Moody s Proposes to Update Its Approach to Rating Securities Backed by FFELP Student Loans

Transcription

1 JANUARY 18, 2012 ASSET-BACKED SECURITIES REQUEST FOR COMMENT Moody s Proposes to Update Its Approach to Rating Securities Backed by FFELP Student Loans Analyst Contacts: SAN FRANCISCO Stephanie Fustar Assistant Vice President - Analyst [email protected] NEW YORK Barbara Lambotte Vice President Senior Credit Officer [email protected] Tracy Rice Assistant Vice President Analyst [email protected] Nicky Dang Assistant Vice President Analyst [email protected] William Black Managing Director [email protected] * Andrew Silver also contributed to this report as a research consultant I. Introduction In this Request for Comment (RFC), we describe our proposed approach to rating securities backed by student loans made under the Federal Family Education Loan Program (FFELP) issued in the United States. The proposed analytical framework detailed in the body of this report does not substantially modify the current approach but is intended to provide a single comprehensive overview of our approach. The proposed cash flow assumptions detailed in the Appendix of this report contain updates and additions to previously published cash flow assumptions, as described in the next section. If we implement the proposed changes, we expect to downgrade the ratings of approximately $20 billion of outstanding FFELP student loan-backed notes, which constitute 10% of the universe of FFELP securitizations that we rate. Some Aaa ratings on senior notes may fall as far as Baa3, and some ratings on subordinate notes will fall as far as B. Higher default rate assumptions drive the downgrade of the ratings on approximately $10 billion of outstanding FFELP student loan-backed notes. The addition of a high interest rate scenario lowers the ratings of an additional $10 billion. Exhibit 1 below shows the detailed impact of the proposed methodology on outstanding ratings. EXHIBIT 1 Impact of the Proposed Methodology on Outstanding Ratings Reason for Rating Change Approximate # of Affected Trusts Approx. O/S Balance of Affected Trusts ($ in Billions) Estimated Rating Migration Increase in Net Losses 15 $10.00 Senior: 1-3 notches; Subordinate: 4-9 notches Implementation of 9% Tbill Scenario 20 $10.00 Senior: 1-9 notches; Subordinate: 1-9 notches; Mezzanine: 1-3 notches A detailed description of the rationale behind certain key proposed changes is available in the following report Companion Report to FFELP U.S. Student Loan-Backed Securitizations Request for Comments. Moody s invites participants in the structured finance market to review and comment on our proposed assumptions. Comments are welcome and should be sent to [email protected] by February 20, The final report is scheduled to be released publicly in March This report was revised on February 24, 2012 to add more detail to the non-consolidation loan default assumptions at the end of Appendix III and to correct a typographical error in Appendix II.D.

2 Contents I. INTRODUCTION 1 II. SCOPE 4 III. EXECUTIVE SUMMARY 5 IV. CONSIDERATIONS IN DEVELOPING AND APPLYING STRESS TESTS 8 V. DETERMINING THE MODEL SUGGESTED OUTPUT BASED ON THE STRESS TESTS 8 VI. ANALYZING CREDIT RISK 9 A. Basis Risk 9 B. Overall Level of Interest Rates 10 C. Prepayment Risk 11 D. Lifetime Net Loss Rates 11 E. Timing and Application of Defaults 12 F. Deferment and Forbearance Rates 12 G. Borrower Benefit Utilization Rates 12 H. Servicing Fee Inflation 13 I. Investment Rate on Cash Accounts 13 J. Remarketing or Rate Reset Risk 13 K. Note Amortization 14 L. Pre-funding or Recycling 14 M. Rehabilitated FFELP Loans 14 N. Adjustments for Seasoned Pools Cumulative Default Rates Default Timing Curves Voluntary Prepayment Rates On-Time Borrower Benefits Utilization Rates 15 VII. ANALYZING LIQUIDITY RISK 16 VIII. OPERATIONAL RISKS 17 A. Operational Risks 17 B. Servicing and Administration Fees 19 IX. LEGAL RISKS 19 A. Bankruptcy of the Sponsor Risk of Sponsor Bankruptcy Differs By Type of Sponsor Evaluating Bankruptcy Risk for Sponsors Who Can Become Bankrupt 21 B. Bankruptcy of the Securitization Vehicle 22 C. Legal Opinions 23 X. PROVISIONS OF THE OPERATIVE DOCUMENTS 23 A. Representations and Warranties and Covenants 23 B. Identifying and Enforcing Breaches of Representations and Warranties 24 C. True Sale Transfer of Loans to the Issuer 24 D. Servicing and Indenture Trustee Standard of Care 24 E. Servicer and Custodial Arrangements 25 F. Eligible Accounts and Investments 25 G. Events of Default 25 H. Amendments 26 I. Provision of Surveillance Information 26 XI. FFELP STUDENT LOAN ABS SURVEILLENCE 26 A. Moody s Surveillance Process 26 APPENDIX I EXPECTED CASE AND STRESS CASE CASH FLOW SCENARIOS 27 APPENDIX II - CASH FLOW ASSUMPTIONS 34 A. Basis Risk 34 B. Interest Rate Level Scenarios 36 C. Cumulative Default Rates 36 D. Voluntary Prepayment Rates 39 E. Net Claim Reject Rates 40 F. Timing of Defaults 41 2 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

3 G. Payment Lags 41 H. Deferment and Forbearance Rates 42 I. Borrower Benefit Utilization Rates 42 J. Servicing Fee Inflation 43 K. Investment Rate on Cash Accounts 43 L. Remarketing or Reset Rate Risk Interest Rate for Variable Rate Demand Bonds (VRDBs) 43 Interest Rate for Auction Rate Securities 43 Interest Rate for Reset Rate Notes 44 M. Adjustments for Seasoned Pools Expected Case Cumulative Default Rates for Seasoned Pools Expected Case Default Timing Curves for Seasoned Pools Non-Consolidation Loan Voluntary Prepayment Rates for Seasoned Pools Consolidation Loan Voluntary Prepayment Rates for Seasoned Pools On-Time Borrower Benefit Utilization Rates for Seasoned Pools 46 APPENDIX III CUMULATIVE DEFAULT RATE MODEL 47 MOODY S RELATED RESEARCH 50 3 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

4 II. Scope This methodology report covers our approach to rating securities backed by student loans made under the Federal Family Education Loan Program (FFELP) issued in the United States. FFELP loans are guaranteed by the U.S. Department of Education (DOE) for at least 97% of defaulted principal and accrued interest. 1 The Health Care and Education Reconciliation Act, enacted March 2010, discontinued FFELP lending as of June 30, 2010 and directed all new government-guaranteed student loans to be funded through the William D. Ford Federal Direct Student Loan Program. Consequently, this methodology only applies to transactions backed by FFELP loans originated prior to June 30, This report combines and updates the following prior methodologies and special reports as follows: Cash Flow Assumption Updates The methodology in this RFC updates our cash flow assumptions for the level, timing and application of cumulative defaults, the level and timing of voluntary prepayments, reimbursement rates on defaulted loans, borrower benefit utilization rates, deferment and forbearance rates and time periods, investment rates on cash accounts, and interest rates on auction rate securities and reset rate notes, described in Moody s Rating Methodology, Moody's Revised Cash Flow Assumptions for FFELP Student Loan-Backed Transactions: Methodology Update, November 30, Cash Flow Assumptions that are Unchanged From Previous Publications Our cash flow assumptions for payment lags and servicing fee inflation remain the same as those published in Moody s Rating Methodology, Moody's Revised Cash Flow Assumptions for FFELP Student Loan-Backed Transactions: Methodology Update, November 30, In addition, the current methodology makes no changes to the following assumptions, described previously in the publications indicated:» three-month LIBOR / 90-day CP basis risk assumptions as described in Moody s Rating Methodology, Methodology Update on Basis Risk in FFELP Student Loan-Backed Securitizations, November 19, 2008;» one-month LIBOR / 90-day CP basis risk assumptions as described in Moody s press release, Moody's Rates Variable Funding Notes Issued by Bluemont Funding I, January 21, 2010;» three-month LIBOR / 91-day T-Bill basis risk assumptions as described in Moody s pre-sale report, Nelnet Student Loan Trust Student Loan Asset-Backed Notes, October 15, 2009;» one-year average of 91-day T-Bill / 90-day CP basis risk assumptions as described in Moody s presale report, State Board of Regents of the State of Utah, Series 2010EE (1993 Indenture), December 16, 2010;» one-month LIBOR / 91-day T-Bill basis risk assumptions as described in Moody s press release, Moody's assigns definitive ratings to SLM Student Loan Trust , April 20, 2010; 1 Under the FFELP, students and their parents could obtain low-cost education loans to help pay for the cost of higher education. The FFELP was authorized by Title IV, part B of the Higher Education Act of 1965 (the Act), as amended, and includes the Federal Stafford, PLUS and Consolidation Loan Programs. Those loan programs were funded by lenders, guaranteed by guarantors, and reinsured by the federal government. The individual programs are defined in 668 of the Act. The federal guarantee percentage depends on when the loans was originated. For loans originated prior to October 1, 1993, the rate is 100%; for loans originated between October 1, 1993 and June 30, 2006, the reimbursement rate is 98%; and for loans originated after July 1, 2006, the reimbursement rate is 97%. 4 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

5 » interest rate (T-Bill) level scenarios as described in Moody s Rating Methodology, Methodology Update on Interest Rate Assumptions in Student Loan-Backed Securitizations, March 31, 2009;» slow repayment stress scenarios as described in Moody s Rating Methodology, Methodology Update: New Repayment Stress Scenario in FFELP Student Loan-Backed Securitizations, November 18, 2009;» interest rate assumptions for auction rate notes, as described in Moody s Special Report, Rating Actions on Student Loan Auction Rate Securities, dated September 10, 2008;» interest rate assumptions for variable rate demand bonds, as described in Moody s Rating Methodology, Updated Cash Flow Methodology for Variable Rate Transactions Backed by Federal Family Education Loan Program (FFELP) Student Loans, October 7, 2009; and» cumulative default rate and default timing curve assumptions for rehabilitated FFELP loans as described in Moody s Special Comment, Rehabilitated FFELP Student Loans: Higher Defaults but Still Guaranteed, September 21, Cash Flow Assumptions and Scenarios Being Published for the First Time:» expected case scenario;» interest rate scenario for transactions with fixed rate liabilities;» 91-day T-Bill / 90-day CP basis risk assumptions;» SIFMA/J.J. Kenny Index assumptions; and» adjustments for seasoned pools (cumulative default rates, default timing curve, voluntary prepayment rates, borrower benefit utilization rates). III. Executive Summary Moody s methodology in rating securities backed by FFELP student loans is based on both quantitative and qualitative analyses. The quantitative analysis focuses on stress testing a cash flow model of the transaction. That is, we determine modeling inputs (i.e., cash flow assumptions ) that reflect our view of the levels of stress that a transaction must withstand to achieve a particular rating and use our cash flow model to determine the extent to which investors would receive timely payments of interest and principal in those stress scenarios, given the transaction structure and underlying FFELP student loan pool composition. We model our cash flow stress scenarios using Structured Finance Workstation (SFW), the cash flow model developed by Moody s Wall Street Analytics. Moody s cash flow stress scenarios and assumptions for each rating level are shown in Appendix I. For each rating category, we apply a variety of stress scenarios, each consisting of a different set of cash flow assumptions, with more severe assumptions the higher the rating. The set of cash flow stress scenarios are designed to cover the range of securitizations, with their differing sensitivities to different factors, that we observe in the market. Appendices I and II describe how we combine the results of the various cash flow runs to determine a model rating. The stress scenarios are designed primarily to test the resilience of FFELP securitizations to withstand credit risk -- the risk that, given the structural features of the securitization and the underlying FFELP student loan pool composition, the student loan cash flows and the securitization s available credit 5 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

6 enhancement and liquidity support 2 will be insufficient to pay investors as promised. The primary components of credit risk in FFELP securitizations are: 1. Basis risk: Basis risk is the main driver of risk in most transactions backed by FFELP student loans. 3 Basis risk exists because the FFELP student loan assets and the securities are tied to different market interest rates and/or there are differences in the periods over which those rates are calculated (i.e., timing mismatches). 4 The lower the interest rate earned by the underlying FFELP loan pool compared with the interest rate earned by the securities, the lower the gross excess spread of the securitization. Similarly, if interest rates were to decline rapidly, the timing mismatch may mean that the interest rate on the assets falls before the interest rate on the liabilities, possibly resulting in negative gross excess spread. 2. The overall level of interest rates: The overall level of interest rates can affect the amount of gross excess spread generated by a securitization. For a particular securitization, the impact of the level of interest rates on excess spread depends on the collateralization level of the trust, the amount of cash in trust accounts. the portion of the underlying collateral pool that has interest rate floors (i.e., floor income loans), and the percentage of fixed-rate bonds in the securitization s capital structure. 3. Prepayment risk Prepayments whether voluntary 5 or involuntary (i.e., defaults) affect the length of time that a securitization remains outstanding, and therefore, the amount of net excess spread generated by a securitization. For securitizations with positive net excess spread, the higher the prepayment rate, the more the credit stress because the loans, on average, are outstanding for a shorter period of time, thereby decreasing the total net excess spread generated over the securitization s life. On the other hand, for securitizations with negative net excess spread, the lower the prepayment rate, the more the credit stress; the lower prepayment rate means the loans, on average, are outstanding for a longer period of time, so that the negative net excess spread erodes the value of the asset base for a longer period of time, making it less likely that there will be sufficient assets to repay the bonds. In FFELP student loan ABS, defaults are considered involuntary prepayments because when a loan defaults, cash from the government guarantee and credit enhancement is used to pay off the balance of the defaulted loan, resulting in a prepayment to investors. 4. Losses on the underlying loans: Losses on the underlying FFELP loans have a direct impact on the credit risk of a securitization since they reduce the asset base of the securitization. The severity of loan losses depends on loan default 2 Credit enhancement is typically provided by overcollateralization or gross excess spread. Liquidity support is typically provided by cash accounts that are fully funded at closing, such as capitalized interest accounts and debt service reserve accounts. 3 Basis risk in most securitizations backed by FFELP student loans is not hedged. 4 For example, some indexes are calculated as an average of daily rates during a specified period while others are calculated as the spot rate on a certain day. 5 A voluntary prepayment is a payment by the borrower in excess of the loan s scheduled payment for a given period. A borrower can prepay any amount in excess of the scheduled payment up to the full balance. 6 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

7 rates, the non-guaranteed portion of loans which can be as high as 3%, and the rate at which claims on defaulted loans are rejected by the guarantors 6 (i.e., the net claim reject rates 7 ). The magnitude of the credit risk arising from the potential default of the underlying loans is low because FFELP loans are guaranteed by guarantors for at least 97% of defaulted principal and accrued interest. provided the loans are originated and serviced according to DOE guidelines. Although the impact of potentially high losses is muted considerably by the federal guarantee, we incorporate the impact by examining the effects of higher-than-expected cumulative default rates and claim reject rates in of our stress scenarios. 5. Borrower benefit risk: Borrower benefits are incentives in the form of reductions in the loan interest rate or loan principal that lenders offer to eligible borrowers. Therefore, borrower benefits either reduce the principal balance of the loans or decrease the amount of interest received from borrowers, thereby reducing the excess spread. The number of eligible borrowers qualifying for these benefits and the rate at which they become disqualified after receiving the benefit drive the magnitude of the reduction in excess spread. The stress scenarios also test the resilience of FFELP securitizations to withstand liquidity risk and maturity risk. Liquidity risk is the risk that the available liquidity support and loan cash flows will be insufficient to make timely payment of bond interest and trust expenses. Maturity risk is the risk that there will not be enough cash to pay all principal by the legal final maturity of the security. For each rating, we examine a variety of combinations of the stressed values of the cash flow variables, since different securitizations are affected differently by the variables; a combination that is particularly stressful for one security may not be very stressful for another. The sensitivity of a securitization to changes in particular cash flow assumptions depends most importantly on the gross excess spread rate earned on the loans (i.e., securitizations that generate positive excess spread are more sensitive to high prepayment rates, whereas securitizations that generate negative excess spread are more sensitive to low prepayment rates.. In addition, the sensitivity of a securitization to certain interest rate scenarios depends on the amount of over- or under-collateralization in the securitization and on the portion of the loan pool comprised of floor income loans. We typically use the same set of generic cash flow assumptions for most securitizations. However, in some cases we may tailor the assumptions to a specific securitization if data provided by a particular issuer or special structural features of the securitization indicate the need for adjustments. If we use assumptions or scenarios other than those published in Appendix I for a new securitization, we will disclose those assumptions in securitization-specific research that we publish at issuance. The cash flow modeling results lead to a modeled rating, which is a key input into a rating committee s consideration before it assigns a final rating. However, the final rating assigned by a rating committee also incorporates our qualitative assessments of the operational 8 and legal risks in a 6 FFELP loans can lose their federal guarantee if they are not originated or serviced in accordance with the DOE s guidelines. When a FFELP loan defaults, the lender (or the servicer on behalf of the lender) submits a claim for reimbursement of the defaulted loan (after 270 days of delinquency) to the loan s guarantor. The guarantor either accepts the claim or rejects it, usually because the loan was not originated or serviced in accordance with DOE guidelines. 7 The net claim reject rate is the number of claims submitted and rejected by the guarantor for reimbursement (net of cures) divided by the total number of claims submitted. Defaulted loans that are guaranteed have a default severity limited to 3.0%. However, those loans that are rejected without a cure (i.e. those that are net claim rejects ) have a loss severity of 100%. 8 Operational risk is the risk of disruption in the performance of key securitization parties during the life of a securitization. In FFELP securitizations, the key securitization parties are the master servicer/administrator, the primary servicer(s) and the trustee. For our global structured finance operational risk guidelines, see Global Structured Finance Operational Risk Guidelines: Moody s Approach to Analyzing Performance Disruption Risk, April 12, Details of the student loan ABS operational risk guidelines appear in the Companion Report on U.S. Student Loan Asset-Backed Transaction (ABS) Operational Risk Guidelines, April 14, JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

8 securitization. Therefore, the actual ratings assigned to a securitization can differ from the model suggested output. IV. Considerations In Developing And Applying Stress Tests To develop our stress scenarios for the different rating levels, we start by formulating an expected case scenario, which is our expectation for each cash flow variable based on up-to-date historical performance information, adjusted for trends in the data and the effects of macroeconomic and legislative developments. We then apply stresses to the expected case scenario which result in what we refer to as the Baa, A, and Aaa stress cases -- to arrive at cash flow modeling assumptions that we deem appropriate for securities to be rated between Aaa and Baa. The relationship between the expected case and the stressed level of a variable depends on the potential variability (i.e., volatility) of future values of the variable. When assessing the appropriate stress level for a cash flow variable for a particular rating, we consider the aggregate effect of all of the stresses in that scenario resulting from the individual stresses applied to each of the variables. As a result, the stress we apply to each individual variable within the overall stress scenario would be moderated to some extent, since the probability that all of the factors would experience extreme variations simultaneously is lower than the probability of any single variable experiencing an extreme event. The effect of a stress can differ depending on other factors present in a securitization. For example, for a securitization that generates positive net excess spread, the higher the prepayment rate, the more the credit stress. On the other hand, for a securitization that generates negative net excess spread, the lower the prepayment rate, the more the credit stress. As a result, we often test the resilience of a securitization in both high and low prepayment rate scenarios. Our stress scenarios typically analyze the effects of a variety of assumed levels for a particular variable. For example, a low prepayment rate is typically more stressful for a securitization s liquidity because there is less cash available to make payments of bond interest and trust expenses early in the life of the securitization. However, for a securitization that generates positive net excess spread (i.e., gross excess spread less net losses of the underlying loan pool), a high prepayment rate is more stressful from a credit perspective. Therefore, we might test the securitization s liquidity by examining a scenario with a low prepayment rate, but test the securitization s credit protection in a scenario with a high prepayment rate. The number and nature of cash flow modeling assumptions and stresses that we use in examining the risks of a proposed FFELP student loan securitization may vary across securitizations and exceptions may apply, depending on the idiosyncrasies of individual securitizations that may have an impact on the risks to investors. For example, the risk reduction provided by an interest rate hedge that is not balance guaranteed, 9 would vary in different economic scenarios. The economic scenarios that would be particularly stressful would depend on the specific structure of the hedge. Therefore, our interest rate and prepayment rate assumptions may differ based on the hedge s specified notional balances during the life of the securitization, which affect the potential mismatches between those notional balances and the actual loan pool balances throughout the life of the transaction. V. Determining The Model Suggested Output Based On The Stress Tests The model suggested output for a given class of FFELP loan-backed securities must satisfy the criteria specified in Exhibit 2. The criteria include break-even stress tests as well as a failure rate that 9 For a balance-guaranteed interest rate hedge, the notional of the hedge adjusts over time to account for principal amortization experienced by the assets. 8 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

9 approximates an expected loss commensurate with the modeled rating category. For example, a Aaa rated security must pass all Aaa stress scenarios. That is, cash flows in the stressed model runs must be sufficient to pay all principal and interest as promised. Securities rated Aa or A must first pass all A stress scenarios. Expected loss in the Aaa stress scenario determines the modeled rating outcome. For example, a model rating of Aa must (1) pass all A stress scenarios and (2) in the Aaa scenarios, cannot lose more than 10% of the promised bond principal and interest. We determine other rating categories in a similar manner. EXHIBIT 2 Criteria for Model Suggested Output Model Suggested Must Pass* all of the following runs Output Approximate failure rate in Aaa scenario ** Aaa Aaa 0% Aa A 0% - 10% A A >10% * To pass a particular stress scenario, the cash flows must be sufficient to pay timely bond interest and securitization expenses, and all bond principal by the legal final maturity date of the securities in that scenario. ** The failure rate is defined as the principal and interest that is not paid as a percentage of the principal and interest that is promised for each bond. VI. Analyzing Credit Risk Moody s credit analysis includes cash flow runs that are used to determine whether the loan pool s cash flows, available credit enhancement, and liquidity support given the underlying collateral pool and structural features of the securitization are sufficient to pay investors of a tranche as promised in scenarios that are stressed to a level consistent with the desired rating. In general, the variables that we stress affect the level of gross excess spread available as credit enhancement and losses on the underlying collateral pool. The level of gross excess spread is primarily limited by interest rate index mismatches between the assets and the liabilities (reflecting basis risk), the overall level of interest rates, prepayments (both voluntary and involuntary), and the utilization of interest rate reductions. The directions in which we stress the cash flow variables from the expected case to analyze credit risk are discussed in the following sections. A. Basis Risk Our basis risk assumptions test the ability of a securitization to withstand bond interest rates that are high relative to loan interest rates. There are multiple combinations of asset and liability interest rate indices found in FFELP securitizations. On the asset side, as prescribed by law, FFELP loans disbursed after January 1, 2000 earn the 90-day financial commercial paper rate 10 (90-day CP) plus a spread, while older loans earn the 91-day T-Bill rate plus a spread. On the liability side, the interest rate on most bonds is pegged to either the three-month USD London Interbank Offered Rate (three-month LIBOR) or to one-month LIBOR. Furthermore, there are a variety of other bond interest rate indices that are used as the base for interest rates on failed auction rate securities, including 90-day CP, the 91-day T-Bill rate (either a spot rate or a quarterly average), the one-year average of the 91-day T-Bill rate (which, we will subsequently refer to as one-year average T-Bill ), and the Securities Industry and Financial Markets 10 The 90-day financial commercial paper rate is sometimes referred to as the three-month financial commercial paper rate. 9 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

10 Association (SIFMA) 11 index. However, liability interest rates tied to LIBOR and loan interest rates tied to 90-day CP are the most common pairing in FFELP securitizations. In Appendix II.A., we show our current assumptions regarding the potential mismatches between the interest rate indices of the assets and liabilities. The assumptions reflect our analysis of the historical levels and trends in the differences (i.e., spreads) between the market interest rates of the FFELP loans and that of the securities, including timing mismatches, and the volatilities of those spreads, as well as our view of those relationships going forward. The assumptions test a securitization s ability to withstand basis risk by examining both long-term periods of stressed interest rate spreads and shorter periods of even more highly stressed spreads (i.e., spikes ), intended to simulate short-term disruptions in the financial markets. In our stress scenarios, we generally model spikes to occur in months 6-12 of the first and fifth year of a securitization, and for a period of six months once the loan-pool factor falls to 10%. The timing of the spike periods is intended to roughly correspond to periods in which securitizations are typically particularly vulnerable to cash shortfalls. For example, the first spike is applied near the start of the securitization when many of the loans are not generating cash flow (i.e., many loans are either in school, grace, deferment or forbearance status). The second spike is intended to stress securitizations at a point in time after the capitalized interest account 12 has stepped-down to a zero balance. The final spike tests whether there is enough credit enhancement available to tolerate a spike occurring at the tail end of a securitization. B. Overall Level of Interest Rates The overall level of interest rates can affect the amount of gross excess spread generated by the trust. For a particular securitization, the impact of the level of interest rates on gross excess spread depends on the collateralization level of the trust, the portion of the underlying collateral pool that has floor income loans, the amount of cash in trust accounts (see I. Investment Rate on Cash Accounts, ) and the percentage of fixed-rate bonds in the securitization s capital structure. The level of interest rates has different impacts on over-collateralized and on under-collateralized securitizations. For a trust that is over-collateralized, a high interest rate environment generates more excess spread because the interest earnings on the excess collateral is higher. In contrast, for a trust that is under-collateralized, a high interest rate environment reduces excess spread because the interest expense of the excess liabilities is higher. In addition, the impact of the overall level of interest rates depends on the portion of the pool comprised by floor income loans. For a floor income loan, the interest earned on the loans is the maximum of (1) the CP rate plus a fixed spread (i.e., the SAP rate) and (2) the borrower s interest rate (the specified floor rate ). When the floating loan rate tied to CP falls below the borrower (floor) rate, the loan generates more income ( floor income ) than it would in the case in which it were a pure floating rate loan. However, in high interest rate environments (i.e., environments in which market interest rates are well above the floor rates), interest rates are unlikely to fall below the floor rates and the trust is unlikely to realize the benefit of floor income, substantially mitigating the benefit of floor income. 11 The SIFMA index is a municipal swap index based on a selected pool of tax-exempt Variable Rate Demand Bonds. 12 Capitalized interest accounts provide liquidity support in the early years of securitizations and are typically fully funded at closing. The account is drawn upon to pay interest on the securities and either all or some of the trust fees and expenses when collections from the underlying assets are insufficient. Amounts drawn from the capitalized interest account are not replenished. Typically, amounts in the capitalized interest account are scheduled to be released at certain dates, upon which the funds are released through the trust waterfall. The account is usually scheduled to be fully released within one to five years after the securitization closing. 10 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

11 For the limited number of securitizations with fixed-rate bonds, the overall level of interest rates can be a key driver of credit risk. The magnitude of the impact on credit risk depends on the portion of the bonds that are fixed rate. The interest rate risk exists because the underlying student loans are floatingrate, and therefore, there is an interest rate mismatch between the fixed-rate notes and floating-rate collateral. When short-term market interest rates fall, excess spread falls as the rates on the loans adjusts, without a corresponding adjustment in the interest rates on the bonds. This negative effect may be offset by floor income loans in a securitization s underlying collateral pool. We address this risk by examining scenarios in which we assume a low level of interest rates for the life of the securitization in our cash flow analysis. To account for these various risks, we examine multiple interest rate level scenarios for 91-day T-Bill, which is the index upon which most of our other interest rate assumptions are anchored, to analyze interest rate risk in FFELP securitizations. 13 Appendix II.B shows our current interest rate level scenario assumptions. C. Prepayment Risk The prepayment rate has two main effects on the credit risk of a FFELP securitization:» For securitizations with positive gross excess spread, faster prepayments pose credit risk since they reduce the outstanding balance of the underlying student loan pools faster, thereby reducing the amount of time that the securitization can generate gross excess spread. As a result, less gross excess spread will be available to (1) offset net losses on the underlying loans, (2) offset basis risk that is inherent in the securitizations, or (3) build up overcollateralization or reduce any existing undercollateralization. Therefore, to account for this risk in our cash flow analysis, we include an assessment of scenarios in which we apply higher-than-expected voluntary and involuntary prepayments (i.e., defaults).» For securitizations with negative gross excess spread, slower prepayments pose credit risk because the loans will be outstanding for a longer period of time, and therefore, the burdensome debt obligations are outstanding for a longer period as well. This would increase the amount of time during which the negative gross excess spread would erode the value of the asset base, making it less likely that there would be sufficient assets to repay the bonds. 14 Therefore, to account for that possibility, we also include in our analysis scenarios in which we apply lower-than-expected prepayments. We examine slow repayments in two ways. In one, we analyze the impact of low voluntary prepayments, but high defaults (i.e., involuntary prepayments). The high defaults offset the impact of the slower voluntary prepayments to some extent, but have a direct negative impact on the securitization through losses on the underlying loans. The second way in which we analyze slow repayments is in scenarios in which we examine the impact of low voluntary prepayments, expectedcase level of defaults on the loans, and longer periods of deferment and forbearance, which results in the loan pool amortizing down the slowest. Appendix II.C. shows our current cumulative default rate assumptions and Appendix II.D. shows our voluntary prepayment rate assumptions. D. Lifetime Net Loss Rates Losses on the underlying loans have a direct impact on the credit risk of a securitization since they reduce the asset base of the securitization. Loan losses depend on loan default rates, the non- 13 See Methodology Update on Interest Rate Assumptions in Student Loan-Backed Transactions, March 31, See Moody s Rating Methodology, Methodology Update: New Repayment Stress Scenario in FFELP Student Loan-Backed Transactions, November 18, JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

12 guaranteed portion of loans, and the rate at which claims are rejected by the guarantors (i.e., the net claim reject rates). Although the impact of potentially high losses is muted considerably by the federal guarantee, we incorporate the impact by examining the effects of higher-than-expected cumulative default rates and net claim reject rates in our stress scenarios. Appendix II.C. shows our current cumulative default rate assumptions based on historical static pool default information of the FFELP student loan industry. Appendix II.E shows our current net claim reject rate assumptions, based on historical performance of net claim reject rates of the FFELP student loan servicing industry. E. Timing and Application of Defaults The timing of defaults matters because defaults result in involuntary prepayments which can impact the excess spread as discussed above in Section VI.C. For all cash flow scenarios, we analyze the securitization with expected default timing (i.e., expected case default timing curve). All loans are subject to default immediately, except those in school or grace. Once loans in school or grace exit that status, they are subject to default. Appendix II.F shows our current default timing curve assumptions. F. Deferment and Forbearance Rates Deferments and forbearances allow borrowers to postpone payments of interest and principal for set periods of time for various reasons, but generally either because the student has returned to school or is facing financial difficulties. 15 Interest that accrues during the deferment and forbearance periods is capitalized and added to the loan balance. Our assumptions regarding the percentages of the pool in forbearance and in deferment and regarding the duration of the time that those loans remain in those statuses are presented in Appendix II.H. For securitizations with negative gross excess spread, the postponement caused by deferment and forbearance increases the negative gross excess spread because of the slower pay-down of the bonds. To capture that potential risk, we include stresses of longer durations of deferment and forbearance utilization in our cash flow scenarios. On the other hand, for securitizations with positive gross excess spread, higher rates of deferment and forbearance act like slower prepayments and generate more assets over time because the excess spread earned during the postponement period is added to the principal balance of the loans, providing additional credit protection to investors. Therefore, for those securitizations, lower rates of deferment and forbearance imply less protection for investors. To capture that potential risk, we include stresses of lower-than-expected deferment and forbearance utilization in our cash flow scenarios. G. Borrower Benefit Utilization Rates Most securitizations contain loans with borrower benefits, which are incentives in the form of reductions in the loan interest rate or loan principal that lenders offer to eligible borrowers. For loans that contain borrower benefits, borrowers typically qualify for the benefits either by having their loan payments deducted automatically from their bank accounts (i.e., Automated Clearing House benefit or ACH benefit ) or by making a specified number of consecutive payments without a delinquency (i.e., on-time benefit ). The number of consecutive on-time payments required to qualify for benefits varies by issuer, but most commonly the benefits vest after making 12, 24, 36 or 48 consecutive ontime payments. In some cases, borrowers qualify for the on-time benefit simply upon entering repayment. However, borrowers typically become disqualified from receiving the benefits if they become delinquent and, for some lenders, if they ever utilize forbearance. 15 Under the FFEL program, deferments are entitlements. Deferment is generally granted for a period of one to three years for economic hardship; however, there is generally no time limit for students returning to school. Forbearance allows the borrower to temporarily cease making payments. Forbearance is typically granted in sixto 12-month increments, not to exceed three years. 12 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

13 The most common type of borrower benefit program offered is the ACH benefit, which typically offers an interest rate reduction of at least 0.25% if a borrower has payments automatically deducted electronically from a bank account. Appendix II.I. Exhibit 1 shows our assumptions for the utilization rate of the ACH benefit; the assumptions are based on historical experience provided by issuers. The effect that on-time borrower benefits have on the credit risk of a securitization is determined by the number of eligible borrowers qualifying for these benefits and the rate at which they become disqualified after receiving the benefit. For on-time benefits, we assume that, of a pool of eligible borrowers entering repayment, an increasing number of borrowers become disqualified from receiving the benefits as the loans become more seasoned, i.e., as the pool ages, an increasing number of borrowers will have been delinquent. In particular, we have observed a significant reduction in the number of borrowers that retain on-time benefits in the first year in active repayment, mainly due to first-payment delinquencies. Since students often move their primary residence upon graduation, a high percentage of first payment delinquencies can be attributed to the inability to locate those students. Therefore, the percentage of a pool of eligible borrowers in active repayment that retains qualification declines as the loans become more seasoned. Appendix II.I Exhibit 2 shows our assumptions, which are specified in terms of that utilization rate, which declines as the loans become more seasoned. Borrower benefits have the largest impact on securitizations with large concentrations of loans originated prior to October This is because the College Cost Reduction and Access Act enacted in October 2007 reduced the profitability of FFELP loans originated after October 1, 2007, and as a result, many lenders discontinued or significantly reduced borrower benefits on new loans after that date. H. Servicing Fee Inflation Our assumptions for servicing fee inflation in the stress cases reflect our analysis of the historical rate of inflation and the likelihood that the cost to service loans will increase over time. In the expected case, our assumptions inflate servicing fees as specified in the servicing agreement. The assumptions are shown in Appendix II.J. I. Investment Rate on Cash Accounts Securitizations typically allow the indenture trustee to invest funds from cash accounts (such as capitalized interest accounts, reserve accounts, loan acquisition accounts, and collection accounts) in certain eligible investments. The funds are generally invested in low-risk, low-yielding securities. For trusts that have large balances in these accounts, the low yield reduces gross excess spread since the interest rate earned on the cash is less than on the loans; in fact, the interest rate on the cash account balances could be even lower than the interest rate paid on the liabilities (i.e., a situation in which we have negative carry ). Our investment rate assumptions are shown in Appendix II.K. J. Remarketing or Rate Reset Risk In some types of securitizations, the interest rate of the liabilities is reset (either periodically or on a pre-defined date) through a remarketing process and is therefore subject to reset or remarketing risk. As a result, the interest rate during the life of the securitization can vary unpredictably. We address this risk through our rate reset assumptions for variable rate demand bonds (VRDBs), 16 auction rate securities, and reset rate notes 17 as shown in Appendix II.L. 16 The assumptions apply to the underlying rating for a VRDB. The underlying rating is in addition to the long- and short-term ratings that address the risks faced by traditional investors that purchase the VRDBs. 17 Reset rate notes have an initial period, for which interest rates are established at closing. After the initial period, the existing bondholders can "put" the bonds to the issuer, who is obligated to repurchase them with proceeds from the remarketed bonds. If the bonds cannot be remarketed after the initial period, the initial bondholders continue to hold the bonds at a pre-set penalty rate until the notes can be successfully remarketed. 13 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

14 K. Note Amortization The classes of notes in a securitization are typically paid on a sequential or pro-rata basis as specified in the legal documents. In the limited number of securitizations issued prior to 2008, in which the amortization of the notes is at the issuer s discretion, we assume that the notes with the lowest cost of funds are paid down first while the notes with the highest cost of funds are paid down last. L. Pre-funding or Recycling Pre-funding or recycling periods both allow for additional student loans to be added to the trust after the securitization s closing date. 18 That introduces the risks that (1) the new loans added could reduce the overall credit quality of the pool and (2) the gross excess spread would be reduced because of the risk of holding cash for long periods of time (i.e., negative carry). In analyzing a securitization with pre-funding or recycling features, we stress the characteristics of loans that can be added over time. Our pre-funding and recycling assumptions are case-specific, based largely on the issuer s historical portfolio mix and consistency in originations and on the terms of the securitization documents that provide concentration limits on the loans that can be added. In those cases where an issuer has identified a specific pool of loans to be acquired, our assumptions will reflect the characteristics of the identified pool. For securitizations with pre-funding or recycling features, we assume that the characteristics of the loans will adhere to any concentration limits set forth in the securitization documents (i.e., loan type, school type, guarantee percentage, SAP rate, etc.) in a manner that is most disadvantageous from a credit perspective. We also stress the timing of the loans to be acquired during the pre-funding or recycling periods. To stress credit, we typically assume that loans will be acquired at later dates than expected. That assumption addresses the possible negative effects on excess spread of holding cash for long periods of time (i.e., negative carry). The specific assumption would be based on our review of the issuer s historical experience in generating new loans and its expected timing for drawing down the prefunding or recycling accounts. If the issuer does not provide data regarding the timing of historical loan acquisitions, we typically assume that the new loans will be acquired on the last day of the prefunding or recycling periods. 19 In some cases, we run non-origination scenarios where acquisition proceeds are held in the acquisition accounts and are not used to acquire any new loans during the prefunding or recycling periods; instead the proceeds are used to pay down outstanding bonds on the last day of the pre-funding or recycling periods. This scenario also addresses negative carry and tests the securitization to determine if it is more stressful to pay down bonds than to acquire loans at the end of the pre-funding or recycling periods. If a particular transaction is structured such that funds available in the pre-funding and/or recycling account can be used to make interest and principal payments, we may require cash flow runs where all of the assets are acquired on the first day of the acquisition period. M. Rehabilitated FFELP Loans A rehabilitated FFELP loan is a loan on which the borrower previously defaulted and has subsequently made at least nine on-time payments in full on the loan. We expect rehabilitated FFELP loan pools to have a higher net loss rate compared with pools of non-rehabilitated FFELP loans because although 18 In a "pre-funded" securitization some of the proceeds from the closing of the securitization are set aside in a pre-funding account to be used to purchase additional receivables during the pre-funding period. In a "recycling" securitization, principal collections from the loans can be used to purchase additional receivables during the revolving period. 14 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

15 the rehabilitated loans benefit from the same degree of federal guarantee, they are expected to default at a significantly higher rate than non-rehabilitated loans. To address the different performance expectation of rehabilitated FFELP student loans in our ratings, we modified our cumulative default rate and default timing curve assumptions, which are shown in Appendix II.C. Exhibit 3 and Appendix II.F. Exhibit 2, respectively. 20 All other assumptions are the same as for non-rehabilitated FFELP securitizations. N. Adjustments for Seasoned Pools Our generic assumptions assume that a pool of loans is unseasoned; however, our cash flow assumptions for a securitized pool are concerned with the performance only during the life of the securitization, which excludes the amortization and performance that would normally occur while the loans became "seasoned" prior to the securitization closing. We typically make adjustments to certain cash flow assumptions to account for the seasoning of the underlying student loan pool; that is, for loans that have been in repayment for a certain number of months before they were securitized. The need for adjustment arises principally from the need to account for the amount of amortization that has occurred on seasoned pools. We typically apply seasoning adjustments to the assumptions for cumulative default rates, default timing curves, voluntary prepayment rates, and the borrower benefit utilization rates; therefore, cash flow assumptions for seasoned pools may differ from those presented in the standard cash flow assumptions in Appendix I and II. For factors that we assume do not typically change over time as loans season, such as ACH benefits, we do not make a seasoning adjustment. 1. Cumulative Default Rates The standard expected case assumptions for cumulative default rates presented in Appendix II.C Exhibit I assume no seasoning. We adjust the cumulative default rate for the seasoning in the pool using the default model in Appendix III. The adjustment accounts for both the amount of defaults and the amount of principal amortization (through scheduled payments, prepayments and defaults) that the seasoned pool likely has already experienced. 2. Default Timing Curves For seasoned loan pools, we adjust the default timing curves based on the weighted average seasoning of the entire loan pool as Appendix II.M.2 shows. 3. Voluntary Prepayment Rates For seasoned loans, we adjust the voluntary prepayment curve based on the weighted average seasoning of the loan pool as Appendix II.M.3 and Appendix II.M.4 shows. 4. On-Time Borrower Benefits Utilization Rates Our assumptions for on-time borrower benefits for seasoned pools are adjusted to account for the pool having aged past at least the first year of active repayment (that is, having made at least 12 months of payments), during which the largest drop in qualification rates tends to occur. Appendix II.M.5. shows our seasoned on-time borrower benefit utilization rates. 19 See Moody s Special Comment, Rehabilitated FFELP Student Loans: Higher Defaults but Still Guaranteed, September 21, JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

16 VII. Analyzing Liquidity Risk The major liquidity risk in FFELP student loan pools is that, early in the life of a securitization, when a significant portion of the loan pool is typically in either school or grace status, the cash flows available to pay bond interest may be further reduced by other factors, resulting in the inability to make the required bond interest payments. Consequently, to assess liquidity risk, we determine whether bond interest will be paid in a timely manner when those factors are stressed in a direction that adds liquidity pressure to the securitization, as follows:» Higher-than-expected cumulative defaults on the underlying loans, which reduce cash flows early in the life of a securitization;» Lower-than-expected voluntary prepayment rates, which reduce cash flows early in the life of a securitization;» Higher-than-expected rates of loan deferment and forbearance; - The suspension of payments during periods of deferment and forbearance - 21 reduces cash flows on the affected FFELP loans until they re-enter repayment status; and» Long payment lags, delaying cash flows. Payment lags can occur for a number of reasons, including delays in default reimbursement, Interest Subsidy Payments (ISP), Special Allowance Payments (SAP), and borrower payments because of delinquencies. Appendix II.G shows our current payment lag stress assumptions. - - Lags in Default Reimbursement Lags in reimbursement for defaults depend on the length of time for the lender to notify the guarantor after a loan defaults and the length of time for the guarantor to make payment after being notified. Under DOE regulations, lenders are generally required to file a default claim with a guarantor within 360 days of delinquency and the guarantor is required to pay the claim or return it to the lender within 90 days of receiving the claim. That implies that the maximum lag for FFELP reimbursements is 450 days after the first day of delinquency, assuming the claim is not returned and re-filed. Actual lags in reimbursement may be less than the maximum. Our assumptions are based on a servicer s historical experience and on the rating desired. Payment Lags for Interest Subsidy Payments (ISP) and Special Allowance Payments (SAP) Student borrowers with subsidized FFELP loans are not required to make interest payments on their loans while they are in school, grace or deferment status. In those cases, the DOE replaces the interest payments with ISP on a quarterly basis. In addition, in high interest rate scenarios, the interest rate that the DOE owes the lender is higher than the interest rate paid by the borrower. In those cases, the DOE makes up the difference with SAP, also paid on a 22 quarterly basis. At the end of each quarter, lenders must submit a report to the DOE to receive ISP or SAP due on their loans. 23 The DOE is required to authorize the payment of any ISP or SAP payments within 30 days of receipt of a complete and accurate report. Our assumptions for the ISP and SAP payment lags address the risk that there may be a delay in 20 Under the FFEL program for "subsidized" loans, the DOE will make interest subsidy payments during in-school, grace, and deferment status. For unsubsidized loans or loans in forbearance, interest accrues while those loans are in non-paying statuses and is capitalized to the loan balance once the loans enter active repayment, therefore these loans may not generate any cash flow for extended periods of time. 21 In practice, servicers submit the report to the DOE on behalf of lenders. 16 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

17 (1) the submission by lenders of complete and accurate reports and/or (2) in the payments sent to the trust by the DOE.» Other factors similar to those that affect credit risk, such as adverse movements in interest rate levels and spreads, low reimbursement rates on defaulted loans, high borrower utilization rates, and high servicer fee inflation. TABLE 1 FFELP Student Loan Trusts VIII. Operational Risks A. Operational Risks Operational risk is the risk of disruption of key securitization parties in structured finance securitizations. The key transaction parties in FFELP student loan-backed securitizations include the master servicer/administrator, the primary servicer(s) and the trustee. Our guidelines for U.S. student loan ABS operational risk vary based on four characteristics: the asset type, the ratings of key transaction parties, the duties of those parties, and whether the parties satisfy certain presumptions. For FFELP securitizations with non-investment-grade master servicers or administrators, the guidelines in general call for a back-up arrangement to mitigate the operational risk. The standard U.S. trustee language may also be able to compensate for the risks if the master servicer or administrator satisfies certain presumptions. Table 1 below lays out the operational risk guidelines for FFELP student loan securitizations. 1-A Investment Grade Master Servicers/Administrators or a State Agency/Public Instrumentality of a State as Master Servicers/Administrators Duties Criterion Aaa/P-1 may be achievable under the following conditions: Master servicer/administrator only performs adminsitrative duties, i.e. third-party primary servicer(s) Master servicer/administrator also collects on the loans as primary servicer Third party primary servicer(s) satisfies primary servicer presumptions Master servicer/adminsitrator satisfies primary servicer presumptions >> Standard U.S. trustee language is preferred but not required» No BUMSi/BUAii at closing» Standard U.S. trustee language is preferred but not required» No BUMS/BUA at closing» Standard U.S. trustee language is preferred but not required 1-B Non-Investment Grade Master Servicers/Administrators Duties Criterion Aaa/P-1 may be achievable under the following conditions: Master servicer/administrator only performs adminsitrative duties, i.e. third-party primary servicer(s) Master servicer/adminsitrator satisfies master servicer/administrator presumptions Third party primary servicer(s) requirements:» Satisfies primary servicer presumptions; if not, a backup servicing arrangement with a widely recognized servicing company Master servicer/administrator requirements:» Named BUMS/BUA that satisfies the master servicer/administraotr presumptions or standard U.S. trustee language 17 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

18 TABLE 1 FFELP Student Loan Trusts Master servicer/administrator also collects on the loans as primary servicer Master servicer/administrator does not satisfy master servicer/administrator presumptions Master servicer/adminsitrator satisfies primary servicer presumptions Master servicer/adminsitrator does not satisfy primary servicer presumptions Third party primary servicer(s) requirements:» Satisfies primary servicer presumptions; if not, a backup servicing arrangement with a widely recognized servicing company Master servicer/administrator requirements:» Named BUMS/BUA that satisfies the master servicer/administraotr presumptions» Named BUMS/BUA that satisfies the primary servicer presumptions or standard U.S. trustee language» Named BUMS/BUA that satisfies the primary servicer presumptions and is one of the widely recognized servicing companies 1-C Non-Investment Grade Master Servicers/Administrators with Large Scale Platform Duties Criterion Aaa/P-1 may be achievable under the following conditions: Administration functions only or both administration and primary servicing Master servicer/administrator s platform has sufficient scale and/or franchise value so as to presumptively remain intact into and during a bankruptcy; servicing fee is sized to be economic» Standard U.S. trustee language or BUMS/BUS» Box 1-A applies to FFELP loan securitizations serviced and administered by investment-grade entities or a state agency or a public instrumentality of a U.S. state. Regardless of whether the master servicer or administrator is actually collecting on the underlying loans, the securitization does not need to have a backup master servicer (BUMS) or backup administrator (BUA) or the standard U.S. trustee language 24 in the transaction documents in order to retain or achieve a Aaa rating. However, the master servicer or administrator should satisfy the primary servicer presumptions 22 Standard U.S. trustee language: in many U.S. securitizations, the indenture or trust agreement includes language to the effect that, in the event of a servicer termination, the trustee will either name a successor servicer, take over the duties itself, or if unable or unwilling to do so, it may petition a court of competent jurisdiction to name a successor servicer. In termination situations in the U.S., the trustee has generally found another party to service. We are unaware of any situations in which a trustee has petitioned a court to appoint a servicer. This language provides reasonable assurance that the securitization will not be left truly rudderless should a servicer termination event occur. 18 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

19 » 38 if it also collects payments on the loans. In the case of third-party primary servicing, the thirdparty primary servicers also satisfy the primary servicer presumptions.» Box 1-B lists the characteristics that are consistent with a Aaa rating for FFELP loan securitizations with non-investment-grade master servicer or administrator. To compensate for the greater uncertainty, a higher standard is necessary relative to Box 1-A. Securitizations can retain or achieve a Aaa rating by having a backup master servicer, backup administrator, or standard U.S. trustee language if the master servicer or administrator satisfies certain presumptions. The backup master servicer or backup administrator needs to satisfy the master servicer or administrator presumptions if the master servicer or administrator only performs administrative functions, or the primary servicer presumptions if the master servicer or administrator also collects on the underlying loans. In the former case, the third-party primary servicers should meet the primary servicer presumptions. If they do not meet the presumptions, a backup servicing arrangement with a widely recognized servicing company can further mitigate the risk.» Box 1-C applies to FFELP loan securitizations whose master servicer or administrator has platform with sufficient scale or franchise value to remain intact into and during a bankruptcy. Either the presence of standard U.S. trustee language or named backup master servicer or backup administrator is sufficient to retain or achieve a Aaa rating. For our global structured finance operational risk guidelines, see Global Structured Finance Operational Risk Guidelines: Moody s Approach to Analyzing Performance Disruption Risk, April 12, 2011, which includes student loan ABS. More details of the student loan ABS operational risk guidelines appear in the Companion Report on U.S. Student Loan Asset-Backed Securitization (ABS) Operational Risk Guidelines, April 14, B. Servicing and Administration Fees Master servicing, administration, and primary servicing fees should be sufficient, in terms of amount (i.e., market rate) and payment priority, to attract a replacement or back-up to perform those roles should the original party resign or be removed at any time during the life of a securitization. A replacement or back-up would likely require a senior position in the waterfall during stressful situations. Therefore, we analyze whether the fees paid senior in the waterfall are adequate to cover actual costs over time in stressed situations. IX. Legal Risks Our legal analysis focuses on two major sources of risk posed by potential bankruptcies of the following securitization parties: 1. the securitization sponsor and any affiliated entities in the chain of title of the assets transferred to the securitization vehicle; this includes the originator and seller if either is a separate legal entity from the sponsor (collectively, such entities are referred to in this section as the sponsor ) and 2. the securitization vehicle. 23 General presumptions for master servicers, administrators, and primary servicers include: (a) no governance concerns (b) audited financials from major CPA firm (c) in business at least 5 years (d) ownership public, or private equity, or institutional, not private individuals (e) net worth at least $5 million or equivalent (f) net worth at least $20 million or equivalent for the primary servicer (g) adequate liquidity. If a master servicer or administrator does not have $5 million net worth, we consider other mitigants such as historical financial information that demonstrates consistent net worth growth in the past and positive projected financial information. For certain primary servicers that have a close linkage with a not-for-profit sister company, we consider consolidated financial information of the two entities to assess whether or not the primary servicer meets the net worth requirement. 19 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

20 In both cases, the risk to investors arises either from the possibility that creditors outside the securitization might make a claim on the securitization assets or that the cash flows to investors will be subject to an automatic stay and be delayed through the bankruptcy proceedings. In the following sections, we describe the legal structure we generally view as necessary to prevent the legal risks from causing the notes ratings to be lower than the model suggested output described above. If any of the legal protections are absent, we determine the extent of the risk and how it may be mitigated in other ways, including additional credit enhancement. In the absence of sufficient mitigating factors, Moody s may assign a rating lower than that indicated by our modeling. A. Bankruptcy of the Sponsor Our legal analysis of the potential bankruptcy of the sponsor is an assessment of three key factors: 1. whether the receivables have actually been sold; 2. whether the owner of the assets (the securitization trust) would be substantively consolidated with the sponsor in the event of the sponsor s bankruptcy; and 3. whether the securitization trustee can enforce its ownership or security interest in the collateral once the sponsor has filed for bankruptcy protection. Therefore, we assess the likelihood that a sponsor bankruptcy proceeding whether voluntary or involuntary would delay or reduce the payments on the notes. The degree to which the securitization is protected against these risks determines the extent to which the transaction ratings can be higher than those of the sponsor s own rating. 1. Risk of Sponsor Bankruptcy Differs By Type of Sponsor In this section, we discuss how the risk of a sponsor bankruptcy varies for each of the major types of sponsors of FFELP securitizations: (a) state agencies, (b) non-profit corporations, and (c) for-profit corporations. a) State Agencies State agency issuers are not subject to involuntary bankruptcy because they typically are deemed to be a governmental units, which are not subject to involuntary proceedings under Chapter 7 or 11 of the Bankruptcy Code. 26 A state agency or municipality may file a voluntary case under Chapter 9 only if it is specifically authorized, in its capacity as a municipality or by name, to be a debtor under such chapter by state law, or by a governmental officer or organization empowered by state law to authorize such entity to be a debtor under such chapter. (11 U.S.C. Section 109.) Usually, counsel to the state agency provides an opinion that the state agency issuer is a governmental unit under the Bankruptcy Code. We review such opinion to determine the state agency s authorization to file a voluntary petition. (also see the Legal Opinions subsection below.) 24 Typically, state agency originators issue student loan-backed notes as limited recourse debt obligations backed by a pool of student loans. The issuer grants an indenture trustee a security interest in the student loan assets, but there is no sale or transfer of the loans. 20 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

21 b) Non-Profit Corporations A non-profit issuer is a corporation that has qualified for a tax exemption under Section 501(c)(3) of the Internal Revenue Code. Non-profit issuers use a legal structure that is similar to that employed by state agency issuers -- an indenture provides for the issuance of limited recourse bonds that are secured by an interest in student loan assets. Moody s bankruptcy risk analysis focuses on the status of the non-profit corporation under the Bankruptcy Code. Nonprofit corporations are generally not subject to involuntary petitions by creditors under the Bankruptcy Code. Section 303(a) of the Bankruptcy Code provides that: An involuntary case may be commenced only under Chapter 7 or 11 of this title and only against a person, except a farmer, family farmer, or a corporation that is not a moneyed, business or commercial corporation, that may be a debtor under the chapter under which such case is commenced. However, the Bankruptcy Code does not provide a clear definition of a moneyed, business or commercial corporation and the analysis of a non-profit s status under the Bankruptcy Code is dependent on an analysis of the relevant state law. Our assessment of the risk typically is informed by a reasoned opinion from the non-profit s counsel that the issuer would not be subject to an involuntary petition. Given the diverse approaches taken by courts in this area, we typically view the opinion to be stronger if its conclusion that a non-profit corporation is not a moneyed, business or commercial corporation is based on analysis under both of the tests that have emerged under case law, as well as on an examination of other relevant factors, such as the corporation s actual conduct and method of operation. 27 Unlike state agencies, non-profit issuers are able to seek the protection of a bankruptcy court by filing a voluntary petition without explicit permission of any state authority. However, given the absence of the threat of involuntary bankruptcy by creditors, a non-profit corporation is less likely to seek protection from creditors through a voluntary bankruptcy proceeding. The risk of a voluntary bankruptcy filing is significantly reduced if the non-profit issuers have limited the scope of their activities to the specified functions of securitizing student loans, have no additional flexibility to engage in other activities, and have several independent directors or independent managers (or a similar governance mechanism). c) For-Profit Originators For-profit originators are subject to both voluntary and involuntary bankruptcy. 2. Evaluating Bankruptcy Risk for Sponsors Who Can Become Bankrupt a) True Sale If the FFELP student loan receivables truly have been sold, then in a sponsor bankruptcy, the bankruptcy court should neither include the loans in the sponsor s bankruptcy estate nor impose an automatic stay to delay payments stemming from the receivables to investors. In determining that the sponsor s asset transfer to the securitization vehicle was, in fact, a true sale, the primary test is whether the risks and benefits of ownership have been transferred. Courts have confronted a number of fact-specific cases. Relevant facts for our analysis of 25 Generally, bankruptcy courts go beyond a simple analysis of whether a non-profit has qualified for an exemption from federal taxation under Section 501(c)(3) and extend their assessment to an ad hoc factual review of the corporation s charitable purpose. Two tests, while not applied universally by bankruptcy courts, have emerged from the case law the state classification rule and the bankruptcy rule. The state classification rule focuses on the corporation s not-for-profit status under state statutes while the bankruptcy rule requires the court to independently analyze the corporation s charitable nature focusing on its charter and its activities. However, courts generally have not relied on one rule exclusively, but rather on an analysis of the corporation s actual conduct and method of operation. In particular, courts have focused on the corporation s capital structure and its historical return to investors. 21 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

22 FFELP student loan securitizations include, but are not limited to, the level of recourse retained by the sponsor, the securitization s economics and the unequivocal intent of the parties. As part of our analysis, we typically review a legal opinion from counsel hired by the sponsor in connection with the securitization that states that the transfer of the collateral to the trust is a legal true sale. b) Substantive Consolidation Substantive consolidation is an extraordinary remedy in which the bankruptcy court views the bankrupt sponsor as being indistinguishable from the securitization vehicle, resulting in confusion about which assets should belong to which creditors, including the ABS noteholders. As a result, the securitization vehicle s assets are consolidated with the sponsor s assets and become subject to the automatic stay in bankruptcy. Our substantive consolidation legal analysis includes an assessment of the likelihood that the securitization might be subject to factors that have led courts to order consolidation in the past, including cases in which (i) the records of the companies were intertwined to such an extent that they could not be separated; (ii) arm s-length dealings between the two companies had not been observed; or (iii) boards of directors overlap. The risk of consolidation of corporate entities is mitigated when the affiliated companies are organized and managed separately and all corporate formalities of separateness are observed in their ongoing governance. As part of our analysis, we typically review a legal opinion from legal counsel hired by the sponsor in connection with the securitization that states that the trust would not become consolidated into the purchaser upon a bankruptcy of the purchaser. c) Perfection Our legal analysis assesses whether the securitization s indenture trustee (on behalf of the noteholders and other secured parties) can enforce its ownership or security interest in the student loans and other collateral in a sponsor bankruptcy. To enforce its interest, the indenture trustee must perfect its security interest -- that is, notify the world of the trust s interest in the asset -- and check to be sure that no other party has a claim upon the collateral superior to the indenture trustee s claim. In other words, the indenture trustee s interest in the student loans must be a first perfected security interest. A first priority perfected security interest serves as back-up protection to the securitization vehicle s rights to the cash flows in the event of the sponsor s bankruptcy, in case the bankruptcy court recharacterized the true sale transfer of the loans to the securitization vehicles as a secured loan. A perfected security interest in FFELP student loans is usually achieved by filing financing statements that are required under the applicable Uniform Commercial Code (UCC). As part of our analysis, we typically review a legal opinion from the issuer s legal counsel in connection with the securitization which concludes that the investors have a perfected security interest in the student loans, which includes payments made thereunder. B. Bankruptcy of the Securitization Vehicle Our analysis of a potential bankruptcy of the securitization vehicle focuses on two main assessments: (i) the likelihood that a third-party creditor would successfully petition the securitization vehicle into an involuntary bankruptcy proceeding or (ii) the likelihood that the securitization vehicle would voluntarily seek bankruptcy protection. We refer to securitizations in which the likelihoods of both an involuntary and a voluntary bankruptcy are very low as bankruptcy remote securitizations. To analyze the likelihood of an involuntary bankruptcy, we assess characteristics that mitigate the risk, such as if the securitization vehicle limits its business activities to (i) issuing debt pursuant to the 22 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

23 indenture and (ii) using the proceeds to acquire eligible student loan collaterals; and (2) enters agreements with the sponsor and the other securitization parties not to petition for the securitization vehicle s dissolution, liquidation or bankruptcy. Similarly, other factors that mitigate the risk of a voluntary bankruptcy include (i) the securitization vehicle s charter documents requiring directors who are independent of the sponsor s and its affiliates and (ii) a unanimous vote of directors as a condition to filing a voluntary bankruptcy. C. Legal Opinions We receive legal opinions provided by the sponsor and use them as a basis to inform our understanding of the legal issues in the securitization. In our analysis, we focus on the extent to which various assumptions, qualifications, and limitations weaken the opinion. We give no credit to TriBar opinions, which are opinions that rely on cross-references to sections from the Legal Opinion Accord of the American Bar Association Section of Business Law (1991). In addition to the opinions mentioned above regarding true sale, non-consolidation, security interest, and the inability to be petitioned into bankruptcy, we typically review the following opinions:» Tax opinions regarding whether the securitization is in compliance with applicable tax law and has been structured so that the issued notes will be treated as debt for United States federal income tax purposes, and that the issuer will be a tax-exempt entity in light of the contemplated securitization;» Enforceability opinions regarding whether the governing securitization documents are legal, valid, binding and enforceable in accordance with their terms against the securitization parties;» Organizational opinions regarding whether (1) the securitization parties were duly formed and authorized to execute the agreements to which they are party, (2) the execution of such documents were permitted under their organizational documents and conflict with any applicable law or rules or regulations of any governmental authority or with other material agreements of such securitization parties and (3) there is material pending or threatened litigation against such parties to the best of their knowledge that would interfere with their performance of their obligations under the securitization documents; and» Investment Company Act opinions providing that the issuer is not subject to the requirements of the Investment Company Act of In addition, Moody s may request additional opinions or legal analysis. This is particularly the case when the regulations or law governing FFELP loans are changed or are contemplated to be changed, including the application and interpretation thereof. Moody s does not rely on legal opinions but uses them as a basis to inform us of the legal issues in the securitization. X. Provisions of the Operative Documents A. Representations and Warranties and Covenants Securitizations typically contain various representations, warranties and covenants from the securitization participants (collectively, R&Ws 28.) Those R&Ws often can reduce both legal and credit risk to investors. The benefit that Moody s ascribes to the R&Ws depends upon (a) the content of the R&Ws, (b) the contractual oversight and remedies in place for breaches of R&Ws, (c) the financial 26 See Moody s report SEC Rule 17G-7 Report of R&Ws: Student Loans Benchmark v1.0, September 23, JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

24 strength and the ability, willingness and explicit commitment of the R&W provider to comply with its contractual and other legal obligations, and (d) the strength and integrity of certain procedures, including compliance with FFELP servicing and origination guidelines. Generally, we regard weak R&Ws (e.g., separateness covenants) as adding to a securitization s legal risk in the event of a sponsor s bankruptcy. B. Identifying and Enforcing Breaches of Representations and Warranties Moody s rating committees review R&Ws on a case-by-case basis to assess remediable breaches, such as the originator s or seller s requirement to buy-back loans later found to have been ineligible collateral at the time of their transfer to the securitization vehicle. 29 In addition to determining the likelihood of discovering such remediable breach, we assess the appropriateness of remedies designed to reimburse the trust for such breach and specific procedures for such remedies enforcement. Typically, under the securitization documents, loans that breach R&Ws are either repurchased or substituted with eligible loans, and the issuer is indemnified for any costs associated with such breach. Moody s reviews those remedies, the specific procedures for enforcing the remedies, and provisions that require reporting of breaches and their cures to the affected market participants (e.g., noteholders or the securitization s controlling party) to determine their credit impact. C. True Sale Transfer of Loans to the Issuer In cases when our analysis relies on a true sale of the loans, we assess whether the sponsor s ownership transfer of the loans to the issuer provides the indenture trustee with the necessary documentation to foreclose against the loans. To achieve this, the securitization s operative documents typically stipulate that all of the loan documents necessary to affect the ownership transfer are transferred to the collateral agent or custodian at closing. Furthermore, we evaluate whether one of the deal parties will be obligated to record any UCC filing or other document whose recording is required to make the assignment of the loan effective against the student. The insufficiency of such provisions may weaken the securitization. D. Servicing and Indenture Trustee Standard of Care Student loan transactions typically set forth the standards pursuant to which a master servicer and/or servicer (collectively, servicer ) will service the collateral and the indenture trustee will act as trustee for the collateral. Moody s views as credit-neutral a standard that obligates the servicer or the indenture trustee to use: 1. customary and usual servicing or trustee practices of prudent institutions that service or provide trustee services for student loans or similar collateral, 2. a degree of care that is no less than what the servicer or indenture trustee itself exercises when servicing or providing trustee services for comparable assets for itself, affiliates and third-parties, and 3. a standard that is without regard to the servicer s, the indenture trustee s or any of their respective affiliate s ownership of any of particular securities. Moreover, we view as credit neutral structures in which servicers and indenture trustees will be held liable for actions that are negligent or involve willful misconduct and are prohibited from seeking indemnification from the collateral or trust estate or the noteholders for any negligence or willful 27 Securitizations usually include provisions that require a breach of a R&W to be remedied only when the breach materially and adversely affects the value of the student loan or the interests of the noteholders therein. 24 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

25 misconduct claims arising from the performance of their duties. Standards and degrees of liability that are less stringent than those described above will be considered by the rating committee and may be viewed as imposing additional credit risk. E. Servicer and Custodial Arrangements Moody s also considers the following factors with regard to the servicing and custodial arrangements for the securitized transferred assets (collateral):» Segregating the collateral (including limitations on the commingling of the collateral s cash flows and reinvestments) from other assets managed, controlled or owned by the originator, servicer and indenture trustee and their respective agents;» Limitations on such parties access to the collateral; and» Explicit guidelines on the fiduciary, agency or service provider terms associated with specific securitization party roles, including any securitization parties ability to delegate their responsibilities to agents, bailees or other parties not explicitly identified in the securitization documentation at closing (e.g., indenture trustee, master servicing, servicing, collateral agent and custodian responsibilities). Such factors have increased importance if the servicer or custodian of the receivables is related to the originator or seller, particularly in the case where a true sale is being relied on in Moody s bankruptcy analysis. Inadequate servicer and custodial provisions may weaken the securitization. F. Eligible Accounts and Investments As mentioned previously, securitizations typically allow the indenture trustee to place cash receipts in certain eligible accounts or invest them in eligible investments. Moody s has published a Request for Comment, Moody s Ratings Guidelines for Eligible Investments and Institutions Holding Trust Accounts for Transactions, October 17, 2011, addressing the temporary use of funds within a securitization structure. G. Events of Default Securitizations are typically structured so that the failure to pay a scheduled amount of interest or principal to the noteholders on a given distribution date is an event of default. Events of default generally include non-payment of rated interest and rated principal (e.g., ultimate principal payment at legal maturity and not unrated sinking fund or turbo redemptions), bankruptcy or insolvency of the issuer (but not any other securitization party), and breaches of covenants and R&Ws of the issuer (but not any other securitization parties). If the securitization contains broad covenants or R&Ws beyond the above mentioned scope, we consider whether 100% of the noteholders are required to vote to accelerate the rated debt and/or liquidate the collateral. This mechanism prevents market value risk caused by certain noteholders acting against the interests of the minority noteholders by deciding to accelerate the debt and liquidating the collateral below the par value of the rated securities. For example, to the extent that the securitization provides for other kinds of events of default, including servicer termination events or financial covenants relating to the sponsor, Moody s reviews whether such events could have a credit impact. 25 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

26 H. Amendments Moody s asks to receive prior written notice (at least ten business days) of any amendment, waiver or other modification to any securitization document, including a change in any securitization party (e.g., appointment of a successor to a securitization party role). If there is flexibility in the documents to amend, waive or otherwise modify material features of the securitization structure, such as increase the securitization expenses or borrower benefits offered to the loans, we incorporate those flexibilities into our ratings analysis. I. Provision of Surveillance Information We review the adequacy of the requirements in the documentation for the periodic reporting of key performance and other securitization data necessary for proper surveillance. XI. FFELP Student Loan ABS Surveillence A. Moody s Surveillance Process In monitoring FFELP student loan ABS, we assess whether they perform in line with our expectations at the time of closing. We continuously monitor key performance trends across rated transactions including parity levels, excess spread, delinquency, deferment and forbearance levels, pool factors, and funding costs. We also monitor any change in servicing and administration functions. We identify transactions whose performance deviates from our expectations for further review. In the process of the detailed review we perform one or more of the following actions: 1. We compare trust performance with either our original performance expectations or the performance of similar deals. 2. We assess the impact of any operational changes. 3. We seek additional insight into deal performance by discussing metrics outside of the expected range and any operational concerns with transaction sponsors. 4. We review all available information, including issuer feedback, and assess whether the negative or positive trends are passing or systemic. 30 Where applicable we perform cash flow analysis to determine the likelihood of investors not receiving timely interest or ultimate principal payments. In other cases we perform excess spread analysis and its impact on parity levels under expected and stressed scenarios. 5. Committee is assembled to conclude on the rating of the transaction. Issuers and trustees should provide adequate and timely reporting on all Moody s-rated FFELP student loan-backed securitizations. We review the adequacy of reported key performance information and other transaction data such as borrower benefit information and will request additional information from the issuers as necessary For example, if the observed change is concentrated in a specific loan type, the effect may be short-term if there are relatively few loans of that type remaining in the pool. 29 A rating action can be a decision to place certain classes of notes on review for possible downgrade or to upgrade or to downgrade, upgrade, or confirm the current ratings on the notes. 26 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

27 Appendix I Expected case and stress case Cash Flow Scenarios Appendix A summarizes the cash flow assumptions for the Aaa, Single-A, Baa and Expected case. FIGURE 1 Aaa Stress Scenarios Assumption Notes I II III IV V VI VII Interest Rate (T-Bill) Level Cumulative Default Rates Annual Voluntary Prepayment Rates (CPR) Use assumption for applicable loan type and school type Slow Repayment Stress 1% 5% 9% 1% 5% 9% Choose most stressful T-Bill scenario in IV-V1 Aaa stress case (see Appendix II.C) Consolidation loans 2% in year 1, rising linearly to 10% over the first 7 years, 10% thereafter Non-Consolidation 10% in year 1, rising linearly to 18% over the first 3 years, 18% thereafter Timing of defaults Consolidation loans 15/15/15/10/10/ 10/10/5/5/5 Interest Rate Spread Non-Consolidation loans (pre-07) Non-Consolidation loans (post-07) 40/20/15/10/5/ 5/5 20/15/15/10/10/ 10/10/10 Aaa spread from Table A Aaa stress case (see Appendix II.C) Expected case (see Appendix II.C) Comments Ramp up from the current T-Bill rate by adding 2.5% per year Dollar amount of defaulted loans as percent of the balance of loans in repayment, deferment and forbearance. Do not default capitalized interest from loans exiting deferment and forbearance 0% 0% Dollars of voluntary prepayments as a percent of the balance of total loans in repayment outstanding at the beginning of each period 0% 0% 15/15/15/10/10/1 0/10/5/5/5 40/20/15/10/5/5 /5 20/15/15/10/10/ 10/10/10 Aaa spread from Table A 15/15/15/10/10/10/10/5 /5/5 40/20/15/10/5/5/5 20/15/15/10/10/10/10/1 0 Aaa spread from Table A Percent of lifetime defaults incurred in each subsequent year Investment Rate Taxable T-Bill T-Bill T-Bill If there is a GIC, use the Tax-Exempt (TE) Lower of T-bill or Lower of T-bill or Lower of T-bill or TE actual contracted rate. TE bond rate TE bond rate bond rate Interest Rate spikes Aaa spike from Table B Aaa spike from Table B Aaa spike from Table B Net Claim Reject Rates 3.00% 3.00% 3.00% Apply the net reject rate to the relevant default percentage. Rejected loans should have 100% severity of losses. Payment Lags Default Reimbursement 450 days 450 days 450 days Number of days cash receipts are delayed after a specified event ISP/SAP 60 days 60 days 60 days No late fees coming in the trust. Net of ACH. 27 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

28 FIGURE 1 Aaa Stress Scenarios Assumption Notes I II III IV V VI VII Comments Forbearance Rates Non-Consolidation Loans 25% for 36 mos. 25% for 36 mos. 25% for 36 mos. Consolidation loans 25% for 36 mos. 25% for 36 mos. 20% for 60 mos. Deferment and forbearance Deferment Rates Non-Consolidation Loans 25% for 24 mos. 25% for 24 mos. 25% for 24 mos. are provided as a percentage of the loan balance in active Consolidation loans 25% for 24 mos. 25% for 24 mos. 20% for 60 mos. repayment, forbearance, and deferment status in the first period of the cash flows, and those loans remain in deferment or forbearance status for the period indicated. On-Time Borrower Benefits ACH after 0 months 50% 50% 50% Percent of outstanding balance of all eligible loans % remaining after 0 months % remaining after 12 months % remaining after 24 months % remaining after 36 months % remaining after 48 months 100% 75% 65% 55% 100% 75% 65% 55% 100% 75% 65% 55% Percent of outstanding balance of all eligible loans that were originally offered benefits. Assume a decline in the utilization rate to the level shown after each subsequent number of months in repayment 50% 50% 50% Servicing Fee Inflation 4.0% per annum 4.0% per annum 4.0% per annum Increase in servicing fee per year as a percent of the servicing fee at the beginning of each period Auction Rate Securities Rate Reset Notes Variable Rate Demand Bonds Failed auction rate for life Penalty rate for life Failed auction rate for life Penalty rate for life Failed auction rate for life Penalty rate for life VRDB Rate Taxable and Tax-Exempt Bank rate for life Bank rate for life Bank rate for life VRDB Spike Taxable LIBOR + 6% during months Tax-Exempt LIBOR + 4% during months LIBOR + 6% during months LIBOR + 4% during months LIBOR + 6% during months LIBOR + 4% during months JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

29 FIGURE 2 Single-A Stress Case Cash Flow Scenarios A Stress Scenarios Assumption Notes I II III IV V Interest Rate (T-Bill) Level Cumulative Default Rates Annual Voluntary Prepayment Rates (CPR) Use assumption for applicable loan type and school type Slow Repayment Stress 1% 5% 1% 5% Choose most stressful T-Bill scenario in III-IV A stress case (see Appendix II.C) Consolidation loans 2% in year 1, rising linearly to 8% over the first 7 years, 8% thereafter Non-Consolidation 10% in year 1, rising linearly to 13% over the first year, 13% thereafter Timing of defaults Consolidation loans 15/15/15/10/10 /10/10/5/5/5 Interest Rate Spread Non-Consolidation loans (pre-07) Non-Consolidation loans (post-07) 40/20/15/10/5 /5/5 20/15/15/10/10 /10/10/10 "A" spread from Table A A stress case (see Appendix II.C) Expected case (see Appendix II.C) Comments Ramp up from the current T- Bill rate by adding 2.5% per year Dollar amount of defaulted loans as percent of the balance of loans in repayment, deferment and forbearance. Do not default capitalized interest from loans exiting deferment and forbearance 0% 0% Dollars of voluntary prepayments as a percent of the balance of total loans in repayment outstanding at the beginning of each period 0% 0% 15/15/15/10/10/ 10/10/5/5/5 40/20/15/10/5/ 5/5 20/15/15/10/10/ 10/10/10 "A" spread from Table A 15/15/15/10/10/10/10/ 5/5/5 40/20/15/10/5/5/5 20/15/15/10/10/10/10/ 10 "A" spread from Table A Percent of total defaults incurred in each subsequent year Investment Rate Taxable T-Bill T-Bill T-Bill If there is a GIC, use the Tax-Exempt (TE) Lower of T-bill Lower of T-bill or Lower of T-bill or TE actual contracted rate or TE bond rate TE bond rate bond rate Interest Rate spikes "A" spike from Table B "A" spike from Table B "A" spike from Table B Net Claim Reject Rates 2.00% 2.00% 2.00% Apply the net reject rate to the relevant default percentage. Rejected loans should have 100% severity of losses. Payment Lags Default Reimbursement 450 days 450 days 450 days Number of days cash receipts are delayed after a specified event ISP/SAP 30 days 30 days 30 days No late fees coming in the trust. Net of ACH Forbearance Rates Non-Consolidation Loans 20% for 24 mos. 20% for 24 mos. 20% for 24 mos. Consolidation loans 20% for 24 mos. 20% for 24 mos. 20% for 60 mos. Deferment and forbearance Deferment Rates Non-Consolidation Loans 20% for 24 mos. 20% for 24 mos. 20% for 24 mos. are provided as a percentage of the loan balance in active Consolidation loans 20% for 24 mos. 20% for 24 mos. 20% for 60 mos. repayment, forbearance, and deferment status in the first 29 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

30 FIGURE 2 Single-A Stress Case Cash Flow Scenarios A Stress Scenarios Assumption Notes I II III IV V On-Time Borrower Benefits Slow Repayment Stress Comments period of the cash flows, and those loans remain in deferment or forbearance status for the period indicated. ACH after 0 months 35% 35% 35% Percent of outstanding balance of all eligible loans % remaining after 0 months % remaining after 12 months % remaining after 24 months % remaining after 36 months % remaining after 48 months 100% 60% 50% 45% 100% 60% 50% 45% 100% 60% 50% 45% Percent of outstanding balance of all eligible loans that were originally offered benefits. Assume a decline in the utilization rate to the level shown after each subsequent number of months in repayment 40% 40% 40% Servicing Fee Inflation 2.5% per annum 2.5% per annum 2.5% per annum Increase in servicing fee per year as a percent of the servicing fee at the beginning of each period Auction Rate Securities Rate Reset Notes Variable Rate Demand Bonds Failed auction rate for life Penalty rate for life Failed auction rate for life Penalty rate for life Failed auction rate for life Penalty rate for life VRDB Rate Taxable and Tax-Exempt Bank rate for life Bank rate for life Bank rate for life VRDB Spike Taxable None None None Tax-Exempt None None None 30 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

31 FIGURE 3 Expected Case and Baa Stress Case Cash Flow Scenarios Baa Stress Scenarios Assumption Notes Base Case I II III IV V Interest Rate (T- Bill) Level Cumulative Default Rates Annual Voluntary Prepayment Rates (CPR) Timing of defaults Interest Rate Spread Use assumption for applicable loan type and school type Consolidation loans Non- Consolidation Consolidation loans Non- Consolidation loans (pre-07) Non- Consolidation loans (post-07) Use 3 mo Forward LIBOR Curve -TED spread (from Table A) Expected case (see Appendix II.C) 0% in year 1, rising linearly to 8% over the first 10 years, 8% thereafter Slow Repayment Stress 1% 5% 1% 5% Choose most stressful T-Bill scenario in III-IV Baa stress case (see Appendix II.C) 1% in year 1, rising linearly to 8% over the first 8 years, 8% thereafter 10% 10% in year 1, rising linearly to 12% over the first year, 12% thereafter 15/15/15/10/10/ 10/10/5/5/5 40/20/15/10/5/ 5/5 20/15/15/10/10/ 10/10/10 Expected Case spread from Table A 15/15/15/10/10/10 /10/5/5/5 Baa stress case (see Appendix II.C) Expected case (see Appendix II.C) Comments Ramp up from the current T-Bill rate by adding 2.5% per year Dollar amount of defaulted loans as percent of the balance of loans in repayment, deferment and forbearance. Do not default capitalized interest from loans exiting deferment and forbearance 0% 0% 0% 0% Dollars of voluntary prepayments as a percent of the balance of total loans in repayment outstanding at the beginning of each period 15/15/15/10/10/10 /10/5/5/5 15/15/15/10/10/10/10/ 5/5/5 40/20/15/10/5/5/5 40/20/15/10/5/5/5 40/20/15/10/5/5/5 20/15/15/10/10/10 /10/10 Baa spread from Table A 20/15/15/10/10/10 /10/10 Baa spread from Table A 20/15/15/10/10/10/10/ 10 Baa spread from Table A Percent of total defaults incurred in each subsequent year Investment Rate Taxable T-Bill T-Bill T-Bill T-Bill If there is a GIC, use Tax-Exempt (TE) the actual contracted rate Interest Rate spikes Net Claim Reject Rates Payment Lags Default Reimbursement Lower of T-bill or TE bond rate Expected case spike from Table B Lower of T-bill or TE bond rate Baa spike from Table B Lower of T-bill or TE bond rate Baa spike from Table B Lower of T-bill or TE bond rate Baa spike from Table B 0.50% 1.00% 1.00% 1.00% Apply the net reject rate to the relevant default percentage. Rejected loans should have 100% severity of losses. 390 days 450 days 450 days 450 days Number of days cash receipts are delayed 31 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

32 FIGURE 3 Expected Case and Baa Stress Case Cash Flow Scenarios Baa Stress Scenarios Assumption Notes Base Case I II III IV V Forbearance Rates Deferment Rates On-Time Borrower Benefits Servicing Fee Inflation Auction Rate Securities Rate Reset Notes Variable Rate Demand Bonds VRDB Rate Slow Repayment Stress Comments after a specified event ISP/SAP 30 days 30 days 30 days 30 days No late fees coming in the trust. Net of ACH Non- Consolidation Loans Consolidation loans Non- Consolidation Loans Consolidation loans 0.25% ACH after 0 months % remaining after 0 months % remaining after 12 months % remaining after 24 months % remaining after 36 months % remaining after 48 months Taxable and Tax-Exempt 15% for 60 mos. 20% for 18 mos. 20% for 18 mos. 15% for 60 mos. Deferment and forbearance are provided as a 15% for 60 mos. 20% for 18 mos. 20% for 18 mos. 20% for 60 mos. 20% for 60 mos. 20% for 60 mos. 20% for 24 mos. 20% for 24 mos. 20% for 60 mos. 20% for 24 mos. 20% for 24 mos. 20% for 60 mos. percentage of the loan balance in active repayment, forbearance, and deferment status in the first period of the cash flows, and those loans remain in deferment or forbearance status for the period indicated. 20% 30% 30% 30% Percent of outstanding balance of all eligible loans 100% 50% 40% 35% 100% 55% 45% 40% 100% 55% 45% 40% 100% 55% 45% 40% Percent of outstanding balance of all eligible loans that were originally offered benefits. Assume a decline in the utilization rate to the level shown after each subsequent number of months in repayment 30% 35% 35% 35% Run according to servicing agreement Failed auction rate for life Penalty rate for life 2.5% per annum 2.5% per annum 2.5% per annum Increase in servicing fee per year as a percent of the servicing fee at the beginning of each period Failed auction rate for life Failed auction rate for life Failed auction rate for life Penalty rate for life Penalty rate for life Penalty rate for life Bank rate for life Bank rate for life Bank rate for life Bank rate for life VRDB Spike Taxable None None None None Tax-Exempt None None None None 32 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

33 TABLE A Interest Rate Spread Assumptions For securitizations with LIBOR-based liabilities only: If > 5% of loans have T-bill based SAP payments, use the following assumptions: Interest Rate Spreads Expected Case Baa A Aaa CP Rate T-Bill + 60bps T-Bill + 85 bps T-Bill bps T-Bill +125 bps Three-month LIBOR T-Bill + 70bps T-Bill bps T-Bill bps T-Bill bps One-month LIBOR (monthly reset liabilities) T-Bill + 60bps T-Bill + 88bps T-Bill + 117bps T-Bill + 135bps If < 5% of loans have T-bill based SAP payments, use the following assumptions: Interest Rate Spreads Expected Case Baa A Aaa CP Rate T-Bill + 45 bps T-Bill + 53 bps T-Bill + 53 bps T-Bill +75 bps Three-month LIBOR T-Bill + 55 bps T-Bill + 68 bps T-Bill + 73 bps T-Bill bps One-month LIBOR (monthly reset liabilities) T-Bill + 45 bps T-Bill + 56 bps T-Bill + 60 bps T-Bill + 85 bps For securitizations with other (or multiple) liability indexes: Interest Rate Spreads Expected Case Baa A Aaa CP Rate (with T-Bill (91day) Liabilities) CP Rate (with T-Bill (one yr) Liabilities) T-Bill + 65 bps T-Bill + 15 bps T-Bill + 10 bps T-Bill T-Bill + 60 bps T-Bill + 10 bps T-Bill + 5 bps T-Bill SIFMA/J.J. Kenny 70% of 3moLIBOR 70% of 3moLIBOR 75% of 3moLIBOR 85% of 3moLIBOR TABLE B Interest Rate Spike Assumptions Interest Rate Spikes Expected Case Baa Case A Case Aaa Case CP Rate T-Bill bps qtrs 1&2 of yr 1 T-Bill bps qtrs 3&4 of yr 1, qtrs 3&4 of yr 5, and for 2 qtrs at 10% pool factor T-Bill bps qtrs 3&4 of yr 1, qtrs 3&4 of yr 5, and for 2 qtrs at 10% pool factor T-Bill bps qtrs 3&4 of yr 1, qtrs 3&4 of yr 5, and for 2 qtrs at 10% pool factor Three-month LIBOR T-Bill bps qtrs 1&2 of yr 1 T-Bill bps qtrs 3&4 of yr 1, qtrs 3&4 of yr 5, and for 2 qtrs at 10% pool factor T-Bill bps qtrs 3&4 of yr 1, qtrs 3&4 of yr 5, and for 2 qtrs at 10% pool factor T-Bill bps qtrs 3&4 of yr 1, qtrs 3&4 of yr 5, and for 2 qtrs at 10% pool factor CP Rate (with T-Bill (one yr) Liabilities) T-Bill - 60 bps qtrs 1&2 of yr 1 T-Bill - 70 bps qtrs 1&2 of yr 1 T-Bill - 80 bps qtrs 1&2 of yr 1 T-Bill bps qtrs 1&2 of yr 1 One-month LIBOR (monthly reset liabilities) T-Bill bps qtrs 1&2 of yr 1 T-Bill bps qtrs 3&4 of yr 1, qtrs 3&4 of yr 5, and for 2 qtrs at 10% pool factor T-Bill bps qtrs 3&4 of yr 1, qtrs 3&4 of yr 5, and for 2 qtrs at 10% pool factor T-Bill bps qtrs 3&4 of yr 1, qtrs 3&4 of yr 5, and for 2 qtrs at 10% pool factor 33 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

34 Appendix II - Cash Flow Assumptions A. Basis Risk TABLE 1 1. Three-month LIBOR-Based Liabilities with 90-day CP-Based Assets Spread Assumptions Expected Case Baa Stress Case A Stress Case Aaa Stress Case Three-Month LIBOR minus 90-day CP 10bps 15bps 20bps 25bps Spike Assumptions* Three-Month LIBOR minus 90-day CP 75bps 100bps 130bps 150bps * In the expected case, the spikes occur in the first six months of the first year. For the Baa, A and Aaa stress cases, the spikes occur in months 6-12 of the first and fifth year and for the first six months after the pool factor reaches 10%. TABLE 2 2. One-month LIBOR-Based Liabilities with 90-day CP-Based Assets Spread Assumptions Expected Case Baa Stress Case A Stress Case Aaa Stress Case One-Month LIBOR minus 90-day CP 0bps 3bps 7bps 10bps Spike Assumptions* One-Month LIBOR minus 90-day CP 50bps 60bps 80bps 100bps * In the expected case, the spikes occur in the first six months of the first year. For the Baa, A and Aaa stress cases, the spikes occur in months 6-12 of the first and fifth year and for the first six months after the pool factor reaches 10%. We expect basis risk between one-month LIBOR and 90-day CP to be lower than basis risk between three-month LIBOR and 90-day CP because (1) one-month LIBOR has tracked more closely to 90- day CP historically than three-month LIBOR has and (2) the lag between the setting of the rate on the loans and the setting of the rate on the bonds is reduced. TABLE 3 3. Three-month LIBOR-Based Liabilities with 91-day T-Bill-Based Assets Spread Assumptions Expected Case Baa Stress Case A Stress Case Aaa Stress Case Three month LIBOR minus 91-day T-bill 70 bps 100 bps 130 bps 150 bps Spike Assumptions* Three month LIBOR minus 91-day T-bill 190 bps 230 bps 260 bps 330 bps * In the expected case, the spikes occur in the first six months of the first year. For the Baa, A and Aaa stress cases, the spikes occur in months 6-12 of the first and fifth year and for the first six months after the pool factor reaches 10%. 34 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

35 Our three-month LIBOR minus 91-day T-Bill basis risk assumptions reflect the lower historical spread and the less frequent and smaller spikes between three-month LIBOR and 90-day CP, compared to the spread between three-month LIBOR and 91-day T-Bill. TABLE 4 4. One-month LIBOR-Based Liabilities with 91-day T-Bill-Based Assets Spread Assumptions Expected Case Baa Stress Case A Stress Case Aaa Stress Case One month LIBOR minus 91-day T-bill 60bps 88bps 117bps 135bps Spike Assumptions* One month LIBOR minus 91-day T-bill 165bps 190bps 210bps 280bps * In the expected case, the spikes occur in the first six months of the first year. For the Baa, A and Aaa stress cases, the spikes occur in months 6-12 of the first and fifth year and for the first six months after the pool factor reaches 10%. Our one-month LIBOR minus 91-day T-Bill basis risk assumptions reflect the lower historical spread and the less frequent and smaller spikes between one-month LIBOR and 91-day T-Bill, compared to the spread between three-month LIBOR and 91-day T-Bill. TABLE 5 5. One-year Average of T-Bill-Based Liabilities with 90-day CP-Based Assets Spread Assumptions Expected Case Baa Stress Case A Stress Case Aaa Stress Case One-year average of T-bill minus 90-day CP -60 bps -10 bps -5 bps 0 bps Spike Assumptions* One-year average of T-bill minus 90-day CP 60 bps 70 bps 80 bps 100 bps * The spikes are applied in the first six months of the first year for all rating categories. Some auction rate trusts have been exposed to basis risk between one-year average T-Bill and either 90-day CP or 91-day T-Bill as a result of the failure of the auction rate market beginning in early Those auction rate trusts have been paying the failed auction rate as specified in the securitization documents, which is sometimes based on the one-year average T Bill rate or the 91-day T-Bill. TABLE day T-Bill-Based Liabilities with 90-day CP-Based Assets Spread Assumptions Expected Case Baa Stress Case A Stress Case Aaa Stress Case 91-day T-bill minus 90-day CP -65 bps -15 bps -10 bps 0 bps Spike Assumptions 91-day T-bill minus 90-day CP None None None None 35 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

36 Our assumptions for 91-day T-bill-based liabilities apply to trusts that use one of the following T- Bill rates in the failed auction rate definition: (1) a spot rate of 91-day T-Bill or (2) a quarterly average of 91-day T-Bill. The stresses in these assumptions are less onerous than for the spread between oneyear average T-bill and 90-day CP. Our assumptions reflect the lower historical average of the 91-day T-bill rate compared to the one-year average T-bill rate and less volatility in the spread between 91-day T-bill and 90-day CP than that of the spread between one-year average T-bill and 90-day CP. TABLE 7 7. SIFMA-Based Liabilities with 90-day CP-Based Assets or 91-day T-Bill Based Assets SIFMA Index (as a percentage of three-month LIBOR)* Expected Case Baa Stress Case A Stress Case Aaa Stress Case 70% of 3-month LIBOR 70% of 3-month LIBOR 75% of 3-month LIBOR 85% of 3-month LIBOR * Our assumptions for the SIFMA index also apply to the J.J. Kenny/S&P High Grade Weekly index, which is comparable to SIFMA. Some auction rate trusts with tax-exempt ARS have been exposed to basis risk between the SIFMA index and either 90-day CP or 91-day T-Bill as a result of the failure of the auction rate market. Our interest rate assumptions for liabilities based on the SIFMA index are presented as a percentage of three-month LIBOR. B. Interest Rate Level Scenarios* TABLE 7 T-Bill Scenarios Expected Case Baa Stress Case A Stress Case Aaa Stress Case 1 Use 3 month Forward LIBOR Curve -TED spread(**) 1% 1% 1% 2 5% 5% 5% 3 9% * For the stress case assumptions, increase the current T-bill rate by 2.5% per year until the assumption is reached. The rate remains constant at the assumption thereafter. ** The TED spread refers to the spread between T-Bill and LIBOR which can be found for the relevant asset-liability combination in Appendix I Table A. For securitizations that are very sensitive to low interest rate scenarios, we run an additional low interest rate scenario that we disclose at the time of issuance. C. Cumulative Default Rates Historical default performance across various issuers consistently shows that defaults vary in magnitude, depending on the year loans enter repayment, loan-type and school-type. Because all FFELP loans are originated under the DOE guidelines with little to no underwriting criteria, if issuer-specific default information is not available, we apply our standard expected case cumulative default rate assumptions for unseasoned pools broken-out by the three key drivers of default performance -- the year loans enter repayment, loan-type and school-type (see Exhibit 1). Our standard expected case assumptions for unseasoned pools are derived from available FFELP student loan default industry data. Exhibit 1 shows our standard expected case cumulative default rate assumptions for unseasoned nonconsolidation FFELP loans and consolidation FFELP loans by the year loans enter repayment, loantype and school-type. We adjust the standard expected case unseasoned cumulative default rates for 36 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

37 loan seasoning (i.e., the number of months the loan pool has been in repayment status). See Appendix II.M.1 and Appendix III for how we adjust the cumulative default rates for the seasoning of the pool. EXHIBIT 1 Cumulative Default Rate Assumptions, Unseasoned Non-Consolidation FFELP Loans* (Standard Expected Case)** School Type*** Loan Type Repayment Year Four-Year Two-Year Prop Stafford Loans < % 22% 35% >= % 30% 45% PLUS Loans**** < % 11% 11% >= % 15% 15% SLS 18% 40% 45% * The cumulative default rate assumptions apply to unseasoned pools. See Appendix II.M.1 and Appendix III for how we adjust the cumulative default rates for the seasoning of the pool. ** Dollar amount of defaulted loans as percent of the balance of loans in repayment, deferment and forbearance status. Do not default capitalized interest from loans exiting deferment and forbearance. *** The school type "Four-year" consists of public and private four-year institutions and graduate/medical schools. The school type "Two-year" consists of public and private junior and community colleges. The school type "Prop" consists of proprietary and vocational schools, which are for-profit institutions. **** Some PLUS loans are graduate PLUS loans. Graduate PLUS loans were first originated after July 1, 2006, so there is very little performance data on them, and therefore, we apply the higher default assumptions of Stafford Four-year loans. Cumulative Default Rate Assumptions, Unseasoned Consolidation FFELP Loans * (Standard Expected Case)** School Type*** Loan Type Repayment Year All Consolidation Loans < % >= % * The cumulative default rate assumptions apply to unseasoned pools. See Appendix II.M.1 and Appendix III for how we adjust the cumulative default rates for the seasoning of the pool. ** Dollar amount of defaulted loans as percent of the balance of loans in repayment, deferment and forbearance status. Do not default capitalized interest from loans exiting deferment and forbearance. *** The majority of FFELP servicers do not provide loan balances by school type for consolidation loans so we provide one default assumption across the various school types. 37 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

38 To arrive at our cumulative default rate stress assumptions for each rating category, we apply a multiple to the expected case cumulative default rate adjusted for seasoning (see Exhibit 2). The multiples vary with the magnitude of cumulative default rates and the seasoning of the loan. For A and Baa Cumulative Defaults, the multiple also varies based on the size of the subordinate class. EXHIBIT 2 Cumulative Default Rate Multiple Tables Non-Consolidation Aaa Stress Case Cumulative Default Rate Multiple Table* Expected Case Seasoned Cumulative Default Rate Weighted Average Months in Repayment Equal to or less than: >= 37 5% % % % % NA Non-Consolidation A Stress Case Cumulative Default Rate Multiple Table* Expected Case Seasoned Subordinate Class < 10% Subordinate Class > 10% Cumulative Default Rate Weighted Average Months in Repayment Weighted Average Months in Repayment Equal to or less than: >= >= 37 5% % % % % NA NA Non-Consolidation Baa Stress Case Cumulative Default Rate Multiple Table* Expected Case Seasoned Subordinate Class < 10% Subordinate Class > 10% Cumulative Default Rate Weighted Average Months in Repayment Weighted Average Months in Repayment Equal to or less than: >= >= 37 5% % % % % NA NA Consolidation Aaa Stress Case Cumulative Default Rate Multiple Table* Expected Case Seasoned Cumulative Default Rate Weighted Average Months in Repayment Equal to or less than: >= 37 5% NA NA % NA % % JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

39 Consolidation A Stress Case Cumulative Default Rate Multiple Table* Expected Case Seasoned Subordinate Class < 10% Subordinate Class > 10% Cumulative Default Rate Weighted Average Months in Repayment Weighted Average Months in Repayment Equal to or less than: >= >= 37 5% NA NA 2.30 NA NA % NA NA % % Consolidation Baa Stress Case Cumulative Default Rate Multiple Table* Expected Case Seasoned Subordinate Class < 10% Subordinate Class > 10% Cumulative Default Rate Weighted Average Months in Repayment Weighted Average Months in Repayment Equal to or less than: >= >= 37 5% NA NA 1.85 NA NA % NA NA % % * Apply the relevant multiple to the expected case seasoned cumulative default rate. EXHIBIT 3 Cumulative Default Rate Assumptions, Rehabilitated FFELP Loans* Rating Category Cumulative Default Rate Expected Case 40% A Stress Case 60% Aaa Stress Case 75% Slow Repayment Stress 30% * Dollar amount of defaulted loans as percent of the balance of loans in repayment, deferment and forbearance status. Do not default capitalized interest from loans exiting deferment and forbearance. D. Voluntary Prepayment Rates Our voluntary prepayment rate assumptions are based on recent prepayment rate performance data and our outlook on the economic environment. The table below provides our voluntary prepayment rate stress assumptions for each rating level, broken out by loan type. 39 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

40 Annual Voluntary Prepayment Rate Assumptions*, ** Expected Case Non-Consolidation 10% Consolidation 0% in year 1, rising linearly to 8% over the first 10 years, 8% thereafter Baa Stress Case Non-Consolidation Consolidation A Stress Case Non-Consolidation Consolidation Aaa Stress Case Non-Consolidation Consolidation (1) 0%; (2) 10% in year 1, rising linearly to 12% over the first year, 12% thereafter 1) 0%; (2) 1% in year 1, rising linearly to 8% over the first 8years, 8% thereafter (1) 0%; (2) 10% in year 1, rising linearly to 13% over the first year, 13% thereafter 1) 0%; (2) 2% in year 1, rising linearly to 8% over the first 7 years, 8% thereafter (1) 0%; (2) 10% in year 1, rising linearly to 18% over the first 3 years, 18% thereafter (1) 0%; (2) 2% in year 1, rising linearly to 10% over the first 7 years, 10% thereafter Slow Repayment Stress Non-Consolidation 0% Consolidation 0% * The voluntary prepayment rate assumptions apply to unseasoned pools. See Appendix II.M for how we adjust the assumptions for the seasoning of the pool. ** Dollars of voluntary prepayments as a percent of the balance of total loans in repayment outstanding at the beginning of each period. E. Net Claim Reject Rates The table below shows our current net claim reject rate assumptions, based on historical performance of net claim reject rates of the FFELP student loan servicing industry. Net Claim Reject Rate Assumptions* Net Reject Rates Expected Case % Baa Stress Case % A Stress Case % Aaa Stress Case % * Apply the net reject rate to the relevant default rate. The rejected defaults will have 100% severity of losses. The non-rejected defaults will have losses equivalent to 100% minus the guarantee percentage. See footnote 1 for details on the guarantee percentage. 40 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

41 F. Timing of Defaults EXHIBIT 1 Standard Default Timing Curve* Assumptions (Expected Case and All Rating Levels)** Loan Type Repayment Year Repayment Month Consolidation All 15% 15% 15% 10% 10% 10% 10% 5% 5% 5% Non-Consolidation < % 20% 15% 10% 5% 5% 5% 0% 0% 0% >= % 15% 15% 10% 10% 10% 10% 10% 0% 0% * Percent of lifetime defaults incurred in each subsequent year. ** All loans are subject to default immediately except those that are in school or grace status. Once loans exit school or grace status they are subject to default. EXHIBIT 2 Default Timing Curve Assumptions* for Rehabilitated FFELP Loans** Repayment Month Rating Level Expected Case 30% 20% 10% 10% 10% 10% 10% A Stress Case 40% 20% 10% 10% 10% 10% Aaa Stress Case Curve 1 70% 20% 10% Curve 2 30% 20% 10% 10% 10% 10% 10% Slow Repayment Stress Expected Case Default Timing Curve * Percent of lifetime defaults incurred in each subsequent year of repayment. ** All loans are subject to default immediately except those that are in school or grace status. Once loans exit school or grace status they are subject to default. G. Payment Lags Payment Lag Assumptions Expected Case Baa Stress Case A Stress Case Aaa Stress Case Default Reimbursement Lag 390 days 450 days 450 days 450 days SAP/ISP Payment Lag 30 days 30 days 30 days 60 days 41 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

42 H. Deferment and Forbearance Rates The assumptions for unseasoned pools are shown in the table below: Deferment and Forbearance Rate Assumptions*, **, *** Loan Type Expected Case Baa Stress Case A Stress Case Aaa Stress Case Forbearance Rate Non-Consolidation 15% for 60 months 20% for 18 months 20% for 24 months 25% for 36 months Consolidation 15% for 60 months 20% for 18 months 20% for 24 months 25% for 36 months Deferment Rate Non-Consolidation 20% for 60 months 20% for 24 months 20% for 24 months 25% for 24 months Consolidation 20% for 60 months 20% for 24 months 20% for 24 months 25% for 24 months * For the slow repayment stress, for non-consolidation loans use the forbearance and deferment assumptions shown in the above table for the appropriate rating category; for consolidation loans use 20% for 60 months for both deferment and forbearance for all rating categories. ** In the expected case and slow repayment rate stress, forbearance and deferment rate assumptions are provided as a percentage of the current loan balance in active repayment, forbearance, and deferment status each period. For the stress cases, deferment and forbearance are provided as a percentage of the loan balance in active repayment, forbearance, and deferment status in the first period of the cash flows, and those loans remain in deferment or forbearance status for the period indicated. *** Deferment and forbearance rates and duration can vary among securitizations due to differences in loan seasoning, loan type, and servicing policies. Therefore, the deferment and forbearance assumptions may be adjusted by the rating committee for a particular securitization. I. Borrower Benefit Utilization Rates EXHIBIT 1 ACH Benefit Utilization Rate Assumptions* Expected Case Baa Stress Case A Stress Case Aaa Stress Case ACH benefit 20% 30% 35% 50% * Percent of outstanding balance of all eligible loans. EXHIBIT 2 On-Time Borrower Benefit Utilization Rate Assumptions Based on Current Balance* Months in Active Repayment** Expected Case Baa Stress Case A Stress Case Aaa Stress Case 0 months 100% 100% 100% 100% 12 months 50% 55% 60% 75% 24 months 40% 45% 50% 65% 36 months 35% 40% 45% 55% 48 months 30% 35% 40% 50% * After loans that are eligible to receive a particular borrower benefit reach the specified time in active repayment, the corresponding utilization rate is applied to the eligible loan balances outstanding at that point in time. As an example of interpreting the numbers in Exhibit 2, in the Expected Case, if loans are eligible to receive an interest rate reduction after making 24 consecutive monthly payments, then after the loans have reached 24 months in active repayment, a 40% utilization rate would be applied to the those outstanding loan balances at that point in time. 42 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

43 J. Servicing Fee Inflation Servicing Fee Inflation Assumptions*, ** Expected Case Baa Stress Case A Stress Case Aaa Stress Case Inflate according to servicing agreement 2.5% per annum 2.5% per annum 4% per annum * Increase in servicing fee per year as a percent of the servicing fee at the beginning of each period. ** The cash flows are modeled with the higher of (i) the inflation assumptions above and (ii) the rate specified in the servicing agreements. K. Investment Rate on Cash Accounts We assume that the tax-exempt investment rate is the tax-exempt bond interest rate to reflect the fact that any yield above the tax-exempt bond rate would not be retained by the securitization, since it would have to be rebated. The taxable investment rate assumption is the risk-free rate of T-Bill. Investment Rate on Cash Accounts Assumptions* Expected Case Baa Stress Case A Stress Case Aaa Stress Case Taxable Bonds T-Bill T-Bill T-Bill T-Bill Tax-Exempt (TE) Bonds Lower of T-Bill or TE Rate Lower of T-Bill or TE Rate Lower of T-Bill or TE Rate Lower of T-Bill or TE Rate * In securitizations with a Guaranteed Investment Contract (GIC), our assumptions will assume the GIC rate, provided, that the GIC meets Moody s criteria. L. Remarketing or Reset Rate Risk 1. Interest Rate for Variable Rate Demand Bonds (VRDBs) Interest Rate Assumptions for Variable Rate Demand Bonds (VRDBs) Expected Case Baa Stress Case A Stress Case Aaa Stress Case Tax-Exempt VRDB Spread Bank rate for life Bank rate for life Bank rate for life Bank rate for life Tax-Exempt VRDB Spike NA NA NA LIBOR + 4% during months Taxable VRDB Spread Bank rate for life Bank rate for life Bank rate for life Bank rate for life Taxable VRDB Spike NA NA NA LIBOR + 6% during months Interest Rate for Auction Rate Securities We assume that auction rate securities will bear interest at the failed auction rate, as specified in the legal documents, for life. Interest Rate Assumptions for Auction Rate Securities Auction Rate Securities Expected Case Baa Stress Case A Stress Case Aaa Stress Case Failed Auction Rate for Life Failed Auction Rate for Life Failed Auction Rate for Life Failed Auction Rate for Life 43 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

44 Interest Rate for Reset Rate Notes We run the fixed interest rate until the bond is remarketed and then the penalty rate, as specified in the legal documents, is applied. Our assumptions assume a successful remarketing will not occur at a spread to LIBOR lower than the penalty rate. 32 Interest Rate Assumptions for Reset Rate Notes Reset Rate Notes Expected Case Baa Stress Case A Stress Case Aaa Stress Case Penalty Rate for Life Penalty Rate for Life Penalty Rate for Life Penalty Rate for Life M. Adjustments for Seasoned Pools 1. Expected Case Cumulative Default Rates for Seasoned Pools Our cumulative default rate model adjusts the unseasoned expected case cumulative default rate assumptions for seasoning (i.e., the number of months the loan pool has been in repayment status). We provide a hypothetical example and an explanation of the cumulative default rate model in Appendix III. 2. Expected Case Default Timing Curves for Seasoned Pools The table below shows the expected case default timing curve assumptions for loan pools by seasoning (months in repayment). The assumptions for unseasoned loan pools (i.e., 0-12 months of seasoning) are the same as those shown previously in Appendix II.F. Default Timing Curve Assumptions for Non-Consolidation Loans by Years of Seasoning (Expected Case and all Rating Levels) Loan Type Consolidation Repayment Year All Months in Repayment Repayment Month (unseasoned) 15% 15% 15% 10% 10% 10% 10% 5% 5% 5% % 20% 10% 10% 10% 10% 10% 5% 5% % 15% 15% 15% 15% 10% 5% 5% % 20% 15% 15% 10% 10% 10% 49+ months 25% 20% 20% 15% 10% 10% Non- Consolidation < (unseasoned) 40% 20% 15% 10% 5% 5% 5% % 25% 15% 10% 10% 5% % 25% 15% 10% 10% % 20% 20% 20% 49+ months 35% 35% 30% Non- Consolidation >= (unseasoned) 20% 15% 15% 10% 10% 10% 10% 10% % 20% 15% 15% 10% 10% 10% % 15% 15% 15% 15% 15% % 20% 20% 20% 20% 49+ months 25% 25% 25% 25% 30 To the extent a rate reset note may be remarketed above the pre-set penalty rate we will consider the likelihood of this occurrence and adjust our stresses accordingly. 44 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

45 3. Non-Consolidation Loan Voluntary Prepayment Rates for Seasoned Pools Non-Consolidation Loan Voluntary Prepayment Rate Assumptions by Months in Repayment Months in Repayment Aaa Stress Case 0-24 (unseasoned) 10% in year 1, rising linearly to 18% over the first 3 years, 18% thereafter % in year 1, rising linearly to 18% over the first 2 years, 18% thereafter % in year 1, rising linearly to 18% over the first 1 year, 18% thereafter Months in Repayment A Stress Case 0-24 (unseasoned) 10% in year 1, rising linearly to 13% over the first 1 year, 13% thereafter % for life % for life Months in Repayment Baa Stress Case 0-24 (unseasoned) 10% in year 1, rising linearly to 12% over the first 1 year, 12% thereafter % for life % for life Months in Repayment any Expected Case 10% for life 4. Consolidation Loan Voluntary Prepayment Rates for Seasoned Pools The tables below shows the consolidation loan voluntary prepayment rate assumptions by seasoning (months in repayment) for each rating category. The assumptions for unseasoned consolidation loan pools (i.e., 0-12 months of seasoning) are the same as those shown previously in Appendix II.D. Consolidation Loan Voluntary Prepayment Rate Assumptions by Months in Repayment Months in Repayment Aaa Stress Case 0-24 (unseasoned) 2% in year 1, rising linearly to 10% over the first 7 years, 10% thereafter % in year 1, rising linearly to 10% over the first 5 years, 10% thereafter 49+ 4% in year 1, rising linearly to 10% over the first 4 years, 10% thereafter Months in Repayment A Stress Case 0-24 (unseasoned) 2% in year 1, rising linearly to 8% over the first 7 years, 8% thereafter % in year 1, rising linearly to 8% over the first 5 years, 8% thereafter 49+ 4% in year 1, rising linearly to 8% over the first 4 years, 8% thereafter Months in Repayment Baa Stress Case 0-24 (unseasoned) 1% in year 1, rising linearly to 8% over the first 8 years, 8% thereafter % in year 1, rising linearly to 8% over the first 5 years, 8% thereafter 49+ 4% in year 1, rising linearly to 8% over the first 4 years, 8% thereafter Months in Repayment Expected Case 0-24 (unseasoned) 0% in year 1, rising linearly to 8% over the first 10 years, 8% thereafter % in year 1, rising linearly to 8% over the first 6 years, 8% thereafter 49+ 4% in year 1, rising linearly to 8% over the first 5 years, 8% thereafter 45 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

46 5. On-Time Borrower Benefit Utilization Rates for Seasoned Pools The table below shows the on-time borrower benefit utilization rate assumptions by seasoning for each rating category. The assumptions for unseasoned loan pools (i.e., 0-11 months of seasoning) are the same as those shown previously in Appendix II.I. These assumptions reflect the same rate of disqualification as the unseasoned assumptions, but are adjusted to account for the pool having aged past at least the first year of active repayment (that is, having made at least 12 months of payments). 33 Seasoned Assumptions for On-Time Borrower Benefit Utilization Rates* Based on Current Balance Months in Active Repayment Expected Case Baa Stress Case A Stress Case Aaa Stress Case Unseasoned (0-11 Months of Payments Made**) 0 months 100% 100% 100% 100% 12 months 50% 55% 60% 75% 24 months 40% 45% 50% 65% 36 months 35% 40% 45% 55% 48 months 30% 35% 40% 50% Months of Payments Made** 0 months 100.0% 100.0% 100.0% 100.0% 12 months 80.0% 82.5% 85.0% 90.0% 24 months 70.0% 72.5% 75.0% 80.0% 36 months 60.0% 65.0% 67.5% 70.0% Months of Payments Made** 0 months 100.0% 100.0% 100.0% 100.0% 12 months 87.5% 90.0% 90.0% 92.5% 24 months 75.0% 77.5% 80.0% 82.5% 36+ Months of Payments Made** 0 months 100.0% 100.0% 100.0% 100.0% 12 months 90.0% 92.5% 92.5% 95.0% * After each loan that is eligible to receive a particular borrower benefit reaches the specified time in active repayment, the corresponding utilization rate is applied to the eligible loan balances outstanding at that point in time. ** Months of Payments Made is defined as the weighted average months of payments made for those loans still eligible to receive on-time borrower benefits. 31 In seasoned pools, our definition of utilization is slightly different than for unseasoned pools. The denominator contains loans that were still eligible for benefits at the beginning of the securitization and who are eligible to receive the benefits, instead of loans that were originally offered benefits and are eligible. 46 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

47 Appendix III Cumulative Default Rate Model To calculate a seasoned pool s expected case cumulative default rate and the cumulative default rates we assume for various rating levels, we use the following methodology: 1) Calculate the pool s expected case unseasoned cumulative default rate as described in Appendix II.C. See Box I below for a Consolidation deal example. 2) Determine the seasoning distribution of the pool based on years in repayment (Box II. Column B). For each seasoning bucket, calculate the seasoned default rate (Column E) which is equal to expected case unseasoned cumulative default rate (from step 1, above) * percentage of lifetime defaults remaining (Column C) / pool factor of seasoning bucket (Column D)). The percentage of lifetime defaults remaining (Column C) is based on our standard (consolidation or non-consolidation) expected case default timing curve and the pool factor (Column D) is based on our standard expected case (consolidation or non-consolidation) amortization curve). 3) Calculate the pool s expected case seasoned cumulative default rate. The pool s expected case seasoned cumulative default rate is a weighted average, using each buckets expected case seasoned cumulative default rates (Column E), weighted by its portion of the pool (Column B). For the example in Box II, the seasoned pool expected cumulative default rate is 9.5%. 4) Calculate the pool s stress case cumulative default rates. To calculate the stress case cumulative default rates, determine the proper multiple for each stress case as found in Appendix II.C. Apply each multiple to the seasoned expected cumulative default rate to determine the default rate. See Box III. The figure below shows an example of the calculations for a consolidation loan pool for demonstration purposes, where we have assumed: (1) 50% of the pool is comprised of consolidation loans that entered repayment before 2007 and 50% is comprised of consolidation loans that entered repayment n or after 2007; (2) 25% of the pool is seasoned two years, 25% is seasoned three years, and 50% is seasoned four years, (3) our standard expected case default timing curve and amortization assumptions for consolidation loans (columns C and D in Box II) and; (4) our standard assumptions for rating multiples (Box III). In the example, our assumptions are in blue and the securitization characteristics are in red. 47 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

48 A B C D E Box I: Calculating the Unseasoned Pool's Expected Default Rate Loan Type Portion of Pool Expected Default Assumption Consolidation, < % 13% Consolidation, >= % 18% 15.5% Calculated expected default rate for unseasoned pool Box II: Adjusting the Pool's Expected Default Rate for Seasoning Years in Repayment Distribution of Pool (100% total) Expected Loss Curve (Assumed % of Lifetime Losses Remaining) Assumed Amortization Curve (Pool Factor) Calculated Bucket Default Rate, as % of original balance 0 0% 100% 100% 15.5% 1 0% 85% 96% 13.8% 2 25% 70% 92% 11.8% 3 25% 55% 90% 9.5% 4 50% 45% 79% 8.8% 5 0% 35% 73% 7.5% 6 0% 25% 68% 5.7% 7 0% 15% 60% 3.9% 8 0% 10% 58% 2.7% 9 0% 5% 49% 1.6% 9.5% Calculated expected default rate for seasoned pool Box III: Calculating the Assumed Default Rates for Different Rating Levels* Multiple Calculated Rating Level Default Rates Expected % Baa % A % Aaa % *Assumes size of subordinate bond is less than 10% For a non-consolidation loan pool, we would perform the same calculations but use assumptions (shown above in blue text) specific to the loan type, as listed below. For securitizations that contain non-consolidation loans from both pre-2007 and repayment vintages, we expect to provide assumptions separately because those vintages have different assumptions. However, at committee s discretion we may provide one assumption reflecting the weighted average of the entire nonconsolidation loan pool. Box I: Expected Default Assumption Please refer to Appendix II.C. Exhibit I Box II: Expected Loss Curve and Assumed Amortization Curve 48 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

49 Repayment Vintage Pre-2007 Repayment Vintage Post-2007 Years in Repayment Expected Loss Curve (Assumed % of Lifetime Losses Remaining) Assumed Amortization Curve (Pool Factor) Expected Loss Curve (Assumed % of Lifetime Losses Remaining) Assumed Amortization Curve (Pool Factor) 0 100% 100% 100% 100% 1 60% 55% 80% 80% 2 40% 33% 65% 71% 3 25% 20% 50% 58% 4 15% 15% 40% 31% 5 10% 12% 30% 20% 6 5% 9% 20% 15% 7 0% 6% 10% 12% 8 0% 4% 0% 9% 9 0% 4% 0% 6% Box III: Multiple Please refer to Appendix II.C. Exhibit 2 49 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

50 Moody s Related Research This methodology can, as appropriate, be more precisely explained by the following guidance on specific cases:» Moody s Approach to Rating Structured Finance Securities in Default, November 2009 (121070). This Rating Implementation Guidance governs the ratings treatment of defaulted structured finance securities. The methodology provides guidance on ratings on securities in default or about to default and is primarily a function of expected recovery rates. Expected recovery rates are calculated by comparing our assessment of likely future cash flows against the cash flows promised to investors. The shortfall is then divided by the bond s outstanding principal balance at the time of the calculation.» Moody s Ratings Guidelines for Eligible Investments and Institutions Holding Trust Accounts for Transactions, October 17, 2011 (SF262353). This Request for Comment proposes to change some existing guidelines outlined in the Rating Implementation Guidance The Temporary Use of Cash In Structured Transactions: Eligible Investment Guidelines, December 2008 (SF149666). We propose clarifications to eligible investment standards for structured finance securities and a minimum ratings threshold of Baa3 for trust account providers that hold a securitization s non-cash property in trust accounts.» The Temporary Use of Cash In Structured Transactions: Eligible Investment Guidelines, December 2008 (SF149666). This Rating Implementation Guidance provides general guidelines for temporary investment of cash within structured finance securitizations. The level of minimum long and/or short term ratings for those eligible investments vary depending on the maximum maturity of investments and the rating level of the structured finance securities for which the cash is used to make payments.» Moody's Approach to Evaluating Distressed Exchanges, March 2009 (115337). This Rating Implementation Guidance provides a description under which an exchange of securities by an issuer would be deemed to be a default. Generally, a situation in which an issuer offers a reduced economic package to investors in order to mitigate a default would be characterized as a distressed exchange and the execution of such agreement would be viewed as a default by the issuer under the promise.» Updated Approach to the Usage of Credit Estimates in Rated Transactions, October 2009 (120461). This Rating Implementation Guidance defines Credit Estimates (CE) and explains the conditions for their use in rated securitizations. The report provides a set of expected case notching assumptions and stress sensitivity tests to ensure that ratings are not overly exposed to the volatility of CEs compared to traditional Moody s credit ratings. For example, CEs on exposures above 3% in a pool (up to a total of 30% of the pool) should be notched down by 2 alphanumeric rating categories as the expected case scenario in rating analysis. Further, a jump-todefault stress test are performed on the two largest CEs (although not simultaneously) to ensure senior class ratings do not move by amounts exceeding certain levels.» Framework for De-Linking Hedge Counterparty Risks from Global Structured Finance Cashflow Transactions, October 2010 (SF73248). This Rating Implementation Guidance sets out a framework for determining whether a securitization rating can be de-linked from a hedge counterparty s ratings if an issuer chooses to aim for such de-linkage. It only applies to standard interest rate and foreign exchange swaps and not to other types of swaps such as credit default swaps or complex swaps that have no secondary market. 50 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

51 » The framework sets out the rating levels for a counterparty to provide a hedge to a cash-flow securitization without posting collateral and provides triggers that would cause the counterparty to take action to further de-link its credit risk. It also provides some guidelines for application and modification of standard ISDA events of default and termination events as well as collateral amounts and types to be posted by a counterparty with lower ratings.» Global Structured Finance Operational Risk Guidelines: Moody s Approach to Analyzing Performance Disruption Risk, April 12, 2011 (SF243145). This Rating Implementation Guidance describes standards for servicers, backup servicers, collateral/paying agents for highly rated securities to ensure continuity of payments. For example securitizations with investment grade servicers in standardized assets have little risk of a payment disruption and no backup servicer is expected. On the other hand, transactions with non-investment grade servicers or nonstandard assets have a higher risk of disruption and need additional protections in the form of backup servicers. Standards for collateral agents are also provided. To access any of these reports, click on the entry above. Note that these references are current as of the date of publication of this report and that more recent reports may be available. All research may not be available to all clients. Moody s publishes a weekly summary of structured finance credit, ratings and methodologies, available to all registered users of our website, at 51 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS

52 Report Number: SF Moody s Investors Service, Inc. and/or its licensors and affiliates (collectively, MOODY S ). All rights reserved. CREDIT RATINGS ISSUED BY MOODY'S INVESTORS SERVICE, INC. (MIS ) AND ITS AFFILIATES ARE MOODY S CURRENT OPINIONS OF THE RELATIVE FUTURE CREDIT RISK OF ENTITIES, CREDIT COMMITMENTS, OR DEBT OR DEBT-LIKE SECURITIES, AND CREDIT RATINGS AND RESEARCH PUBLICATIONS PUBLISHED BY MOODY S ( MOODY S PUBLICATIONS ) MAY INCLUDE MOODY S CURRENT OPINIONS OF THE RELATIVE FUTURE CREDIT RISK OF ENTITIES, CREDIT COMMITMENTS, OR DEBT OR DEBT-LIKE SECURITIES. MOODY S DEFINES CREDIT RISK AS THE RISK THAT AN ENTITY MAY NOT MEET ITS CONTRACTUAL, FINANCIAL OBLIGATIONS AS THEY COME DUE AND ANY ESTIMATED FINANCIAL LOSS IN THE EVENT OF DEFAULT. CREDIT RATINGS DO NOT ADDRESS ANY OTHER RISK, INCLUDING BUT NOT LIMITED TO: LIQUIDITY RISK, MARKET VALUE RISK, OR PRICE VOLATILITY. CREDIT RATINGS AND MOODY S OPINIONS INCLUDED IN MOODY S PUBLICATIONS ARE NOT STATEMENTS OF CURRENT OR HISTORICAL FACT. CREDIT RATINGS AND MOODY S PUBLICATIONS DO NOT CONSTITUTE OR PROVIDE INVESTMENT OR FINANCIAL ADVICE, AND CREDIT RATINGS AND MOODY S PUBLICATIONS ARE NOT AND DO NOT PROVIDE RECOMMENDATIONS TO PURCHASE, SELL, OR HOLD PARTICULAR SECURITIES. NEITHER CREDIT RATINGS NOR MOODY S PUBLICATIONS COMMENT ON THE SUITABILITY OF AN INVESTMENT FOR ANY PARTICULAR INVESTOR. MOODY S ISSUES ITS CREDIT RATINGS AND PUBLISHES MOODY S PUBLICATIONS WITH THE EXPECTATION AND UNDERSTANDING THAT EACH INVESTOR WILL MAKE ITS OWN STUDY AND EVALUATION OF EACH SECURITY THAT IS UNDER CONSIDERATION FOR PURCHASE, HOLDING, OR SALE. ALL INFORMATION CONTAINED HEREIN IS PROTECTED BY LAW, INCLUDING BUT NOT LIMITED TO, COPYRIGHT LAW, AND NONE OF SUCH INFORMATION MAY BE COPIED OR OTHERWISE REPRODUCED, REPACKAGED, FURTHER TRANSMITTED, TRANSFERRED, DISSEMINATED, REDISTRIBUTED OR RESOLD, OR STORED FOR SUBSEQUENT USE FOR ANY SUCH PURPOSE, IN WHOLE OR IN PART, IN ANY FORM OR MANNER OR BY ANY MEANS WHATSOEVER, BY ANY PERSON WITHOUT MOODY S PRIOR WRITTEN CONSENT. All information contained herein is obtained by MOODY S from sources believed by it to be accurate and reliable. Because of the possibility of human or mechanical error as well as other factors, however, all information contained herein is provided AS IS without warranty of any kind. MOODY'S adopts all necessary measures so that the information it uses in assigning a credit rating is of sufficient quality and from sources MOODY'S considers to be reliable including, when appropriate, independent third-party sources. However, MOODY S is not an auditor and cannot in every instance independently verify or validate information received in the rating process. Under no circumstances shall MOODY S have any liability to any person or entity for (a) any loss or damage in whole or in part caused by, resulting from, or relating to, any error (negligent or otherwise) or other circumstance or contingency within or outside the control of MOODY S or any of its directors, officers, employees or agents in connection with the procurement, collection, compilation, analysis, interpretation, communication, publication or delivery of any such information, or (b) any direct, indirect, special, consequential, compensatory or incidental damages whatsoever (including without limitation, lost profits), even if MOODY S is advised in advance of the possibility of such damages, resulting from the use of or inability to use, any such information. The ratings, financial reporting analysis, projections, and other observations, if any, constituting part of the information contained herein are, and must be construed solely as, statements of opinion and not statements of fact or recommendations to purchase, sell or hold any securities. Each user of the information contained herein must make its own study and evaluation of each security it may consider purchasing, holding or selling. NO WARRANTY, EXPRESS OR IMPLIED, AS TO THE ACCURACY, TIMELINESS, COMPLETENESS, MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE OF ANY SUCH RATING OR OTHER OPINION OR INFORMATION IS GIVEN OR MADE BY MOODY S IN ANY FORM OR MANNER WHATSOEVER. MIS, a wholly-owned credit rating agency subsidiary of Moody s Corporation ( MCO ), hereby discloses that most issuers of debt securities (including corporate and municipal bonds, debentures, notes and commercial paper) and preferred stock rated by MIS have, prior to assignment of any rating, agreed to pay to MIS for appraisal and rating services rendered by it fees ranging from $1,500 to approximately $2,500,000. MCO and MIS also maintain policies and procedures to address the independence of MIS s ratings and rating processes. Information regarding certain affiliations that may exist between directors of MCO and rated entities, and between entities who hold ratings from MIS and have also publicly reported to the SEC an ownership interest in MCO of more than 5%, is posted annually at under the heading Shareholder Relations Corporate Governance Director and Shareholder Affiliation Policy. Any publication into Australia of this document is by MOODY S affiliate, Moody s Investors Service Pty Limited ABN , which holds Australian Financial Services License no This document is intended to be provided only to wholesale clients within the meaning of section 761G of the Corporations Act By continuing to access this document from within Australia, you represent to MOODY S that you are, or are accessing the document as a representative of, a wholesale client and that neither you nor the entity you represent will directly or indirectly disseminate this document or its contents to retail clients within the meaning of section 761G of the Corporations Act Notwithstanding the foregoing, credit ratings assigned on and after October 1, 2010 by Moody s Japan K.K. ( MJKK ) are MJKK s current opinions of the relative future credit risk of entities, credit commitments, or debt or debt-like securities. In such a case, MIS in the foregoing statements shall be deemed to be replaced with MJKK. MJKK is a wholly-owned credit rating agency subsidiary of Moody's Group Japan G.K., which is wholly owned by Moody s Overseas Holdings Inc., a wholly-owned subsidiary of MCO. This credit rating is an opinion as to the creditworthiness or a debt obligation of the issuer, not on the equity securities of the issuer or any form of security that is available to retail investors. It would be dangerous for retail investors to make any investment decision based on this credit rating. If in doubt you should contact your financial or other professional adviser. 52 JANUARY 18, 2012 REQUEST FOR COMMENT: MOODY S PROPOSES TO UPDATE ITS APPROACH TO RATING SECURITIES BACKED BY FFELP STUDENT LOANS