Capital Speedboat Session 2. Charting your way through troubling waters FARIN & Associates Inc. Agenda

Transcription

1 Capital Speedboat Session 2 Charting your way through troubling waters 1 Agenda Session 2 Defining Stress Tests Stress vs. Scenario Testing Sensitivity Testing Scenarios Silos Scenario Testing Building Scenarios Probability vs. Stress Cross Risk Considerations Setting Capital Goals Determine Impact of Stress Test Contingency Capital Plan Incorporating Stress Scenarios with Business Plan 2 1

2 STRESS TESTING DEFINED 3 OCC Community Bank Stress Testing Issued 10/18/2012 Lays Out Supervisory Expectations Primarily focused on Credit Stress Testing Generally Stress Tests Should Lay Out Plausible What If Questions Assess Impact of What-If Incorporate Results into Institution Risk Management Process If Major Problems are Uncovered at the Basic Level, Additional Detailed Analysis May be Necessary 4 2

3 FDIC Stress Testing - Defined Stress testing is a forward-looking quantitative evaluation of stress scenarios that could impact a banking institution s financial condition and capital adequacy. These risk assessments are based on assumptions about potential adverse external events, such as changes in real estate or capital markets prices, or unanticipated deterioration in a borrower s repayment capacity. Stress tests are most useful when customized to reflect the characteristics particular to the institution and its market area, and can be used to evaluate credit risk in the overall loan portfolio, segments of portfolios, or individual loans. Stress tests also can be used to evaluate whether existing financial (such as capital and liquidity) and operational (such as staffing and internal systems) resources are sufficient to withstand an economic downturn or unexpected event. Focus of all Regulatory Documents is on CRE Lending though can be applied universally in Top Down Testing 5 5 Key Stress Testing Principles: 1. The framework should include activities and exercises that are tailored to and sufficiently capture the organization's material exposures, activities and risks. 2. An effective stress testing framework employs multiple conceptually sound stress testing activities and approaches 3. An effective stress testing framework is forward-looking and flexible 4. Stress test results should be clear, actionable, well supported, and inform decision-making. 5. Strong governance and effective internal controls help ensure that the framework contains core elements, from clearly defined stress testing objectives to recommended ones. 6 3

4 Two Approaches to Stress Testing Sensitivity Analysis: refers to assessment of risk when certain variables, parameters, and inputs are "stressed" or "shocked." Unlike scenario analysis, this is performed without an explicit underlying reason or narrative in order to explore what occurs under a range of inputs and at extreme or highly adverse levels. The guidance state: "It can help to assess a combined impact on several variables, parameters, factors, or drivers. For example, an organization could better understand the impact on its credit losses from a combined increase in default rates and a decrease in collateral values...an organization can also explore the impact of highly adverse capitalization rates, declines in net operating income, and reductions in collateral when evaluating risk from CRE exposure." 7 Two Approaches to Stress Testing Scenario Analysis: a technique where you apply a historical or hypothetical scenario to assess the impact of various events and circumstances, including the most extreme situations. Examples include severe recession, failure of a major counterparty, loss of major clients, localized economic downturn, or a sudden change in interest rates brought about by unfavorable inflation developments. This is what we tie into other ALCO risks we monitor 8 4

5 Sensitivity vs. Scenario Testing: A Baking (not banking) Analogy When Baking a cake, Directions Tell you Bake in a 350 degree oven for 30 Minutes If you are in High Altitude, bake 22 minutes at 325 degrees Scenario Test: 2 scenarios based on altitude Sensitivity Test Altitude was tested over time to determine changes in scenario needed. Assumptions changed based on tracking and control But, what is assumed is that when you set your oven to 350 degrees, the oven ACTUALLY gets to 350 What if your oven is off by 10% (35 degrees hot or cold)? Scenario and Sensitivity testing combine known relationships with different potential environments and assumptions. 9 Consider this Stress Testing is to be done both within a risk area AND across risk areas. (Inter-related risks) So, where do we use sensitivity testing vs. scenario testing in Capital Planning? Sensitivity Testing: What key variables have significant impact on results. Run first in the silo How confident are the assumptions? What are the outer boundaries? Scenario Testing: Run across Silos to see combined effect on performance Stress scenarios with sensitivity tests for planning Scenario & Sensitivity Testing: Assess Combined Plan results with major sensitivity assumptions 10 5

6 Business Planning Scenarios Sensitivity Test: Scenario A Rates Funding sources & costs Credit Quality Sensitivity Test: Scenario B Projected Growth 11 SILO TESTING RISK AREAS 12 6

7 Sensitivity Testing Within Silos Typical Sensitivity Testing Within Silos Interest Rate Risk Examples What If Prepayment Speeds are wrong? Test with Faster/Slower Speeds Assumes current model is not accurately modeling performance No indication on why the speeds would change. What is the cause? Liquidity Risk Example: What if runoff factors are too low. Run at 2x and recalculate ratio What do I do then? What if Securities Can t be sold? 13 Silo Stress Testing Example of Sensitivity Testing the loss ratio impact on earnings Note that all of previous example slides (session 1) showed the estimated impact of changes in a single key variable (loss ratio) Similar testing in place for other key risk silos Liquidity Risk Liquidity Coverage Ratio (LCR) or Net Stable Funding Ratio (NSFR) Interest Rate Risk Impact of Immediate Movement in Rates on a Projects Static Balance Sheet (Income at Risk) Change key variable on Non-maturity balances and retest results 14 7

8 Sensitivity Testing: Liquidity LCR Test Short Term, Asset Based Needs Modify variables for LCR Test Loan Repay/Prepayment Investment Repay/Prepayment Stable Non-Maturity Deposit Sensitivities Less Stable NMD Sensitivities Stable Maturing CD Runoff Less Stable CD runoff Maturing Wholesale Payoff Amounts Draws on Existing Lines/Letter of Credit Run Testing Across 2-4 tests Sensitivity Link to scenarios? 15 Sensitivity Testing: Liquidity Developed a series of potential factors given an Economic projection 16 8

9 Sensitivity Testing: Liquidity Given potential changes in LCR Risk Factors, short term liquidity appears to be OK Remember this is a point in time analysis Ideally, measure this out 3, 6 and/or 12 months 17 Interest Rate Risk Stress Testing Traditional IRR Testing Run interest rate shocks/ramps/forecasts Test Margin performance for volatility Make adjustments to plan for high volatility levels Balance sheet projections mostly static Testing the existing balance sheet (Bologna!) Assumed no change in volumes for all rates! EVE tests mostly done on prior balance sheets Some forward looking EVE being done by more progressive ALCOs 18 9

10 Static Interest Rate Risk: Income With immediate movements in rates, institution earns more until rates are up by 300 bp or more how would you explain that change in direction? With this test we see how a major move in all rates impacts the future earnings assuming no change in balance sheet size or mix What s missing is how credit risk would be impacted if rates moved and/or liquidity sources 19 Static Interest Rate Risk: Current EVE Minimal volatility Implies we aren t going to see large swings in earnings regardless of rates movements. We are matched

11 Benefits: Sensitivity Testing Allows all to understand how critical assumptions can move the findings Easy to implement Disadvantages: No Enterprise-wide impact Not a realistic approach to assessing institutional risks 21 Building Stress Tests in the Speed Boat First, this is a model where earnings are an INPUT to the plan, not a calculated level Good to help establish goals and risks True ALM/Forecasting Models tell us if we can meet the goals Building the Stress Test: Credit Loss Assumptions Step 1: Stratify Loans by Category/Type Step 2: Determine Loss Ratio for Stress Test Step 3: Apply Loss Ratio(s) by Category/Type and Future Period to Projected Loan Balances Can Modify Loss Rates Over Time Step 4: 22 11

12 Stress Testing in the Speedboat Projections from Plan Goals Note that corresponding Asset Growth and Balance Sheet Mix May Change with Scenario 23 Stress Testing in the Speedboat Loan Call Report Classifications Note that corresponding Asset Growth and Balance Sheet Mix May Change with Scenario 24 12

13 Stress Testing in the Speedboat Enter Loss Rates for Scenario Enter Projected Mix % of Loans by Type Note that corresponding Asset Growth and Balance Sheet Mix May Change with Scenario 25 Stress Testing in the Speedboat Enter Tax Rate and Annual PLL Expense in Base Plan Note that corresponding Asset Growth and Balance Sheet Mix May Change with Scenario 26 13

14 Stress Testing in the Speedboat Enter addition to ALLL from Loan Stress Test Enter planned dividends Note that corresponding Asset Growth and Balance Sheet Mix May Change with Scenario 27 Stress Testing in the Speedboat Enter addition to ALLL from Loan Stress Test Bottom of sheet shows Stressed Capital Ratio on all 4 measures Compare to Base Plan Results without stress How big is the difference? Is that the buffer we need? 28 14

15 Capital Ratio Comparisons 29 BUILDING A BETTER STRESS TEST 30 15

16 Baseline Projections: Build Your Capital Plan General Economic Forecasts Base Case Key Forecast Assumptions (60%) Fiscal Policy Assumptions: Discretionary Spending. real nondefense federal government spending falls 1.7% in calendar 2012 and 2.6% in 2013 as budget cuts bite. Fiscal Policy: Expiring Stimulus. assume 2% payroll tax cut and emergency unemployment insurance benefits are extended for 2013, and phased out over several years, rather than disappearing overnight. Oil Prices Housing Demand Add in relevant parts to your planning Local Market & Institution Assumptions Asset Growth Assumed at 2-3% Loan Growth sluggish 3-5% Net Deposits growing to replace maturing Nonaccrual remaining at 5% for year 1 & 2 of plan Institution Capital Plan Goals Maintain > 7% Capital/Asset Increase ROA Define Alternative Business Plan Scenarios Economic Recovery Scenario Triggers: Improving job market Stable oil prices Housing market recovery Local business bring back 3 rd shift, increase production Impact on Capital Plan Anticipate Interest Rates increase in 6-12 months dramatically (Global Insight High Rate Forecast) Expect 10% increase in loan originations Expect non-accruals to rise to 10% before dropping late in Yr. 2 of the plan Anticipate a 5-10% runoff in existing non-maturity deposits in Yr

17 Building Stress Scenarios Create Stress Test Scenarios Must be based on potential risks in your situation Move beyond economic focus of business plan and stress key variables. Base tests on Balance sheet concentrations Local Market Risks Non-financial Risks Modify Test Relationships Duration Key impacts on other business lines Breaking of old relationships Review Results of Stress Test to Business Plan How much risk is in the plan? 33 Documenting Stress Scenarios Step 1: Describe the Scenario: Clear depiction of the issues that might cause impact on current and planned business plan objectives Step 2: Describe the Events and Impacts of Scenario on Business Plan Detailed description of the issues, events and stress impacts Silo based stress testing may drive the risk factors used Step 3: Model Results of Events on Business Plan Step 4: Review Impact and Build Potential Management Plans 34 17

18 Sample Stress Scenarios Major Loss of Key Employer Due to Weak Housing Sector: This scenario is based on a severe downturn in the local/national economy that results in significant drops in property values, employment, increased delinquency and charge-offs in the loan portfolio and possibly reductions in available wholesale funding sources Step 2: Describe the Events and Impacts of Scenario on Business Plan Anticipate increase in loan losses and delinquency to build over months Anticipate an additional outflow of non-maturity deposits of 5% in 12 months Expect a reduction in loan applications/originations We anticipate that the risk is valid in ALL of the GI Rate Forecasts 35 Alternative Scenarios Once you have defined the major events to model and assigned various assumption changes, Run all scenarios in the ALM model Knock out the irrelevant combinations to arrive at a better risk position Compare the results of the combined risk assessment to the stand alone 36 18

19 Step 3 - Comparing Scenarios: Reporting Time Series Trends Compares the key metrics over the forecast horizon Consistent with Capital Planning Goals ROA All 4 Capital Ratios Growth Rates Balance Sheet Mix Others Must watch all risk measures Interest Rate Risk Earnings at Risk Value at Risk Liquidity Risk Asset Based Cash Flow Credit Risk Decision Matrix: A good overview of relative performance between stress scenarios and Business Plan results Compares the financial results for Various measures and Projections down the left Across Various Interest Rate Projections (Economic Projections) along the top Key is to only focus on the combinations of results that are matches between forecast assumptions and rate/economic projections 37 Step 3: Reviewing Results Note that the Common Equity declined in both plans. Base plan results (solid lines) show slightly more capital if rates rise. Stress results (dashed lines) show lower results in all rates 38 19

20 Step 3: Reviewing Results Note that the Total Risk based declines below Well Capitalized levels, with stress costing roughly 25 bp more in worse case scenario. However, they remain well above Adequately Capitalized until the buffer is added at later date How much capital is enough? 39 Stress Testing Liquidity Cash Flow Gap The scenario above shows book capital declining below 7% within 2 years and a full 83 bp below base forecast Note the impact on ROA/ROE is inverted ROE = ROA * Leverage 40 20

21 Stress Testing Liquidity Cash Flow Gap Stress testing liquidity gap for economic recovery shows this institution dropping to -30.3% Liquidity Gap/Assets by End of Yr 2. What is ALCO s next move on this information? So we hold 30% of assets more in capital? We burned through available borrowing and saleable securities? 41 Alternative Scenarios Decision Matrix The following dashboard concept presents the Base Business Plan with each individual risk component, then the combined effect Incorporates all changes into financial results combined Green shaded areas are good combinations Red are not meant for review Note that these values are 1 year forecast results

22 Alternative Scenarios Decision Matrix This reporting offers opportunity to see impact of stress on many risks. 1 Yr Impact on Interest Rate Risk No significant difference in Margin volatility between pre/post stress After PLL levels decline due to increasing PLL expense building in Yr. 1 Cost of credit stress in yr 1 on Margin is less than 10 bp under nonparallel movements in rates Under Immediate rate movements, cost rises to roughly 10bp due to difference in asset size Under stress, assets actually increase as non-performing loans build up and stay on books increasing denominator for NIM 43 Alternative Scenarios Decision Matrix Let s Review Capital Levels at 12 Months Out Capital levels for the 3 ratios shown stay above Well Capitalized levels inside 1 year Stress Test hits during year 1 Illustrates need for minimum of 2 year horizon for forecasts If we are attempting to determine how much more capital we need to cover risks, is this the right way to measure, even if we looked at 2 year forecasts? 44 22

23 Looking out Long In order to evaluate the long-term effect of the risks, we believe the forward looking EVE runs should take place on valid combinations of risk parameters and rate projections Let s compare 1 year results GI High Forecast With and Without the combined stress events If we have real assumptions about growth and risk, is this a better measure of hedging needed? 45 Forecast EVE Results Under Traditional Business Plan Test Up 200, 300 and 400 bp EVE Ratio near minimum levels Stressed Plan incorporates effects of Income Projections on real capital ratio and future EVE Short on EVE by 2.11% in up 400 bp shock Would it make sense for the shop to hold at least an additional capital buffer of 2% based on EVE results? 46 23

24 Putting This To Work First, define business plan scenarios for multiple economic events Basis for dynamic management of the real business plan Movement away from static balance sheet Second, define stress events based on your concentrations and business plan What are the major vulnerabilities How bad could things get even if you don t think they will? How quickly could it happen and what might drive it Note we are focusing on the financial management risk areas, other risks like reputation risk from fraud or event risk from a natural disaster could be problematic. 47 Putting This To Work Third, assess the capital plan (macro level) with the stress events to see how plans and goals are impacted Fourth, run the plans through ALM model to determine impact of stress on earnings, capital & growth Run sensitivity testing within the scenarios Determine Buffer for capital ratio above regulatory minimums Based on Results, refine capital plan or set in place contingency plans for maintaining capital goals 48 24