The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX A EXAMPLES OF QUARTERLY STATEMENTS OF ACCOUNT TO ACCOUNT HOLDERS January 6, 2003 A-1 The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX A. EXAMPLES OF QUARTERLY STATEMENTS OF ACCOUNT TO ACCOUNT HOLDERS As the TFAS system replaced the IR/vIS system, Interior commenced sending Quarterly Statements of Accounts to IIM account holders. The final TFAS conversions, which were principally done on a BIA agency by agency basis, were completed in early calendar 2000. Thus, IIM account holders have been receiving Quarterly Statements for a number of quarters since the respective TFAS conversions were completed. The Quarterly Statements reflect an opening balance (the closing balance of the pr/or Quarterly Statement - for the first such Quarterly Statement, the opening balance was the closing balance in the former IRMS system), receipts, disbursements, other changes, and an ending balance. Each Quarterly Statement has a telephone number shown which IIM account holders can call if they have questions on their Quarterly Statement. Following are sample redacted Quarterly Statements. January 6, 2003 A-2 '2T. " )u HLVt AHT _uEsTIo_L_ ;E CS,L: OFFICE OF TRUST FUNDS MANAGEMENT APE_Y ;_8 .qqO5 . _?_,?,_ . 'INDIVIDUAL INDIAN MONIES Statement at Account · - _._/_.__.._._ ' :_.__.. , SZCTION ACTZVITY FOR JCCOUNTZH_ PERIOD: 09126/02 THROL_H 12/2.__01 TIE 3'RJkNSAC'T Z ON D_.%CRZPTZOH r.. Jk _N - I_[b-'_NNZN_ b.ILANCES 7.29 S/Il CASH REC:IE'rPT 0.oS HON'THLY ZNTEREST EARNZNGS PAZD FROM ACCDUNT I DOCunL.q'Fr I JV 5/11 CASH RECEZPT O.o& t_OPITKL.„ XHTEREST EAKNIHGS PAZD FRDH ACCO_qT I OCT 2IIZ EARNI_ _/I_ CASH RECEZPT _.3_ FARM PASTURE PAZD FRDt_ ACCDUHT I DOCUMENT I JV- 5/e1 CASH RECEZPT I.O& m3m_HLY 2NTEREST EARNINGS !'AZD FROPI ACCOUNT I NOV 2002 EARH2_ ENDZNG _ALANCE$ 10.74 ^-3 .._ 1.s tuu nA,C AFl1 _Ut_I.LL]_NL_ _.-_2oux_ASECALL:AF_ OFFICE OF TRUST FUNDS MANAGEMENT 2.373.10S2 INDIVIDUAL INDIAN MONIES ...... Statement of Aecount AHS,kCTZDH ACTIVITY FOR ACCOUNTINC PERIOD: ]O/14/O! THR'OUGH 01/13/02 bATE TRANSACTION DESCRI?TION CASH Z_EGINHIHG EAL)NCES 11, ol /15/ml CaSH RECEIPT 8.85 FARM PAS_TURE PAID FROI'I ACCOUITT m DOCUMENT II JR/' /25/111 CASH DIS t'Dcr"cu"r '1¶.86 PAID TO ACCOUNT lB ALAJ4C£ T,/A C_IECK II /IG/OZ CASH RECEIPT o,o/ FARM PASTURE PAID FROM ACCOUNT I DOCOMENT I /17/o). CISH RECEIPT 3.75 FARrl PASTURE raid FR[_'t accote_l' I _. DOCUllENT II JIL_ f_2/01 CASH R£CEIPT 0.08 FARH PASTURE PAID FROM accoIJHT · DOCl.n_HT il JV' /oS/il CaSH RECEIPT 0.03 IqONTHLY INTEREST EARNZNI_S PA.TD FROH accoUNT · OCT 2ooz E.ARNII'J¢$ ,'ll._J'o).. CASH laECEIPT MONTHLY INTEREST E,IRH21,.iC$ 11.02 P,IID FROn ACCOUNT I NOV 2o02 EARNING'S ,o_/02 ClSx RECEIPT 0,02 HONTHLY 2NTEREST EARNZN'GS PAID FROM ACCOLJ_r II DEC 2001 EARNINg'S ENDINg BALANCES 3.91 A-4 . ,_.._ .,Al ,,r,,.c_ I _OIq_' [ _ALL: OFFICE OF TRUST FUNDS MANAGEMENT _CENCY 38._40_ INDIVIDUAL INDIAN MONIES Statement of Account .... __,_ ,,_. ;ACTION ACTIVITY FOR ACCOUNTING PERIOD: 0_/26/11 THROUGH 127_,S/01 TE TRANSACT'IoN DET>C_IPTION CASH BEG'iNNING BALANcEs 6O.30 5/ol CASH RECEIPT e.37 _ONTHLY INTEREST EARNINGS PATD FRei ACCOUNT l DOCUMENT ! ,JV S/o! CASH RECEIPT e.a_ _ONTHLy INTEREST EARNINGS PAID FROM ACCDUNT I OCT 2Oil EARNINGS S/gl CASH RECEIPT O.3& HONTHLY 3NTEREST EA_NING_ PAID FROM ACCDUNT I NOV 20B2 EARNIN_ _/ll CASH RECEIPT X25.10 PER CAPITA - N_N TA_JBLE PAID FROM ACCOUNT P ee EHDZHG _ALANCES 2O6.*q A-5 15 you ,_vE _y _uEsT_o.s _,,s_ C, LL: OYFICE OY TRUST FUNDS MANAGEMENT' _CKFFET AGENCY ;._a.7518 " INDIVIDUAL INDIAN MONIES Statement of Account _NSACT_[ON ACTIVITY FOR ACCOUNTINC PERIOD: $9/2_]Q1 THRDUCN 12/2S/01L DATE TRANSAC_ON DE%CRIPTZD_I CASH i_E_ZNHZ:HC BALANCES 7.96 /a3/oI CASH RECEIPT t_ONTHLY INTEREST EARHIHG'S o. 14 PAID FROM ACCOUNT II DOCUMENT m ,.A/ /17/02 CASH RECEIPT OIL I gAS ROYALTY q.:L_; PA_D FRDM ACCOUHT I1 DOCLV_EHT II .pC.. /27/1! CASH RECEIPT OZL I CAS ZHTEREST 1.01 PA_D FR()_ ACCOUNT Il DOCUHEHT tl JVT /17/0"_ CASH RECE-/PT C_TL _t ;AS ROYALTY g,?5 PA2D FROM ACCOL_T 41 DOCUHENT _ JV' · .y /o$/02 CASH RECEZPT ' * tlONTHLY THTEREST EARNT:H&_ O.QS PAID tROll ._CT:OU_,IT f OCT 2e01 EARN1N_S /1S/ol CASH RECE'tPT FARt1 PASTURE 6.6g PA_D FRD_ ACCDUtJT II DOCUttEHT il I15/02 CASH RECEZ_'T FJ, RI9 PASTURE 13 -"2 g PA2D FROM ACCOUNT !1 DOCUMENT II JV /]-5/01 CASH RECEIPT INTEREST Q. O1 ...."_'-,. PAZZ) Fr. oH ACCOC,'HT # DOCU_EHT I J' *.. A-6 'f'_ i OFFICE OF TRUST i;'UiNq)S MANAGEMENT INDIVIDUAL INDIAN MONIES Statement of Account 'RANSACT_ON ACT'TVITY FOR ACCOUN'T/HC PER2DD: e9/26/01 _HROUCH 32./_S/01 DATE TRANSACTION DESC_ZPq'Z ON CASH L1/iS/e2 CASH BLS" ..... ,,.,ru.'e '_2.'7'T PAID TO ACCOUNT BALANCE T/A [:NECK II L3/3o/o:L CI.SH RECEIPT It.$0 OIL 8 CAS ROYALTY Paid IDEM ACCOUNT Il DOCUI'IENT tl 12./eS/'ll CASH RECEIPT O.O3 hOHTHLY INTEREST EARNZHb'_ PAID FRDH ACCDUHT II Nov Zeml [ARHINOS E_2_ BAL,_ NCE$ O.__B A-7 The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX B IIM ACCOUNT DATA January 6, 2003 B-1 The Historical Accounting Plan for Individual Indian Money Accounts Table B-1 - Balances as of December 31, 2000, and Dollar Throughput of Land-Based IIM Accounts for 1985 Through 2000 ' I "'""_-._._oughput -4 Less Than $100 to 'i $1,000 to $10,000 to $100,000 or $100 $999 _ $9,999 $99,999 More Total Balance 4, __ Less Than $100 $ Balance 878,018 306,977 J (71,788) _ {20,404 ) ' 623 1,093,426 # of.Accounts 70,057 52,000 28,683 3,889 219 154,848 # of Transactions 2,474,934 7,276,837 8,909,869 1,919,173 231,018 20,81 1,83 I $ Throughput 14,056,318 192,077,324 831,149,758 805,415,006 322,418,395 2,165,116,801 $100 to $999 $ Balance 2,326,5711 4,741,907 1,534,775 177,770 6,368 8,787,391 tt of Accounts 8,135 12,230 3,540 422 16 24,34__ it_ of Transactions 407,718, 1,387,110 1,095,85C 155,401 2,306 3,048,391 _; Throughput 2,583,963 ! 36,355,784 I 11,510,574 86,857,665: 22,777,091 260,085,07 / '$1,000 to $9,999 [ "_; Balance 3,505,981 I4,805,93 I' 15,484,495 2,059,153' 87,44!) 35,943,009 # of Accounts 1,472 5,886 4,280 51 I 27 12,17fi of Transactions 54,413 500,137 1,053,130 _ .241,411 15,664 1,864,755 $Throughput 410,875 20,021,408 118,527,900 133,366,427 79,484,0911 351,810,70 I $10,000 to $99,999 $ Balance 3,533,206!I 5,178,956 27,486,793 t 9,458,476 2,048,607 57,706,038 # of Accounts 153 267 1,119 545 46 2,130 # of Transactions 832 34,403; 286_.! 08! 229,231 37,666 588,240 iS'Throughput 17,794 1,694,884, 42,373,70C 126,471,119 156,252,991 326,810,488 '$100,000 or More I i $ Balance 1,340,99C! 1,024,084 10,083,398 37,395,176' 44,601,914 94,445,562 # of Accounts 7' fi 57 136! 61 267 # of Transactions 34 ' 105 I1,249 94,502 45,406 151,296 t; Throughput 111 -- 16,427 3,670,97{3 61,341,603 165,743,160 230,772,271 [. [ .„otal $ Balance 11,584,766 26 057,855 54,517,673 59,070,171 46,744,961 197,975,424 !Total # of Accounts 79,826 70,389 37,679 5,503 369 193,766 Total # of Tra nsa...c.!,!ons 2,937,931 9,198,592 I 1,356,212 2,639,718 332,060 26,464,513 Total $ Throughput 17,069,061 250,165,827 1,107,232,902 1,213,451,820 746,675,728 3,334,595,33'3 Source: CD&L based on data from OSI/OIFM. January 6, 2003 B-2 Tile Historical Accounting Plan for Individual Indian ]Money Accounts Table B-2 -Summary of IIM Accounts at December 31, 2000 a Number of.Accounts $ Balance (in millions) Land-based Accounts 193,766 $ 198.0 Judgment Accounts 33,205 80.8 Per Capita Accounts 9,013 69.5 Total IIM Accounts 235,984 $ 348.3 Special Deposit Accounts 21,415 67.9 Total Current Accounts 257,399 $ 416.2 House, Foreign, to be investigated and other accounts 4,071 Total Open Accounts 261,470 Inactive Accounts b 17,840 Total IIM Accounts c 27%310 a Source: Office of Trust Funds Management December 3 I, 2000, ItM Statistics Report. bThese accounts have no recent activity, but are expected to have future activity. c Whereabouts unknown accounts number 61,673 and represent $65.4 million. These accounts are included in the total 279,310 accounts. They represent account holders whose addresses are unknown. These account holders do not presently receive Quarterly Statements of Account or periodic disbursements. January 6, 2003 B-3 The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX C QUALITY CONTROL January 6, 2003 C-I The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX C. QUALITY CONTROL As noted in its July 2, 2002, ReFort to Congress, OHTA is implementing quality control (QC) checks to achieve best practices at each phase of the historical accounting process, including data inputs, systems and infrastructures, processing, and outputs. This Appendix describes details in five areas of OHTA's quality control plan. · Basic approachto quality · Overall quality check process · Record collection/recordimaging · Accounting results · Information systems Quality performance has two basic premises--planning and supervision, and sufficient relevant data in working papers. These premises are included in each of the foregoing areas. Basic Approach to Quality High standards intended to result in high quality results are central to conducting an adequate historical accounting. Another key factor is appropriately focusing quality efforts. The overall quality focus is _k-based, i.e., determining what major risks significantly impact the quality of results and what strategies can minimize these risks. One risk is inconsistent results. OHTA is striving for consistency through standardization on common systems, procedures, and processes across the various orgmfizations involved. For example, OHTA is centrally developing a historical accounting support system, which will be used by the accounting firms to produce accounting results. Another example is OHTA's issuance of guidance, such as demonstrated by OHTA's Accounting Standards Manual (Manual). OHTA conducts training programs to obtain common understandings and has implemented centralized issue tracking to resolve issues timely and appropriately. Besides establishing a common infrastructure for the various firms involved, OHTA is conducting QC checks and tracking results. Trends will be analyzed to determine whether additional guidance, training or other action is necessary to achieve quality objectives. OHTA senior management receives statistical information regarding the quality of work packages along with other key performance data, such as transactions and accounts reconciled. Performance data gathered will be used to refocus QC testing strategies. For example, the data may indicate increased or decreased QC testing is desirable in selected activities. January 6, 2003 C-2 The Historical Accounting Plan for Individual Indian Money Accounts Quality objectives are being achieved through both internal and external quality checks. Each team involved w/th producing inputs or outputs of the historical accounting bears primary responsibility for the quality plan for its products, such as responsibility to maintain internal quality processes, staff the project with appropriately skilled personnel, conduct staff training, and perform supervisory review. Additionally, certain subject-matter expert consultants are conducting independent quality checks on their performance. Overall Quality Check Process An OHTA external QC check typically involves six steps. 1. Plan -Begins with identifying the area that will undergo a QC check and determining the objective, scope and timing of the QC check. A QC program is developed and documented in a QC checklist to maintain a consistent process. The plan takes into account previous QC results to develop adaptive strategies, such as the need to further test additional activity in populations. 2. Conduct Analysis- Consists mainly of activities related to conducting interviews with performing consultants or OHTA staff, and inspecting documentation. The analysis step also involves identification and discussion of preliminary observations with the team that performed the work undergoing a QC check. 3. Draft Report- The QC firm documents the results of analysis in a draft report. The report is then provided to OHTA senior management, OHTA's contracting officer representatives, and to the contractor whose work underwent the QC check. The draft report contains a description of QC results, key observations and suggestions for improvement. 4. Resolve Fi_Tdings - The contractor whose work underwent the QC check has an opportunity to respond formally to the QC draft: report. Should differences of opinion exist between the QC firm and the contractors as to the proper resolution of a comment, both parties present their rationale to OHTA senior management for its input. 5. Issue Final Report - After comments are discussed, the QC firm will revise its report as necessary and distribute the report. 6. Conduct Follow-up, as Necessary - In some cases, the contractor may have 'action items' needed to complete the work product that underwent the QC check. The QC firm tracks these 'action items' and conducts a follow up QC check to verify the action item is completed successfully. Additionally, as a result of the QC check, issues may arise that are communicated to the parties involved in performing the historical accounting, such as sharing of 'best practices', or updating of the Manual or other guidance used to conduct the historical accounting. January 6, 2003 C-3 The Historical Accounting Plan for Individual Indian Money Accounts Record Collection/Record Imaging Interior plans to implement QC checks at each phase of the records collection and imaging process. This includes checks on the design of processes, procedures, and systems to conduct the collection and imaging. For example, a QC check will be performed on the document imaging and coding structure. Once implemented, periodic QC checks will be conducted to verify the design is being implemented as per management's intent. QC checks are anticipated to cover the following areas: · Imaging and coding structure; · Completeness of documents collected, imaged, and returned to storage; · Legibility of images; · Accuracy of coding; · Record/box inventory; · Security; and · Document preparation and re-assembly (pre-and post imaging). Accounting Results The accounting results produced by the accounting firms undergo QC checks by another accounting finn. The QC finn samples results for consistency with the Manual. The Manual provides guidance, such as documentation standards. For transactions sampled, the QC firm assesses the judgment made by the accounting finn as to adequacy of supporting documentation, classification of the reconciliation results, and calculation of errors (if any). Additionally, the QC firm assesses historical accounting results in light of professional standards established by the American Institute of Certified Public Accountants. These standards include professional competence, due professional care, planning and supervision, and sufficient relevant data. Information Systems The support system under development for the historical accounting is focused on meeting OHTA and its stakeholder requirements for account reconciliation. This will be accomplished by following a structured development life cycle approach. This approach encompasses trust requirements that are clearly defined and traced throughout the life cycle of the historical accounting project. At each life cycle milestone, products are tested against requirements and communicated with the users for their information and feedback. To further assure that OHTA has an information system that answers its needs, a contractor has been engaged to provide independent verification and validation that deliverables meet the needs of the users. January 6, 2003 C--4 Tile Historical Accounting Plan for Individual Indian Money Accounts Key system life cycle deliverables include: · Requirements documentation- What the system is to provide; · Design documentation- How the system works; · Implementation- Installation of the system; · Test plans/results - Periodic verification of system reliability; · Data conversion - Processes used for conversion of data from other information systems; and · Training- Ensure that system users understand this tool. Additionally, once systems are in operation, periodic post-implementation reviews will be conducted to verify they are meeting management objectives, such as data security. January 6, 2003 C-$ The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX D SAMPLE DESIGN PLANNING REPORT PREPARED BY NORC January 6, 2003 D-1 The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX D. SAMPLE DESIGN PLANNING REPORT PREPARED BY NORC _ 1. Main Coal In the historical accounting of the Individual Indian Money (IIM) Accounts, various problems arise which are to be addressed by statistical sampling methods. 2 This Appendix discusses one particular application of sampling, that of audit sampling, 3 in the historical accounting: The use of sampling to test the accuracy and completeness of the land-based IIM account statements. An important goal of the historical accounting effort is to provide each IIM account holder with a Statement of all transactions in the account. Accompanying this Statement is a description of its completeness and accuracy, and an estimate of the needed adjustment (if any) for each account. For large transactions something like the following may be said. "Enclosed is a copy of your Historical Statement of Account. All transactions with a value of $5, 000 or more have been checked against the supporting documents; any discrepancies that were found are shown." For smaller transactions, where sampling occurs, the resulting statement and conclusions depend on what is learned. However, if the results are good, then something like the following might be said. "For transactions under $5,000, statistical sampling procedures were used to verify the accuracy of the process. With 99 percent confidence, we can say that more than 99 percent of the transactions are accurate." Many errors may be clerical in nature, for example data entry errors or errors caused by two pages in a document sticking together. After a modest amount of checking, such errors may be treated as random and not systematic. This error rate should be small, but it may vary by BIA agency or location, type of transaction, and time period. The sample is designed to capture these errors when they exist and to quantify their importance with high accuracy. This Sample Design Planning Report represents NORC's recommendations. Therefore, in places, for example, there may be minor timing differences between this Report and the Plan. 2 Sampling is to be used in assessing and controlling the quality of the searching for records and even, as Appendix C states, in the accountings themselves (For a general view, see Deming, W.E. 1986). 3 Statistical sampling is routinely used in audit practice, as described in texts such as Guy (1994) and Wilbum (1984). See also AICPA (2001). January 6, 2003 D-2 The Historical Accountlng Plan for Individual Indian Money Accounts The sample must also be designed to do more. Not all errors are random and independent. There could be "pockets" of errors that are systematic, that are due to a procedure that was not well explained, or an algorithm that was incorrectly programmed, or an ownership interest that was misapplied. These errors can cluster together within an IIM account, among owners of the same tract of land, in the same BIA agency or during a particular time period. By using an adaptive approach (Thompson 1996), the sample procedures are designed to deal with unexpected or systematic patterns or "pockets," if these are detected. When an error is found on a randomly selected transaction, and it is not just clerical (e.g., a transposition of digits), then additional work is performed "around" this error. These additional sample tests are chosen with the specific purpose of pursuing and correcting sysiematic mistakes. The adaptive aspect of the sample design is quite challenging technically because it combines forensic accounting ideas with rigorous probability sampling (e.g., Thompson, 1996). Section 7 below sketches the issues briefly. It may be enough to say here that we expect this two-step sampling process to be frugal in its use of resources, while still providing great benefit to account holders. 2. Soecific Scope The transaction population oflIM land-based accounts has a natural division in terms of the information that is currently available. For the recent transactions, from the mid- 1980's fo_vard, information on the transactions is available electronically - via a database with transaction information, n These transactions are referred to as being in the "Electronic Records Era." Transactions prior to the mid-1980's are referred to as being in the "Paper Records Era," because paper ledgers need to be found and an electronic ledger created. The sampling of Electronic Records Era transactions occm-s first, and the design of that process is described in this Appendix. The design process for the Paper Records Era is also discussed, but only briefly (see Section 8 below). Details on sampling in the Paper Records Era are to be covered in a subsequent design report, when the data are available. 4 The conversion of the IIM to the Electronic Record Era was done over several years, at different points in time in different places. Gaps exist too. For these, electronic records must be recreated by retrieving original paper records or still available computer printouts fi.om files that are no longer electronic. See Zhang and Scheuren (2002) for dates and locations when the Electronic Record Era started. Also found there is a summary of the gaps. January 6, 2003 D-3 The Historical Accounting Plan for Individual Indian Money Accounts Land-based IIM accounts are those that derive income from allotted tracts of land. The transactions for IIM land-based accounts can be classified into three general types: collections, disbursements, and interest accruals, s It is our understanding that interest accruals are being recalculated and compared with the actual interest amounts that are posted; hence, interest verification is not in scope for sampling. Judgment and Per Capita IIM accounts are also out of scope here, because sampling is not required. Typically the same amount is provided to each authorized individual in setting up a Judgment or Per Capita account. Because the record searching and verification efforts are minimal in these cases, they are being reconciled on a 100- percent basis, without utilizing sampling. Special Deposit Accounts are also out of scope for sampling, as described in the Plan. While the final sample design cannot be determined until further data analysis has been performed, this Appendix describes the process in broad outline. In order to make the process concrete, a specific example is provided of a sample design for the Anadarko Agency, given hypothetical numbers. 3. Ma]or Types of Samples Four main approaches to drawing samples from the IIM Trust Fund activities involving IIM land-based accounts 6 are considered in detail in this Appendix. These are -- · Selecting Individual Transactions · Selecting Unique Receipts · Drawing Supporting Documents · Taking Entire Accounts The different sampling options grow out of the complexity of the transaction population and the different ways that one can approach verification. Each has advantages and disadvantages- all of which are take into consideration in the design of the final samples. 5 Transfers from one account to another within the IIM are a fourth transaction type that needs to be verified by sampling. Transfers can be either a disbursement or a collection. It is hardto make any generalizations about them for this reason and because their treatment varied over time in different locations, including during the software changeover from IRMS to TFAS. The best way to treat them remains to be determined. Almost certainly they will be put into separate strata. 6 The accounts being considered for sampling in this Appendix are the accounts in 11 regions (the Eastern Region is excluded) that derive income from allotted land. Eastern Region accounts are being fully reconciled on a transaction-by-transaction basis, because they are so few in number. January 6, 2003 D-.4 The Historical Accounting Plan for Individual Indian Money Accounts A. Statistical Sampling oflIM Transactions Description: Statistical sampling of transactions begins by randomly selecting transactions posted on an electronic IIM database (for gaps in the Electronic Records Era after 1985 and for periods prior to the Electronic Records Era but after June 24, 1938, such a database would be created from paper ledgers prior to sampling). Goal: To determine, with an acceptable level of confidence, that the error rate among the transactions posted to an IIM account is below an acceptable maximum. Several ways are under consideration that would lead to an estimated maximum dollar error for each land- based IIM account. Several techniques in the statistical literature are to be examined. What we will recommend, however, depends on what is learned from the data. Some of the statistical references being studied are cited in Section 9. Process: The sampling process involves first stratifying the transactions by dollar size, location, and era. For each randomly selected transaction within the established stratum, an attempt is made to obtain supporting document(s). The document(s) found are then checked by accountants to determine whether the posted amount is accurate as to dollar amount, IIM account number, and in accord with the supporting documentation (e.g., leases, journal vouchers). Strength: The method is conceptually simple and easy to communicate. It allows a scientifically defensible statement about the transaction error rate, including the dollar error rate. If there are known sources of accounting errors (e.g., in certain types of transactions or transactions at a specific location) the approach can be adapted accordingly (e.g., by selecting 100 % of these transactions, i.e., taking them all with certainty). Weakness: The approach is unable to detect erroneous omissions on the IIM account (receipts that have never been posted to the proper IIM account). It can also be difficult to find all supporting documentation (leases, vouchers) or "supports" for a given selected transaction. Methods are available to address this concern. The approach to finding supporting documentation is similar to that used earlier in the "Paragraph 19" document search in CobelL It is anticipated from this earlier work, therefore, that it may be quite expensive to locate all the supports. Recommendation: We recommend that this approach be used for statistically sampling disbursements. It can also be used, as adapted elsewhere, for other types of transactions, except receipts testing, which is described next. B. Statistical Sampling of Unique IINI Receipts Description: Transactions for land-based revenue can be grouped by 'the type of transaction (e.g., collection) and the type of revenue (e.g., timber sale) using the reference code of the transaction. In this way, income from one source (e.g., a bill for collection) that is distributed to several IIM accounts can be related back together as a unit, based on January 6, 2003 D-5 The Historical Accounting Plan for Individual Indian Money Accounts the IIM transaction coding information. Such a grouping of individual revenue transactions is called a unique receipt.* The statistical sampling under this test begins by randomly selecting unique receipts and then verifies one or more individual IIM transactions. Goal: As stated above, to determine, with an acceptable level of confidence, that the error rate among the transactions posted to an IlM account is below an acceptable maximum. Such a result, when used in combination with other statistical testing described herein, might permit the computation of an estimated maximum dollar error for each land-based IIM account (see Section 9 below). Process: Unique receipts can be stratified by the total dollar value and by the number of individual account distributions within each transaction. For a randomly selected unique receipt, the accountants verify that the amount of the receipt ,,,,'as credited appropriately to the proper IIM accounts. Another option, when there are multiple contracts or leases involved with the same unique receipt, is to verify a random subsample of the IIM transactions on the selected unique receipt. Strength: For land-based receipts, this approach can be cost-effective because a single unique receipt affects a number of land-based IIM accounts. This method can more efficiently test large-dollar amounts coming into the IIM Trust Fund that are subdivided into many accounts so that, at the IIM transaction level, the amounts are small. Strata based on dollar amounts of the unique receipts can be separately tested. Weakness: This approach applies only to certain types of receipts, and does not include any disbursements. In addition, it does not target the largest individual transaction amount. Furthermore, as with the direct sampling of individual IIM transactions, this approach is unable to detect erroneous omissions. Recommendation: We recommend that unique receipts be the sampling unit for land- based IIM receipts. The stratification of unique receipts is, however, to be by the size of the largest individual transaction they contain (see Section 7 below for an example). C. Statistical Sampling of Supporting Interior Documents Description: In all the samples being drawn, accountants use supporting documentation to verify the accuracy of a transaction.' Documents may be in active use in file drawers at an agency office. Documents, not in active use, are in boxes stored at various locations 7 Not all "unique receipts" pertain to a single contract or lease. This is especially true in TFAS where what appear to be unique receipts may actually involve many different contracts on different properties. For cases when only a single contract exists, only one IIM transaction needs to be tested, when testing one is testing all. When there are multiple leases more than one transaction has to be given a chance for testing. This subsampling is done once the documents haYe been located and separate contracts identified. 8 Some examples of supporting documents include negotiated checks, leases, rights-of-way, journal vouchers and so forth. January 6, 2003 D-6 The Historical Accounting Plan for Individual Indian Money Accounts around the country. As part of this third approach, these paper files are going to be directly subjected to sampling. The statistical sampling begins by selecting a random sample of storage units (file drawers, boxes) containing supporting transaction documents. The contents found in each storage unit are then subsampled. In this case, randomly selected subsamples of support documents are matched to appropriate postings on related IIM account(s); that is, the matching goes in the opposite direction, from supporting documents to IIM electronic records, rather than the reverse. Goal: As with the direct sampling of transactions, sampling support documents determines, with an acceptable level of confidence that the error rate among the transactions posted to an IIM account is below an acceptable maximum. Such a result, when used in combination with other statistical testing described herein, might permit the computation of an estimated mcccimum dollar error for each account. Process: The sampling process involves first stratifying the storage units by agency/area location and time period. Every selected storage unit or box of records is triaged and all required land-based supporting records found are imaged and coded so they can be selected for matching to a transaction posted in the related IIM account. An attempt is then made to verify that activities documented in the land based supporting documents are consistent with and support related IIM transactions. Strength: This is the only approach that is able to detect omitted transactions on IIM ledgers (i.e., receipts or disbursements that have never been posted). A sampling approach that incorporates some document sampling is, therefore, much more robust and easier to defend than a pure account/transaction-based method which is incapable of detecting a missing posting (a major concern expressed by IIM account holders). Directly sampling documents is expected to be cost-efficient because it is much easier to match a document to its related IIM account posting than it is to take an IIM transaction and find the pertinent supporting documentation. Weakness: If there are agencies/time periods/income types for which supporting documentation is particularly incomplete or poorly maintained, the corresponding transactions are less likely to be included in the sample. Also, if used alone, accounts with few transactions are less likely to be included within the sampling process. Finally, while it is straightforward to determine the probability of selecting a particular document, determining the selection probability of a transaction is challenging since an unknown number of copies of the same document may exist in the files. The Cobell record collection is to be examined to assess how to handle this. Recommendation: We recommend that this approach be relied on heavily for sampling small transactions. It should be supplemented by methods that permit stratification by transaction size and, hence, is combined with the first two methods explained above. The search costs should be less for this approach because the search starts with one of the supporting documents already on hand, making the overall "find" rate for documents higher. January 6, 2003 D-7 The Historical Accounting Plan for Individual Indian Money Accounts D. Statistical Sampling oflIM Accounts Description: The sampled population consists of the land-based IIM accounts (using the population definition of accounts open on or after October 24, 1994). The sampling begins by randomly selecting IIM accounts. Goal: To estimate, with a certain level of confidence, that the account error rate, in terms of dollars orpercemage of dollars, is below an acceptable maximum. Process: A stratified sample of accounts is selected (stratified possibly by location, throughput, number of transactions, ending balance, etc.). For the selected IIM accounts, supporting documents are found for all transactions and the accuracy of the Historical Statement of Account is verified. The dollars in error are then measured directly. Strength: Account sampling provides complete accounting for the selected IIM accounts and the sample verifications for these accounts can be used directly. It is also easy to explain. Weakness: Account sampling is the most expensive sampling method in terms of time, cost, and required work force. Furthermore, the total cost cannot be readily determined with precision. The variability in cost is high, depending on the number and complexity of transactions in the selected land-based IIM accounts. Costs depend not only on the number and type of transactions in the sampled IIM account, but also on how diverse the account is (some accounts cover many allotments, over several agencies, even agencies in different regions). Finally, it is difficult to predict the sample size necessary for estimating the dollars in error with specified precision. At this point, there is minimal information available to allow reasonable predictions. Recommendation: We recommend that this approach be used sparingly because it is not cost effective. As already noted, it is very difficult to estimate the (transaction) sample sizes necessary for specified (account-based) confidence levels without prior information regarding the properties of the errors - information that only becomes available after the project is completed. 4. Supplementary Checkinl_ A large number of transactions are fully reconciled directly through sampling. Ideally, all sampled cases are completely resolved, with no room for controversy or disagreement. However, this rarely happens in practical situations; hence we are developing plans for handling those cases for which incomplete or contradictory supporting evidence may be all that can be found in Interior files. In particular, we concur with the policy that asks third parties to notify the Department of the Interior before destroying potentially relevant records - notably, small local banks in January 6, 2003 D-8 The Historical Accounting Plan for ]ndividuai Indian Money Accounts Indian Country that may have kept canceled checks for longer than the now standard seven years. There are other tests, too, that must be made of certain processes or databases. For example, in order to check that the income from land is correctly allocated, the distribution of receipts by degree of allotment ownership must be verified. This step is part of the sample verification process within IRMS but the IRMS recorded ownership is carried back to still other sources, such as the Land Record Information System (LRIS). 5. Basic Activities The samples of transactions, unique receipts, and supporting documents that were described earlier in Section 3 are integrated in stages to determine the overall sample design. In this section we describe the process for determining the final sample design. The more information that is available about the population, obviously the better one is able to design an effective and efficient sample. The first step, therefore, is the analysis of the electronically available data (e.g., as in Zhang and Scheuren 2002). Analysis is important for determining that the proper population definitions and data are being used, and for learning as much as possible about the population, so that the sample can be stratified and balanced effectively. The following provides an outline of the procedure for developing the specific sample design and the considerations that need to be analyzed or researched. Transactions are treated differently, depending on their type: disbursements versus receipts, the different income sources for receipts, the year of the transaction, whether the individual transaction was part of a unique receipt, where the income was earned, the type of account (e.g., administrative versus individual), and any other information that might reveal special problems or characteristics. Because individuals often own interests in more than one allotment, the income to an account may be generated from many different agencies and even from different regions. Therefore, the geography of the accounts and the geography of the transactions need to be considered. For each agency, we expect to use the following definition of the population of transactions: ,, Ail disbursement transactions for accounts in the agency · All TFAS collection transactions for accounts in the agency · All IRMS collection transactions that originated in the agency (i.e., transactions where the source of the income was in the agency of interest). Once the population is defined by agency, the distribution of the transactions by value can be examined. January 6, 2003 D-9 The Historical Accounting Plan for Individual Indian Money Accounts For a receipt or disbursement, more than one supporting document is usually needed. For income based on the land, the source for at least one of the supporting documents is the agency/area where the land is located. Every individual transaction valued at $5,000 or more is to be reconciled to its related supporting documentation. This means that, based on accounts open through December 31,2000, about 73,500 large IIM transactions are to be reconciled in that one stratum alone. For transactions of less than $5,000, two additional strata based on the transaction value are used - transactions under $500 and transactions $500 to $5,000. Each of these has a sample size provisionally set at 80,000 (see table 1). These transaction sample sizes are chosen to provide agency-by-agency information such that for each stratum, statements about transaction error rates can be made with high confidence (if the error rates are as low as the 1 percent designed for). Table 1. Stratification by Size, based on Accounts Open on December 31, 2000 Size of Percentage of Approximate Number of Number of Transactions Total Dollars Transactions in the ' Transactions to be Population Sampled _ _ Under $500 25 25,592,000 80,000 $500 to $5,000 30 800,000 80,000 $5,000 or more 45 73,500 73,500 Total 100 26,465,500 233,500 Note: The final population subject to sampling should be similar in terms of the distribution by size to the population shown above, which is based on accounts that were open on December 31, 2000. Sample sizes of about 400 per agency, per stratum, would suffice for this purpose if no errors are found. These sample sizes are then doubled to afford protection against the possibility that, in some agencies and strata, the observed error rates might be somewhat higher. At a sample size of about 800 per stratum out of a total population of, say, 25,000, it is possible to make a 99 percent assurance statement that the error rate is less than about 1 percent if up to 2 errors are found. The estimated total sample size per stratum is calculated by allowing for the fact that there are about 100 agencies or agency-like organizational units that are to be sampled separately. Other sample sizes were examined, both smaller and larger, ranging from 10,000 to 320,000 nationally (or from 100 to 3,200 by agency). These are shown in Table 2. Gains in higher levels of assurance from samples above 160,000 are very modest under the asst_mptions being made here. For samples below 160,000, on the other hand, assurance levels drop to what we consider unacceptable levels, particularly so with samples under 80,000. One final comment, the design being recommended here has the advantage of balancing assurance levels and accuracy levels at the same high level of 99%. January 6, 2003 D-10 The Historical Accounting Plan for lndMdual Indian Money Accounts Table 2. Assurance Level that the Error Rate is less than 1%, if No Errors Are Found Total Sample Agency/Area Stratum Sample Assurance Level* Size Sample Size Size (Percent) 10,000 100 50 40 20,000 200 100 64 40,000 400 I 200 87 80,000 800 400 98 160,000 1,600 800 99+ 320,000 3,200 1,600 I 99+ *Source: Table 3.8, Audit Sampling by D. Guy et al, 1994. For disbursement transactions of less than $5,000, after further analysis (to explore other possibly important strata characteristics), strata sample sizes are determined and the sample balanced at least across time and so that transactions are drawn from as many different accounts as possible. For receipts of less than $5,000, a more complex structure is needed. Rather than using the transaction as the sampling unit, unique receipts are used, as described in Section 3. First, unique receipts are stratified by the largest indi„idual receipt they contain. Then, unique receipts are subsequently stratified or balanced, as much as possible, based on the type of income (oil and gas, surface lease, etc.) and time, within each agency. An additional complexity arises, as already mentioned, when, a bill of collection or unique receipt includes, for example, multiple grazing permits or farming leases. For cases when only a single contract exists, only one IIM transaction may need to be tested, when testing one is testing all. When there are multiple leases more than one IIM transaction has to be given a chance for testing. This subsampling is done once the documents have been located and the existence of multiple contracts identified. Finally a separate sample is to be selected using the supporting documents as the sampling unit. In Section 7 below we provide a complete illustration for Anadarko. 6. Implementation Slruelure For every BIA agency/area, all large transactions (with a value of $5,000 or more) are tested. By agency separately stratified random samples are selected from all transactions with values less than $5,000. All the cases for each BIA agency are then randomly divided into batches that are balanced across the three strata. Once the supporting documents are found for the selected transactions, accountants determine the accuracy of the respective land-based IIM account postings, beginning by looking at the first random batch. If the transactions are reviewed in random batches in this way, then after each batch of transactions is completed in an agency, an estimated error rate and confidence interval can be calculated. This approach can provide progressively better estimates of January 6, 2003 D-11 Thc Historical Accounting Plan for Individual Indian Money Accounts accounting costs and error yields, as the work proceeds. Moreover, if the errors are few, it might make sense to stop the sample verifications before the entire sample is completed, in order to provide the account holders with information more quickly. All of the transactions with an absolute value of $5,000 or more are still completed, in any case, because these are not being sampled but are to be reconciled on a 100 percent basis. Perhaps the most important feature of the design is that the results from batches completed early can be used to change the stratification for later batches. This adaptive approach can make it possible to increase the efficiency of the design as time goes on. The sample sizes described earlier in Section 5 allow for the pursuit of systematic errors and adaptive approaches, and are therefore upper limits, developed to make it possible to budget for the effort. It is expected that a frugal use of these resources can make the undertaking less expensive than the upper limit, unless the error rate in the underlying data is very high, or there are other unexpected results. 7. Approach to Anadarko Pilot Samplinff This Section describes how the sampling plan described in the preceding sections is performed. The Anadarko Agency is selected to use as an example because it contains a variety of income sources, covering most types except income from forestry. In what follows concerning Anadarko, we rely heavily on an analysis of Anadarko Agency files using transactions on the 7,206 accounts open as of December 31, 2000 (even though the final accounts to be sampled in the Electronic Records Era are those open on or after October 25, 1994). 9 Before the final sample can be selected, the electronic data will be reanalyzed so that the "geography" of the IRMS collections is defined by the source of the income, rather than the location of the land-based IIM account, l° The process defined here will remain about the same, but the population numbers will change. 9 The detailed analysis conducted of the Anadarko transaction file is found in the report entitled, NORC Sample Design Planning Report Part I1 -- Anadarko Agency IIM Transactions Described (available separately). 10 As noted earlier, for disbursements the account and transaction geography should be identical but not for receipts, where income may come from widely separated locations, far from the home agency of the account holder. For example, by marriage or migration, an individual undivided interest in a tract may be held by someone who no longer lives near the land she or he partly owns. In IRMS, which covers most of the Electronic Era (until mid-1998 or even as late as early 2000), the paper documents supporting a transaction are filed based on the agency or area where the income was earned, so the work must be organized on that basis. In particular, the IRMS portion of the electronic record transaction database is first sorted by transaction geography to facilitate the accounting work. For TFAS, which begins in 1998 or later, transactions are filed in Albuquerque no matter where the income was generated. To fred the supporting documentation for a transaction in, say, late 2000, the paper files in Albuquerque have to be searched by date and, then, once located the source of the income can be obtained. Some of the supports are present with the back-up materials in Albuquerque but the leases and contracts are going to have to be obtained fi.om the agency where the income is located. January 6, 2003 D-12 The Historical Accounting Plan for Individual Indian Money Accounts For the Anadarko Agency, Electronic Records Era data generated from IRMS and TFAS show there are 457,195 recorded disbursement transactions and 684,992 recorded receipt transactions. The data used here are for illustrative purposes only; certain complexities in the database are not addressed at this time. Most disbursements (417,023) are under $500 (91%). Of the remainder (40,172), most are between $500 and $5,000, but there were 2,495 in amounts of $5,000 or more. The distribution of IIM receipts is similarly skewed with many small transactions and few very large ones. Most receipts (668,566) are under $500 (98%). Of the remainder (16,426), most are between $500 and $5,000. Again, as with disbursements, a much smaller number (728) had dollar amounts of $5,000 or more. For sampling purposes, the disbursements can be selected directly, within the three strata, balancing, as already noted, on type of account and transaction date, including a separate treatment of disbursements before and after the date Treasury records n are likely to be complete and before and after the start of TFAS (in April 2000 for Anardako). Before sampling the 684,992 individual receipt transactions, we group them by type, code, reference, and date into 81,646 unique or combined receipts, as discussed earlier. A. IIM Disbursement Certainty Stratum In the sample described here, all individual disbursements of $5,000 or more are included with certainW and the rest are sampled. This means that within the certainty stratum under this design there is a complete accounting for 2,495 transactions involving $32.4 million in recorded disbursements. Thus, about 25.8 percent of the $125.6 million disbursed in the Electronic Records Era by the Anadarko Agency are tested directly by this means. B. IIM Disbursement Noncertainty Strata An initial sample of 200 is taken of the remaining IIM disbursements, stratified by size of disbursement (under $500 versus $500 to $5,000). The selections are balanced by the year of the transaction (for each year 1985 to 2000) and spread as widely as possible across IIM account holders. The remaining 200 initial disbursement sample cases are drawn from the sample of supporting documents described below. 12 _ The Department of the Treasury now retains records of all Treasury checks cashed. This source is believed to be complete over at least the last seven years but is still expected to be useful in earlier years. _2 The ability to stratify the support sample before use is almost certainly going to be imperfect, hence the fraction of the total sample drawn from the two sources may be uneven. The goal of achieving combined initial samples of 400 transactions fi.om each stratum can still be met, of course, but more of the cases in the $500 under $5,000 class may be coming from the IIM frame and less from the support sample. January 6, 2003 D-13 The Historical Accounting Plan for Individual Indian Money Accounts C. 115/I IndMdual Receipt Certainty Stratum For Anadarko IIM receipts we also propose to include all unique receipt transactions that have one or more individual transactions of $5,000 or more for the Electronic Records Era. In Anadarko there are 728 such individual transactions of $5,000 or more, as indicated earlier. Since some of these occurred as part of the same unique receipt, there are fewer unique receipts (only 487) involved. It is our understanding that, because of the nature of the testing being done, in the process of confirming one of the individual receipts in a unique transaction, other of the individual receipts may be considered tested. In that case, it turns out that even though receipts for indMdual transactions of $5,000 or more amounted to $7.6 million, we may be able to test with certainty up to $8.8 million in unique receipts (on 2,496 individual transactions) or 18.1 percent of all Anadarko receipts available electronically (see figure 1 on the following page). D. IIM Individual Receipt Noncertainty Strata A sample of 200 unique receipts is taken of the remaining IIM unique receipts, stratified by size of the largest individual receipt (under $500 versus $500 to $5,000). The selections are then balanced by year of the transaction (for each year 1985 to 2000), unique receipt class, and type of income. Again, as with the disbursement selections, the sample is again balanced to be sure that the receipt transactions examined are spread as widely across IIM account holders as possible. The remaining 200 initial receipt sample cases are drawn from the sample of supporting documents described next. E. Direct Support Sampling A second sample is drawn directly from the supporting documentation underlyin g the Anadarko transactions that are supposed to be recorded on the account ledger. This sample begins by selecting a stratified sample of 150 storage units from the Southern Plains Region. The selections are made at the Anadarko Agency, the FRC at Lee's Summit, Missouri, and various Interior record storage locations in Albuquerque, New Mexico. Because of concerns about records being misfiled, the sample is drawn regionally in Albuquerque and at Lee's Summit. (This same sample can later be used in part for other Southern Plains Agencies, of course. That is why it is so large.) Using the inventory of the storage units or the related BIA paperwork for information on the contents, the boxes or storage units are stratified to oversample those with a high likelihood of containing land-based collection documents or disbursement documents (e.g., negotiated checks). Each selected box or storage unit is then triaged to prepare the contents for imaging, if appropriate. After scanning (onsite)the images are coded for matching to any applicable IIM transactions. The intent, at this point, is only to match the main collection and disbursement supports, although this could change as we learn more. January 6, 2003 D-14 Figure 1: Distribution of Individual Transactions and Unique Receipt Groupings in Anadarko Agency 684,992 transactions transactions$5'O00or more:formingI__/_ "*""'"'"'*"%'""...I Remaining 1 ii 728 ',_ not in Stratum A / $500 to $5,000: "%. 15,153 transactions *'.% forming 9,048 unique %% receipts _ "... ,/ IRemaining 599,878 I Transactions from 487 ' 1 [Transactions from 9,048 ] Itransactions, all under unique receipts: 2,496 lunique receipts: 82,6t8 ! 15500, contained in (Stratum A) / | (Stratum e) / 172,111 unique receipts t / [ (Stratum C) Notes; As displayed above, the population of individual receipt transactions is first divided into those of $5,000 or more (of which there are 728) and the remainder under $5.000. The 728 individual receipts are included in 487 unique receipt groups with 2.496 individual transactions. All other transactions except for 2,496 are divided into those between $500 and $5,000 (15,153) and under $500 (667,343). The 15,153 individual receipts are included in 9,048 unique receipts, with 82,618 individual transactions. The remaining individual receipts (all under $500) ere included in 72,111 unique receipts and total 599,878. Thc Historical Accounting Plan for Individual Indian Money Accounts A collection document is matched to an IIM transaction to ensure that all money entering the Trust actually ended up as an IIM receipt transaction or that there was a good reason why not. Disbursement supporting documents are obtained directly, so as to supplement the IIM disbursement samples. There is some considerable concern that not all the supports needed for the IIM disbursement transaction sample are found and this second sample provides a chance to examine whether failing to find all supports might make a difference in the accountants' overall assessment of the soundness of the Anadarko books and records. Priority for leases and collection documents is given to agency files as far back as these extend, then to documents at Lee's Summit and Albuquerque. We expect to have to image and code a fairly large sample of these records before selecting the cases to be matched. For disbursements the support sampling concentrates primarily on the period before the Treasury retained all copies of canceled checks. The samples for receipt and disbursement documents are both planned initially to be approximately 200 per agency or 400 in total. Again, as with the sampling from the IIM, additional cases may be included using rigorous probability selection methods, ifa systematic error pattern is detected. 8. Paper Records Era Samplin? In this brief sample design Appendix, we have yet to specifically consider how the transactions not currently available electronically are to be selected for sampling. Analysis efforts have not progressed very far yet on these samples. Broadly, though, we expect the approach to be similar to what has just been described. Anticipating that a full discussion will be needed, a separate sample design report is planned, when we know more. Obviously, before a transaction sample can be selected, and before a sample of supporting documents can be matched to an IIM transaction, the IIM transactions must be located from the paper ledgers and entered into an analyzable electronic database. It is anticipated that this process will begin early in 2003 and will be ongoing as the work on the Electronic Records Era transactions is proceeding. 9. Error Ad[ustment Approaches There are many possible ways of calculating an estimated maximum dollar error for an account (e.g., Guy 1994; Wilburn 1984). The maximum dollar error can be expressed as a percent of the transactions or in dollar terms. Obviously, in the present application, accounts with more transactions have more exposure, ceteris paribus, to the chance for error and this must be taken into consideration. It is premature to make specific error adjustment recommendations at this time. The comparison and evaluation of appropriate estimation procedures will be included in the final analysis. January 6, 2003 D-16 The Historical Accounting Plan for Individual Indian Money Accounts References American Institute of Certifi ed Public Accountants (2001). Audit Sampling, AICPA, New Jersey. Deming, W.E. (1986). Out of the Crisis, MIT Press. Cambridge. Guy, D.M. (1994). ,4 udit Sampling -,4 n Introduction, John Wiley and Sons, New York. Scheuren, F.J. (2001). NORC Statistical and Economic Series Report: Pre-Design Report. Thompson, S. et al. (1996). ,4daptive Sampling, John Wiley & Sons. Wilbum, A.J. (1984). Practical Statistical Sampling for ,4 uditors, Marcel Dekker, New York, Zhang, H. and Schcuren, F. (2002). NORC Sample Design Planning Report Part II- ,4nadarko Agency IIM Transactions Described. January 6, 2003 D-I 7 The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX E INDEXING January 6, 2003 E-I The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX E. INDEXING OTR in Albuquerque, New Mexico, estimates that there are approximately 195,000 boxes or containers of Indian trust records, an estimated 300-500 million pages, held by the United States Government. These records are part of the history of Amer/can Indians, and, in part, contain the records of the management of the I1M Trust Fund. These records vital to conducting a historical accounting directed by the Court in the Cobell v. Norton litigation. Indian trust records are stored in a number of different locations. Working and active records are stored at BIA regional and agency offices around the United States. OTR has several local facilities in Albuquerque that store approximately 51,000 boxes of records. The National Archives and Records Administration operates the Federal Records Center at Lee's Summit, Missouri, which stores approximately 42,000 boxes of Indian trust records, and also has 64,000 containers of records stored at National Archives. The Trust Accounting Division of the General Services Administration has approximately 20,000 boxes of records at its facility in Lanham, Maryland. In addition, the Department of the Treasury and the General Accounting Office have approximately 18,000 boxes of records at storage facilities. Interior must rely on these records to perform the historical accounting. However, the inventories and indexes of these records, essential to locating needed records effectively and efficiently, are not consistently detailed and accurate for this purpose. OTR's approach is to create an electronic database for the records, much like transforming a traditional library catalog card index for books into an electronic reference database. This database will be created by entering data on the contents of each box in the collection onto a data-entry screen that will be integrated into a database for access by the users. Information in the database can then be sorted and indexed for easy access by document types, dates, collections, agency, etc. OTR has engaged the firm Labat-Anderson, Incorporated to index all of the trust-related records that are currently stored in its facilities in Albuquerque and at Lee's Summit. This endeavor will require the contractor to examine every file contained within a box or container of records, and accurately enter key information about the records into a database. The indexing is working at a "file level," meaning that every file title will be entered, verbatim, into the database. Where there are no file folders for the contents in a box, the entire box will be considered the folder and the contractor will indicate such as an "implied folder" in the database. January 6, 2003 E-2 The Historical Accounting Plan for Individual Indian Money Accounts In addition, the contractor will include the following information in the index: · Data Appearing on the Exterior of a Box _ This will allow searchers to identify the physical location of the boxes based on earlier inventory records. Also, where tribal names are identified on the exterior of boxes, that information will be captured in the database · Source Agency-- This will denote the agency or regional office to which the records belong and the originators of the records. · Document Type and Date Range _ This information will enable users who are searching for a specific document with a known or approximate date to identify the appropriate box containing the document without having to request an unnecessarily large number of boxes. OTR's contractor, Labat-Anderson, has started the indexing project and is training two teams to work on the records simultaneously in Albuquerque and at Lee's Summit. OTR and OHTA are coordinating on document needs for the historical accounting, and OTR is prioritizing which groups of records to index first by BIA agency and by tribe. This will ensure that the database index will be populated and available to search for the first records needed for the historical accounting. The indexing project is expected to be completed in FY 2004. January 6, 2003 E-3 The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX F COST ASSUMPTIONS January 6, 2003 F-1 The Historical Accounting Plan for Individual Indian Money Accounts APPENDIX F. COST ASSUMPTIONS The scope and complexity of the historical accounting require an assessment of costs for each of the tasks associated with the accounting work. The objective is to establish the cost of the historical accounting based on significant cost drivers (activities), such as whether transactions are in electronic or paper form, the level of document location, collection, and scanning required, and the number of transactions that must be reconciled. However, there is a high level of uncertainty in the cost estimates for the historical accounting project because many parameters continue to be investigated and could significantly change the estimates developed. Preliminary Estimate of Accounting Costs The total preliminary estimated cost for conducting the historical accounting is $335 million, of which 89 percent of the estimated cost is the accounting work. The following cost estimating assumptions were made to construct the cost for the historical accounting project. Cost per Transaction Electronic Paper Records Records Era Era $ 80 $ 80 - to locate 4 documents (on average) to support the transaction 25 25 - to prepare, image and digitize (scan) the documents (6 pages per document, on average) 125 185 - for accounting about I hour per transaction on average 230- 529____9_0 - total estimated cost per transaction Digitized Ledger $1,000 - to create a digitized ledger from paper ledgers cost including locating documents to support transactions, preparing, imaging, and digitizing the documents (Note: approximately 30 percent of the land-based IIM account holders have Paper Records Era portions of their accounts or approximately 60,000 ledgers) January 6, 2003 F-2 The Historical Accounting Plan for Individual Money Accounts APPENDIX G ACRONYMS January 6,2003 G-1 The Historical Accounting Plan for Individual Money Accounts APPENDIX G. ACRONYMS ACRONYMS BIA Interior's Bureau of Indian Affairs BLM Interior's Bureau of Land Management CD&L Chavarria, Dunne & Lamey LLC IIM Individual Indian Money IRMS Integrated Records Management System LRIS Land Records Information System LTRO BIA Land Title and Records Office MMS Interior's Minerals Management Service NORC National Opinion Research Center of the University of Chicago OHTA Interior's Office of Historical Trust Accounting OST Interior's Office of the Special Trustee for American Indians OTFM Office of the Special Trustee Office of Trust Funds Management OTR Interior's Office of Trust Records QC Quality Control SDA Special Deposit Account TFAS Trust Funds Accounting System January 6, 2003 G-2 CERTIFICATE OF SERVICE I declare under penalty of perjury that, on January 6, 2003 I served the foregoing Notice of Filing by hand upon: Dennis M Gingold, Esq. Keith Harper, Esq. Native American Rights Fund 1712 N Street, N.W. Washington, D.C. 20036-2976 By U.S. Mail upon: Elliott Levitas, Esq. 1100 Peachtree Street, Suite 2800 Atlanta, GA 30309-4530 By Hand the Morning of January 7, 2003 upon: Alan L. Balaran, Esq. Special Master 1717 Pennsylvania Avenue, N.W. 12th Floor Washington, D.C. 20006 (202) 986-8477 By Hand upon: Joseph S. Kieffer, III Special Master Monitor 420 7 th Street, N.W. Apartment 705 Washington, D.C. 20004 (202) 478-1958 Kevin P. Kingston' 7