3900 Wisconsin Avenue, NW Washington, DC 20016-2892 | |
November 26, 2013
Via EDGAR
Ms. Stephanie J. Ciboroski
Senior Assistant Chief Accountant
Division of Corporation Finance
United States Securities and Exchange Commission
100 F Street, NE
Washington, DC 20549
Re: Federal National Mortgage Association
Form 10-K for the Fiscal Year Ended December 31, 2012
Filed April 2, 2013
File No. 0-50231
Dear Ms. Ciboroski,
The Federal National Mortgage Association (“Fannie Mae”) provides the following information in response to the comments contained in the correspondence of the staff (the “Staff”) of the Division of Corporation Finance of the U.S. Securities and Exchange Commission (the “Commission”), dated November 7, 2013, relating to the above-referenced filing. Below we have repeated the text of your comments and followed each with our response.
Form 10-K for the Fiscal Year Ended December 31, 2012
FHFA Advisory Bulletin Regarding Framework for Adversely Classifying Loans, page 34
As stated in our letter to the Staff dated August 29, 2013, the Advisory Bulletin requires us to charge off (in full or in part) any loan classified as a “loss” when the loan becomes 180 days or more past due, except for instances where the delinquent loan is well secured (loans with loan-to-value ratios of 60% or less) or is in an open trial period for a loan modification. The Advisory Bulletin establishes a standard and uniform methodology for classifying assets based on their credit quality. Additionally, our regulator, the Federal Housing Finance Agency (“FHFA”), has indicated that the Advisory Bulletin was issued to provide guidelines for management of credit risk, to ensure that we take effective and expedient action to address delinquent and impaired loans, and to appropriately realize losses in a timely fashion.
As discussed in our previous letter, the adoption of the charge-off guidance in the Advisory Bulletin is a change from our existing practice for determining when a loan is deemed uncollectible. The primary justification for this change is the fact that substantially all loans that are 180 days past due have been referred to foreclosure under our current default servicing timeline and the majority of this population is ultimately foreclosed upon. Additional considerations include:
U.S. Securities and Exchange Commission
November 26, 2013
Page 2
• | A borrower that is 180 days delinquent has demonstrated an inability or lack of willingness to meet their obligation to make timely payments on their mortgage loan and has not been responsive to efforts to cure the delinquency. |
• | The Advisory Bulletin is consistent with long-standing regulatory requirements imposed on other federally regulated financial institutions. |
Finally, FHFA has indicated that charging off the loans that are 180 days or more delinquent to the fair value of the collateral (less costs to sell) plus expected recoveries from credit enhancements and representations and warranties (collectively referred to as “Collateral”) provides the most meaningful financial reporting for the loans that are 180 days or more delinquent and will help drive vigorous implementation of enhanced credit risk management and loss mitigation activities.
We believe that it would be beneficial to accelerate the timing of our loss mitigation activities. We have recently implemented new programs (a streamlined modification program and default servicing standards) that are designed to achieve this goal. However, it is too early to determine the impact that these programs will have on our current loss mitigation timelines. If recent trends continue, we expect the population of loans that will be subject to the charge-off provisions of the Advisory Bulletin will decline either through liquidation or re-performance. As a result, we believe the impact upon adoption of the guidance on January 1, 2015 will be less than $1.4 billion (pre-tax). For additional information on the adoption of the Advisory Bulletin, see our disclosure related to the Advisory Bulletin on pages 14 and 15 of our quarterly report on Form 10-Q for the quarter ended September 30, 2013 (our “Third Quarter 2013 Form 10-Q”).
Note: Our credit models generally consider the anticipated benefits from borrower re-performance, which serve to reduce our exposure to credit losses, when we establish our allowance for loan losses. Borrower re-performance in this context includes the likelihood that the following activities could occur after a loan becomes 180 days past due: payoff, full or partial repayment, a successful loan modification, and/or no loss default. For additional information on each of these outcomes, see our letter dated August 29, 2013.
1. | We note your response to prior comment 1 where you describe the use of an econometric model for loans that are collectively evaluated for impairment, along with your expectations that you will continue to use this approach for estimating expected losses for collectively evaluated single-family loans upon the adoption of the Advisory Bulletin. You also state that upon the adoption of the Advisory Bulletin, the amount of the charge-off at 180 days is expected to exceed your best estimate of incurred losses going forward. Please address the following for loans past due 180 days and collectively evaluated for impairment: |
• | Tell us whether you believe that the amount charged-off upon implementation of the Advisory Bulletin is consistent with the guidance in ASC 450-20-30-1. |
ASC 450-20-30-1 requires us to measure and recognize the best estimate of a loss as a valuation allowance on a pool of loans. However, when no amount within a range is a better estimate than any other amount within the range, the minimum incurred loss within the range shall be recognized. For any given loan, the range of possible losses is between zero (e.g., borrower performs in accordance with contractual terms) and the amount computed as the difference between the recorded investment in the mortgage loan and the expected recoveries from Collateral.
Our allowance methodology uses a probability-weighted outcome to determine our best estimate of the loss we have incurred on a pool of loans. As described in our letter dated August 29, 2013, we use our historical experience, an econometric model, and the current attributes of each loan to estimate the probability that a given loan will default (go to foreclosure) and to estimate the likely severity of loss using the expected
U.S. Securities and Exchange Commission
November 26, 2013
Page 3
recoveries from Collateral. Our models are calibrated to historical data and back-tested. We believe that the loss estimates from this process represent the best estimate within the range of possible losses.
In contrast, the measurement of the charge-off upon adoption of the Advisory Bulletin is on an individual loan basis and will not consider all of the elements of our current definition of borrower re-performance (which is the same as assuming that all loans that are 180 days or more past due will default). The Advisory Bulletin requires that re-performance be recorded on a cash basis as recoveries when received. Therefore, losses are estimated using the difference between each loan’s recorded investment and expected recoveries from Collateral. This approach, which excludes the pool-level expected re-performance, is consistent with the manner in which a charge-off on an individual mortgage loan is customarily measured. As the measurement of the charge-off under the Advisory Bulletin does not consider the likelihood that loans within the loan pool will re-perform, this amount does not represent the best estimate within the range of loss on a pool basis when the percentage of loans within the pool that re-perform after becoming 180 days past due is more than insignificant.
In certain states that require a judicial foreclosure process, backlogs have extended the timeline to a foreclosure for up to several years. Additionally, some borrowers do not engage in a timely manner. Therefore, we continue to experience loan modifications in very deep stages of delinquency, until well after the loans are a year past due. As of September 30, 2013, the average probability of default for a loan that was collectively evaluated for impairment and past due 180 days or more was 70%. As the level of expected re-performance on this population is not insignificant, the high end of the range would not represent the best estimate of our ultimate incurred losses. Therefore, as of September 30, 2013, the charge-off measured under the Advisory Bulletin may not be consistent with the guidance that is the basis for our measurement of the allowance for loan losses in ASC 450-20-30-1.
The charge-off provisions of the Advisory Bulletin will not be adopted until January of 2015. As we believe that it would be beneficial to accelerate the timing of our loss mitigation activities, we have recently implemented new programs (a streamlined modification program and default servicing standards) that are designed to achieve this goal. However, it is too early to determine the impact that these programs will have on our current loss mitigation timelines. If these programs accelerate the timing of our loss mitigation activities such that re-performance after 180 days is insignificant, then the incurred losses on the loans that are 180 days or more past due may approximate the charge-off recorded under the Advisory Bulletin. In the current economic environment, the difference between these measurement approaches is not expected to be material to our financial statements.
Note: In the credit model that is used to determine the collective reserve, a loan modification is not considered to be an incurred loss as of the balance sheet date if the modification has performed for at least one year in accordance with the loan’s modified terms. See Table 36 in our Third Quarter 2013 Form 10-Q for information on the performance of our modifications, specifically, the percentage of single-family loan modifications that were current or paid off at one and two years post modification.
• | Tell us whether you believe collection based on the fair value of the underlying collateral reflects the most likely outcome for loans past due 180 days and subject to charge-off. |
We have assumed for the purpose of this response that “most likely” means more than 50% likely. Assuming that threshold and given our 70% probability of default noted above based on our historical data, collection based on the expected recoveries from the underlying Collateral reflects the most likely outcome for loans that are past due 180 days, subject to charge-off, and collectively evaluated for impairment.
• | Tell us whether you believe measuring incurred losses for single-family loans past due 180 days based on an individual assessment of the most likely outcome, as opposed to your collective methodology |
U.S. Securities and Exchange Commission
November 26, 2013
Page 4
which factors in the possibility of loan re-performance, is consistent with the guidance in ASC 450-20 and ASC 310-10-35-9.
While ASC 310-10-35-9 allows for the recognition criteria of ASC 450-20 to be applied to an individual loan, it does not provide criteria for measurement of such loan. ASC 450-20 requires an entity to measure the impairment based on its best estimate of the ultimate incurred loss on a pool basis. Measuring impairment based on the “most likely outcome” (foreclosure) would not result in Fannie Mae’s best estimate of incurred loss for the pool of single-family loans past due 180 days. Further, this approach would not consider information available to Fannie Mae (as discussed in ASC 310-10-55-5), including historical collection experience and recovery rates for loans in similar situations. As previously stated, Fannie Mae’s best estimate of incurred loss is based on an average probability that 70% of loans 180 days past due will go to foreclosure with the other 30% re-performing recently, primarily via successful loan modifications, but historically also driven by borrowers paying off their loan after sale of the property. If measurement were based on the most likely outcome, the incurred loss would be based on a probability that 100% of the loans would go to foreclosure, resulting in higher loss recognition because it would not consider re-performance. This result would not appear to be consistent with the best estimate of incurred loss.
As noted above, we have recently implemented several new loss mitigation approaches. If these programs accelerate the timing of our loss mitigation activities such that re-performance after 180 days is insignificant, then the incurred losses on the loans that are 180 days or more past due may approximate the charge-off recorded under the Advisory Bulletin.
2. | We note that loans individually evaluated for impairment principally include troubled debt restructurings (TDRs) and acquired credit-impaired loans. Please address the following for loans that are past due 180 days and individually evaluated for impairment: |
• | Tell us whether you believe measuring the incurred losses for loans past due 180 days based on an individual assessment of the most likely outcome, as opposed to a pool basis, is consistent with the guidance in ASC paragraph 310-10-35-21. |
We believe that it would be appropriate to measure impairment under ASC 310-10-35-21 using either each loan as the unit of account or a pool of loans aggregated based on common risk characteristics. Our current practice is to measure impairment using each individual loan as the unit of account and we have not elected to aggregate the loans on a pool basis as allowed under ASC 310-10-35-21. We do not believe that it is consistent with the requirements in ASC 310-10-35 to measure impairment based upon an assessment of the most likely outcome. The guidance in ASC 310-10-35 requires impairment to be measured based on the present value of expected future cash flows discounted at the effective interest rate with the following exceptions:
(1) | as a practical expedient, the guidance allows impairment to be measured using either the loan’s observable market price or the fair value of the collateral if the loan is a collateral dependent loan; and |
(2) | the guidance requires impairment to be measured based on expected recoveries from collateral when foreclosure is deemed probable. |
For our impairment calculations, we have concluded that foreclosure is probable on the majority of single-family loans that are individually evaluated for impairment and are 180 days past due. For this population, we measure impairment based upon the fair value of each loan’s specific collateral (less costs to sell), as well as the benefit associated with the credit enhancements that are specific to that loan. This measurement approach is primarily used for loans previously accounted for as troubled debt restructurings and for which
U.S. Securities and Exchange Commission
November 26, 2013
Page 5
loss mitigation efforts (e.g., modification, forbearance, etc.) have failed. For this population, our measurement of impairment is consistent with the most likely outcome (i.e. foreclosure).
A minority of the single family loans that are 180 days past due are individually measured for impairment using a cash flow approach as the borrowers on those loans are in trial modifications or have other fact patterns that would qualitatively support the conclusion that foreclosure is not probable. For these loans, we measure impairment as the present value of the expected future cash flows for each loan discounted at the loan’s effective interest rate. The cash flow estimates under this measurement approach have to consider all likely outcomes. As re-performance after the loan is 180 days past due is one of those outcomes, the estimates that are derived using this approach would only be consistent with the most likely outcome when re-performance is expected to be insignificant. As of September 30, 2013, the average probability of default for a loan that was individually evaluated for impairment using a cash flow approach and past due 180 days or more was 70%, based on our historical data. As the level of expected re-performance on this population is not insignificant, we believe that measuring impairment based upon an individual assessment of the most likely outcome may not be appropriate.
As noted above, we have recently implemented several new loss mitigation approaches. If these programs accelerate the timing of our loss mitigation activities such that re-performance after 180 days is insignificant, then the impairment estimated on the loans that are 180 days or more past and evaluated for impairment using a discounted cash flow methodology may approximate the charge-off recorded under the Advisory Bulletin.
• | Tell us whether you believe that collection based on the fair value of the underlying collateral reflects the most likely outcome for loans past due 180 days. |
We have assumed for the purpose of this response that “most likely” means more than 50% likely. The combined probability of default for loans that are individually impaired and for which impairment is measured based upon either a discounted cash flow approach or the fair market value of the underlying collateral is 81%, based on our historical data. As a result, collection based on the expected recoveries from the underlying collateral represents the most likely outcome for loans that are individually impaired and 180 days past due.
3. | We note you provide certain quantitative information in your response to prior comment 1 regarding the population of loans that would be affected by the adoption of the Advisory Bulletin. Please clarify the following as it relates to this information and provide an update of the analysis as of September 30, 2013: |
• | Confirm that the individually impaired loans measured based on collateral value on page 5 are included in the total 220,000 loans that impairment was measured based on the fair value of the underlying collateral and that the 135,000 loans are the only loans that were delinquent 180 days or more. |
See the table below to further clarify the composition of loans for which impairment is based on the fair value of the underlying collateral as of June 30, 2013:
U.S. Securities and Exchange Commission
November 26, 2013
Page 6
Unpaid Principal | ||||||
Approximate | Balance (UPB) | |||||
Loan Count | (in billions) | |||||
Loans measured based on the fair value of the underlying collateral: | ||||||
Not scoped into the Advisory Bulletin(1) | 85,000 | $ | 15 | |||
Scoped into the Advisory Bulletin | 135,000 | 27 | ||||
Total | 220,000 | $ | 42 | |||
__________
(1) | Loans that are not scoped into the Advisory Bulletin include loans that are less than 180 days past due and loans that are 180 days or more past due that have a mark-to-market loan-to-value ratio of less than 60% or that are in an open trial period as part of a loan modification. |
The 135,000 loans disclosed above do not represent the entire population of loans that are 180 days or more past due. Loans that have a mark-to-market loan-to-value ratio of less than 60% and loans that are in an open trial period as part of a loan modification are excluded from that number as they are not within the scope of the Advisory Bulletin.
• | Provide us with the related allowance for each line item in the table at the top of page 5 and the impact the Advisory Bulletin charge-off policy would have on each population if adopted as of September 30, 2013. |
The table below reflects the approximate population of loans, including the related allowance and estimated charge-off, which would have been subject to an accelerated charge-off if the Advisory Bulletin had been implemented as of September 30, 2013:
Percentage Of | |||||||||||||||||||||||
Total Single-Family | |||||||||||||||||||||||
Conventional | |||||||||||||||||||||||
Guaranty Book | |||||||||||||||||||||||
Estimated | Difference | of Business | |||||||||||||||||||||
Loan Count | UPB | Allowance | Charge-Off | (pre-tax) | (based on UPB) | ||||||||||||||||||
$ in Billions | A | B | A - B | ||||||||||||||||||||
Collective reserve | 116,000 | $ | 20.8 | $ | 4.4 | $ | 5.8 | $ | (1.4 | ) | 0.7 | % | |||||||||||
Individually impaired loans (cash flow) | 28,000 | 4.9 | 1.1 | 1.1 | * | 0.2 | % | ||||||||||||||||
Individually impaired loans (collateral value) | 129,000 | 26.2 | 6.0 | 6.0 | — | 0.9 | % | ||||||||||||||||
Total | 273,000 | $ | 51.9 | $ | 11.5 | $ | 12.9 | $ | (1.4 | ) | 1.8 | % | |||||||||||
__________
* Denotes amounts less than $0.1 billion.
If recent trends continue, we expect the population of loans that will be subject to the charge-off provisions of the Advisory Bulletin will decline either through liquidation or re-performance. As a result, we believe the impact upon adoption of the guidance on January 1, 2015 will be less than $1.4 billion (pre-tax). In the aggregate, this amount is not expected to be material to our consolidated financial statements.
Our expectation regarding the financial statement impact of the Advisory Bulletin upon adoption is a forward-looking statement, and actual results may differ, possibly materially. Factors that could cause actual results to differ include economic or market conditions or other factors resulting in our having a greater number of seriously delinquent loans in our book of business than we currently anticipate or home prices being lower than we currently expect.
U.S. Securities and Exchange Commission
November 26, 2013
Page 7
4. | We note the portion of your response to prior comment 1 where you discuss how you plan to account for recoveries and you note the two industry methods, cost-recovery and cash-basis, and you provide an alternative approach. For the alternative approach, you allocate the cash received between principal and interest based on the contractual amortization schedule and recognize a recovery of the previous charge-off on the performing loan. Under both the cost-recovery and cash-basis methods the amount recorded is equal to the cash payment received. Please explain how you determined it was reasonable to record more than the cash payment received (i.e. a recovery of the previous charge-off) under your proposed alternative approach and consider providing an illustrative example of how your alternative approach would operate. |
Upon adoption of the Advisory Bulletin, we plan on using the approach outlined in Office of the Comptroller of the Currency (“OCC”) Bank Accounting Advisory Series (Topic 2B, Question 3) to account for any recoveries we receive. Specifically, the OCC has provided the following guidance:
• | When doubt exists about the ultimate collectability of principal in full, payments received should be applied to principal to the extent necessary to reduce such doubt. Once the principal balance is reduced to zero, subsequent payments should be recorded as recovery of prior charge-offs, and then as interest income. |
• | When doubt about the ultimate collectability of principal no longer exists, payments received should be applied on a cash basis by allocating the payment to interest income, reduction of principal and recovery of prior charge-offs. Our discussions with banks that are subject to this guidance have indicated that cash is allocated between principal and interest in accordance with the contractual amortization schedule until the recorded investment in the loan is reduced to zero. Then, subsequent cash receipts are allocated as a recovery of the prior charge-off. |
As noted in our August 29, 2013 letter, the alternative approach that we believe is consistent with the accounting guidance allocates the cash received between principal and interest based on the contractual amortization schedule for the mortgage loan while concurrently recognizing a recovery of the previous charge-off as payments are received over the remaining life of the mortgage loan. This approach does not record more than the cash received; rather the cash received is effectively allocated between interest income, net reduction in recorded investment and recovery of previously charged-off amounts (see an illustrative example in the Appendix to this letter). A portion of the principal payment is a recovery of a previous charge-off. The alternative approach is based upon our interpretation of the guidance in ASC 310-10-35. Specifically, this guidance requires recoveries of loans and trade receivables that were previously charged off to be recorded when received. The judgment in the interpretation of this guidance relates to the question of when a creditor has received a recovery. The guidance in the OCC Bank Accounting Advisory Series is based upon the premise that the charge-off has not been received until the recorded investment in the loan has been reduced to zero.
The alternative approach is based upon the presumption that each payment received from the borrower includes a partial recovery of the amount that was previously charged off. To reach this conclusion, one would have to assume that the write-off on a mortgage loan represented a proportional reduction of each of the mortgage loan’s contractual payments. This assumption is not unreasonable given the fact that each loan’s recorded investment was initially set equal to the net present value of the loan’s contractual payments. Given the limited accounting guidance on this topic, we believe the linkage of a recovery to the cash receipts to be reasonable and consistent with the guidance on recoveries in ASC 310-10-35.
Although we plan to initially adopt the OCC approach described above, our regulator has requested that we align accounting policies with Freddie Mac. Therefore, we may have to adopt the alternate approach for the recognition of recoveries.
U.S. Securities and Exchange Commission
November 26, 2013
Page 8
Measurement of Interest Rate Risk, page 156
5. | We note your response to comment 3. Please further enhance the disclosure to explain the anomalies that may occur in the sensitivity tables as they relate to the duration gap and interest rates. Specifically address the following in your future filings: |
• | Tell us if the duration gap of your net portfolio is close to zero and discuss the anomalies that exist under this circumstance or other inconsistencies that may exist for a particular duration gap; and |
• | When applicable, disclose the fact that the market value sensitivities to parallel changes in rates are estimates as of a point of time and that the duration gap estimates are an average of daily values throughout the quarter. |
We will consider revisions to our disclosure in our Form 10-K for the year ended December 31, 2013 (our “2013 Form 10-K”) in response to your comments, based on facts and circumstances relevant to that reporting period. The revised disclosures that follow incorporate various points of clarification described in our letter to you dated August 29, 2013 and have been prepared to illustrate how such enhancements would have been implemented in our Form 10-K for the year ended December 31, 2012. (Revisions to our disclosures presented in this letter are shown in strike-through (deletions) and bold and underline (additions).) Please note that additions below that provide expanded analysis of market sensitivity results reflect facts and circumstances that are unique to the reporting periods presented. In this regard, we expect these disclosures may change in future reporting periods in light of changes in underlying modeling assumptions, future changes in the composition of our net portfolio and other relevant factors.
Interest Rate Sensitivity to Changes in Interest Rate Level and Slope of Yield Curve
As part of our disclosure commitments with FHFA, we disclose on a monthly basis the estimated adverse impact on the fair value of our net portfolio that would result from the following hypothetical situations:
•A 50 basis point shift in interest rates.
•A 25 basis point change in the slope of the yield curve.
In measuring the estimated impact of changes in the level of interest rates, we assume a parallel shift in all maturities of the U.S. LIBOR interest rate swap curve.
In measuring the estimated impact of changes in the slope of the yield curve, we assume a constant 7-year rate and a shift of 16.7 basis points for the 1-year rate and 8.3 basis points for the 30-year rate. We believe the aforementioned interest rate shocks for our monthly disclosures represent moderate movements in interest rates over a one-month period.
The market value sensitivities of our net portfolio are a function of both the duration and the convexity of our net portfolio. Duration provides a measure of the price sensitivity of a financial instrument to changes in interest rates while convexity reflects the degree to which the duration of the assets and liabilities in our net portfolio changes in response to a given change in interest rates. We use convexity measures to provide us with information about how quickly and by how much our net portfolio’s duration may change in different interest rate environments. The market value sensitivity of our net portfolio will depend on a number of factors, including the interest rate environment, modeling assumptions and the composition of assets and liabilities in our net portfolio, which vary over time.
The sensitivity measures presented in Table 64, which we disclose on a quarterly basis as part of our disclosure commitments with FHFA, are an extension of our monthly sensitivity measures. There are
U.S. Securities and Exchange Commission
November 26, 2013
Page 9
three primary differences between our monthly sensitivity disclosure and the quarterly sensitivity disclosure presented below: (1) the quarterly disclosure is expanded to include the sensitivity results for larger rate level shocks of plus or minus 100 basis points; (2) the monthly disclosure reflects the estimated pre-tax impact on the market value of our net portfolio calculated based on a daily average, while the quarterly disclosure reflects the estimated pre-tax market value impact calculated based on the estimated financial position of our net portfolio and the market environment as of the last business day of the quarter; and (3) the monthly disclosure shows the most adverse pre-tax impact on the market value of our net portfolio from the hypothetical interest rate shocks, while the quarterly disclosure includes the estimated pre-tax market value impact of both up and down interest rate shocks.
Duration Gap
Duration gap measures the price sensitivity of our assets and liabilities in our net portfolio to changes in interest rates by quantifying the difference between the estimated durations of our assets and liabilities. Our duration gap analysis reflects the extent to which the estimated maturity and repricing cash flows for our assets are matched, on average, over time and across interest rate scenarios to those the estimated cash flows of our liabilities. A positive duration gap indicates that the duration of our assets exceeds the duration of our liabilities. We disclose duration gap on a monthly basis under the caption “Interest Rate Risk Disclosures” in our Monthly Summary, which is available on our website and announced in a press release.
While our goal is to reduce the price sensitivity of our net portfolio to movements in interest rates, various factors can contribute to a duration gap that is either positive or negative. For example, changes in the market environment can increase or decrease the price sensitivity of our mortgage assets relative to the price sensitivity of our liabilities because of prepayment uncertainty associated with our assets. In a declining interest rate environment, prepayment rates tend to accelerate, thereby shortening the duration and average life of the fixed rate mortgage assets we hold in our net portfolio. Conversely, when interest rates increase, prepayment rates generally slow, which extends the duration and average life of our mortgage assets. Our debt and derivative instrument positions are used to manage the interest rate sensitivity of our mortgage assets. As a result, the degree to which the interest rate sensitivity of our mortgage assets is offset will be dependent upon, among other factors, the mix of funding and other derivative instruments we use at any given point in time.
Convexity reflects the degree to which the duration of the assets and liabilities in our net portfolio changes in response to a given change in interest rates. We use convexity measures to provide us with information about how quickly and by how much our net portfolio’s duration gap may change in different interest rate environments. The market value sensitivity measures presented in Table 64 below are impacted by both the duration and convexity of our net portfolio and, depending upon our position at that time, may not move symmetrically in response to hypothetical interest rate shocks.
The sensitivity measures presented in Table 64, which we disclose on a quarterly basis as part of our disclosure commitments with FHFA, are an extension of our monthly sensitivity measures. There are three primary differences between our monthly sensitivity disclosure and the quarterly sensitivity disclosure presented below: (1) the quarterly disclosure is expanded to include the sensitivity results for larger rate level shocks of plus or minus 100 basis points; (2) the monthly disclosure reflects the estimated pre-tax impact on the market value of our net portfolio calculated based on a daily average, while the quarterly disclosure reflects the estimated pre-tax impact calculated based on the estimated financial position of our net portfolio and the market environment as of the last business day of the quarter; and (3) the monthly disclosure shows the most adverse pre-tax impact on the market value of our net portfolio from the hypothetical interest rate shocks, while the quarterly disclosure includes the estimated pre-tax impact of both up and down interest rate shocks.
U.S. Securities and Exchange Commission
November 26, 2013
Page 10
Results of Interest Rate Sensitivity Measurements
Table 64 displays the pre-tax market value sensitivity profile of our net portfolio as measured on the last day of each period presented of December 31, 2012 and 2011. In addition, Table 64 also provides the daily average, minimum, maximum and standard deviation values for duration gap for the three months ended December 31, 2012 and 2011 and for the most adverse market value impact on the net portfolio for non-parallel and parallel interest rate shocks as of the last day of each period presented.for the three months ended December 31, 2012 and 2011. The effective duration gap was (0.1) months for the last three months of 2012, which was the same average duration for the last three months of 2011.
Table 64: Interest Rate Sensitivity of Net Portfolio to Changes in Interest Rate Level and Slope of Yield Curve(1)
As of December 31, | |||||||||||
2012 | 2011 | ||||||||||
(Dollars in billions) | |||||||||||
Rate level shock:(2) | |||||||||||
-100 basis points | $ | 0.8 | $ | 0.3 | |||||||
-50 basis points | 0.2 | 0.1 | |||||||||
+50 basis points | 0.1 | (0.1 | ) | ||||||||
+100 basis points | — | (0.4 | ) | ||||||||
Rate slope shock:(2) | |||||||||||
-25 basis points (flattening) | — | — | |||||||||
+25 basis points (steepening) | — | 0.1 | |||||||||
For the Three Months Ended December 31, 2012 | |||||||||||||
Duration Gap(3) | Rate Slope Shock 25 bps(2) | Rate Level Shock 50 bps(2) | |||||||||||
Exposure | |||||||||||||
(In months) | (Dollars in billions) | ||||||||||||
Average | (0.1) | $ | — | $ | 0.1 | ||||||||
Minimum | (0.8) | — | — | ||||||||||
Maximum | 0.4 | — | 0.2 | ||||||||||
Standard deviation | 0.3 | — | 0.1 | ||||||||||
For the Three Months Ended December 31, 2011 | |||||||||||||
Duration Gap(3) | Rate Slope Shock 25 bps(2) | Rate Level Shock 50 bps(2) | |||||||||||
Exposure | |||||||||||||
(In months) | (Dollars in billions) | ||||||||||||
Average | (0.1) | $ | — | $ | (0.1 | ) | |||||||
Minimum | (0.7) | — | (0.2 | ) | |||||||||
Maximum | 0.4 | — | — | ||||||||||
Standard deviation | 0.2 | — | 0.1 | ||||||||||
__________
(1) | Computed based on changes in U.S. LIBOR interest rate swap curve LIBOR swap rates. |
(2) | Measured on the last day of each period presented. |
(3) | Average duration gap represents the average of duration gaps that were measured on each day during each period presented. |
U.S. Securities and Exchange Commission
November 26, 2013
Page 11
The market value sensitivity of our net portfolio varies across a range of interest rate shocks depending upon the duration and convexity profile of our net portfolio. The average duration gap was (0.1) months for the last three months of 2012, which is consistent with the average duration gap for the last three months of 2011. Because the effective duration gap of our net portfolio was close to zero months in the periods presented, convexity risk was the primary driver of the market value sensitivity of our net portfolio in those periods.
In response to the Staff’s comment, we will also include a footnote to the following effect on our table regarding the derivative impact on interest rate risk in our 2013 Form 10-K :
Table 65: Derivative Impact on Interest Rate Risk (50 Basis Points)(1)
As of December 31, | |||||||||||
2012 | 2011 | ||||||||||
(Dollars in billions) | |||||||||||
Before Derivatives | $ | (0.5 | ) | $ | (1.3 | ) | |||||
After Derivatives | 0.1 | (0.1 | ) | ||||||||
Effect of Derivatives | 0.6 | 1.2 | |||||||||
__________
(1) | Measured on the last day of each period presented. |
U.S. Securities and Exchange Commission
November 26, 2013
Page 12
* * *
Fannie Mae acknowledges that:
• | Fannie Mae is responsible for the adequacy and accuracy of the disclosure in the filing; |
• | Staff comments or changes to disclosure in response to Staff comments do not foreclose the Commission from taking any action with respect to the filing; and |
• | Fannie Mae may not assert Staff comments as a defense in any proceeding initiated by the Commission or any person under federal securities laws of the United States. |
If you have any questions regarding our responses, you may contact me at (202) 752-5830.
Sincerely,
/s/ David C. Benson
David C. Benson
Executive Vice President and Chief Financial Officer
cc: | Bradley E. Lerman, Executive Vice President, General Counsel |
and Corporate Secretary, Fannie Mae
Gregory A. Fink, Senior Vice President and Controller, Fannie Mae
Robert Plesnarski, O’Melveny & Myers, LLP
U.S. Securities and Exchange Commission
November 26, 2013
Page 13
APPENDIX — Recovery Example
Background:
The following illustrative example is provided for a hypothetical loan that became 180 days past due and for which a charge-off was recorded. In this scenario, we assume that, prior to foreclosure, the borrower engaged with the servicer on loss mitigation alternatives and a modification was executed, for which the borrower has successfully completed the trial period.
Assumptions:
Outstanding Unpaid Principal Balance (UPB) | $ | 100,000 | |
Charge-Off Amount | 20,000 | ||
Recorded Investment | $ | 80,000 | |
Contractual Monthly Payments: | |||
Principal | $ | 300 | |
Interest | 700 | ||
Total | $ | 1,000 | |
Additional assumptions regarding the prior accounting period:
1) | Borrower has successfully completed the trial modification. Consistent with our current policy, we plan to use the cash basis method to account for recoveries during the trial period of the loan. This method would allocate payments based on the contractual schedule and would not accrete any of the charge-off. |
2) | Loan is current under its modified terms. |
3) | Period of re-performance has been sufficient to demonstrate borrower’s ability and willingness to repay. |
Example entries to record the contractual monthly payment of $1,000 received in the current accounting period:
1) Apply cash receipt according to the loan’s original contractual schedule
Dr. | Cash | $ | 1,000 | |||||
Cr. | Outstanding UPB | $ | 300 | |||||
Cr. | Interest Income | $ | 700 | |||||
2) Recognize Pro Rata Recovery of Charged-off Amount(1)
2a) Calculate the percent of UPB paid down with contractual payment
Payment applied | $ | 300 | A | |
Outstanding UPB | $ | 100,000 | B | |
UPB recovery percent | 0.30 | % | C = A/B | |
2b) Calculate Recovery Amount
UPB recovery percent | 0.30 | % | C | |
Charge-off Amount | $ | 20,000 | D | |
Pro Rata Recovery Amount | $ | 60 | E = C*D | |
2c) Record Entry
Dr. | Loan Charge-off Contra(2) | $ | 60 | |||||
Cr. | Allowance - Recoveries(3) | $ | 60 | |||||
U.S. Securities and Exchange Commission
November 26, 2013
Page 14
__________
(1) | We are using a pro rata recovery method as an illustrative example only. While we believe the pro rata method of recognizing recovery of previous charge-offs would be reasonable and appropriate, we acknowledge that other approaches that would recognize the recovery on a systematic and rationale basis may also be acceptable. |
(2) | This represents the cumulative amount of the charge-off for a loan which, for operational ease, is separately tracked from the principal due from the borrower. Loan Charge-off Contra is a component of the Loans Held for Investment line item in our Consolidated Balance Sheets. |
(3) | Recoveries of the amounts previously charged-off are recorded as income through an adjustment to the allowance for loan losses. |
Resulting Impact on Recorded Investment:
Prior to | Journal Entry | Subsequent to | |||||||||
Contractual Payment | Impact | Contractual Payment | |||||||||
Outstanding UPB | $ | 100,000 | $ | (300 | ) | $ | 99,700 | ||||
Charge-off Amount | (20,000 | ) | 60 | (19,940 | ) | ||||||
Recorded Investment | $ | 80,000 | $ | (240 | ) | $ | 79,760 | ||||
Reconciliation to Cash Received:
Interest Income Recognized | $ | 700 | |
Net Reduction in Recorded Investment | 240 | ||
Charge-off Recovery (income recognized through adjustment to the allowance for loan losses) | 60 | ||
Total (Equal to $1,000 Cash Receipt) | $ | 1,000 | |