What are CECL gaps in the current ALLL process?

Transcription

1 What are CECL gaps in the current ALLL process? Considerations for implementing the forthcoming Accounting for Financial Instruments: Credit Losses standard Zions Bancorporation Alexander Hume Controller Brooks Brady - SVP ALLL May 13, 2015

2 Disclaimers All information provided is of a general nature and is not intended to address the circumstances of any particular individual or entity. Although we endeavor to provide accurate and timely information, there can be no guarantee that such information is accurate as of the date it is received or that it will continue to be accurate in the future. No one should act upon such information without appropriate professional advice after a thorough examination of the particular situation. Any similarity between any depiction in this course and any actual event, person or entity is purely coincidental. 2

3 Agenda CECL Status Implementation Checklist People Integration between Credit and Finance Data and Process Internal Controls Models Technology Landscape Next Steps Q&A Appendix CECL Overview Disclosures Key implementation considerations 3

4 CECL Status FASB developments April 22 nd FASB meeting determined to - Amend the definition of PCI assets by replacing the word significant with more than insignificant. And allow only PCI assets purchased in a business combination to be accounted for under the gross-up approach expanding the amount of purchased loans to be considered PCI. Basel Committee consultative document on Guidance on accounting for expected credit losses Issued to provide regulatory view on IFRS 9 implementation and comment period is open for all stakeholders US Bank regulators to vote adoption of standard in the future Next Steps: FASB issuance of final standard - Clarity on terms, expectations, etc. PCAOB and Bank Regulator expectations for the Standard 4

5 Partnership between Credit and Finance CECL requires the partnership of specialized knowledge set that Credit and Finance can contribute to the solution Criteria for a successful CECL implementation partnership: Periodic team meetings Delegation of responsibilities to functional knowledge expertise Leading a cross functional loan operations, IT, accounting, credit, etc. Shared success mindset Education for teams message to a diverse crowd Internal Loan/Relationship Officers and Loan operations Implementers (people directly impacted) External Investor Relations Management reporting Board reporting 5

6 Data and Process CECL requires effective data capture over originated loan data and other significant credit quality factor events Data Inventory to meet this requirements Data Supply Chain - data needs are so new even though this may not be required now it may be in the future - from time customer walks in to origination to payments to partial payments to modifications Key operational steps for CECL accumulation, computation, iteration and; results If there is a gap in data, or management judgment is used, document, document, document. 6

7 Some Basic Data Requirements for CECL model Servicing System Data: 1. Loan characteristics a. Term renewals, modifications and extension dates b. Rate fixed, variable, and index 2. Loan "credit enhancements 3. TDR events 4. Collateral fair values 5. Transaction codes in loan servicing systems 6. Loan geography 7. Contractual cash flows for each loan or pool of loans Risk System Data: 1. Loan classes and grades 2. Modification history by loan class 3. Loan renewal and funding history by loan class 4. Loan prepayments history by loan class 5. Loan cash flows not collected 6. Loan grade LGD and PD rates 7. Loan grade charge-off history 8. Loan grade migration history 9. External historical loss rate by loan class 7

8 TOP 10 loan servicing system data issues 1. Missing, incorrect, or modified maturity dates 2. Variable rate loans (interest rate caps/floors) 3. Carve outs and other optionality on commercial loans 4. Participations 5. Balance Transfers 6. Foreclosure events (OREO) 7. Sale events 8. Modifications - often see these as a new loan and payoff on the old loan with no connection 9. Charge-offs (partial and full) 10. Accounting related: 1. non accrual identification (loan passed due beyond x days per the policy and still on accrual) 2. correct contractual level yield for fee amortization 3. proper non accrual accounting, including re-accrual 4. intent (Held for sale, Held for Investment) 8

9 Zions - Present Incurred ACL Process Loan Source Systems Historic and current period loan balance information Enterprise Data Warehouse Historic and current period loan balance information Loss Migration Database Charge-off and recovery data collection Current period loan balance information Loss factors based on losses and recoveries over the loss emergence period Enterprise Data Warehouse Non-impaired loans ACL quantitative calculation: produces a range of potential ACL levels ACL qualitative adjustment: estimate point in the range based on current and past information ACL Impaired loans Impairment Documentation Tool: Probability-weighted impairment estimate based on current and past information 9

10 Data Considerations for CECL Future Expected ACL Process Loan Source Systems Current period loan balance information and contractual cash flows EDW Historic and current period loan balance information and life of loan contractual cash flows Evaluated as a pool EDW ACL quantitative calculation: produces a range of potential ACL levels Historic and current period loan balance information and life of loan contractual cash flows Loss Migration Database and prepayment model Loss factors based on expected events, discounted cash flow shortfalls for the life of each loan, prepayment, and modification ACL: estimate based on current and past information and future expectations Charge-off and recovery data collection and cash flow information collection for all historic loans ACL Evaluated on loan-by-loan basis Impairment Documentation Tool: : Probabilityweighted impairment estimate based on current and past information and future expectations Shaded text indicates changes in current data processes and controls including: 1. Collection of life of loan contractual cash flow information for current and historic loans (including loans with non-standard payment schedules) 2. Collection of realized cash payment information for all loans 3. Calculating loss factors based on cash flow shortfalls for life of loan, prepayment, and modification 4. Developing a view of future expectations 5. Incorporating future expectations at the loan level if required 6. Developing a prepayment model 10

11 Internal Control Development and Operations for CECL SOX certified CECL process needs to consider a record to report mindset Record to Report - management process for providing strategic, financial and operational feedback to understand how a business is performing from start to finish. Accountable functions for the data capture, accumulation, calculation, integration, and reporting activities/control owners Loan Operations data capture controls IT data capture, accumulation, and integration controls Accounting accumulation and reporting controls Credit accumulation and calculation controls Iteration process to calculation the allowance will require controls impacted with iteration 11

12 Record to Report Key Controls for the CECL process Data Capture Controls Need to ensure data is captured properly from origination if interest rate is incorrect at origination, contractual cash flows will be incorrect. Integration Controls Need to ensure proper controls across integration points and interfaces between systems. Limit transitions and need for unnecessary integration Accumulation Controls Models and calculations need an aggregation point and a platform so that there is no need for a manual process to bring results together Calculation Controls Calculations should be embedded and integrated into the overall process and not manual and in excel Reporting Controls Limit manual pivots and slice & dice. Need loan level data mart to ensure multiple views of the data 12

13 Record to Report Controls Map Groupings Data Capture Integration Accumulation Calculation Reporting 5/13/

14 CECL in the Context of Other Credit Risk Models Parameter attribute Proposal (CECL) ALLL (current GAAP) Basel A-IRB CCAR/DFAST When is a loss recognized? (recognition threshold) No recognition threshold, updated at each reporting date When a loss is probable, or incurred as of the financial statement date No recognition threshold, updated at each reporting date Same as current GAAP, based on the economic scenario How much of a loss is recognized? current estimate of all contractual cash flows not expected to be collected The amount of unconfirmed loss, estimated as of the financial statement date, for loss events occurring prior to that date (loss emergence period concept) 12 months of defaults (either a throughthe-cycle average or a current point-intime estimate ) + LGDs 9 quarters of expected losses + change in ALLL What information set is used in determining a loss? Past events and current conditions + reasonable and supportable expectations about future Past events and current conditions Past events and current conditions Past events, current conditions, and forwardlooking scenarios determined by the bank and regulators 14

15 Common Misconceptions about CECL Models If you like your [data and allowance models], you can keep your [data and allowance models]. (quote adapted from a politician) The measurement of expected losses for regulator capital purposes may be a starting point for estimating ECL for accounting purposes. (BCBS) However: ECL can have many meanings, depending on the context! For example, let s just focus on PD. PD parameter attribute Marginal or cumulative Loss forecast resolution Time horizon Modeling goal or bias Proposal (CECL) ALLL (current GAAP) Basel A-IRB CCAR/DFAST Term structure of marginal Cumulative Cumulative Term structure of marginal default rates default rates Monthly Single point Single point Quarterly contractual term, considering prepayments but not extensions, renewals, modifications unless TDR expected Conditional on reasonable and supportable forecasts Loss emergence period Point-in-time as of today + prudent, conservative, but not excessive 12 months 9 quarters Point-in-time or through-the-cycle average Conditional on potential scenarios 15

16 Integrated technology and data landscape 1 Business hosts the CECL data party not IT Upstream Data Sources Operational Data Store (ODS) 3 2 Data User Interface Business Rule Engine 4 Applications / Calculation Engines 1 2 ODS, Data warehouse need everything to talk via a common financial language Who is responsible for this map - Chief Data Officer Technology tool configuration a function of Bank s environment integrated tools to solve CECL record to report process: Loan and Risk systems Other loan data capture systems 5 3 Operational Data Store & Data Warehouse Data Warehouse Reports and Extracts 5/13/ Business Rule Engine/Calculation Engines Reports Platforms/Systems -

17 Next Steps to consider to operationalize the CECL proposed process - Define key terms such as credit enhancements to reduce confusion over meaning of term as it relates to CECL process. Clarify expectations for usage of assumptions in the CECL standard to ensure preparers and auditors understand and meet the standard s expectations. Enhance implementation guidance to provide other illustrative examples that provide greater insight into how to adjust contractual cash flows for prepayments, but exclude extensions, renewals, and modifications. Reduce Complexity through additional practical expedients which do not depend on an expanded SOX controlled data environment where appropriate. 17

18 Current Bank Regulator CECL expectations: Regulators will not be [publicly] prescriptive regarding methodologies Regulators will expect a greater focus on origination data (i.e., vintage ) to influence CECL estimates Regulators will expect larger institutions to use more sophisticated methodologies than smaller institutions Regulators expect to issue guidance shortly after FASB issues the standard Regulators expect the CECL learning curve to extend beyond the implementation date Banks should expect regulator criticism (similar to CCAR) Banks will probably make significant post-implementation methodology changes Engage with you local Regulatory examination team to set expectations for CECL assumptions and implementation plan 18

19 Q&A on Key Operationalization topics Modeling approach: existing risk mgmt. systems or different? TDR: operational challenges of permanent write-down Time value of money: discount rate for different cash flows Reasonable and supportable forecasts: internal vs. external sources and mean reversion Renewals Qualitative factors 19

20 APPENDIX 20

21 Disclosure Requirements for CECL model Disclosure Requirements Roll forward of loan balances Schedule of collateral values Forecasted economic assumptions embedded in ALLL model Loan-to-value ratios ALLL segmentation by loan grade Policy elections Vintage analysis Loan originations 21

22 Data Expectations in the Audit Process and Regulatory review Auditors and Regulators expect to audit/examine a well controlled data environment that will support the future CECL calculations. Consistent, reliable, well defined, well controlled, accessible, and high quality data is key for regulatory reporting associated with CECL. Existing CECL proposal will require additional investments to establish this new back office process to ensure a SOX compliant data process Federal Reserve, OCC, FDIC and Basel Committee Basel Committee on Bank Supervision - Principles for effective risk data aggregation and risk reporting Call Report requirements per the FDIC to preserve data integrity and accuracy OCC Bulletin on Risk Modeling including expectations for data integrity THIS HAS BEEN UPDATED Auditor / SEC/ PCAOB - STAFF AUDIT PRACTICE ALERT NO Considerations for audits of Internal Control over Financial Reporting October 24, 2013 SEC Financial Reporting and Audit Task Force accurate and reliable financial reporting July 2,

23 Key impact considerations for implementation of the proposed CECL model 1. Creating a well controlled, consistent, and SOX compliant data environment to support CECL model 2. Expansion of the SOX environment over a wider data environment in the back office 3. Application of assumptions and data into the CECL model for: a. Prepayment analysis by product type b. Application of data elements such as Credit enhancements 4. How will preparers set up an process to be efficiently maintained and audited? What are the audit expectations for this new process? 23