The Chronicle of Higher Education reported that the proposed loan subsidy cuts would allow lenders to remain profitable ["Loan-Subsidy Cuts Sought by Congress Would Still Leave Profits for Lenders, Study Finds", July 11, 2007, http://chronicle.com/daily/2007/07/2007071101n.htm]. That analysis compared the 0.72% gain Sallie Mae receives, according to a Congressional Research Service analysis, with a 50 to 55 basis point cut in special allowance payments (SAP), correctly emphasizing that the subsidy cuts should be subtracted from the before-tax gain (and the taxes then recomputed) and not compared with the after-tax gain of 45 bp. However, pending legislation involves cuts to more than just SAP. It also involves decreases in the lender insurance percentage and increases in the lender-paid origination fees. This message analyzes the combined impact of these subsidy cuts, which are greater than the 0.50% subsidy cut considered in the Chronicle of Higher Education article. The Congressional Research Service report did not consider the impact on yield of any proposed lender subsidy cuts, but rather provided a snapshot of Sallie Mae's current yield on FFELP loans. The following analysis discusses the impact on yield of all aspects of the proposed subsidy cuts, including aspects that were previously overlooked. [Addendum 7/19/07: The conclusion of this analysis is that including the impact of increases in risk sharing and lender-paid origination fees adds 7 bp to 15 bp to the impact on lender yield from the SAP cuts (probably closer to the 7 bp end of the range). This brings the total impact on for-profit lenders to somewhere between 57 bp and 62 bp for the President's proposal, 65 bp to 71 bp for the House proposal (60 bp to 62 bp for small and not-for-profit lenders), and 58 bp to 64 bp in the Senate (43 bp to 49 bp for not-for-profit lenders). While this will leave larger lenders like Sallie Mae profitable (e.g., due to remaining before-tax loan spread, economies of scale involving a lower marginal cost for additional loan originations not considered by the Congressional Research Service report and increases in loan volume due to higher annual and aggregate FFELP loan limits in 2007 and beyond), these cuts are severe enough to significantly reduce their profits and may leave smaller lenders unprofitable. This analysis is theoretical and limited by numerous assumptions discussed at the end of this document. The author feels that there is sufficient uncertainty concerning the impact on lender yield that it would be best to implement a somewhat smaller cut initially, evaluate the impact empirically through analysis of subsequent lender 10Q and 10K filings, and then potentially revisit the issue of additional cuts through subsequent legislation. Accordingly the author recommends limiting the subsidy cuts to just the 50 bp SAP cuts, omitting the increases in risk sharing and lender-paid origination fees. This would yield results that are consistent with the analysis published in the Chronicle of Higher Education. The Nelson-Burr amendment proposes applying a lower 35 bp SAP cut while retaining the increases in lender-paid origination fees and risk sharing, yielding an estimated impact on yield of 43 bp to 49 bp. While this is lower than the 50 bp considered in the Chronicle of Higher Education report, it would eliminate the uncertainty inherent in analysis of the increases in lender-paid origination fees and risk sharing. When Congress subsequently considered additional SAP cuts, there would be a greater degree of clarity on the impact on lender yields.] [Addendum 9/5/07: Added analysis based on the conference bill released today.] The following analysis depends, in part, on figures derived from Sallie Mae's 2006 10K filing with the SEC, various prospectuses for Sallie Mae's 2006 student loan securitizations and on figures obtained directly from Sallie Mae. In some cases I have genericized the data in order to consider a range of potential values applicable to other lenders. This analysis also depends on the Congressional Research Service's report and on default rate data from the President's FY2008 budget. *** Proposed Subsidy Cuts *** There are four proposals for cutting lender subsidies: [a] President's Budget - SAP cut by 0.50% - 2% reduction in insurance percentage (99% to 97% for exceptional performers, 97% to 95% for all else) - Increases lender-paid origination fee from 0.50% to 1.0% [b] House College Cost Reducation Act of 2007 (H.R.2669) - SAP cut by 0.55% (Stafford, Consolidation) and 0.85% (PLUS) - Eliminates exceptional performer status and cuts insurance percentage to 95% for all lenders (4% reduction for exceptional performers, 2% for all others) - Increases lender-paid origination fee from 0.50% to 1.0% but reduces them to 0.0% for small lenders and nonprofit/state lenders (small lenders = bottom 15% by loan volume) [c] Senate Higher Education Access Act of 2007 (S.1762) - SAP cut by 0.50% (Stafford, Consolidation) and 0.80% (PLUS) for for-profit lenders - SAP cut by 0.35% (Stafford, Consolidation) and 0.65% (PLUS) for not-for-profit lenders - Eliminates exceptional performer status and cuts insurance percentage to 97% for all lenders (2% reduction for exceptional performers, 0% for all others) - Increases lender-paid origination fee from 0.50% to 1.0% for all lenders [d] Conference Bill (9/5/2007) - SAP cut by 0.55% (Stafford, Consolidation) and 0.85% (PLUS) for for-profit lenders effective 10/1/07 - SAP cut by 0.40% (Stafford, Consolidation) and 0.40% (PLUS) for not-for-profit lenders effective 10/1/07 - Eliminates exceptional performer status. Delays change until end of year for lenders who had exceptional performer status as of 10/1/07. - Cuts insurance percentage to 95% for all lenders (4% reduction for exceptional performers, 2% for all others). Effective 10/1/2012. - Increases lender-paid origination fee from 0.50% to 1.0% for all lenders effective 10/1/07. *** Analysis *** The first step in the analysis is to translate the decreases in insurance percentages and increases in lender-paid origination fees into the equivalent of SAP cuts. This involves making assumptions about the composition of the loan portfolio and the average life of each component. From that it is possible to amortize the subsidy cuts over the life of the loans. According to page 337 of the President's FY2008 budget for the US Department of Education, long-term projected FFEL default rates for FY2006 are as follows: 12.30% Subsidized Stafford 11.13% Unsubsidized Stafford 5.20% PLUS 13.31% Consolidation with a weighted average of 12.27%. Based on 2006 subsidized Stafford loan volume of $19.9 billion and unsubsidized Stafford loan volume of $20.1 billion, I calculate a weighted average default rate for Stafford loans of 11.71%. Based also on PLUS loan volume of $7.4 billion and consolidation loan volume of $72.0 billion, I calculate the weighted average default rate, excluding PLUS loans, at 12.74%. The weighted average default rate for Stafford and PLUS loans, but excluding consolidation loans, is 10.70%. According to page 59 of Sallie Mae's 2006 10K filing, as of 12/31/2006 the distribution of loans in Sallie Mae's loan portfolio was as follows: 28% FFELP Stafford and PLUS Loans 56% FFELP Consolidation Loans 16% Private Education Loans Of Sallie Mae's total FFELP loan volume, the breakdown is as follows: 33% FFELP Stafford and PLUS Loans 67% FFELP Consolidation Loans Of the Stafford and PLUS loans, 48% are subsidized Stafford, 40% are unsubsidized Stafford, and 12% are PLUS. When amortizing a one-time fee over the life of a loan to determine the impact on yield, one needs to consider that the loan balance varies over the life of the loan, which includes the in-school, grace period and repayment period, as well as any periods of deferment and forbearance, reduced by any prepayments. One can calculate the average life of the loan by dividing the sum of the monthly loan balances from origination to payoff by the amount originated. For example, if we assume that a student borrows four years of unsubsidized Stafford loans at the July 1, 2007 annual limits with even disbursements at months 1 and 5 of a 9 month academic year, capitalization at repayment, 6.8% interest rate, a six month grace period and a 10 year repayment term, the average life of the loan is 8.9 years (107.1 months). 6.5 years (78.2 months) of that will be during repayment and 2.4 years (28.9 months) during the in-school and grace periods. If Congress increases the junior and senior loan limits to $7,500, the average life will drop to 8.7 years (104.7 months) because more of the debt will be incurred closer to repayment. For a subsidized Stafford loan, where the government pays the interest during the in-school and grace periods, the average life using current annual limits is 8.0 years (96 months). Using the proposed higher limits for juniors and seniors the average life drops to 7.9 years (94 months). For PLUS loans, where repayment begins 60 days after full disbursement, the average life is 6.2 years (74.1 months), assuming a 10 year repayment term and 8.5% interest rate. Combining these figures yields a weighted average for Stafford and PLUS loans of 8.2 years (97.8 months) assuming current annual limits, 8.0 years (95.9 months) assuming the higher proposed annual limits. If we assume that students consolidate immediately at repayment, we have the following mapping from loan term to average life, including an average of 2.4 years before the loan is consolidated and enters repayment, assuming a 48:40 mix of subsidized and unsubsidized Stafford loans at current annual limits: Loan Term Average Life --------------------------------- 10 years 8.4 years (101 months) 12 years 9.8 years (117 months) 15 years 11.9 years (142 months) 20 years 15.6 years (187 months) 25 years 19.5 years (234 months) 30 years 23.7 years (284 months) Sallie Mae reports that the average life of a Stafford loan after entering repayment is about 6.5 years and the average life for a consolidation loan is about 11 years. The prospectuses for their 2006 securitizations report the following weighted average remaining terms to scheduled maturity: 276 months at $3.7 billion (2006-10) 261 months at $2.6 billion (2006-9) 236 months at $3.1 billion (2006-8) 269 months at $2.6 billion (2006-7) 264 months at $1.1 billion (2006-6) 264 months at $3.1 billion (2006-5) 289 months at $1.5 billion (2006-4) 123 months at $2.6 billion (2006-3) 290 months at $3.1 billion (2006-2) 119 months at $2.5 billion (2006-1) The volume-weighted average of these figures yields an overall weighted average remaining term for all 2006 securitizations at 239 months (19 years 11 months). The weighted average remaining term for just consolidation loans may be as high as 290 months (24 years 2 months) based on the 2006-2 and 2006-4 securitizations which appear to consist primarily of consolidation loans. The maximum loan term on an unconsolidated Stafford loan is 120 months (10 years), not considering the ability to obtain extended repayment on unconsolidated Stafford loans with total borrower balances of $30,000 or more. A remaining term of 290 months corresponds to an average life of 18.8 years (226 months). This figure includes 2.4 months during the in-school and grace periods; otherwise it would be 16.4 years. It strikes me that average life figures will be somewhere in between the Sallie Mae figures and my calculations because low interest rates on FFELP loans and increasing loan balances may reduce future prepayment activity. Accordingly, I will calculate the impact of both sets of figures: 6.5 to 8.4 years for Stafford Loans and 11 years to 18.8 years for consolidation loans. (I will substitute 16.4 years in some amortization calculations as appropriate.) Calculating the impact on yield of the cuts in the insurance percentage involves multiplying the default rate by the amount of the cut in risk sharing divided by the average life. We do not need to normalize for differences between loan balances at default and the average loan balance for the portfolio because default rates are expressed in terms of loan volume, not number of loans. Accordingly, the impact of a 1% cut in the insurance percentage is: Stafford & PLUS: 10.7% * 1.0% / (6.5 years to 8.4 years) = 1.65 bp to 1.27 bp Consolidation: 13.31% * 1.0% / (11 years to 18.8 years) = 1.21 bp to 0.71 bp The weighted average, assuming a portfolio that is 33% Stafford/PLUS and 67% Consolidation is 1.36 bp to 0.89 bp. Focusing on the impact on exceptional performers, which represent approximately 80% of all education lenders, the President's Budget and the Senate would cut the insurance percentage by 2% and the House by 4%. This yields the following impact on yield from the cuts in insurance percentage: Impact of Risk Sharing on Yield: [a] President's Budget: 2.72 bp to 1.79 bp [b] House: 5.44 bp to 3.58 bp [c] Senate: 2.72 bp to 1.79 bp [d] Conference Bill: 5.44 bp to 3.58 bp With regard to increases in lender-paid origination fees, the President's Budget and the Senate would increase it by 0.50%, while the House would increase it by 0.50% for larger lenders and reduce it by 0.50% for nonprofit and small lenders (bottom 15% by loan volume). I estimate that the House proposal is equivalent to about a 0.35% increase overall. Calculating the impact of lender-paid origination fees on yield involves amortizing the up-front fee over the life of the loan. Assuming a 5.25% discount rate (the origination fees are paid in advance unlike the risk sharing which is incurred in arrears), I calculate that a 0.50% increase in fees is the equivalent of 9.06 bp to 7.39 bp for Stafford/PLUS loans (assuming 6.5 years to 8.4 years average life, respectively) and 5.97 bp to 4.54 bp to 4.18 bp for Consolidation loans (assuming 11 years to 16.4 years to 18.8 years average life). The weighted average of the impact on yield corresponding to these fees is 6.99 bp to 5.24 bp, assuming a 0.50% increase in the lender-paid origination fees. Sallie Mae notes that it pays the origination fee twice on consolidation loans, since most of its consolidation loans involve its own borrowers, and so feels that the impact on yield for consolidation loans should be doubled. It is not quite that simple, since the two fees have different applicable average lifes, and loans that are securitized and subsequently consolidated can be securitized again, yielding a second premium on the loan. Adjusting for this would increase the weighted average of the impact on yield by 4.00 bp to 3.04 bp. I believe that the latter is more accurate, so I will allow a 3.04 bp increment on the upper end of the range, yielding a weighted average of 10.03 bp to 5.24 bp. This yields the following impact on yield from the increases in lender paid origination fees: Impact of Lender-Paid Origination Fees: [a] President's Budget: 10.03 bp to 5.24 bp [b] House: 7.02 bp to 3.67 bp (average) 10.03 bp to 5.24 bp (large for-profit) -10.03 bp to -5.24 bp (small and not-for-profit) [c] Senate: 10.03 bp to 5.24 bp [d] Conference Bill: 10.03 bp to 5.24 bp Assuming 4% PLUS loan volume and 96% Stafford/Consolidation loan volume, the weighted average of the SAP cuts are as follows: Impact of the SAP cuts: [a] President's Budget: 50.00 bp [b] House: 56.20 bp [c] Senate: 51.20 bp (for-profit) 36.20 bp (not-for-profit) [d] Conference Bill: 56.20 bp (for-profit) 41.20 bp (not-for-profit) Combining the impact of the risk sharing and lender-paid origination fees with the SAP cuts yields the following total impact on yield: Total Impact of Subsidy Cuts on Yield: [a] President's Budget: 62.75 bp to 57.03 bp [b] House: 68.66 bp to 63.45 bp (average) 71.67 bp to 65.02 bp (larger for-profit) 51.61 bp to 54.54 bp (small and not-for-profit) [c] Senate: 63.95 bp to 58.23 bp (for-profit) 48.95 bp to 43.23 bp (not-for-profit) [d] Conference Bill: 71.67 bp to 65.02 bp (for-profit) 56.67 bp to 50.02 bp (not-for-profit) The Congressional Research Service report pegged the Sallie Mae excess spread at 2.18%, the difference between an "average coupon rate" paid by borrowers of 7.43% and the 5.25% paid to investors. The 2.18% figure seems reasonable, considering that page 62 of the Sallie Mae 2006 10K lists a student loan yield of 8.09% and a cost of funds of 5.45% for all loans, including private student loans, and page 64 reports a 1.26% FFELP loan spread and 5.13% private education loan spread. Normalizing the 0.80% in fees and costs subtracted in Sallie Mae's spread calculation by 85%, to account for the normalization of those costs by the inclusion of the private student loan portfolio, yields 0.94% in costs for just the FFELP loans. 1.26% plus 0.94% is 2.20%, roughly the same as the 2.18% figure in the CRS report. The CRS report then subtracts 0.70% in origination and servicing costs and borrower benefits, and 0.75% in average lender origination and consolidation offset fees, yielding 0.72% in before-tax profit. The average tax rate is 37.7%. The after-tax profit is 0.45%. (Note: Sallie Mae has stated recently in newspaper interviews that "Student loan providers make roughly $.0056 on every $1 they lend." This is equivalent to a rate of 0.56%, somewhat higher than the rate indicated by the CRS report.) To calculate the impact of the proposed lender subsidy cuts on after-tax profit, it is necessary to subtract the impact on yield from the before-tax figure and recalculate the taxes on the new figure. This yields the following impact on larger for-profit lenders: After-Tax After-Cut Profit: [a] President's Budget: 5.76 bp to 9.33 bp [b] House: 0.21 bp to 4.35 bp [c] Senate: 5.02 bp to 8.58 bp [d] Conference Bill: 0.21 bp to 4.35 bp Percentage Cut in Profits on New Originations After Subsidy Cuts: [a] President's Budget: 87% to 79% [b] House: 99% to 90% [c] Senate: 89% to 81% [d] Conference Bill: 99% to 90% Since these cuts represent a 81% to 99% cut in Sallie Mae profits from new loan originations, it is highly likely that most of Sallie Mae's profit from newly originated loans after the subsidy cuts will derive from servicing the loans and collecting defaulted loans, along with revenue from private student loans. However, it is worth noting that the Congressional Research Service's report does not consider that some of Sallie Mae's costs are fixed, so the marginal cost for additional loans is lower, perhaps by as much as 30 bp. Limiting the subsidy cuts to just the SAP cuts of 0.50% for Stafford and Consolidation loans and 0.80% for PLUS loans (weighted average of 51.2%) as proposed by the Senate, without any decrease in lender insurance percentage or increase in lender-paid origination fees would result in after-tax profit of 12.96 bp, a 71% reduction in profit. The following table illustrates the impact on after-tax profit of various SAP-only cuts: Cut After-Tax Profit ---------------------------------- 70 bp 1.25 bp (97% cut) 65 bp 4.36 bp (90% cut) 60 bp 7.48 bp (83% cut) 55 bp 10.59 bp (76% cut) 50 bp 13.71 bp (70% cut) 45 bp 16.82 bp (63% cut) 40 bp 19.94 bp (56% cut) 35 bp 23.05 bp (49% cut) 30 bp 26.17 bp (42% cut) 25 bp 29.28 bp (35% cut) *** Limitations *** This analysis has several limitations: - It does not consider profit associated with last fees and collection of defaulted loans. - It does not consider potential reductions in the premium paid to acquire education loans. Those premiums would likely drop due to the lower value of newly originated loans. These are currently 10 bp to 16 bp amortized. - It relies on long-term projections of national default rates, which are highly sensitive to variations in model parameters. In addition, Sallie Mae has stated publicly that its default rate is "well below the national rate". The impact of a lower default rate, however, is likely to be limited due to the low impact of risk sharing on lender yield. - It does not consider potential reductions in loan discounts (borrower benefits), which are currently about 9 bp. Lenders might cut these back end benefits, although they are unlikely to eliminate them entirely due to the competitive environment. Likewise, front end benefits such as fee waivers may also be reduced due to the lower premium paid for such loans. - It does not consider improvements in the economy of scale from increases in loan volume, where increases in loan volume may compensate somewhat for declines in per-loan profit, nor the lower marginal costs associated with additional loan volume. Increases in loan volume may stem partly from reduced competition and partly from current and proposed increases in annual and aggregate Stafford loan limits (e.g., 29% to 36% increases in aggregate limits from $23,000 to $30,500, freshman annual limits from $2,625 to $3,500, sophomore annual limits from $3,500 to $4,500 and junior/senior annual limits from $5,500 to $7,500). The new loan limits were, however, considered in the analysis of average loan life. - It considers only the impact of the subsidy cuts on newly originated loans. Existing loans will not be affected by these cuts, nor will private student loans. This will dilute the immediate impact of the cuts on lender profitability, causing a several year delay before the lenders feel the full impact of the cuts. - It does not consider the impact of potential future legislative changes that might shift loan volume from private student loans to federal loans (e.g., allowing undergraduate students to borrow PLUS loans without parental involvement) or the potential impact of market based loan mechanisms on FFELP loan volume. - It is based on 2006 statistics and as such does not consider changes subsequent to 12/31/2006. *** Conclusion and Recommendations *** The proposed subsidy cuts, especially those passed by the House, represent a severe cut in profits for education lenders. The cuts are severe enough that they may leave many smaller lenders unprofitable. This is especially true of consolidation-only lenders and marketers and lenders that do not service their own loans. The cuts are severe enough that industry-wide profit levels may fall significantly below comparable benchmarks for consumer finance companies. This may be contrary to Congress's intent to focus public dollars on students while still allowing education lenders to retain a reasonable level of profit. If that were not Congress's intent, it would be simpler for Congress to repeal the FFELP program. This analysis is complicated and involves many simplifying assumptions that can significantly affect the results. Although I am reasonably confident in the results, I recognize that the results may overstate or understate the impact of the subsidy cuts on lender profits. There are a lot of fudge factors. Most likely the impact is toward the less severe end of the range, which is still more severe than previously indicated. Due to the uncertainty associated with this type of analysis, I recommend proceeding with caution. A staged approach, in which some cuts are implemented now and others pursued later depending on the results of an empirical evaluation of the impact of the first set of cuts, might be advisable. Accordingly, Congress may wish to consider scaling back the subsidy cuts to just the 50 bp SAP cut, eliminating the proposed reduction in insurance percentage and increase in lender-paid origination fees. Congress can then evaluate the actual impact of the SAP cuts on subsequent lender profitability and potentially revisit the other subsidy cut proposals during the 2013 reauthorization of the Higher Education Act. The impact of SAP cuts on lenders is much easier to evaluate than increases in risk-sharing and lender-paid origination fees. Mark Kantrowitz Publisher, FinAid.org -------------------------------------------------- Mark Kantrowitz Publisher, FinAid Director of Advanced Projects, FastWeb Author, FastWeb College Gold FinAid Page LLC PO Box 2056 Cranberry Township, PA 16066-1056 Tel: 1-724-538-4500 Fax: 1-724-538-4502 Email: mkant@finaid.org, mkant@fastweb.com www.fastweb.com www.finaid.org www.collegegold.com