTA on whether application of the rule to CLO managers is valid under § 941, to vacate summary judgment on the issue of how to calculate the 5 percent risk retention, and to vacate the rule insofar as it applies to open-market CLO managers.

So ordered.