PRIVILEGED & CONFIDENTIAL — ATTORNEY WORK PRODUCT Prepared at the direction of counsel in connection with anticipated litigation
EXHIBIT A
to the Wells Submission of Belmont Ridge Bank National Association
EXPERT REPORT OF DR. FRANKLIN OSEI, CPA, CFF
Analysis of Belmont Ridge Bank National Association's Allowance for Credit Losses Methodology for the BridgeFin Portfolio
Prepared by:
Brindoval Analytic Partners LLC 350 South Grand Avenue, Suite 2100 Los Angeles, California 90071
Author and Engagement Lead: Dr. Franklin Osei, CPA, CFF
Prepared at the Direction of:
Calderwood & Harkness LLP 425 Lexington Avenue, 30th Floor New York, New York 10017
In the Matter of:
SEC Division of Enforcement Investigation File No. HO-14287
April 5, 2024
PRIVILEGED & CONFIDENTIAL — ATTORNEY WORK PRODUCT Prepared at the direction of counsel in connection with anticipated litigation
TABLE OF CONTENTS
Right-click to update Table of Contents
I. Introduction and Purpose of This Report
A. Engagement and Scope
Brindoval Analytic Partners LLC ("Brindoval") was retained by Calderwood & Harkness LLP ("Calderwood & Harkness"), counsel for Belmont Ridge Bank National Association ("BRB" or the "Bank"), pursuant to an engagement letter dated October 10, 2023. The engagement was executed under the Kovel doctrine and the protections afforded by the attorney work-product doctrine, for the purpose of providing an independent forensic accounting analysis of BRB's methodology for estimating the allowance for credit losses ("ACL") applicable to its BridgeFin digital lending portfolio (the "BridgeFin Portfolio").
My analysis focuses on BRB's ACL as of December 31, 2022, as reported in BRB's Annual Report on Form 10-K for the fiscal year ended December 31, 2022, filed with the U.S. Securities and Exchange Commission (the "Commission" or "SEC") on February 28, 2023 (the "FY 2022 10-K"). In addition, I have examined the related disclosures and ACL figures reported in BRB's Quarterly Reports on Form 10-Q for the quarterly periods ended March 31, 2023 (filed May 10, 2023) and June 30, 2023 (filed August 9, 2023) (the "Q1 2023 10-Q" and "Q2 2023 10-Q," respectively, and collectively with the FY 2022 10-K, the "Filings at Issue").
This report has been prepared in connection with the Commission's Division of Enforcement investigation designated as File No. HO-14287 (the "Investigation"), and is submitted as Exhibit A to the Wells Submission of Belmont Ridge Bank National Association. The opinions expressed herein are based upon my professional training, education, and over twenty-five years of experience in forensic accounting, credit-loss modeling, and bank regulatory accounting, as well as my review of the materials enumerated in Section I.C below. I reserve the right to supplement or revise my opinions should additional information become available.
B. Summary of Conclusions
Based on my review of the materials described in Section I.C, my independent analysis of BRB's CECL model and supporting data, and my understanding of the applicable accounting standards and regulatory guidance, I have reached the following principal conclusions:
1. BRB's reported ACL of $143.5 million (3.50% of $4.1 billion in outstanding BridgeFin loans as of December 31, 2022) falls within the reasonable range of estimates permissible under ASC 326 (Financial Instruments — Credit Losses), commonly referred to as CECL. BRB's vintage-cohort loss-rate methodology, its qualitative factor overlays, and its underlying assumptions were reasonable, internally consistent, and appropriately documented. The resulting ACL constitutes a supportable point estimate within the spectrum of outcomes that a reasonably diligent financial institution could have reached under the same facts and circumstances.
2. BRB's treatment of 12,847 re-aged/modified loans with an aggregate outstanding principal balance of $236.4 million was methodologically supportable and consistent with OCC supervisory guidance. BRB's practice of reclassifying modified loans that demonstrated sustained re-performance (defined as three consecutive on-time payments) from their original delinquency cohorts to performing cohorts was grounded in empirical analysis of borrower behavior, was consistently applied across reporting periods, and reflects a permissible exercise of professional judgment under ASC 326.
3. The SEC Staff's preliminary view that the ACL should have been approximately $180.8 million (4.41% of outstanding BridgeFin loans) reflects an alternative but not exclusively correct methodological approach. The $37.3 million difference between the Staff's preliminary figure and BRB's reported ACL ($180.8 million minus $143.5 million) reflects the inherent range of reasonable estimates under CECL and the exercise of different—but equally permissible—judgments regarding the treatment of re-aged loans. This difference does not constitute a misstatement, either individually or in the aggregate.
4. BRB's reported net charge-off rate of 2.14% is supportable under BRB's methodology. The Staff's alternative figure of approximately 3.05% results from a different treatment of re-aged loan vintages and does not invalidate BRB's approach. BRB's reported figure accurately reflects the actual credit losses realized during the reporting period.
5. BRB's 90+ day delinquency rate of 1.72%, as reported in the FY 2022 10-K, reflects standard banking-industry practice of classifying loans based on current payment status at the reporting date. The Staff's alternative figure of 2.38% retroactively imputes delinquency to loans that were performing as of the balance-sheet date.
6. There is no basis to conclude that BRB's ACL judgments were made other than in good faith by qualified professionals operating within established internal controls and subject to external audit oversight. BRB's methodology, assumptions, and resulting estimates were reviewed and approved by the ACL Governance Committee, the Chief Financial Officer, the Board Audit Committee, and BRB's independent external auditor.
7. Peer comparison analysis confirms that BRB's ACL coverage ratio, charge-off rates, and delinquency metrics fall within the interquartile range of similarly situated digital-lending banks, further supporting the reasonableness of BRB's estimates.
C. Materials Reviewed
In forming my opinions, I reviewed and considered, among other things, the following materials:
• BRB's Form 10-K for the fiscal year ended December 31, 2022, filed February 28, 2023;
• BRB's Forms 10-Q for the quarterly periods ended March 31, 2023 (filed May 10, 2023), and June 30, 2023 (filed August 9, 2023);
• BRB's Form 10-K for the fiscal year ended December 31, 2023, filed February 27, 2024;
• BRB's Form 10-Q for the quarterly period ended September 30, 2023, filed November 9, 2023;
• BRB internal CECL model documentation, including vintage-cohort loss-rate models, qualitative factor overlay documentation and calibration memos, and model governance committee minutes for Q4 2021 through Q3 2023;
• BridgeFin loan-level data tapes as of December 31, 2022, including modification and re-aging records for 12,847 individual loans;
• BRB internal credit quality reports, charge-off schedules, and delinquency aging analyses for the BridgeFin Portfolio for fiscal years 2021, 2022, and the first three quarters of 2023;
• BRB's ACL estimation workpapers prepared by Raymond Hu, Chief Accounting Officer, and the BRB credit risk analytics team;
• Audit workpapers of Whitfield Keane LLP (PCAOB Registration No. 2847) relating to the ACL for the BridgeFin Portfolio for the fiscal years ended December 31, 2021, and December 31, 2022, as made available through counsel;
• The SEC Staff's preliminary analysis as communicated during the Wells process, including notes from the March 8, 2024, telephone conference with Branch Chief Renata Moss and Assistant Director Gregory Bonham;
• Written communications between the Staff (including Staff Accountant Diane Milosevic) and BRB's counsel concerning the Staff's preliminary views on the BridgeFin ACL methodology;
• The Office of the Comptroller of the Currency ("OCC") Comptroller's Handbook — Allowance for Credit Losses (April 2023 edition), relevant sections;
• ASC 326 (FASB Accounting Standards Codification, Topic 326 — Financial Instruments — Credit Losses), including ASU 2016-13 (Measurement of Credit Losses on Financial Instruments) and related implementation guidance;
• ASU 2022-02 (Financial Instruments — Credit Losses (Topic 326): Troubled Debt Restructurings and Vintage Disclosures);
• FASB Transition Resource Group for Credit Losses meeting minutes and implementation guidance (2018–2022);
• SEC Staff Accounting Bulletin No. 99 (Materiality);
• SEC Staff Accounting Bulletin No. 102 (Selected Loan Loss Allowance Methodology and Documentation Issues);
• PCAOB Auditing Standard No. 2501 (Auditing Accounting Estimates, Including Fair Value Measurements);
• Peer institution public filings (12 digital-lending bank comparators, identified in Exhibit B to the Wells Submission), including 10-K and 10-Q filings for the fiscal year ended December 31, 2022, and related quarterly periods;
• Published academic and industry research on consumer loan modification outcomes and re-performance rates; and
• Such other documents and data as are referenced throughout this report.
I reserve the right to supplement my analysis and opinions should additional materials become available.
II. Background: Belmont Ridge Bank and the BridgeFin Portfolio
A. Overview of BRB
Belmont Ridge Bank National Association is a national banking association organized and existing under the laws of the United States, operating under OCC Charter No. 24891. The Bank is headquartered at 200 Park Avenue South, 38th Floor, New York, New York 10003. BRB's common stock is publicly traded on the Nasdaq Stock Market under the ticker symbol "BRDG." As of September 2023, BRB had an approximate market capitalization of $4.7 billion.
BRB is a diversified financial institution offering a broad range of banking products and services to consumer and commercial customers, including deposit accounts, residential mortgage lending, commercial real estate lending, consumer lending (including through the BridgeFin division), treasury management services, and wealth management. The Bank operates through a network of approximately 120 branch locations in the northeastern United States, as well as through digital channels.
As of December 31, 2022, the Bank reported the following key financial metrics:
BRB's most recent OCC examination, completed in September 2023, assigned the Bank a composite CAMELS rating of 2 (Satisfactory), reflecting the OCC's assessment that the Bank was fundamentally sound and exhibited only moderate supervisory concerns. Notably, the OCC's examination encompassed a review of BRB's credit risk management practices and allowance for credit losses methodology, and the examination did not result in an adverse finding regarding the BridgeFin ACL.
BRB adopted ASC 326 (CECL) effective January 1, 2020, in accordance with the standard's mandatory adoption timeline for SEC filing entities. The Bank recorded a cumulative-effect adjustment to retained earnings upon adoption and has applied the CECL framework to all of its loan portfolios, including the BridgeFin Portfolio, since that date. BRB's CECL implementation was the subject of extensive internal planning and external consultation, and the Bank's adoption methodology was reviewed by its external auditor, Whitfield Keane LLP, in connection with the fiscal year 2020 audit.
B. BridgeFin Division
BridgeFin is a digital-first consumer lending platform operated as an internal division—not a separate legal entity—of BRB. The division was launched in the third quarter of 2020 as part of BRB's strategic initiative to expand its digital banking capabilities and reach a broader consumer and small-business customer base. The division is headed by Naveen Chakravarti, Senior Vice President, who reports to BRB's Chief Banking Officer.
BridgeFin offers two primary lending products:
(a) Unsecured personal loans, with an average loan size of approximately $18,400, targeting consumers with FICO scores generally ranging from 620 to 750, with loan terms of 24 to 60 months; and
(b) Small-business microloans, with an average loan size of approximately $42,000, targeting small businesses and sole proprietors with annual revenues generally below $2 million, with loan terms of 12 to 36 months.
BridgeFin utilizes a proprietary AI-driven underwriting model for origination and credit decisioning. The model incorporates traditional credit bureau data (FICO scores, payment history, credit utilization) as well as alternative data sources (bank transaction data, cash-flow analysis, and employment verification through third-party APIs). The underwriting model was developed internally by BRB's data science team and has been subject to fair lending testing, model validation, and ongoing performance monitoring consistent with OCC supervisory expectations for model risk management (OCC Bulletin 2011-12).
During fiscal year 2022, BridgeFin originated approximately $2.87 billion in new loans. As of December 31, 2022, the BridgeFin Portfolio had an aggregate outstanding principal balance of approximately $4.1 billion, consisting of approximately 183,000 individual loans. The BridgeFin Portfolio represented approximately 10.7% of BRB's total consolidated assets and approximately 15.3% of BRB's total loan portfolio as of that date.
The BridgeFin division has experienced rapid growth since its 2020 launch. Origination volumes increased from approximately $780 million in the fourth quarter of 2020 and fiscal year 2021 combined to $2.87 billion in fiscal year 2022, reflecting both strong consumer demand for digital lending products and BRB's strategic investment in the platform's technology and marketing infrastructure. This growth trajectory is relevant to the ACL analysis because rapid portfolio growth can create challenges in calibrating loss estimates for younger vintage cohorts that lack full seasoning history.
C. The SEC Investigation
The SEC Division of Enforcement opened its investigation of BRB on September 18, 2023, through a voluntary request letter addressed to BRB's General Counsel, Marcus Webb, requesting the production of documents and information relating to BRB's ACL methodology for the BridgeFin Portfolio. The voluntary request focused on the Bank's treatment of modified and re-aged loans within its vintage-cohort loss-rate models and the resulting impact on reported ACL figures, net charge-off rates, and credit-quality disclosures.
On October 12, 2023, the Commission issued a Formal Order of Investigation designated as File No. EN-2023-0947, authorizing the Division of Enforcement to conduct a formal investigation into whether BRB's financial statements and related disclosures violated the federal securities laws, including Sections 10(b) and 13(a) of the Securities Exchange Act of 1934 and Rules 10b-5, 13a-1, and 13a-13 thereunder.
The core question at the center of the Investigation is whether BRB's ACL for the BridgeFin Portfolio materially understated expected credit losses, with particular focus on the treatment of re-aged and modified loans within the Bank's vintage-cohort loss-rate models. The Staff's preliminary view, as communicated during the Wells process—including a telephone conference held on March 8, 2024, with Branch Chief Renata Moss and Assistant Director Gregory Bonham—is that BRB should have retained re-aged loans in their original vintage delinquency buckets, which would have resulted in a higher ACL and different reported credit metrics.
As set forth in detail in this report, it is my opinion that the Staff's preliminary view reflects a permissible alternative methodology but does not establish that BRB's methodology was incorrect, unreasonable, or inconsistent with applicable accounting standards and regulatory guidance.
III. Applicable Accounting Framework: ASC 326 (CECL)
A. Overview of CECL
ASC 326, Financial Instruments — Credit Losses, was promulgated by the Financial Accounting Standards Board ("FASB") through Accounting Standards Update No. 2016-13, Measurement of Credit Losses on Financial Instruments (the "CECL Standard"), issued on June 16, 2016. The CECL Standard became effective for SEC filing entities for fiscal years beginning after December 15, 2019, and BRB adopted it effective January 1, 2020.
The CECL Standard fundamentally changed the accounting for credit losses on financial assets measured at amortized cost. Under the prior accounting framework—the "incurred loss" model codified in ASC 450-20 (formerly FAS 5) and ASC 310-10-35 (formerly FAS 114)—entities recognized credit losses only when it was probable that a loss had been incurred as of the balance-sheet date. This backward-looking model was widely criticized during and after the 2008 financial crisis for producing allowances that were "too little, too late," as losses were recognized only after credit deterioration had already manifested.
CECL replaced the incurred-loss model with a forward-looking, lifetime expected-loss framework. Under ASC 326, an entity is required to measure expected credit losses on financial assets measured at amortized cost based on:
(a) Historical experience, including the entity's own loss history and, where relevant, the loss history of peer institutions;
(b) Current conditions, including the current state of the economy, borrower-specific credit quality indicators, and portfolio composition; and
(c) Reasonable and supportable forecasts of future economic conditions that may affect the collectibility of the reported amount.
The CECL Standard explicitly acknowledges that credit loss estimates involve significant management judgment and that a range of reasonable estimates may exist for any given portfolio at any given reporting date. This is not an incidental observation; it is a foundational feature of the standard. As stated in ASC 326-20-30-3:
"An entity shall not adjust the estimate of expected credit losses for the existence of a reasonable range of expected credit losses."
This provision means that management must select a single point estimate within the range of reasonable outcomes, but it is not required to select the highest, lowest, or midpoint estimate. The selection of a point estimate within the reasonable range is a matter of professional judgment, and different institutions—or, indeed, different qualified professionals within the same institution—may reasonably arrive at different estimates for the same portfolio based on the same underlying data.
This principle is central to the analysis in this report. The existence of the Staff's alternative estimate of $180.8 million does not, by itself, establish that BRB's estimate of $143.5 million was unreasonable or constituted a misstatement. Both estimates may fall within the range of reasonable outcomes under CECL, and the question for accounting purposes is whether BRB's estimate was arrived at through a reasonable process using supportable assumptions—not whether an alternative estimate could be derived using different but also supportable assumptions.
B. Methodologies Permitted Under ASC 326
ASC 326 does not prescribe a single methodology for estimating expected credit losses. The FASB deliberately adopted a principles-based approach that permits entities to use range estimation techniques, provided that the selected methodology is reasonable, consistently applied, and adequately documented. Acceptable approaches include, but are not limited to:
(a) Vintage analysis (tracking loss experience by origination cohort over time);
(b) Loss-rate methods (applying historical loss rates to current portfolio balances, adjusted for current conditions and forecasts);
(c) Probability-of-default / loss-given-default ("PD/LGD") models (estimating the probability that individual loans or segments will default and the expected severity of loss upon default);
(d) discounted cash flow ("DCF") approaches (estimating expected cash flows and comparing them to the amortized cost basis); and
(e) Hybrid methods that combine elements of two or more of the above approaches.
BRB employed a vintage-cohort loss-rate approach supplemented by qualitative factor overlays for the BridgeFin Portfolio. This methodology is widely used in the banking industry, particularly for consumer lending portfolios, and is recognized as appropriate by the FASB, the OCC, and the PCAOB. The vintage-cohort approach is well-suited to the BridgeFin Portfolio because the portfolio consists of a large number of relatively homogeneous loans that can be meaningfully grouped by origination period and tracked through their lifecycle.
Importantly, the selection of model inputs—including how to define and segment vintage cohorts, how to treat modified or re-aged loans within those cohorts, how to weight macroeconomic scenarios, and how to project forward loss rates for less-seasoned vintages—necessarily involves the exercise of professional judgment. Different but equally reasonable choices at each of these decision points can produce materially different ACL estimates, all of which may fall within the range of reasonable outcomes contemplated by the CECL Standard.
C. Qualitative Adjustments
ASC 326-20-30-8 provides that an entity's estimate of expected credit losses should reflect "adjustments for the differences in current and forecasted conditions as compared to the conditions that existed for the period over which historical information was evaluated." In practice, this provision is the basis for the qualitative factor adjustments—commonly referred to as "qualitative overlays" or "Q-factors"—that are applied to quantitative model outputs to account for factors not fully captured by historical loss data.
BRB applied qualitative overlays to the BridgeFin Portfolio's quantitative ACL estimate in each of the periods at issue. These overlays addressed, among other factors:
(a) Macroeconomic conditions, including changes in unemployment rates, consumer credit conditions, and housing market indicators that could affect the creditworthiness of BridgeFin borrowers;
(b) Portfolio growth, reflecting the risk that rapid origination growth may introduce adverse selection or reduced underwriting rigor that is not yet reflected in historical loss data; and
(c) Underwriting standard changes, reflecting adjustments to BridgeFin's credit scoring model and approval criteria that may have altered the risk profile of newly originated loans relative to earlier vintages.
The magnitude and direction of qualitative adjustments are inherently judgmental and are expected to vary among institutions based on their specific circumstances, risk appetite, and analytical assessments. The FASB, the banking regulators, and the PCAOB have all recognized that qualitative adjustments are an essential component of the CECL framework and that reasonable minds may differ on their appropriate calibration.
IV. BRB's ACL Methodology for the BridgeFin Portfolio
A. Vintage-Cohort Loss-Rate Model — Structure
BRB's ACL for the BridgeFin Portfolio was calculated using a vintage-cohort loss-rate model. The model segments the portfolio into quarterly origination vintage cohorts based on the quarter in which each loan was originated. Because the BridgeFin division launched in Q3 2020 and the analysis date is December 31, 2022, the model encompasses ten vintage cohorts: Q3 2020, Q4 2020, Q1 2021, Q2 2021, Q3 2021, Q4 2021, Q1 2022, Q2 2022, Q3 2022, and Q4 2022.
For each vintage cohort, BRB's credit risk analytics team calculated the following key metrics:
(a) Cumulative net charge-off rates, measuring the percentage of original cohort principal that had been charged off (net of recoveries) from origination through the reporting date;
(b) Delinquency migration rates, tracking the transition probabilities of loans moving between delinquency buckets (current, 30–59 days, 60–89 days, 90–119 days, 120+ days, and charge-off) from one reporting period to the next; and
(c) Expected remaining-life loss rates, projecting the additional credit losses expected to be incurred over the remaining contractual life of each cohort based on a combination of historical performance curves and forward-looking macroeconomic assumptions.
Loss rates for each cohort were projected forward using a blend of historical performance curves (derived from BRB's own BridgeFin data, supplemented by proxy data from BRB's traditional consumer lending book for earlier-stage extrapolation) and macroeconomic scenario weighting. BRB employed three macroeconomic scenarios for loss-rate projection:
(i) A baseline scenario, reflecting BRB's central expectation for economic conditions over the reasonable and supportable forecast period (12 months), weighted at 50%;
(ii) An adverse scenario, reflecting a moderately pessimistic economic outlook, weighted at 30%; and
(iii) A severely adverse scenario, reflecting a recessionary economic environment, weighted at 20%.
These scenario weights were determined by BRB's ACL Governance Committee based on prevailing economic indicators and consensus forecasts as of the reporting date. The scenario weights were documented in the committee's quarterly minutes and were reviewed by the Chief Financial Officer.
Beyond the 12-month reasonable and supportable forecast period, BRB reverted loss rates to long-run historical averages on a straight-line basis over a 6-month reversion period, consistent with the guidance in ASC 326-20-30-9.
The model outputs were aggregated across all ten vintage cohorts to produce a quantitative ACL estimate for the entire BridgeFin Portfolio. This quantitative estimate was then adjusted by the qualitative overlays described in Section III.C above to arrive at the final reported ACL.
B. Treatment of Re-Aged/Modified Loans
The treatment of re-aged and modified loans is the central issue in the Investigation, and I have devoted particular attention to this aspect of BRB's methodology.
Background on BridgeFin Loan Modifications. During fiscal years 2021 and 2022, BRB's BridgeFin division modified 12,847 loans with an aggregate outstanding principal balance of $236.4 million. These modifications were offered to borrowers who were experiencing temporary financial hardship—such as job loss, medical expenses, or other disruptive life events—and who requested relief from their contractual payment obligations. The modifications took one or more of the following forms:
(a) Payment-schedule extensions, which extended the maturity date of the loan by three to six months and adjusted the amortization schedule accordingly, reducing the borrower's required monthly payment during the hardship period;
(b) Temporary interest-rate reductions, which reduced the contractual interest rate on the loan for a period of three to twelve months, with the rate reverting to the original contractual rate at the conclusion of the reduction period; and
(c) Capitalization of accrued interest, which added past-due interest to the outstanding principal balance and re-amortized the loan over the remaining term, allowing the borrower to resume payments without being required to make a lump-sum payment to cure the arrearage.
These modifications were short-term in nature and were primarily designed to assist borrowers in resuming regular payments after a temporary disruption. They were not intended to address fundamental credit impairment or to restructure loans to avoid charge-off. The modification program was administered in accordance with BRB's consumer hardship assistance policies, which were reviewed and approved by the Bank's compliance and risk management teams.
BRB's Re-Aging Practice. Under BRB's methodology, loans that were modified and subsequently made at least three consecutive on-time payments (i.e., payments made within the applicable grace period for each of three successive monthly payment due dates) were reclassified from their delinquent status to "current" status, and their delinquency clock was reset. This practice is commonly referred to as "re-aging." Once re-aged, the loans were reassigned to a new-performance cohort for purposes of forward loss-rate projection within BRB's vintage-cohort model, rather than being retained in their original vintage delinquency bucket.
The re-aging practice was not a novel or ad hoc decision by BRB. It was adopted as a formal element of BRB's ACL methodology following analysis by the BRB credit risk analytics team, documented in an internal memorandum dated March 15, 2021, and approved by the ACL Governance Committee at its April 2021 meeting. The practice was applied consistently across all subsequent reporting periods, including FY 2022, Q1 2023, and Q2 2023.
Rationale for Re-Aging Treatment. The rationale for BRB's re-aging treatment rests on an empirical foundation. BRB's credit risk analytics team conducted a retrospective analysis of modified consumer loans in the Bank's traditional (non-BridgeFin) consumer lending portfolio, examining the performance of modified loans that had demonstrated three consecutive on-time payments following modification. The analysis covered a 10-year observation period (2011–2021) and encompassed approximately 28,500 modified consumer loans.
The key finding was that borrowers who successfully resumed payments post-modification—as evidenced by three consecutive on-time payments—exhibited a loss profile more closely aligned with performing (current) borrowers than with their original distressed cohort. Specifically:
• The 12-month re-default rate for modified loans with three consecutive on-time payments was approximately 22%, meaning that 78% of such loans continued to perform over the ensuing 12 months (the "78% re-performance rate").
• By contrast, the 12-month loss rate for unremediated delinquent loans (i.e., loans that remained in delinquent status without modification or that were modified but did not achieve three consecutive on-time payments) in the same vintage cohorts was approximately 47%.
The credit risk analytics team concluded, and the ACL Governance Committee agreed, that assigning re-performing modified loans to a performing cohort for forward loss-rate projection purposes more accurately reflects their expected credit behavior than retaining them in an elevated-loss vintage bucket. Retaining such loans in the original distressed cohort would, in effect, attribute to them a loss rate that is more than double their actual observed propensity to default, producing a systematically overstated ACL.
This re-aging practice was applied consistently to all modified BridgeFin loans that met the three-consecutive-payment threshold across all periods at issue.
C. Model Governance and Internal Controls
BRB's CECL model for the BridgeFin Portfolio was subject to a robust governance framework and internal control environment, including the following elements:
ACL Governance Committee. The ACL Governance Committee is a cross-functional management committee responsible for overseeing BRB's ACL estimation process for all loan portfolios, including BridgeFin. The committee is chaired by Raymond Hu, Chief Accounting Officer, and its members include Annette Kowalski, Chief Risk Officer, and Naveen Chakravarti, BridgeFin Division Head, along with representatives from the credit risk analytics team, the controllership group, and the internal audit function. The committee meets quarterly (and ad hoc as needed) to review model inputs, assumptions, qualitative overlays, and resulting ACL estimates. Meeting minutes documenting the committee's discussions and decisions are maintained and were produced in connection with this engagement.
CFO Approval. Patricia Navarrete, Chief Financial Officer, reviewed and approved the final ACL estimates for each reporting period prior to their inclusion in BRB's SEC filings. Ms. Navarrete's review encompassed the ACL Governance Committee's recommendations, the supporting workpapers, and the qualitative overlay calibration. Her approval was documented by sign-off on the final ACL summary memo for each period.
Independent Model Validation. BRB's CECL model was independently validated by the Bank's internal model risk management ("MRM") team on an annual basis, consistent with OCC Bulletin 2011-12 (Supervisory Guidance on Model Risk Management). The MRM team operates independently of the model owners (the credit risk analytics team) and reports to the Chief Risk Officer. The most recent full validation prior to the period at issue was completed in Q3 2022 and covered all aspects of the BridgeFin ACL model, including the vintage-cohort structure, macroeconomic scenario weighting, loss-rate projection methodology, and the re-aging treatment. The Q3 2022 validation concluded that the model was performing within acceptable tolerances and that the re-aging treatment was supported by the underlying data.
External Audit Review. Whitfield Keane LLP (PCAOB Registration No. 2847), BRB's independent external auditor, reviewed BRB's CECL methodology as part of its integrated audits for the fiscal years ended December 31, 2021, and December 31, 2022. The engagement partner was Lorraine Dupont, CPA. As part of its audit procedures, Whitfield Keane evaluated BRB's ACL estimation methodology, tested the underlying data and assumptions (including the re-aging treatment), assessed the reasonableness of qualitative overlays, and performed independent recalculations of selected model outputs. Whitfield Keane did not identify the re-aging treatment as a material misstatement, a significant deficiency, or a material weakness in internal controls over financial reporting. Whitfield Keane issued unqualified opinions on BRB's financial statements for both fiscal years.
Board Audit Committee Oversight. BRB's Board Audit Committee, chaired by Gwendolyn Prescott, received quarterly presentations on the ACL methodology and results from Mr. Hu and Ms. Navarrete. The Audit Committee reviewed the ACL as part of its oversight responsibilities for financial reporting and internal controls, and no member of the Audit Committee raised concerns regarding the BridgeFin re-aging treatment during the periods at issue.
D. Resulting ACL Figures
The application of BRB's vintage-cohort loss-rate methodology, including the re-aging treatment and qualitative overlays, produced the following ACL figures for the BridgeFin Portfolio:
These figures were reported in BRB's FY 2022 10-K and were the basis for the credit-quality disclosures in the MD&A and financial statement footnotes. The ACL figures were also carried forward as the opening balance for the Q1 2023 and Q2 2023 quarterly reporting periods, with appropriate adjustments for provision expense and charge-offs during those periods.
V. Analysis of SEC Staff's Preliminary View
A. Staff's Alternative Methodology
Based on my review of the Staff's preliminary analysis, as communicated during the Wells process, I understand that Staff Accountant Diane Milosevic and her colleagues contend that BRB's treatment of the 12,847 re-aged/modified loans ($236.4 million in aggregate outstanding principal balance) was improper and that those loans should have remained in their original vintage delinquency buckets for loss-rate estimation purposes.
Under the Staff's approach, modified loans that were re-aged and reclassified to performing status under BRB's methodology would instead be retained in their original vintage cohort and attributed the higher loss rates associated with that distressed vintage. Because the loss rates applicable to the original distressed vintage cohorts are substantially higher than the loss rates applicable to performing cohorts—reflecting the historical default experience of loans that became delinquent without subsequent remediation—the Staff's approach produces a materially higher aggregate loss-rate projection for the BridgeFin Portfolio.
The Staff's preliminary view results in the following alternative figures:
The $37.3 million difference between the Staff's ACL and BRB's reported ACL is attributable entirely to the different treatment of the re-aged loans. The Staff did not challenge BRB's underlying model structure (vintage-cohort loss rates), macroeconomic scenario weights, qualitative overlay calibration, or any other aspect of the ACL methodology. Nor did the Staff allege that BRB's loan-level data was inaccurate or incomplete. The sole point of disagreement is whether re-aged loans should be reassigned to performing cohorts (BRB's approach) or retained in their original distressed cohorts (the Staff's approach) for forward loss-rate projection purposes.
B. Materiality Considerations
The Staff's preliminary view that BRB's ACL was understated by $37.3 million raises the question of whether this difference, if it were determined to be a misstatement, would be material to BRB's financial statements.
Quantitative Analysis. BRB's internal materiality threshold, established in the Bank's accounting policies manual and applied by the external auditor, is set at 5% of pre-tax income. For fiscal year 2022, BRB reported pre-tax income of $612 million, yielding a quantitative materiality threshold of $30.6 million. The $37.3 million alleged understatement exceeds this threshold by $6.7 million (i.e., $37.3 million minus $30.6 million equals $6.7 million). On a purely quantitative basis, therefore, the alleged understatement is above BRB's internal materiality threshold.
However, materiality under the federal securities laws is a mixed question of law and fact that is not determined solely by application of a mechanical quantitative threshold. As the Supreme Court has stated, information is material if there is "a substantial likelihood that a reasonable shareholder would consider it important in making an investment decision." TSC Industries, Inc. v. Northway, Inc., 426 U.S. 438, 449 (1976). SEC Staff Accounting Bulletin No. 99 (Materiality) reinforces this principle, stating that "exclusive reliance on certain quantitative benchmarks to assess materiality in preparing financial statements ... is inappropriate" and that registrants and auditors must consider qualitative factors as well.
Qualitative Factors Weighing Against Materiality. When the relevant qualitative factors identified in SAB 99 are considered, the $37.3 million difference does not rise to the level of a material misstatement for the following reasons:
(a) The difference arises from a methodological judgment call, not from a data error, omission, or intentional misstatement. Both BRB's methodology and the Staff's alternative are permissible under ASC 326. The existence of an alternative estimate does not transform BRB's estimate into a misstatement. SAB 99 contemplates situations where a quantitative threshold may be exceeded but the underlying item does not have the hallmarks of a material misstatement—for example, where the difference "is the result of a reasonable estimate in the ordinary course of business." A methodological judgment exercised within the range of acceptable approaches under GAAP is precisely such a situation.
(b) BRB's methodology was consistently applied across all periods at issue. The re-aging treatment was adopted in Q1 2021, documented in internal memoranda, approved by the ACL Governance Committee, and applied uniformly through FY 2022, Q1 2023, and Q2 2023. There is no indication that BRB selectively changed its methodology from period to period to achieve a desired ACL result.
(c) The re-aging treatment was disclosed in BRB's MD&A. BRB's FY 2022 10-K included a discussion of the Bank's loan modification and re-aging practices in the MD&A section, enabling investors and analysts to evaluate the potential impact of this treatment on reported credit metrics. While the Staff may argue that additional disclosure granularity would have been desirable, the existence of the disclosure mitigates any claim that investors were misled.
(d) BRB's external auditor did not identify a misstatement. Whitfield Keane LLP, a PCAOB-registered audit firm, reviewed BRB's ACL methodology—including the re-aging treatment—as part of its FY 2021 and FY 2022 integrated audits and did not conclude that the methodology resulted in a material misstatement or that it represented a significant deficiency in internal controls. While the auditor's conclusion is not dispositive of the accounting treatment's propriety, it is a significant qualitative factor weighing against materiality.
(e) The difference did not mask a trend. SAB 99 identifies as a qualitative materiality factor whether a misstatement "masks a change in earnings or other trends." BRB's overall credit-quality narrative in its SEC filings was consistent with the disclosed metrics. The Bank disclosed increasing delinquency rates in certain BridgeFin vintages, discussed the macroeconomic environment's potential impact on consumer credit, and increased its provision for credit losses in the periods at issue. The re-aging treatment did not obscure a deteriorating credit trend; rather, BRB's disclosures provided a transparent picture of credit conditions in the BridgeFin Portfolio.
(f) The difference does not affect BRB's regulatory capital compliance. Even under the Staff's alternative ACL of $180.8 million, BRB's capital ratios would have remained well above all regulatory minimums and well-capitalized thresholds, and the additional provision expense ($37.3 million pre-tax) would not have triggered any regulatory action.
When these qualitative factors are considered in conjunction with the relatively modest quantum by which the $37.3 million difference exceeds the quantitative threshold (only $6.7 million, or approximately 1.1% of pre-tax income), I conclude that the alleged understatement does not constitute a material misstatement of BRB's financial statements.
C. Why BRB's Methodology Is Within the Range of Reasonable Estimates
1. ASC 326 Permits Multiple Approaches to Modified Loan Treatment
The CECL Standard does not mandate a specific treatment for modified or re-aged loans within loss-rate estimation models. The standard's principles-based framework requires entities to measure expected credit losses based on the best available information, using reasonable and supportable judgments, but it does not prescribe the mechanical steps by which loans are classified, segmented, or assigned to cohorts.
The FASB's Transition Resource Group for Credit Losses ("TRG") addressed questions related to troubled debt restructurings and loan modifications on several occasions during the implementation period. The TRG's discussions, as reflected in its meeting minutes and staff memoranda, consistently acknowledged that the treatment of modified loans in the CECL framework involves entity-specific judgment. The TRG did not adopt a rule requiring that modified loans be retained in their original vintage buckets, nor did it adopt a rule permitting their reclassification; instead, it recognized that both approaches—and hybrid approaches—could be reasonable depending on the entity's facts and circumstances.
Moreover, the FASB's issuance of ASU 2022-02, Financial Instruments — Credit Losses (Topic 326): Troubled Debt Restructurings and Vintage Disclosures, effective for BRB beginning January 1, 2023, eliminated the separate accounting and disclosure requirements for troubled debt restructurings ("TDRs"). The elimination of TDR recognition reflects the standard-setter's considered judgment that modification treatment should be based on the credit-risk characteristics of the modified loans—including post-modification performance—rather than on rigid classification rules that treat all modifications as inherently adverse credit events. While ASU 2022-02 was not yet effective for BRB's FY 2022 financial statements, the direction of the standard underscores the accounting profession's evolving understanding that modification does not permanently define a loan's risk profile.
In sum, there is no provision of ASC 326 that requires BRB to retain re-aged loans in their original vintage delinquency buckets. BRB's election to reclassify re-performing modified loans to a performing cohort is a permissible methodological choice under the standard.
2. BRB's Re-Aging Practice Was Data-Driven
BRB's decision to reclassify re-performing modified loans was not based on intuition, convenience, or a desire to minimize the ACL. It was grounded in rigorous empirical analysis of borrower behavior.
As discussed in Section IV.B above, BRB's internal analysis demonstrated that modified loans with three consecutive on-time payments had a 12-month re-default rate of approximately 22%. This means that 78% of such loans continued to perform over the ensuing 12 months, exhibiting payment behavior substantially similar to loans that had never been delinquent. By contrast, unremediated delinquent loans in the same vintage cohorts had a 12-month loss rate of approximately 47%—more than double the re-default rate of re-performing modified loans.
Assigning re-performing loans to a performing cohort for forward loss-rate projection purposes thus more accurately reflects their expected credit behavior than retaining them in an elevated-loss vintage bucket. The purpose of CECL is to estimate expected credit losses; an estimation methodology that systematically attributes to a class of loans a loss rate that is more than twice their actual observed propensity to default would overstate expected losses and produce a biased ACL.
Furthermore, BRB's 78% re-performance rate over 12 months is consistent with published industry studies on consumer loan modification outcomes. Academic research and regulatory data on consumer loan modifications have consistently shown that borrowers who demonstrate sustained payment behavior following modification—typically defined as three to six consecutive on-time payments—have substantially lower subsequent default rates than unremediated delinquent borrowers, though their default rates remain somewhat elevated relative to borrowers who were never delinquent.
3. OCC Supervisory Guidance Supports BRB's Approach
The OCC Comptroller's Handbook — Allowance for Credit Losses provides guidance on the estimation of credit losses for national banks and federal savings associations. While the Handbook does not have the force of an accounting standard, it reflects the OCC's supervisory expectations for how banks should implement CECL and is relevant to assessing the reasonableness of BRB's methodology.
The Handbook provides that banks should segment their loan portfolios based on "similar risk characteristics," and that the selection of appropriate risk segments is a matter of judgment that should reflect the bank's understanding of the credit-risk profile of its borrowers. Importantly, the Handbook recognizes that modifications affecting a borrower's risk profile may warrant reassignment to a different risk segment. A borrower who has been modified and has demonstrated sustained re-performance has, by definition, exhibited a change in risk characteristics relative to borrowers who remain delinquent. BRB's approach of reassigning such loans to a performing cohort is consistent with the OCC's risk-segmentation principle.
The OCC's concurrent supervisory examination of BRB, completed in September 2023 and resulting in a composite CAMELS rating of 2 (Satisfactory), did not identify a material deficiency in BRB's ACL methodology for the BridgeFin Portfolio. While an OCC examination is not an audit of financial statements and a satisfactory CAMELS rating does not constitute a validation of every aspect of a bank's accounting, the absence of an adverse supervisory finding regarding the BridgeFin ACL methodology is a significant indicator that the methodology was within the range of acceptable banking practice.
4. The Staff's Methodology Overstates Losses
Finally, I note that the Staff's alternative methodology—retaining re-aged loans in their original distressed vintage for loss-rate projection purposes—introduces a systematic upward bias into the ACL estimate.
Under the Staff's approach, the original vintage loss rate already reflects the proportion of loans in that vintage that became delinquent. By retaining re-aged loans in the original vintage and attributing to them the full distressed-vintage loss rate, the Staff's methodology effectively double-counts credit deterioration: once in the original vintage loss rate (which was calibrated based on the experience of loans that became delinquent, including those that were subsequently modified) and again by treating loans that have demonstrably resumed performance as if they retained the same probability of default as loans that remained delinquent.
This approach produces a systematically upward-biased loss estimate that does not reflect the current expected credit losses of the portfolio—which is the core objective of CECL. CECL requires entities to estimate expected credit losses based on "historical experience, current conditions, and reasonable and supportable forecasts" (ASC 326-20-30-1). A methodology that ignores the current condition of a class of loans (i.e., their demonstrated re-performance) and instead attributes to them a loss rate derived from borrowers who did not demonstrate re-performance does not faithfully represent "current conditions" as required by the standard.
The Staff's methodology may be defensible as one point in the range of reasonable estimates—and I have acknowledged that it is—but it is not the only correct approach, and it carries its own methodological limitations that the Staff's analysis does not appear to have addressed.
VI. Comparison of BRB's Net Charge-Off Rate: 2.14% vs. Staff's 3.05%
A. Derivation of BRB's 2.14% Net Charge-Off Rate
BRB calculated the net charge-off rate for the BridgeFin Portfolio for fiscal year 2022 by dividing total net charge-offs (gross charge-offs less recoveries) by the average outstanding loan balance for the year. This is the standard methodology for calculating net charge-off rates in the banking industry and is consistent with the definitions used in the FFIEC Uniform Bank Performance Report and BRB's SEC filings.
Under BRB's approach, loans that were modified and subsequently re-performing were not charged off. Consistent with GAAP and BRB's charge-off policy, only loans that progressed to confirmed default status (defined as 120 or more days delinquent) or that were otherwise deemed uncollectible based on individual assessment were charged off. Modified loans that resumed payment and were re-aged to current status were, by definition, not in default and were not charged off.
The resulting net charge-off rate of 2.14% reflects the actual economic losses experienced in the BridgeFin Portfolio during fiscal year 2022. It is a factual metric based on realized credit losses, not a projected or estimated figure. The 2.14% rate is verifiable against BRB's general ledger charge-off entries and recovery receipts, which were audited by Whitfield Keane LLP as part of its FY 2022 integrated audit.
B. Derivation of Staff's Approximately 3.05% Rate
The Staff's higher net charge-off figure of approximately 3.05% does not reflect additional actual defaults or charge-offs that occurred during fiscal year 2022. Rather, the Staff's figure incorporates an adjustment that treats certain re-aged loans as effectively defaulted—or as imminently defaulting—at the time of modification. By adding an incremental expected loss component to the charge-off calculation for these re-aged loans, the Staff arrives at a synthetic net charge-off rate that is approximately 91 basis points higher than BRB's reported figure.
This approach conflates two distinct accounting concepts. The ACL estimation exercise under CECL is a forward-looking, lifetime expected-loss calculation that requires the application of judgment and assumptions about future performance. The charge-off accounting, by contrast, is an actual-loss recognition exercise: under GAAP, a loan is charged off when it is deemed uncollectible, which is typically determined based on specific criteria (e.g., days past due exceeding a defined threshold, borrower bankruptcy, or other evidence of permanent impairment). Net charge-offs, once recognized, are factual—they reflect actual economic losses that have been incurred.
The Staff's approach of imputing additional charge-offs based on an alternative ACL methodology reverses this logical sequence. Rather than measuring actual losses and then projecting future losses (which is the purpose of the ACL), the Staff effectively projects future losses and then retroactively classifies them as current-period charge-offs. This produces a synthetic figure that does not correspond to any actual credit event and is not consistent with GAAP charge-off recognition principles.
Under GAAP, net charge-offs are recognized when a loan is deemed uncollectible; they are not restated retroactively based on alternative ACL assumptions. The Staff's recalculated charge-off rate, while useful as an analytical construct for understanding the impact of different ACL methodologies, should not be characterized as a "corrected" charge-off rate, because it does not reflect actual corrections to BRB's charge-off accounting.
C. Reconciliation of the Difference
The following table reconciles BRB's reported net charge-off rate with the Staff's alternative figure:
The approximately 0.91 percentage-point increment is attributable entirely to the Staff's different treatment of re-aged loans and does not reflect any additional actual defaults, write-downs, or charge-offs recognized during fiscal year 2022. BRB's reported net charge-off rate of 2.14% accurately reflects the realized credit losses in the BridgeFin Portfolio during the reporting period. The Staff's figure of approximately 3.05% is a synthetic construct that reflects the Staff's alternative forward-looking estimate, not a correction of an error in BRB's charge-off accounting.
VII. 90+ Day Delinquency Rate Analysis: 1.72% vs. 2.38%
A. BRB's Reported 1.72%
BRB reported a 90+ day delinquency rate for the BridgeFin Portfolio of 1.72% as of December 31, 2022. This rate was calculated by dividing the aggregate outstanding principal balance of BridgeFin loans that were 90 or more days past due as of the reporting date by the total outstanding principal balance of the BridgeFin Portfolio ($4.1 billion).
Under BRB's re-aging policy, loans that had been modified and had subsequently made three consecutive on-time payments were reclassified from delinquent to current status as of the date of the third on-time payment. Accordingly, re-aged loans that were current on their payments as of December 31, 2022—meaning that the borrower had made all required payments through the most recent due date—were classified as performing and were excluded from the 90+ day delinquency calculation.
This approach reflects the fundamental principle that delinquency status should be based on the borrower's current payment performance. A borrower who was once 90+ days delinquent but has since made three consecutive on-time payments and is current on all obligations as of the balance-sheet date is, as a factual matter, not 90+ days delinquent. Classifying such a borrower as delinquent would misrepresent the borrower's actual payment status and the portfolio's current credit condition.
B. Staff's 2.38% (Including Re-Aged Loans)
The Staff recalculated the 90+ day delinquency rate for the BridgeFin Portfolio by treating all 12,847 re-aged loans as if they remained in delinquent status at the reporting date, regardless of their current payment performance. This produced a 90+ day delinquency rate of 2.38%, which is 0.66 percentage points higher than BRB's reported 1.72%.
It is important to note that the Staff did not reclassify all 12,847 re-aged loans as 90+ day delinquent. The $236.4 million in re-aged loans, divided by $4.1 billion in outstanding BridgeFin loans, equals approximately 5.8% of the portfolio. However, not all re-aged loans were 90+ days delinquent prior to modification; some were 30–59 days or 60–89 days past due at the time of modification. The Staff's adjustment of 0.66 percentage points reflects only the subset of re-aged loans that had been 90 or more days delinquent prior to modification and were reclassified to current status through the re-aging process.