1045-00007 · Client 1045 · Witness Preparation

post-testimony-debrief-chakravarti.docx DOCUMENT Litigation

30 / 35 ← → ✕

ATTORNEY-CLIENT PRIVILEGED ATTORNEY WORK PRODUCT CONFIDENTIAL

CALDERWOOD & HARKNESS LLP INTERNAL MEMORANDUM

PURPOSE: This memorandum summarizes the investigative testimony of Naveen Chakravarti taken on January 30, 2024 at the SEC Division of Enforcement offices, 100 F Street NE, Washington, D.C. 20549, and provides an assessment of the Staff's areas of focus, Chakravarti's performance as a witness, and the implications of his testimony for the defense strategy going forward. This memorandum is intended solely for the internal use of the defense team and should be treated in accordance with the document-retention note set forth in Section VIII below.

I. SUMMARY OF PROCEEDINGS

A. Logistics. The investigative testimony of Naveen Chakravarti was taken on January 30, 2024 at the offices of the SEC Division of Enforcement, Home Office, 100 F Street NE, Washington, D.C. 20549. Testimony commenced at approximately 10:00 a.m. and concluded at approximately 4:45 p.m. Eastern Time. During the course of the day, a one-hour lunch break was taken (approximately 12:15 p.m. to 1:15 p.m.), along with two shorter breaks of approximately ten to fifteen minutes each — one in the mid-morning and one in the mid-afternoon. Total on-the-record time was approximately five hours and forty-five minutes.

A court reporter was present throughout the proceeding and administered the oath to the witness at the outset of the testimony. A certified transcript has been ordered through the SEC's designated reporting service. Based on the reporter's estimate, the transcript is expected to be available within ten to fourteen business days, which would place the expected delivery date on or about February 13–14, 2024. We have requested expedited delivery if available and will follow up as appropriate.

The testimony was taken pursuant to the testimony subpoena issued to Chakravarti on December 5, 2023, commanding his appearance for investigative testimony before the Staff of the Securities and Exchange Commission in the above-referenced matter. The subpoena was issued pursuant to the SEC's Formal Order of Investigation, No. EN-2023-0947, entered on or about July 12, 2023.

B. SEC Staff Present

The following members of the SEC Division of Enforcement Staff were present during the testimony:

• Andrea Felton, Senior Staff Attorney — Led the questioning during the first half of the testimony, from approximately 10:00 a.m. through the lunch break and into the early afternoon. Felton's questioning focused primarily on Chakravarti's background and role, the re-aging and loan modification practices, communications with senior management, and the disclosure review process.

• Christopher Yuen, Staff Attorney — Led the questioning during the second half of the testimony, from approximately 2:00 p.m. through the conclusion of the session. Yuen's questioning concentrated on credit quality metrics, charge-off rate calculations, and the relationship with the external auditor.

• Diane Milosevic, Staff Accountant — Present throughout the entire testimony session. Milosevic did not lead any extended line of questioning but interjected at multiple points to ask targeted technical questions regarding credit metrics, the re-aging methodology, modification-flag data, and re-default rates. Her questions were precise and reflected a detailed understanding of the underlying loan-level data produced in earlier document productions.

• Gregory Bonham, Assistant Director, Financial Fraud Unit — Observed the first two hours of testimony (approximately 10:00 a.m. to 12:00 p.m.). He departed the testimony room before the lunch break, returned briefly at the close of the session (approximately 4:30 p.m.), and did not ask any questions. His presence is noted for the reasons discussed in the demeanor and strategy assessment sections below.

C. Counsel Present

Counsel for Belmont Ridge Bank National Association ("BRB" or the "Bank") present during the testimony included:

• Victoria Chambers, Partner, Calderwood & Harkness LLP (Responsible Partner for the engagement). Ms. Chambers attended in her capacity as lead outside counsel for BRB and was present throughout the entire testimony session. She made objections as appropriate and conferred with the witness during breaks.

• Liam Donahue, Associate, Calderwood & Harkness LLP (Working Associate). I attended in my capacity as the primary working associate on the matter and was present throughout the entire session. I took contemporaneous notes, which form the basis for this memorandum.

Note: Jasmine Thornton, Lead Associate, remained at the Calderwood & Harkness office during the testimony to coordinate any real-time document requests from the Staff and to serve as a point of contact should additional materials from the production set be needed during the proceeding. No such requests were made by the Staff during the testimony.

No separate counsel appeared on behalf of Chakravarti individually.

D. Witness. The witness was Naveen Chakravarti, who holds the title of Senior Vice President and Division Head, BridgeFin Division, Belmont Ridge Bank National Association. Chakravarti appeared for testimony in his capacity as a current officer of BRB. He has been employed by BRB continuously since 2018, and has served as Division Head of BridgeFin since the division's launch in the third quarter of 2020.

At the outset of the testimony, and consistent with established practice, the Upjohn warning was reiterated to Chakravarti on the record. Specifically, Ms. Chambers confirmed that Calderwood & Harkness LLP represents BRB the entity, and not Chakravarti in his individual capacity. Chakravarti acknowledged on the record that he understood this representation, that the attorney-client privilege with respect to the representation belongs to BRB and not to him individually, and that BRB, as the holder of the privilege, could choose to waive it in the future. Chakravarti confirmed that he understood these terms and wished to proceed with the testimony under these conditions. This Upjohn colloquy was consistent with the written Upjohn acknowledgment previously executed by Chakravarti on November 15, 2023, a copy of which is maintained in the matter file.

II. KEY TOPICS COVERED DURING TESTIMONY

The testimony covered seven principal subject areas, which are discussed in detail below. The topics are presented in approximate chronological order as they arose during the testimony session, though the Staff occasionally revisited earlier subjects during the later portions of the examination.

A. Background and Role at BRB / BridgeFin

The testimony opened with foundational questions regarding Chakravarti's professional background, his role at Belmont Ridge Bank, and the structure and operations of the BridgeFin Division. This portion of the testimony was led by Felton and consumed approximately thirty to forty minutes.

Chakravarti described his role as Senior Vice President and Division Head of BridgeFin, a position he has held since the division's launch in the third quarter of 2020. He explained that BridgeFin is an internal division of Belmont Ridge Bank NA and is not a separately organized legal entity. BridgeFin operates under BRB's national bank charter and is subject to BRB's enterprise-wide risk management framework, compliance infrastructure, and governance structure. Chakravarti testified that BridgeFin was established as a strategic initiative to expand BRB's lending footprint in the consumer and small-business segments, with a particular focus on digitally originated lending products.

With respect to BridgeFin's product suite, Chakravarti described two principal product lines: (1) unsecured personal loans, with an average loan size of approximately $18,400, and (2) small-business microloans, with an average loan size of approximately $42,000. Both products are originated primarily through digital channels, utilizing BridgeFin's proprietary technology platform. He noted that the personal loan product constitutes the larger share of origination volume by both number of loans and dollar amount.

Chakravarti provided financial data regarding BridgeFin's scale. For fiscal year 2022, BridgeFin's total origination volume was approximately $2.87 billion, and the year-end outstanding loan balance was approximately $4.1 billion. He noted that this represented significant growth from BridgeFin's first full year of operations in fiscal year 2021 and reflected both organic demand and the expansion of BridgeFin's geographic footprint across additional states.

Regarding his reporting structure and interactions with other BRB executives, Chakravarti testified that he reports directly to Chief Risk Officer Annette Kowalski and ultimately to Chief Executive Officer Marcus R. Ellingham. He further described regular working relationships with Chief Financial Officer Patricia Navarrete and Chief Accounting Officer Raymond Hu, particularly with respect to the reporting of BridgeFin credit quality data that feeds into BRB's allowance for credit losses ("ACL") estimation process. He described these as collaborative but functionally independent relationships, with the business unit (BridgeFin) providing data and the accounting function (Hu's team) making the modeling and estimation judgments.

B. BridgeFin Underwriting Model and AI-Driven Origination

Felton questioned Chakravarti at length about BridgeFin's proprietary AI-driven underwriting model. This line of questioning occupied approximately forty-five minutes and was methodical in nature, proceeding from general to specific.

Chakravarti described the model as a machine-learning-based credit decisioning system that incorporates both traditional and non-traditional data inputs. Traditional inputs include FICO scores, debt-to-income ratios, and bureau-reported credit history. Alternative data inputs include bank transaction data (with borrower consent), employment verification through third-party data aggregators, and cash-flow analysis based on deposit account patterns. Chakravarti explained that the use of alternative data was a central design principle of BridgeFin's lending model, intended to extend credit to borrowers who might not qualify under traditional-only underwriting criteria while maintaining appropriate risk controls.

When asked about the model's governance and validation framework, Chakravarti described a multi-layered approach. The model undergoes a full independent validation on an annual basis, conducted by BRB's enterprise Model Risk Management ("MRM") group. This annual validation includes an assessment of model performance, stability, and discriminatory ability. In addition, the BridgeFin data science team performs quarterly back-testing, comparing the model's predicted default and loss rates against actual observed outcomes. Chakravarti stated that the results of these validations and back-tests are reported to the BRB Model Risk Committee, which he described as a sub-committee of the Executive Risk Committee.

The Staff appeared particularly interested in whether the AI model's risk segmentation adequately captured deteriorating credit quality in certain post-Q2 2021 vintage cohorts. Felton asked several questions about whether the model's performance metrics had shown any degradation in predictive accuracy for loans originated after the second quarter of 2021. Chakravarti acknowledged that the model's Kolmogorov-Smirnov ("KS") statistic had shown modest decline in certain vintages, which he attributed to the rapid macroeconomic shifts (rising interest rates, persistent inflation) that affected borrower repayment capacity beginning in late 2021 and accelerating through 2022. He emphasized, however, that the model remained within acceptable performance thresholds as defined by the MRM validation framework.

Chakravarti further testified that the model was recalibrated in the first quarter of 2023 to incorporate updated loss experience from the 2021 and 2022 vintage cohorts. He stated clearly and affirmatively that this recalibration was a scheduled event, consistent with the model's established governance cadence, and was not undertaken in response to the SEC's investigation or any inquiry from the Staff. He stated that the recalibration decision had been approved by the Model Risk Committee in December 2022, prior to BRB's receipt of the SEC's initial document request in this matter.

C. Re-Aging / Loan Modification Practices — Central Focus of Examination

The re-aging and loan modification practices employed by the BridgeFin Division constituted the most heavily examined topic of the entire testimony session. By our estimate, approximately two hours and thirty minutes of on-the-record time were dedicated to this subject, spanning portions of both Felton's and Yuen's examinations, with frequent interjections from Milosevic. This section of the memorandum warrants the most detailed treatment given the centrality of this topic to the Staff's investigation.

Scope of Modified Loan Population. The Staff focused on the population of 12,847 BridgeFin loans that had been re-aged or modified during the periods under examination, representing an aggregate outstanding balance of approximately $236.4 million. This figure had been identified by the Staff through their review of loan-level data produced in earlier document productions and was not disputed by Chakravarti.

Policy Framework. Felton presented Chakravarti with two internal BridgeFin policy documents that had been previously produced to the Staff. These were marked as Exhibit 47 and Exhibit 48, respectively, in the Staff's exhibit numbering scheme from the document production. (Bates numbers were not recorded during the testimony and will need to be confirmed against the transcript and exhibit list.) Exhibit 47 appeared to be BridgeFin's Loan Modification and Workout Policy, and Exhibit 48 appeared to be the related Standard Operating Procedure governing the re-aging process at the servicing-system level. Felton walked Chakravarti through several provisions of these documents, asking him to confirm their accuracy and currency.

Modification Program Structure. Chakravarti testified that BridgeFin's modification program was available to borrowers who had experienced a temporary hardship — such as job loss, medical emergency, or other documented financial disruption — and who demonstrated a reasonable capacity and willingness to resume payments. The typical modification involved one of two forms of relief: (1) a three-month payment deferral, during which the borrower's obligation to make monthly payments was suspended and the deferred amounts were added to the loan's remaining term, or (2) a temporary or permanent interest rate reduction designed to lower the borrower's monthly payment to an affordable level. In some cases, both forms of relief were combined.

The Re-Aging Treatment. Upon execution of a modification agreement, the loan's delinquency status in BRB's loan servicing system (ServicePoint, maintained by BRB's operations group) was reset to "current." This is the re-aging treatment that the SEC Staff has identified as the core issue in the investigation. The effect of the re-aging was that a loan that had been, for example, 90 days past due at the time of modification would thereafter be reported as current, with its delinquency clock effectively restarted. Chakravarti confirmed this treatment on the record and explained that it was the standard system treatment applied whenever a modification was executed, consistent with BRB's enterprise-wide loan modification policy applicable to all lending divisions.

Staff's Theory of Improper Purpose. Yuen pressed Chakravarti on whether the re-aging practice was adopted, in whole or in part, for the purpose of improving BRB's reported delinquency metrics. This was a critical line of questioning. Chakravarti firmly and unequivocally denied that the re-aging treatment was motivated by any desire to manage or manipulate reported metrics. He testified that the practice followed BRB's enterprise-wide loan modification policy, which had been in place since at least 2017, and that it was consistent with guidance published in the Office of the Comptroller of the Currency's Comptroller's Handbook regarding troubled debt restructurings and loan modifications. He characterized the re-aging as a standard servicing-system treatment that reflected the economic reality of the modification: once a borrower entered into a modification agreement and the new contractual terms took effect, the loan was current under those new terms.

Modification-Flag Data and CECL Model Inputs. This exchange represented one of the more significant moments of the testimony. Milosevic asked whether the re-aged loans were flagged or tagged in any manner in the data provided to the ACL modeling team within the CAO's function. Chakravarti stated that a "modification flag" did exist in the ServicePoint loan servicing system and that this flag was populated at the time of modification. However, when asked whether the CECL model team used that modification flag for the purpose of vintage-cohort bucketing or loss-rate estimation, Chakravarti acknowledged that he was not certain. He stated that the CECL model was developed and maintained by the CAO's team under Raymond Hu's direction and that, while BridgeFin provided the underlying loan data, the modeling decisions — including which data fields to incorporate as model inputs — were made by the accounting function. He deferred to Hu on this point, stating that Hu would be the appropriate witness to address the CECL model's architecture and input specifications.

Delinquency Rate Discrepancy. The Staff then turned to the quantitative impact of the re-aging treatment on reported delinquency rates. Yuen stated that BRB's publicly reported 90+ day delinquency rate for the BridgeFin portfolio was 1.72% as of the relevant measurement date (year-end 2022). He then stated that the Staff's preliminary analysis indicated that if re-aged loans were included in their original vintage delinquency buckets — i.e., if their delinquency status were calculated based on their pre-modification contractual status rather than their post-modification reset status — the 90+ day delinquency rate would be approximately 2.38%. This represents a difference of 66 basis points.

Chakravarti stated that he was not personally aware of the Staff's 2.38% figure prior to the testimony and that he had not independently calculated what the delinquency rate would be under the Staff's alternative methodology. He maintained that BRB's reported 1.72% rate was accurate based on the loans' current contractual status at the measurement date. He did not concede that the Staff's methodology was more appropriate, nor did he affirmatively challenge it — he simply stated that BRB's reported figure reflected its established policy for measuring delinquency.

Modification Volume Trends. The Staff exhibited a spreadsheet that had been marked as Exhibit 62. This exhibit appeared to be a re-aging trend analysis prepared by the Staff, which showed modification volumes on a quarterly basis. The analysis indicated that modification volumes increased materially in the third and fourth quarters of 2022, relative to prior quarters. Felton asked Chakravarti to comment on this trend.

Chakravarti confirmed that modification volumes did increase during Q3 and Q4 2022. He attributed this increase to macroeconomic headwinds, specifically the rapid rise in interest rates and persistent inflationary pressures, which had placed stress on certain segments of BridgeFin's borrower population. He described the increased modification activity as a normal and expected credit-cycle response — BridgeFin was providing relief to borrowers experiencing temporary hardship in a deteriorating macroeconomic environment, which is precisely the purpose for which the modification program had been designed. He drew a parallel to modification activity during the COVID-19 pandemic in 2020, when BRB's enterprise-wide modification volumes had similarly increased across all lending divisions.

D. Credit Quality Metrics and Charge-Off Rates

Yuen led the questioning on credit quality metrics and charge-off rates during the second half of the testimony session. This line of questioning consumed approximately forty-five minutes to one hour.

Net Charge-Off Rate. Yuen questioned Chakravarti about the reported net charge-off ("NCO") rate of 2.14% for the BridgeFin portfolio for fiscal year 2022. Yuen stated that the Staff's preliminary view was that the NCO rate should have been approximately 3.05% when accounting for the performance of re-aged loans — specifically, when modified loans that subsequently re-defaulted were attributed to their original vintage cohort rather than treated as new delinquencies measured from the post-modification date. This represents a difference of approximately 91 basis points between the reported and Staff-adjusted figures.

Chakravarti testified that the 2.14% rate was calculated in a manner consistent with BRB's enterprise-wide charge-off policy. Under that policy, loans are charged off at 120 days past due. For modified loans that subsequently re-default — i.e., that become delinquent again after the modification resets their status to current — the 120-day charge-off clock begins running from the date of the new delinquency, not from the original pre-modification delinquency date. Chakravarti stated that this treatment was consistent with the terms of the modification agreement, under which the borrower's obligations are contractually restructured and the prior delinquency is effectively cured.

Re-Default Rate. In what proved to be one of the most significant exchanges of the entire testimony, Milosevic asked whether BRB tracked re-default rates for modified loans separately from the general default rate for the BridgeFin portfolio. Chakravarti confirmed that BridgeFin did track this metric internally and stated that the re-default rate within twelve months of modification was approximately 28%. That is, approximately 28% of BridgeFin loans that were modified subsequently became delinquent again within one year of the modification date.

This figure appeared to be new information to the Staff, or at least new in the context of the testimony record. The 28% re-default rate had not previously been the subject of a specific Staff question in prior testimony sessions (including the Navarrete testimony on January 18, 2024). Chakravarti stated that the re-default metric appeared in internal BridgeFin management reports, which were within the document production set but which the Staff may not yet have identified or reviewed in detail given the volume of the production.

The Staff appeared to note this 28% re-default figure with significant interest. Milosevic wrote extensively during this portion of the testimony, and Yuen paused to confer with Felton before proceeding to the next line of questions. This data point warrants careful attention in our defense strategy, as discussed in Sections V and VI below.

E. Communications with Senior Management and Board

Felton led the questioning on Chakravarti's communications with other BRB senior executives regarding BridgeFin credit quality. This line of questioning was interspersed with the re-aging and credit quality topics discussed above and occupied approximately thirty to forty minutes of testimony time.

Executive Risk Committee Reports. Chakravarti described a regular cadence of credit-quality reporting. On a monthly basis, he or his team prepared a credit-quality report for the BridgeFin Division, which was presented to the BRB Executive Risk Committee. The Executive Risk Committee is composed of the CEO (Ellingham), CRO (Kowalski), CFO (Navarrete), CAO (Hu), and other senior executives. Chakravarti testified that the monthly reports included key credit metrics such as delinquency rates by vintage, modification volumes, and the re-default rate for modified loans.

Board Audit Committee Reports. On a quarterly basis, credit-quality information was also included in reports presented to the Board Audit Committee, which is chaired by independent director Gwendolyn Prescott. Felton asked whether these quarterly Audit Committee reports included the re-default rate for modified loans. Chakravarti stated that the monthly Executive Risk Committee reports included the re-default metric but that he was less certain about the specific format and content of the Audit Committee presentations, which were compiled and curated by the CRO's office and the CAO's office for presentation to the Board. He indicated he would need to review the actual presentations to confirm what specific metrics were included.

The October 2022 Email (Exhibit 71). Felton showed Chakravarti an email that had been marked as Exhibit 71. The email was sent from Chakravarti to CFO Navarrete, dated October 2022 (the specific date within October was not noted in my contemporaneous notes and will need to be confirmed in the transcript). In the email, Chakravarti flagged "elevated modification activity" within the BridgeFin portfolio and recommended a review of qualitative overlays in the CECL model to ensure that the increased modification volumes were being appropriately reflected in the ACL estimation.

Chakravarti confirmed the authenticity of the email and that he had authored it. He testified that Navarrete had acknowledged receipt and had forwarded the email to CAO Hu with a directive to evaluate whether qualitative overlay adjustments were warranted. Chakravarti stated that he understood Hu had subsequently reviewed the matter, although Chakravarti was not directly involved in the decisions made by the CECL model team regarding qualitative overlays.

Assessment of This Email. As discussed further in the defense strategy section below, this email is a double-edged evidentiary item. On one hand, it demonstrates awareness at a senior level within BRB that modification activity was increasing and that there might be implications for the ACL estimation. The Staff could attempt to use this as evidence that management was on notice of a potential understatement. On the other hand, the email demonstrates affirmative steps by Chakravarti to raise the issue through normal governance channels, including an explicit recommendation that the CECL model be evaluated — which is precisely the kind of responsible internal communication that undermines any theory of concealment or willful disregard. The email is consistent with a narrative of good-faith management judgment exercised through established processes.

F. Relationship with External Auditor

Yuen briefly questioned Chakravarti regarding his interactions with BRB's external auditor, Whitfield Keane LLP ("WK"), during the fiscal year 2022 audit cycle. This line of questioning was comparatively brief, occupying approximately fifteen to twenty minutes.

Chakravarti stated that he participated in two meetings with the WK engagement team during the FY 2022 audit cycle. The engagement partner was Lorraine Dupont. During these meetings, BridgeFin credit quality and modification practices were discussed. Chakravarti recalled that WK had asked questions about the modification program's structure and volume trends, and that WK had requested and received the modification-flag data from BridgeFin's operations team. He understood that this data was provided to WK in connection with their audit procedures over the ACL.

When asked whether WK had raised any concerns regarding the re-aging treatment or its effect on reported delinquency and charge-off metrics, Chakravarti stated that he was not aware of any such concerns. He noted, however, that the primary point of contact for WK's ACL-related audit work would have been the CAO's team, not BridgeFin directly, and that any audit findings or issues would more likely have been communicated through Hu's office.

G. Disclosure Review Process

The final substantive area of questioning addressed Chakravarti's role in the review and preparation of BRB's public filings with the SEC. Felton led this portion of the testimony, which occupied approximately twenty to twenty-five minutes.

Felton identified the three filings at issue in the investigation: (1) BRB's Annual Report on Form 10-K for fiscal year 2022, filed on February 28, 2023; (2) BRB's Quarterly Report on Form 10-Q for Q1 2023, filed on May 10, 2023; and (3) BRB's Quarterly Report on Form 10-Q for Q2 2023, filed on August 9, 2023. Felton asked Chakravarti to describe his role, if any, in the review of these filings.

Chakravarti testified that his involvement in the filing review process was limited to the BridgeFin-specific portions of the Management's Discussion and Analysis ("MD&A") section. He stated that he reviewed the draft MD&A language related to BridgeFin credit quality, origination volumes, and portfolio performance, and that he provided factual comments and corrections to ensure the accuracy of the business-unit-level information. He stated that he did not review the financial statements themselves, the notes to the financial statements, or the ACL-related disclosures in the notes, all of which fell within the responsibility of the CAO's team.

Chakravarti confirmed that BRB's regular outside securities counsel, Caldwell Pratt & Stein LLP (lead partner Martin Caldwell), coordinated the overall filing review process, including the distribution of drafts to the relevant executives for review and comment. He further confirmed that he provided a sign-off on a sub-certification as part of BRB's Sarbanes-Oxley certification framework. This sub-certification was a component of the disclosure controls and procedures framework and fed into the principal certifications executed by the CEO and CFO pursuant to Exchange Act Rules 13a-14(a) and 13a-14(b). Chakravarti described the sub-certification as a confirmation that the information within his area of responsibility had been accurately reported and that he was not aware of any material misstatements or omissions.

III. WITNESS DEMEANOR AND CREDIBILITY ASSESSMENT

Chakravarti presented as a knowledgeable, composed, and credible witness throughout the testimony. He demonstrated a thorough command of BridgeFin's operations, products, and credit-management practices, which was consistent with his position as the Division Head. His responses were well-organized and substantive, reflecting both his familiarity with the subject matter and the effectiveness of the pre-testimony preparation sessions.

He was appropriately careful not to speculate on accounting methodology questions that fell outside his direct area of responsibility. On several occasions, when asked questions relating to CECL model architecture, qualitative overlay decisions, or ACL estimation methodology, he consistently and naturally deferred to Raymond Hu as the responsible executive for those functions. Similarly, when asked about financial reporting judgments, he appropriately deferred to CFO Navarrete. This pattern of deference was credible and reflected genuine organizational boundaries within BRB rather than evasiveness or an effort to deflect responsibility.

Chakravarti did not volunteer information beyond what was asked in each question. He listened carefully to each question, paused before responding, and provided responsive and direct answers. There were no instances of apparent evasiveness, internal inconsistency, or hostility toward the Staff. He maintained a professional and cooperative tone throughout, which we believe left a favorable impression.

Area of Mild Concern — Modification Flag Uncertainty. One area that warrants monitoring is Chakravarti's uncertainty regarding whether the CECL model team used the modification flag for vintage-cohort bucketing purposes. While his deference to Hu on this point was entirely appropriate — the modeling decision was not his to make — the gap in his knowledge could be perceived by the Staff as evidence of insufficient coordination between the BridgeFin business unit (which generated the loan data) and the CAO's accounting function (which used that data for ACL estimation). The Staff may attempt to characterize this as a control deficiency — i.e., that the business unit was re-aging loans without ensuring that the downstream accounting treatment appropriately reflected the re-aged status. However, we believe this is a defensible organizational boundary: the separation of responsibilities between business-unit operations and accounting estimation is a feature of sound internal controls, not a deficiency.

The 28% Re-Default Rate Disclosure. Chakravarti's candid acknowledgment of the 28% re-default rate for modified loans was notable. He provided this figure without hesitation and in a straightforward manner, which was consistent with his overall approach of providing responsive, factual answers. While the substance of this figure is potentially adverse (as discussed in Sections V and VI), Chakravarti's manner of providing it reinforced his credibility as a witness who was not attempting to conceal unfavorable information. This is a significant data point that the Staff will likely pursue aggressively in Raymond Hu's upcoming testimony, which is currently scheduled for February 14, 2024.

Bonham's Presence. The presence of Gregory Bonham, Assistant Director of the Financial Fraud Unit, during the first two hours of testimony is noted. Bonham also attended a portion of CFO Navarrete's testimony on January 18, 2024, where he similarly observed the initial portion and departed before the session concluded. Bonham's pattern of attending portions of key witness testimony sessions suggests continued senior-level interest in the matter within the Division of Enforcement. However, his departure before lunch and only brief return at the close of the Chakravarti testimony does not indicate escalated concern or the kind of sustained senior-level engagement that would be expected if the matter were being treated as a high-priority fraud case. His presence is more consistent with routine supervisory oversight of a significant investigation rather than a signal of imminent enforcement action.

IV. COMPARISON WITH NAVARRETE TESTIMONY (JANUARY 18, 2024)

A critical objective of the post-testimony review is to assess the consistency between Chakravarti's testimony and the prior testimony of CFO Patricia Navarrete, which was taken on January 18, 2024. We have conducted an initial comparison based on our contemporaneous notes from both sessions. A more detailed comparison will be conducted once the official transcripts of both testimony sessions are available. Based on the information currently available, our preliminary assessment is as follows:

Consistency on Core Issues. Chakravarti's testimony was consistent with Navarrete's testimony on the key substantive points. Both witnesses described the re-aging practice as following BRB's existing enterprise-wide loan modification policy, which was established well before the BridgeFin Division was launched. Both characterized the loan modification program as a standard credit-management tool — a normal and expected response to borrower hardship — rather than a mechanism designed to manage reported delinquency metrics. Both witnesses confirmed that the ACL estimation was the responsibility of the CAO's team under Hu's direction, using the approved CECL methodology, and that the business units and the CFO's office relied on Hu's team for the technical accounting judgments embedded in the ACL.

The October 2022 Email (Exhibit 71). Chakravarti's testimony regarding the October 2022 email to Navarrete (Exhibit 71) was consistent with Navarrete's prior testimony. In her testimony on January 18, Navarrete confirmed that she had received the email from Chakravarti flagging elevated modification activity and that she had directed Hu to evaluate whether qualitative overlay adjustments were warranted. Chakravarti's account on January 30 corroborated this sequence of events. The two accounts are mutually reinforcing and establish a documented chain of communication through which the credit-quality concern was escalated to the appropriate decision-maker (Hu) through normal governance channels.

No Material Inconsistencies. We have not identified any material inconsistencies between the two witnesses' accounts on any substantive topic. The accounts are consistent with respect to organizational structure, roles and responsibilities, the modification program's design and purpose, the reporting cadence, and the division of responsibilities between the business unit and the accounting function. Minor differences in level of detail — for example, Chakravarti provided more granular information about the modification program's operational mechanics, while Navarrete provided more detail about the financial reporting process — are attributable to the witnesses' different roles and perspectives, not to any substantive disagreement.

This consistency is important from a defense perspective. Inconsistencies between witnesses on key factual points would provide the Staff with material for further inquiry and could undermine the credibility of BRB's account. The alignment between Chakravarti and Navarrete on the central narrative — that the re-aging practice was a policy-compliant credit-management tool and that management took appropriate steps to ensure the accounting function was aware of the increased modification activity — strengthens BRB's overall defensive posture.

V. STAFF'S APPARENT THEORY AND AREAS OF CONTINUED FOCUS

Based on the lines of questioning pursued during the Chakravarti testimony, as well as the prior Navarrete testimony, the Staff's apparent theory and areas of continued focus can be summarized as follows:

Core Theory — Material ACL Understatement. The Staff's theory remains centered on the contention that the re-aging of the 12,847 modified BridgeFin loans (aggregate outstanding balance of $236.4 million) caused BRB to materially understate its allowance for credit losses. Specifically, the Staff's preliminary view appears to be that the ACL was understated by approximately $37.3 million: BRB reported an ACL of $143.5 million as of the relevant measurement date, whereas the Staff's preliminary analysis suggests the ACL should have been approximately $180.8 million. The difference of $37.3 million exceeds BRB's internal materiality threshold of $30.6 million, which is calculated as 5% of BRB's pre-tax income of approximately $612 million, by a margin of approximately $6.7 million.

Delinquency Rate Suppression. The Staff appears focused on establishing that the re-aging treatment artificially suppressed BRB's reported 90+ day delinquency rate. As noted above, BRB reported a 90+ day delinquency rate of 1.72%, whereas the Staff's preliminary analysis — which places re-aged loans in their original vintage delinquency buckets rather than treating them as current — produces a rate of approximately 2.38%. The 66 basis point difference represents a 38% relative increase in the delinquency rate.

Net Charge-Off Rate Impact. Similarly, the Staff is focused on the contention that the re-aging treatment resulted in an understated net charge-off rate. BRB reported an NCO rate of 2.14% for fiscal year 2022, whereas the Staff's preliminary analysis — which attributes re-defaulting modified loans to their original vintage cohort — produces a rate of approximately 3.05%. The 91 basis point difference represents a 43% relative increase in the charge-off rate.

Milosevic's Quantitative Analysis. Milosevic's targeted technical questions throughout the testimony suggest that she is building a quantitative model to demonstrate that had re-aged loans been maintained in their original vintage buckets for CECL modeling purposes, the model would have produced a materially higher ACL. Her questions about the modification flag, the re-default rate, and the charge-off timeline for re-defaulting modified loans all appear designed to gather the factual inputs needed for this alternative calculation.

The 28% Re-Default Rate as a Focal Point. The 28% re-default rate disclosed by Chakravarti will almost certainly become a central element of the Staff's analysis going forward. The Staff may argue that a 28% re-default rate within twelve months of modification demonstrates that re-aged loans carried significantly higher credit risk than "current" loans and should not have been treated equivalently for loss-estimation purposes. In other words, the Staff's position will likely be that resetting these loans to "current" status in the CECL model, without applying a separate risk adjustment or qualitative overlay to account for the elevated re-default risk, resulted in a systematic underestimation of expected credit losses for the modified loan population.

Absence of Scienter-Based Questioning. Notably, and significantly, there was no line of questioning during the Chakravarti testimony — or, for that matter, during the Navarrete testimony — that suggested the Staff is pursuing a scienter-based theory. The Staff did not ask whether management intended to deceive investors, analysts, or regulators. The questions were framed around whether BRB's accounting methodology was reasonable and in compliance with GAAP, not around whether management knowingly or recklessly adopted an improper methodology for the purpose of misleading the market.

This framing is consistent with our assessment that the potential charges, if any, would be brought under Sections 13(a) and 13(b)(2)(A) of the Securities Exchange Act of 1934 and Rules 13a-1 and 12b-20 thereunder. These are non-scienter provisions governing periodic reporting obligations and books-and-records requirements. A violation under these provisions requires a showing that the issuer filed materially inaccurate reports or failed to maintain books and records that accurately reflected its transactions, but does not require proof that the issuer acted with an intent to defraud. This is a meaningfully less severe enforcement posture than a Section 10(b) / Rule 10b-5 fraud action, which would require proof of scienter.

VI. IMPLICATIONS FOR DEFENSE STRATEGY

A. Preparation for Raymond Hu Testimony (February 14, 2024)

The upcoming testimony of Chief Accounting Officer Raymond Hu, currently scheduled for February 14, 2024, will be the most critical remaining testimony event in the investigation. Hu is the witness best positioned to address the Staff's core questions regarding the CECL model's treatment of re-aged loans, which is the lynchpin of the Staff's theory. Both Chakravarti and Navarrete deferred to Hu on the technical accounting and modeling questions, and the Staff will expect Hu to provide definitive answers.

Hu should be thoroughly prepared to address the following topics:

(a) Modification Flag Availability. Whether the modification flag in the ServicePoint loan servicing system was available as an input to the CECL model, and if so, whether and how it was used in the model's vintage-cohort bucketing and loss-rate estimation processes.

(b) Treatment of Re-Aged Loans as Current. The rationale for treating re-aged loans as "current" for CECL modeling purposes, including whether this treatment was consistent with ASC 326 (Financial Instruments — Credit Losses), BRB's established CECL methodology, and applicable regulatory guidance.

(c) Qualitative Overlays. What qualitative overlays or management adjustments, if any, were applied to account for the re-default risk associated with modified loans. This is a critical question because even if the quantitative component of the CECL model treated re-aged loans as current, the application of an appropriate qualitative overlay could have mitigated or eliminated any resulting understatement.

(d) The Navarrete-to-Hu Referral (October 2022). Hu should be prepared to describe his response to the October 2022 referral from Navarrete — specifically, what analysis he conducted in response to Navarrete's directive to evaluate qualitative overlay adjustments, what conclusions he reached, and what changes (if any) were made to the CECL model or qualitative overlays as a result.

In addition, the 28% re-default rate should be addressed proactively in Hu's preparation. Hu should be prepared to explain how that rate was or was not factored into the ACL estimation process. If the re-default rate was considered by the CECL model team — for example, as a basis for a qualitative overlay or as a data point in the model's loss-rate calibration — Hu should be able to articulate how and to what extent. If it was not considered, Hu should be prepared to explain why, and to defend that decision as reasonable under the circumstances.

Jasmine Thornton is the lead associate responsible for Hu's testimony preparation. I recommend that a preparation working session be scheduled no later than February 7, 2024, to allow sufficient time for multiple preparation rounds before the testimony date.

B. Forensic Accounting Work — Brindoval Analytic Partners

Our forensic accounting consultants, Dr. Franklin Osei and the team at Brindoval Analytic Partners, should be briefed immediately on the 28% re-default figure disclosed during Chakravarti's testimony and its significance in the context of the Staff's emerging theory.

The Brindoval analysis, which is being prepared as attorney work product under the Kovel doctrine and is therefore protected by the work-product privilege, should be expanded to address whether incorporating the 28% re-default rate into a modified vintage-cohort model would produce an ACL within a reasonable range of BRB's reported figure of $143.5 million. Specifically, the analysis should model multiple scenarios:

(i) A scenario in which re-aged loans are maintained in their original vintage buckets but no qualitative overlay is applied (the Staff's apparent approach);

(ii) A scenario in which re-aged loans are treated as current but a qualitative overlay reflecting the 28% re-default rate is applied to the modified loan population;

(iii) A scenario in which re-aged loans are maintained in their original vintage buckets and a qualitative overlay is applied to account for the re-default rate; and

(iv) A scenario reflecting BRB's actual reported methodology and ACL, with supporting analysis of why that methodology falls within the range of reasonable estimates under ASC 326.

This analysis will be essential for any Wells Submission if the matter progresses to that stage. The analysis should also be prepared in a format that could support expert testimony if the matter were to proceed to an administrative proceeding or litigation, although we do not believe that stage is imminent.

C. Document Review Priorities

In light of the Chakravarti testimony, the following document review priorities should be addressed promptly by the review team:

(a) Re-Default Metric in Internal Reports. Prioritize the identification and review of all internal BridgeFin management reports containing the 28% re-default metric (or any re-default metric for modified loans). The review should determine which reports were shared with the Executive Risk Committee, which were shared with the Board Audit Committee, and whether there are any reports in which the re-default metric was discussed in the context of ACL estimation or qualitative overlays.

(b) October 2022 Email Chain. Confirm the chain of custody on the October 2022 Chakravarti-to-Navarrete email (Exhibit 71) and locate any responsive communications from Hu regarding qualitative overlay adjustments. If Hu responded by email to Navarrete's referral, that response will be a critical document in the Hu testimony preparation and in the overall defense narrative. If Hu's response was communicated orally or through other channels, that fact should be documented.

(c) Whitfield Keane Audit Workpapers. To the extent available without implicating bank-examination privilege issues, verify that WK's audit workpapers for the FY 2022 audit cycle reflect their review of the modification-flag data that Chakravarti testified was provided to the WK engagement team. The auditor's treatment of this data — whether WK independently assessed the re-aging treatment and concluded it was appropriate — could be a significant element of BRB's defense, as it would demonstrate that the external auditor reviewed the relevant information and did not object to BRB's methodology.

D. Overall Assessment

Chakravarti's testimony was, on balance, a net positive for BRB's defense positioning. Several factors support this assessment:

First, Chakravarti's testimony established that the re-aging practice was a standard credit-management tool employed pursuant to a longstanding enterprise-wide policy, not a scheme devised to manipulate reported delinquency metrics. The fact that the policy predates BridgeFin's existence by approximately three years undercuts any suggestion that it was adopted for an improper purpose.

Second, the October 2022 email demonstrates that Chakravarti took affirmative steps to flag credit-quality concerns through normal governance channels, including a specific recommendation that the CECL model's qualitative overlays be reviewed. This is inconsistent with a narrative in which management sought to conceal deteriorating credit quality or to suppress unfavorable metrics.

Third, Chakravarti's credibility as a witness was strong. His demeanor, responsiveness, and appropriate deference on questions outside his expertise enhanced the overall persuasiveness of his account.

The 28% re-default figure is a double-edged data point that requires careful handling. On the adverse side, the figure provides the Staff with a specific, quantifiable metric to argue that re-aged loans carried substantially higher risk than the "current" classification would suggest, and that BRB's CECL model should have accounted for this risk. On the favorable side, Chakravarti's candid disclosure of the figure — and his testimony that it was tracked in internal management reports and reported to the Executive Risk Committee — demonstrates transparency, not concealment. The key question going forward is whether BRB's CECL methodology adequately accounted for this re-default risk through qualitative overlays or other model adjustments. This is squarely within Raymond Hu's testimony territory and will be the central issue in his preparation.

We continue to assess that the matter is defensible. The absence of scienter-based questioning in both the Navarrete and Chakravarti testimony sessions reinforces our view that the Staff's most likely charge, if any, would be a non-fraud reporting violation under Sections 13(a) and 13(b)(2)(A) of the Exchange Act. While such a charge would still carry significant reputational and financial consequences for BRB, it represents a fundamentally different posture than a fraud action. Our defense strategy — centered on demonstrating that BRB's methodology was reasonable, consistently applied, and informed by appropriate governance processes — is well-suited to this legal framework.

VII. NEXT STEPS

The following action items should be completed in the near term:

1. Hu Testimony Preparation. Complete the preparation memorandum for the testimony of Raymond Hu, currently scheduled for February 14, 2024. Jasmine Thornton is the lead associate on Hu's preparation. An initial preparation session with Hu should be scheduled no later than February 7, 2024, with a follow-up session scheduled for February 12, 2024.

2. Document Productions. The fifth rolling production to the SEC Staff, consisting of 39,881 documents, is scheduled for February 15, 2024. The sixth and final rolling production, consisting of 38,912 documents, is scheduled for March 1, 2024. The review team should ensure that both productions are completed on schedule and that any documents responsive to the priority review topics identified in Section VI.C above are flagged for the defense team's attention prior to production.

3. Brindoval Analytic Partners Working Session. Schedule a working session with Dr. Franklin Osei and the Brindoval Analytic Partners team to integrate the insights from the Chakravarti testimony — particularly the 28% re-default rate — into the ongoing forensic accounting analysis. This session should occur no later than February 9, 2024.

4. Transcript Review. Request the official transcript of the Chakravarti testimony upon availability (estimated delivery on or about February 13–14, 2024). Upon receipt, conduct a detailed comparison against the Navarrete transcript to confirm consistency on all key factual points. Any discrepancies should be immediately flagged for partner review.

5. Audit Committee Status Memorandum. Prepare an interim status memorandum for the BRB Board Audit Committee. The target date for this memorandum is following the completion of the Hu testimony and the final document production, which would place the target in mid-to-late March 2024. The memorandum should summarize the status of the investigation, the key themes emerging from the testimony sessions, and the defense team's assessment of BRB's position.

VIII. DOCUMENT RETENTION NOTE

This memorandum constitutes attorney work product prepared in anticipation of litigation and is protected by the work-product doctrine. It reflects the mental impressions, conclusions, opinions, and legal theories of counsel. This memorandum also reflects attorney-client privileged communications between Calderwood & Harkness LLP and its client, Belmont Ridge Bank National Association. This memorandum should not be disclosed to any person outside of Calderwood & Harkness LLP without the express authorization of Victoria Chambers, Responsible Partner.

This memorandum is filed in the matter file for Engagement No. SC-MAT-2023-4172.

ATTORNEY-CLIENT PRIVILEGED / ATTORNEY WORK PRODUCT / CONFIDENTIAL

FROM:	Liam Donahue, Associate, Calderwood & Harkness LLP
TO:	Victoria Chambers, Partner; Howard Quinn, Partner
CC:	Jasmine Thornton, Lead Associate
DATE:	January 31, 2024
RE:	Post-Testimony Debrief — Investigative Testimony of Naveen Chakravarti, SVP & Division Head, BridgeFin Division, Belmont Ridge Bank National Association
MATTER:	SEC Investigation File No. HO-14287; SEC Formal Order No. EN-2023-0947
ENGAGEMENT NO.:	SC-MAT-2023-4172

ATTORNEY-CLIENT PRIVILEGED ATTORNEY WORK PRODUCT CONFIDENTIAL
CALDERWOOD & HARKNESS LLP INTERNAL MEMORANDUM
PURPOSE: This memorandum summarizes the investigative testimony of Naveen Chakravarti taken on January 30, 2024 at the SEC Division of Enforcement offices, 100 F Street NE, Washington, D.C. 20549, and provides an assessment of the Staff's areas of focus, Chakravarti's performance as a witness, and the implications of his testimony for the defense strategy going forward. This memorandum is intended solely for the internal use of the defense team and should be treated in accordance with the document-retention note set forth in Section VIII below.
I. SUMMARY OF PROCEEDINGS
A. Logistics. The investigative testimony of Naveen Chakravarti was taken on January 30, 2024 at the offices of the SEC Division of Enforcement, Home Office, 100 F Street NE, Washington, D.C. 20549. Testimony commenced at approximately 10:00 a.m. and concluded at approximately 4:45 p.m. Eastern Time. During the course of the day, a one-hour lunch break was taken (approximately 12:15 p.m. to 1:15 p.m.), along with two shorter breaks of approximately ten to fifteen minutes each — one in the mid-morning and one in the mid-afternoon. Total on-the-record time was approximately five hours and forty-five minutes.
A court reporter was present throughout the proceeding and administered the oath to the witness at the outset of the testimony. A certified transcript has been ordered through the SEC's designated reporting service. Based on the reporter's estimate, the transcript is expected to be available within ten to fourteen business days, which would place the expected delivery date on or about February 13–14, 2024. We have requested expedited delivery if available and will follow up as appropriate.
The testimony was taken pursuant to the testimony subpoena issued to Chakravarti on December 5, 2023, commanding his appearance for investigative testimony before the Staff of the Securities and Exchange Commission in the above-referenced matter. The subpoena was issued pursuant to the SEC's Formal Order of Investigation, No. EN-2023-0947, entered on or about July 12, 2023.
B. SEC Staff Present
The following members of the SEC Division of Enforcement Staff were present during the testimony:
• Andrea Felton, Senior Staff Attorney — Led the questioning during the first half of the testimony, from approximately 10:00 a.m. through the lunch break and into the early afternoon. Felton's questioning focused primarily on Chakravarti's background and role, the re-aging and loan modification practices, communications with senior management, and the disclosure review process.
• Christopher Yuen, Staff Attorney — Led the questioning during the second half of the testimony, from approximately 2:00 p.m. through the conclusion of the session. Yuen's questioning concentrated on credit quality metrics, charge-off rate calculations, and the relationship with the external auditor.
• Diane Milosevic, Staff Accountant — Present throughout the entire testimony session. Milosevic did not lead any extended line of questioning but interjected at multiple points to ask targeted technical questions regarding credit metrics, the re-aging methodology, modification-flag data, and re-default rates. Her questions were precise and reflected a detailed understanding of the underlying loan-level data produced in earlier document productions.
• Gregory Bonham, Assistant Director, Financial Fraud Unit — Observed the first two hours of testimony (approximately 10:00 a.m. to 12:00 p.m.). He departed the testimony room before the lunch break, returned briefly at the close of the session (approximately 4:30 p.m.), and did not ask any questions. His presence is noted for the reasons discussed in the demeanor and strategy assessment sections below.
C. Counsel Present
Counsel for Belmont Ridge Bank National Association ("BRB" or the "Bank") present during the testimony included:
• Victoria Chambers, Partner, Calderwood & Harkness LLP (Responsible Partner for the engagement). Ms. Chambers attended in her capacity as lead outside counsel for BRB and was present throughout the entire testimony session. She made objections as appropriate and conferred with the witness during breaks.
• Liam Donahue, Associate, Calderwood & Harkness LLP (Working Associate). I attended in my capacity as the primary working associate on the matter and was present throughout the entire session. I took contemporaneous notes, which form the basis for this memorandum.
Note: Jasmine Thornton, Lead Associate, remained at the Calderwood & Harkness office during the testimony to coordinate any real-time document requests from the Staff and to serve as a point of contact should additional materials from the production set be needed during the proceeding. No such requests were made by the Staff during the testimony.
No separate counsel appeared on behalf of Chakravarti individually.
D. Witness. The witness was Naveen Chakravarti, who holds the title of Senior Vice President and Division Head, BridgeFin Division, Belmont Ridge Bank National Association. Chakravarti appeared for testimony in his capacity as a current officer of BRB. He has been employed by BRB continuously since 2018, and has served as Division Head of BridgeFin since the division's launch in the third quarter of 2020.
At the outset of the testimony, and consistent with established practice, the Upjohn warning was reiterated to Chakravarti on the record. Specifically, Ms. Chambers confirmed that Calderwood & Harkness LLP represents BRB the entity, and not Chakravarti in his individual capacity. Chakravarti acknowledged on the record that he understood this representation, that the attorney-client privilege with respect to the representation belongs to BRB and not to him individually, and that BRB, as the holder of the privilege, could choose to waive it in the future. Chakravarti confirmed that he understood these terms and wished to proceed with the testimony under these conditions. This Upjohn colloquy was consistent with the written Upjohn acknowledgment previously executed by Chakravarti on November 15, 2023, a copy of which is maintained in the matter file.
II. KEY TOPICS COVERED DURING TESTIMONY
The testimony covered seven principal subject areas, which are discussed in detail below. The topics are presented in approximate chronological order as they arose during the testimony session, though the Staff occasionally revisited earlier subjects during the later portions of the examination.
A. Background and Role at BRB / BridgeFin
The testimony opened with foundational questions regarding Chakravarti's professional background, his role at Belmont Ridge Bank, and the structure and operations of the BridgeFin Division. This portion of the testimony was led by Felton and consumed approximately thirty to forty minutes.
Chakravarti described his role as Senior Vice President and Division Head of BridgeFin, a position he has held since the division's launch in the third quarter of 2020. He explained that BridgeFin is an internal division of Belmont Ridge Bank NA and is not a separately organized legal entity. BridgeFin operates under BRB's national bank charter and is subject to BRB's enterprise-wide risk management framework, compliance infrastructure, and governance structure. Chakravarti testified that BridgeFin was established as a strategic initiative to expand BRB's lending footprint in the consumer and small-business segments, with a particular focus on digitally originated lending products.
With respect to BridgeFin's product suite, Chakravarti described two principal product lines: (1) unsecured personal loans, with an average loan size of approximately $18,400, and (2) small-business microloans, with an average loan size of approximately $42,000. Both products are originated primarily through digital channels, utilizing BridgeFin's proprietary technology platform. He noted that the personal loan product constitutes the larger share of origination volume by both number of loans and dollar amount.
Chakravarti provided financial data regarding BridgeFin's scale. For fiscal year 2022, BridgeFin's total origination volume was approximately $2.87 billion, and the year-end outstanding loan balance was approximately $4.1 billion. He noted that this represented significant growth from BridgeFin's first full year of operations in fiscal year 2021 and reflected both organic demand and the expansion of BridgeFin's geographic footprint across additional states.
Regarding his reporting structure and interactions with other BRB executives, Chakravarti testified that he reports directly to Chief Risk Officer Annette Kowalski and ultimately to Chief Executive Officer Marcus R. Ellingham. He further described regular working relationships with Chief Financial Officer Patricia Navarrete and Chief Accounting Officer Raymond Hu, particularly with respect to the reporting of BridgeFin credit quality data that feeds into BRB's allowance for credit losses ("ACL") estimation process. He described these as collaborative but functionally independent relationships, with the business unit (BridgeFin) providing data and the accounting function (Hu's team) making the modeling and estimation judgments.
B. BridgeFin Underwriting Model and AI-Driven Origination
Felton questioned Chakravarti at length about BridgeFin's proprietary AI-driven underwriting model. This line of questioning occupied approximately forty-five minutes and was methodical in nature, proceeding from general to specific.
Chakravarti described the model as a machine-learning-based credit decisioning system that incorporates both traditional and non-traditional data inputs. Traditional inputs include FICO scores, debt-to-income ratios, and bureau-reported credit history. Alternative data inputs include bank transaction data (with borrower consent), employment verification through third-party data aggregators, and cash-flow analysis based on deposit account patterns. Chakravarti explained that the use of alternative data was a central design principle of BridgeFin's lending model, intended to extend credit to borrowers who might not qualify under traditional-only underwriting criteria while maintaining appropriate risk controls.
When asked about the model's governance and validation framework, Chakravarti described a multi-layered approach. The model undergoes a full independent validation on an annual basis, conducted by BRB's enterprise Model Risk Management ("MRM") group. This annual validation includes an assessment of model performance, stability, and discriminatory ability. In addition, the BridgeFin data science team performs quarterly back-testing, comparing the model's predicted default and loss rates against actual observed outcomes. Chakravarti stated that the results of these validations and back-tests are reported to the BRB Model Risk Committee, which he described as a sub-committee of the Executive Risk Committee.
The Staff appeared particularly interested in whether the AI model's risk segmentation adequately captured deteriorating credit quality in certain post-Q2 2021 vintage cohorts. Felton asked several questions about whether the model's performance metrics had shown any degradation in predictive accuracy for loans originated after the second quarter of 2021. Chakravarti acknowledged that the model's Kolmogorov-Smirnov ("KS") statistic had shown modest decline in certain vintages, which he attributed to the rapid macroeconomic shifts (rising interest rates, persistent inflation) that affected borrower repayment capacity beginning in late 2021 and accelerating through 2022. He emphasized, however, that the model remained within acceptable performance thresholds as defined by the MRM validation framework.
Chakravarti further testified that the model was recalibrated in the first quarter of 2023 to incorporate updated loss experience from the 2021 and 2022 vintage cohorts. He stated clearly and affirmatively that this recalibration was a scheduled event, consistent with the model's established governance cadence, and was not undertaken in response to the SEC's investigation or any inquiry from the Staff. He stated that the recalibration decision had been approved by the Model Risk Committee in December 2022, prior to BRB's receipt of the SEC's initial document request in this matter.
C. Re-Aging / Loan Modification Practices — Central Focus of Examination
The re-aging and loan modification practices employed by the BridgeFin Division constituted the most heavily examined topic of the entire testimony session. By our estimate, approximately two hours and thirty minutes of on-the-record time were dedicated to this subject, spanning portions of both Felton's and Yuen's examinations, with frequent interjections from Milosevic. This section of the memorandum warrants the most detailed treatment given the centrality of this topic to the Staff's investigation.
Scope of Modified Loan Population. The Staff focused on the population of 12,847 BridgeFin loans that had been re-aged or modified during the periods under examination, representing an aggregate outstanding balance of approximately $236.4 million. This figure had been identified by the Staff through their review of loan-level data produced in earlier document productions and was not disputed by Chakravarti.
Policy Framework. Felton presented Chakravarti with two internal BridgeFin policy documents that had been previously produced to the Staff. These were marked as Exhibit 47 and Exhibit 48, respectively, in the Staff's exhibit numbering scheme from the document production. (Bates numbers were not recorded during the testimony and will need to be confirmed against the transcript and exhibit list.) Exhibit 47 appeared to be BridgeFin's Loan Modification and Workout Policy, and Exhibit 48 appeared to be the related Standard Operating Procedure governing the re-aging process at the servicing-system level. Felton walked Chakravarti through several provisions of these documents, asking him to confirm their accuracy and currency.
Modification Program Structure. Chakravarti testified that BridgeFin's modification program was available to borrowers who had experienced a temporary hardship — such as job loss, medical emergency, or other documented financial disruption — and who demonstrated a reasonable capacity and willingness to resume payments. The typical modification involved one of two forms of relief: (1) a three-month payment deferral, during which the borrower's obligation to make monthly payments was suspended and the deferred amounts were added to the loan's remaining term, or (2) a temporary or permanent interest rate reduction designed to lower the borrower's monthly payment to an affordable level. In some cases, both forms of relief were combined.
The Re-Aging Treatment. Upon execution of a modification agreement, the loan's delinquency status in BRB's loan servicing system (ServicePoint, maintained by BRB's operations group) was reset to "current." This is the re-aging treatment that the SEC Staff has identified as the core issue in the investigation. The effect of the re-aging was that a loan that had been, for example, 90 days past due at the time of modification would thereafter be reported as current, with its delinquency clock effectively restarted. Chakravarti confirmed this treatment on the record and explained that it was the standard system treatment applied whenever a modification was executed, consistent with BRB's enterprise-wide loan modification policy applicable to all lending divisions.
Staff's Theory of Improper Purpose. Yuen pressed Chakravarti on whether the re-aging practice was adopted, in whole or in part, for the purpose of improving BRB's reported delinquency metrics. This was a critical line of questioning. Chakravarti firmly and unequivocally denied that the re-aging treatment was motivated by any desire to manage or manipulate reported metrics. He testified that the practice followed BRB's enterprise-wide loan modification policy, which had been in place since at least 2017, and that it was consistent with guidance published in the Office of the Comptroller of the Currency's Comptroller's Handbook regarding troubled debt restructurings and loan modifications. He characterized the re-aging as a standard servicing-system treatment that reflected the economic reality of the modification: once a borrower entered into a modification agreement and the new contractual terms took effect, the loan was current under those new terms.
Modification-Flag Data and CECL Model Inputs. This exchange represented one of the more significant moments of the testimony. Milosevic asked whether the re-aged loans were flagged or tagged in any manner in the data provided to the ACL modeling team within the CAO's function. Chakravarti stated that a "modification flag" did exist in the ServicePoint loan servicing system and that this flag was populated at the time of modification. However, when asked whether the CECL model team used that modification flag for the purpose of vintage-cohort bucketing or loss-rate estimation, Chakravarti acknowledged that he was not certain. He stated that the CECL model was developed and maintained by the CAO's team under Raymond Hu's direction and that, while BridgeFin provided the underlying loan data, the modeling decisions — including which data fields to incorporate as model inputs — were made by the accounting function. He deferred to Hu on this point, stating that Hu would be the appropriate witness to address the CECL model's architecture and input specifications.
Delinquency Rate Discrepancy. The Staff then turned to the quantitative impact of the re-aging treatment on reported delinquency rates. Yuen stated that BRB's publicly reported 90+ day delinquency rate for the BridgeFin portfolio was 1.72% as of the relevant measurement date (year-end 2022). He then stated that the Staff's preliminary analysis indicated that if re-aged loans were included in their original vintage delinquency buckets — i.e., if their delinquency status were calculated based on their pre-modification contractual status rather than their post-modification reset status — the 90+ day delinquency rate would be approximately 2.38%. This represents a difference of 66 basis points.
Chakravarti stated that he was not personally aware of the Staff's 2.38% figure prior to the testimony and that he had not independently calculated what the delinquency rate would be under the Staff's alternative methodology. He maintained that BRB's reported 1.72% rate was accurate based on the loans' current contractual status at the measurement date. He did not concede that the Staff's methodology was more appropriate, nor did he affirmatively challenge it — he simply stated that BRB's reported figure reflected its established policy for measuring delinquency.
Modification Volume Trends. The Staff exhibited a spreadsheet that had been marked as Exhibit 62. This exhibit appeared to be a re-aging trend analysis prepared by the Staff, which showed modification volumes on a quarterly basis. The analysis indicated that modification volumes increased materially in the third and fourth quarters of 2022, relative to prior quarters. Felton asked Chakravarti to comment on this trend.
Chakravarti confirmed that modification volumes did increase during Q3 and Q4 2022. He attributed this increase to macroeconomic headwinds, specifically the rapid rise in interest rates and persistent inflationary pressures, which had placed stress on certain segments of BridgeFin's borrower population. He described the increased modification activity as a normal and expected credit-cycle response — BridgeFin was providing relief to borrowers experiencing temporary hardship in a deteriorating macroeconomic environment, which is precisely the purpose for which the modification program had been designed. He drew a parallel to modification activity during the COVID-19 pandemic in 2020, when BRB's enterprise-wide modification volumes had similarly increased across all lending divisions.
D. Credit Quality Metrics and Charge-Off Rates
Yuen led the questioning on credit quality metrics and charge-off rates during the second half of the testimony session. This line of questioning consumed approximately forty-five minutes to one hour.
Net Charge-Off Rate. Yuen questioned Chakravarti about the reported net charge-off ("NCO") rate of 2.14% for the BridgeFin portfolio for fiscal year 2022. Yuen stated that the Staff's preliminary view was that the NCO rate should have been approximately 3.05% when accounting for the performance of re-aged loans — specifically, when modified loans that subsequently re-defaulted were attributed to their original vintage cohort rather than treated as new delinquencies measured from the post-modification date. This represents a difference of approximately 91 basis points between the reported and Staff-adjusted figures.
Chakravarti testified that the 2.14% rate was calculated in a manner consistent with BRB's enterprise-wide charge-off policy. Under that policy, loans are charged off at 120 days past due. For modified loans that subsequently re-default — i.e., that become delinquent again after the modification resets their status to current — the 120-day charge-off clock begins running from the date of the new delinquency, not from the original pre-modification delinquency date. Chakravarti stated that this treatment was consistent with the terms of the modification agreement, under which the borrower's obligations are contractually restructured and the prior delinquency is effectively cured.
Re-Default Rate. In what proved to be one of the most significant exchanges of the entire testimony, Milosevic asked whether BRB tracked re-default rates for modified loans separately from the general default rate for the BridgeFin portfolio. Chakravarti confirmed that BridgeFin did track this metric internally and stated that the re-default rate within twelve months of modification was approximately 28%. That is, approximately 28% of BridgeFin loans that were modified subsequently became delinquent again within one year of the modification date.
This figure appeared to be new information to the Staff, or at least new in the context of the testimony record. The 28% re-default rate had not previously been the subject of a specific Staff question in prior testimony sessions (including the Navarrete testimony on January 18, 2024). Chakravarti stated that the re-default metric appeared in internal BridgeFin management reports, which were within the document production set but which the Staff may not yet have identified or reviewed in detail given the volume of the production.
The Staff appeared to note this 28% re-default figure with significant interest. Milosevic wrote extensively during this portion of the testimony, and Yuen paused to confer with Felton before proceeding to the next line of questions. This data point warrants careful attention in our defense strategy, as discussed in Sections V and VI below.
E. Communications with Senior Management and Board
Felton led the questioning on Chakravarti's communications with other BRB senior executives regarding BridgeFin credit quality. This line of questioning was interspersed with the re-aging and credit quality topics discussed above and occupied approximately thirty to forty minutes of testimony time.
Executive Risk Committee Reports. Chakravarti described a regular cadence of credit-quality reporting. On a monthly basis, he or his team prepared a credit-quality report for the BridgeFin Division, which was presented to the BRB Executive Risk Committee. The Executive Risk Committee is composed of the CEO (Ellingham), CRO (Kowalski), CFO (Navarrete), CAO (Hu), and other senior executives. Chakravarti testified that the monthly reports included key credit metrics such as delinquency rates by vintage, modification volumes, and the re-default rate for modified loans.
Board Audit Committee Reports. On a quarterly basis, credit-quality information was also included in reports presented to the Board Audit Committee, which is chaired by independent director Gwendolyn Prescott. Felton asked whether these quarterly Audit Committee reports included the re-default rate for modified loans. Chakravarti stated that the monthly Executive Risk Committee reports included the re-default metric but that he was less certain about the specific format and content of the Audit Committee presentations, which were compiled and curated by the CRO's office and the CAO's office for presentation to the Board. He indicated he would need to review the actual presentations to confirm what specific metrics were included.
The October 2022 Email (Exhibit 71). Felton showed Chakravarti an email that had been marked as Exhibit 71. The email was sent from Chakravarti to CFO Navarrete, dated October 2022 (the specific date within October was not noted in my contemporaneous notes and will need to be confirmed in the transcript). In the email, Chakravarti flagged "elevated modification activity" within the BridgeFin portfolio and recommended a review of qualitative overlays in the CECL model to ensure that the increased modification volumes were being appropriately reflected in the ACL estimation.
Chakravarti confirmed the authenticity of the email and that he had authored it. He testified that Navarrete had acknowledged receipt and had forwarded the email to CAO Hu with a directive to evaluate whether qualitative overlay adjustments were warranted. Chakravarti stated that he understood Hu had subsequently reviewed the matter, although Chakravarti was not directly involved in the decisions made by the CECL model team regarding qualitative overlays.
Assessment of This Email. As discussed further in the defense strategy section below, this email is a double-edged evidentiary item. On one hand, it demonstrates awareness at a senior level within BRB that modification activity was increasing and that there might be implications for the ACL estimation. The Staff could attempt to use this as evidence that management was on notice of a potential understatement. On the other hand, the email demonstrates affirmative steps by Chakravarti to raise the issue through normal governance channels, including an explicit recommendation that the CECL model be evaluated — which is precisely the kind of responsible internal communication that undermines any theory of concealment or willful disregard. The email is consistent with a narrative of good-faith management judgment exercised through established processes.
F. Relationship with External Auditor
Yuen briefly questioned Chakravarti regarding his interactions with BRB's external auditor, Whitfield Keane LLP ("WK"), during the fiscal year 2022 audit cycle. This line of questioning was comparatively brief, occupying approximately fifteen to twenty minutes.
Chakravarti stated that he participated in two meetings with the WK engagement team during the FY 2022 audit cycle. The engagement partner was Lorraine Dupont. During these meetings, BridgeFin credit quality and modification practices were discussed. Chakravarti recalled that WK had asked questions about the modification program's structure and volume trends, and that WK had requested and received the modification-flag data from BridgeFin's operations team. He understood that this data was provided to WK in connection with their audit procedures over the ACL.
When asked whether WK had raised any concerns regarding the re-aging treatment or its effect on reported delinquency and charge-off metrics, Chakravarti stated that he was not aware of any such concerns. He noted, however, that the primary point of contact for WK's ACL-related audit work would have been the CAO's team, not BridgeFin directly, and that any audit findings or issues would more likely have been communicated through Hu's office.
G. Disclosure Review Process
The final substantive area of questioning addressed Chakravarti's role in the review and preparation of BRB's public filings with the SEC. Felton led this portion of the testimony, which occupied approximately twenty to twenty-five minutes.
Felton identified the three filings at issue in the investigation: (1) BRB's Annual Report on Form 10-K for fiscal year 2022, filed on February 28, 2023; (2) BRB's Quarterly Report on Form 10-Q for Q1 2023, filed on May 10, 2023; and (3) BRB's Quarterly Report on Form 10-Q for Q2 2023, filed on August 9, 2023. Felton asked Chakravarti to describe his role, if any, in the review of these filings.
Chakravarti testified that his involvement in the filing review process was limited to the BridgeFin-specific portions of the Management's Discussion and Analysis ("MD&A") section. He stated that he reviewed the draft MD&A language related to BridgeFin credit quality, origination volumes, and portfolio performance, and that he provided factual comments and corrections to ensure the accuracy of the business-unit-level information. He stated that he did not review the financial statements themselves, the notes to the financial statements, or the ACL-related disclosures in the notes, all of which fell within the responsibility of the CAO's team.
Chakravarti confirmed that BRB's regular outside securities counsel, Caldwell Pratt & Stein LLP (lead partner Martin Caldwell), coordinated the overall filing review process, including the distribution of drafts to the relevant executives for review and comment. He further confirmed that he provided a sign-off on a sub-certification as part of BRB's Sarbanes-Oxley certification framework. This sub-certification was a component of the disclosure controls and procedures framework and fed into the principal certifications executed by the CEO and CFO pursuant to Exchange Act Rules 13a-14(a) and 13a-14(b). Chakravarti described the sub-certification as a confirmation that the information within his area of responsibility had been accurately reported and that he was not aware of any material misstatements or omissions.
III. WITNESS DEMEANOR AND CREDIBILITY ASSESSMENT
Chakravarti presented as a knowledgeable, composed, and credible witness throughout the testimony. He demonstrated a thorough command of BridgeFin's operations, products, and credit-management practices, which was consistent with his position as the Division Head. His responses were well-organized and substantive, reflecting both his familiarity with the subject matter and the effectiveness of the pre-testimony preparation sessions.
He was appropriately careful not to speculate on accounting methodology questions that fell outside his direct area of responsibility. On several occasions, when asked questions relating to CECL model architecture, qualitative overlay decisions, or ACL estimation methodology, he consistently and naturally deferred to Raymond Hu as the responsible executive for those functions. Similarly, when asked about financial reporting judgments, he appropriately deferred to CFO Navarrete. This pattern of deference was credible and reflected genuine organizational boundaries within BRB rather than evasiveness or an effort to deflect responsibility.
Chakravarti did not volunteer information beyond what was asked in each question. He listened carefully to each question, paused before responding, and provided responsive and direct answers. There were no instances of apparent evasiveness, internal inconsistency, or hostility toward the Staff. He maintained a professional and cooperative tone throughout, which we believe left a favorable impression.
Area of Mild Concern — Modification Flag Uncertainty. One area that warrants monitoring is Chakravarti's uncertainty regarding whether the CECL model team used the modification flag for vintage-cohort bucketing purposes. While his deference to Hu on this point was entirely appropriate — the modeling decision was not his to make — the gap in his knowledge could be perceived by the Staff as evidence of insufficient coordination between the BridgeFin business unit (which generated the loan data) and the CAO's accounting function (which used that data for ACL estimation). The Staff may attempt to characterize this as a control deficiency — i.e., that the business unit was re-aging loans without ensuring that the downstream accounting treatment appropriately reflected the re-aged status. However, we believe this is a defensible organizational boundary: the separation of responsibilities between business-unit operations and accounting estimation is a feature of sound internal controls, not a deficiency.
The 28% Re-Default Rate Disclosure. Chakravarti's candid acknowledgment of the 28% re-default rate for modified loans was notable. He provided this figure without hesitation and in a straightforward manner, which was consistent with his overall approach of providing responsive, factual answers. While the substance of this figure is potentially adverse (as discussed in Sections V and VI), Chakravarti's manner of providing it reinforced his credibility as a witness who was not attempting to conceal unfavorable information. This is a significant data point that the Staff will likely pursue aggressively in Raymond Hu's upcoming testimony, which is currently scheduled for February 14, 2024.
Bonham's Presence. The presence of Gregory Bonham, Assistant Director of the Financial Fraud Unit, during the first two hours of testimony is noted. Bonham also attended a portion of CFO Navarrete's testimony on January 18, 2024, where he similarly observed the initial portion and departed before the session concluded. Bonham's pattern of attending portions of key witness testimony sessions suggests continued senior-level interest in the matter within the Division of Enforcement. However, his departure before lunch and only brief return at the close of the Chakravarti testimony does not indicate escalated concern or the kind of sustained senior-level engagement that would be expected if the matter were being treated as a high-priority fraud case. His presence is more consistent with routine supervisory oversight of a significant investigation rather than a signal of imminent enforcement action.
IV. COMPARISON WITH NAVARRETE TESTIMONY (JANUARY 18, 2024)
A critical objective of the post-testimony review is to assess the consistency between Chakravarti's testimony and the prior testimony of CFO Patricia Navarrete, which was taken on January 18, 2024. We have conducted an initial comparison based on our contemporaneous notes from both sessions. A more detailed comparison will be conducted once the official transcripts of both testimony sessions are available. Based on the information currently available, our preliminary assessment is as follows:
Consistency on Core Issues. Chakravarti's testimony was consistent with Navarrete's testimony on the key substantive points. Both witnesses described the re-aging practice as following BRB's existing enterprise-wide loan modification policy, which was established well before the BridgeFin Division was launched. Both characterized the loan modification program as a standard credit-management tool — a normal and expected response to borrower hardship — rather than a mechanism designed to manage reported delinquency metrics. Both witnesses confirmed that the ACL estimation was the responsibility of the CAO's team under Hu's direction, using the approved CECL methodology, and that the business units and the CFO's office relied on Hu's team for the technical accounting judgments embedded in the ACL.
The October 2022 Email (Exhibit 71). Chakravarti's testimony regarding the October 2022 email to Navarrete (Exhibit 71) was consistent with Navarrete's prior testimony. In her testimony on January 18, Navarrete confirmed that she had received the email from Chakravarti flagging elevated modification activity and that she had directed Hu to evaluate whether qualitative overlay adjustments were warranted. Chakravarti's account on January 30 corroborated this sequence of events. The two accounts are mutually reinforcing and establish a documented chain of communication through which the credit-quality concern was escalated to the appropriate decision-maker (Hu) through normal governance channels.
No Material Inconsistencies. We have not identified any material inconsistencies between the two witnesses' accounts on any substantive topic. The accounts are consistent with respect to organizational structure, roles and responsibilities, the modification program's design and purpose, the reporting cadence, and the division of responsibilities between the business unit and the accounting function. Minor differences in level of detail — for example, Chakravarti provided more granular information about the modification program's operational mechanics, while Navarrete provided more detail about the financial reporting process — are attributable to the witnesses' different roles and perspectives, not to any substantive disagreement.
This consistency is important from a defense perspective. Inconsistencies between witnesses on key factual points would provide the Staff with material for further inquiry and could undermine the credibility of BRB's account. The alignment between Chakravarti and Navarrete on the central narrative — that the re-aging practice was a policy-compliant credit-management tool and that management took appropriate steps to ensure the accounting function was aware of the increased modification activity — strengthens BRB's overall defensive posture.
V. STAFF'S APPARENT THEORY AND AREAS OF CONTINUED FOCUS
Based on the lines of questioning pursued during the Chakravarti testimony, as well as the prior Navarrete testimony, the Staff's apparent theory and areas of continued focus can be summarized as follows:
Core Theory — Material ACL Understatement. The Staff's theory remains centered on the contention that the re-aging of the 12,847 modified BridgeFin loans (aggregate outstanding balance of $236.4 million) caused BRB to materially understate its allowance for credit losses. Specifically, the Staff's preliminary view appears to be that the ACL was understated by approximately $37.3 million: BRB reported an ACL of $143.5 million as of the relevant measurement date, whereas the Staff's preliminary analysis suggests the ACL should have been approximately $180.8 million. The difference of $37.3 million exceeds BRB's internal materiality threshold of $30.6 million, which is calculated as 5% of BRB's pre-tax income of approximately $612 million, by a margin of approximately $6.7 million.
Delinquency Rate Suppression. The Staff appears focused on establishing that the re-aging treatment artificially suppressed BRB's reported 90+ day delinquency rate. As noted above, BRB reported a 90+ day delinquency rate of 1.72%, whereas the Staff's preliminary analysis — which places re-aged loans in their original vintage delinquency buckets rather than treating them as current — produces a rate of approximately 2.38%. The 66 basis point difference represents a 38% relative increase in the delinquency rate.
Net Charge-Off Rate Impact. Similarly, the Staff is focused on the contention that the re-aging treatment resulted in an understated net charge-off rate. BRB reported an NCO rate of 2.14% for fiscal year 2022, whereas the Staff's preliminary analysis — which attributes re-defaulting modified loans to their original vintage cohort — produces a rate of approximately 3.05%. The 91 basis point difference represents a 43% relative increase in the charge-off rate.
Milosevic's Quantitative Analysis. Milosevic's targeted technical questions throughout the testimony suggest that she is building a quantitative model to demonstrate that had re-aged loans been maintained in their original vintage buckets for CECL modeling purposes, the model would have produced a materially higher ACL. Her questions about the modification flag, the re-default rate, and the charge-off timeline for re-defaulting modified loans all appear designed to gather the factual inputs needed for this alternative calculation.
The 28% Re-Default Rate as a Focal Point. The 28% re-default rate disclosed by Chakravarti will almost certainly become a central element of the Staff's analysis going forward. The Staff may argue that a 28% re-default rate within twelve months of modification demonstrates that re-aged loans carried significantly higher credit risk than "current" loans and should not have been treated equivalently for loss-estimation purposes. In other words, the Staff's position will likely be that resetting these loans to "current" status in the CECL model, without applying a separate risk adjustment or qualitative overlay to account for the elevated re-default risk, resulted in a systematic underestimation of expected credit losses for the modified loan population.
Absence of Scienter-Based Questioning. Notably, and significantly, there was no line of questioning during the Chakravarti testimony — or, for that matter, during the Navarrete testimony — that suggested the Staff is pursuing a scienter-based theory. The Staff did not ask whether management intended to deceive investors, analysts, or regulators. The questions were framed around whether BRB's accounting methodology was reasonable and in compliance with GAAP, not around whether management knowingly or recklessly adopted an improper methodology for the purpose of misleading the market.
This framing is consistent with our assessment that the potential charges, if any, would be brought under Sections 13(a) and 13(b)(2)(A) of the Securities Exchange Act of 1934 and Rules 13a-1 and 12b-20 thereunder. These are non-scienter provisions governing periodic reporting obligations and books-and-records requirements. A violation under these provisions requires a showing that the issuer filed materially inaccurate reports or failed to maintain books and records that accurately reflected its transactions, but does not require proof that the issuer acted with an intent to defraud. This is a meaningfully less severe enforcement posture than a Section 10(b) / Rule 10b-5 fraud action, which would require proof of scienter.
VI. IMPLICATIONS FOR DEFENSE STRATEGY
A. Preparation for Raymond Hu Testimony (February 14, 2024)
The upcoming testimony of Chief Accounting Officer Raymond Hu, currently scheduled for February 14, 2024, will be the most critical remaining testimony event in the investigation. Hu is the witness best positioned to address the Staff's core questions regarding the CECL model's treatment of re-aged loans, which is the lynchpin of the Staff's theory. Both Chakravarti and Navarrete deferred to Hu on the technical accounting and modeling questions, and the Staff will expect Hu to provide definitive answers.
Hu should be thoroughly prepared to address the following topics:
(a) Modification Flag Availability. Whether the modification flag in the ServicePoint loan servicing system was available as an input to the CECL model, and if so, whether and how it was used in the model's vintage-cohort bucketing and loss-rate estimation processes.
(b) Treatment of Re-Aged Loans as Current. The rationale for treating re-aged loans as "current" for CECL modeling purposes, including whether this treatment was consistent with ASC 326 (Financial Instruments — Credit Losses), BRB's established CECL methodology, and applicable regulatory guidance.
(c) Qualitative Overlays. What qualitative overlays or management adjustments, if any, were applied to account for the re-default risk associated with modified loans. This is a critical question because even if the quantitative component of the CECL model treated re-aged loans as current, the application of an appropriate qualitative overlay could have mitigated or eliminated any resulting understatement.
(d) The Navarrete-to-Hu Referral (October 2022). Hu should be prepared to describe his response to the October 2022 referral from Navarrete — specifically, what analysis he conducted in response to Navarrete's directive to evaluate qualitative overlay adjustments, what conclusions he reached, and what changes (if any) were made to the CECL model or qualitative overlays as a result.
In addition, the 28% re-default rate should be addressed proactively in Hu's preparation. Hu should be prepared to explain how that rate was or was not factored into the ACL estimation process. If the re-default rate was considered by the CECL model team — for example, as a basis for a qualitative overlay or as a data point in the model's loss-rate calibration — Hu should be able to articulate how and to what extent. If it was not considered, Hu should be prepared to explain why, and to defend that decision as reasonable under the circumstances.
Jasmine Thornton is the lead associate responsible for Hu's testimony preparation. I recommend that a preparation working session be scheduled no later than February 7, 2024, to allow sufficient time for multiple preparation rounds before the testimony date.
B. Forensic Accounting Work — Brindoval Analytic Partners
Our forensic accounting consultants, Dr. Franklin Osei and the team at Brindoval Analytic Partners, should be briefed immediately on the 28% re-default figure disclosed during Chakravarti's testimony and its significance in the context of the Staff's emerging theory.
The Brindoval analysis, which is being prepared as attorney work product under the Kovel doctrine and is therefore protected by the work-product privilege, should be expanded to address whether incorporating the 28% re-default rate into a modified vintage-cohort model would produce an ACL within a reasonable range of BRB's reported figure of $143.5 million. Specifically, the analysis should model multiple scenarios:
(i) A scenario in which re-aged loans are maintained in their original vintage buckets but no qualitative overlay is applied (the Staff's apparent approach);
(ii) A scenario in which re-aged loans are treated as current but a qualitative overlay reflecting the 28% re-default rate is applied to the modified loan population;
(iii) A scenario in which re-aged loans are maintained in their original vintage buckets and a qualitative overlay is applied to account for the re-default rate; and
(iv) A scenario reflecting BRB's actual reported methodology and ACL, with supporting analysis of why that methodology falls within the range of reasonable estimates under ASC 326.
This analysis will be essential for any Wells Submission if the matter progresses to that stage. The analysis should also be prepared in a format that could support expert testimony if the matter were to proceed to an administrative proceeding or litigation, although we do not believe that stage is imminent.
C. Document Review Priorities
In light of the Chakravarti testimony, the following document review priorities should be addressed promptly by the review team:
(a) Re-Default Metric in Internal Reports. Prioritize the identification and review of all internal BridgeFin management reports containing the 28% re-default metric (or any re-default metric for modified loans). The review should determine which reports were shared with the Executive Risk Committee, which were shared with the Board Audit Committee, and whether there are any reports in which the re-default metric was discussed in the context of ACL estimation or qualitative overlays.
(b) October 2022 Email Chain. Confirm the chain of custody on the October 2022 Chakravarti-to-Navarrete email (Exhibit 71) and locate any responsive communications from Hu regarding qualitative overlay adjustments. If Hu responded by email to Navarrete's referral, that response will be a critical document in the Hu testimony preparation and in the overall defense narrative. If Hu's response was communicated orally or through other channels, that fact should be documented.
(c) Whitfield Keane Audit Workpapers. To the extent available without implicating bank-examination privilege issues, verify that WK's audit workpapers for the FY 2022 audit cycle reflect their review of the modification-flag data that Chakravarti testified was provided to the WK engagement team. The auditor's treatment of this data — whether WK independently assessed the re-aging treatment and concluded it was appropriate — could be a significant element of BRB's defense, as it would demonstrate that the external auditor reviewed the relevant information and did not object to BRB's methodology.
D. Overall Assessment
Chakravarti's testimony was, on balance, a net positive for BRB's defense positioning. Several factors support this assessment:
First, Chakravarti's testimony established that the re-aging practice was a standard credit-management tool employed pursuant to a longstanding enterprise-wide policy, not a scheme devised to manipulate reported delinquency metrics. The fact that the policy predates BridgeFin's existence by approximately three years undercuts any suggestion that it was adopted for an improper purpose.
Second, the October 2022 email demonstrates that Chakravarti took affirmative steps to flag credit-quality concerns through normal governance channels, including a specific recommendation that the CECL model's qualitative overlays be reviewed. This is inconsistent with a narrative in which management sought to conceal deteriorating credit quality or to suppress unfavorable metrics.
Third, Chakravarti's credibility as a witness was strong. His demeanor, responsiveness, and appropriate deference on questions outside his expertise enhanced the overall persuasiveness of his account.
The 28% re-default figure is a double-edged data point that requires careful handling. On the adverse side, the figure provides the Staff with a specific, quantifiable metric to argue that re-aged loans carried substantially higher risk than the "current" classification would suggest, and that BRB's CECL model should have accounted for this risk. On the favorable side, Chakravarti's candid disclosure of the figure — and his testimony that it was tracked in internal management reports and reported to the Executive Risk Committee — demonstrates transparency, not concealment. The key question going forward is whether BRB's CECL methodology adequately accounted for this re-default risk through qualitative overlays or other model adjustments. This is squarely within Raymond Hu's testimony territory and will be the central issue in his preparation.
We continue to assess that the matter is defensible. The absence of scienter-based questioning in both the Navarrete and Chakravarti testimony sessions reinforces our view that the Staff's most likely charge, if any, would be a non-fraud reporting violation under Sections 13(a) and 13(b)(2)(A) of the Exchange Act. While such a charge would still carry significant reputational and financial consequences for BRB, it represents a fundamentally different posture than a fraud action. Our defense strategy — centered on demonstrating that BRB's methodology was reasonable, consistently applied, and informed by appropriate governance processes — is well-suited to this legal framework.
VII. NEXT STEPS
The following action items should be completed in the near term:
1. Hu Testimony Preparation. Complete the preparation memorandum for the testimony of Raymond Hu, currently scheduled for February 14, 2024. Jasmine Thornton is the lead associate on Hu's preparation. An initial preparation session with Hu should be scheduled no later than February 7, 2024, with a follow-up session scheduled for February 12, 2024.
2. Document Productions. The fifth rolling production to the SEC Staff, consisting of 39,881 documents, is scheduled for February 15, 2024. The sixth and final rolling production, consisting of 38,912 documents, is scheduled for March 1, 2024. The review team should ensure that both productions are completed on schedule and that any documents responsive to the priority review topics identified in Section VI.C above are flagged for the defense team's attention prior to production.
3. Brindoval Analytic Partners Working Session. Schedule a working session with Dr. Franklin Osei and the Brindoval Analytic Partners team to integrate the insights from the Chakravarti testimony — particularly the 28% re-default rate — into the ongoing forensic accounting analysis. This session should occur no later than February 9, 2024.
4. Transcript Review. Request the official transcript of the Chakravarti testimony upon availability (estimated delivery on or about February 13–14, 2024). Upon receipt, conduct a detailed comparison against the Navarrete transcript to confirm consistency on all key factual points. Any discrepancies should be immediately flagged for partner review.
5. Audit Committee Status Memorandum. Prepare an interim status memorandum for the BRB Board Audit Committee. The target date for this memorandum is following the completion of the Hu testimony and the final document production, which would place the target in mid-to-late March 2024. The memorandum should summarize the status of the investigation, the key themes emerging from the testimony sessions, and the defense team's assessment of BRB's position.
VIII. DOCUMENT RETENTION NOTE
This memorandum constitutes attorney work product prepared in anticipation of litigation and is protected by the work-product doctrine. It reflects the mental impressions, conclusions, opinions, and legal theories of counsel. This memorandum also reflects attorney-client privileged communications between Calderwood & Harkness LLP and its client, Belmont Ridge Bank National Association. This memorandum should not be disclosed to any person outside of Calderwood & Harkness LLP without the express authorization of Victoria Chambers, Responsible Partner.
This memorandum is filed in the matter file for Engagement No. SC-MAT-2023-4172.
ATTORNEY-CLIENT PRIVILEGED / ATTORNEY WORK PRODUCT / CONFIDENTIAL
FROM: | Liam Donahue, Associate, Calderwood & Harkness LLP
TO: | Victoria Chambers, Partner; Howard Quinn, Partner
CC: | Jasmine Thornton, Lead Associate
DATE: | January 31, 2024
RE: | Post-Testimony Debrief — Investigative Testimony of Naveen Chakravarti, SVP & Division Head, BridgeFin Division, Belmont Ridge Bank National Association
MATTER: | SEC Investigation File No. HO-14287; SEC Formal Order No. EN-2023-0947
ENGAGEMENT NO.: | SC-MAT-2023-4172