May 1, 2014
Mr. John P. Nolan, Senior Assistant Chief Accountant
Securities and Exchange Commission
Division of Corporation Finance
Washington, D.C. 20549
Re: First Citizens BancShares, Inc.
Form 10-K for the Year Ended December 31, 2013
Filed February 26, 2014
File No. 001-16715
Dear Mr. Nolan,
This letter is provided in response to your letter dated April 4, 2014 on behalf of the Securities and Exchange Commission (the Commission) based on a review of the First Citizens BancShares, Inc.’s (“BancShares”) Form 10-K for the fiscal year ended December 31, 2013 (Form 10-K).
We acknowledge that:
Ÿ | BancShares is responsible for the adequacy and accuracy of the disclosures contained in the Form 10-K; |
Ÿ | Staff comments or changes to disclosure in response to staff comments do not foreclose the Commission from taking any action with respect to the Form 10-K; |
Ÿ | BancShares may not assert staff comments as a defense in any proceeding initiated by the Commission or any person under the federal securities laws of the United States. |
Comments from your April 4, 2014 letter and our responses are provided below, numbered to conform with the letter.
Form 10-K for Fiscal Year Ended December 31, 2013
Noninterest income, page 35
1. You state: “beginning on page 35 that during 2013 you experienced large recoveries of loan balances that were previously charged off and that these recoveries were recorded as noninterest income rather than as an adjustment to the allowance for loan and lease losses since charge-offs on acquired loans are primarily recorded through the nonaccretable difference. Please tell us in further detail how you accounted for these loans subsequent to your acquisition, including but not limited to interest accretion, recognition of provisions loan losses and charge-offs, and upon receipt of payments subsequent to charge-off. Please consider providing us with journal entries to better understand your accounting along with the authoritative guidance you relied upon to support your accounting if different than the guidance in ASC 310-30-35-10 and 35-11.”
BancShares is following the guidance included in ASC 310-30-35-10 and 35-11 for the acquired loans with deteriorated credit quality.
Specifically related to the recoveries referenced in the comment letter, these are all acquired loans with deteriorated credit quality that have been fully charged off. These loans are a combination of loans charged off prior to acquisition, loans valued at or close to zero in day-one accounting, and more recently, loans fully charged off and no longer on the balance sheet, which have provided subsequent recoveries. BancShares estimates cash flows and collections on an individual loan basis for all of the acquired loans rather than by homogeneous loans pools. As a result, there is no remaining allowance or loan balance for which the subsequent recoveries could be applied. BancShares has chosen to recognize subsequent recoveries through noninterest income since the loans are no longer on the balance sheet, and unlike originated loans, they do not represent an ongoing portfolio with continuous originations or renewal activity. Therefore, recognition of recoveries on these loans as yield adjustments or recoveries to the allowance could potentially mislead users of the financial statements about the level of recoveries to expect on the ongoing, originated portfolio in subsequent periods. Based on these facts, management concluded that recovery amounts should be immediately recognized in income as collected. In making this determination, management considered the following alternatives:
• | Accretion income - Management concluded that since any accretable yield or nonaccretable difference was removed at the time the loan was charged off, or the loans were never recorded on the books (i.e. charged off at acquisition), it was inappropriate to attribute the recovery-related income as accretion income. Treating the recoveries as accretion income would also inappropriately inflate margins as there would be an increase to income (numerator) without a corresponding increase to total loans (denominator); |
• | Allowance for loan and lease losses - Management concluded that since there was no allowance for loan and lease losses related to a fully charged off loan, it was inappropriate to adjust the allowance for a recovery; an adjustment to the allowance would result in an overstatement of the allowance for loan and lease losses that would require an immediate reduction in provision expense for an event that did not relate to an outstanding loan balance; |
• | Noninterest income - Having determined that adjustments to accretion income and the allowance were inappropriate, management concluded that the recoveries related to previously charged off acquired loans are appropriately classified as a component of noninterest income; |
The Company will revise the discussion regarding the treatment of these revenues in future filings to more clearly explain that these recoveries are related to acquired loans with deteriorated credit quality that have been fully charged off.
Note C - Loans and Leases, page 81
2. “We note disclosure on page 86 of the changes in the carrying value of acquired loans. Please tell us and consider revising your future filings to disaggregate the reductions to the carrying value separately for repayments, foreclosures, and decreases in fair value.”
We believe that the acquired loan disclosures in our 2013 Form 10-K conform to the disclosure requirements under ASC 310-30. Prospectively, we will expand the acquired loan rollforward to reflect the impact of accretion on the acquired loan balance. We believe this will help the reader recognize the rate at which these loans are being repaid.
Following is a pro forma presentation of the 2013 acquired loan rollforward reflecting the expanded presentation.
2013 | 2012 | ||||||
Balance at January 1 | $ | 1,809,235 | $ | 2,362,152 | |||
Accretion | 224,672 | 304,023 | |||||
Payments received and other changes, net | (1,004,481 | ) | (856,940 | ) | |||
Balance at December 31 | $ | 1,029,426 | $ | 1,809,235 |
Note D - Allowance for Loan and Lease Losses, page 87
3. “We note disclosure on page 88 of the quantitative reclassification adjustments made during the second quarter of 2013 as a result of enhancements to your allowance for loan losses (ALLL) as further described in Note A and on pages 17 and 44. Due to the significant fluctuations that have resulted from these ALLL reclassifications, please tell us and revise your future filings to provide a more detailed understanding, by portfolio segment, of the reasons for these fluctuations. Please also explain the impact, if any, to the provision for loan losses recorded during 2013 as we note significant differences in provisions for each portfolio segment when compared to the prior period and in total for 2013.”
As recognized in the comment letter, during 2013 we implemented enhancements to the process by which we estimate the allowance for loan and lease losses (ALLL) for our originated loans. More specifically, during the second quarter of 2013, we implemented an enhanced model which resolved some limitations of the prior model. Key improvements in the enhanced model include loan level loss assumptions incorporating historical detail, improved documentation and maintenance of key assumptions, and methodologies more closely aligned with industry leading practices.
In calculating reserves for originated commercial and noncommercial loan segments, our prior model included a qualitative amount not representative of losses on any specific loans and was not allocated to individual loan segments. Reserves for the commercial segment were calculated based on the total population of commercial loans at the aggregate level, and incorporated historical loss rates for all commercial loans based on the risk rating of the loans, regardless of type. The resulting total reserve was then allocated by management to the individual loan classes for disclosure purposes. Management calculated the loss rates for the higher risk-rated loans based in part on limited migration data focusing on historical net charge-offs for those commercial loans. Loss rates for lower risk-rated loans were also based on historical net charge-offs, but included qualitative factors and considerations. Similarly, noncommercial segment reserves were calculated on an aggregated basis through loss rates based on historical net charge-offs, but also employing other qualitative considerations. The unallocated reserve amounts applied to the entire portfolio were based on external and internal factors (e.g., economic conditions, loan concentrations) applied to our total loan portfolio, not individual segments or classes.
Our enhanced model provides more granular reserve estimates for the originated commercial and noncommercial loan segments. Originated commercial loan and lease reserves are estimated using an average loss rate methodology, by loan type, using historical individual loan detail. This methodology captures default trends and loss severities observed in our loan portfolios, and uses that data to develop loss rates based on the risk ratings of the loans. The enhanced model employs more sophisticated migration techniques for commercial loans, allowing us to build loss rates for all risk ratings, by loan class, incorporating historical individual loan level detail. Reserves for originated noncommercial loans and leases are computed through a delinquency-based roll rate methodology, by loan type, again using historical individual loan level detail. As with commercial, this methodology captures delinquency trends and loss severities occurring in the portfolios to calculate reserves. The enhanced model also considers the impact of changes in external and internal qualitative factors at a loan class level; under the prior model,
such factors were applied at the portfolio level. Qualitative reserves are added to the quantitative results of our model based on changes in qualitative factors specific to each loan class.
In implementing the enhanced model, we noted an increase in the allowance for the originated commercial loan segment and a reduction in the allowance for the originated noncommercial loan segment. The enhanced model drove this shift, identifying more inherent risk in the construction and land development classes - commercial, commercial mortgage, and commercial and industrial - than our prior model. We concluded this shift was appropriate based on the improved calculations of default probabilities within these segments. Conversely, our enhanced model identified less inherent risk in the loan classes of revolving mortgage and consumer. Again, we attributed the shift to the enhanced model’s superior use of recent delinquency aging trends at the individual loan level together with historical net charge-offs. The less-sophisticated prior model focused primarily on historical net charge-offs.
Through analysis of the results of our enhanced model as discussed, we determined that these reclassifications were appropriate. Our total allowance did not materially change with these enhancements as we stated in our filings. We do believe, however, that we now have an allocation of reserves that is better supported by historical data related to our loan segments and classes, rather than defaulting to aggregate data, enabling us to allocate the previously unallocated reserves to specific loan classes, and to provide a more detailed assignment of the total required reserves.
The significant differences in provision expense for each segment and class (compared to the prior period and in total for 2013) are driven by both improvements in credit quality trends during 2013 and the enhanced model. Once we reclassified the existing and unallocated reserves across the loan segments and classes, provisions for loan losses were determined based on the results of the enhanced modeling process. We noted generally improving credit trends across both commercial and noncommercial segments, as evidenced by significantly lower net charge-offs in 2013 compared to 2012, resulting in a lower total provision expense for 2013. Management believes the stated provisions for loan losses by loan class is appropriate based on the credit trends within each class and our improved modeling process.
The Company will enhance its disclosures in future filings regarding significant changes in the allowance.
4. “As a related matter, we note that the ALLL for your revolving mortgage loans, as a result of your reclassification adjustments decreased $9.8 million during 2013. Noting that overall loan balances and delinquency trends have remained consistent during 2013 when compared to 2012 in addition to the revolving mortgage loans disclosures on page 41, please tell us and revise your future filings to fully explain the reasons for the significant reclassification adjustment to your revolving mortgage loan ALLL during 2013.”
As discussed in our response to Comment #3, we implemented an enhanced modeling process for both our originated commercial and noncommercial loan segments. Our prior model primarily focused on historical net charge-offs in determining reserves at a portfolio level along with other qualitative considerations, but could not distinguish between 1st and 2nd lien loans. As such, the prior model included higher qualitative considerations for the revolving mortgage class due to its limitations which among other items included the inability to capture recent credit trends and inherent losses at a loan class level. Higher qualitative adjustments in the prior model were deemed necessary based on loss experience during the recession, real estate valuation trends in our markets, increased regulatory focus on revolving mortgages, and model imprecision risk.
The enhanced modeling process allows us to better capture inherent risk at the loan class level by a more sophisticated use of trend data. For the revolving mortgage class in the noncommercial segment, we were able to incorporate recent delinquency-aging trends at the individual loan level and historical net charge-offs as key assumptions. Using individual loan level detail, we now are also able to incorporate lien position into the reserve calculation for revolving mortgages and better identify inherent risk. Our first lien revolving mortgages total $903.2 million, for which we have $2.7 million in reserves or 0.30% as of December 31, 2013. Our second lien revolving mortgages total $1.2 billion for which we have $13.2
million in reserves, or 1.10% as of December 31, 2013. This reduction in reserves was due to the implementation of the enhanced model and improved credit quality, as evidenced by a reduction in net charge-offs of almost 50% below the prior year.
Due to the increased granularity and precision provided by the enhanced model, we are able rely more heavily on our historical data, and less on qualitative factors. The results of this analysis provided increased transparency allowing us to further reduce qualitative considerations within the first lien loan class.
As is apparent, the drivers behind the $9.8 million reclassification in revolving mortgages are similar to, if not the same as, the factors cited in our response to Comment #3. In future filings, we will enhance disclosures regarding significant changes in the allowance.
5. “We note disclosure on page 90 that the total recorded investment for your originated impaired loans was $202.3 million as of December 31, 2012. We note, however, that the total recorded investment for your originated impaired loans disclosed on page 94 of your Form 10-K for the fiscal year ended December 31, 2012 totaled $172.9 million. Please tell us and consider revising your future filings to explain the reason(s) for the difference between these disclosed impaired loans totals and all of your other impaired loan disclosure amounts (i.e. unpaid principal balances, related allowance recorded, YTD average balance, interest income recognized YTD).”
In its 2012 Form 10-K, BancShares reported $173.0 million in originated impaired loans, which included restructured loans deemed to be troubled debt restructurings (‘TDRs’) for regulatory reporting purposes as of December 31, 2012. The 2012 10-K excluded, however, $29.3 million of previously restructured loans that were no longer considered to be TDRs under regulatory reporting rules based upon sustained post-modification performance. Upon further consideration during 2013, we determined it appropriate to include the $29.3 million in our impaired loan disclosures. The December 31, 2012, data was therefore adjusted in the third quarter 2013 10-Q and the subsequent Form 10-K to include those TDRs. We further concluded that these impaired loan disclosure adjustments were not material and did not warrant amendment of prior periodic reports.
We hope these responses resolve the Staff’s questions and concerns. Please contact our Chief Accounting Officer Lorie K. Rupp at 919/716-4152 if you have any further questions or we can provide further information.
Sincerely,
Glenn D. McCoy
Chief Financial Officer