Evaluating the Credit Loss Estimates in M&A / Capital Raise Scenarios. David H. Ruffin Director DHG Credit Risk Management

Transcription

1 Evaluating the Credit Loss Estimates in M&A / Capital Raise Scenarios David H. Ruffin Director DHG Credit Risk Management

2 Agenda (September 14, 2017) A decade of credit mark / portfolio loss estimates What do we mean by credit mark / loss estimate, etc.? The community bank conundrum: problem loans Quantitative & qualitative compares in M&A Questions

3 A decade of credit mark / portfolio loss estimates

4 Loan Portfolio Due Diligence (Loss Estimates) ~260 Due Diligence Transactions since '08 + Capital Raise 43% Data current as of Q M&A 47% Loss Share 10% Accurate and back-tested results Expertise in both sides: Credit Mark & Loan Review Efficient process 2-Phases to support Go / No-Go Decision Time Frame Range of Credit Marks* % % % Recent for CRMa / % DHG-CRM *Ranges only general estimates from CRM work and not necessarily representative of a national averages

5 Key Data Needed for Credible Loss Estimates Flat File (Interagency or Loan Data file) Impairment Calculations Past Due List flat file date Watch List/Special Assets Update Reports (detailed workout plans, etc.) Relationship Groupings TDR & Non-Accrual List Letters of Credit List Policy Manual with RG Definitions Look Back Study Loan data (i.e. core system download, snapshot ) at each month s end over a TBD month look back period ending on the Estimate Date ( Look Back Period ). Individual charge-offs and recoveries recorded, by note number, over the Look Back Period along with the date of the respective charge-off or recovery.* Impairments recorded by the Bank on loans evaluated under ASC 310, Receivables (formerly SFAS 114), by individual note number. *Individual charge-offs and recoveries should be produced as a list of "Note Number / Charge- Off Amount / Charge-Off Date" (with multiple partial charge downs being listed multiple times) or a list of "Note Number / Total Charge-Off Amount / First Charge-Off Date" (with multiple partial charge downs being aggregated).

6 What do we mean by credit mark / loss estimate, etc.?

7 What Do We Mean by the Following? Credit mark Portfolio loss estimate Yield mark Day one (valuation of loans at purchase) Day two CECL

8 The community bank conundrum: problem loans

9 NPA s Structurally Larger at Smaller Banks NPA S+90 PD / Tang Equity+ALLL (%) NCO s / Average Loans (%) Small Banks Big Banks Big Banks Small Banks Smaller Banks couldn t flush and made bet to keep more capital and more NPA s

10 Loan Portfolio Composition by Asset Size March 31, 2016 Small Institutions Larger Institutions 37% 63% Loans Tied To Real Estate Source FDIC

11 Smaller the Bank: Greater the Loss Given Default From Blind Pool Data

12 Possible Disproportionate CECL Impact? Community Banks have higher CECLrisk than industry as a whole Source FDIC

13 Quantitative & qualitative compares in M&A

14 M&A: Common Fears / Risks of Credit DD s Explaining easily to staff unlike capital raises Satisfying multiple suitors, wanting to do their own work Under-(or significantly over-) shooting the credit mark Confusing the credit and the accounting Navigating between the wink & nods and burn-downs Drilling down below call report data How bifurcated is risk / production? Using regulatory examinations as proxies for credit quality Short-cutting (or just estimating) the credit mark Syncing lending philosophies, delivery modes, and talent levels? Addressing the social issues conundrum objectively? Planning credible integration strategies

15 What to Look for on a Bank s Balance Sheet Quality of Risk Management Transparency (not just relying on regulators / other audit functions) Bifurcation between production and risk Credit skills Board (involvement in the lending process & borrowing) % Speculative RE Unmonitored C&I History with stress testing Risk grade migrations Growth rates in loan types ALLL / new guidance (CECL) Losses imbedded in OLEM / Substandard (and Weak Pass) Traditional ratios (TDR s / NPL s / TDR s / Texas Ratio) Miss-read of Weighted Average Risk Grades OREO (low-to-medium loss rates w/ large pool remaining)

16 Portfolio: Granular Segmentation Granular model segmentation for both PD and LGD seeks to establish both risk homogeneity and meaningfulness for the bank. Segmentation typically along product / call report lines. Further sub-product segmentation if mix has varied over time. Migration performed by risk grade or delinquency cycle. A vintage-based segmentation may provide further differentiation across borrowers.

17 Another Use of Macro Data: Creating the Targeted Credit Mark or Loss Estimate A comprehensive loan review, combined with: Risk grade migrations Probabilities of default (PD s) Loss given default (LGD s) PD s X LGD s = EL (Estimated Losses) Probabilistic modeling Comparative analysis

18 The Sampling Process: Modern Loan Review Line of business / new loans / vintage / collateral type / perceived rogue lender or market Large exposures Risk-based Random

19 Risk Grade Migration Analysis

20 Loss Distribution Estimation and Stress Testing

21 M&A? Comparatives: PD s Commercial (2 Year)

22 LGD Parameter Quantification

23 M&A? Comparatives: LGD s

24 Comparatives: Discounted Potential Credit Loss ( dpcl ) Estimates

25 Risk Grade and Impairment Distribution

26 Policy Observations

27 Concentration Comparison

28 Prospective Credit Culture Issues Impacting Merger Deliberations Create process for inventorying (horse-trading / best practices) each bank s: Credit policies Credit talent / redundancies Underwriting / delivery (hunter-skinner dynamic) Credit approval process Quality / format of loan review Pro-forma combined concentration (loan product / geographic) tolerances Pro-forma light DFAST impacts on Capital / IS / BS Hold on to classically trained/deep-experienced credit talent: it s a declining resource.

29 Conventional Due Diligence Process

30 Another Approach: Two Phase Due Diligence Phase I Remote Portfolio Assessment and Risk Identification Loss estimate modeling on bank(s) loan portfolios to develop PD s and LGD s and potential credit loss estimates based on the bank(s) own experience Credit Policy Observations Concentrations Portfolio Compositions Distribution of Risk Grades Phase II On-site Credit File Review with targeted sampling informed by Phase I Can be applied consistently to both banks in a potential MOE, or to one bank as part of a due diligence or reverse due diligence.

31 Time Line for Two Phase Due Diligence

32 Questions?

33 Contact Information