Appeal from the judgment of the United States District Court for the Southern District of New York (Victor Marrero, J.) granting on remand the Government’s petition to enforce two Internal Revenue Service summonses, one sent to Fridman in his personal capacity and one sent to him in his capacity as a trustee, based on the foregone conclusion and collective entity exceptions to the Fifth Amendment’s self-incrimination clause. We agree with the district court that the Government has shown with reasonable particularity the documents’ existence, Fridman’s control of the documents, and an independent means of authenticating the documents such that the foregone conclusion doctrine applies. We also agree with the district court that, as a matter of first impression in our Circuit, a traditional trust is a collective entity subject to the collective entity doctrine. Accordingly, we affirm the grant of the Government’s petition to enforce the summonses. Affirmed. ROSEMARY POOLER, C.J. Natalio Fridman appeals from the judgment of the United States District Court for the Southern District of New York (Victor Marrero, J.) granting on remand the Government’s petition to enforce two Internal Revenue Service (“IRS”) summonses, one sent to Fridman in his personal capacity and one sent to him in his capacity as a trustee, based on the foregone conclusion and collective entity exceptions to the Fifth Amendment’s self-incrimination clause. We agree with the district court that the Government has shown with reasonable particularity the documents’ existence, Fridman’s control of the documents, and an independent means of authenticating the documents such that the foregone conclusion doctrine applies. We also agree with the district court that, as a matter of first impression in our Circuit, a traditional trust is a collective entity subject to the collective entity doctrine. Accordingly, we affirm the grant of the Government’s petition to enforce the summonses. BACKGROUND The IRS has long been investigating the use of offshore bank accounts to improperly conceal federally taxable income. As part of these efforts, the IRS sought to investigate Fridman for the 2008 tax year. While Fridman reported only three foreign financial accounts maintained in a personal capacity for the 2008 tax year, the IRS became aware of at least five additional personal accounts: a UBS account in Switzerland; a Credit Suisse account in Switzerland; a Bank Leumi account in Switzerland; a Bank Leumi account in Israel; and a Bank Safra account in Luxembourg (collectively, the “Personal Accounts”). The IRS learned of these accounts in 2012 when Fridman’s representative provided the IRS with these accounts’ statements or other bank records pertaining to the 2008 tax year. From these documents, the IRS identified the account numbers and other information related to the Personal Accounts. Besides the Personal Accounts, the IRS became aware of seven companies affiliated with or controlled by Fridman that maintained foreign bank accounts. The IRS has identified seventeen such corporate foreign bank accounts (collectively, the “Corporate Accounts”).1 The IRS learned of these accounts through Fridman’s filing of Foreign Bank and Financial Accounts (“FBARs”). A person or entity is required to file an FBAR if they have “a financial interest in, or signature or other authority over” a foreign financial account. 31 C.F.R. §1010.350(a). In 1991, Fridman filed FBARs for the Consist Teleinformatica Argentine, Consist Consultoria Systemast Repre, and Wanstat Systemar DE Computacao CTDA accounts. In 1998, he filed FBARs for the Consist France and the Consist Asia Pacific accounts. Fridman filed FBARs for one Mak Data System account in 1998 and from 2000-2004, and for a second Mak Data System account in 1998 and from 2000-2005. For four other Mak Data System accounts, he filed FBARs in 2004 and 2005. For the Consist International Inc. account, Fridman filed FBARs from 2000-2003. Finally, the IRS learned that Fridman controls a number of trusts, including at least one domestic trust that has a foreign financial account. This trust is the David Marcelo Trust, named for Fridman’s son, of which Fridman is both a trustee and beneficiary. Fridman’s sister is a second trustee, and Fridman’s son is a second beneficiary. Though Fridman has not filed a tax return for the David Marcelo Trust, the IRS learned that the David Marcelo Trust has an HSBC account in Switzerland (the “HSBC Account”). The IRS came to know of this account after Fridman’s representative provided some bank statements and records related to the account in 2012 and 2013. One record shows a transfer of $2.4 million from the HSBC Account into a domestic Citibank account belonging to Fridman’s wife and son. In 2013, the IRS issued the two summonses (“the Summonses”) that are central to this case. One summons was sent to Fridman in his personal capacity, and the other was sent to him in his capacity as Trustee of the David Marcelo Trust. The documents sought (the “Requests”) were the same, and they largely pertain to the Personal Accounts, Corporate Accounts, and HSBC Account (collectively, the “Known Accounts”). The relevant Requests, as revised during the proceeding below, are: Request 1: For all of the foreign bank accounts listed on “Exhibit B” [a list of 24 accounts]2 over which you have signatory authority since 1999, please provide existing documents sufficient to show the opening date and closing date of each. Request 2: Please provide the 2006 and 2007 bank statements for all your foreign bank accounts listed on “Exhibit B.” Request 3: All bank statements and all the account opening documents, including but not limited to, Know Your Customer Account information including signature cards, opening deposit slips, passport copies, certificates of beneficial ownership, letters of reference, certificates of clean funds and/or other source of funds documentation for accounts held under the name of Consist Teleinformatica Argentine; Consist Consultoria Systemast Repre; Wanstst Systemar DE Computacao CTDA; Consist France; Consist Asia Pacific; Mak Data System; Consist International Inc. from opening date to 1/31/2009. Request 4: All bank statements and account opening documents, including but not limited to, Know Your Customer Account information including signature cards, opening deposit slips, passport copies, certificates of beneficial ownership, letters of reference, certificates of clean funds and/or other source of funds documentation for the following accounts from opening date to 1/31/2009: UBS 29, Credit Suisse 3, Leumi Bank 49, Leumi Bank 02, Safra 37, HSBC 15, Republic National Bank of New York (Suisse) S.A. Request 7: Existing bank documents including but not limited to cancelled checks, wire transfer instructions, wire transfer slips, deposit slips that show the flow of the following funds, the account numbers and account holders’ names for the Citibank accounts and any other bank accounts, sufficient to show the flow of funds in five transactions transferring funds between the HSBC account and an unknown New York-based Citibank account on 7/26/2004, 8/30/2004, 9/28/2004, 1/31/2008, and 2/28/2010. Request 11: Please refer to the Trust Agreement dated April 30th, 1990 — “the Grantor hereby transfers to the Trustees the property described in the annexed Schedule A, which transfers the Trustees hereby confirm.” See App’x at 81. Provide existing documents that explain how the funds were transferred. For example: a. Provide existing documents to show whether Fridman was added to the existing account as trustee. b. Provide existing documents relating to a redacted request concerning the transfer of funds. Provide existing documents sufficient to show the transfer of property described in Schedule A of the Trust Agreement. Request 12: All bank records for HSBC Private Bank (Suisse) SA [redacted account number] from 1999 to 2002. Request 13: All bank records for any account held under the name of any trust for which Fridman is a Trustee or Beneficiary since the inception of such trusts. Request 14: In reference to first clause of Trust Agreement dated 04/30/1990, the Trust Agreement states “The Grantor wishes to record that she intends by this Trust Agreement to create two separate trusts…” See App’x at 81. a. Please provide existing documents that show the name of the trusts and all bank records for any account holding assets from the inception of the trusts. Request 15: Trust Agreement for any Trust for which Fridman is a Trustee or Beneficiary. Request 16: All books and records for the Trusts referenced in Clause 1 of the Trust Agreement dated 04/30/1990 and any other Trusts for which Fridman is a Trustee or Beneficiary, including but not limited to David Marcelo Trust. Request 17: All records pertaining to property in which David Marcelo Trust and/or any other trust have an interest. Request 20: All correspondence between Fridman and any other trustees, trustors, beneficiaries, and any other persons involved with the trust(s) for which Fridman is a Trustee or Beneficiary related to such a trust, its property and/or its administration. Fridman asserted his Fifth Amendment privilege and refused to produce the requested documents. On March 11, 2015, the Government filed a petition to enforce the summonses. United States v. Fridman, 337 F. Supp. 3d 259, 262 (S.D.N.Y. 2018). On November 25, 2015, the district court granted the petition on the grounds that the documents were relevant and certain exceptions to the Fifth Amendment act-of-production privilege applied. Id. Fridman appealed, and a panel of this Circuit vacated and remanded by summary order. United States v. Fridman, 665 F. App’x 94, 94 (2d Cir. 2016). While we affirmed the determination on relevance, we remanded on the Fifth Amendment issue because the record was insufficiently developed. Id. at 96-97. In a footnote, we directed the district court to look to an intervening decision from our Circuit, United States v. Greenfield, 831 F.3d 106 (2d Cir. 2016), with respect to its analysis of the foregone conclusion doctrine. Fridman, 665 F. App’x at 96 n.1. On remand, the district court again granted the petition. Fridman, 337 F. Supp. 3d at 264. The district court applied the foregone conclusion doctrine to Requests 1, 2 (limited to the Known Accounts only), 4, and 7. Id. at 268-71. Because the Government “has provided the name of the account holder, the bank, the country, and the account number for each of the Known Accounts,” the district court was satisfied that the Government knew of the existence, location, and authenticity of the requested documents, as required by Greenfield. Id. at 268-71. The district court also held that so-called “traditional trusts” were collective entities for purposes of the collective entity rule, and therefore that Requests 1-4, 7, 11-17, and 20 were excepted as to all trust-related documents. Id. at 270-72. The district court applied the collective entity rule to Request 3 as well because the entities listed are collective. Id. at 272-74. Rejecting Fridman’s argument that the Summonses were not served on him in his capacity as representative for any of those entities, the district court reasoned that “[r]egardless of the capacity in which Fridman was served, the result is the same: [he] is required to produce documents responsive to Document Request No. 3 to the extent he possesses such documents in his individual capacity or his capacity as the custodian of a collective entity, including the corporations listed in the Document Request.” Id. at 273. Fridman timely appealed. DISCUSSION “We review de novo the District Court’s determination of questions of law as to the Fifth Amendment privilege. But [we] will overturn the District Court’s determination as to whether the act of producing the documents would involve testimonial self-incrimination only where such a finding has no support in the record.” Greenfield, 831 F.3d at 114 (internal quotation marks and citations omitted). The Fifth Amendment provides that no person shall be compelled in a criminal case to be a witness against himself. U.S. Const. amend. V. In Fisher v. United States, 425 U.S. 391, 409-11 (1976), the Supreme Court defined the contours of the Fifth Amendment as it applies to document requests. The Court held that documents voluntarily prepared prior to the issuance of a summons were not compelled testimony, so there was no Fifth Amendment protection for the contents of these records. Id. at 410-11. At the same time, however, the Court recognized a narrow privilege against the act of production. Because producing documents “tacitly concedes the existence of the papers demanded and their possession or control by the taxpayer…[as well as] the taxpayer’s belief that the papers are those described in the subpoena,” the Court concluded that the act of production could, in some cases, communicate incriminatory statements and thus may fall under the Fifth Amendment’s protection against self-incrimination; but the Court hinted that such a determination would be conditioned on the “facts and circumstances of particular cases.” Id. at 410-11. Similarly, when a defendant must “make extensive use of the contents of his own mind in identifying the hundreds of documents responsive to the requests in the subpoena,” he or she contributes to a “link in the chain” of their prosecution in violation of the Fifth Amendment privilege. United States v. Hubbell, 530 U.S. 27, 42-43 (2000) (internal quotation marks and citations omitted). The act-of-production privilege is not an absolute one. Fridman challenges the district court’s ruling that two exceptions to the act-of-production privilege permitted enforcement of the requests at issue in this case. I. The Foregone Conclusion Doctrine Under one exception, when “[t]he existence and location of the papers are a foregone conclusion and the taxpayer adds little or nothing to the sum total of the Government’s information by conceding that he in fact has the papers,” production does not run afoul of the Fifth Amendment. Fisher, 425 U.S. at 411. This principle has been aptly called the foregone conclusion doctrine. As referenced above, our Circuit most recently addressed the foregone conclusion doctrine in United States v. Greenfield, 831 F.3d 106, 110 (2d Cir. 2016). There too, we dealt with a summons related to tax evasion. Id. at 110. In Greenfield, we explained that for the foregone conclusion exception to apply, the Government must establish “with reasonable particularity” its knowledge as to “(1) existence of the documents, (2) the taxpayer’s possession or control of the documents and (3) the authenticity of the documents.” Id. at 115; see also id. at 119.3 To meet the existence requirement, the “Government is not required to have actual knowledge of the existence and location of each and every responsive document.” Id. at 116. When a summons seeks “customary account documents related to financial accounts that [the Government] knew existed,” the documents’ existence is a foregone conclusion. Id. at 118. Although we did not explicitly define “control” in the context of the foregone conclusion doctrine in Greenfield, we start from the premise that a taxpayer’s “possession or control of the [requested] documents” is one of the “communicative elements” protected by the Fifth Amendment. Id. at 115 (emphasis added); see also Fisher, 425 U.S. at 409-11. The Government may therefore satisfy the control requirement by establishing its knowledge of the physical possession of the requested documents by the subpoenaed individual, see Fisher, 425 U.S. at 411-12, or the subpoenaed individual’s control over the requested documents, see Greenfield, 831 F.3d at 119 (explaining that an individual’s ability or authority to receive the requested documents is an essential part of being able to control the documents); id. at 199 n.10 (“[B]oth a general power of attorney and the power to give instructions would suffice to provide [the individual] with control over [the requested] documents.”); see also Matter of Marc Rich & Co., A.G., 707 F.2d 663, 667 (2d Cir. 1983) (“The test for the production of documents is control, not location); Control, Black’s Law Dictionary (11th ed. 2019) (defining “control” as “[t]o exercise power or influence over”). Thus, if the Government can prove it knows that an individual controls the disposition of assets in an account, it follows that that individual controls the requested documents associated with that account — given the Government has satisfied the existence requirement. With respect to the authenticity requirement, documents may be “implicitly” authenticated if an individual complies with a summons demanding production of documents, and the Government establishes that “th[ose] documents are in fact what they purport to be” and “the taxpayer [was] not…forced to use his discretion in selecting…the responsive documents.” Greenfield, 831 F.3d at 118 (internal quotation marks and citation omitted). Additionally, the Government can “independently establish[]” authenticity in several ways: “ a) through the testimony of third parties familiar with that type of document, b) by comparison to a prior version of the document, or c) by comparison to other related documents.” Id. (citations omitted). There is a critical temporal requirement as well. The Government must prove its knowledge at the time the summons was issued. Id. at 119. “Relevant” to this inquiry is whether “the Government can demonstrate that the documents ever existed.” Id. at 119. “[I]n many circumstances, the Government’s ability to establish existence and control as of an earlier date does permit an inference of existence and control as of the date of the Summons.” Id. at 125 (emphasis added). When considering whether such an inference of continued existence and control is available,4 we have borrowed and applied a balancing test from the Eighth Circuit that examines: “(1) the nature of the documents, (2) the nature of the business to which the documents pertained, (3) the absence of any indication that the documents were transferred to someone else or were destroyed, and (4) the relatively short time period…between the date as of which possession was shown and the date of the ensuing IRS summons.”5 Id. at 125-26 (internal quotation marks and citation omitted). Greenfield provides a bifurcated approach to the foregone conclusion analysis when the “inference” of continued existence and control is at play. First, we examine “whether the Government can establish the existence, control, and authenticity of each category of sought documents” at some point prior to the issuance of the summons. Id. at 119. For those documents for which the Government’s evidence suffices, we then examine “whether it is a foregone conclusion that these documents remained in [the individual's] control through the issuance of the Summons…. Only if that retention is a foregone conclusion will the issuance of the Summons not violate [the] Fifth Amendment privilege.” Id. at 123.6 In this case, Fridman challenges the district court’s determination that the foregone conclusion doctrine allows enforcement of Requests 1, 2, 4, and 7. He makes two primary arguments. First, he claims that the district court misapplied our decision in Greenfield. Second, he asserts that enforcing these Requests would require him to provide testimonial information because the Government has failed to establish its knowledge of the relevant documents with reasonable particularity. We are not persuaded by either argument. 1. Existence There is no doubt that the Government knows that the Known Accounts existed prior to the issuance of the Summonses. Exhibit B, which lists the Known Accounts, specifies the account holder, account number, the bank, and the location (by country) for each account. Based on filings from Fridman or his representative, the Government also knows the following about the listed accounts: (1) the Personal Accounts existed in 2008; (2) the HSBC Account existed in the years 2003-2010; (3) the Consist Consultoria Systemast Repre Corporate Accounts and Wanstat Systemar Corporate Account existed in 1991; (4) the Consist France Corporate Accounts existed in 1998; (5) the Consist Asia Pacific Corporate Account existed in 1998; and (6) two of the Mak Data System Corporate Accounts existed in 1998 and in the years 2000-2004 while four others existed in 2004 and 2005. Requests 1, 2, 4, and 7 seek “customary account documents” from the Known Accounts: (1) records sufficient to show the opening and closing dates of the Known Accounts (listed in Exhibit B)7; (2) bank statements for the Known Accounts for 2006 and 2007; (3) bank statements and account opening documents for certain Personal Accounts and the HSBC Account, from the account opening date through to January 31, 2009; and (4) bank records for transactions that the Government knows occurred between the HSBC Account and an unknown Citibank account. These requested documents — bank statements, account opening documents, transaction records, and account closing documents — are all quintessential customary account documents. Because of this, and because these documents pertain to “financial accounts that [the Government] knew existed,” their existence is a foregone conclusion. Greenfield, 831 F.3d at 118. Fridman offers a number of unavailing counterarguments. First, Fridman argues that the Government has not proven that it knows the specific requested documents existed. But it was not required to do so. Greenfield makes clear that even if “the Government does not have specific knowledge of every document that is responsive to the Summons, such specific knowledge exceeds what is required under a ‘reasonable particularity’ standard.” Id. at 119. And when dealing with customary account documents, proof of the account’s existence is sufficient to show knowledge of the account documents with a reasonable particularity. Id. Therefore, because the Government knows of the accounts’ existence, it has met its burden. Fridman’s argument has the most traction with respect to Requests 1 and 2. As to the account closing documents sought in Request 1, it is true that the Government does not know for certain that any of the Known Accounts, other than the HSBC Account, have been closed. And unlike account opening documents, which must necessarily exist for all the Known Accounts, account closing documents will only exist for those accounts that have been closed. Therefore, it is not as obvious whether knowledge of closing documents can be predicated on knowledge of the accounts’ existence. Similarly, although Request 2 seeks bank documents for 2006 and 2007, the Government does not know for a fact that the Personal or Corporate Accounts existed in those years; the IRS has documents showing the Personal Accounts existed in 2008, and it knows the FBAR filings for the relevant Corporate Accounts were made at various points from 1991 to 2005. But the Government does not need specific knowledge that the accounts had been closed or that the accounts existed in 2006 and 2007. Such a requirement does not square with our decision in Greenfield. In Greenfield, the summons called for, inter alia, all documents in Greenfield’s possession for a certain account. Id. at 112-13. There, we held that the Government had sufficiently proven its knowledge of the existence of those documents because of its knowledge of the account’s existence in 2001 and because the documents sought were customary account documents. Id. at 119. It was no matter that the category of documents sought — “all documents in [Greenfield's] possession” pertaining to the account — was sweeping and significantly broader than Requests 1 and 2 here. See id. at 113. Thus, even though the Government in this case does not have “specific knowledge of every document that is responsive to the Summons,” such a heightened standard is not required under our “reasonable particularity” standard, nor does it align with our analysis in Greenfield. Id. Fridman also contends that the Government does not have specific knowledge of the facts contained in the documents soughtBut in arguing that the Government must know of the facts contained in the documents, Fridman conflates the documents’ contents with their existence. The distinction is a consequential one: the contents of documents are not covered by the Fifth Amendment privilege. Fisher, 425 U.S. at 409-10. Therefore, the Government need not prove its knowledge of the documents’ contents. Further, In re Katz, 623 F.2d 122, 126 (2d Cir. 1980), is not to the contrary. In In re Katz, the Government sought “all documents relating to any dealings or business with…any company owned, operated or controlled” by an individual named Benjamin Jamil. Id. at 123. We denied enforcement of the subpoena as written because the Government did not know the identity of the corporations, Jamil’s relationship with them, or whether some documents may be protected by attorney-client privilege. Id. at 126-27. We remanded to the district court to either conduct in camera review or limit enforcement of the subpoena to public documents related to known entities. Id. But here, the Government knows the accounts whose records are sought and their connection to Fridman. Because the Government knew the accounts existed and requested only customary account documents, we conclude that there is sufficient evidence in the record to support the Government’s knowledge as to the documents’ existence for Requests 1, 2, 4, and 7. 2. Prior Control The record establishes that Fridman’s representative provided account statements or other bank records to the IRS in 2012 for the Personal and HSBC Accounts. Moreover, Fridman held the Personal Accounts individually in his name, and the HSBC Account was held in the name of the David Marcelo Trust, for which Fridman serves as Trustee. This is sufficient to establish the Government knew Fridman had control over the Personal and HSBC Accounts. See id. at 119-20. With respect to the relevant Corporate Accounts, Fridman made FBAR filings at various points from 1991 to 2005. An individual must file an FBAR if he has “a financial interest in, or signature or other authority over” a foreign financial account. 31 C.F.R. §1010.350(a). By filing FBARs for the Corporate Accounts, Fridman indicated that he had control over those accounts’ records as of the years of those corresponding filings. See id. §1010.350(a), (e)-(f). Although Fridman argues that certain requested documents would be in the possession of offshore banks, not the account owner, this argument is inapposite. We first note that the Government was not required to establish that Fridman possessed the documents; rather, it was required to prove possession or control of the documents. See Fisher, 425 U.S. at 410; see also Greenfield, 831 F.3d at 119-21. Moreover, we have previously explained that “[t]he test for the production of documents is control, not location.” Matter of March Rich & Co., A.G., 7070 F.2d at 667 (holding that a witness cannot resist the production of documents on the ground that the documents are located abroad). Because Fridman or his representative filed FBARs and copies of documents pertaining to the Known Accounts, there is sufficient evidence in the record to establish the Government’s knowledge of Fridman’s control over the requested documents. 3. Inference of Continued Existence and Control Having concluded that the Government has met its burden of showing that the documents existed and were in Fridman’s control at one point, we now turn to the question of existence and control at the time the Summonses were issued in 2013. See Greenfield, 831 F.3d at 123. In doing so, we are permitted to infer existence and control if the four factors from the Greenfield balancing test weigh in favor of such an inference. Id. at 125-26. After considering these factors, we conclude that the record permits an inference of continued existence and control in the present case. The first three factors support an inference of continued existence and control. First, as we explained in Greenfield, “bank documents are more likely to be retained long term as compared to documents like receipts or prosaic emails.” Id. Second, banks “tend to maintain consumer records.” Id. Third, unlike in Greenfield, here there is no “significant intervening event[] that might well have resulted in transfer or destructions of the sought documents.” Id. Although Fridman suggests that the absence of continued FBAR filings for the Corporate Accounts indicates that he ceased having an interest in those accounts, we are not convinced. This absence could be a product of Fridman’s failure to make requisite filings consistently.8 His suggestion that he lost control of his Corporate Accounts during certain years is just that — a suggestion. It does not amount to evidence of intervening factors like those encountered in Greenfield. Id. at 126-27. The final factor is less clear-cut. We have little difficulty agreeing with the Government that the time period between the last date of possession for the Personal and HSBC Accounts and the date of the Summonses was relatively short. Fridman’s representative provided the IRS with documents relating to the Personal Accounts in May 2012. The same representative last provided the IRS with records relating to the HSBC Account in November 2012. Thus, only nineteen and thirteen months, respectively, eslaped between this evidence of control and the date of the Summonses (December 2013). Both lengths of time are relatively short. See United States v. Rylander, 460 U.S. 752, 761 n.3 (1983) (concluding that an “inference of continuing possession” over twenty-one months was reasonable). The same cannot be said for the Corporate Accounts. The last known date of control for some of these accounts was more than a decade before the issuance of the Summonses. The most recent last-known date of control was 2005, approximately eight years prior to the Summonses’ date. This factor is not dispositive, however, for we weigh all four factors in determining if an inference of existence and control is acceptable. The Summonses seek “customary account documents,” the requested documents are related to financial accounts, and there is no evidence suggesting that the documents were transferred or destroyed.9 Here, while the time lapse for the Corporate Accounts is significant, considering all the factors together, the inference is still applicable. 4. Authentication We are also satisfied that the Government could independently authenticate the documents it seeks without relying on Fridman’s act of production. It could do so through the testimony of third parties familiar with records, by comparisons to other bank records it already possesses, or by comparisons to similar bank documents. Greenfield, 831 F.3d at 118. For the Personal and HSBC Accounts, the Government possesses copies of statements and other account documents that Fridman’s representative produced to it, and the Government can use those as comparators for authentication purposes. Although Greenfield did not reach the issue of whether such comparisons can sufficiently authenticate documents, we so hold here. As to the relevant Corporate Accounts, the Government argues that it can use bilateral tax treaties to authenticate the documents. It points to a declaration from an IRS employee noting that through certain treaties, the United States can request authenticated records from Switzerland, Israel, Brazil, France, Spain, the Philippines, and Germany — countries where the relevant Corporate Accounts are located. Through those treaties, the Government may also obtain sworn testimony or statements of authentication from bank officials to authenticate bank records.10 In our view, these tax treaties provide the Government with a sufficient independent method of authentication. Although in Greenfield we said that authentication through use of the Hague Evidence Convention was insufficient, we emphasized that this was “in light of the controversy surrounding the source of the documents” and the Government’s failure to provide anything beyond “a conclusory statement” that authentication was likely to occur. 831 F.3d at 120. Because whether treaties can be used to authenticate documents depends on the facts and circumstances of a particular case, and because the record before us presents neither the unusual circumstances existing in Greenfield nor any other indication that the bilateral tax treaties to be used here are too complicated or ineffective to permit authentication, we conclude that the Government has sufficiently established that it can authenticate the Corporate Account documents. In sum, we are satisfied that the Government has met its burden of proving knowledge of the existence of the documents, knowledge of Fridman’s control over the documents, and an independent means of authenticating of the documents as required by Greenfield. Therefore, the district court’s decision that the foregone conclusion doctrine applies to Requests 1, 2, 4, and 7 for documents relating to the relevant Known Accounts is affirmed. II. The Collective Entity Doctrine Because the Fifth Amendment privilege protects only natural persons, collective entities such as corporations or partnerships may not invoke it to evade document requests. See, e.g., Braswell v. United States, 487 U.S. 99, 102, 104-08 (1988); Bellis v. United States, 417 U.S. 85, 88-89, 93-101 (1974). Nor may an individual custodian holding a collective entity’s records in a representative capacity refuse to produce documents. Braswell, 487 U.S. at 108-12; In re Grand Jury Subpoena Issued June 18, 2009, 593 F.3d 155, 158 (2d Cir. 2010). “The critical issue [for determining what is a collective entity] is whether the organization had an institutional identity separate from that of its individual members.” In re Two Grand Jury Subpoenae Duces Tecum Dated Aug. 21, 1985, 793 F.2d 69, 72 (2d Cir. 1986); see also Bellis, 417 U.S. at 94-95. Pursuant to this rule, we have held that sole proprietorships are covered by the Fifth Amendment’s privilege against compulsory self-incrimination, because “a sole proprietorship has no legal existence apart from its owner.” United States v. Fox, 721 F.2d 32, 36 (2d Cir. 1983); see also Braswell, 487 U.S. at 104. We have been strict in limiting the privilege to only sole proprietorships and even holding that one-person corporations are collective entities. In re Grand Jury Subpoena Issued June 18, 2009, 593 F.3d at 158-59. Our conclusion was motivated in part by the desire to “avoid[] creating a category of organizations effectively immune from regulation by virtue of being beyond the reach of the Government’s subpoena power.” Id. at 159. Fridman first challenges the district court’s holding that trusts are collective entities and that Fridman is therefore required to produce all trust-related documents contemplated by Requests 1-4, 7, 11-17, and 20. Fridman also challenges the district court’s conclusion that he is required to produce documents responsive to Request 3 that he possesses in his capacity as custodian of certain collective entities. A. The collective entity rule applies to trusts, including traditional common law trusts, like the David Marcelo Trust. Whether a trust is a collective entity for purposes of the Fifth Amendment is an issue of first impression for our Circuit. Every other circuit to address the issue has answered in the affirmative. See In re Grand Jury Subpoena, 973 F.2d 45, 48 (1st Cir. 1992); Watson v. Comm’r of Internal Revenue, 690 F.2d 429, 431 (5th Cir. 1982); United States v. Harrison, 653 F.2d 359, 361-62 (8th Cir. 1981); In re Grand Jury Proceedings, 633 F.2d 754, 757 (9th Cir. 1980). Our sister circuits have focused on facts like the circumscribed discretionary authority of trustees, formal status of the trust, and that a defendant may not be the sole beneficiary, In re Grand Jury Subpoena, 973 F.2d at 48-51; that trusts are artificial entities, Watson, 690 F.2d at 431; that a trust is formally organized and legally distinct from the trustees, Harrison, 653 F.2d at 361-62; and that a trust has independent functions even when it is grantor-controlled, In re Grand Jury Proceedings, 633 F.2d at 757. We join these circuits and hold that a trust is a collective entity. We conclude so for a number of reasons. Notably, a trust, including a traditional common law trust like the David Marcelo trust,11 has a separate legal existence from the trustee, which is the critical hallmark of a collective entity. See Aug. 21, 1985, 793 F.2d at 72; see also Harrison, 653 F.2d at 361-62; In re Grand Jury Proceedings, 633 F.2d at 757. Unlike a sole proprietorship, which cannot survive without the owner or creator, when a trustee resigns or is removed, the trust may still continue under New York law. Cf. 106 N.Y. Jur. 2d Trusts §§294, 298, 303. This could not be so if the trust and trustee were one and the same. In addition, a trust “represent[s] a formal institutional arrangement.” Cf. Bellis, 417 U.S. at 95. A trust is “relatively well organized and structured,” much like the other collective entities previously recognized by Supreme Court precedent on the collective entity exception. See id. at 92-93; see also Braswell, 487 U.S. at 104-12 (detailing the collective entity rule’s “lengthy and distinguished pedigree”). One example is that a trustee’s discretion is limited by the governing document and the principle that a trustee must act in the beneficiaries’ best interest. See 106 N.Y. Jur. 2d Trusts §§347, 360; see also In re Grand Jury Subpoena, 973 F.2d at 48-50. This clearly distinguishes a trust from an entity that “embodies little more than [a] personal [venture].” See Bellis, 417 U.S. at 94-95. Another example is that a trust’s records are “distinct from the personal books and records of” the trustee. Id. at 93; 106 N.Y. Jur. 2d Trusts §373. Finally, our holding that trusts are collective entities recognizes that the decision to create a trust “is freely made and generates benefits, such as limited liability, and burdens, such as the need to respond to subpoenas” for records. See In re Grand Jury Subpoena Issued June 18, 2009, 593 F.3d at 159. As the Eighth Circuit has said, “those who form a separate business entity [and] hold that entity out as distinct and apart from the individuals involved…are estopped from denying the existence and viability of that entity for Fifth Amendment purposes.” Harrison, 653 F.2d at 361-62. We agree that it would be inequitable to allow individuals to create a separate entity like a trust for favorable treatment while simultaneously denying that entity’s separate existence for subpoena purposes. Fridman’s arguments to the contrary are meritless. He first argues that because a trustee can be held personally liable, a trust is more akin to a sole proprietorship. Although a trustee is personally liable “in the first instance” when acting on behalf of the trust, the trustee may be “entitled to be reimbursed or indemnified…from [trust] assets.” 106 N.Y. Jur. 2d Trusts §356. Fridman next argues that because a trust cannot take legal action on its own behalf, and because recent case law establishes that a trustee’s individual citizenship governs jurisdiction in diversity suits, a trust and trustee are one and the same. See Americold Realty Tr. v. ConAgra Foods, Inc., 136 S. Ct. 1012, 1016 (2016); Raymond Loubier Irrevocable Tr. v. Loubier, 858 F.3d 719, 731 (2d Cir. 2017). But that “state and federal law do not treat [trusts] as distinct entities for all purposes” does not alter the conclusion that a trust is a collective entity because trusts “bear enough of the indicia of legal entities to be treated as such for the purpose of our analysis of the Fifth Amendment issue presented in this case” for the reasons discussed. See Bellis, 417 U.S. at 97 n.7. We hold that trusts, including a traditional common law trust like the David Marcelo Trust, are collective entities for purposes of the Fifth Amendment. As such, we affirm the district court’s grant of the petition to enforce the Summonses with respect to any documents Fridman retains in his capacity as a trustee in response to Requests 1-4, 7, 11-17, and 20.12 B. Fridman must also produce responsive corporate documents he retains in a representative capacity. Fridman also challenges the district court’s ruling that if he possesses or controls records of any of the entities listed in Request 3 in a representative capacity, he must produce these records under the collective entity doctrine. Fridman argues that it is the Government’s burden to prove a custodial relationship exists between him and the corporations. While our Circuit has not articulated a specific framework for establishing a custodial relationship, we adopt the D.C. Circuit’s burden-shifting framework. See, e.g., In re Sealed Case, 877 F.2d 83, 87 (D.C. Cir. 1989). The Government need only show a reasonable basis to believe a defendant has the ability to produce records; once the Government has done so, the burden shifts to the defendant to explain or justify refusal. See id. In the case at hand, the Government has established Fridman’s control over the records for the Corporate Accounts, as we discussed in our analysis of the foregone conclusion doctrine. This control provides a reasonable basis for believing that Fridman has the ability to produce these records, and the burden has thus shifted to Fridman to prove otherwise. See id. Fridman fails to meet his burden. He cites to two cases from our Circuit, In re Grand Jury Subpoenas Duces Tecum Dated June 13, 1983 & June 22, 1983, 722 F.2d 981, 986-87 (2d Cir. 1983), and In re Three Grand Jury Subpoenas Duces Tecum Dated January 29, 1999, 191 F.3d 173, 181 (2d Cir. 1999), for the proposition that former employees possess records only in an individual, not representative, capacity. But Fridman’s reliance is misplaced. It is clear that former employees could not serve in a custodial capacity and therefore would hold records only in a personal capacity. See In re Three Grand Jury Subpoenas Duces Tecum Dated January 29, 1999, 191 F.3d at 181. However, Fridman has not shown that he is no longer affiliated with any of the corporate entities at issue here. Therefore, these cases are inapposite. We accordingly affirm the district court’s ruling that Fridman must produce the corporate documents he holds in a representative capacity. CONCLUSION For the reasons given above, we affirm the district court’s grant of the petition to enforce the Summonses.
