Cobell v. Kempthorne

532 F. Supp. 2d 37, 2008 U.S. Dist. LEXIS 6658, 2008 WL 253035

• Court: District Court, District of Columbia
• Decided: January 30, 2008
• Docket: Civil Action 96-1285 (JR)
• Status: Published

Opinion

FINDINGS OF FACT AND CONCLUSIONS OF LAW

JAMES ROBERTSON, District Judge.

These findings and conclusions are the result of a 10-day bench trial in October 2007. The central purpose of the trial was to determine whether the Department of Interior has remedied or is remedying what Judge Lamberth found in Cobell v. Babbitt, 91 F.Supp.2d 1, 58 (D.D.C.1999) (Cobell V), aff'd, Cobell v. Norton, 240 F.3d 1081 (D.C.Cir.2001) (Cobell VI), to be a breach of its duty under the Indian Trust Fund Management Reform Act of 1994 to produce an accounting for Individual Indian Money (IIM) account holders. In setting the matter for trial, I said that, although the details of the trial remained to be worked out, it was both appropriate and prudent to review the Interior Department’s historical accounting project in detail, and to do so in open court, where the government might present, and plaintiffs might test or challenge, its methodology and results up to the time of the hearing [Dkt. 3312]. The end product of such a proceeding was to include the answers to at least the following questions:

• Have the defendants cured (or are they curing) the breaches of their fiduciary duty that were found in Cobell V?

• Do the defendants’ historical statements of account ... satisfy defendants’ duties “rooted in and outlined by the relevant statutes and treaties ... [and] defined in traditional equitable terms”? Cobell VI, 240 F.3d at 1099.

• Have the defendants unreasonably delayed the completion of the required accounting?

• What further relief, if any, should be ordered?

By the time the trial began, the issues for trial had been distilled to these four:

First, it’s going to be about what you’re doing and what you’re not doing.... It’s going to be about both of those things. Second, what would it cost to do the things that they say that you should be doing and you’re not doing?

Third, taking into account the cost, because that, I think, I’m required to do by the Court of Appeals, is what you’re doing adequate? Is it an adequate accounting?

And fourth — and this is what you don’t want to hear, but I think Mr. Gingold is entitled to at least a record on this point, fourth, what does it all add up to? Throughput versus what you can prove, what are the big numbers?

H’rg Tr. 76:23-77:10 (6/18/07). The question of what further relief, if any, should be ordered was left to another day.

These findings and conclusions, derived not only from the trial, but also from the extensive record that preceded it, support and explain my decision (i) that, although the defendants have attempted and continue to attempt to cure the breach of their fiduciary duty that was found in Cobell V and affirmed by Cobell VI, they have not succeeded in doing so; (ii) that the historical statements of account contemplated by defendants’ latest accounting plan will not satisfy defendants’ duties “rooted in and outlined by the relevant statutes and treaties ... [and] defined in traditional equitable terms,” Cobell VI, 240 F.3d at 1099; and (iii) that the defendants have unreasonably delayed the completion of the required accounting. Indeed, it is now clear that completion of the required accounting is an impossible task.

BACKGROUND

To say that the histories of the IIM trust and of this lawsuit have been exhaustively chronicled in district court and appellate opinions is to stretch the limits of understatement. See, e.g., Cobell v. Babbitt, 30 F.Supp.2d 24, 27-29 (D.D.C.1998); Cobell v. Babbitt, 91 F.Supp.2d 1, 6-12 (D.D.C.1999); Cobell v. Norton, 240 F.3d 1081, 1086-94 (D.C.Cir.2001); Cobell v. Norton, 226 F.Supp.2d 1, 11-20 (D.D.C. 2002); Cobell v. Norton, 283 F.Supp.2d 66, 72-86 (D.D.C.2003). Those seeking Cliffs-Notes can even consult the Cobell v. Kempthome Wikipedia entry (though the Court, of course, cannot vouch for its accuracy). At this date, there are 3,504 entries on the Cobell v. Kempthome docket. Appellate panels hearing Cobell arguments have engaged ten of our Circuit judges, some of them more than once. Upon publication, this opinion will have the shorthand title Cobell XX. Nevertheless, those histories must be retold at least briefly in order to provide context for today’s opinion.

Plaintiffs are a certified class of present and former IIM account holders numbering in excess of 300,000. Some account holders have more than one IIM account. Hundreds of thousands of IIM accounts exist, managed for the United States by its trustee-delegates, the Department of Interior and the Department of Treasury. Most of these IIM accounts exist to receive income the government collects for leasing or selling Indian-owned lands and then to distribute it to account holders when account balances reach certain thresholds (usually fifteen dollars). A small percentage of the funds flowing through the IIM trust are in “Judgment” and “Per Capita” accounts, which were created to hold funds derived from litigation settlements (Judgment accounts) and tribal revenues allocable to individual Native Americans (Per Capita accounts). By far the largest amount of trust funds flow through the “land-based” IIM accounts that contain lease, royalty, and land sale payments tied to individual land allotments.

Individual Indian land allotments date to a period between the late 1800’s and 1934 when the federal government attempted to dismantle tribes and instill the Anglo-American concept of private ownership in Native Americans by carving reservation land into individually owned parcels of up to 160 acres (now known as “tracts” or “allotments”). See, e.g., County of Yakima v. Yakima Indian Nation, 502 U.S. 251, 254, 112 S.Ct. 683, 116 L.Ed.2d 687 (1992) (“The objectives of allotment were simple and clear cut: to extinguish tribal sovereignty, erase reservation boundaries, and force assimilation of Indians into the society at large.”), quoted in Cobell v. Norton, 240 F.3d 1081, 1087 (D.C.Cir.2001). The government’s pursuit of the allotment policy occurred alongside its official abandonment of treaty-driven relationships with tribes in favor of “governing] [tribes] by acts of Congress.” United States v. Kagama, 118 U.S. 375, 382, 6 S.Ct. 1109, 30 L.Ed. 228 (1886); see Act of March 3, 1871, ch. 120, § 1, 16 Stat. 566 (1871) (codified as amended at 25 U.S.C. § 71). This policy shift and its corollary acts— such as coercive assimilation — were carried out without so much as the pretense of tribal consent. See, e.g., Lone Wolf v. Hitchcock, 187 U.S. 553, 565-68, 23 S.Ct. 216, 47 L.Ed. 299 (1903).

The allotment policy was first codified in the Indian General Allotment Act (Dawes Act), ch. 119, 24 Stat. 388 (1887) (codified as amended at 25 U.S.C. § 331 et seg.), and was reflected in several subsequent allotment acts. In the Dawes Act, Congress granted unilateral authority to the executive branch to divide reservation land west of the Mississippi into plots for individual tribal members and families. It also allowed non-Indian settlement upon and exploitation of some reservation land, resulting in the alienation of millions of acres from tribal ownership. The statute required the federal government to hold the allotted land in trust for the individual allottees and their heirs for a period of 25 years — a period subject to extension at the government’s discretion — after which fee patents would issue to the allottees. See Cobell VI, 240 F.3d at 1087; Cobell v. Norton, 283 F.Supp.2d 66, 74 (D.D.C. 2003). While held in trust, allotted lands were to be immune from state taxation. The expectation was that, during that time, Indians would establish self-sufficient farms and earn enough money to pay their own taxes. At the close of the 19th Century, Congress passed several acts allowing the government to lease allotments that had not been successfully cultivated. See, e.g., An Act Making Appropriations for Current and Contingent Expenses of the Indian Department and Fulfilling Treaty Stipulations with Various Indian Tribes for the Fiscal Year Ending June Thirtieth, Eighteen Hundred and Ninety-Five, and for Other Purposes, ch. 290, 28 Stat. 286 (1894). 1 Any income generated from the land was to flow into IIM accounts established by the government. Native American landowners could not (and today still cannot) sell or lease allotted land without the government’s consent.

Congress soon realized that the allotment policy was responsible for innumerable problems, see, e.g., 1915 Congressional Report to the Joint Commission of the Congress of the United States, PX-681, not least of which was the phenomenon known as ‘fractionation’. Fractionation occurs when Indian allotments are divided and divided again by inheritance through succeeding generations, diluting the ownership interests of allottees and causing enormous administrative difficulties for the BIA. The Indian Reorganization Act (Wheeler-Howard Act), ch. 576, 48 Stat. 984 (1934), (codified as amended at 25 U.S.C. § 461 et seq.), was supposed to reconsolidate Indian lands and reverse the allotment process, but the land reclamation effort prescribed by that statute was never properly funded and never materialized. Instead, the Act succeeded only in ending the creation of new allotments and, for allotted lands already held in trust, extending the trust period indefinitely. Cobell v. Norton, 283 F.Supp.2d 66, 75 (D.D.C.2003). As of 1990, some eleven million acres were held in trust for the heirs of allottees, by now several generations removed. 2 Many trust allotments are owned in common by hundreds or even thousands of beneficiaries — each with undivided (“fractionated”) interests in the whole parcel.

The statute that gave rise to this litigation was enacted in 1994 as the Indian Trust Fund Management Reform Act, Pub.L. No. 103-412, 108 Stat. 4239 (codified at 25 U.S.C. § 4001 et seq.) (hereinafter “the 1994 Act”). The 1994 Act reflected many years of congressional frustration over Interior’s handling of the IIM trust. It commanded Interior, among other things, to provide an historical accounting to trust beneficiaries. 25 U.S.C. § 4011(a). Two years after the enactment of the 1994 Act, concerned that the required accounting had been neither accomplished nor even begun, the plaintiffs filed this suit. They alleged that the defendants were in breach of their fiduciary duties, and they prayed for an accounting and various other forms of declaratory, injunctive, and other equitable relief.

Judge Lamberth certified the plaintiff class on February 4, 1997 [Dkt. 27]. On November 5, 1998, he ruled that plaintiffs were not entitled to mandamus relief, but otherwise rejected the government’s motion to dismiss the complaint. Cobell v. Babbitt, 30 F.Supp.2d 24 (D.D.C.1998) (Cobell I). On June 7, 1999, he denied defendants’ motion for summary judgment [Dkt. 317], and, six months later, he declared defendants to be in breach of their statutory trust obligations, dismissed plaintiffs’ common law claims, and remanded the case to the agency with an injunction to bring its actions into conformity with its fiduciary duties and to submit quarterly reports on its progress (which quarterly reports it has been submitting ever since). Cobell v. Babbitt, 91 F.Supp.2d 1 (D.D.C. 1999) (Cobell V). The Court of Appeals affirmed the declaration of breach, but vacated several of Judge Lamberth’s more specific findings. Cobell v. Norton, 240 F.3d 1081 (D.C.Cir.2001) (Cobell VI).

Judge Lamberth then divided the proceedings into two phases: one addressing the “to-be plan,” or the agency’s plans to remedy its breach going forward, and another addressing defendants’ historical accounting work. After a 44-day trial, he issued an historical accounting opinion, a “fixing the system” opinion, and a structural injunction. Cobell v. Norton, 283 F.Supp.2d 66 (D.D.C.2003) (Cobell X). The historical accounting portion of the structural injunction was reversed by the D.C. Circuit on December 10, 2004, Cobell v. Norton, 392 F.3d 461 (D.C.Cir.2004) (Cobell XIII); reinstated by Judge Lam-berth after Congress failed to meet its self-imposed deadline for achieving the legislative settlement on which the Court of Appeals had relied in deciding Cobell XIII, Cobell v. Norton, 357 F.Supp.2d 298 (D.D.C.2005) (Cobell XIV); stayed by the Court of Appeals, Cobell v. Norton, 2005 U.S.App. LEXIS 5788 (D.C.Cir. Apr. 7, 2005); and eventually reversed, Cobell v. Norton, 428 F.3d 1070 (D.C.Cir.2005) (Co-bell XVII). On July 11, 2006, the Court of Appeals issued its eighth and ninth published opinions in the case, and directed that the case be reassigned. Cobell v. Kempthorne, 455 F.3d 301 (D.C.Cir.2006) (Cobell XVIII); Cobell v. Kempthorne, 455 F.3d 317 (D.C.Cir.2006) (Cobell XIX). On December 12, 2006, the case was assigned to me [Dkt. 3278].

During Judge Lamberth’s heroic stewardship of this case, he was beset by a host of important but ancillary issues: vulnerabilities within the Interior Department’s information technology systems, civil contempt proceedings concerning several government employees, retaliatory action within the agency against agency employees testifying on plaintiffs’ behalf, objections to communication between the agency and the plaintiff class, and the appointment and removal of court monitors and special masters, to name a few. The nineteen published opinions in this case have yielded no definitive, undisturbed ruling on the core question that looms over this dispute, which is: What is the scope or nature of the accounting that is required by the 1991 Act? This opinion seeks to answer that question. The outline set forth below emerges from the testimony taken in October 2007 on the trial questions identified above, and from the enormous record that has been compiled over the past eleven years.

I.THE GOVERNMENT’S ACCOUNTING EFFORT

1. Evolution of the current historical accounting project.

2. Activities the government has planned — and now plans — to discharge its accounting duty.

3. Activities the government has declared beyond the scope of its plan.

4. Cost to do what is not being done.

5. “Throughput” — amount of money received into the IIM trust; amount distributed to IIM beneficiaries.

II. COMPLIANCE WITH THE FIDUCIARY DUTY TO ACCOUNT

1. Does this court have jurisdiction?

2. Taking cost into account, will Interior’s 2007 Historical Accounting Plan result

in an adequate accounting that is compliant with the 1994 Act, prior Cobell opinions, and other precedent?

I. THE GOVERNMENT’S ACCOUNTING EFFORT

1.1. Evolution of the current historical accounting project

No real accounting, historical or otherwise, has ever been done of the IIM trust. What is now called the historical accounting project would have been unthinkable before the advent, some time in the 1980s, of automated systems that began to offer the possibility of merging the records and procedures that had been used for nearly 100 years in the various regions and agencies of the Bureau of Indian Affairs. This section describes the legacy computer systems that first held out some promise of doing an historical accounting, some of the problems that were encountered soon after the 1994 Act mandated an historical accounting, the exaggerated promises of Interior’s 2003 historical accounting plan, various ancillary IIM-related projects that Interior has undertaken, and the establishment in 2004 of the American Indian Records Repository, which has brought some degree of order out of the chaos that was IIM trust record-keeping.

I.I.A. Systems used in administering the IIM trust

The government uses an array of data and financial systems, referred to mostly by acronyms, 3 to collect, manage, and distribute IIM trust funds. The two major types of systems relevant to historical accounting contain land ownership data and financial data.

i. Integrated Records Management System (IRMS)

A major premise of the government’s accounting effort is that the transition from paper to electrons took the accuracy, completeness and reliability of IIM trust data to a level that far surpassed the “paper ledger era.” The agencies refer to February 1985 as the beginning of the new “electronic ledger era.” The first electronic ledger was the Integrated Records Management System (IRMS). IRMS contained four platforms for the input and access of lease data, ownership data, account data, and transactional data by BIA employees. Tr. 2041:1-23 (Christie); Tr. 343:22-344:5 (Ramirez). 4 The “IIM platform” within IRMS contained most of the accounting data. Tr. 343:23-25 (Ramirez). The government considers IRMS the beginning of the “electronic ledger era,” an era when data became putatively more uniformly handled and more easily accessible. Some BIA locations began using IRMS long before February 1985, however — some as early as the 1970s. Tr. 288:13-20 (Ramirez). The first regional office to use IRMS was Billings, Montana (one of the twelve regional BIA offices, now called the Rocky Mountain regional office). Tr. 1743:24-1744:2 (Infield). Interior did not require all BIA regions to use IRMS until the late 1980s. Tr. 1743:17-21 (Infield). Because IRMS was implemented on a rolling basis, it is impossible to draw a hard line between the paper ledger era and electronic ledger era.

Up until 1991, moreover, Interior deleted from IRMS data that were more than six years old, so as not to overwhelm their rudimentary computer systems. Data recorded before 1985 has been erased. Tr. 1744:22-1746:11 (Infield). And the post-1985 data on the IRMS system is incomplete and difficult to analyze, since some agencies 5 and regional offices did not convert to IRMS until after 1985, Tr. 693:1-6 (Herman), at least one agency (Osage) handled money collected from oil and gas leases in a unique way, Tr. 1737:4-1741:6 (Infield), bookkeeping within IRMS varied across regions, Tr. 572:19-573:15 (Herman), and even offices that did use IRMS did not always keep complete records on the system. Tr. 789:22-790:13 (Herman).

ii. Trust Funds Accounting System (TFAS)

In August 1998, Interior began its conversion from IRMS to the Trust Funds Accounting System (TFAS). The conversion was complete by March 2000. Tr. 450:18-24 (Herman). The upgrade to TFAS gave BIA the ability to send quarterly IIM statements, Tr. 866:15-867:9 (Winter), which it began to do in 2000. Those statements reflect opening and ending balances as well as posted transactions. Tr. 867:1-4 (Winter). Consequently, the agency uses that year as the end point of its HSA project. TFAS also contains data regarding the IIM funds that are pooled for investment purposes. Tr. 872:3-12 (Winter). Though transactional information has been recorded in many different ways over the life of the trust, the HSA project focuses primarily on information contained in paper records, and later in IRMS and TFAS.

iii. Land Records Information System (LRIS) and Trust Asset and Accounting Management System (TAAMS)

The Land Records Information System (LRIS) contains historical land ownership information that is used in Interior’s historical accounting project, National Opinion Research Center (NORC) Analysis of LRIS Tract History Reports, AR-405 at 5, but at least some agencies determined that the ownership information in LRIS was out of date, and that LRIS data could not be relied upon for distributing trust funds. Tr. 1317:17-1321:25 (Redthunder).

Interior began converting from LRIS to a “more robust land title system” called the Trust Asset and Accounting Management System (TAAMS) in 2005. Tr. 69:6-13 (Cason). As of September 2007, all BIA agency offices had at least partially converted to TAAMS. Tr. 868:6-11 (Winter). TAAMS contains two different systems: realty and title. Tr. 903:15-904:2 (Winter). The TAAMS title system produces invoices for lease holders, who then send their payments along with the invoice to a lockbox. Tr. 868:6-11 (Winter). The invoice and lockbox procedures are very new: they accompanied the TAAMS title roll-out in 2005. Tr. 903:15-904:12 (Winter). It is not clear whether all agencies have fully implemented the TAAMS title system at this date, but those that have now have an accounts receivable system for the first time in the history of the IIM trust. Tr. 902:23-904:12 (Winter). This accounts receivable system is referred to as the Trust Funds Receivable System (TFR), and permits the agency to follow up on leases to be sure that payments are made. Tr. 902:23-903:14 (Winter). Prior to the TAAMS conversion, using something like an honor system, Interior simply relied on lease holders to submit accurate, timely payments.

iv.Other data systems

Other systems, while not prominently featured in the historical accounting pro ject, contribute to the functioning of the IIM trust. For example, the Treasury Department has systems that record IIM trust funds and balances. The Minerals Management Service within Interior (MMS) has a data system for processing payments received from lessees and sending payment and lease-level information (not broken down by beneficiary or allotment) to Treasury and the Bureau of Indian Affairs. BIA processes oil and gas monies from MMS in a system called the Royalty Distribution Reporting System (RDRS) in order to allocate oil and gas revenue among land owners. Tr. 919:6-920:1 (Winter). BIA relied on Osage computer systems to process quarterly annuity payments and some oil and gas income. Tr. 1736:2-32 (Infield). In the past, BIA used several different systems to distribute trust revenues, such as the ownership module within IRMS, the GLAD distribution system in the Great Plains Agency, and the MAD system used in the Aberdeen and Standing Rock Agencies. Tr. 1756:12-18 (Infield); Tr. 1321:20-1322:2 (Redthunder).

I.l.B. Inconsistencies between agency and regional offices

The BIA is divided into twelve regions, under which a total of 93 agency offices operate. Feb.2002 OHTA Accounting Records Conference Materials at 213, AR-058. As mentioned above, the use of electronic records systems varied among the twelve regional offices, and even among agency offices within a single region. Former Special Trustee Paul Homan recalled a contractor describing the information systems used by the regional offices as “twelve islands of information without a ferry in between.” Tr. 1552:21-23 (Ho-man). The regional offices, and agencies within those regions, each had their own ways of using — or not using — IRMS. Most agencies used the MMS system to process oil, gas, and mineral payments, but Osage was again an exception to this rule. Tr. 1737:4-24 (Infield).

I.l.C. Destruction of documents

Until very recently, Interior had no system for the preservation of trust records. Over the life of the IIM trust, trust documents were routinely destroyed in accordance with record retention schedules at National Archives and Records Administration (NARA) centers. Tr. 2054:10-2056:18 (Christie). Most documents and data were retained for only six years pursuant to standard NARA policy, until 1995, when Interior instructed NARA to cease destroying trust records according to the six year schedule. Trust records from 1989 forward have not been subject to the document destruction schedules.

The loss of trust records before 1989 has been mitigated — to what extent is a matter of dispute — by the fact that, in many cases, Interior created several copies of IIM records. See Defendants’ Corrected Proposed Findings of Fact and Conclusions of Law [3459-1] at 8 (table listing types of IIM transactional documents, number of copies produced, and locations stored). Interior’s expert contends that the existence of multiple copies of trust records stored in different locations renders the loss of trust records immaterial, as the loss or destruction rarely impacted all copies of a given document. Tr. 1264:8-11, Tr. 1282:3-7, (Angel). Defendants have located and centralized 43 miles of Indian records potentially relevant to the accounting at the National Archives and the American Indian Records Repository (AIRR) in Lenexa, Kansas, Tr. 1198:14-21 (Angel), and they have access to another 10,000 cubic feet of MMS and United States Geological Survey (USGS) records potentially useful to those conducting the HSA accounting. Tr. 1181:5-10 (Angel). Problems related to the disorganized or poor condition of records were noted early in this litigation and have been addressed by defendants’ contractors at the AIRR. Interior’s Office of Historical Trust Accounting (OHTA) continues to work with NARA to collect data for the AIRR. Defendants have been surprised to discover that the records needed to perform the accounting are indeed available and accessible, Interior’s 2007 HSA Plan Part 1, AR-565 at 33-02-08, and they contend that plaintiffs’ complaints about the existence and condition of trust records are essentially based on outdated information.

Plaintiffs acknowledge, as they must, that an enormous volume of records has been collected at Lenexa, but their response, essentially, is “Compared to what?” They suggest that the volume of records is meaningless without some assessment of the total number of trust documents produced over the history of the trust. They point out that in 1998, the BIA estimated that a total of approximately 1.425 billion pages of IIM trust records existed. (Government experts later determined that that number had been overestimated by some 1.205 billion pages, partially because they had not adequately accounted for destroyed documents. DOI memo stamped June 2002, AR-80 at 54-21-32; Tr. 1266:13-1268:15 (Angel)). Plaintiffs’ witnesses testified that a great deal of data was missing from the IRMS database, Tr. 1746:9-1747:14 (Infield), that the manual entry of information into BIA’s Royalty Distribution and Reporting System (RDRS) was unreliable, Tr. 1749:24-1751:13 (Infield), and that some copies of documents necessary for determining the quantity of resources exploited from allotted land — -such as oil and timber — were never recorded electronically and had been routinely destroyed. Tr. 2043:19-2049:17 (Christie).

It is likely that, in some cases, all copies of timber scaling tickets, distribution settlement worksheets, and run tickets were destroyed, since relatively few copies of these documents were created. Tr. 2044:3-2045:6 (Christie). Joe Christie, a 27-year veteran of the Interior Department, opined that distribution worksheets are essential if the historical accounting project seeks to “verify who was considered the owners [of an allotment], the amount of funds that were collected, and who [the funds were] distributed to. It’s the only document, by the way, that does that.” Tr. 2042:5-8 (Christie). Plaintiffs’ witness Mona Infield, an Interior employee, also indicated that lease information was frequently destroyed. Tr. 1750:18-1751:13 (Infield). Before 1990, cleared disbursement checks that had been issued to IIM beneficiaries were destroyed by Treasury. Motions Hearing Tr. 171:14-172:4 (Locks, 11/24/98) (no checks available predating Oct. 1990); Oct. 19, 1994 Letter from Sandra Chambers (FMS) to James Parris (DOI), PX-3340 (records prior to 1987 not available); Nov. 22, 1994 Letter from Leo Warring (FMS) to Jim Parris (DOI), PX-3342 (records unavailable prior to 1988).

Interior has an abysmal record of failing to prioritize the maintenance and preservation of trust documents. Many court opinions, audits, and congressional committee reports have catalogued that record. As late as 1995, the Deputy Commissioner of Indian Affairs discovered that agencies were destroying “financial and lease documents ... required in the reconciliation process.” 3/07/95 Memo from Hilda Manuel to Area Directors and Agency Superintendents, PX-0350. After decades of neglect, it is impossible to imagine that all documents necessary to perform a complete historical accounting are presently accessible to Interior. Nevertheless, the agency has made an impressive (and very expensive) effort in recent years to find, scan, and preserve whatever documents still exist. If the government has failed to prove that all documents necessary for the completion of its historical accounting have been or will be found, 6 plaintiffs have also failed to establish that the problem of lost or destroyed documents renders the historical accounting project entirely pointless.

I.l.D. Predecessor historical accounting plans

The 1994 Act did not tell the Interior Department how, or how thoroughly, to perform the historical accounting that it mandated. Associate Deputy Secretary James Cason said that because the agency had no model to follow, it had to engage in an “iterative process where we learn and redesign, learn and redesign.” Tr. 65:2-9 (Cason).

The Department’s initial compliance effort was the work of the “Tiger Team;” a group of Officer of Trust Fund Management (OTFM), BIA, and MMS employees assembled soon after the passage of the 1994 Act to study trust accounting issues. Tr. 1753:16-1754:4 (Infield). In August 1995, the team issued a report entitled “IIM Related Systems Improvement Project: The Tiger Team,” which was to be implemented by the end of 1996. August 1995 Draft Tiger Team Report, PX-607 at 157. The Tiger Team report noted uncertainty about the accuracy of data entered in IRMS, and it criticized the internal controls and inadequate automated systems relied upon in managing the IIM trust. Id. at 2-8. The Team’s efforts appear to have focused more on reforms going forward than on historical accounting.

Two years after the issuance of the Tiger Team report, the first Special Trustee for American Indians Paul Homan finalized a Strategic Plan to Implement the Reforms Required by the American Indian Trust Fund Management Reform Act of 1994, PX-615. Homan submitted the report in April 1997 to the Secretary, who incorporated many of its recommendations into the 1998 High Level Implementation Plan (HLIP) for the Trust Management Improvement Project (TMIP), filed with the Court in August 1998. High Level Implementation Plan, PX-4154. The HLIP identified thirteen projects Interior planned to undertake to improve trust management and to correct shortcomings in data maintenance, records management, training, policies and procedures, system weaknesses and internal controls. Id. at 3-8. The 1998 HLIP called for the completion of the Internal Control Sub-project by June 1999, and of the whole HLIP within three years. As with the Tiger Team report, this project was focused on trust reform, though it devoted significant attention to historical weaknesses, particularly in the area of internal controls.

A revised HLIP was filed in this litigation on March 1, 2000 [Dkt. 438]. The expected dates of completion were extended, but many of the objectives remained the same, such as the intention of reviewing trust records to determine accurate title and resource management information. The updated HLIP remained focused on documenting and rectifying historical weaknesses, inconsistencies, and failures in the Department’s trust management operations.

In November 2001, after the Court of Appeals affirmed Judge Lamberth’s finding that defendants were in breach of their trust duties, the government filed an additional plan with the court titled “DOI Trust Reform: Interim Report and Road-map for TAAMS and BIA Data Cleanup” [Dkt. 990]. That report, prepared by Electronic Data Systems, Inc., identified significant problems with missing data, inconsistent records, and conflicting information, and it recommended reforms that were to be completed within three years. Id. at 21-30.

On July 2, 2002, after requests from House and Senate Appropriations Committees made during the FY 2001 and 2002 budgeting processes, Interior submitted a report to Congress that detailed its historical accounting plan. Report to Congress on the Historical Accounting of Individual Indian Money Accounts, AR-561; Tr. 58:5-9 (Cason). The report informed Congress that Interior planned to conduct a transaction-by-transaction reconciliation of all funds in the IIM accounts through December 31, 2000, and that it expected the historical accounting project to cost $2,425 billion to complete. AR-561 at 25-02-15-16. The report indicated that the historical accounting would result in transaction histories for all accounts that resembled the account reconstructions done for the named plaintiffs in this case. Id. at 25-02-30. The historical statements of account envisioned in that report were to include ownership information for each account: allotment number and ownership interest. Id. at 25-02-87. Statements would be provided for the estates of deceased beneficiaries whose IIM accounts and allotment interests had been probated, for closed accounts, and for transactions dating back apparently from the inception of the trust (no beginning date is listed in the report). Id. at 25-02-40. The historical accounting process would involve the reconciliation of leases, land ownership interests, and the collection and disbursement of all IIM funds. Id. at 25-02-92.

That report to Congress was prepared in response to House Appropriations Committee concerns about the costs associated with the historical accounting. Id. at 25-02-21. After its submission, the House Committee on Resources addressed a December 9, 2002 letter to then Interior Secretary Gale Norton, describing the report as “troubling in several areas,” and requesting that the Secretary “promptly consider ways to reduce the costs and the length of time necessary for an accounting [including] alternative accounting methods.” AR-184.

A month later, on January 6, 2003, the Interior Department issued an Historical Accounting Plan for Individual Indian Money Accounts (the “2003 HSA Plan”), PX-507, that was strikingly different from the accounting project envisioned in the 2002 Report to Congress. The scope of the accounting was narrowed to exclude accounts closed before October 25, 1994, transactions occurring after December 31, 2000, and transactions in closed accounts or in the accounts of deceased beneficiaries. Id. at 8-9. While the accounting described to Congress in July 2002 would have cost approximately $2.4 billion, the effort described in the 2003 HSA Plan was projected to cost $335 million over five years. Id. at 1. The implementation of the 2003 Plan began, though not quite according to plan, with the commencement of the Litigation Support Accounting Project in the fall of 2003. Tr. 95:20-98:2 (Cason).

The 2003 Plan — with its abandonment of a total transaction-by-transaction approach to land-based accounts in favor of a mixture of transaction-by-transaction and sta tistical sampling reconciliations — reflected Interior’s acquiescence to the House Committee’s request that it “consider all available options regarding the use of alternative accounting methods.” AR-184; 2003 Historical Accounting Plan, PX-507 at 17-18. Specifically, the Department indicated that, within the population of transactions occurring during or after 1985, all transactions over $5,000 would be reconciled, 10% of transactions between $500-$5,000 would be reconciled, and .31% of transactions under $500 would be reconciled. PX-507 at 63. In total, 160,000 transactions less than $5,000 would be sampled, and the sampled transactions would be pulled from every agency within Interior. For transactions occurring before 1985 (the “paper ledger era”), all transactions of $5,000 or greater would be reconciled, and transactions below $5,000 would be sampled. The sampling plan for these transactions was not yet fully developed at the time of the 2003 HSA Plan’s publication. Id. at 69. The government’s 2003 estimate was that this stripped-down historical accounting would be completed in about four years: Judgment and Per Capita accounts were to be reconciled by June 30, 2004, land-based accounts by September 30, 2006, IIM systems tests by September 30, 2006, and Special Deposit Accounts by December 31, 2006. 7 Id. at 32. The historical statements of account prepared for IIM account holders pursuant to the 2003 HSA Plan were to include Interior’s assessment of the accuracy of the account transaction history and sufficient information for IIM beneficiaries to “ascertain whether Interi- or has faithfully carried out its IIM Trust Fund duties.” Id. at 4.

On the same day it issued its 2003 HSA Plan, Interior filed a court-ordered Fiduciary Obligations Compliance Plan, PX-508. The purpose of this plan was to bring the Department into compliance with the 1994 Act. This plan recognized shortcomings in the HLIP and acknowledged the importance of verifying the accuracy of account balances, a step the government noted was crucial “no matter how carefully future transactions may be recorded.” Id. at 5. Defendants next issued a Comprehensive Trust Management Plan, which conceded failures to implement prior trust reform plans and announced a new strategic plan for doing so in the future [Dkt. 2050], The government filed its “To-Be” Trust Business Model and its Fiduciary Trust Model on March 15, 2005 [Dkt. 2882, attachments 2, 3]. According to the Special Trustee, implementation of the Fiduciary Trust Model would ensure future compliance with fiduciary obligations owed to IIM beneficiaries.

I.l.E. IIM-related projects before the 2007 plan

i. Paragraph 19 project

In 1999, after Secretaries Babbitt and Rubin and Director Gover had been found in contempt for their noncompliance with orders of this Court, Cobell II, 37 F.Supp.2d 6, 17 (D.D.C.1999), the government hired Arthur Andersen to help both the Treasury and Interior Departments comply with Paragraph 19 of the First Order for Production of Information [Dkt. 16]. Under Paragraph 19, defendants were to produce all “documents, records or tangible things which embody, refer to, or relate to the IIM accounts of the named Plaintiffs or their predecessors in interest.” Id. at 519. Treasury produced over 2,000 documents pursuant to Paragraph 19, and Andersen ultimately concluded that Treasury’s Paragraph 19 search had been “thorough [and] well-executed [at or above] industry practices.” Trial 1.5 Tr. 51:15-52:9 (Brunner 6/6/03). Interior searched approximately 80 facilities for documents responsive to Paragraph 19 and produced around 160,000 documents. Trial 1.5 Tr. 54:6-55:8, Tr. 66:5-7 (Brunner 6/6/03). A total of 37 accounts were analyzed during this project. The documents collected dated back to 1914. Interior’s 2007 HSA Plan Part 2, AR-566 at 33-03-12.

The records produced in response to Paragraph 19 were reviewed by Ernst & Young partner Joseph Rosenbaum in 2001. Trial 1.5 Tr. 53:3-10 (Rosenbaum 6/9/03). Rosenbaum’s report analyzed a virtual ledger of transactions reflecting the documents Interior had collected in response to Paragraph 19. Rosenbaum determined that the documents necessary for assembling transaction histories for the named plaintiffs and their predecessors were available, and that TFAS balances from December 31, 2000 were sufficiently supported by supplemental documentation. Trial 1.5 Tr. 56:15-22 (Rosenbaum 6/9/03). Supporting documentation was discovered for 86 percent of the 12,617 transactions reviewed, representing 93 percent of the total dollar value of those transactions, or approximately $1.12 million. Trial 1.5 Tr. 75:5-76:19, Tr. 77:6-19 (Rosenbaum 6/9/03). Only small variances were noted. The total cost of the project, however, was around $20 million. AR-566 at 33-03-12. Interior concluded that, although the documents necessary to complete adequate accountings are available, the accounting process is extremely expensive, often dwarfing the dollar amounts reflected in beneficiaries’ accounts.

Rosenbaum also performed an “expected versus actual” analysis by comparing information about leases, transactions, and ownership interests. Trial 1.5 Tr. 54:17-55:10 (Rosenbaum 6/9/03). He identified and analyzed the majority of leases associated with the named plaintiffs that related to farming and oil and gas extraction, Trial 1.5 Tr. 57:20-58:3 (Rosenbaum 6/9/03), finding a net of only $32.04 in unexplained differences between transaction ledger entries and leases. During the October 2007 bench trial, Associate Deputy Secretary James Cason testified that Interior understood the results of the Paragraph 19 project as indicating that, although there were errors in the accounts, the errors were relatively few, the errors tended to be small, and the errors were on both sides of the ledger. Tr. 62:12-21 (Cason).

The Paragraph 19 project ignored direct pay transactions, Trial 1.5 Tr. 44:13-23 (Rosenbaum), escheated interests, Trial 1.5 Tr. 7:18-23 (Rosenbaum), and other types of transactions. The analysis of disbursements typically ended when funds were disbursed from IIM accounts, not when beneficiaries received the funds. Trial 1.5 Tr. 84:25-85:3 (Rosenbaum 6/10/03).

ii. Mass cancellation project

Historically, Treasury checks were of unlimited payability — in other words, there were no temporal limits on when they could be cashed. That changed when Congress passed the Competitive Equality Banking Act of 1987 (CEBA), Pub.L. No. 100-86, 101 Stat. 552 (1987), which provided that, as of October 1, 1989, Treasury checks would be of “limited payability” and could be cashed for only one year from the date of issuance. Treasury Bulletin No. 90-03, DX-231 at 1, 18; Tr. 323:17-324:4 (Ramirez). CEBA also mandated the “mass cancellation” of all checks issued by Treasury that were at least one year old by April 1, 1991. DX-231 at 1, 3; Tr. 323:17-324:4, Tr. 325:13-20 (Ramirez). At that time, some 10 million Treasury checks from as early as 1954 remained outstanding; approximately 60,000 of these checks were IIM checks with a combined value of approximately $1.9 million. Mass Cancellation Project, Analysis of Treasury Listing, DX-225 at 8; Tr. 324:8-328:16 (Ramirez); BIA Office of Trust Funds Management (OTFM) Instructions on Completing Mass Cancellation Project, DX-217 at 1-2.

Certain provisions of CEBA were problematic as applied to IIM trust funds. For example, CEBA broadly prohibited re-crediting to agencies the funds associated with the cancelled checks. DX-231 at 4. Generally, an agency will deposit funds at Treasury to cover the payments of checks issued pursuant to its authority. Under CEBA, all funds that had been deposited to cover mass cancelled checks were to remain in the Treasury general account rather than revert back to the agencies. At the time, legislators apparently were not focused on the fact that some funds — like IIM and tribal trust funds — were not tax dollars, but were monies that were only being held by Treasury after their collection and before their disbursement.

To address this problem, BIA put Katherine Ramirez in charge of a “Mass Cancellation Project,” Tr. 323:11-15, Tr. 332:8-20 (Ramirez), whose mission was to identify IIM checks that had been mass can-celled and to re-credit their amounts to the proper accounts. Tr. 323:11-15, Tr. 331:8-14 (Ramirez); July 30, 1992 Letter from Mary Sandoval to Donald Gray, DX-207 at 1. Interior had to obtain a special appropriation for the purpose of re-crediting IIM accounts, and, in 1992, Congress appropriated $3 million for this purpose. See Department of the Interior and Related Agencies Appropriations Act, 1993, Pub.L. No. 102-381, 106 Stat. 1374, 1391 (1992). The mass cancellation project discovered that approximately 22,000 of the approximately 60,000 mass-cancelled IIM checks were zero dollar instruments, 8 and attempted to trace the 38,554 remaining checks to specific IIM accounts. Tr. 349:3-350:22, Tr. 352:18-353:4 (Ramirez).

BIA traced $616,736.31 — almost a third of the value of the 38,554 non-zero dollar, mass cancelled IIM checks — to specific accounts or to voided checks, 1993 OTFM CEBA Report, DX-221 at 2; Tr. 353:5-8, Tr. 354:11-356:7 (Ramirez), leaving approximately $1.3 million in mass cancelled IIM checks unresolved. DX-221 at 2. Of that $616,726.31, BIA re-credited approximately $278,000 to IIM accounts; $338,000 was apparently attributable to voided checks. Tr. 356:12-357:9 (Ramirez).

At trial, no evidence was presented as to Interior’s post-1993 progress in resolving the remaining approximately $1.3 million worth of mass cancelled IIM checks. Ramirez testified that Interior continues to maintain a fund totaling approximately $500,000 to pay claims on mass cancelled checks, although no claims against the fund have been made recently. Tr. 364:4-365:11 (Ramirez). It appears that reimbursement for mass cancelled checks occurs now only if an IIM beneficiary or BIA agency presents a mass-cancelled trust check to the agency for payment, and that unless and until such a check is presented, the trust funds associated with the mass cancelled check are not disbursed. Examining unrestored funds from mass can-celled checks is not part of Interior’s 2007 HSA project.

iii. 20-year tribal reconciliation project

The original concept of the reconciliation project was to reconcile both the IIM trust and the tribal trust. After a preliminary assessment revealed significant problems relating to missing IIM records, however, Arthur Andersen informed Interior that reconciling the IIM trust would be infeasible under the $12 million contract awarded by the government. May 1996 GAO Report to the Senate Committee on Indian Affairs Re: The Tribal Reconciliation Project, PX-710 at 3. Instead, Arthur Andersen endeavored to reconcile the tribal trust between the dates of July 1, 1972 and September 29, 1992. Tr. 2074:9-10 (Christie). The 20-year reconciliation project began in December 1992 and was terminated before completion in 1995, at a total cost of approximately $21 million. Tr. 2073:8-17, Tr. 2084:6-20 (Christie); PX-710 at 2.

The tribal reconciliation project was not an audit, but a contract governed by “agreed upon procedures” — in other words, a contract in which the client defines the scope and nature of the project. Interior’s tribal reconciliation project manager Joe Christie testified that the project accomplished its goal of reconciling tribal trust fund transactions “to the extent possible ... within the time frame we were given.” Tr. 2076:14-19 (Christie). However, because they encountered “lots of missing documents,” Tr. 2076:21 (Christie), they were typically unable to put together reconcilable packages for Arthur Andersen’s analysis that met the highest standard set forth in the agreed-upon procedures (referred to as the “C” standard). Indeed, the “vast majority” of the packages assembled by BIA workers did not meet this standard. Tr. 2076:21-2080:2 (Christie). When the reconciliation packages were ultimately presented to the relevant tribes, the tribes were told that they could accept or reject packages reconciled at levels other than the “C” level. Tr. 2080:3-9 (Christie).

The results of the reconciliation project were mixed and inconclusive. Approximately 320 reconciliation packages were provided to tribes after Interior terminated the project. At least some supporting documentation was located for 86% of the non-investment transactions reviewed, but the government could not verify the remaining 14% of those transactions — representing $2.4 billion — and abandoned an effort to test investment transactions after being stymied by instances of missing records. PX-710 at 5-6. The 1996 GAO Report on the Tribal Trust Fund Reconciliation Project reveals that BIA failed to certify that the reconciliation was performed in compliance with the contract’s agreed upon procedures (a modest certification standard). PX-710 at 3. Notably, the reconciliation did not “address the completeness of records,” nor calculate “receipts and disbursements that should have been recorded,” and it revealed that “BIA did not know the [complete] universe of leases.” Id. at 5-6.

This project did not involve an analysis of the IIM trust beyond Arthur Andersen’s initial conclusion that the problem of missing documentation within the universe of IIM trust records was more severe than within the universe of tribal trust records.

iv. TIME project

In 2000, Interior hired DataCom Sciences, Inc., to “examine the accuracy of the current ownership document informa tion in LRIS by comparing mandatory elements.” DataCom Time Project Report, PX-4352 at 4. This became known as the “TIME project.” DataCom received a random sample of 93 of the 239,311 tracts within the LRIS system at the Rocky Mountain Land Title Records Office, PX-4352 at 2, and analyzed 541 documents reflecting ownership information stored in LRIS at the time of the project. Operating on the assumption that the paper records used in the comparative analysis were accurate, Contempt II Tr. 3363:13-3364:2 (Nessi 2/01/02), DataCom determined that over 30% of the 541 LRIS documents analyzed contained errors of different types. PX-4352 at 2.

Interior later hired NORC to conduct a pilot study of the accuracy of probate entries contained in the Tract History Reports housed within LRIS. NORC Analysis of LRIS Tract History Reports, AR-405. After reviewing 99 probated IIM accounts, NORC reported no material errors and no evidence of an error rate as high as 33%, but the NORC report was narrower in scope than the TIME report and indicated that it was not to be used to extrapolate an “error rate” among probate entries. AR-405 at 4.

The differences between the NORC and DataCom findings appear attributable — at least in part — to different definitions of “error”. NORC criticizes DataCom for over-identifying errors, because Data-Com’s definition of error was tied to ‘conformance to requirements’ rather than ‘fitness for use.’ A misspelling would be a ‘conformance to requirements’ error, even if it did not affect the entry’s fitness for use in the administration of the trust. NORC Analysis of LRIS Tract History Reports, AR-405 at 50-02-06-08; Tr. 1079:3-9 (Scheuren). The DataCom report indicates only that “inconsistencies” between hard copy documents and LRIS entries were treated as errors. The NORC pilot study operated with a more inquiring definition of “error”. For example, documents missing from LRIS due to the probate backlog were not considered errors, nor were incorrect IIM numbers. AR-405 at 50-2-6.

Neither definition of “error” can be judged “right” or “wrong.” They are merely more and less sensitive to mistakes within the LRIS system. The error rate calculations of both approaches can be useful, if for different purposes. The record will not support plaintiffs’ submission that the high error rate found by DataCom is indicative of material errors throughout current LRIS ownership information. It is reasonable to conclude, however, that further analysis of LRIS ownership records should precede total reliance on that database for purposes of auditing or accounting for ownership interests.

v. St raw Man project

Dr. Scheuren drafted a document concerning a proposed “Straw Man” approach to the historical accounting during his early days as an Interior contractor, when he was “just learning about this work.” Tr. 1080:4-21 (Scheuren); Undated document titled “Adaptive Testing Approach for Phase 1 Straw Man Design,” AR-167; Undated Scheuren document titled “Straw Man Pretest,” AR-170; see also Tr. 1080:4-12 (Scheuren); Exchange containing 07/08/02 email from Scheuren to Edwards mentioning straw man proposal, AR-304 at 14-02-04. The “Straw Man Pretest” was prepared after meetings between NORC and Interior and included several questions for Interior to consider as it planned its data gathering efforts. AR-170; AR-304.

The documents describing the “Straw Man Pretest” focus on the benefits of adaptive sampling — an approach that is heavily utilized in both the 2003 and 2007 HSA plans. In one document, the government notes that “[t]he main benefit emphasized in an adaptive approach is obviously reduced expense,” and identifies other benefits such as scalability, speed, suitability for targeting of efforts, likelihood that results can be reported as “a series of successes,” use of pilots to benefit from lessons learned by prior pilots, and increased accuracy. AR-167 at 57-29-01-02. Adaptive sampling “links up nicely with the legitimate payment of a cash settlement, since inherently an adaptive approach admits error, even though it may well manage to keep it less than would have been the case in a full accounting.” Id. at 57-29-02. The “Straw Man” approach appears to reflect Interior’s initial reconsideration of the 100% transaction-by-transaction approach proposed in the 2002 Report to Congress: “the adaptive strategy ... is completely compatible with the ... [July 2, 2002] Historical Accounting Report. The only difference is that we do not necessarily have to test every transaction.” Id.

vi. Recent audits of the IIM trust

To the court’s knowledge, there has never been an unqualified independent audit of the IIM trust. “Qualified” audits are issued when the auditor is unable to comply with generally accepted auditing standards. In the case of the IIM trust, auditors have issued qualified audits after discovering cash balances that conflicted with those reported by the U.S. Treasury, internal control deficiencies, records management problems, and irreconcilable ledgers, among other things. See, e.g., AR-343; AR-347; AR-350; AR-352; AR-355; AR-369; AR-374; PX-575 at 3, 10-12. The first IIM trust audit performed by an independent auditor was the Arthur Andersen audit conducted in 1988. PX-575 at 3. That audit documented internal controls weaknesses attributable to inadequate training, lack of experienced supervisors, understaffing, and out-of-date accounting manuals, and concluded that the “accounting systems [were] unreliable.” PX-575 at 10-12. These concerns have been reflected in more recent audits as well. See PX-695; AR-633; AR-377; AR-374. The most recent independent audit — prepared by KPMG for the year ending on September 30, 2006 — noted that OST relied upon unreliable BIA data and unresolved financial reporting from prior periods, that OST’s processing of trust data relies upon the accuracy of information from BIA, MMS, and other bureaus and offices, and that “current management is burdened with the ongoing impact of decades of accumulated discrepancies in the accounting records.” AR-343 at 60-02-34.

I.l.F. Establishment of the AIRR in Lenexa, Kansas

In 2004, the government opened the American Indian Records Repository (AIRR) in an underground limestone mine in Lenexa, Kansas. AR-563 at 11-12; Tr. 378:19-379:17 (Ramirez). The AIRR is within NARA’s Federal Records Center, which is one of many government offices and private businesses occupying space in the 90 acre mine. Tr. 368:3-8; Tr. 370:13-14 (Ramirez). Retired Indian records are sent to Lenexa for storage and as a resource for research. The government now reports that 165,825 boxes of Indian records are stored at the AIRR (not all of which contain IIM trust records). OTR Activity Report for December 2007 (Jan. 15, 2008) [Dkt. 3479].

Two Interior offices have outposts at the AIRR: the Office of Historical Trust Accounting (OHTA) and the Office of Trust Records (OTR). Tr. 367:1-8 (Ramirez). OTR is within the Office of the Special Trustee for American Indians (OST). The AIRR is under the overall management of OTR, but OHTA requires access to trust records that are stored on-site, so OHTA sublets space from OTR and employs on-site contractors who search for (“searchers”) and code (“coders”) documents. Id. OTR also has an off-site annex facility, where boxes of records destined for the AIRR are initially received from BIA, OST, and regional offices across the United States. These sealed boxes are indexed, inventoried, and labeled according to strict procedures before they are sent into the cave and officially transferred from the custody of the DOI into the custody of NARA. Tr. 372:12-373:7 (Ramirez). Kathy Ramirez — the former manager of the mass cancellation project — manages OHTA’s contract staff at the AIRR. Tr. 366:25-367:1 (Ramirez).

Two primary systems are used for indexing, storing, searching for, coding, and scanning specific documents at the AIRR: the Box Index Search System (BISS) and Account Reconciliation Tool (ART). After they have been delivered and labeled, the boxes’ numbers and indexing information are entered into the BISS by annex contractors before the records are transferred to NARA for storage. Tr. 372:18-373:7, Tr. 377:5-378:9, Tr. 385:19-386:19 (Ramirez). OHTA’s accounting contractors (Clifton Gunderson) use the BISS system to identify records that will be potentially useful in conducting historical accounting work. Tr. 381:3-25 (Ramirez). After identifying potentially relevant boxes on the BISS, these “searchers” submit requests to NARA, and NARA delivers the requested boxes to OHTA’s searchers. Tr. 382:1-383:9 (Ramirez). Once the searchers identify relevant documents within the requested boxes, another group of contractors (the “coders”) images and codes the responsive documents and loads them into the ART system where they are accessible to OHTA contractors performing historical accounting work. Tr. 383:20-384:11, Tr. 385:3-11, Tr. 387:2-389:15 (Ramirez). Quality control measures are observed throughout the process.

It is evident from the photographs presented and testimony given at trial that the AIRR is a state of the art, climate-controlled, organized, and sizable facility suitable to the storage and research obligations of the Interior Department. Storage bays containing Indian trust records were constructed in accordance with the highest standards, the facility maintains low temperature and low humidity, and particulate matter and ultraviolet light are controlled. Tr. 370:17-19 (Ramirez).

While the facility reflects a significant improvement in the conditions and treatment of the records that have survived the cramped, disorganized, flooded, rat-infested storage facilities in which some of them were found, it is unclear from the evidence presented at trial what percentage of total IIM records that should have been maintained over the life of the trust have actually been recovered and shipped to the AIRR. Document requests still present vexing challenges for the considerable number of “searchers” working in the AIRR. At the time of trial, of the 80,000 requests FTI has submitted to AIRR in connection with its DCV work, approximately 35,000 requests remained pending, and some requests can take over a year to fill. Tr. 505:3-14 (Herman). As recently as June 2007, the government estimated that it could cost up to $37.4 million to search the AIRR for documents related to plaintiffs’ request for the trust records of less than one hundred specific beneficiaries. [Dkt. 3340] at 3-4. The development of the AIRR has streamlined the analysis of trust records to some extent, but difficulties in locating documents and the inability to certify that all relevant documents have been consolidated in the facility prevent a finding that the records management problems chronicled in earlier opinions have been fully rectified.

1.2. The 2007 Historical Accounting Plan

The historical accounting plan announced by Interior in May 2007 — the 2007 HSA Plan — is dramatically different from the far more ambitious plan announced four years earlier. It reflects Interior’s decisions, sometimes made for cost reasons and sometimes because of its legal interpretation of the statute, to rely much more heavily on sampling and statistics. Interior will reconcile only a small sample of IIM transactions (which it believes will yield statistically significant results), test its existing electronic data for completeness, test the process by which revenues collected from allotments are transferred into IIM accounts (again on a sample basis), and then begin issuing historical statements of account to IIM holders — with Interior’s assurance that they are correct.

I.2.A. Major changes to the Historical Accounting Project reflected in the 2007 HSA Plan

The Interior Department’s 2007 HSA Plan “builds upon and replaces” the 2003 HSA Plan. Interior’s 2007 Historical Ac-

counting Plan, AR-566 at 33-03-03. Many significant changes are reflected in the Department’s newest plan.

i. Cost and Schedule

When it adopted the 2003 HSA Plan, Interior estimated that the IIM historical accounting project would cost $335 million to perform and would be completed within five years. Between 2003-2007, however, not only did Interior receive only $127.1 million in appropriations for its IIM historical accounting work, but it also discovered that the accounting process it had envisioned would be both more costly and more time-consuming than it had anticipated. Tr. 67:24-68:7 (Cason). Even with the dramatic reduction in the number of land-based transactions that will be sampled under the 2007 Plan, Interior expects that the 2007 Plan work will not be completed until the end of 2011. That estimate is contingent upon congressional appropriations for IIM historical accounting totaling $144 between 2007 and 2011.

ii. Transactions and Accounts to be Reconciled

The 2003 plan relied on statistical sampling of lower value transactions in land-based accounts, but the number of transactions to be reconciled under the new plan is considerably smaller. A chart may help to illustrate the differences:

2003 HSA Plan_ 2007 HSA Plan_

Reconciliations of transactions from land-based accounts post-1985 (the “Electronic Ledger Era”) • transactions > $5,000: 100% transaction-by transaction approach (73,000 transactions) • transactions > $100,000: 100% transaction-by transaction approach (2,099 transactions)

• transactions btw $500-$5,000: 80,000 transactions sampled • transactions < $500: 80,000 • transactions < $100,000: 4,500 transactions sampled

transactions sampled samples must be drawn from every agency all agencies subject to sampling, but no requirement that samples be drawn from each agency

• projected total « 233,000 transactions reconciled— • projected total = 6,599 transactions reconciled—

AR-566 at 33-03-22 — 33-03-23 AR-566 at 33-03-22 — 33-03-23

Reconciliations of transactions from land-based accounts pre1985 (the “Paper Ledger Era”) • transactions > $5,000: 100% transaction-by transaction approach (100,000 transactions) • transactions of all dollar amounts subject to statistical sampling; unknown number of sampled transactions

• transactions < $5,000: 160,000 transactions sampled

• projected total ~ 260,000 transactions reconciled AR-566 at 33-03-22 • projected total = unknown AR-566 at 33-03-23 — 33-03-24

Reconciliations of transactions 100% transaction-by-from Judgment and Per Capita transaction approach accounts 100% transaction-by-transaction approach 86% complete; further reconciliations deferred indefinitely

• projected total ~ all transactions in 96,823 accounts reconciled • projected total = unknown (reconciliations of all transactions in at least 83,226 accounts complete)

AR-566 at 33-03-24 AR-566 at 33-03-24

The bulk of the reconciliation work that remains to be done under Interior’s 2007 HSA plan involves pre-1985 transactions in land-based accounts. Because most of the materials necessary for this work are in paper form, the process is considerably more time consuming. Even though only approximately 65,000 of the total 267,949 land-based accounts Interior intends to reconcile were opened prior to 1985, analyzing them is by far the most expensive and labor intensive effort. At the time of the trial, preparations for “paper ledger era” land-based account reconciliations had begun, but the work had not commenced in earnest.

Most of the other reconciliation work contemplated in the 2007 HSA Plan is complete. Interior has finished its analysis of 6,599 postAL985 transactions within land-based accounts, and has reconciled all transactions within 86% of the Judgment and Per Capita accounts. It has suspended reconciliation work on Judgment and Per Capita accounts in order to focus on the more time-consuming “paper ledger era” land-based accounts, and has deferred consideration of whether to complete transaction-by-transaction reconciliations of the remaining 14% of Judgment and Per Capita accounts. Interior plans to produce HSAs for all Judgment and Per Capita account holders, but it may conclude that reconciliation of transactions within the remaining approximately 13,600 accounts is unnecessary. In other words, its reconciliation work on Judgment and Per Capita accounts may also be complete.

iii. Special Deposit Accounts

Special Deposit Accounts (SDAs) are temporary accounts used to hold trust funds until they are identified for transfer to specific accounts. AR-566 at 33-03-27. The 2003 Plan revealed that $67.9 million in residual balances were in SDAs awaiting distribution. Id. As part of the historical accounting project described in the 2003 Plan, Interior planned to identify the proper owners of those funds and distribute the funds accordingly. Id. Interior reports that it has since distributed over $51 million in residual SDA balances at a cost of $48 million. Id. No documentation supporting those numbers is contained in the Administrative Record submitted for the October 2007 bench trial, but testimony at trial echoed the total figures reported in the 2007 Plan. Tr. 158:12-19 (Cason).

Interior’s current position is that “SDA work ... is not truly part of the historical accounting,” as money distributed from SDAs are “part of current accounting distributions.” AR-566 at 33-03-27. Consequently, while OHTA will continue to work on distributing funds in SDAs as part of its trust reform efforts, the resolution of these accounts is no longer an element of the HSA project. Id.

iv. Asset statements

The 2003 HSA plan reflected Interior’s intent to include, as part of its historical statements of all land-based IIM accounts, statements of land assets owned as of December 31, 2000. AR-566 at 33-03-28. In the 2007 HSA plan, Interior has changed its mind. Now statements of land assets will not be provided to beneficiaries in HSAs. Id. Instead, Interior indicated that it would begin to send combined statements that would include current land ownership, encumbrance, and financial information to IIM account holders in late 2007, after the TAAMS lease management module is fully implemented and the TFAS/TAAMS interface enables the agency to produce such statements. Id. This change reflects Interior’s judgment that producing land asset ownership statements as of December 31, 2000 would be costly and complicated, and would “add little value” to the historical statements of account. Id.

I.2.B. Rationale for changes to the historical accounting project

Several interrelated factors caused Interior to revise the plan it had adopted in 2003. Original cost and time estimates were off by several multiples, and Congress failed to appropriate the funds Interior had requested and expected. Interior initially estimated that reconciliation work would cost around $100 per transaction, but experience in recent years has revealed that reconciling a single transaction costs between $3,000-$3,500 when searching, imaging, coding, computer support, accounting services, quality review, and statistical analysis are taken into account. AR-566 at 33-03-09. While Interior estimated that the 2003 Plan would cost $335 million to implement, its received only $127 million in congressional appropriations for IIM historical accounting between 2003-2007. In any event, the agency now estimates that the 2003 Plan would have cost approximately $1,675 billion to complete. See AR-566 at 33-03-05 (noting that “2003 Plan is now estimated to cost about five times the original cost estimate”).

Interior expected that it could complete the historical accounting activities described in the 2003 plan — including the reconciliation of hundreds of thousands of transactions in land-based accounts — by 2007, but by year’s end it had completed a small fraction of those reconciliations. Indeed, at the time of trial, it had not even finalized its accounting plan for pre-1985 transactions in land-based accounts and the remaining Judgment/Per Capita accounts.

Judicial decisions and congressional action following the issuance of the 2003 HSA Plan affected the agency’s implementation of its historical accounting project. Shortly after the plan’s issuance, Judge Lamberth issued a structural injunction rejecting the Department’s plan and mandating a significant expansion of the historical accounting project. Cobell X, 283 F.Supp.2d 1 (D.D.C.2003). On November 10, 2003, less than two months after the issuance of that opinion (and after receiving estimates from the Interior Department that compliance with the structural injunction would cost the agency approximately $3 billion, Tr. 200:17-201:1 (Ca-son)), Congress included a provision in an appropriations act that granted Interior relief from performing an historical accounting of the IIM trust until “the earlier of the following shall have occurred: (a) Congress [amends the 1994 Act] to delineate the specific historical accounting obligations of the Department of the Interior with respect to the Individual Indian Money Trust; or (b) December 31, 2004.” Department of the Interior and Related Agencies Appropriations Act, 2004, Pub.L. No. 108-108, 117 Stat. 1241, 1263 (2003).

On December 10, 2004, exactly one year after the passage of the appropriations bill, the Court of Appeals for the D.C. Circuit reversed Cobell X, holding that the structural injunction had no legal basis after the passage of the 2004 appropriations bill. Cobell XIII, 392 F.3d 461 (D.C.Cir.2004). Congress failed to step up to the problem, however — it did not “delineate the specific historical accounting obligations of the Department of the Interior with respect to the Individual Indian Money Trust” before December 31, 2004 — so, on February 23, 2005, Judge Lamberth reinstated his structural injunction. Cobell XIV, 357 F.Supp.2d 298 (D.D.C.2005). The reinstated injunction was stayed soon thereafter, Cobell v. Norton, 2005 U.S.App. LEXIS 5788 (D.C.Cir. Apr. 7, 2005), and eventually reversed in late 2005, Cobell XVII, 428 F.3d 1070.

During this turbulent period, defendants did perform some historical accounting work, even if that work did not always conform to the scope or methodology described in the 2003 Plan. The results of work done during that period informed the agency’s adjustments to its historical accounting project. Defendants began to implement the 2003 plan for statistical sampling of post-1985 land-based transactions within the Alaska Region, but that beginning was aborted because of the judicial decisions and congressional actions described above. Tr. 971:16-972:9 (Scheuren). The land-based transaction reconciliation work performed after the issuance of the 2003 Plan was largely part of a project entitled “Litigation Support Accounting,” or LSA, which was conducted between October or November 2003 and November 2004. Tr. 199:1 (Cason); Tr. 974:7-12 (Scheuren). Interior borrowed the term “litigation support” from the appropriations bill discussed above, which temporarily halted both the agency’s obligation to perform an historical accounting of the IIM trust and, arguably, temporarily prohibited any work other than “litigation support” on land-based IIM accounts:

[0]f the amounts available under this heading not to exceed $45,000,000 shall be available for records collection and indexing, imaging and coding, accounting for per capita and judgment accounts, accounting for tribal accounts, reviewing and distributing funds from special deposit accounts, and program management of the Office of Historical Trust Accounting, including litigation support.

Pub.L. 108-108 at 23, (emphasis added). Noticeably absent from this list of permissible expenditures in the period between November 10, 2003 and December 31, 2004 was any allowance for reconciliation of land-based accounts. Interior interpreted the reference to “litigation support,” however, as authorizing an accounting project that involved analyzing — as well as reducing liability arising out of — a small sample of land-based accounts. Tr. 199:11-23, Tr. 205:16-19 (Cason). Part of the rationale for the Litigation Support Accounting project was that it would develop information that would facilitate. Congressionally directed settlement discussions that were oc curring during that time. Tr. 62:2-64:5 (Cason).

The government retained NORC to draw samples for the LSA project, which ultimately involved the reconciliation of 6,599 transactions from land-based accounts in the electronic era. NORC’s goal in performing the LSA project was to deepen its understanding of the system in order to better target future historical accounting efforts. Tr. 974:18-975:1 (Scheuren). NORC — along with various accounting firms assisting with the LSA project— learned many lessons while performing this relatively small number of reconciliations. First, it learned that the cost of reconciling transactions in land-based accounts is significant. Tr. 964:23-965:16 (Scheuren). Second, however, NORC encountered far fewer errors and missing records than it had expected to discover. Tr. 977:15-22 (Scheuren). The large sample sizes contained in the 2003 Plan had been based on Interior’s assumption — an assumption informed by historical reports, anecdotal evidence, and decades of criticism — 'that records would be missing, erroneous, and in disarray. Those assumptions, NORC concluded, were overblown and incorrect.

The positive results of the LSA project caused NORC to suggest that the Interior Department reassess the sample design in the 2003 Plan. Tr. 1003:3-11 (Scheuren). In fact, although NORC had reconciled less than 3% of the transactions to be reconciled under the 2003 HSA Plan, AR-566 at 33-3-22, it recommended that Interior cease conducting reconciliation work on post-1985 land-based transactions. NORC based this recommendation on the small number of errors it had encountered, and on its estimation that additional work would be expensive and would not meaningfully lower the margin of error around estimated error rates. Tr. 1004:17-1005:7 (Scheuren); 3/15/07 email from Scheuren to Edwards, AR^127 at 4-5. NORC did make certain recommendations for additional historical accounting work, which will be described in further detail below.

1.2.C. Producing the historical statements of account

Interior’s 2007 historical accounting project utilizes several different computer systems, contracting firms, and methodologies. It involves six discrete steps, which are set out in the 2007 HSA Plan in deceptively simple terms. AR-565 at 33-02-08-09. They will be described broadly below, and in further detail in the discussion that follows.

The first step is to gather credit and debit ledgers and to assemble data from each IIM account. Id. Next, the agency reconciles selected transactions (see chart above) to determine accuracy — comparing ledgers to contemporaneous records — and tests the integrity of the data systems. Id. After completing that process, the DOI petitions the court for permission to assemble and mail HSA packages to beneficiaries. Id. Upon receipt of the court’s permission, the agency assembles HSA packages and mails them to beneficiaries. Id. Each HSA package will contain a list of all transactions in the beneficiary’s account along with available references for each transaction, as well as a statement regarding the accuracy and completeness of the information provided. Id. Finally, the Department will implement an administrative appeal process, by which account holders may challenge information contained within the HSAs within a specific period of time. Id.

1.2.D. Information Contained in the Historical Statements of Account

Historical Statements of Account are to include two main elements: a report of the transaction history within each account, listing all transactions posted to IRMS or TFAS, and “Interior’s conclusions” regarding the accuracy and completeness of the accounting provided. AR-565 at 33-02-03-04; 33-02-06-11. Information that will not be included in the HSAs will be discussed in the further detail below.

1.2.E. Beneficiaries Receiving Historical Statements of Account

Under the 2007 HSA Plan, Interior intends to prepare HSAs for all IIM accounts that were open at any time between October 25, 1994 and December 31, 2000. For all IIM accounts open at any point on or between those dates, the agency intends to prepare an HSA listing transactions dating back to the opening of the account or June 24, 1938, whichever is later. Defendants will not prepare HSAs for accounts closed before October 25, 1994 or opened after December 31, 2000. AR-565 at 33-02-03-06. Defendants estimate that HSAs may be prepared for approximately 364,772 accounts. AR-565 at 33-02-11. It is worth noting that there is not a 1:1:1 relationship among accounts, beneficiaries, and allotments. A beneficiary may have multiple accounts relating to a single allotment, or a beneficiary may own multiple allotments that have never been leased and never led to the creation of IIM accounts. For this reason, the number of accounts reconciled will not be equal to the number of beneficiaries receiving accounting statements.

1.2.F. Transactions Reconciled under the 2007 HSA Plan

i. Judgment and Per Capita accounts

In its 2007 Historical Accounting Plan, Interior explains that

When a tribe receives money as part of a legal judgment or negotiated settlement, the tribe may elect or be required to pass along some, or all, of the money to its enrolled members or their descendents [sic]. Similarly, other income to the tribe may be distributed to tribal members in accordance with a tribal resolution specifying the amount of the payment and the eligibility of members to receive payment. In most cases, money is distributed directly to tribal members; however, for minors or other tribal members who are not eligible to receive direct payment, the money is deposited on their behalf into an IIM account.

Interior’s 2007 HSA Plan, AR-565 at 33-02-11.

The majority of the Judgment and Per Capita accounts — indisputably the easiest funds to reconcile — have been reconciled under an 100% transaction-by-transaction approach. Interior has no immediate plans to reconcile the remaining 13,122 funds; it has put that work on hold indefinitely in order to focus on higher priority accounting work. AR-565 at 33-02-13. The reconciliations that have already been completed were conducted pursuant to a contract with Chevarria, Dunne & Lamey and reviewed by the government’s quality control contractor, Grant Thornton. The steps followed by CD & L in reconciling Judgment and Per Capita accounts were as follows: (1) query IRMS and TFAS to produce account transaction histories; (2) use document indexes to identify record boxes; (3) search for records necessary to reconcile transactions; (4) compare account details with supporting documentation; (5) calculate interest based on historical interest factor and compare with interest posted; (6) produce and seek authorization to mail HSAs. Reconciliations were performed according to the Accounting Standards Manual. Some of the funds were listed in CD & L’s reports as “fully reconciled,” while others were listed as “partially reconciled.” Scant attention was paid to the reconciliation of these accounts in the October 2007 trial.

ii. Land-Based Accounts: 1985-2000

The LSA project reconciled non-interest transactions in land-based accounts posted on IRMS and TFAS between 1985-2000. The sample design for the LSA project involved a 100% reconciliation of the 2,099 transactions of $100,000 or more. The sample of lower value transactions reconciled consisted of 289 transactions taken from the Eastern Region, 442 transactions from the Alaska region, and, initially, a sample of approximately 20,000 transactions from the remaining 10 regions (the “national sample”). Tr. 990:12-994:10 (Seheuren); see chart § I.2.A.Ü. The 20,-000+ transactions in the national sample were drawn from 1,020 accounts across 10 regions, and were divided into four replicates of approximately 5,000 transactions each. In the end, the LSA contractors only analyzed transactions from one of the four replicates. Tr. 993:7-994:7 (Scheuren). The transactions analyzed were within a replicate of 5,128 transactions, but contractors attempted to reconcile only 3,769 of those transactions because the others were apparently outside the scope of the analysis (interest transactions, transactions occurring after 2000, etc.). Between the Eastern Region sample (289 transactions), the Alaska Region sample (442 transactions), and National Sample replicate (3,769 transactions), contractors attempted to reconcile 4,500 lower-value transactions pursuant to the LSA project.

Accounting firms working with the sampled transactions provided by NORC were able to reconcile 2,363 of the 2,372 debit transactions sampled, and 2,117 of 2,128 credit transactions sampled. Tr. 997:1-12, Tr. 1000:2-12 (Seheuren); AR-438 (tables 4 and 7). A posted transaction from the IRMS or TFAS database was considered “reconciled” if contractors found supporting documents sufficient to determine whether the transaction was correct or erroneous. Tr. 997:5-8 (Seheuren); see Section I.2.G. below. No errors were found among the reconciled debit transactions, and 25 errors were found among the reconciled credit transactions. Tr. 997:1-12, Tr. 1000:7-20 (Seheuren). For statistical purposes, NORC also considered all unreconciled transactions erroneous (nine debit transactions and eleven credit transactions). 9/30/2005 NORC LSA Report, AR-438 at 14, 17. Both variable and attribute sampling were employed in the LSA project, as contractors noted both the presence of errors as well the total amount of dollars in error.

The population from which samples were drawn for the LSA — the “target population” — was not the entire population analyzed under the 2007 plan, nor was it the entire “electronic era” population. Tr. 2149:1-8, Tr. 2169:119-22 (Hinkins). Certain transactions from the electronic ledger era were not in the population sampled in the LSA project, as not all transactional data from the “electronic era” had been restored to the data set when the samples were drawn. Tr. 689:11-24 (Herman).

NORC acknowledges that the positive inferences that can be drawn from the LSA work are only applicable to the population sampled — that is, to transactions between 1985-2000 in land-based accounts available in electronic form at the time the sampling was conducted. Tr. 2152:9-16, Tr. 2152:17-2153:3 (Hinkins). Interior believes that other projects will address transactions missing from the target population in the LSA work, such as the Data Completeness Validation and Land-to-Dollars tests discussed below. Tr. 2171:5-8, Tr. 2172:1-5 (Hinkins). Interior contractor Michelle Herman also suggested that Interior plans to draw another sample after the entire data set is restored, but no defense witnesses testified as to the details of such plans at trial. Tr. 574:8-574:1, Tr. 702:16-24, Tr. 704:10-14 (Herman).

Indeed, defendants’ position on post-LSA sampling and reconciliation of land-based transactions within the “electronic era” after data restoration projects is confusing. In some instances, Interior indicates that it has completed its sampling and transaction reconciliation of “electronic era” land-based transactions through the Litigation Support Accounting project described above. See, e.g., AR-566 at 33-03-14 (“no further reconciliation is necessary for the electronic era”). Elsewhere, the agency expresses an intention to conduct further sampling after data gaps are restored to the IRMS and TFAS databases. See, e.g., AR-566 at 33-03-19 (“Interior plans to reconcile a sample of these restored transactions to determine if they have an error rate significantly different from that found in the LSA sample.”); Susan Hinkins Responding Expert Report [Dkt. 3393-2] at 23 (“[T]he 2007 Plan clearly shows that DOI has not claimed to have completed tests for the Electronic Ledger Era.”).

A close reading of the 2007 HSA Plan reveals that further testing of restored, “electronic era” transactions from land-based accounts is not a task that Interior will necessarily accomplish before land-based account HSAs from the “electronic era” are mailed. According to the plan, the completion of the Data Completeness Validation, Interest Recalculation, and Land-to-Dollars Posting Tests constitute “all the [remaining] steps necessary to mail land-based account HSAs[.]” AR-566 at 33-03-18. Because the government represents (1) that the mailing of HSAs discharges its accounting duty and (2) that HSAs can be mailed to beneficiaries prior to farther reconciliation of land-based transactions within the electronic era after the data set is restored, my reading of the 2007 Plan is that Interior considers the land-based electronic era reconciliation work to be complete.

iii. Pre-1985 transactions in land-based accounts

Interior’s plan for sampling and reconciling “paper era” transactions — a project Interior has yet to begin — remains vague. Defendants estimate that approximately 65,000 of the nearly 268,000 land-based IIM accounts within the historical accounting population were open before 1985. AR-565 at 33-02-10. This reconciliation work will be more expensive, since the records are older. They may have been destroyed. If they can be located at all, they must be translated into digital form and uploaded into an electronic database before reconciliations can be attempted.

On May 25, 2007, NORC recommended to the Interior Department that it begin its “paper ledger era” land-based reconciliation work by testing the hypothesis that the error rate in the “paper ledger era” is similar to the error rate in the “electronic era.” 5/25/07 memo from Scheuren to Edwards, AR-426. Six days later, DOI officially adopted this approach by including it in its 2007 HSA Plan. AR-565 at 33-02-15.

The initial hypothesis recommended was influenced by, among other things, a meta-analysis conducted by NORC. A “meta-analysis” is a process by which statisticians consider studies from several different sources on similar or related subjects “to draw together what is known about the subject”. Tr. 1019:14^19 (Scheuren). The purpose of NORC’s meta-analysis was to examine whether historical studies of the IIM trust — an incomplete but voluminous collection of around 900 documents, including some 263 audits — suggested that the proposed hypothesis was way off base. Tr. 1020:9-1021:6, Tr. 1043:14-21, Tr. 1066:21-25 (Scheuren). NORC’s meta-analysis played a minor role in the adoption of the initial hypothesis for “paper ledger era” testing: it did not result in the hypothesis, but was instead one technique NORC used in considering whether the hypothesis was within the realm of possibility. Tr. 1021:16-1022:2 (Scheuren). Dr. Scheuren described it as a “due diligence step.” Tr. 1020:3 (Scheuren). Nothing NORC uncovered in its meta-analysis caused the statisticians to reconsider their initial hypothesis that the error rate in the “paper ledger era” resembles that of the “electronic ledger era”. Tr. 1022:4-7 (Scheuren). Complete copies of all documents relied upon in the meta-analysis have not been disclosed, nor has NORC’s internal work product summarizing those documents. See, e.g., Tr. 1443:25-1444:6 (Duncan).

Following NORC’s recommendation, Interior will draw a small sample of accounts from a virtual electronic ledger compiled from the “paper ledger era,” and a small sample of transactions from -within those accounts. AR-565 at 33-02-15; Tr. 1018:9-17 (Scheuren). The exact size of those samples has not been revealed. If Interior discovers that the error rate in the “paper ledger era” is similar to that of the “electronic ledger era,” the conclusion will be that “the findings of the electronic ledger era apply to the paper ledger era as well” and no further sampling will be conducted. AR-565 at 33-02-15. If, however, they are not, the Interior Department plans to broaden the sample of paper era transactions, and will probably have to extend its projected schedule to complete its work. Tr. 269:7-19 (Cason).

I.2.G. Reconciliation Process

i. Contractors involved in the reconciliation work

Interior’s historical accounting project now underway is a classic case of government outsourcing. It employs five accounting firms, two historian firms, and firms to assist in statistical matters and project design. AR-565 at 33-02-07. Of the five accounting firms, three (Clifton Gunderson LLC, Deloitte & Touche LLP, and Reznick Group, P.C.) are performing reconciliations and account analyses; one (FTI Consulting, Inc.) is conducting forensic accounting; and one (Grant Thornton LLP) is performing quality control reviews of all the historical accounting activities. Id.

FTI worked on the first three steps in the historical accounting as detailed in the 2007 Plan, Tr. 438:2-441:20 (Herman), and has been primarily responsible for gathering account data on IIM transactions, Tr. 438:2-10 (Herman). FTI has also done reconciliation work as part of the LSA project, along with the other accounting firms. Tr. 438:18-439:13 (Herman). Michelle Herman is a Managing Director of FTI Consulting. Tr. 434:7-11 (Herman). Over the past ten years — practically her entire post-graduate career — she has worked for three contractor firms involved with the historical accounting project. Tr. 437:9-17, Tr. 435:4-6 (Herman).

The National Opinion Research Center (NORC) works on statistical matters. AR-565 at 33-02-7. Doctors Fritz Scheuren and Susan Hinkins have led most of the historical accounting work at NORC. Morgan, Angel & Associates, LLC, and Historical Research Associates Inc., which specializes on Indian issues, provide information on leasing, the allotment process, reservation histories, and other matters. AR-565 at 33-02-07. The five accounting firms and NORC have contributed to the development and updating of the Accounting Standards Manual. AR-8 at 2.

■ NORC was engaged to address measurement and quality issues for Interior’s July 2, 2002 report to Congress; statistical sampling was not a significant consideration at that time. Tr. 982:12-20 (Scheuren); AR-561. NORC was thereafter engaged to design and propose a sampling approach for Interior’s January 2003 Plan. Tr. 934:11-15 (Scheuren). NORC designed the sample for testing by Grant Thornton with respect to performing quality assurance reviews of reconciliation work performed by Interior’s accounting contractors. Tr. 1032:9-17 (Scheuren).

ii.Accounting Standards Manual

Contractors doing the reconciliation work are guided by the Accounting Standards Manual (ASM), a 581-page compilation of historical accounting standards. AR-566 at 33-03-11. Its purpose is to ensure quality and uniformity among the various firms performing historical accounting work; it is also used by the firm performing quality control services. Tr. 2109:20-24 (Dunne). The ASM governs both “paper ledger era” and “electronic ledger era” reconciliations. Tr. 580:2-12 (Herman); Tr. 2111:9-19 (Dunne). It details a reconciliation process in three steps: (1) identifying document types that might support a given transaction and searching for those documents; (2) tracing the transaction back to its initial entry in the IIM system using documents collected in step (1); and (3) quality control procedures, both within the firm conducting the reconciliation and by the QC firm, Grant Thornton. Accounting Standards Manual, AR-8 at 44-01-13. Contractors are to assess reconciliation results on a scale of 1 to 4, depending on how conclusively the supporting documents verify the transaction posted. The ASM makes reference to auditing standards, but as the historical accounting work is a “consulting engagement” and not an audit, accountants and other consultants using the ASM are not bound by those standards. Tr. 2119:16-19 (Dunne).

iii. Materials Relied Upon in the Reconciliation Process

Pursuant to the Accounting Standards Manual, accountants reconciling transactions are to specify whether documentation supporting a particular transaction is “level 1,” “level 2,” “level 3,” or “level 4.” All documents within levels one and two are utilized or created by the federal government in the process of managing the IIM trust; a document meets the level one standard only if it is “utilized or created by the government in the normal operations of processing [IIM] transactions.” Tr. 2112:15-19 (Dunne). Level three is the subject of considerable controversy, as it may be assigned case-by-case, using “alternative procedures” undefined by the manual, thus allowing accountants to use their judgment in validating transactions with atypical supporting documentation. Tr. 2113:8-23 (Dunne).

Well over half of the 581-page ASM is comprised of examples of supporting documentation that may be used in the reconciliation process. Different types of supporting documents are used for reconciliation, depending on whether the transaction is a receipt or a disbursement. Tr. 2121:16-2122:5 (Dunne). A receipt (or credit transaction) must be linked to a document reflecting expectation of a collection, such as a lease. Tr. 2121:21-23 (Dunne). A disbursement (or debit transaction) requires evidence that the government issued the check, such as a check register. Tr. 2122:2-18 (Dunne).

iv. Systems relied upon

The reconciliation work uses transactions systems (IRMS and, later, TFAS), and ownership systems (LRIS and, later, TAAMS). In order to reconcile “electron ic era” transactions, DOI contractor FTI loaded data from TFAS and the IIM module of IRMS' — along with restored transactions discovered in the course of the Data Completeness Validation project— into a database called the “SQL server”. Tr. 449:13-24, Tr. 450:5-17, Tr. 453:1-454:17 (Herman). The financial information has also been loaded onto the Account Reconciliation Tool, or ART. Tr. 596:15-25 (Herman). The ART, demonstrated at trial, is a powerful database that allows for extensive searching, linking, and reconciling of transactions with supporting documentation. Tr. 599:1-644:14 (Herman). Because the IRMS database was inconsistently used across agencies and some transactions are missing from all systems, the database is incomplete. Tr. 1036:8-15 (Scheuren); Tr. 2042:1-4 (Christie). The Data Completeness Validation project (described below) aims to find “electronic ledger era” data that is not available on IRMS or TFAS and then input the “restored” data into the ART as it is discovered. Ownership information utilized in performing transaction reconciliations is typically derived from the LRIS excerpt BIA provided OHTA at the beginning of its historical accounting work. Tr. 614:10-17, Tr. 639:24-640:5, Tr. 760:6-8 (Herman). Contractors engaged in reconciliation work rely on that excerpt in determining ownership interests in allotments, although the LRIS ownership data is incomplete.

v. Definition of error

The validity of defendants’ definition of “errors” in its historical accounting analysis was the subject of considerable testimony, confusion, and disagreement at trial. It is not for me to determine the best “error” definition for purposes of statistical analysis, but it is worth sorting out what defendants’ error rates do and do not reflect.

Defendants’ plan employs both attribute and variable sampling. Attribute sampling answers yes or no questions (such as whether there is a discrepancy between a transaction recorded on IRMS and the supporting documentation), while variable sampling records the value of the error/difference (such as, “recorded transaction is for $10 less than supporting documentation would indicate”). As I understand it, there are differences in the definition of “error” between the different historical accounting tests. For example, in discussing the LSA project, Dr. Scheuren testified that any transactions LSA contractors attempted but were unable to reconcile were treated as errors, even though they had no reason to believe that the inability to reconcile those transactions reflected errors. Tr. 997:13-1001:15 (Scheuren). Dr. Hinkins also testified that, when contractors informed NORC that they were unable to reconcile transactions, NORC’s statisticians treated those unreconciled transactions as errors. Tr. 2155:9-22 (Hinkins). Michelle Herman testified that, in conducting transaction reconciliations, any disparity of two cents or more was recorded as an error. Tr. 615:21-616:2 (Herman). After some internal debate within Interior, the agency decided that, in the transaction-by-transaction reconciliations, the failure to locate supporting documentation should be recorded as an error. Tr. 1949:15-20, Tr. 1958:15-17 (Zippin).

In other instances, however, it appears that the failure to locate supporting data is not recorded as an error. The 2007 HSA Plan notes that an error indicates only that “the actual transaction posting to an account is different from the amount expected to be posted based on the contemporaneous records.” AR-565 at 33-02-30. James Cason testified that an inability to find supporting data is not automatically recorded as an error. Tr. 187:14-188:8, Tr. 190:20-191:1 (Cason). Michelle Herman testified that she “consider[s] an error a transaction that wasn’t posted correctly, not transactions that are missing from the record set.” Tr. 789:15-20 (Herman). It appears that the definition of error employed by NORC in its statistical analysis of the LSA reconciliation work, and relied upon for the claimed 1% error rate, is broader than the definition used elsewhere in the historical accounting project. Error statistics, however, do not reflect the failure to reconcile transactions intentionally left out of the LSA sample because of expected reconciliation difficulties, NORC Interim LSA Report dated 12/28/04, AR-416 at 51-09-17, or transactions left out of the sample because they had not yet been “restored” through the DCV project.

I.2.H. Data Completeness Validation Project

Michelle Herman has designed and directed the Data Completeness Validation (DCV) Project for Interior: a battery of tests on IIM data that together comprise a major component of the historical accounting project referenced in the 2007 Plan. According to Interior’s plan, the project goals are “(1) identifying and resolving gaps in electronic data, (2) verifying the transfer of accounts and balances through system conversions (from paper ledgers to IRMS and from IRMS to TFAS), and (3) assessing the integrity of the underlying electronic data.” AR-566 at 33-03-15. Ms. Herman testified that the primary goal is “to assess the completeness of the electronic ledger era for historical statements, and when I mention completeness, it’s of the data today versus the data as posted historically.” Tr. 458:2-6 (Herman).

This project has been a massive undertaking. At FTI, between four and eight employees have worked between three and four years conducting DCV testing on some 113 million transactions, or 50,000 annual man-hours, Tr. 502:19-503:9, Tr. 465:9-21 (Herman), not including the work of other contractors who support DCV testing through projects like scanning and coding, etc.

The DCV project carries out tests for electronic data gaps and system conversion that were contemplated in the 2003 HSA Plan. Tr. 451:1-10 (Herman). The data gaps test is to determine whether the electronic data stored in the IRMS system was complete, and the system conversion test was to make sure that data were successfully transferred from the IRMS system to the TFAS system when Interior rolled out TFAS between 1998-2000. Id.; Tr. 450:20-21 (Herman). The DCV project primarily relies on information taken from the IIM module of IRMS from between 1985-2000 and information from TFAS from between 1998-présent. To some extent, the DCV project has also relied upon TAAMS in the account number analysis. Tr. 453:9-11 (Herman).

FTI has produced a national report describing its DCV work in all 12 BIA regions, and an additional six more detailed reports on six specific regions. Tr. 455:1-12 (Herman). The six regions for which DCV work is largely complete are the largest ones, including the first three to convert from IRMS to TFAS. Tr. 456:15-18; 21-24 (Herman). Within each of the regional reports, FTI describes the months and years of available data for each agency within the region, the transactions restored to the database, accounts out of balance, results of the “reused accounts” test, and the results of transaction mapping efforts. Tr. 482:2-483:12 (Herman); see DX-153A, DX-154A, DX-155A, DX-156A, DX-157A, DX-158A (updated regional reports).

Transaction mapping is another component of the DCV project. Tr. 465:N4 (Herman). Transfer transactions — such as a debit from an SDA account and a credit to one or more IIM accounts— should sum to zero, so when FTI identifies instances where that has not happened, it attempts to identify documentation explaining the apparent (or actual) discrepancy. Tr. 484:15-484:17 (Herman). Efforts to map transactions can result in one of several outcomes: (1) the transfer can be mapped successfully (sum of transfers is zero); (2) the transfer can be an “explained difference” (sum of transfers is not zero, but not because an error was made); (3) the transfer can be unresolved (sum of transfers is not zero, FTI awaiting further documentation to complete forensic analysis); (4) the transfer can reveal a posting error; or (5) the transfer can indicate that data is missing from the system. Tr. 496:14-20 (Herman). Bookkeeping mistakes uncovered during transaction mapping will be reflected on HSAs in the “statement of known difference” column. Tr. 520:1-7 (Herman). Transaction mapping takes place electronically in almost all cases. Tr. 498:l-499:-l, Tr. 500:4-7 (Herman). When discrepancies cannot be mapped with electronic information on the SQL server, FTI will put in a request (through OHTA) for searchers to look for supporting documents at the AIRR in Lenexa. Tr. 504:16-23 (Herman). If the searchers find supporting documentation, the documents are loaded onto the ART system which can be accessed by FTI in Los Angeles. Transaction mapping was helpfully demonstrated by Michelle Herman during the October 2007 trial.

Examples of mapping activities include comparing balance files and transactions within a single account (the “roll-forward analysis”), Tr. 525:13-528:15, Tr. 552:2-21 (Herman), analyzing the transfer of account balances from paper ledgers to the IRMS system, Tr. 523:23-525:2 (Herman), and check disbursement mapping, to follow disbursement data available in the database through to Treasury records. Tr. 535:4-536:13 (Herman). When FTI workers (or other DCV contractors) discover a hard copy document (such as a bill for collection, journal voucher, or a transaction register print out) indicating that a certain transfer took place that is not reflected on the IRMS and TFAS systems available on the SQL database, they input the transaction into the database in an effort to complete the data set: these are referred to as “restored transactions.” Tr. 529:19-530:9 (Herman). Transactions occurring both before and after 1985 have been restored to the electronic data set, so that the database will ultimately be utilized for “paper era” accounting work, too. At the time of trial, approximately 113 million transfers had been mapped as part of this effort. Tr. 551:16-19 (Herman). FTI had placed approximately 80,000 DCV-related documents requests of the AIRR searchers; 35,000 remained outstanding. Tr. 505:3-10 (Herman). It can take searchers from a week to longer than a year to find a document at AIRR. Tr. 505:11-14 (Herman).

Because account numbers have been reused over time, HSAs cannot be prepared according only to account numbers. The DCV project thus assigns each account a new, unique “Native American Account Number” or NAAN, and each beneficiary a unique “Native American Beneficiary Number” or NABN. Tr. 460:1-461:11 (Herman).

FTI discovered approximately 645,000 IIM accounts that are open, or have been open, on the IRMS and TFAS systems between 1985 and present. Tr. 459:11-22 (Herman). All but 786 accounts had been analyzed in the re-used account test at the time of trial. Tr. 487:15-24 (Herman). This set of accounts is broader than the set of accounts for which Interior is performing an historical accounting. Tr. 487:25-288:3 (Herman).

ii. Results ofDCV Testing

FTI and other contractors had restored 451,875 transactions to the electronic data set at the time of trial: 276,118 of those transactions occurred after February 1985, and 175,757 occurred before February 1985. 9/30/07 DCV Interim Report, DX-152A at 20; Tr. 721:21-722:15 (Herman). The record does not establish what percentage of originally missing transactions this figure represents. Plaintiffs’ expert suggested that the number may be in the realm of only 7%, but because significant pieces of information were not included in his calculations, I do not consider his estimate reliable. Tr. at 1387:13-1390:14, Tr. 1469:19-1476:18 (Duncan). FTI has also mapped over 6 million of the 6.4 million checks issued to IIM beneficiaries during the “electronic era.” Tr. 535:2-536:13 (Herman). More than 97% of the checks analyzed between 1987 and 2002 were successfully mapped. Tr. 565:2-566:6 (Herman). At the time of trial, FTI had confirmed that 266,285 accounts successfully transitioned from IRMS to TFAS. DX-152A at 20. Of the 355,320 accounts FTI has identified as falling within the scope of the historical accounting Interior now plans to accomplish, 54,618 accounts still have unreconciled differences or data gaps, which FTI believes is largely attributable to the unavailability of many balance files before 1989. DX-152 at 271. FTI has completed mapping the majority (93.9%) of IRMS and TFAS transactions from the electronic era. DX-152A.

I.2.I. Posting Test/Land-to-Dollars Test

i. Test purpose

The LSA project looks only at transactions posted to IRMS and TFAS, it does not detect the “failure to collect, deposit, and record collection transactions.” 3/31/07 memo from Jeffrey Zippin to Susan Hinkins re Land to Dollar Completeness Test at Horton Agency, AR-435 at 38-01-01. To identify any such omissions, Interior’s historical accounting plan includes a test designed to start with the land and examine whether revenue generated by IIM allotments has actually made it into the proper IIM accounts. Tr. 2121:13-15 (Dunne). The goal of this Land-to-Dollars (or “posting”) tesh — which is to be conducted on a sample basis — is to confirm that funds generated by leased allotments were actually received and properly entered into the IIM system. AR-565 at 33-02-21. Since Interior has only recently instituted an accounts receivable system, this test responds to a very credible concern that not all payments owed by lessees were received into the IIM trust.

ii. Initial effort: the Alaska prototype

In 2003, NORC documented its initial effort to design a sample for such a test. AR-389 at 47-03-04. Thirty-three land-related documents were selected from Lee’s Summit in the Alaska Region. The process of linking these documents to transactions proved to be challenging, as the documents necessary for the test were in multiple locations, in formats and maintained according to systems that were subject to wide variations between regions. Id. at 47-03-10. NORC’s 2003 conclusion was that the sampling process utilized in the Land-to-Dollars testing needed to be “rethought.” Id.

iii. Pilot Test at the Horton Agency

In 2006, Interior conducted a pilot Land-to-Dollars test at the Horton Agency in Kansas. 3/31/07 memo from Susan Hinkins to Jeffery Zippin re: Land to Dollar Completeness Test at Horton Agency, AR-435 at 38-01-01. The Horton Agency was chosen because it processed farming and grazing leases, making more straightforward document collection possible, and because agency staff were not already burdened by requests from the LSA project. Id. at 38-01-02. Interior does not claim that the results of this test are representative of collection and posting patterns throughout the IIM trust, and expects that further testing in other locations will be more challenging. Id. Interi- or does not have a complete list of land records, and compiling a complete list of such documents would be, at the very least, extremely difficult. Id. at 38-01-01. Trial testimony revealed, indeed, that lease information in some cases has been routinely deleted from Interior’s electronic systems. Tr. 1750:18-1751:13 (Infield). NORC, however, believes that a complete inventory is not necessary to successfully conduct these sorts of tests, AR-435 at 38-01-02, and Interior has adopted this view. Tr. 247:5-17 (Cason).

NORC discovered that land records at the Horton Agency were in good order. AR-435 at 38-01-02. NORC analyzed 21 of the 812 allotments within the Horton Agency, AR-435 at 38-01-03, reviewing agency data to determine what sort of revenue generating activity was processed by Horton between 1985-2000. Of the 21 allotments analyzed, only ten showed evidence of farming and grazing revenue generation during that time period; two showed evidence of potential revenue from subsurface interests which were not analyzed. Id. For the ten allotments that did appear to have farming/grazing revenue during those 16 years, NORC drew a sample (a sample of a sample) of expected annual revenues. Id. at 38-01-04. Revenue generating documents were located for 33 of 35 samples analyzed, and for 32 of those 33 samples (all of the ones within the temporal scope of the study), accountants verified that “revenue was received for the listed owners.” Id. There is no indication that the test considered the accuracy of revenue postings, or whether the “listed owners” were the correct owners. The test was performed on a very small sample of allotments, and no statistical conclusions can be drawn from it. AR-435 at 38-01-03.

iv. Plans for future posting tests

OHTA plans to conduct more land-to-dollars tests: at least one test for each of the twelve BIA regions. Michelle Herman testified that some were underway at the time of the evidentiary hearing, but no evidence reflecting these tests was introduced or contained in the Administrative Record. Tr. 587:3-17 (Herman); AR-566 at 18. Nor was any evidence presented to demonstrate how (or whether) Interior plans to address apparently missing lease data: a problem encountered in two of the 35 samples analyzed in the Horton Agency pilot. AR-435 at 38-01-04.

1.2. J. Interest Recalculation Project

Each historical statement of account will include interest posted to the beneficiary’s account in the electronic record. Tr. 115:22-25 (Cason). Interior will perform independent recalculations of the interest posted in the ledgers to confirm that the interest listed in the HSAs reflects the interest actually earned. Tr. 114:15-19 (Cason). The interest postings listed in each account will be recalculated using historical interest factors. Tr. 1963:23-1964:4 (Zippin). IIM funds earn interest in the following manner: monies from IIM accounts that are entitled to earn interest are pooled for investment in certain types of securities or bonds that pay a dividend or earn interest. Tr. 1966:18-24, Tr. 1967:7-13 (Zippin). At certain intervals, the interest earned by the pooled funds is transferred into a single Special Deposit Account. In accordance with applicable interest payment rules, the dollars in that SDA are then divided by the total IIM dollars in accounts to which the interest is applicable. That fraction is then used to calculate interest owned to eligible IIM accounts. Tr. 1965:12-1967:13 (Zippin). At trial, Jeffrey Zippin hesitantly confirmed that the “calculation assumes ... that the pooled funds that were invested reflects [sic] accurately the total of all of the individual IIM accounts.” Tr. 1967:14-18 (Zippin) (“I’ll have to say that would be my understanding. I don’t have direct knowledge of that.”). The pooled investments did not always reflect the complete sum of the IIM account balances, however, and the Interest Recalculation Project does not address any problems relating to understated interest resulting from this discrepancy. Tr. 1967:22-1969:7 (Zippin); see also 914:17-917:25 (Winter).

The ASM instructs contractors to recalculate interest by applying the interest factor described above to historical balances in IIM accounts. Tr. 2122:8-10 (Dunne). In their HSAs, beneficiaries will be informed of the new, independent calculation of interest, and any limitations on that calculation. Tr. 114:21-115:2 (Cason); Tr. 1968:23-1969:4 (Zippin). This project responds to OHTA’s concerns over whether interest was calculated and credited properly to IIM accounts. Tr. 1964:5-9 (Zippin); 9/13/02 Memo from Chevarria to Edwards, “Issues Regarding Interest Yields,” AR-367 at 60-26-02; 11/1/02 Memo from Chevarria to Zippin, “IIM Trust Funds System Level Issues,” AR-361 at 60-20-02-05. After applying this test to Judgment and Per Capita accounts, Interior concluded that most discrepancies in individual accounts were minor, typically only a dollar over or two dollars under the posted interest amount. Tr. 116:4-9 (Ca-son).

The interest recalculation test proposed in the 2007 HSA Plan has significant limitations. Problems with pooled IIM investments have been extensively documented. See, e.g., AR-361 at 60-20-02. Unfortunately, Interior’s approach “is of minimal value if the actual balance on which the interest is calculated [is] believed to be inaccurate.” Office of Historical Trust Accounting, Bank of America Contract # 1435-01-02-CT-85121 (undated), AR-579 at 31-06-01. Actual investments will not be reconstructed. The interest factor “is not about interest earned; it’s about the total amount that was available to be paid from the interest, and the factor is what [Interior] use[d] to apply to the money in the accounts.” Tr. 1966:10-17 (Zip-pin). At trial, OHTA acknowledged that recalculating the historical interest factor will only reflect how much interest should have been posted to IIM accounts if the pooled funds accurately reflect the IIM funds available for investment. Tr. 1968:15-25 (Zippin).

I.2.K. Land title records

i. Problems with land title records

Concerns over the accuracy of ownership records have plagued the IIM trust for years. Early in the historical accounting project, Interior’s accountants advised that “project teams will need the ownership records for subject land and will need to know that the ownership records are correct.” 3/12/02 Ernst & Young “OHTA Accounting Methods” memo, AR-25 at 1-18-03. NORC has also advised defendants that “land ownership records must be verified” in order to confirm the correct distribution of income derived from land allotments as part of the historical accounting. 12/2/02 NORC memo re: sampling plan, AR-235 at 08-02-07. Unfortunately, testimony at trial echoed findings in the Misplaced Trust Report (as well as other studies) suggesting that Interior’s electronic land ownership records are inconsistently updated, inaccurate, and unreliable. Tr. 2064:6-11 (Christie). There was anecdotal testimony that, at least in some regional offices, LRIS was not timely implemented; that the IRMS ownership module was outdated, inconsistently used, or and not relied upon at all, and that, when agencies sent their ownership data to Land Title Records Offices (LTROs), it was not always uploaded to LRIS in a timely fashion. Tr. 1742:21-1743:15 (Infield); Tr. 1318:21-1322:14 (Redthunder).

Interior’s contractors have discovered, however, that although LRIS is the only electronic source of historical ownership information, some data that cannot be found on LRIS can be found elsewhere in hard copy. NORC Analysis of LRIS Tract History Reports, AR-405 at 50-02-10. Note that LRIS is not used for disbursing IIM funds, but instead as an historical record of allotment ownership. AR-405 at 50-02-05. For that reason, LRIS data is critical to an historical accounting.

ii. Land Title Records Office test

Beginning in 2001, defendants started testing land title records kept at BIA Land Title Records Offices (LTROs). See NORC’s 11/25/01 “Design Report on Sampling and Economic Applications,” AR-399 at 49-03-17. After completing this project, Interior’s “contractor determined that the LRIS data examined adequately reflected land ownership reported by the Bureau of Indian Affairs’ (BIA) LTROs[;] [and] [a]s a result of the project, Interior is confident in the information held within the LTROs for use in the historical accounting.” Interior’s Thirteenth Status Report dated May 1, 2003, AR-549 at 24-18-02.

The land title records study was performed by NORC and OHTA. Four reports in the administrative record reveal the study’s design and major findings. See AR-404; AR-405; AR-406; AR-407. The purpose of the test was to determine whether the land title systems could be relied upon in the historical accounting. AR-407 at 50-04-05. Using statistical sampling, NORC examined randomly selected allotments and traced a small number of tract ownership histories from LRIS, through probate proceedings for some of those tracts. AR-407 at 50-04-03; AR-565 at 33-02-19. Between November 2001 and May 2002, NORC and OHTA visited LTROs in Aberdeen, Alaska, Albuquerque, Anadarko, Billings, Muskogee, Portland, and Sacramento. AR-407 at 50-04^-03-05. The intention was to examine: (1) the completeness and accuracy of the tract records underlying IIM accounts; (2) whether ownership transfers resulting from probate were accurately captured; and (3) how the land and lease records might be used in the historical accounting. Id.

NORC encountered difficulties in executing these tests. This court’s December 5, 2001, temporary restraining order related to IT security shut down the LRIS computer database and significantly affected NORC’s performance of these tests. Also, because each LTRO was different, NORC had to customize its approach to each office’s filing systems. AR-407 at 50-04-06.

In November 2001, Interior conducted its first review of an LTRO in Albuquerque, AR-407 at 50-04^05, before the court’s shutdown of Interior’s computers. Albuquerque was the only LTRO where Interior was fully able to examine LRIS reports. LRIS was unavailable in Alaska by March 2002, but NORC was able to review the LTRO’s offline data. AR-407 at 50-04-07. Aberdeen, the largest LTRO, had a title service area covering the Great Plains and Midwest BIA regions, comprising approximately 7.2 million acres of allotted lands (individual and tribal). AR-404 at 50-01-06. NORC sampled 50 tracts at the Aberdeen LTRO and then drew a sub-sample of 12 tracts within that population. AR-404 at 50-01-07. NORC reported that “the contents of the Aberdeen LTRO systems that we examined appear complete and accurate, within the scope of the checking we did. We therefore are ready to recommend their fuller use in the historical accounting to come.” AR-404 at 50-01-08.

NORC reported similar findings in other LTROs. AR-407 at 50-04-07-13; AR-565 at 33-02-19; AR-404 at 50-01-13 (Anchorage), 50-01-25-26 (Anadarko), 50-01-32-33 (Billings), 50-01-44-45 (Portland). Tests examining ownership changes by operation of probate orders were narrowed to focus on recent beneficiaries, however, using the electronic database that covered from the mid-1980s forward. The results of the ownership testing were based on a random sample taken from LRIS Tract History Reports (“THRs”) in the Aberdeen, Albuquerque, Anadarko, Billings, and Portland LTROs. AR-405 at 50-01-04. NORC found no material errors in the LRIS probate entries it sampled, AR-405 at 50-01-04, but the sample used for this testing addressed only 99 probated estates and was, therefore, not large enough for estimating an “error rate” among the probate entries at conventional assurance levels. AR-405 at 50-01-04-05.

NORC qualified its report on the LTROs in the following manner: “Given the importance of LRIS for the historical accounting, it is clearly desirable to use the knowledge gained so far to design a larger statistical sample which will allow us to estimate the accuracy of the data on LRIS with a higher level of confidence.” AR-405 at 50-02-22-23. For reasons unexplained on this record, no further testing has been conducted, or, apparently, is contemplated.

1.2.L. Mailing historical statements of account

Interior asserts in its 2007 Plan that its duty to account is discharged by the mailing of HSAs. AR-565 at 33-02-09. Interi- or acknowledges that, while HSA packages will be mailed to the most current known address for each beneficiary, it will encounter difficulty identifying current addresses for former account holders. Even many current account holders are “where-about unknowns,” for whom Interior lacks accurate mailing addresses. AR-565 at 33-02-08. Consequently, Interior expects that many HSAs will be returned. Id. at 33-02-09. Interior’s quality control procedures for the HSA mailing process involve sampling HSAs to confirm that the envelope includes the relevant inserts and that the inserts are consistent with the listed addresses. Id. at 57-14-04-05.

1.2.M. Anticipated administrative appeals process

Interior is currently drafting regulations for an administrative appeals process under which a beneficiary who objects to his or her HSA can petition OHTA for review. AR-566 at 33-03-20-21. Generally, the agency would like to establish a process by which beneficiaries must challenge information in the HSA within a certain number of days from the date of the HSA, in order to allow the agency to fix the mistake in the first instance; the failure to raise a challenge within the set time period would likely result in the HSA becoming “conclusively deemed accurate and com píete for all purposes.” See Draft Administrative Appeals Proposed Rule, [Dkt. 3356-2] at 26. The agency plans to issue a notice of proposed rule-making in the Federal Register to seek public comment on the rule. Tr. 88:2-15, 221:24-222:11 (Ca-son). At the present time, the Department is unable to issue a notice of proposed rule-making in the Federal Register because of the class communication orders in this case, which defendants have moved to vacate [Dkt. 3348]. Plaintiffs have strenuously opposed the administrative appeals process they expect defendants to propose based on previous drafts released [Dkt. 3356]. A ruling on that motion will follow the release of this opinion.

1.3. Areas Outside the Historical Accounting Plan

A central purpose of the October 2007 bench trial was to allow a more thorough consideration of whether the historical accounting underway is legally sufficient. That process requires consideration, not only of the activities that are underway in the historical accounting project, but also of the accounting activities that will not be performed under the 2007 Plan. A review of areas that Interior has declared outside the scope of its accounting effort follows.

1.3.A. Account sampling

Interior has no plans to reconcile, each and every transaction within any single land-based account: the sampling design in the 2007 Plan involves reconciling a small number of transactions from within a small number of land-based accounts. Tr. 2164:15-17, Tr. 2167:10-16 (Hinkins); Tr. 1084:13-1085:5, Tr. 1105:7-10 (Scheuren); Tr. 111:8-10 (Cason). It is Interior’s judgment that this approach is the best balance between the twin aims of controlling costs and providing reasonable assurances of the accuracy of transactions within all land-based accounts in the sampled pool. Tr. 1085:3-7 (Scheuren).

I.3.B. Reconciliation of account balances

The focus of the historical accounting work is on the reconciliation of sampled transactions. Defendant’s contractor confirms that “[t]he 2007 plan clearly states that an accuracy and completeness statement will be provided with respect to transactions, not with respect to account balances.... NORC was never asked to design a sample which would provide the individual with a statistical statement about his or her account balance.” Hinkins’ Response to Duncan’s Expert Report, DX-4 at 36; see also Tr. at 2161:11-23, Tr. 2163:3-16, Tr. 2167:10-16 (Hinkins). Interior’s position appears to be that, as a result of sampled transaction reconciliations and tests designed to probe the reliability and completeness of the systems themselves, beneficiaries’ HSAs should provide meaningful assurance that account balances are accurate.

At trial, much was made of the discrepancy between the total balance of all subsidiary IIM accounts and Interior’s general ledger balance, or control account. Tr. 907:19-917:25 (Winter). A Comptroller General Report from 1982 indicated a $25 million aggregate out-of-balance condition between the two totals. Comptroller General’s Sept. 8, 1992 Report to Congress, PX-4174; Tr. 909:5-24 (Winter). Defendant’s witness testified that the agency has reduced this out-of-balance condition to the point where the current aggregate out-of-balance condition is $5.2 million, Tr. 912:13-917:22 (Winter), and that that balance discrepancy results in $5.2 million more in beneficiaries’ accounts than the total of funds in the asset pool for investment. Tr. 914:21-914:23 (Winter). Interi- or has determined the cause of a portion of that remaining out-of-balance condition, Tr. 915:22-916:5 (Winter), but has no plans to reconcile the remaining out-of-balance condition between the subsidiary and general ledger as part of the historical accounting.

I.3.C. Reconciliation of opening balances

The reconciliation of opening balances in Judgment and Per Capita accounts is a relatively simple matter. Transaction-by-transaction reconciliation of these accounts is 86% complete. AR-565 at 33-03-12. Land-based accounts are not so simple. If an account was opened during the sampling frame for the LSA project, its initial credit — reflecting its opening balance— was subject to possible reconciliation as part of the sampled population. Tr. 2154:17-20, Tr. 2178:24-2179:16 (Hinkins); Tr. 1105:11-1107:21 (Lasater). If an account was opened before the beginning of the sampling frame (June 24, 1938), however, its opening balance will not be reconciled, even when reconciliation is accomplished in the future “paper era” sampling test, the design of which has not been revealed to the court. Tr. 1105:11-1107:21 (Lasater). No other testing of opening balances has occurred or is contemplated under the 2007 Plan.

Interior concedes that it is impossible to determine the accuracy of an account balance without confirming the accuracy of the account’s opening balance. Tr. 112:25-113:6 (Cason). Interior nevertheless intends to include within HSAs its conclusions about “the accuracy of the [beneficiary’s] account balance as of December 31, 2000,” apparently based on its conclusions regarding the overall integrity of transactions within the IIM trust system and the possibility of sampling opening balances in the reconciliation samples. For opening balances established with proceeds of probated estates, no accounting will be performed for the funds in those probated estates.

1.3.D. Reconciliation of DPI records and Treasury records

The long-standing discrepancy between IIM balances at the Department of Interi- or and the Department of Treasury has been one of the principal reasons that all independent auditors have qualified their audits of IIM management. When the BIA assumed exclusive control over IIM accounting functions in 1971, there was a net difference in account balances of over $70 million between the ledgers maintained by BIA and Treasury, respectively. Tr. 1557:5-1558:21 (Homan). In 2002, Interior’s IIM balance exceeded Treasury’s IIM balance by $33 million. 12/09/03 Memo from DOI Asst. I.G. for Audits to OST, AR-352 at 60-11-22. In 2004, OST resolved a $4 million out-of-balance condition between IIM balances at Interior and Treasury: Treasury simply reduced “their reported balances, as requested by OST, to amounts that reflect the trust asset balances maintained by OST.” AR-352 at 60-11-22; 906:5-8 (Winter). No effort to reconcile the historical differences that have admittedly been present through the majority of the life of the trust is being made a part of the historical accounting project. Tr. 904:16-905:7 (Winter). Any reconciliation of aggregate balance differences reflects only the consistency between the balances recorded within the two departments, and does not reflect on the accuracy of either total balance. Tr. 906:9-909:25 (Winter).

1.3.E. Asset statements

Defendant no longer plans to provide asset statements as part of the historical statements of account. Under the 2003 Plan, beneficiaries were to be informed of their land ownership interests as of December 31, 2000. That element was removed from the 2007 Plan. Interior’s explanation is that the change of plan allows OHTA to focus on providing beneficiaries information about funds. Tr. 148:23-150:7 (Cason). Interior states that OST will include asset statements with periodic statements of account mailed to beneficiaries once TAAMS is fully implemented, but submits that asset statements are part of Interior’s trust reform efforts, not part of its historical trust accounting.

I.3.F. Special Deposit Accounts

Special Deposit Accounts (SDAs) are used to hold funds until the proper owner is identified and transfer to IIMs can be effected. When income is received from trust property with multiple owners, it is typically deposited in an SDA until the funds are allocated and transferred to IIM accounts. Tr. 560:11-23, Tr. 743:19-745:8 (Herman). Over the years, there has been a build-up of unidentified funds in SDAs. Money sometimes remains in SDAs for extended periods before it is disbursed. Griffin and Associates, Interior’s independent auditors, identified $104 million in 24,000 SDAs within the IIM system in 1998 and $81 million in 18,000 SDAs within the IIM system in 1999. Independent Audit for FY 1998 and FY 1999, AR-375 at 60-34-44. Interest earned on funds held in SDAs often has not been disbursed to the beneficiaries. Interior representatives testified that significant progress has been made in distributing historically accumulated funds held in Special Deposit Accounts, but this work is not reflected in the administrative record. As with asset statements, Interior plans to continue pursuing this project, but no longer considers it to be part of the historical accounting work. AR-566 at 33-03-27.

I.3.G. Predecessor accounts

Interior is not going to reconcile accounts that were closed before October 25, 1994, the effective date of the 1994 Act. Tr. 182:11-15 (Cason). This decision was preceded by considerable discussion between Interior officials, contractors, and accountants. Several Interior contractors or employees have opined that a reconciliation of predecessor accounts is the only way to accurately confirm opening balances of present accounts, and is therefore a necessary part of the accounting. See, e.g., Tr. 1575:9-1576:15 (Homan) (discussing opinion of Deloitte & Touche); March 2002 OHTA accounting conference materials, AR-056 at 04-02-33; 04/03/02 letter from Ross Swimmer to Secretary Norton, AR-511 at 61-37-02. The decision, however, was that probate orders would be considered final and conclusive with respect to the balances of probated IIM accounts. Tr. 141:5-10 (Cason). The Interior Department does not contend that the Interior Board of Indian Appeals (IBIA) probate process amounts to an accounting. Tr. 139:18-144:24 (Cason).

I.3.I. Cadastral surveys

Cadastral surveys, conducted for the federal government by the Bureau of Land Management, are land surveys detailed enough to provide exact property boundaries and locations. Tr. 174:1-4 (Cason); June 2003 Cadastral Resurvey Pilot, AR-394 at 48-01-04. Cadastral surveys of some allotted lands were last performed during the allotment period, that is, between the late 1800s and the 1930s. Cadastral surveys performed before 1910 were crude, resulting in typical error rates of between negative 8.9% to positive 20.7%. June 2003 Cadastral Resurvey Pilot, BLM Acreage and Location Results, AR-395 at 48-02-09-11. Though Interior developed a small scale cadastral resurvey pilot project in 2001 in connection with the historical accounting work, it does not intend to review cadastral surveys as part of its 2007 HSA plan. Tr. 174:24-181:18 (Ca-son).

1.3.J. IIM trust funds managed by compacting/contracting tribes iitilizing their own record systems

Pursuant to the Indian Self-Determination Act of 1975, Pub.L. No. 93-638, 88 Stat. 2203, as amended, 25 U.S.C. § 450, Interior defendants permit compacting and contracting of “the management or administration, at a local level, of certain aspects of IIM trust management.” Cobell X, 283 F.Supp.2d at 180. According to Associate Deputy Secretary James Cason, Interior has an obligation to account for IIM funds managed by tribes pursuant to compact or contract. Tr. 122:20-123:6 (Cason). Ca-son stated that five tribes manage their own IIM funds (by collecting cash, processing cash, and managing distributions) but three of those tribes use Interior’s systems and are therefore included within the current scope of the 2007 HSA Plan. Tr. 119:6-11, Tr. 120:8-16 (Cason). As to the two remaining tribes who process their own IIM funds, Interior has yet to develop an historical accounting project for their IIM monies. Cason did not know whether other tribes since 1975 had collected, processed, and distributed their own IIM funds on their own accounting systems that might not be captured in the 2007 HSA project. Tr. 122:15-19 (Cason).

1.3.K. Direct pay accounts

Some leases of IIM land allotments are structured so that IIM allottees receive proceeds from their land directly from lessees. These arrangements are referred to as “direct pay,” where the lease payments never pass through the government’s IIM trust system but instead are sent straight from the entity renting the property to the IIM property owner or owners. Tr. 2087:14-21 (Christie). IIM accounts within Interior’s system are not established for direct pay arrangements. Tr. 268:8-18 (Cason). The funds transferred in direct pay arrangements retain the character of IIM trust funds as they derive from land in trust, but they are not in any literal sense “taken into trust by the Interior IIM Trust Fund” system. July 2002 Report to Congress, AR-561 at 25-02-25.

Only the Interior Department may enforce the terms of direct pay leases; beneficiaries are not empowered to do so. Tr. 135:2-10 (Cason). Regulations require Interior to collect statements from lessees on direct pay leases, 25 C.F.R. § 13.18, but Associate Deputy Secretary Cason could not confirm that this is in fact being done. Tr. 134:19-23 (Cason). There was anecdotal testimony that at least some agency superintendents demand copies of can-celled direct pay checks to verify payment. Tr. 2037:16-2038:2, Tr. 2039:5-7 (Christie); Tr. 1338:20-23 (Redthunder). Mr. Cason acknowledged that the government retains a “duty to verify” direct payments when the Department is informed of an error, but that he was “not clear about whether this carries over to the historical accounting duty that we have and what its implications would be.” Tr. 127:3-8 (Cason). Ca-son was also unable to definitively state whether direct pay leases are ever initiated by area directors without a beneficiary’s consent, but he said the “approach” of the Department is to initiate direct pay leases only when the beneficiaries affected have agreed. Tr. 135:11-136:16 (Cason). Interior has consistently maintained that it is not required to account for funds on direct pay leases as part of its historical accounting effort. AR-561 at 25-02-15.

I.3.L. Routine collection of third party records

While the ASM contains guidance on collecting records from third parties as necessary, the historical accounting project does not typically collect third party records because Interior has determined that its own documentation is sufficient to perform the historical accounting under the 2007 Plan. Tr. 139:7-13 (Cason). Eight years ago, Judge Lamberth ordered defendants to “establish written policies and procedures for collecting from outside sources missing information necessary to render an accurate accounting of the IIM trust.” Cobell V, 91 F.Supp.2d at 58. The Court of Appeals observed that “[tjhere may not literally be a duty to have such written policies and procedures. Were there a means of ensuring discharge of appellants’ fiduciary obligations absent such steps, there would be no breach. Nonetheless, though the failure to take such steps may not constitute a breach, it surely provides substantial evidence that such a breach has occurred.” Cobell VI, 240 F.3d at 1105-06.

Some third party records maintained in Interior’s files — such as copies of marriage or death certificates — have been part of Interior’s historical accounting analysis. Tr. 490:5-14 (Herman). However, the benefit of collecting third party records from other entities is not only that the documents and document types will be different from those maintained by Interior, but that documents stored by third parties may not have been subject to the same problems with internal controls that have plagued Interior’s record retention systems. For example, the Navajo Nation consults third party records to verify production and revenue data, not because it does not have access to that type of information in MMS systems, but because of concerns over the unreliability of data in MMS systems. Tr. 1775:16-1776:20 (Gambrell). Although Mr. Cason testified that Interior has not “found the need to [pursue third party documents] so far,” Tr. 137:18-21 (Cason), in many instances, the incompleteness of Interior’s data may not be evident unless it is compared to third party records. In the case of Judgment and Per Capita account reconciliation, CD & L did not consult tribal rolls to confirm that “every enrolled participant [of a given tribe] received payment.” 7/16/02 letter from Chevarria to Edwards, AR-314 at 19-07-02. This example illustrates how the failure to collect third party records, in certain instances, may result in an incomplete accounting.

I.3.M. Analysis of other IIM-related data systems

The Minerals Management Service (MMS) is a bureau within the Interior Department that processes oil, gas, and mineral lease payments on behalf of the Federal Government. When oil, gas, and mineral lease payments are delivered into the Department of the Treasury, before the funds are posted to individual accounts, MMS receives the production reports. Tr. 160:21-161:6 (Cason). MMS systems contain lease records and resource production information, but do not contain information regarding how proceeds from such leases are allotted among individual Indians. Tr. 1182:23-1183:9 (Angel); Tr. 161:23-25, Tr. 162:18-19 (Ca-son). To process IIM oil, gas, and mineral revenue, MMS collects the data regarding production and reports to OST a statement of the amounts collected on each lease. Tr. 919:15-19, Tr. 920:11-13 (Winter). OST posts those amounts to “a general holding account.” Tr. 919:19-20 (Winter) (it was not clear whether the witness was referring to a SDA). Then MMS sends data to BIA and BIA consults its Royalty Distribution and Reporting System (RDRS) to determine ownership. Tr. 919:21-24 (Winter). BIA then conveys the ownership information to OST, where the disbursements are made to the account holders from the holding account. Tr. 919:24-920:1 (Winter). OST relies on BIA for the accuracy of the ownership information, Tr. 920:17-21 (Winter), and on MMS for the accuracy of its transactional data.

Approximately $500 million to a $1 billion a month pass through the MMS system. Most of those funds are payments to the Federal Government; only one to two percent of those funds belong to Indian oil and gas lessors. Tr. 132:13-17, Tr. 165:11-16 (Cason). Because of the magnitude of data processed by MMS, Interior has determined that it is neither worthwhile nor feasible to perform an “historical accounting” of MMS systems. Accordingly, the 2007 Plan does not address the issue of whether MMS lease-level information is correct. Interior sees its job as preparing HSAs on an account-by-account basis, as opposed to auditing the entire process, including MMS. Tr. 165:22-166:2 (Cason).

I.3.N. Administrative fees deducted from IIM accounts

The Interior Department is authorized to charge administrative fees for certain transactions related to IIM lands. See generally 25 U.S.C. § 413. Interior represented at trial that “[i]n large part, for most of our program [sic], we don’t charge fees,” Tr. 231:9-10 (Cason), but it is undisputed that administrative fees were charged to beneficiaries (deducted from revenues) from the sale of timber — up to 10% of the total timber sale. Tr. 1343:11-21 (Redthunder); Tr. 230:24-233:14 (Ca-son). In some instances, administrative fees are charged when Interior leases or sells IIM allotted land. Tr. 1702:16-1704:12 (McCarthy). Fees have also been deducted from IIM accounts in connection with the probate process. 8/21/02 email from Edwards re: meeting with OHA, AR-190 at 56-21-02. Interior does not intend, as part of its 2007 Historical Accounting Plan, to account for administrative fees charged to IIM beneficiaries, or to confirm that such fees were in compliance with applicable law. Mr. Cason testified that the possibility of accounting for administrative fees charged to beneficiary’s accounts was not discussed within Interior as part of the historical accounting project planning process. Tr. 233:1-14 (Cason). HSAs will reflect only transactions within a beneficiary’s account, not any gross amount collected before the deduction of an administrative fee. Tr. 231:4-12 (Cason).

1.3.0. Youpee escheated interests

The Indian Land Consolidation Act (ILCA), Pub.L. No. 97-459, tit. II, 96 Stat. 2517 (1983), was intended to address the problem of fractionation. It provided that allotted land interests that fell below a minimum threshold should not be inherited by individual Indians but instead escheated to the tribes. See 25 U.S.C. § 2206. ILCA escheatments were declared unconstitutional takings without just compensation in violation of the Fifth Amendment by the U.S. Supreme Court in Hodel v. Irving, 481 U.S. 704, 107 S.Ct. 2076, 95 L.Ed.2d 668 (1987), and an amendment to the ILCA was subsequently rejected by the Supreme Court in Babbitt v. Youpee, 519 U.S. 234, 117 S.Ct. 727, 136 L.Ed.2d 696 (1997). See also 2/28/03 NORC Analysis of Tract History Reports, AR-405 at 50-02-19-21.

During these challenges to the escheatment process, BIA was instructed to place the money in question in a special account. Tr. 1335:21-1336:8 (Redthunder). After Youpee was issued, BIA was instructed to transfer ownership of the land back to the proper heirs and to get the money back to the proper landowners. Tr. 1336:9-16 (Redthunder); AR^405 at 50-02-21.

Interior estimates that 775,000 fractional land interests belonging to IIM beneficiaries escheated to tribes prior to the Youpee opinion. Tr. 235:17-238:8 (Cason). Interior is presently “acting on” the escheatment issue, albeit somewhat begrudgingly, since the interests are tiny, generally of very low value, and the cost of reversing the escheatments is high. Tr. 236:2-246:8 (Cason). Fractionation has significantly increased the number of You-pee escheated interests that Interior has to try to recapture and return. Tr. 237:13-238:8 (Cason).

Mr. Cason could not testify in any detail as to the Department’s present approach to the escheatment issue, but it is clear that Interior does not intend to account for the escheated interests as part of its historical accounting project. The rationale is that to do so would constitute an accounting for land, and the historical accounting project accounts only for funds. Tr. 239:20-25 (Cason). Interior has not addressed whether or how it will account for any income that has derived from the lands that were covered by the escheated interest. James Cason testified that this issue has not been discussed directly within Interior, and that no decision has been made. Tr. 240:3-14 (Cason).

I.3.Q. Accounts and transactions outside the temporal scope of the 2007 HSA Plan

As noted above, § I.3.G., the Department of the Interior does not consider that the 1994 Act requires it to provide an historical accounting to beneficiaries for accounts that were closed prior to October 25, 1994. Tr. 90:3-91:1, Tr. 92:3-9 (Ca-son). Nor does Interior plan to prepare HSAs for accounts that were opened after Dec. 31, 2000, or for any transactions occurring after that date — the date it began issuing periodic statements of account pursuant to the 1994 Act. If accounts were open as of October 25, 1994, historical transactions within those accounts will be reflected in HSAs back to June 24, 1938, if the accounts contain transactions that old. Transactions earlier than that date are outside the temporal scope of the 2007 Plan, even if the account was open during the period that is within scope. 1957:24-1958:2 (Zippin); AR-565 at 33-02-03. This decision was driven by legal conclusions rather than cost estimates, see 10/28/02 Agenda: Historical Accounting Planning Meeting with Secretary Norton, AR-485 at 61-11-02, but Interior considered the costs and benefits of performing an historical accounting for transactions and accounts outside the temporal scope of the accounting, recognizing that the courts might disagree with its legal interpretation of its duties. Tr. 83:17-84:10, Tr. 85:10-18, Tr. 90:17-91:24 (Cason). Its conclusion was that it is less important to provide accountings for individuals over the entire course of the last 100 years than to provide accountings for individuals with whom it has a continuing trust relationship. Tr. 83:17-25 (Cason).

1.4. Cost Of Performing The Historical Accounting

The high cost of historical accounting is attributable to a host of factors, such as inconsistent practices over numerous agency offices and twelve regions, missing records or data gaps in electronic systems, and the time and effort required to locate necessary documents. The reconciliation of a single transaction costs on average between $3,000-$3,500. AR-566 at 33-03-09. OHTA has approximately 35 employees and 400 contractors. Document searches at AIRR in Lenexa for a single document supporting a transaction can take between a week and a year to complete. Tr. 505:11-14 (Herman).

1.4.A. Cost of historical accounting work

Interior estimates that it has spent $127,000,000 on its historical accounting work since the adoption of the 2008 HSA Plan. AR-565 at 33-02-03; Tr. 1133:6-9 (Haspel); AR-566 at 33-03-09. 9 Appropriations for historical accounting work have been lower than Interior requested or expected. The outsourcing of historical accounting projects has been expensive: in FY 2006, for example, Interior paid over $6 million to the firm charged with coding documents in Lenexa, and over $16 million to one of its accounting firm contractors, CD & L. AR-175 at 56-06-06.

Completion of the 2007 HSA Plan, Interior estimates, will cost another $144,000,000. AR-565 at 33-02-03; AR-566 at 33-02-09. The total estimated cost of historical accounting work from the adoption of the 2003 Plan to the expected completion of the 2007 Plan is thus $271,000,000. AR-565 at 33-02-03. The estimate of future costs is received with several grains of salt, however. Past estimates have proven to be unrealistically low; the defendants do not yet have a complete understanding of the work that will be involved in the “paper ledger era.”

1.4.B. Cost estimates for an expanded historical accounting

Given the importance the Court of Appeals has attached to cost — with its observation that Interior must strike a “balance between exactitude and cost,” Cobell XVII, 428 F.3d at 1076 — I posed as one of the trial questions, “How much would it cost to do what is not being done?” I expected the parties, and particularly the defendants, to devote considerable energy to that subject. What was presented, instead, was the high level, generalized, and essentially unchallenged testimony of Assistant Deputy Secretary Abraham Haspel. The estimated cost to provide HSA packages in the 2007 Plan for all IIM accounts that existed between 1909 and

2006 is between $2,921 and $3,410 billion, almost all of which would be for land-based accounts. DX-098 at 9; Tr. 1124:20-1125:11 (Haspel). If the historical accounting period were limited to the years 1938 to 2006, the cost would run between $2,304 and $2,586 billion. DX-099-COR; Tr. 1125:12-16 (Haspel). These numbers reflect only the expansion in temporal scope, and not the inclusion of methodologies or accounts that have been excluded from the 2007 Plan. Thus, the cost of performing cadastral surveys is estimated at another $1.1 billion. Tr. 1138:14-19 (Haspel). The cost of accounting for direct pay transactions, IIM funds managed by compacting or contracting tribes, or funds never collected is not estimated at all. Tr. 1138:20-1139:8 (Haspel). If historical accounting were extended back as far as 1909, it would cost an estimated $1,365 billion to $1,817 billion to link closing balances in predecessor accounts to successor accounts. DX-098 at 9; Tr. 1120:18-1121:7 (Haspel). (Linking closing balances in predecessor accounts to successor accounts is not an element of the 2007 Plan, even within its limited temporal scope, Tr. 1121:9-11 (Haspel), but it would be necessary in analyzing closed accounts to ensure that money was accurately transferred to successor accounts. Tr. 1121:12-23 (Haspel).) The other major elements responsible for this expected steep increase in costs would be imaging and coding all boxes of ledgers, probates, and financial documentation ($766 million) and digitizing transactions not included in sampling ($637 million). DX-098 at 9; DX-098 at 2; Tr. 1119:16-1120:16 (Haspel).

In 2005, Interior estimated that the cost of complying with the Judge Lamberth’s structural injunction (which did not require accounting for direct pay transactions) would have been approximately $12.9 billion. DX-100; Tr. 1125:19-1126:14 (Haspel). Haspel said that estimate would be higher today. Tr. 1126:12-14 (Haspel).

I.4.C. Congressional appropriations for HSA work

The Office of Historical Trust Accounting was created by Secretarial Order in July 2001. In its first two fiscal years of operation, it was funded from the budgets of other Interior offices. Tr. 78:12-24, Tr. 79:3-19 (Cason). The first congressional appropriation for the Office of Historical Trust Accounting was $15.9 million for FY 2003. For each of the next three years (FY 2004, FY 2005, and FY 2006), Interior submitted budget requests of over $100 million for historical accounting work, but Congress appropriated much less: $44 million (FY 2004), $57 million (FY 2005), and $56 million (FY 2006). Congress “basically sent the message back in their appropriations decisions that they were not willing to spend that much money on the historical accounting process.” Tr. 73:23-74:10 (Cason); DX-102. Based on this experience, Interior submitted a FY 2007 budget request of only $54 million, so that it would have more control over its budget. See Tr. 74:11-22, Tr. 75:15-76:16 (Cason). The estimated 2007 OHTA appropriation at the time of trial was $56 million.

1.5. Throughput

One of the core questions the October 2007 trial was to address was the dollar amount of the IIM trust “throughput,” 10 i.e., the funds that have flowed into and out of the IIM trust. Sorting out throughput involves two distinct but related questions. The first is how many of the total dollars are accounted for in some way by the procedures proposed in the 2007 plan. The second is whether the work done so far permits any reliable estimate of the difference (if any) between dollars in and dollars out. Defendants resisted the idea of exploring throughput, on the ground that throughput is an issue that relates to damages. So it may be, but knowledge of how much money has been received in the IIM trust, and how much has been, will be, or can be accounted for, is also relevant to the question of whether or not Interior’s historical accounting has been, is, or ever can be adequate.

The response of both parties to the throughput question can best be described as desultory. The development of throughput data was limited by Interior’s historical failure to record aggregate IIM figures of receipts and disbursements, and by the refusal of plaintiffs’ expert to utilize Interior records or reveal documentation supporting his work. What follows is an attempt to make sense of the sparse and largely unsupported evidence of throughput that was presented at trial.

I.5.A. Aggregate trust fund vahie estimates

The government’s records of IIM receipts and disbursements have been kept primarily at the individual level rather than the aggregate level. Tr. 1161:10-17 (Angel). In fact, Interior did not begin tracking annual trust receipts and disbursements until 1997. DX-260. Thus, for 111 years, from 1887 to 1997, IIM trust account throughput data is limited to year-end trust balances. Tr. 1166:17-1169:6 (Angel). In 1997, the first year the United States began tracking all government receipts and disbursements, Indian trust receipts and disbursements were not divided into IIM and tribal categories. From 1998 until the present these reports have separated the two trusts in the summation of receipts and disbursements. First annual report of receipts and disbursements, DX-259; Tr. 1167:24-1168:17 (Angel). Official government throughput calculations of the IIM trust therefore date back only to 1998.

Interior’s initial attempt to estimate throughput was reflected in its 2002 Report to Congress. At that point, Interior’s contractors estimated the aggregate receipts of the IIM trust between 1909 and 2001 (or between 1877 and 2001) 11 at approximately $13 billion. On June 30, 1909, the recorded trust fund balance (the earliest one that has been found) was $6.5 million. Interior reported to Congress that IIM Trust receipts totaled somewhere around $3 billion between 1909-1971 (or 1877-1971), and approximately $10 billion between 1972-2001. AR-561 at 25-02-64. The $13 billion estimate did not include direct pay transactions or IIM funds managed by tribes pursuant to compact or contract. Tr. 124:15-25 (Cason).

The estimate set forth in the 2002 report to Congress was a “rough one,” and it has been refined since that time. AR-430. Two significant assumptions informed the initial $13 billion figure. First, to reach the $3 billion revenue estimate for the earlier years, NORC assumed that, as with the years 1972-2001, annual balances and annual receipts were roughly the same. AR-368 at 60-27-02. No one at Interior verified these historic annual balances, however, nor does Interior seriously posit that they are good proxies for collections or receipts. 12 Second, for the 32 years for which NORC lacked year-end balances, missing balance values were interpolated using the end points of each gap. Id.

Michelle Herman did not work on the original $13 billion estimate but has since worked on refining it. Tr. 648:18-650:1 (Herman). At trial, Ms. Herman presented and discussed a detailed chart, color coded by source, reflecting throughput estimates between 1909 and 2005. See AR-171. In the government’s most recent throughput calculations, no estimates are made for the period 1887 to 1909, as defendants could find no data with which to estimate year end IIM balances during the earliest years of the trust. Tr. 841:2-6 (Herman); Tr. 1278:3-14 (Angel). The time periods 1909 to 1922 and 1933 to 1971 include the revenue estimates calculated by NORC, apparently in the dubious fashion outlined above. For the years 1972 to 2005, Herman lists the annual IIM beginning balance, and separates estimated trust revenue into the following five categories: (1) interest revenue; (2) Osage quarterly annuity funds; (3) Judgment and Per Capita (where available); (4) tribal IIM revenue; and (5) all other IIM trust revenue. Id. For the years 1972 to 2005, the chart also contains: (1) disbursements by year; (2) the closing balance listed in the accounting system at the close of those years; (3) the closing balance calculated by adding beginning balance to all receipts and subtracting disbursements; (4) the difference between those two closing balances; and (5) the closing balance reported to Congress in July 2002. Id. The numbers in the chart were provided by Morgan Angel & Associates and NORC, and also include IRMS and TFAS values from FTI’s historical accounting work.

Tribal IIM receipts and Osage quarterly annuity figures reflect two instances where non-IIM funds were commingled with IIM accounts. Tribal trust money has been deposited into the IIM system over the years, even though it is not attached to IIM allotments. Tr. 659:22-660:10 (Herman). Tribes utilized IIM accounts for convenience — as a relatively easy way to obtain checks from the IIM system. Such accounts, oxymoronically, are referred to as “tribal IIM accounts” containing “tribal IIM money.” Tr. 342:14-343:18 (Ramirez). FTI estimates that between 1934-2005, approximately $1,513 billion worth of revenue in the overall IIM system was, in fact, tribal IIM money. Notably missing from the chart is disbursement estimates between 1909-1971: years for which the chart estimates some $3.2 billion in collections. The bottom line, which becomes clear only after the reader performs the critical calculations that are curiously missing from the chart itself, is that between 1909 and 2005, the total estimate of IIM receipts (including Osage and “Tribal IIM”) is approximately $14.3 billion, the total disbursements is around $10.7 billion, and the difference between the two is $3.6 billion. If one were to assume the unproven proposition that the $3.2 billion collected before 1971 was, in fact, disbursed, the unaccounted for collections would total only $400 million — a number that government attorneys have repeatedly said represents the average “float” in the IIM trust (primarily constituting funds in restricted accounts).

These figures are of course estimates, calculated according to unverified hypotheses, using data that are opaque. They provide only a starting point in assessing the size of the IIM trust fund that cannot be accounted for.

I.5.B. Dollars covered by accounting procedures

In a series of charts prepared by Abe Haspel and produced in the midst of trial, the government breaks down estimates of “collections” into the IIM trust (excluding “tribal IIM”) between 1909-2006, and “credits” into individual IIM accounts between those dates. DX-365. The eras are further broken down into 1909-1937, 1938-1984, 1985-2000, and 2001-2006. The total estimated collections between 1909 and 2006 reflected in this chart is $13,186 billion. DX-365 at 1-3. For the period between 1938 and 2000 (the years addressed in the 2007 plan) the estimate is $10,696 billion. Id. This is where it gets complicated.

Haspel prepared nine different charts, each of which assumes different average life spans for IIM accounts. Recall that Interior only plans to sample from and reconcile transactions within accounts that were open on or after October 25, 1994 but not later than December 31, 2000. That means that many transactions will neither be mapped nor within any population sampled for reconciliation. Depending on the average account life span (that is, the average “paper tail” of accounts open during the frame described above), Interior estimates that at least one of the 2007 Plan’s tests will address between 52% and 68% of the total funds collected between 1909-2006 — in other words, somewhere between 52% and 68% of the throughput revenue dollars will be part of a sampled population, or subject to DCV tests, etc. These estimates would be lower if not for the fact that the bulk of the trust’s revenue is of more recent vintage — most of the dollars have entered since 1972, and the 2007 Plan involves much more testing in more recent years. Sampling of older dollars is expected to be extremely sparse, see, e.g., DX-365 at 4 (positing a 15-year average lifespan, transaction reconciliation and interest recalculation would only reach 18% of revenue between 1938 and 1984), and even then, it is the more recent dollars within that group that are the most likely to be tested. This is despite the fact that it is the older dollar transactions that may be more likely to prove either erroneous or impossible to reconcile, and which in real (ie., inflation adjusted) terms, could account for a much more considerable portion of the value that has passed through the trust.

Roughly summarizing all these calculations, Cason testified that, as he understood it, the 2007 Plan will “account for” approximately 50-55% of the nominal dollars received into IIM accounts. Tr. 258:10-20 (Cason). Interior’s estimate of the funds covered under its accounting plan is not reassuring.

I.5.C. Dollars in vs. dollars out

Perhaps the most striking, though entirely unexplained, throughput evidence is Interior’s “current[] estimate[]” that, between 1909 and 2006, 77% of total IIM revenue collected (not including tribal revenue) was posted to IIM beneficiaries’ accounts. DX-365. This, defendants note, is a provisional estimate, subject to revision as historical accounting work continues. Id. Applying the 77% factor to the Interior’s estimate of IIM collections between 1909 and 2006 ($13,186 billion) results in IIM account postings of $10,101 billion. The difference between collections and postings, depending on the temporal scope, is therefore again in the range of $3 billion.

These numbers, presented by Mr. Haspel’s demonstrative exhibit, were neither explained nor effectively cross-examined. Mr. Haspel did not elaborate on the striking disparity that the numbers suggested. He only noted that “if one is talking about collections, that tends to be a larger number than credits, and so if you’re doing a division between the amount of collections or the amount of credits that are proven to be — which are proven coverage, then depending on what you use as the denominator changes the answer.” Tr. 1112:3-7 (Haspel); see also Tr. 154:14-156:14 (Ca-son) (distinguishing between receipts and collections, and describing some collections that IIM beneficiaries are not entitled to, such as the deposits of potential lessees). The estimated 23% of IIM collections that have not entered IIM accounts is only mentioned in a bold footnote at the bottom of each page of the nine-page exhibit, and the court is left to wonder what this disparity represents. Presumably the government did not mean to suggest that between $2.5 and $3 billion of funds intended for IIM accounts were lost or misdirected, as it has resisted such suggestions throughout this litigation. Perhaps Mr. Haspel’s assumption means that the government expects there to be supporting documentation for $13 billion in receipts, but only documentation for $10 billion of the expected $13 billion in credits. All this is speculation — suffice it to say that although on the face of Haspel’s exhibit there appears to be a shortfall of about $3 billion, I do not believe that is what the exhibit intended to communicate.

Plaintiffs’ calculations of throughput and shortfall are no more tangible or helpful. Tellingly, plaintiffs chose not to argue the validity of — or even mention — their expert’s financial model in their proposed findings of fact. That was a wise choice, as I found some of his criticisms of defendants’ model helpful but found his own estimates largely impenetrable. Since plaintiffs have elected not to rely upon that financial model, I will not review it in any detail. It is sufficient to note that it was an attempt to calculate revenue generated by IIM land historically. It contained no disbursement data and assumed no disbursements, and Mr. Fasold attempted to ignore Interior’s records while depending entirely upon the $13 billion figure presented in Interior’s 2002 report to Congress. (After he estimated revenues from natural resources, he calculated “other” collections by subtracting his total from Interior’s $13 billion figure. Tr. 1674:8-1675:3 (Fasold).) He was frank in acknowledging there were no details supporting his model’s estimate of “other” collections. Plaintiffs did not produce worksheets supporting Mr. Fasold’s revenue estimates. Tr. 1659:10-1660:14 (Fa-sold).

We are therefore left with defendants’ throughput estimate, which uses odd assumptions to calculate total throughput of about $13 billion and appears to show an expected shortfall of about $3 billion between receipts and postings. There are also heaps of records suggesting that the numbers informing these calculation are unreliable, and considerable evidence — anecdotal and otherwise — of instances in which IIM beneficiaries did not receive funds to which they were apparently entitled. It remains to be seen whether more accurate or more reliable throughput estimates can be achieved by any of the knowledgeable and diligent professionals working on this case. If the mysterious $3 billion shortfall in Haspel’s charts demonstrates anything, it is that even the broadest questions — and therefore, the easiest questions to answer with statistical sampling — do not seem to have been answered so far by the 2007 Plan, and that there is little reason to be confident that answers are forthcoming.

From the undersigned judge, at least, the Interior Department deserves substantial credit for its effort to strike a “balance between exactitude and cost.” Cobell XVII, 428 F.3d at 1076. The requirement of “exactitude” comes from Interior’s preexisting fiduciary duty to account and from the mandate of the 1994 Act. The “cost” factor comes from common sense — it would indeed be “nuts” to spend several billion dollars to account for a trust fund worth around the same amount — and from the limits Congress has placed on what it will spend for the project. Cobell XIII, 392 F.3d at 466, quoting 149 CONG. REC. S13,784, 13,786 (2003) (statement of Sen. Dorgan). Interior’s 2007 plan reflects the efforts of dedicated public servants to do the impossible. There have been substantial improvements in the administration of the trust: TFAS and TAAMS are clearly superior to the obsolete financial and land ownership systems they replaced. Those improvements are for future beneficiaries, however. Them implementation at the close of the 20th Century has not remedied and cannot remedy the failures of the past, nor do they — together with all the other projects Interior has implemented— amount to the accounting required by the 1994 Act and by Interior’s fiduciary duty as trustee.

II.1. Jurisdiction

Three jurisdictional requirements must be satisfied before a federal court may entertain an action against an officer or agency of the United States. The first is subject matter jurisdiction. That requirement is undisputed, as plaintiffs’ claims arise under the laws of the United States. See 28 U.S.C. § 1331; Cobell VI, 240 F.3d at 1094. The second is the waiver of sovereign immunity. It has long been settled that the Administrative Procedure Act waived the government’s immunity to this suit. 5 U.S.C. § 702; see also Cobell I, 30 F.Supp.2d. at 38-42; Cobell V, 91 F.Supp.2d at 24-28; Cobell VI, 240 F.3d at 1094-94; Cobell XVIII, 455 F.3d at 304. Plaintiffs assert a legal wrong because of agency action or inaction (Interior’s breach of its fiduciary duty to account), and their claim is for relief other than money damages. (The dispute about the nature of the relief plaintiffs seek has not been resolved. The agency contends that plaintiffs would not be entitled to any sort of monetary relief since their action relies on the waiver sovereign immunity in 5 U.S.C. § 702. Plaintiffs insist that the recovery of one’s own unaccounted for trust funds is not money damages but equitable disgorgement, relying on Bowen v. Massachusetts, 487 U.S. 879, 897, 108 S.Ct. 2722, 101 L.Ed.2d 749 (1988). These arguments have been considered in past Cobell decisions, see, e.g., Cobell V, 91 F.Supp.2d at 28 n. 20, and will surely be addressed again. For jurisdictional purposes, it is enough to note that the government has waived sovereign immunity at least insofar as the relief sought by plaintiffs is not money damages.)

The third jurisdictional requirement for suits under the Administrative Procedures Act is that the subject matter be either “final agency action” or a claim that final agency action has been unreasonably delayed. This case certainly has one or another of those elements, and perhaps both.

To be sure, the nineteen published decisions in this case have left its precise jurisdictional underpinnings less than clear. In Cobell I, Judge Lamberth considered claims beyond the APA, noting that plaintiffs

primarily assert a claim that does not implicate the APA. Instead, they sue under the law of trusts for breaches of trust duties by the trustee who, in this case, is the government. Again, the Supreme Court has looked to this substantive body of law to define the duties and remedies afforded to the beneficiaries of the Indian trust relationship. See Mitchell II, 463 U.S. at 226, 103 S.Ct. 2961. The defendants have not contested the premise that the plaintiffs can sue for breaches of trust duties based on legal authority outside the context of the APA, and the APA clearly allows such suits to be brought.

Cobell I, 30 F.Supp.2d at 33. In Cobell V, however, Judge Lamberth dismissed plaintiffs’ pure common law claims, 91 F.Supp.2d at 28-31, and concluded that, “[bjecause the court ultimately holds today that plaintiffs’ proved breaches of trust are actionable under the [APA], the court will not address the merits of plaintiffs’ non- statutory review allegations.” Id. at 24 n. 18.

The Court of Appeals has generally referred to plaintiffs’ claims as arising under the APA, but recent opinions have qualified that description without clarifying the situation: “Although plaintiffs’ core claim is under the APA, Cobell VI, 240 F.3d at 1095, this is not an ordinary APA case,” Cobell XVII, 428 F.3d at 1074 (emphasis added); “we looked primarily to administrative law concepts in resolving jurisdictional issues,” Cobell XVIII, 455 F.3d at 304 (emphasis added). These slight qualifications apparently refer to the incorporation of common law duties into the statutory obligations at issue. See, e.g., Cobell XVIII, 455 F.3d at 307 (“We rely on the common law to flesh out the statutory mandates and determine the precise contours of the government’s responsibilities.”) (internal quotations omitted). In Cobell XVIII, the court explained that the “jurisdiction is limited to addressing specific agency action or inaction,” a clear reference to the APA. Id. Nevertheless, and even though the Court of Appeals stated six years ago that the agency’s “reasonable time to discharge its fiduciary obligations has expired,” Cobell VI, 240 F.3d at 1095, the existence of APA jurisdiction cannot be taken for granted. Cobell VI contains an admonition — one that the government has quoted many times:

Nonetheless, we expect the district court to be mindful of the limits of its jurisdiction. It remains to be seen whether in preparing to do an accounting the Department takes steps so defective that they would necessarily delay rather than accelerate the ultimate provision of an adequate accounting, and the detection of such steps would fit within the court’s jurisdiction to monitor the Department’s remedying of the delay; beyond that, supervision of the Department’s conduct in preparing an accounting may well be beyond the district court’s jurisdiction.

Cobell VI, 240 F.3d at 1110 (emphasis added).

The questions that were explored at the October 2007 bench trial were thus designed, in part, to permit an assessment of whether agency action has been “unlawfully withheld or unreasonably delayed.” They were also designed to facilitate review of Interior’s 2007 historical accounting plan, as final agency action, and to determine whether it is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). The trial record establishes that the 2007 HSA Plan “mark[s] the consummation of the agency’s decisionmaking process.” Bennett v. Spear, 520 U.S. 154, 178, 117 S.Ct. 1154, 137 L.Ed.2d 281 (1997) (internal quotations omitted). It specifies which beneficiaries will receive what Interior calls an accounting and what sort of accounting will be performed, so that the “rights ... [of plaintiffs] have been determined” by its adoption. Id. Defendants resist calling the 2007 HSA plan final agency action, suggesting that only the plan’s completion will be final agency action. Defendants’ Proposed FF/CL [3459-1] at 190. The Court of Appeals did reject the finding in Cobell V that the “preexisting accounting system used to administer the IIM trust constituted a final agency action,” since “an on-going program or policy is not, in itself, a ‘final agency action’ under the APA,” Cobell VI, 240 F.3d at 1095, but the panel did not consider plaintiffs’ argument that the High Level Implementation Plan then under review was final agency action. In any event, the 2007 Plan now reflects a final agency plan to complete — rather than initiate — its historical accounting obligation.

II.2. Taking cost into account, Interior’s 2007 Historical Accounting Plan will not result in an adequate accounting that is compliant with the 1994 Act, prior Cobell opinions, and other precedent.

II.2.A. Accounting methodology, scope, and results

This is “not an ordinary APA case.” Cobell XVII, 428 F.3d at 1074. The deferential standard that typifies judicial review of an agency’s statutory interpretation is inapplicable here. Cobell XVIII, 455 F.3d at 303-07. I am to give Interior’s interpretation of the 1994 Act “careful consideration,” but Chevron deference is “trumped by the requirement that statutes are to be construed liberally in favor of the Indians, with ambiguous provisions interpreted to their benefit.” Cobell XVIII, 455 F.3d at 304 (internal quotations and citations omitted); see also Rosebud Sioux Tribe v. Kneip, 430 U.S. 584, 586, 97 S.Ct. 1361, 51 L.Ed.2d 660 (1977); Montana v. Blackfeet Tribe of Indians, 471 U.S. 759, 766, 105 S.Ct. 2399, 85 L.Ed.2d 753 (1985); Albuquerque Indian Rights v. Lujan, 930 F.2d 49, 59 (D.C.Cir.1991); Muscogee (Creek) Nation v. Hodel, 851 F.2d 1439, 1445 n. 8 (D.C.Cir.1988).

Because this is not only an Indian law case, but a case involving the actions of a fiduciary, Interior’s administrative discretion in performing the historical accounting is “somewhat constrained.” Co-bell XVIII, 455 F.3d at 307. It may not choose “any reasonable option” in designing its plan, because its “actions must satisfy fiduciary standards.” Id. (internal quotations omitted). Accordingly, “[t]he common law of trusts limit[s] the deference [afforded] to Interior’s interpretation of the Act.” Id. at 306. When it comes to striking “a definitive balance between exactitude and cost in performing the accounting,” however, “neither statutory language nor trust principles” provide clear guidance, and the court owes “substantial deference to Interior’s plan.” Id. (quoting Cobell XVII, 428 F.3d at 1076).

In short, “both the APA and the common law of trusts apply in this case; the specific question to be addressed determines which body of law becomes most prominent.” Cobell XVIII, 455 F.3d at 303-04. Obedient to that direction, I proceed to consider which body of law is more prominent with respect to specific aspects of Interior’s 2007 accounting plan: (1) its methodology; (2) its scope; and (3) its result in terms of the information it proposes to include in the IIM account holders’ historical statements of account.

The methodology of Interior’s plan is owed the greatest deference. The use of statistical sampling — especially in light of the discussion of the method in Cobell XVII — reflects a reasonable effort to balance cost and accuracy. See 428 F.3d at 1077-79. By Interior’s own account, however, its understanding of the proper scope of the accounting obligation is not the result of a cost-benefit analysis. It is the result, rather, of a legal interpretation of the 1994 Act and other statutes governing the IIM trust. 13 See Tr. 89:20-91:6, Tr. 92:3-9, Tr. 129:15-130:4 (Cason).

Unlike accounting methodology, the proper interpretation of the statute’s temporal scope is not a matter committed to agency discretion and informed by agency expertise. In interpreting the statute’s scope, I must look toward the common law of trusts and be guided by the canon of construction directing that ambiguous statutes be resolved in favor of Indians. See, e.g., County of Oneida v. Oneida Indian Nation of New York State, 470 U.S. 226, 247-48, 105 S.Ct. 1245, 84 L.Ed.2d 169 (1985).

Nor is scope only a temporal matter. It also relates to the elements that are present within and missing from the statements of account Interior proposes to issue — for example, past transactions (within scope, as Interior sees it) and asset statements (out of scope). These choices were not dictated by administrative cost-benefit analyses to which judicial deference is owed, but by Interior’s idea of what is meant by “historical accounting.” Interior sometimes blurs this line by arguing that the information owed to beneficiaries under the accounting is a function of how much Congress has provided for the project. Cost is certainly a factor in the agency’s assessment of how an accounting should be done. But the core concept of an “accounting” is a matter of law, and not some kind of Procrustean bed that may be stretched or shrunk to fit available resources. It is ultimately the responsibility of the courts to determine whether the thing that Interior proffers as an “accounting” is “sufficient to serve the purposes for which a trust accounting is typically conducted.” Cobell VI, 240 F.3d at 1103; see also, Cobell XIII, 392 F.3d at 472 (“once a statutory obligation is identified, the court may look to common law trust principles to particularize that obligation.”).

I inquire into the scope of the accounting — and the necessary elements of an “adequate” accounting statement — with caution, mindful that “common law precedents don’t map directly onto the context.” Cobell XVII, 428 F.3d at 1078. I “may to a degree use the common law of trusts as a filler of gaps left by the statute, but in doing so [I] may not assume a fictional plaintiff class of trust beneficiaries completely and uniformly free of bars or limitations that the common law may provide.” Id. at 1079. My consideration of common law will be guided by discussions of its applicability in prior Cobell opinions. Ultimately — despite confusion over the concept of a trust accounting, the unique qualities of the IIM trust, and the awkward fit of common law trust principles with the trust under consideration — it is clear for reasons that will be set forth below that the results contemplated by the 2007 Plan fall short of this standard. The HSAs contemplated by the plan will not “contain sufficient information for the beneficiary readily to ascertain whether the trust has been faithfully carried out.” Cobell VI, 240 F.3d at 1103.

i. Methodology

Because the Interior Department is the “actor with primary responsibility for ‘working out compliance with the broad statutory mandate,’ ” Cobell XVII, 428 F.3d at 1076 (quoting Norton v. Southern Utah Wilderness Alliance, 542 U.S. 55, 66-67, 124 S.Ct. 2373, 159 L.Ed.2d 137 (2004)), the district court cannot “bluntly treat[] the character of the accounting as its domain.” Id. Neither the statutes governing the IIM trust nor the common law of trusts delineates the specific “accounting methodology” for Interior to use in performing the historical accounting. Cobell XVIII, 455 F.3d at 306. Discussions of administrative discretion in previous Cobell opinions suggest that, of the three elements of Interior’s plan currently under review, the agency is entitled to the greatest latitude in selecting its accounting methodology. See generally Cobell XVII, 428 F.3d at 1074-79. Indeed, according to the testimony of Interior Department officials, it is the essentially the only one of the three elements (methodology, scope, and results) that arises out of an administrative balancing of cost, time, and accuracy. I must approach Interior’s “judgment regarding the allocation of scarce resources ... under the same [deferential] principles” the Court of Appeals applied to the question of sampling, 14 Cobell XVII, 428 F.3d at 1079.

I am instructed to defer to policies reasonably resulting from cost-benefit analyses, as “neither congressional language nor common law trust principles (once translated to this context) establish a definitive balance between exactitude and cost.” Co-bell XVII, 428 F.3d at 1076; see also Co-bell VI, 240 F.3d at 1104 (noting approvingly that Judge Lamberth “explicitly left open the choice of how the accounting would be conducted, and whether certain accounting methods, such as statistical sampling or something else, would be appropriate. Such decisions are properly left in the hands of administrative agencies”). Choices regarding how to perform the accounting “require[] both subject-matter expertise and judgment about the allocation of scarce resources, classic reasons for deference to administrators.” Co-bell XVII, 428 F.3d at 1076.

Interior’s reliance on statistical sampling remains controversial. It means that only a tiny fraction of land-based transactions under $100,000 will be reconciled, see chart, section I.2.A.Ü., and plaintiffs’ objections to the practice have only grown more strident since the agency reduced the sample size. As Judge Lamberth noted in Cobell X, defendants have not identified a single instance apart from the present case where a trust accounting was performed by statistical sampling, and the use of such a procedure in any conventional accounting would likely be frowned upon. Cobell X, 283 F.Supp.2d at 188. This is not a conventional accounting case, however, and the “absence of precedent [for relying on statistical sampling in performing an accounting] ... [cannot] properly be deemed controlling.” Cobell XVII, 428 F.3d at 1078. The treatment of statistical sampling in Cobell XVII, along with evidence presented at trial, informs my conclusion that the statistical sampling in Interior’s plan is not a per se violation of Interior’s accounting duty. 15

While the number of land-based, “electronic ledger era” transactions sampled and reconciled has dropped from 233,000 under the 2003 plan to 6,599 under the 2007 plan, and the number of “paper ledger era” transactions sampled may be lower yet, the agency now indicates that the average cost of reconciling a single transaction is not $100, as it had thought, but $3,000-$3,500. AR-566 at 33-03-09. In reviewing the sampling ban in the structural injunction, the Court of Appeals seemed particularly struck by evidence that “the average cost of accounting, per transaction, would exceed the average value of the transactions.” Cobell XVII, 428 F.3d at 1078. That observation is likely even more true today.

Unlike when the statistical sampling issue was last before the Court of Appeals, the bulk of the sampling in Interior’s current plan is part of the Litigation Support Accounting (LSA) project. Plaintiffs have been quite critical of this project and have fixated on its title, contending that its purpose was not to satisfy Interior’s fiduciary duty to account, but rather to limit the agency’s liability for settlement negotiation purposes. Interior responds that it chose the name as a way of fitting its activities within the narrow compass of the FY 2004 appropriations bill, which authorized land-based historical accounting only for “litigation support,” Department of the Interior and Related Agencies Appropriations Act, 2004, Pub.L. No. 108-108, 117 Stat. 1241, 1263 (2003), and that estimating the government’s liability exposure was only one purpose of the project. Tr. 62:2-64:5 (Ca-son). The main purpose of the LSA project, according to the contractor who designed it, was to deepen the government’s understanding of “electronic ledger era” transactions. Tr. 974:18-975:1 (Scheuren). Indeed, it was not Interior’s original intent for the LSA project to satisfy its reconciliation responsibilities, but, as the principal statistician began analyzing the data, he realized that further reconciliation of “electronic ledger era” land-based transactions would be unnecessary. Tr. 1004:17-1005:7 (Scheuren); AR-427 at 4-5. Interior essentially came to the conclusion that, without meaning to, it had satisfied its accounting duty with respect to “electronic ledger era” transactions in land-based accounts through its LSA work.

Plaintiffs offered only the opinion testimony of an economist with some statistical training, Dwight Duncan, to criticize NORC’s work on the LSA project. One of Duncan’s primary objections — that NORC inappropriately applied inferences drawn from the sampled population to transactions beyond the population set — was deflated by Dr. Susan Hinkins from NORC, both in her rebuttal report and at trial. She openly acknowledged that the LSA project drew samples from an incomplete population of transactions, and that no statements regarding the reliability of electronic era transactions as a whole could be made on the basis of the sampling conducted to date. She clarified the definition of error — noting that unreconciled transactions were treated as errors — Hinkins Responding Report [Dkt. 3393] at 4, she demonstrated that errors were not netted to reduce total dollars in error, and she illustrated that the LSA project employed both attribute and variable sampling.

It is therefore conceded that the population sampled in the LSA project was incomplete. No assurance has been offered, either that the entire population can be restored, or that additional sampling of the restored population will precede completion of the historical accounting project. Since Interior represents that the HSAs may be mailed before testing of the complete population takes place, see discussion in § I.2.F.Ü, AR-566 at 33-03-18, and since neither “paper ledger era” transaction sampling nor further sampling of the restored “electronic ledger era” population is described in any detail in the HSA Plan, the ultimate statistical validity of the HSA plan has yet to be demonstrated. As described in §§ I.2.I. and I.2.J., the completeness of the transactions recorded in government ledgers cannot be established until several other tests are done, chief among them, the Land-to-Dollars tests, see § I.2.I, which so far have only been performed on the simplest of leases and resulted in no statistical conclusions.

Nevertheless, I am not prepared to state that the agency’s methodological approach to historical accounting is wrong or invalid. It does appear that the various projects reviewed in the findings of fact will generally produce helpful information concerning the history of the IIM trust. It is the Department’s prerogative to decide, as a routine matter, not to collect third party trust records (for example) to aid in the performance of its accounting. That decision may limit the degree to which the agency is able to restore missing transactions from the data set, which would in turn limit the extent to which sampling exercises can reflect accurately the characteristics of the IIM trust. It is not methodology, however, but scope and expected results, that are ultimately fatal to the adequacy of Interior’s accounting project.

ii. Scope and Coverage

a. Scope: years covered

Nothing in the 1994 Act limits the temporal scope of Interior’s accounting obligation. See e.g., Cobell VI, 240 F.3d at 1104. The historical aspect of that accounting duty has been declared by this Court and affirmed by the Court of Appeals. It was established by the statutory language:

(a) Requirement to account

The Secretary shall account for the daily and annual balance of all funds held in trust by the United States for the benefit of an Indian tribe or an individual Indian which are deposited or invested pursuant to the Act of June 24, 1938 (25 U.S.C. § 162a).

25 U.S.C. § 4011(a). The Interior Department reads limitations into those words that are not there. The reference to “the Act of June 24, 1938” does not set a start date for the accounting. All it does is identify the funds for which Congress has mandated an accounting. The parties’ hunt for the meaning of the 1994 Act’s reference to the 1938 Act was diverted by the Court of Appeals’ notation, in Cobell VI, that “all funds means all funds, irrespective of when they were deposited (or at least so long as they were deposited after the Act of June 24,1938)” 240 F.3d at 1102 (emphasis in original). That reference to the date of the Act’s passage turns out to be a red herring. The Act of June 24, 1938, also known as the Indian Reorganization Act and codified at 25 U.S.C. § 162a, sets forth the Interior Department’s authority to deposit Indian trust funds in banks and to invest them subject to certain limitations. 25 U.S.C. §§ 162a(a)-(c); see also Tr. 1278:15-1279:1 (Angel) (defendants’ historian confirms that the central purpose of the 1938 act was to “provide assurance that if funds were to be deposited in commercial banks, that they would be in interest-bearing accounts;” and that the Act “provides for investment in Treasury securities as well”). The 1938 Act expressly contemplated the management of funds already held in the Treasury for Indian trust beneficiaries at the time of the Act’s passage. See Tr. 1276:17-1277:14 (Angel) (testifying that the government has been holding and managing funds in trust for allotted landowners since the late 1800s). It also outlined some of the “[t]rust responsibilities of the Secretary of the Interior[,]” including the duty to account. 25 U.S.C. § 162a(d). The 1938 Act was not the first statute affirming Interior’s duty to account for Indian Funds, see, e.g., 30 Stat. 495, 55th Cong., Sess. II, Ch. 545 at 1399 (1899), but, even if the duty to account had not been mentioned in any statute, “the government’s preexisting duty to provide an accounting to IIM trust beneficiaries ... inheres in the trust relationship itself.” Cobell VI, 240 F.3d at 1103; United States v. Mitchell, 463 U.S. 206, 224-26, 103 S.Ct. 2961, 77 L.Ed.2d 580 (1983) (Mitchell II) (statutes granting United States control over Indian resources and land establish fiduciary relationship).

The most logical reason for the 1994 Act’s reference to the 1938 Act is that it more fully describes the government’s current handling of IIM trust funds. “[B]y its very terms the 1994 Act identified a portion of the government’s specific obligations and created additional means to ensure that the obligations would be carried out.” Cobell VI, 240 F.3d at 1100. The specific obligations emphasized in the statute are the duty to perform an historical accounting, and the duty to reform the trust going forward. A legislative command to trace IIM trust funds only as far back as June 24, 1938 would have been nonsensical, and no legislative history supports such a reading. June 24, 1938 was neither the date when the duty to account originated nor the date of any prior accounting. The correct reading of the 1994 Act’s reference to the 1938 Act is that it instructs Interior to account for “all funds” the Secretary invests pursuant to the authority given him under 25 U.S.C. § 162a: it clarifies the global command to account for “all funds.” 16

Interior has decided to conclude its historical accounting period at December 31, 2000, not because of any statutory interpretation, but because December 31, 2000 is “the beginning of the current accounting period in which IIM account holders receive quarterly statements of account pursuant to the 1994 Act.” Interior’s HSA Plan, Part 1, AR-565 at 33-02-06 n. 10. But fixing the start date for quarterly statements of account should not have also fixed the stop date for historical accounting- — for accounts opened after December 31, 2000. The mailing of periodic statements that were compliant with the 1994 Act was a milestone, to be sure, but it did not terminate the Department’s obligation to assure current and future account holders of the accuracy of their assets and opening balances. Periodic statements of account issued in the 21st Century are to identify: “(1) the source, type, and status of the funds; (2) the beginning balance; (3) the gains and losses; (4) receipts and disbursements; [and] (5) the ending balance.” 25 U.S.C. § 4011(b). Because the Department’s historical accounting work was unfinished in January 2001 and remains unfinished today, however, its 2007 HSA plan gives no assurance to the holders of IIM accounts opened after December 31, 2000 that their opening balances or asset statements are accurately stated.

It would be absurd, of course, to read the 1994 Act as requiring that, for as long as the IIM trust exists, new IIM account holders must be provided with voluminous records tracing their assets and initial credits back to their origins in the late 19th Century. Surely, after the entirety of the IIM trust is properly accounted for, opening balances for new accounts can be presumed accurate and new IIM beneficiaries will not be entitled to historical accountings. If a complete and final accounting can never be done, as I have concluded, how long will an historical accounting requirement exist for new IIM accounts? That will be a question for the next, remedial phase of this case.

b. Scope: funds accounted for

Interior’s accounting project does not cover every account that existed and every transaction that occurred, even between June 24, 1938 and December 31, 2000.

1. “Held in trust ”

It is Interior’s position that certain funds derived from allotted lands are beyond the scope of the accounting mandated by the 1994 Act because those funds were never “held in trust.” 25 U.S.C. § 4011(a).

Allotment owners who receive payments directly from lessees, in so-called “direct pay” arrangements, are not covered by the 2007 HSA Plan; payments that do not pass through the government’s custody will not be analyzed, and beneficiaries will not receive HSA statements for them. AR-565 at 33-03-05. Associate Deputy Secretary Cason testified that while he was aware that the Department has a duty to verify that direct payments are in fact being made, he was “not clear about whether this carries over to the historical accounting duty that we have and what its implications would be.” Tr. 126:16-128:3 (Cason). The Department’s position is clear, however: the Secretary has announced the conclusion that “direct pay funds were never held in trust by the United States and therefore do not fall within the requirement in the 1994 Act that Interior account for all funds held in trust by the United States.” AR-565 at 33-03-05 (internal quotations omitted); see also Tr. 129:15-130:4 (Cason).

The Secretary appears to be correct on this point. The funds that must be accounted for pursuant to the 1994 Act are “all funds held in trust.” 25 U.S.C. § 4011(a) (emphasis added). Direct payments, by definition, are never “held” in trust. Even if the statute were ambiguous — which it is not — the common law of trusts is silent on this question since there is no analogue for direct payments in the world of private trusts. See Trial 1.5 Tr. 72:18-73:7 (Langbein 6/02/03). The government may have duties relating to such funds, but they do not arise out of the statute, and they are not cognizable in this suit.

Similarly, funds that would ordinarily be IIM account funds but are instead managed by tribes (apparently only two tribes handle such funds, Tr. 121:8-18 (Cason)), while undoubtedly trust funds, are never “held” in trust by the United States, and are thus outside the purview of the 1994 Act. As with direct payments, the statute is not ambiguous on this point, and even if it were, there is no common law precedent for trust funds processed by a third party that could conceivably serve as a “filler of gaps left by the statute.” Cobell XVII, 428 F.3d at 1079.

2. Administrative fees & Youpee escheatments

Plaintiffs complain that the HSAs will be incomplete if they fail to account for IIM ownership interests that unlawfully es-cheated to tribes pursuant to the Indian Land Consolidation Act before the Supreme Court declared such escheatments unconstitutional (for a second time) in Babbitt v. Youpee, 519 U.S. 234, 243-45, 117 S.Ct. 727, 136 L.Ed.2d 696 (1997). The 2007 Plan contains no mention of escheated Youpee interests, but the subject was discussed somewhat at trial. In addition, plaintiffs object to the government’s failure to account for administrative fees occasionally deducted from gross IIM revenues before net IIM funds are posted in beneficiaries’ accounts (for services such as timber sales). Tr. 232:19-24 (Cason).

Associate Deputy Secretary Cason’s testimony was that Interior officials had not decided whether or not ultimately to account for administrative fees deducted from beneficiaries’ funds, or for escheated interests that had not yet been returned to IIM accounts, Tr. 232:19-233:14, Tr. 239:1-240:12 (Cason), but it is undisputed that no accounting for such funds will be done pursuant to the 2007 Plan, which, according to the Department, discharges its historical accounting obligation.

Administrative fees and unconstitutional escheatments likely amount to a tiny fraction of the monies that pass through the IIM trust. It is the Department’s prerogative to prioritize its accounting activities, and Interior is, of course, free to decide that it should reconcile high value transactions first and escheatments last, or even to calculate the value of administrative fees and escheatments after sampling. I may not, however, simply ignore two categories of “funds held in trust” under the 1994 Act, both of which operate to diminish the value of IIM accounts or assets, however slight the diminution may be. Neither unrestored Youpee interests nor administrative fees will be captured in the accounting activities as they are now designed, as the fees are not reflected as specific IIM account transactions, Tr. 231:9-14 (Cason), and the escheatments are invisible on the IIM systems if they have not been reversed. The absence of any discussion of such topics, Tr. 233:7-14, Tr. 240:3-8 (Cason), reinforces my conclusion that Interior cannot perform a complete accounting.

3. Special Deposit Accounts

At first blush, it may appear that the same criticisms could be applied to Interi- or’s decision to exclude Special Deposit Accounts from the 2007 HSA Plan. In the 2003 Plan, Interior treated Special Deposit Accounts as a fourth type of account to be reconciled as part of its historical accounting exercise (in addition to land-based accounts, Judgment accounts, and Per Capita accounts). PX-507 at III — 1. By the time it adopted its 2007 HSA Plan, however, the Department had decided to eliminate the reconciliation of SDAs from its historical accounting project, because the funds in those accounts are, by definition, not associated with particular IIM accounts. IIM funds held in SDAs are indeed “held in trust” under the 1994 Act. My conclusion, however, is that Interior’s treatment of SDA monies is the result of a considered cost-benefit analysis, is entitled to deference, and is not improper.

SDAs are used to hold funds temporarily, until the proper recipients are identified. SDA accounts include funds belonging to IIM beneficiaries, tribes, lease bidders, and others. Tr. 2065:2-24 (Christie). Thus, while current balances in SDAs are not reconciled as part of the historical accounting project, to the extent that SDAs consist of IIM funds, those funds will be within the populations tested in the LSA and DCV projects after they have been transferred to IIM accounts.

At times the amount of money held in SDA accounts has been excessively high. In July 2002, there were approximately 21,000 open SDAs, containing a total of almost $68 million funds waiting to be identified and distributed. Tr. 157:8-12 (Cason). As a result of Interior’s efforts to assess, reconcile, and distribute the SDA funds, the total balance of funds in SDAs as of October 2007 had been reduced to about $16-17 million, and Interior had distributed 75% of the funds that had been held in SDAs in July 2002. AR-566 at 33-03-27; Tr. 158:12-19 (Cason). Interior continues to work on the remaining funds in Special Deposit Accounts, and has chosen to prioritize its SDA work by reconciling high value accounts before reconciling lower value accounts. Tr. 158:20-159:2 (Cason). So far, Interior has spent approximately $48 million distributing residual balances from Special Deposit Accounts. AR-566 at 33-03-27.

Interior intends to complete its distribution of the SDA funds; the Office of Historical Trust Accounting, Bureau of Indian Affairs, and Office of the Special Trustee are all working on the project. Tr. 159:3-6 (Cason); Tr. 752:4-7, 22-25 (Herman). At Trial, Associate Deputy Secretary Cason could not state for certain when this project would be complete. Tr. 159:3-160:3 (Cason). As Interior gets into the smaller value SDAs, it reaches a point where the cost of reconciling the accounts exceeds the funds in the accounts themselves. Tr. 159:11-17 (Cason). When that occurs, the Department plans to make “pacing decisions” in order to “move the historical accounting process along more rapidly.” Tr. 159:18-22 (Cason).

Interior’s approach to SDA accounting is entitled to deference because the record reflects extensive consideration of the problem by three bureaus or offices within the Department and a determination to deal with it after projects with higher priorities are complete. This stands in sharp contrast to the Youpee escheatment question, for example, about which Associate Deputy Secretary Cason could give only vague testimony, and which may be meeting with internal bureaucratic resistance. Tr. 2016:24-2022:1 (Willett).

4. Closed accounts and probated estates

Interior’s duty to account for “all funds” in the IIM trust is enforceable in this suit, not by or on behalf of every beneficiary who ever walked the earth, but by a plaintiff class, certified by Judge Lamberth on February 4, 1997, and consisting of “present and former beneficiaries” of the IIM trust [Dkt. 27]. That expansive definition of the class, appropriate at that early stage of the litigation, now needs to be refined.

On October 25,1994, a statutory right to an accounting accrued for all then-living IIM beneficiaries: those who held or at any point in their lives had held IIM accounts. Even if the government sold allotted land owned by Beneficiary X in 1992 and closed X’s account by mailing her a check for X’s share of the sale price, X is entitled to an accounting under the statute. If beneficiary Y, the owner of a 10% interest in an IIM allotment, died on October 24, 1994, the night before the 1994 Act became law, his right to an accounting never accrued — the government is not required to prepare an accounting and mail it to his grave — but Y’s heirs do have a right to an accounting of any IIM funds and assets they inherited, and since the probate process does not produce an accounting, Tr. 139:21-144:24 (Cason), Interi- or’s duty to those heirs includes accounting for the allotment assets and IIM funds that were their inheritance from Y. The rationale for including predecessor accounts in the historical accounting process is simply that beneficiaries are entitled to know where their money came from.

The only limitation on the entitlement of the members of the plaintiff class to demand an accounting for “all funds” are those imposed by the law of standing, by reasonable methodological shortcuts taken to conserve funds (e.g., sampling), and by whatever restrictions may be present in the common law. See Cobell XVII, 428 F.3d at 1079. Limiting the historical accounting process by excluding accounts that were closed prior to October 25, 1994, and by refusing to examine IIM accounts behind probated estates, narrows the scope of the accounting. This adjustment is substantive, not methodological. It is not sanctioned by the common law of trusts or the plain terms of the 1994 Act. And the members of the plaintiff class— IIM beneficiaries alive on October 25, 1994 — have standing to challenge it.

5. Asset statements

The most serious deficiency of the historical accounting project reflected in the 2007 Plan is that it will not provide beneficiaries with information about the assets from which IIM funds have been derived. A review of statutory and common law trust accounting principles reveals why HSAs so limited cannot meet the adequacy and fairness standard. 17

“The most relevant statute for ascertaining the defendants’ duty to provide an historical accounting is the 1994 Act[J” Cobell XVII, 428 F.3d at 1074. The Act itself contains few clues indicating what the historical accounting must entail and the what accounting statements must contain, but it is clear from the Special Trustee’s oversight duties that account holders are entitled to a “fair and accurate accounting of all trust accounts.” 25 U.S.C. § 4043(b)(2)(A). The Court of Appeals has spoken on this subject, noting that “[defendants] never explain how one can give a fair and accurate accounting of all accounts without first reconciling the accounts, taking into account past deposits, withdrawals, and accruals. Indeed, the government’s own expert acknowledged that one could not determine an accurate account balance without confirming historical account balances.” Cobell VI, 240 F.3d at 1103 (emphasis added).

The 1994 Act amended 25 U.S.C. § 162a and

identified some of the Interior Secretary’s duties to ensure ‘proper discharge of the trust responsibilities of the United States.’ 25 U.S.C. § 162a(d). These ‘include (but are not limited to) the following’:

‘Providing adequate systems for accounting for and reporting trust fund balances’;

‘Providing adequate controls over receipts and disbursements’;

‘Providing periodic, timely reconciliations to assure the accuracy of accounts’;

‘Preparing and supplying ... periodic statements of ... account performance’ and balances to account holders; and

‘Establishing consistent, written policies and procedures for trust fund management and accounting.’

Cobell VI, 240 F.3d at 1090 (quoting 25 U.S.C. § 162a(d)). The “periodic statements of ... account performance” are described in greater detail in 25 U.S.C. § 4011(b). They are to be provided to tribes and individual beneficiaries “[n]ot later than 20 business days after the close of a calendar quarter,” and are to “identify (1) the source, type, and status of the funds; (2) the beginning balance; (3) the gains and losses; (4) receipts and disbursements; (5) the ending balance.” Id.

These periodic statements of performance — though they are part of the 1994 Act’s trust reform mandate — have an important bearing on the elements of an adequate historical accounting statement. “[P]eriodic statements of ... account performance” are simply accountings going forward, rather than historical accountings. 25 U.S.C. § 162a(d). For that reason, their required contents are probative of what a “fair and accurate” historical accounting of “all funds” is.

Notably, the first element that must be “identified]” in the “periodic statement of performance” is “the source ... of the funds,” 25 U.S.C. § 4011(b)(1), meaning, obviously, information sufficient to allow a recipient to connect her receipt of funds to specific assets. This requirement to include historical asset ownership in historical statements of account was identified and declared by the Court of Appeals years ago, in Cobell VI:

The 1994 Act requires that the Interior Department perform an ‘adequate’ accounting. This indicates that the accounting must be sufficient to serve the purposes for which a trust accounting is typically conducted. By this standard, the district court’s conclusion that the management of a trust and rendering of an adequate accounting requires the locating and retention of records, operational computer systems, and adequate staffing was, in plaintiffs’ words, ‘self-evident.’ Anything less would produce an inadequate accounting.

It is black-letter trust law that “[a]n accounting necessarily requires a full disclosure and description of each item of property constituting the corpus of the trust at its inception.” Engelsmann v. Holekamp, 402 S.W.2d 382, 391 (Mo. 1966); see also BLACK’S LAW DIC TIONARY (7th ed.1999) (defining accounting as “the report of all items of property, income, and expenses” prepared by the trustee for the beneficiary).

240 F.3d at 1103 (emphasis added). The requirement to include asset information in historical accounting statements is consistent with both the statute and the common law. 18

iii. Summary of legal deficiencies

It is Interior’s responsibility to “work[ ] out compliance with the broad statutory mandate,” Cobell XVII, 428 F.3d at 1076 (internal quotations omitted), but in reviewing that compliance, I must determine whether the accounting will be “sufficient to serve the purposes for which a trust accounting is typically conducted,” and whether the HSAs will contain “sufficient information for the beneficiary readily to ascertain whether the trust has been faithfully carried out.” Cobell VI, 240 F.3d at 1103. My answers to both questions are in the negative: the accounting will not serve the purposes of a trust accounting, and the HSAs will not enable beneficiaries to determine whether their trustee has discharged its fiduciary duties.

The absence of a verified opening balance and an asset statement is the most serious defect in the 2007 Plan. A simple illustration explains why this is so. If I demand an accounting for funds in my non-interest earning checking account, a list of transactions — credits and debits with supporting documentation — would probably be a sufficient response. Presumably, I would have firsthand knowledge that the account was indeed opened with, for example, a $500 deposit. By contrast, IIM plaintiffs who receive HSAs containing a list of transactions, however accurate, will have no ability to contextualize those transactions. Beneficiaries are provided no records indicating their historical ownership interests. In most cases, the source of a beneficiary’s funds is a plot of land that was carved out of a tribal reservation, allotted to her ancestors, and placed in trust. The beneficial interest in that land, and perhaps the size of the allotment, likely diminished over time (due to fractionation and land sales) as it worked its way down to the current beneficiary. Without information tying revenue collected to her land ownership, it will be utterly impossible for her to determine whether the duties of the trust have been faithfully carried out and whether her assets, funds, and transactions are correctly stated, and she will be prejudiced from “enforc[ing][her] rights under the trust or [ ] preventing] or redressing] a breach of trust.” RESTATEMENT (THIRD) OF TRUSTS § 173. That, as I understand it, is the purpose of an accounting.

Relatedly, the proposed accounting falls short of its promise to provide beneficiaries reasonable assurance that their account balances are accurate, as opening balances are not confirmed and the funds are not traced back through predecessor accounts. See Tr. 2167:10-16 (Hinkins). The confirmation of accurate account balances is surely one of “the basic fiduciary responsibilities which we would expect out of any trustee,” and one that the 1994 Act was passed to enforce. 140 CONG. REC. H10483, 10488 (1994) (statement of Rep. Synar). Even in its letter expressing concern over the estimated cost of the accounting in Interior’s July 2002 Report to Congress — at that time, $2.4 billion — the House Committee on Resources qualified its request that the agency consider alternative accounting methods with the proviso, “[cjlearly, any such accounting should be sufficient to ensure beneficiaries of the trust that they can rely on their account balances.” 12/9/02 Letter from House Committee on Resources Chairman James Y. Hansen to Interior Secretary Gale Norton, AR-184.

The basic purpose of a trust accounting — to provide sufficient information for beneficiaries to discover whether the trust has been properly administered — is not thrown into doubt by the absence of a uniform definition for the term “accounting,” nor by the unique qualities of the IIM trust. Trust law is typically state law, and there is very little federal common law that addresses the concept of “an accounting.” Defendants’ trust law expert Professor John Langbein — while noting that he was “not a trust accounting person,” Trial 1.5 Tr. 110:18 (Langbein 6/02/03 PM), but rather an expert on the “relative importance of statute and common law,” id. at 121:22-23 — stressed that there are at least five separate legal definitions of “an accounting.” Trial 1.5 Tr. 87:20-22 (Langbein 6/02/03 PM). The basic dictionary definition of an accounting, however, is simple enough: it is “the report of all items of property, income, and expenses” prepared by the trustee for a beneficiary. BLACK’S LAW DICTIONARY (7th ed.1999) (cited in Cobell VI, 240 F.3d at 1103). There may be up to five different definitions of an accounting to suit different contexts, but of the five described in Professor Langbein’s testimony, only two describe the sort of accounting plainly contemplated by the 1994 Act. The first is described in the RESTATEMENT (THIRD) OF TRUSTS § 172, and involves an “internal audit.” Trial 1.5 Tr. 88:1-4 (Langbein 6/02/03 PM). The second is described in the RESTATEMENT (THIRD) OF TRUSTS § 173, and is the sort of accounting to which the Court of Appeals has determined plaintiffs are entitled under the 1994 Act. That section is described by Professor Langbein as “reporting and disclosure to the beneficiaries,” Trial 1.5 Tr. 88:6 (Langbein 6/02/03 PM), and provides that

[t]he trustee is under a duty to the beneficiary to give him upon his request at reasonable times complete and accurate information as to the nature and amount of the trust property, and to permit him or a person duly authorized by him to inspect the subject matter of the trust and the accounts and vouchers and other documents relating to the trust ... the beneficiary is always entitled to such information as is reasonably necessary to enable him to enforce his rights under the trust or to prevent or redress a breach of trust.

RESTATEMENT (THIRD) OF TRUSTS § 173. Professor Langbein stresses that the “reasonableness” concept referenced in this section is tied to the situational “prudence norm,” which would take account of the government’s budgetary restrictions. Langbein Phase 1.5 Expert Report [Dkt. 3395-2] at 6-7 (citing RESTATEMENT (SECOND) OF TRUSTS § 174 (1959)). Moreover, the accounting duty described in § 173, like all other common law trust duties, may be abrogated by the trust instrument (in this case, the statutes setting forth the governance of the IIM trust).

To be sure, as Professor Langbein notes, “[t]he common law of trusts is default law, which yields to the contrary terms of the particular trust.” Id. at 3. He submits that “legislation bearing on the trust” trumps any invocation of common law, id. at 6, and asserts that, if an Act of Congress may reasonably be interpreted as overriding common law, it does, even if Congress has not expressly abrogated traditional trust law duties. Trial 1.5 Tr. 108:14-110:5 (Langbein 6/02/03 PM). It was his testimony that “various trust law duties yield to contrary congressional expression, which is contained in the statutes, and which includes in my view funding as well.” Trial 1.5 Tr. 110:3-5 (Langbein 6/02/03 PM). “[C]ontrary congressional expression,” however, cannot be inferred from the subtext of appropriations bills without some evidence that Congress understood itself to be amending the 1994 Act. Both the canon against implied repeal and the canon construing legislation in favor of Indians recommend strongly against it. Indeed, Congress’s own behavior reveals that it never believed itself to have accomplished so much: it once derailed this litigation by suspending certain accounting activities and giving itself one year to clarify the precise accounting duties owed. That action reflects congressional awareness that, in order to alter the legal entitlements of the beneficiaries arising out of the 1994 Act, it must do so explicitly.

In any event, the trial judge’s ear is more sensitive to what the Court of Appeals has said on such matters — in this case to the dicta of Cobell XVII that, while the funds Congress has appropriated for historical accounting work “can’t be completely disregarded,” the “significance of appropriations bills is of course limited.” 428 F.3d at 1075. The Court, indeed, has been strongly cautioned against identifying wholesale changes to fiduciary duties in ambiguous congressional directives:

While the government’s obligations are rooted in and outlined by the relevant statutes and treaties, they are largely defined in traditional equitable terms. ‘Where Congress uses terms that have accumulated settled meaning under either equity or the common law, a court must infer, unless the statute otherwise dictates, that Congress means to incorporate the established meaning of these terms.’ NLRB v. Amax Coal Co., 453 U.S. 322, 329, 101 S.Ct. 2789, 69 L.Ed.2d 672 (1981). Courts ‘must infer that Congress intended to impose on trustees traditional fiduciary duties unless Congress has unequivocally expressed an intent to the contrary.’ Id. at 330, 101 S.Ct. 2789. Much as the Supreme Court has regularly turned to the Restatement and other authorities to construe trust responsibilities, it is appropriate for the district court to consult similar sources.

Cobell VI, 240 F.3d at 1099. This case must be governed by Congress’s plain demand for an accounting of “all funds held in trust by the United States for the benefit of an ... individual Indian.” In its refusal to appropriate enough money to pay for such an accounting, Congress has not amended that demand or the common law of accounting. What it has done, instead, is to render a real accounting impossible — or, perhaps, to recognize that such an accounting is impossible, unless it is “nuts” enough to pay more than $3 billion to hunt down perhaps $3 billion of unexplained variances in the government’s accounts. 19

CONCLUSION

This case has been in this courthouse for over eleven years. A “long procession of [judges] has come in and gone out” during that time. The “suit has, in course of time, become so complicated” that “no two lawyers can talk about it for five minutes without coming to a total disagreement as to all the premises.” It has been on my docket for one year, during which time I have dismissed persons who were still “parties in [the suit] without knowing how or why,” resolved dozens of motions, enforced an attorneys’ fee award that predated the invasion of Iraq, and studied the case enough to be among the few people “alive [who] know[ ] what it means.” “Innumerable children have been born into the cause,” and, as plaintiffs have reminded us on occasion, “innumerable old [plaintiffs] have died out of it.” I held the October 2007 bench trial in order to review defendants’ historical accounting work and to demonstrate that a just resolution of this dispute, despite what has been said, is not “perennially hopeless.” 20

My conclusion that Interior is unable to perform an adequate accounting of the IIM trust 21 does not mean that a just resolution of this dispute is hopeless. It does mean that a remedy must be found for the Department’s unrepaired, and irreparable, breach of its fiduciary duty over the last century. And it does mean that the time has come to bring this suit to a close.

The Court in this case “retains substantial latitude, much more so than in the typical agency case, to fashion an equitable remedy,” Cobell XII, 391 F.3d at 257, but the 1994 Act does not “have language in any way appearing to grant courts the same discretion that an equity court would enjoy in dealing with a negligent trustee.” Cobell XVII, 428 F.3d at 1075. Nor do I have the authority to “order broad, programmatic reform” or to provide the agency with a “detailed plan of action.” Cobell XVIII, 455 F.3d at 307. What has been determined to this point is that the Department of the Interior has not — and cannot — remedy the breach of its fiduciary duty to account for the IIM trust. The Clerk is directed to schedule a hearing for about 30 days after the issuance of this opinion for the purpose of discussing a process for determining an appropriate remedy.

APPENDIX

A List of Acronyms

AIRR: American Indian Records Repository

ASM: Accounting Standards Manual

ART: Account Reconciliation Tool

BIA: Bureau of Indian Affairs (within DOI)

BISS: Box Index Search System

BLM: Bureau of Land Management (within DOI)

CEBA: Competitive Equality Banking Act of 1987

CD & L: Ghevarria, Dunne, & Lamey

DCV: Data Completeness Validation

HLIP: High Level Implementation Plan

HSA: Historical Statement of Account

IBIA: Interior Board of Indian Appeals

IIM: Individual Indian Money

ILCA: Indian Land Consolidation Act of 1983

IRMS: Integrated Resource Management System

LRIS: Land Records Information System

LSA: Litigation Support Accounting

LTRO: Land Titles Records Office (within BIA)

MMS: Minerals Management Service (within DOI)

NARA: National Archives and Records Administration

NORC: National Opinion Research Center

OHTA: Office of Historical Trust Accounting (within OST)

OST: Office of the Special Trustee (within DOI)

OTFM: Office of Trust Funds Management (within OST)

OTR: Office of Trust Records (within OST)

RDRS: Royalty Distribution Reporting System SDA: Special Deposit Account

TAAMS: Trust Asset and Accounting Management System

TFAS: Trust Fund Accounting System

TMIP: Trust Management Improvement Project

TFR: Trust Funds Receivable

USGS: United States Geological Survey (within DOI)

1 . This Act provided in part that

[d]uring the trust period of twenty-five years, such part of the lands which have been allotted to members of the Yankton tribe of Indians in severalty, as the owner thereof can not cultivate or otherwise use advantageously, may be leased for one or more years at a time. But such leasing shall be subject to the approval of the Yank-ton Indian agent by and with the consent of the Commissioner of Indian Affairs; and provided that such leasing shall not in any case interfere with the cultivation of the allotted lands by the owner thereof to the full extent of the ability of such owner to improve and cultivate his holdings. The intent of this provision is to compel every owner of allotted lands to cultivate the same to the full extent of his ability to do so, before he shall have the privilege of leasing any part thereof, and then he shall have the right to lease only such surplus of his holdings as he is wholly unable to cultivate advantageously.

28 Slat. 286, 316.

2 . Seventy years earlier, in 1920, 35.897 million acres of allotted land were owned by individual Indians. FRANCIS PAUL PRU-CHA, 2 THE GREAT FATHER: THE UNITED STATES GOVERNMENT AND AMERICAN INDIANS 865 (1984).

3 . See Appendix A.

4 . Transcript citations to the October 2007 bench trial will be referenced only by page and line number and witness name. Transcript citations to other proceedings will also include the abbreviated title of the proceeding and the date of the testimony.

5 . Underneath the BIA’s twelve regional offices are 93 "agencies” or "agency offices” responsible for, among other things, processing IIM transactions. Feb.2002 OHTA Accounting Records Conference Materials at 213, AR-058.

6 . Defendants’ contractors were still looking for 35,000 out of 80,000 requested documents at the date of trial, Tr. 466:4-11 (Herman), Assistant Deputy Secretary Haspel acknowledged that, "until we open all those boxes," the extent to which historical trust records may be missing remains unknown, Tr. 1138:4-11 (Haspel), and defendants’ historian Edward Angel testified that he could not say whether there are sufficient records to account for any particular year of the IIM trust. Tr. 1261:5-10 (Angel).

7 . All IIM funds in Special Deposit Accounts— temporary holding accounts used to deposit funds before they are credited to specific accounts — were to be transferred into the appropriate IIM accounts as part of the historical accounting process.

8 . Ramirez testified that the zero dollar instruments were not checks issued to beneficiaries in the amount of zero dollars, but check stock that had been voided, for any number of reasons. Tr. 349:3-351:15 (Ramirez). For example, the check stock may have contained typos, or may have been destroyed in the printer, or could not be used because it did not contain the limited payability instruction (“void after one year”). In those cases, Treasury would not be informed that a particular check number was used, so Treasury officials would input a zero dollar amount. Tr. 349:6-350:24 (Ramirez).

9 . The record does not reveal the cost to Interior between 1994 and 2003 of responding to the requirements of the 1994 Act.

10 . I suggested at trial that one of plaintiffs' witnesses may not have used the term in his report because he prefers to speak English, Tr. 1676:10-13 (Fasold), but no better term has been offered to represent this concept, and “throughput” is now firmly established in the Cobell lexicon.

11 . In AR-368, NORC's report to OHTA on the 2002 report's throughput estimate, the $13 billion estimate covers the years 1909-2001; $3 billion between 1909-1971. However, in notes following the chart on AR-561 at 25-02-64 — the actual report to Congress in 2002 — the $3 billion revenue estimate in the early years of the trust is associated with the time period between 1877-1971. This appears to be a typographical error, however, as defense witnesses testified at trial that the 2002 estimate dated back only to 1909. See Tr. 217:10-19 (Cason); Tr. 657:8-12, Tr. 840:23-841:6 (Herman).

12 . Indeed, this is a bizarre assumption. It amounts to a suggestion that, in any given year, the IIM trust is likely to take in the same number of dollars that remain un-disbursed at the end of the year. But these two variables should be based on different factors and have no obvious relationship: the amount that comes into the trust is based on receipts that the trust made (or didn’t) in that year; the annual balance is based primarily on the difference between receipts and disbursements (as well as opening balance).

13 . Interior has calculated cost estimates relevant to various interpretations of the 1994 Act in designing its accounting plan, see Tr. 81:12-87:5 (Cason), in part because of its awareness that "there are two other major parties within government who have a say in [the historical accounting] process, and that’s both the Court and the Congress.” Tr. 91:2-6 (Cason). The testimony of Interior officials at trial as well as defendants’ brief regarding the nature and scope of the historical accounting reveal, however, that decisions regarding the scope of the accounting are the result of the Department's legal interpretation of the 1994 Act [Dkt. 3339].

14 . During a pre-trial conference, plaintiffs insisted that, despite the clear statement of the Court of Appeals that cost is a valid methodological consideration, the common law of trust dictates that the full cost of a proper accounting must be borne by the trustee if the high cost of the accounting is entirely due to his malfeasance. H’rg Tr. 27:5-28:24 (6/18/07). Plaintiffs insisted that the Court of Appeals did not have a record on which to consider this argument when it stressed the importance of cost. Id. at 27:5-9. This argument must be rejected, as defendants’ responsibility for the difficulties it faces in performing an accounting was extensively documented by Judge Lamberth in several published opinions, and the Court of Appeals has already considered (and rejected) the suggestion that the common law of trusts could be invoked here in that manner:

Where a trustee has by misconduct or negligence made a proper accounting more difficult, the trustee may be charged for the accounting's cost, and no precept of common law constrains the cost of such an accounting, see GEORGE GLEASON BOGERT & GEORGE TAYLOR BOGERT, THE LAW OF TRUSTS AND TRUSTEES § 963, at 459 n. 36 (rev.2d ed.1983) (citing Haas v. Wishmier's Estate, 99 Ind.App. 31, 190 N.E. 548 (Ind.App. 1934)), though obviously bargaining between trustee and beneficiaries might eliminate some excesses. Absent such misconduct or negligence, however, the costs of an accounting would fall on the trust estate itself, which, as we said before, would automatically give private beneficiaries an incentive not to urge extravagance. Cobell XIII, 392 F.3d at 473. While Congress in the 1994 Act plainly faulted the United States’ management, see, e.g., H.R. REP. NO. 103-778, at 9-11 (1994), U.S.Code Cong. & Admin.News 1994, pp. 3464, 3468-70, the Act’s general language doesn't support the inherently implausible inference that it intended to order the best imaginable accounting without regard to cost.

Cobell XVII, 428 F.3d at 1075. In short, despite plaintiffs' presentation of countless historical reports illustrating the Interior Department's poor handling of the IIM trust, I find that the Court of Appeals was no less aware of these problems when it instructed that cost was a relevant consideration in designing and reviewing the historical accounting plan.

15 . The reason given by the Court of Appeals when it vacated Judge Lamberth's ban on statistical sampling was merely that it “reflected no deference to defendants' expertise or to their judgment regarding the allocation of scarce resources,” Cobell XVII, 428 F.3d at 1078, but the court showed no reluctance to apply its own expertise and judgment to the subject, concluding, after considering reconciliation cost estimates, that the "decision to use statistical sampling seems especially reasonable.” Id.

16 . As I mentioned in court on June 18, 2007, I am aware that Judge Lamberth has given the scope of the accounting more attention than anyone. 6/18/07 H’rg Tr. 6:2-16. Though his rulings in Cobell X are not the law of the case, I treat his reasoning as presumptively correct and depart from it only with caution. See Action Alliance of Senior Citizens v. Sullivan, 930 F.2d 77, 83 (D.C.Cir. 1991) (vacated holdings, in the absence of contrary authority, remain persuasive precedent). In this instance, I agree with his ruling that the accounting obligation precedes June 24, 1938, but my reasoning is somewhat different. He found that while the 1994 Act does not itself require an accounting for transactions before 1938, Interior must nevertheless account for all funds deposited since the passage of the Dawes Act in 1887 because of its pre-existing duty to account. My conclusion that the express terms of the statute place no temporal limits on the accounting is based on a reading of the statute that was apparently not suggested to Judge Lamberth.

17 . Note that an HSA is not the historical accounting itself but rather a by-product of the historical accounting. Interior’s position is that the mailing of the HSAs discharges the government’s duty to account.

18 . The description of the trust corpus in Co-bell XII as "consisting] of the revenues derived from land that was carved out of preexisting Indian reservations under the 1887 Act,’’ 391 F.3d at 254, was accurate but incomplete. Mitchell II clarifies that the Indian trust corpus consists of “[natural resources], lands, and funds.” United States v. Mitchell, 463 U.S. 206. 225. 103 S.Ct. 2961. 77 L.Ed.2d 580 (1983). Since Cobell XVIII tells us that the "alleged conflict” between prior Court of Appeals opinions in this case is "illusory,” 455 F.3d at 303, and the Cobell XII passage cited above explicitly cited the history of the trust laid out in Cobell VI, the declaration in Cobell VI that the trust corpus includes property remains controlling precedent.

19 . Defendants may object that they are unfairly surprised by my finding of impossibility, because I resisted plaintiffs’ invitation to refer to the October 2007 trial as an impossibility trial and have given little weight to plaintiffs' impossibility model. See Tr. 9:15-20 (Interi- or's opening statement). The impossibility issue nonetheless emerges irresistibly from this record as a conclusion of law, not because the 2007 Plan does not outline the "accounting that the plaintiffs believe should be performed,” id., but because the 2007 plan does not, and defendants cannot, achieve an accounting that passes muster as a trust accounting.

20 . The quotations are from Charles Dickens's opening description of Jamdyce and Jamdyce in Bleak House.

21 . I do not reach the conclusion urged by plaintiffs: that an adequate accounting is impossible because of the problem of missing records. The record before me is inconclusive on that point. The record is not inconclusive, however, on the tension between the expense of an adequate accounting and congressional unwillingness to fund such an enterprise.