VIA EDGAR
December 17, 2015
Ms. Stephanie L. Sullivan
United States Securities and Exchange Commission
Division of Corporation Finance
100 F Street, N.E.
Washington, D.C. 20549
Re: | Santander Consumer USA Holdings Inc. |
Form 10-K for the Fiscal Year Ended December 31, 2014
Filed March 2, 2015
Form 10-Q for the Quarter Ended September 30, 2015
Filed October 29, 2015
File No. 001-36270
Dear Ms. Sullivan:
This letter includes the responses of Santander Consumer USA Holdings Inc. (the “Company”) to the comments of the staff of the Securities and Exchange Commission (the “Commission”), Division of Corporation Finance (the “Staff”), regarding its Form 10-K for the fiscal year ended December 31, 2014 (the “Form 10-K”) and its Form 10-Q for the quarter ended September 30, 2015 (the “Form 10-Q”), which were delivered in a letter dated November 18, 2015. For your convenience, in this letter the text of the Staff’s comments is set forth in bold, italicized text followed by the responses of the Company.
Form 10-K for the Year Ended December 31, 2014
Item 7 – Management’s Discussion and Analysis of Financial Condition and Results of Operations, page 36
Deferrals and Troubled Debt Restructurings, page 52
1. | We note your response to prior comment 2 regarding the presentation of deferrals and modifications in certain of your tables. Please respond to the following: |
• | Clarify whether the balances shown in the table showing the principal balance of your loans that have received modifications and concessions represent the new modifications/concessions granted during the period, or the cumulative amount of loans that have received modifications and concessions existing in your loan portfolio at the balance sheet date. |
Ms. Stephanie L. Sullivan
December 17, 2015
Page 2 of 7
• | Clarify when you would remove loans that have been modified from the disclosure in this table. For example, are the loans only removed from this disclosure once they have been written-off, paid-off or sold, or could they be removed once the temporary modification was no longer in place? |
• | To the extent that loans that have been modified are removed from this disclosure once the term of the modification has expired, please consider providing a rollforward showing the movement in the modified loans during the period. For example, the rollforward could show the beginning balance, new modifications, transfers out due to expiration of modification period, write-offs, sales, and pay-downs. |
Response: The amounts shown in this disclosure reflect the current balance of all loans that have ever received a modification or concession. Once a loan has received a modification or concession, it is included in this disclosure until it is no longer included in our consolidated balance sheet due to charge-off, payoff, or sale. Therefore, a modified loan would not be removed from this table until its carrying value is zero.
Please note that most of these modified loans meet the accounting criteria to be classified as troubled debt restructurings (TDRs). Effective with our 10-Q for the quarter ended September 30, 2015, we now include a rollforward of TDRs in the footnotes to the financial statements. Because loans that have been classified as TDRs retain that classification until they are no longer included in our consolidated balance sheet, a line in this rollforward for “transfers out due to expiration of modification period” is not applicable.
Form 10-Q for the Quarter Ended September 30, 2015
Note 4 – Credit Loss Allowance and Credit Quality, page 15
2. | We note your disclosure that during the third quarter of 2015 you removed the volatility associated with seasonality assumption for the loss provisioning model for the individually acquired retail installment contracts, resulting in a $134.0 million decrease in the provision for credit losses and credit loss allowance. On page 46 of your 10-Q, you state that you believe removing the seasonality assumptions from your forecast will provide more meaningful information in evaluating the impact that estimated future credit losses have on your business on a periodic basis. Please respond to the following: |
• | Clarify your accounting basis for this change. For example, clarify how you determined that you either no longer experienced this seasonality, or were somehow timely capturing this effect in another manner to ensure that at the balance sheet date your allowance for loan losses captured all incurred losses as of that date. In this regard, we note statements from your Interim CFO on the Q3 earnings call that you are pleased with your year-to-date performance as it positions you well for the remainder of the year when you have “historically experienced seasonally worse credit performance” and statements from your CEO that “losses and operating expenses are expected to continue to increase through the end of the year in line with typical seasonal patterns.” |
Ms. Stephanie L. Sullivan
December 17, 2015
Page 3 of 7
• | Tell us all of the factors and data you considered in concluding that the seasonality assumption was no longer necessary. |
• | Clarify whether any other model changes were made during the period, or whether any underlying data sources were changed in connection with this change. |
• | We note that your interim CFO noted on the Q3 earnings call that total provisions for loan losses for the year would not change due to removal of the seasonality assumption, but provision expense would shift from quarter-to-quarter. Please explain in detail how you determined this shift was appropriate, and resulted in a methodology that timely captures all incurred losses in the portfolio at the appropriate time. |
Response: The credit loss allowance is maintained to cover incurred losses that are probable and estimable on the date of evaluation. We estimate incurred losses using the concept of a Loss Emergence Period (LEP), or estimated average length of time from occurrence of a loss trigger to charge-off. Most recently, this LEP was estimated to be 16.5 months, and we therefore record a credit loss allowance equal to the next 16.5 months of projected net charge-offs on the portfolio existing as of the balance sheet date.
Our business experiences significant seasonality in net charge-offs, particularly in the lower credit quality portions of our portfolio. A 16.5-month forecast of actual net losses amplifies these seasonal effects by including two 4.5-month stretches of the calendar year in the forecast (the same calendar months appeared in months 1 through 4.5 of the forecast and again from 13 through 16.5) while including the other 7.5 months of the calendar year only once. Although this approach is
Ms. Stephanie L. Sullivan
December 17, 2015
Page 4 of 7
consistent with applicable guidance, which does not specify whether actual or cyclically adjusted expected losses should be included, this computational feature of our legacy provisioning methodology emphasized higher loss periods of the year at certain balance sheet dates and emphasized lower loss periods of the year at others. Said another way, even assuming consistency in portfolio size and credit quality, a large movement in credit loss allowance would occur from quarter to quarter simply due to the varying emphases at various points in the year. These large swings in credit loss allowance levels were confusing to financial statement users and made it more difficult to discern underlying credit trends.
Because the average life of our retail installment contracts extends over multiple years, not quarters, we determined that it would be more appropriate, and more transparent to investors, to take an annualized view of losses when forecasting for purposes of credit loss provisioning rather than focusing on quarter-by-quarter moves that are not as relevant to the long-term performance of the assets. We believe that an annualized view of losses is most appropriate to capture our forecasted incurred losses through a cycle, as our higher-loss and lower-loss periods are consistent from year to year. It is important to note that the forecast still includes 16.5 months of projected net losses, and still captures an element of seasonality, but the seasonality effect is captured based on an annual outlook rather than the more volatile quarterly outlook.
Further, our actual net charge-offs, which also effectively comprise a portion of our provision as the allowance associated with the charge-offs are generally replenished, will continue to exhibit the seasonality described in the above-referenced comments from our CEO and interim CFO. As a result, our provision for credit losses will continue to exhibit a quarterly seasonality trend, but our credit loss allowance (balance sheet amount) no longer will exhibit quarterly trends in seasonal charge-offs. When considering the impact of seasonality on our operating results, we believe that utilizing the new approach, combining an annualized view of charge-off expectations for the allowance with actual quarterly charge-offs, is most appropriately reflective of our seasonal charge-off trends.
During the third quarter, we also enhanced the methodology used to estimate the amount of TDR impairment, as further described in our response to Comment 3 below. There were no other model changes made during the period, and underlying data sources did not change.
It is important to note that although there was a benefit to third quarter 2015 earnings from making the change in seasonality impact as of September 30, this benefit purely represents timing and does not represent a permanent increase to cumulative reported earnings. For example, because the quarterly effect of seasonality on the allowance has now been removed, the benefit, typically seen in
Ms. Stephanie L. Sullivan
December 17, 2015
Page 5 of 7
our historical fourth quarter financial statements, of moving from emphasizing the high net charge-off ratio for the October-February period twice in the LEP as of September 30 to emphasizing the lower net charge-off ratio for the January-May period twice in the LEP as of December 31 will not recur this year. Said more broadly, the amount of credit loss allowance level in any given four-quarter period, all else being equal, is not impacted by the removal of seasonality, but rather simply spread differently (and, importantly, we believe more appropriately) among the four quarters. We continue to record a credit loss allowance covering estimated incurred losses, but on an annualized view more appropriate for assets with a multi-year life, rather than with the seasonal volatility previously inherent in the allowance methodology.
3. | We note per your disclosure on page 19 that the allowance for loan losses on your retail installment contract TDR portfolio has increased by 76% or $605.0 million during the nine months ended September 30, 2015, while your outstanding recorded investment in retail installment contract TDRs has only increased $178.4 million during the same period. We also note that the total delinquent TDR principal for your retail installment contracts has decreased by 10% during the same period. Given your policy disclosure on page 19 that states that once a loan is classified as a TDR, it is assessed for impairment based on the present value of expected future cash flows discounted at the loan’s original effective interest rate considering all available evidence, please tell us in detail the specific factors and drivers of the increase in the allowance for your TDRs. As part of your response, please also tell us how you concluded that the factors driving the larger allowance for your TDRs did not also result in the need for higher allowances in your non-TDR portfolio, which decreased by 9% during the same period, despite the substantial growth in the portfolio. |
Response: We recently began the development of updated models for eventual use in our credit loss provisioning process, including changes to the methodologies for determining both the impairment for TDRs and the credit loss allowance for non-TDRs. Historically, we have applied the LEP to our entire retail installment contract portfolio, then separately determined the appropriate impairment for the TDR subset of that portfolio, and deducted that impairment from the total allowance to arrive at a credit loss allowance estimate for our non-TDRs. We have determined that an appropriate alternate approach would be to segment the portfolio into two populations that each have unique loss timing characteristics, applying an LEP to only the non-TDR portion of the portfolio and adding this result to the separately determined TDR impairment amount to arrive at an aggregate credit loss allowance.
As of the third quarter of 2015, the new approach had not yet been fully implemented, as the new models were pending the final steps of our model
Ms. Stephanie L. Sullivan
December 17, 2015
Page 6 of 7
governance process, including validation of all models and completion of parallel testing. However, in the course of the models’ development we identified certain aspects that could provide immediate enhancement to our financial reporting and disclosure process, and implemented those aspects during the third quarter. One such aspect was the seasonality change described in our response to Comment 2 above. The other aspect implemented during the third quarter was an enhancement of the methodology by which we estimate TDR impairment. We continue to apply the guidance prescribed in ASC 310-40 of comparing the present value of expected future net cash flows, discounted using the loan’s original contractual interest rate considering all available evidence, to the net recorded investment in the loan, including accrued interest. However, when segmenting the TDR population as described above, during the third quarter we made enhancements to our TDR cash flows forecasting methodology that impacted the expected timing of cash flows and consequently the resulting impairment figure.
The primary enhancement that changed the expected timing of cash flows was an alignment of the net loss forecast with other aspects of our provisioning methodology. We now are using a discrete, TDR-specific segmentation of our overall retail installment contract vintage loss model to estimate the amount and timing of losses that should be input to the cash flow model used to determine TDR impairment. Previously, we modeled net losses for the TDR impairment calculation using an approach that had different loss timing and recovery rate assumptions.
The legacy provisioning methodology, adjusted for seasonality as described in our response to Comment 2 above, and the new models being developed and tested produce materially consistent results for the aggregate credit loss allowance for our retail installment contract portfolio as a whole. This consistency of results is attributable to the proposed disaggregated approach to LEP, such that the non-TDR portion of the portfolio uses a different LEP than the 16.5 months applied to the portfolio as a whole. Our aggregate credit loss allowance continues to be adequate and appropriate, and the increase in percentage coverage on our TDRs and decrease in percentage coverage on our non-TDRs simply reflect a shift in allocation among these portfolios.
* * *
Ms. Stephanie L. Sullivan
December 17, 2015
Page 7 of 7
The Company acknowledges that:
• | the Company is responsible for the adequacy and accuracy of the disclosure in the Form 10-K and Form 10-Q; |
• | the Staff’s comments or changes to disclosure in response to the Staff’s comments do not foreclose the Commission from taking any action with respect to the Form 10-K and Form 10-Q; and |
• | the Company may not assert the Staff’s comments as a defense in any proceeding initiated by the Commission or any person under the federal securities laws of the United States. |
The Company trusts that the foregoing has been responsive to the Staff’s comments. If you have any questions related to this letter, please contact the undersigned at (214) 634-1110 or by email at jmdavis@santanderconsumerusa.com.
Sincerely, | ||
/s/ Jennifer Davis | ||
Name: | Jennifer Davis | |
Deputy Chief Financial Officer | ||
Santander Consumer USA Holdings Inc. |
cc: | Jason A. Kulas |
Chief Executive Officer
Santander Consumer USA Holdings Inc.