UNITED STATES DEPARTMENT OF EDUCATION OFFICE OF INSPECTOR GENERAL Audit Services December 15, 2011 MEMORANDUM TO: James Runcie Chief Operating Officer Federal Student Aid FROM: Keith West /s/ Assistant Inspector General for Audit SUBJECT: FINAL CONSULTING REPORT Title IV Additional Servicers Capacity Assessment ED-OIG/S15L0001 Attached is the subject final consulting report prepared by Ernst & Young, LLP, as a result of its assessment of the current status of the Title IV Additional Servicers (TIVAS) to handle the volume of servicing new Direct Loan Program originations and consolidations and the servicing of Federal Family Education Loan Program loan purchases under the Ensuring Continued Access to Student Loans Act of 2008. The assessment was performed on our behalf by Ernst & Young, LLP, under the American Institute of Certified Public Accountants’ Consulting Standards. The scope of the assessment was established in a Statement of Work agreed to by the Office of Inspector General and Ernst & Young, LLP. The report’s Executive Summary contains an overview of the assessment’s scope and procedures performed by Ernst & Young, LLP. The report is intended solely for the information and use of the Office of Inspector General and management of the Department. It is not intended and should not be used by anyone other than these specified parties. In accordance with the Freedom of Information Act (5 U.S.C. § 522), reports issued by the Office of Inspector General are available to members of the press and general public to the extent information contained therein is not subject to exemptions in the Act. We appreciate the cooperation given Ernst & Young, LLP, during this assessment. If you have any questions, please contact Patrick Howard, Deputy Assistant Inspector General for Audit, at (202) 245-6949 or Kenneth Smith, Director, Student Financial Assistance Advisory and Assistance Team, at (202) 245-6968. Attachments The Department of Education's mission is to promote student achievement and preparation for global competitiveness by fostering educational excellence and ensuring equal access. James Runcie Page 2 of 2 Electronic Distribution List: John Hurt, Chief Financial Officer, FSA Patrick Bradfield, Director, Acquisition Group, FSA Marge White, Audit Liaison Officer, FSA Dawn Dawson, Audit Liaison Officer, FSA U.S. Department of Education TIVAS servicing capacity assessment May 2011 Ernst & Young LLP Westpark Corporate Center 8484 Westpark Drive McLean, VA 22102 Tel: +1 703 747 1000 Fax: +1 703 747 0100 www.ey.com May 16, 2011 To the Office of Inspector General and Management of the U.S. Department of Education: We have completed our engagement to assess the current status of the Title IV Additional Servicers to handle the volume of servicing for all new Direct Loan originations, consolidations, and Ensuring Continued Access to Student Loans Act loan purchases. Our engagement was performed in accordance with our contract task order dated September 24, 2010, and our procedures were limited to those described in the Statement of Work dated September 17, 2010 and in the Executive Summary of the attached report. The engagement was performed under the American Institute of Certified Public Accountants (AICPA) Consulting Services: Definitions and Standards (CS Section 100). Our findings and recommendations resulting from our procedures are provided in the attached report. We appreciate the cooperation and assistance provided to us by the management of Federal Student Aid, the Office of Inspector General and each of the four TIVAS servicers during the course of our work. This report is intended solely for the information and use of the Office of Inspector General and management of the Department of Education and is not intended to be and should not be used by anyone other than these specified parties. The nature and scope of our services was determined solely by the agreement between Ernst & Young and the Department of Education. Our work was performed only for the use and benefit of the Department of Education, and others who read this report that were not a party to our agreement with respect to the nature and scope of such services do so at their own risk. The services we performed were advisory in nature. Ernst & Young did not render an assurance report or opinion under our contract with the Department of Education, nor did our services constitute an audit, review, examination, or other form of attestation as those terms are defined by the AICPA. None of the services we provided constituted any legal opinion or advice. We did not conduct a review to detect fraud or illegal acts. Notwithstanding anything to the contrary in the contract, we do not assume any responsibility for any third-party products, programs or services, their performance or compliance with the specifications of the Department of Education or otherwise. Ernst & Young LLP May 16, 2011 Contents ► Executive summary 3 ► Background 3 ► Engagement objectives 3 ► Scope and approach 4 ► Assessment results 5 ► Findings and recommendations 6 ► Supporting information ► Borrower volume estimates 10 ► Allocation methodology 18 ► Servicing capacity and per-borrower estimates 24 ► Contingency planning 31 ► Fulfillment of contract requirements 33 ► Appendix A - FSA responses to findings and recommendations 38 ► Appendix B - FSA original responses to findings and recommendations 45 Page 2 TIVAS servicing capacity review Executive summary Background In June 2009, FSA awarded four new contracts to acquire additional The U.S. Department of Education (Department), through Federal Title IV student loan management servicing under Federal Acquisition Student Aid (FSA), administers programs that are designed to provide Regulations (FAR) Part 12. The new servicers are commonly referred to financial assistance to students enrolled in post-secondary education as TIVAS and began servicing FFEL loans in September 2009. TIVAS institutions as well as collect outstanding student loan balances. In include Sallie Mae (SLM), Nelnet (NN), Great Lakes (GL), and 2008, Congress enacted the Ensuring Continued Access to Student Pennsylvania Higher Education Assistance Agency (PHEAA). Loans Act (ECASLA), which authorized the Department to purchase or enter into forward commitments to purchase certain Federal Family On March 30, 2010, Congress enacted the Student Aid and Fiscal Education Loan (FFEL) loans. The Department implemented three Responsibility Act (SAFRA) – legislation eliminating the origination of activities under this temporary loan purchase authority, which were new FFEL loans. Beginning July 1, 2010, all student loans are effective for the 2008-2009 and 2009-2010 academic years: originated through the William D. Ford Federal Direct Loan (DL) ► Loan Participation Program – the Department purchased a program. As a result, all schools participating in the FFEL program participation interest in FFEL loans, if the loans had been at least transitioned to the DL program. Previously, FSA had one servicing partially disbursed. Loans in the participation facility could be either contractor, ACS, to service its DL portfolio. The ACS contract is (1) sold (PUT) to the Department once fully disbursed by the lender, expected to expire in 2014. As a result, as of September 2010, FSA or (2) bought back by the lender. The program closed on September started to allocate newly originated DL to TIVAS. 30, 2010. ► Loan Purchase Commitment Program – Lenders could PUT fully disbursed FFEL loans to the Department. Once sold to the Engagement objectives Department, they were considered Federally held FFEL loans, and Both laws (ECASLA and SAFRA) resulted in additional servicing were no longer guaranteed loans. The Loan Purchase Commitment volume, which had to be managed by FSA or its contractors. Ernst & program closed on September 30, 2010. Young was engaged by the Office of Inspector General (OIG) to provide ► Asset-Backed Commercial Paper Conduit – the Department is consulting services to assess the current status of the TIVAS servicers considered the buyer of last resort to purchase the FFEL loans if the to handle the volume of servicing for all new DL originations, conduit is unable to refinance commercial paper as it matures. consolidations and ECASLA loan purchases. Ernst & Young was engaged to consider the following throughout its assessment: the The Affiliated Computer Services (ACS) – Commercial application was borrower and loan volume estimation process, servicer allocation used to service certain FFEL loans purchased under ECASLA between methodology, servicing capacity, information technology (IT) and September 2008 and June 2010. Its FFEL volume has been since business capacity planning, servicer contingency plans and compliance transferred to Title IV Additional Servicers (TIVAS). with contract requirements. Page 3 TIVAS servicing capacity review Executive summary Ernst & Young gathered information on current operations and from FSA and TIVAS. Through the collection of information from FSA processes across all TIVAS and provided findings and and TIVAS, Ernst & Young compared information across servicers to recommendations to the OIG to be used as baseline information for its identify gaps between expectations and requirements set by FSA that future work. may affect satisfactory service delivery. The scope of our work did not require testing and validation of data, therefore, information provided to Scope and approach Ernst & Young by the TIVAS and FSA was assessed as provided and Ernst & Young conducted the assessment from February 2011 to April was not subject to additional testing and validation. 2011 in accordance with scope and objectives defined in agreement with OIG. Site visits and meetings were conducted with both FSA and The following are the high level procedures performed for each in-scope each TIVAS. The scope of the assessment was defined as follows: area: FSA • Borrower and loan volume estimates (budgetary estimates) Borrower • Understand drivers, methodologies and • Servicer allocation methodology and loan assumptions for both FSA and TIVAS borrower • Fulfillment of contract requirements volumes and loan volume costing estimates. (both FSA • Understand estimate monitoring frequency and TIVAS • Borrower and loan volume estimates (performed by and TIVAS) process for FSA and TIVAS. each servicer) • Compare point-in-time estimates to actual • Servicing capacity volumes allocated to servicers. Understand • Per-borrower estimates/IT and business capacity factors for differences including various planning programs affecting volume. • Contingency plans and processes Servicer • Understand various legislation, programs, related In order to gain an understanding of the processes in place for volume allocation factors and contract terms affecting allocations. estimates, capacity planning, and the ability of servicers to provide methodology • Understand allocation methodology as it varied adequate service to FSA in light of requirements set forth in the between award years. contracts, Ernst & Young reviewed documentation provided by FSA and • Understand the ongoing allocation methodology TIVAS. Documentation that was reviewed included process flows and used for new DL originations. procedures, relevant system policies, contract requirements and system • Review the drivers influencing the ongoing specifications. In addition, Ernst & Young interviewed process owners allocation methodology: surveys, calculations, frequency and sample sizes. Page 4 TIVAS servicing capacity review Executive summary Fulfillment of • Understand the contract requirements review Servicer • Review servicer contingency plans against contract and tracking process managed by FSA contingency NIST standards (SP 800-34 and 53). requirements (phases two and three only). Understand planning • Understand backup procedures in place and involvement of process owners effected by ability to restore IT systems. each requirement (e.g., users of required • Compare servicer contingency plans. reports, system owners of interfaced • Discuss business operation contingency applications, etc). planning as well as excess volume planning. • Discuss contract requirement implementation with TIVAS to understand process for submitting and obtaining acceptance of Assessment results submitted artifacts (deliverables provided by Ernst & Young assessed the current status of the TIVAS servicers to TIVAS supporting contract requirements). handle the volume of servicing for all new DL originations, • For a sample of contract requirements, obtain consolidations and ECASLA loan purchases. Detailed findings and evidence that artifacts for each TIVAS were recommendations have been provided in the next section, pages 6 submitted, tracked, retained and ultimately through 9. As a result of our assessment, we found that: accepted by FSA. • While FSA has been able to on-board the four TIVAS, FSA should • Review the FSA Contract Monitoring Plan for develop more formal retention and management of documentation activities performed and frequency of related to contract requirements and clarifications in order to allow monitoring. FSA improved oversight of the contract requirements. • FSA has developed a formal methodology to determine the amount Servicing • Document TIVAS servicing systems of borrowers to be allocated to each of the TIVAS every year; capacity and configuration (hardware and software). however, FSA should develop baseline servicing standards and per-borrower • Understand approach for development of the metrics in order to support FSA’s goal of improving services and to estimates/ FSA servicing platform. provide better customer service. capacity • Understand capacity planning process for planning TIVAS. • Currently, no TIVAS is able to support 50 million borrowers, the • Determine current maximum capacity. maximum volume of borrowers for the basic ordering period of five • Obtain servicing capacity estimates given years as required by the contract. However, the TIVAS currently specific hypothetical scenarios. support 13.5 million borrowers collectively and FSA does not anticipate the total amount of borrower volume to grow to 50 million by 2014. FSA should actively monitor the constraints at each TIVAS as it relates to the TIVAS’ ability to service borrowers. Page 5 TIVAS servicing capacity review Findings and recommendations # Area Finding Recommendation 1 Borrower Growth assumptions used by the Department Budget Service FSA should obtain information related to the estimates office in developing borrower volume growth – as influenced by growth assumptions used by the Department the macroeconomic environment – are not formally communicated Budget Service office for deriving borrower [Page 11] to FSA. Knowledge related to applying factors of the volume growth. The drivers and assumptions and macroeconomic environment to borrower volume growth related impact to FSA TIVAS loan allocation estimates is held by individuals involved in the process and is estimates should be documented in order to based on experience in the field rather than a documented set of increase process sustainability for FSA and economic indicators or specific population indicators. increased transparency to stakeholders. 2 Borrower FSA performs its own borrower volume estimates. While TIVAS FSA should work with each TIVAS to understand estimates are not contractually required to prepare independent borrower the specific impact on loan servicing and estimates, FSA expects TIVAS to do so in order to plan their customer satisfaction levels caused by [Page 17] operations. unexpected increases in servicing volumes. Through the normal contract monitoring Due to several unpredictable variables in the first 18 months of communication channels, FSA should use the program, some large discrepancies occurred between TIVAS’ information from its monitoring activities to update estimates of borrowers and actual volumes received from FSA. estimates as the information becomes available. Despite the volume discrepancies, no major impact was noted to These estimates and their revisions should be loan servicing, largely due to the servicers’ monitoring of their communicated on a regular basis to allow TIVAS operations and the scalability of their systems. to manage acceptable customer service levels. While ongoing allocations are expected to be more predictable While FSA may choose to communicate (due to allocation methodology put in place and less activity estimates of the number of borrowers, and around PUT and split borrower transfers), a communication changes to those estimates, it is understood that mechanism has not been defined to relay unexpected changes in these communications do not imply a volume to volumes to the TIVAS in order to prevent disruptions to borrower each servicer and are provided for assistance in servicing. each servicer’s planning process. Page 6 TIVAS servicing capacity review Findings and recommendations # Area Finding Recommendation 3 Capacity Each of the TIVAS indicated that they are not currently able to FSA should work closely with each TIVAS to planning support 50 million borrowers, the maximum volume of borrowers understand their specific constraints and for the basic ordering period of five years as required by the understand the required lead time necessary to [Page 30] contract. However, the TIVAS currently support 13.5 million on-board specific loan and borrower volumes. In borrowers collectively and FSA does not anticipate the total addition, FSA should implement a process to amount of borrower volume to grow to 50 million by 2014, the end actively monitor these constraints at each TIVAS of the basic ordering period. and the appropriateness of the required lead time for on-boarding of loan and borrower volumes. TIVAS stated that they use scalable IT infrastructure to process loans, which would allow them to accommodate loan volume processing in excess of current capacity. While the servicers have indicated the ability to increase their computer processing capacity, their ability to increase staffing and physical space and to deliver training pose constraints in the short term (three to six months). More than six months’ notice would be needed for any single servicer to accommodate 50 million borrowers, given the staffing and training constraints. 4 Contract Contract requirements as defined by FSA were documented at a For future on-boarding of servicers, FSA should requirements very high level. Detailed requirements were not defined by FSA timely communicate the details of system and until after the contract award in June 2009, which provided functionality requirements in order to allow for [Page 36] servicers with a short time frame to understand and implement the sufficient planning, development and testing by initial set of requirements and develop appropriate system design. new servicers prior to the implementation phase. Page 7 TIVAS servicing capacity review Findings and recommendations # Area Finding Recommendation 5 Contract A final set of detailed requirements, which would aid in on- FSA should document the contract requirements requirements boarding a new servicer, was not maintained by FSA. and related clarifications in one location. Contract Modifications that occurred following the initial implementation requirements and clarifications should then be [Page 36] were not tracked by FSA and compiled as part of the set of used by FSA for oversight and enforcement of contract requirements; instead they were only captured within the contract terms and conditions. meeting minutes of status meetings between the TIVAS and FSA. For future implementations of servicers, FSA In addition, while FSA noted that all servicers provided evidence should develop documentation that supports a of their compliance with each requirement clarification, that sustainable on-boarding process. This includes documentation was not available for each requirement. The tracking of all clarifications and/or modifications, source, date and acceptance of each contract requirement and/or formal identification of key stakeholders within clarification was not individually tracked by FSA. both TIVAS and FSA, and documented verification of artifact submissions by those business users responsible for various elements of the program. Such documentation could be used in the future to onboard new servicers, as necessary, without incurring similar issues and clarification procedures as in the past. 6 Contingency Server capacity on one servicer’s backup server is smaller than FSA should implement a process to monitor the planning the production server, which may result in the slower processing contingency planning of each servicer. Monitoring of transactions in the event the backup server was required for should include capabilities of back up servers and [Page 32] processing during an outage of the production server. facilities in the event the servicer is required to Management at the servicer is in the process of upgrading its utilize its non-production environment. backup servers to meet processing capacity of its production servers. Page 8 TIVAS servicing capacity review Findings and recommendations # Area Finding Recommendation 7 Contingency Two of four servicers do not have formally documented processes FSA should evaluate the processes in place at planning or flowcharts that identify detailed steps that should be followed to each TIVAS to support the on-boarding of excess accommodate an unplanned increase in borrower volume borrower volumes in an organized and [Page 32] transferred from FSA. Currently, these steps would be defined on sustainable manner. FSA’s process evaluation an ad hoc basis when such an event arises. Both servicers have should include details on the coordination of procedures in place for monitoring system and staffing capacity. specific stakeholders to be notified within FSA and TIVAS in the event of a TIVAS receiving borrower volume in excess of plan and/or capacity. 8 Allocation FSA has not defined minimum survey ratings or maximum FSA should define minimum and maximum methodology delinquency rates for servicers. With defined minimum customer servicing standards to support the FSA’s goal of satisfaction levels and maximum delinquency rates, servicers will improving services to provide better customer [Page 23] better understand the ultimate goals of FSA. service levels. Servicers should be notified of these standards; a monitoring and probation program should be considered to resolve issues before servicers are below the defined minimum standards. 9 Allocation All metrics used for the ongoing allocation methodology are FSA should weight evaluation metrics for future methodology weighted equally for calculation of the final score for servicers; allocations to give greater emphasis to metrics however, not all appear to be equally important (i.e., number of that reflect the long-term goals of the Department [Pages 21-23] defaulted borrowers and dollar amounts in default appear to be and FSA. Survey sample sizes should be defined more significant than results of surveys). In addition, the sampling to achieve representative samples from each methodology for school survey population does not take into distinct population group, which would take into account number of schools or different types, as such public account the varying needs of each customer schools tend to be over-sampled. population. Page 9 TIVAS servicing capacity review Borrower volume estimates Supporting information Borrower volume estimates Federal Student Aid ► Related to loan servicing, estimates developed by FSA are used for FSA’s budgetary purposes (i.e., estimating the cost per month by servicer for loan servicing), not to determine appropriate servicing capacity levels. ► FSA relies on TIVAS to perform their own borrower estimations and servicing capacity planning. ► FSA prepares estimates of borrower volumes based on loan status (e.g., in-school, in- grace/current repayment, deferment/forbearance or delinquent) to determine the cost of serviced loans. ► The estimation process is performed twice a year (January and June) showing projections over three years ► Estimates are reviewed monthly by FSA CFO, FSA Business Operations (BOPS), and the Department Budget Service office and may be updated if discrepancies are noted that have a significant impact on budgeted amounts. ► FSA utilizes the data provided by the Department Budget Service office, which is based on Common Services for Borrowers (CSB) and National Student Loan Data System (NSLDS) data as well as growth assumptions ► Growth assumptions are developed by the Department Budget Service office, while the monthly distribution assumptions are developed by FSA BOPS. While certain specific growth figures are provided to FSA BOPS, Department Budget Service does not provide formal documentation of its assumptions and drivers to FSA BOPS. [See Finding #1] Page 11 TIVAS servicing capacity review Borrower volume estimates Federal Student Aid ► Summary of the process used to budget for TIVAS servicing payments: 1 The Department Budget Service office prepares a baseline loan volume estimate (total dollar amount of loans, as well as loan borrower count) for DL and FFEL loans based on information from CSB and NSLDS. Using these figures, the Department Budget Service office then applies assumptions on growth. 2 FSA BOPS uses the borrower count from the Department Budget Service and reduces this number by a percentage based on historical information (CSB, NSLDS) to estimate a number of borrowers expected to be new to the system. 3 FSA BOPS distributes new borrower estimate over 12 months using monthly distribution historical trends. 4 FSA BOPS uses current actual volume and adds in new volume estimates as they are expected to enter the system to determine a total borrower volume for each month of the year. 5 FSA BOPS uses historical trends to estimate a percentage of borrowers expected to be in specific types of loan status (in-school, in-grace, current repayment, deferment/forbearance or delinquent) to estimate costs, as cost varies by loan status. Page 12 TIVAS servicing capacity review Borrower volume estimates TIVAS ► TIVAS estimate the number of borrowers who will be new to their system and total number of borrowers being serviced at a point in time. ► TIVAS use their estimates to project an expected volume of loans (based on estimated volume of borrowers) and required processing capacity. ► Loan volume is significant for computer processing and storage due to the data and processing associated with each loan. ► Borrower volume is significant to both staffing levels and servicer revenue. Staffing is significant as one borrower, regardless of the number of loans likely, can be assisted by one individual servicer representative. The borrower volume is significant to revenue because servicers are only paid based on the number of borrowers in their system in a specific status. Servicers are not paid additional amounts if the borrower has multiple loans. ► In the first 13 months (September 2009 – September 2010), loan transfers by FSA to TIVAS were mainly the result of qualified FFEL loans sold (PUT) to the Department by various lenders. Newly originated DL started to be allocated in June 2010. ► PUT volume could be either: ► “On the system” – Each TIVAS performs servicing for various lenders as a part of their commercial business. In addition the TIVAS may be lenders themselves. In the case where one of the TIVAS was the current servicer (either as a lender, or a contracted servicer for a third-party lender) of the loan being PUT – loans were kept with the same servicer to limit disruption to the borrower. For these loans, the TIVAS would migrate the loans from its commercial servicing systems to its federal servicing system (an internal transfer). ► “Off the system” – loans being PUT that had not been previously serviced by one of the TIVAS. NOTE: “System” refers to the servicing system or application used by TIVAS. Page 13 TIVAS servicing capacity review Borrower volume estimates TIVAS – FFEL PUTs ► Servicers used different methods to estimate FFEL PUT volume based on the knowledge and relationships they may have had with lenders on their commercial system. ► TIVAS A ► TIVAS B ► TIVAS C ► TIVAS D Page 14 TIVAS servicing capacity review Borrower volume estimates TIVAS estimation process summary TIVAS A TIVAS B TIVAS C TIVAS D Estimate drivers Timing/ frequency Monitoring Page 15 TIVAS servicing capacity review Borrower volume estimates Estimates vs. actuals FSA/TIVAS October 2010 November 2010 December 2010 * TIVAS C FSA estimate 12,277,492 12,539,333 12,674,665 FSA actual 13,002,494 13,265,127 13,378,910 Percentage difference 5.91% 5.79% 5.56% ** TIVAS D TIVAS A estimate TIVAS A actual Percentage difference TIVAS B estimate*** TIVAS B actual Percentage difference TIVAS C estimate TIVAS C actual *** TIVAS B Percentage difference TIVAS D estimate TIVAS D actual Percentage difference ► 2010 Q1-Q3 estimates had a number of variables due to the type of allocations made during that period. Therefore, the table compares only 2010 Q4 estimates to actuals. ► The above estimates were made in late 2009, typically December 2009. Estimates are for the borrower volume expected for October – December 2010 by servicer, or in total for FSA. Page 16 TIVAS servicing capacity review Borrower volume estimates Estimates vs. actuals ► Factors to consider when evaluating accuracy of borrower volume estimates performed by FSA and TIVAS: ► Allocations in Year 1 (Sep. 1, 2009 – Aug. 14, 2010) included mostly FFEL and delinquent conduit transfers from lenders, based on existing relationships between servicers and sellers/ lenders. TIVAS only knew the volume of FFEL loans/borrowers on their systems (“PUT” by lenders serviced on their commercial system). ► Transfers of split borrowers between servicers could not be practically estimated by the TIVAS, as it would be difficult to identify a borrower on two separate systems while determining which of the servicers with a split would ultimately end up servicing the borrower. FSA determined that in the case of split borrowers, the servicer with the lower allocation of borrowers at the time the split was addressed would ultimately service that borrower. ► DL allocations were not considered for 2010 estimates. SAFRA legislation was not passed until March 2010, and most initial annual estimates were completed at the end of 2009. ► While some large differences between estimates and actuals for individual servicers are observed estimates for all servicers in total are very close to FSA’s estimates [See Finding #2] ► The servicing capacity review shows that servicers are able to accommodate large increases in volumes given sufficient notice and lead time. Page 17 TIVAS servicing capacity review Allocation methodology Supporting information Allocation methodology Transfer types ► Allocation of borrowers to TIVAS addresses various transfer types: ► Purchases and Participation (PUTs) – Fully disbursed FFEL loans purchased by the Department as part of the “loan participation program” or the “loan purchase commitment program” from various lenders. This transfer is applicable only between 9/2009 – 10/2010. ► Conduit delinquent – FFEL loans that were purchased by the Department, as a buyer of last resort, when the loan becomes delinquent and is removed from the conduit. (Note: Conduit is also subject to other PUT events until 1/2014) ► FFEL transfers from ACS-FFEL – FFEL loans serviced by ACS (Long Beach, CA). This transfer is mainly applicable for year one. Very low volume in year two. ► DL transfers from ACS-DL – DL serviced by ACS (Germantown, MD) for “split borrowers.” This transfer is mainly applicable for year two onward. ► COD originations – New originations of DL. This transfer is mainly applicable for year two onward. Page 19 TIVAS servicing capacity review Allocation methodology Allocation types Transfer type Year one – September 1, 2009-August 14, 2010 Year two – August 15, 2010 – August 14, 2011 Allocation of borrowers based on calculated allocation Overall allocation Conclude year with equal borrowers for each TIVAS percentages. See slide 22 for details on allocation year goal calculation Based on TIVAS existing relationship with the seller, Same as year one - Sales closed after 9/30/2010 are Purchases and current workload and expected count at year-end; if no allocated and counted in year two (earlier sales are participation relationship – assigned to TIVAS with lowest number considered year one sales). of borrowers Based on TIVAS existing relationship with the seller, Each conduit seller/servicer was assigned a TIVAS Conduit current workload and expected count at year-end; if no servicer to handle all of that seller/servicer's sales delinquent relationship – assigned to TIVAS with lowest current during year one. That relationship was retained in year volume two. Based on TIVAS existing relationship with the seller; if no relationship – assigned to TIVAS A, TIVAS B or FFEL transfers All transfers to a single TIVAS (TIVAS A) due to very TIVAS C based on the servicer with the lowest number from ACS-FFEL low volume (anticipated < 15,000 borrower). of borrowers (TIVAS D – none due to large PUT transfers) Only borrowers with loans already at one of the TIVAS DL transfers from are transferred. These transfers are executed to Not applicable in year one. ACS-DL resolve “split borrowers” and move borrower's loans to a single servicer. If borrower is an existing borrower – loan assigned to servicer with existing borrower relationship; if a new Same as year one; however, FSA adjusted LDE borrower – Loan Distribution Engine (LDE) assigns percentages with the goal of allocating a specific COD originations based on percentages entered by FSA, which were percentage (of the total borrower volume) to each adjusted periodically with the goal of making an even servicer based on their allocation methodology score. distribution among TIVAS Note: Starting August 15, 2011, the ongoing allocation methodology will be used for allocation. See slides 21 and 22 for more details. Page 20 TIVAS servicing capacity review Allocation methodology “Ongoing” allocation ► Contracts were designed to incentivize TIVAS to obtain higher satisfaction rates and lower default rates. ► “Ongoing” contracted allocation method ► Only new DL, FFEL purchase/participation borrowers (allocated after August 15, 2010 or “year two”) and conduit delinquent loans are counted for “ongoing” allocation distribution1). ► Allocation is based on five factors: ► Default2 dollar amount (as a percentage of total dollars for each servicer) – Encouraged TIVAS to maintain a low value for borrowers in default. ► Default borrower rate (as a percentage of total borrowers for each servicer) – Incentivizes TIVAS to not only focus on large dollar amount defaults but to aid all borrowers. ► School survey – Measures school satisfaction when interacting with TIVAS. ► Borrower survey – Measures borrower satisfaction when interacting with TIVAS. ► FSA representative survey – Measures FSA satisfaction when interacting with TIVAS. ► TIVAS are ranked on each of the five factors equally (i.e., 20% each) [See Finding #9] 1: Historical information is required to rank servicers (e.g., survey results). As a result, year one did not utilize the standard allocation methodology, so allocation for each loan type varied (see slide 20). 2: Borrowers in default are defined as loans that have been sent to Debt Management and Collection System (DMCS) or are greater than 360 days delinquent at the end of the quarter. Page 21 TIVAS servicing capacity review Allocation methodology “Ongoing” allocation Metric1 Description Factors considered Value of Percentage of “In Repayment” portfolio • Principal balance outstanding and interest of loans transferred to DMCS or portfolio dollars that go into default more than 360 days delinquent during the period considered in (percent in public schools, private • Principal balance outstanding and interest of all loans in “repayment” default schools, proprietary schools) status Unique Percentage of unique “In Repayment” • Number of delinquent borrowers transferred to DMCS or more than 360 borrowers portfolio borrowers that go into default days during the period considered in (percent in public schools, private • Number of all borrowers in “repayment” status default schools, proprietary schools) Surveys – Measure of borrower satisfaction with • Phone surveys for borrowers are conducted by the CFI Group, an borrowers servicer (0-100%) independent survey company, across all borrowers (target is 900-1000 (In school, in grace, and in repayment borrower surveys for each servicer per year) borrowers) • Phone surveys for schools conducted the by CFI Group for each school type (target is 300 school surveys for each servicer per year) Surveys – Measure of school satisfaction with • Online FSA surveys are administered by the CFI Group and sent to all schools servicer (0-100%) personnel interacting with TIVAS (there are about 90 FSA employees (public schools, private schools, surveyed, with an average response rate of 42%) proprietary schools) • All surveys are designed to prompt respondents to think about the type of Surveys – FSA Measure of FSA satisfaction with servicer service they receive by asking specific questions on individual personnel (0-100%) occurrences; however, only three questions are used for ranking.2 These questions provide a standardized system for rating customer satisfaction in various industries. 1 Each metric is measured quarterly. Quarterly scores are averaged as of July 1 of each year to provide a year end score. 2 American Customer Satisfaction Index (ACSI) Survey Ranking Questions: ► Using a 10-point scale on which “1” means “very dissatisfied” and 10 means “very satisfied,” how satisfied are you with [servicer]? ► Using a 10-point scale on which "1" now means "falls short of your expectations" and "10" means "exceeds your expectations," to what extent has [servicer] fallen short of or exceeded your expectations? ► Imagine what an ideal process would be for dealing with your loan servicer. How well do you think [servicer’s] current process compares with that ideal you just imagined? Please use a 10-point scale on which "1" means "not very close to the ideal," and "10" means "very close to the ideal." Page 22 TIVAS servicing capacity review Allocation methodology “Ongoing” allocation calculation Year two allocation metrics Metric TIVAS 1 TIVAS 2 TIVAS 3 TIVAS 4 Default borrower rate (points awarded) .46% (2) .55% (1) .27% (3) .25% (4) Default amount rate .19% (2) .25% (1) .14% (3) .12% (4) Borrower survey 69.44 (4) 65.67 (1) 67.11 (2) 68.78 (3) School survey 74.78 (2) 75.78 (3) 79.00 (4) 74.67 (1) Federal personnel survey 65.00 (1) 69.33 (2) 73.33 (4) 71.67 (3) Total points awarded 11 8 16 15 Allocation percentage 22% 16% 32% 30% ► Points are awarded based on rank as related to other servicers. Allocation percentage is determined by dividing the servicer’s total points by the total available points (50). ► Lowest allocation percentage (assuming one point in all metrics) is 10% ► Highest allocation percentage (assuming four points in all metrics) is 40% ► FSA does not define a minimum score in any category or in total to continue servicing. ► Servicers will only be discontinued if they do not comply with contract requirements; customer satisfaction and default rate are not included as cause for discontinuing contracts [See Finding #8] Page 23 TIVAS servicing capacity review Servicing capacity and per-borrower estimates Supporting information Servicing capacity and per-borrower estimates Background ► We discussed with each servicer the capacity constraints and scalability of each servicer’s operations. ► We obtained documentation from each servicer and held meetings to become familiar with the IT infrastructure and systems used by the servicers. In addition, we discussed the process to develop the systems used for FSA portfolio processing. ► We discussed with each servicer the business planning and monitoring process as it relates to IT infrastructure and operational factors (personnel, physical space, etc.). ► We obtained responses from each servicer on hypothetical scenarios for accommodating additional volume at defined intervals to determine the key factors and considerations each servicer makes when addressing excess volume. Page 25 TIVAS servicing capacity review Servicing capacity and per-borrower estimates IT infrastructure and systems – current state TIVAS TIVAS TIVAS TIVAS A B C D Servicing system Hardware CPU in use, remaining capacity on demand Average CPU utilization Database type Current disk utilization Average CPU utilization refers to the percentage of the processor in use, the excess capacity is not used and remains idle. Additionally, Current disk utilization refers to the percentage of disk space used, the remaining percentage is idle and is not in use for any other purposes. Note: Individual components of servicing systems are presented above; however, the adequacy of each servicing system depends on the entire system which includes the people, process and technology driving the servicers operations. Page 26 TIVAS servicing capacity review Servicing capacity and per-borrower estimates Systems capacity summary ► Each mainframe system may host multiple services (other than FSA loan processing). Utilization figures noted relay the total utilization for each processor or disk array inclusive of processing for FSA. ► Each servicer currently has excess disk space and processing capacity, therefore, each servicer is able to handle current loan volumes as well as growth projected by each servicer. ► The nature of each servicer’s IT environment appears to be scalable to accept greater servicing volume with appropriate lead time. ► ► ► Data storage for all servicers is , which appears to allow enough head room for additional volume. ► The nature of the systems also allows for near overnight data capacity expansion. In an emergency situation, it appears the systems could scale within a matter of days. Page 27 TIVAS servicing capacity review Servicing capacity and per-borrower estimates Systems development summary ► Each servicer leveraged existing commercial systems to meet FSA requirements for servicing. ► ► ► ► FSA data appeared to be segregated for all servicers either using separate databases for each portfolio (FSA vs. commercial) or through controls that allow access only to cleared individuals. ► Modifications made to TIVAS systems were designed to follow a change management process that included systems testing, user testing, and business/technical approval for modifications. Note: Ernst & Young obtained the above information through review of servicer reports and discussions with servicers. Ernst & Young did not independently verify the change management or logical access processes. Page 28 TIVAS servicing capacity review Servicing capacity and per-borrower estimates Capacity scenarios ► Three hypothetical scenarios were proposed to each TIVAS to allow for comparison of their ability to increase FSA servicing capacity. We proposed scenarios in which the servicer was notified on December 31, 2010, of an increase in volume to occur on one of these three dates: ► January 1, 2011 – On this day, with no system upgrades or increased capacity on demand, what volume of borrowers could be serviced by the current servicing platform with no change in full-time employee (FTE) count? This scenario is based on strictly servicing system capacity. ► March 1, 2011 – Given 60 days notice, what volume of borrowers could be serviced? This scenario takes into account system capacity as well as need to onboard FTEs. ► June 30, 2011 – Given six months’ notice, what volume of borrowers could be serviced? This scenario takes into account system capacity as well as need to onboard FTEs. ► Main assumptions related to scenarios: ► January 1, 2011 scenario: ► No additional staffing needs were taken into account, resulting in possible decrease in customer satisfaction level. ► March 1, 2011 and June 30, 2011 scenarios: ► Servicers were required to keep an equal ratio of borrowers to FTE in each department serving borrowers (inbound and outbound call center, and back office operations affected by number of borrowers served), as compared to their ratios as of December 31, 2010. Servicers took into account their ability to acquire staff and physical space, as well as training. ► Servicers were required to assume an equal ratio of loans “in repayment” to “not in repayment” when compared to their current, December 31, 2010, portfolio. ► Servicing system upgrades were allowed that could be completed prior to the specified date. Page 29 TIVAS servicing capacity review Servicing capacity and per-borrower estimates Capacity scenario results ► Summary results of the Ernst & Young proposed scenarios for each TIVAS: TIVAS A TIVAS B TIVAS C TIVAS D Borrowers Current Total loans 12/31/2010 Inbound/outbound call center FTE [See Finding #3] Support FTE Total FTE Borrowers One day Total loans 2 1/1/2011 FTE (no change from 12/31/10) Increase in borrowers over current volume Borrowers 60 days Total loans 3/1/2011 Total FTE Increase in borrowers over current volume Borrowers in repayment Six months Total loans 6/30/2011 Total FTE Increase in borrowers over current volume ► Without set FTE ratio constraints, all servicers are able to scale their IT system to meet larger volume increases than stated in the table above. 2: Increases in capacity were constrained by: inability to upgrade (one-day), or inability to onboard and train adequate number of employees (60 days and six months). Page 30 TIVAS servicing capacity review Contingency planning Supporting information TIVAS contingency planning TIVAS A TIVAS B TIVAS C TIVAS D NIST 800-34 compliance Yes Yes Yes Yes Business impact analysis Yes Yes Yes Yes Recovery testing Test meets required 72- Yes Yes Yes Yes hour recovery objective Backup method Primary data center site Recovery site Additional notes [See Finding #6] Note: Ernst & Young did not perform testing of contingency planning performance. Contingency plans and servicer-provided test summaries were used to prepare the information above. Ernst & Young reviewed contingency plans to determine if it appeared the servicers had select components of a NIST 800-34 contingency plan. [See Finding #7] Page 32 TIVAS servicing capacity review Fulfillment of contract requirements Supporting information Contract requirement fulfillment Q1 2009 ► Timeline showing the key dates TIVAS submitting proposals begin systems development associated with the TIVAS contracts award and fulfillment June 2009 of contract terms. Contracts awarded August 31, 2009 Initial requirements due (phase 1) March 31, 2010 NOTE: Per the contract requirements, each Intermediate requirements target due date (phase 2) TIVAS is required to be able to service up to 50 million borrowers over the base period of the contract (five years). Based on existing staffing, April 30, 2010 each TIVAS is not currently able to support 50 Actual phase 2 completion million borrowers without sacrificing customer service level. However, in December 2010, the TIVAS supported 13.5 million borrowers August 31, 2010 collectively (with another 12 million serviced by Full requirements target due date(phase 3) ACS) and FSA does not anticipate the total amount of borrower volume to grow to 50 million by 2014, the end of the basic ordering November 30, 2010 period. Actual phase 3 completion Page 34 TIVAS servicing capacity review Contract requirement fulfillment Requirement development ► Phase two and phase three contract requirements were comprised of approximately 110 requirements related to general/legal, financial reporting, internal controls, reconciliation, other reporting, security (physical, systems, etc.), records management and DL. ► Requirements were documented in attachments to the original TIVAS contracts and were ultimately refined with detailed clarifications. ► Servicers were required to implement the terms of the contract by March 31, 2010 (phase two) and August 31, 2010 (phase three). Actual implementation lagged to April 30, 2010 and November 30, 2010, respectively. ► Most original requirements had clarifications or further details added, therefore, clarifications became de facto requirements. ► FSA provided all clarifications through a single implementation team to TIVAS, and met with all TIVAS regularly on each requirement area to determine if additional information was required. Page 35 TIVAS servicing capacity review Contract requirement fulfillment Requirement implementation ► The FSA implementation team served as liaisons to distribute clarifications as well as obtain artifacts validating that requirements were met. ► The FSA implementation team often validated that submissions met expected standards; for certain requirements, artifacts were validated by appropriate stakeholders. ► Documentation identifying requirements stakeholders (e.g. report owner, system interface owner, data set owner) was not maintained by FSA. Additionally, sign-off and user acceptance of artifacts submitted by TIVAS to FSA were not formally documented and maintained by the FSA implementation team. ► Contracts were awarded in June of 2009, and TIVAS were required to start initial servicing in September 2009. Therefore in order to begin servicing and meet all requirements by September 2009, some of the system development had to be initiated prior to contract award. This resulted in the following: ► Detailed contract requirements (i.e., clarifications) were not available timely for TIVAS, therefore, all servicers had to work under accelerated systems development timetables. [See Finding #4] ► Clarifications were not ultimately consolidated to create a list of future requirements for potential new servicer onboarding. [See Finding #5] Page 36 TIVAS servicing capacity review Contract requirement fulfillment Review of artifacts submitted ► The sample selection for contract requirements was performed as follows: ► A haphazard sample of 10 detailed contract requirements were selected. ► Selection focused on items that required artifact submission rather than confirmations and/or acknowledgements provided by the servicer in response to a requirement (e.g., legal requirements). The population (approximately 361) was from the full set of phase two and three contract clarifications. ► For each selected requirement we requested supporting evidence from FSA for each servicer’s fulfillment of that requirement. ► FSA did not maintain a detailed tracking document noting the dates of submission, where data was submitted from, and who validated the data. ► Evidence of stakeholder (e.g., report owner, system interface owner, data set owner) acceptance of artifacts was not always maintained by the FSA implementation team, only summary level acceptance from the implementation coordinator was maintained. ► Evidence of artifacts for some requirements was not consistently maintained by FSA. ► For one of the ten requirements selected, no artifacts were provided supporting the requirement for any servicer. ► Requirement 631: Requirement for servicers to report each collection activity (e.g., Collection of Principal, Collection of Interest, etc.) using unique identifiers in transaction level data. ► For one requirement, artifacts were not maintained for a single servicer. ► Requirement 697: Requirement for servicers to report on Teacher Education Assistance for College and Higher Education (TEACH) Grants that have been converted to Direct Unsubsidized TEACH Loans separately from other DL. Page 37 TIVAS servicing capacity review Appendix A FSA responses to findings and recommendations FSA responses to findings and recommendations # Area Recommendation FSA response EY/OIG response 1 Borrower FSA should obtain information related FSA concurred with the recommendation. None. estimates to the growth assumptions used by the Department Budget Service office for deriving borrower volume growth. The drivers and assumptions and related impact to FSA TIVAS loan allocation estimates should be documented in order to increase process sustainability for FSA and increased transparency to stakeholders. 2 Borrower FSA should work with each TIVAS to FSA did not specifically disagree with the Based on discussions with officials at estimates understand the specific impact on loan recommendation to work with each FSA, estimates of borrower volumes servicing and customer satisfaction TIVAS to understand the impact caused were communicated to TIVAS. levels caused by unexpected increases by unexpected increases in servicing However, based upon discussions in servicing volumes. Through the volumes. FSA noted that it recognizes with officials at the TIVAS, we found normal contract monitoring the importance of working closely with the that not all TIVAS could identify such communication channels, FSA should TIVAS and has established multiple communication occurring. As a use information from its monitoring channels of communicating with the result, FSA may not have received activities to update estimates as the TIVAS to share information and concerns adequate feedback from the TIVAS information becomes available. These related to ongoing or potential issues, into the impact that unexpected estimates and their revisions should be including unexpected volume growth. increases in borrower volumes could communicated on a regular basis to FSA concurred with the balance of the have on each TIVAS’ loan servicing allow TIVAS to manage acceptable recommendations. capability and customer service level. customer service levels. While the finding has been revised to clarify this information, the recommendation remains unchanged. Page 39 TIVAS servicing capacity review FSA responses to findings and recommendations # Area Recommendation FSA response EY/OIG response 3 Capacity FSA should work closely with each FSA believes that under the contracts While each TIVAS is responsible for planning TIVAS to understand their specific each TIVAS is responsible for ensuring ensuring adequate servicing capacity constraints and understand the adequate capacity planning. If an pursuant to the terms of the contract, required lead time necessary to on- individual TIVAS fails to have adequate FSA should proactively work with and board specific loan and borrower capacity, the Department can shift the monitor each TIVAS to ensure that no volumes. In addition, FSA should loan volume to other servicers. In capacity issues arise and to avoid any implement a process to actively addition, no concerns have been raised disruptions in the proper servicing of monitor these constraints at each regarding the one-month lead time borrowers’ loans. For example, if a TIVAS and the appropriateness of contained in the contracts. FSA believes TIVAS receives a loan that it is the required lead time for on- that the TIVAS can quickly increase unprepared to service, a borrower’s boarding of loan and borrower servicing capacity. first payment date may be delayed. volumes. 4 Contract For future onboarding of servicers, FSA concurred with the recommendation. None. requirements FSA should timely communicate the details of system and functionality requirements in order to allow for sufficient planning, development and testing by new servicers prior to the implementation phase. Page 40 TIVAS servicing capacity review FSA responses to findings and recommendations # Area Recommendation FSA response EY/OIG response 5 Contract FSA should document the contract FSA concurred with the None. requirements requirements and related clarifications in recommendation to document the one location. Contract requirements and contract requirements and related clarifications should then be used by FSA clarifications in one location. In for oversight and enforcement of contract addition, FSA noted that it has terms and conditions. developed a sustainable onboarding process to be used in the onboarding For future implementations of servicers, of not-for-profit loan servicers. FSA should develop documentation that supports a sustainable onboarding process. This includes tracking of all clarifications and/or modifications, formal identification of key stakeholders within both TIVAS and FSA, and documented verification of artifact submissions by those business users responsible for various elements of the program. Such documentation could be used in the future to onboard new servicers, as necessary, without incurring similar issues and clarification procedures as in the past. Page 41 TIVAS servicing capacity review FSA responses to findings and recommendations # Area Recommendation FSA response EY/OIG response 6 Contingency FSA should implement a process to FSA noted that each TIVAS was While each TIVAS is responsible for planning monitor the contingency planning of required to meet back-up and ensuring adequate contingency each servicer. Monitoring should disaster recovery requirements, and planning pursuant to the terms of the include capabilities of back-up servers that each TIVAS’ contingency plan contract, FSA should proactively work and facilities in the event the servicer is was reviewed and approved by FSA. with each TIVAS to ensure that required to utilize its non-production The contracts were designed to adequate plans and capabilities are in environment. eliminate the need for excessive place or provided for in order to avoid monitoring, instead relying on any disruptions in the proper servicing competition across vendors to of borrowers’ loans. FSA should not ensure performance. If a TIVAS is solely or predominately rely upon unable to perform as required, FSA competition among the servicers as a can transfer the borrower volume to means of ensuring the adequate another servicer. performance of each TIVAS. For example, improper contingency planning could result in the loss of data concerning a borrower’s payment history; the loss of such data could make the loan unenforceable. 7 Contingency FSA should evaluate the processes in FSA believes that the TIVAS While each TIVAS is responsible for planning place at each TIVAS to support the on- maintain both the incentive and ensuring adequate servicing capacity boarding of excess borrower volumes in capability to ensure that increases in pursuant to the terms of the contract, an organized and sustainable manner. borrower volume can be serviced in FSA should proactively work with FSA’s process evaluation should accordance with the contract terms. each TIVAS to ensure that adequate include details on the coordination of plans are in place to address specific stakeholders to be notified unexpected volume increases so as within FSA and TIVAS in the event of a to avoid any disruptions in the proper TIVAS receiving borrower volume in servicing of borrowers’ loans. excess of plan and/or capacity. Page 42 TIVAS servicing capacity review FSA responses to findings and recommendations # Area Recommendation FSA response EY/OIG response 8 Allocation FSA should define minimum and FSA disagreed with this As noted in the report, an individual methodology maximum servicing standards to recommendation. The performance- TIVAS could rate poorly in all metrics support the FSA’s goal of improving based contracts established a relative to the other TIVAS and still services to provide better customer performance structure to motivate each receive a borrower allocation of 10% service levels. Servicers should be TIVAS to improve customer service, under the contract structure. The notified of these standards; a rather than meet a specific minimum contracts do not contain specific monitoring and probation program performance level. FSA is satisfied provisions to mitigate and address a should be considered to resolve with the TIVAS’ overall performance contractor’s poor customer service issues before servicers are below the and its reduced monitoring and default prevention. defined minimum standards. responsibilities and costs. Page 43 TIVAS servicing capacity review FSA responses to findings and recommendations # Area Recommendation FSA response EY/OIG response 9 Allocation FSA should weight evaluation metrics FSA disagreed with the Under the current weighting of methodology for future allocations to give greater recommendation to weight evaluation evaluation metrics, 60% of total points emphasis to metrics that reflect the metrics. FSA believes that the are based on customer surveys (that long-term goals of the Department contracts reflect FSA’s Strategic Goals is, borrower, school, and FSA and FSA. Survey sample sizes to (1) provide superior service and personnel surveys are weighted at should be defined to achieve information to students and borrowers, 20% each) and 40% of total points are representative samples from each (2) provide efficient processes and based on default measurements (that distinct population group, which would effective capabilities that are among is, dollars in default and borrowers in take into account the varying needs of the best in the private and public default are weighted at 20% each). each customer population. sectors, and (3) ensure program As such, the allocation methodology integrity and safeguard taxpayers’ under weights the input of the ultimate interest. In regards to the customer, that is the borrower, and recommendation concerning sample the negative financial impact on the sizes, the sample sizes used provide a Department associated with defaulted statistically valid representation of the loans. overall population of customers. FSA did not explain why all the evaluation metrics are of equal importance. FSA will consider changes in the sampling methodology to obtain the most reliable information available given constraints related to costs, resources and customer response rates. Page 44 TIVAS servicing capacity review Appendix B FSA original responses to findings and recommendations
S15L0001 - Title IV Additional Servicers Capacity Assessment. - Date Issued: 12/15/2011 PDF (3.99M)
Published by the Department of Education, Office of Inspector General on 2011-12-15.
Below is a raw (and likely hideous) rendition of the original report. (PDF)