REGROC or RAROC or ROCVAR Practical Optimal Credit Portfolio - - PowerPoint PPT Presentation

REGROC or RAROC or ROCVAR

Practical Optimal Credit Portfolio Management

Risk Practitioners Conference October, 2012

Randy Miller Senior Vice President Global Portfolio Strategies Bank of America

Practical Optimal Credit Portfolio Management

Forming a Strategic and Tactical Partnership to Influence the Business

All Data is Indicative Only 2

Manage Risk Well Deliver for Shareholders Maintain a Strong Balance Sheet

Challenges

• Bias for short term growth and earnings versus

future earnings volatility

• Conflicting Risk Metrics
• Pressure to compromise first line of defense
• Rear view risk measurement and management
• Rear view migration risk
• Limited liquidity
• Silo Business Plans

Opportunities

• Optimize for stable growth and earnings
• Optimal Risk Metrics
• More nimble first line of defense
• Forward-looking risk measurement and

management

• Forecast migration risk
• Manage to extended rebalancing times
• Coherent Optimal Business Plans

Basel

All Data is Indicative Only 3

Understand The Risk Metrics

EC versus RC versus CVAR

All Data is Indicative Only 4

Risk Metric Correlation Which is Binding?

Junk Speculative Grade Investment Grade

CVAR B2 RWA Ratings $$ RC or EC

IG SG Junk

Optimal Portfolio Modeling - Overview

All Data is Indicative Only 5

Maximize (Return ~ Accrual, RANIM, MTM, Relationship) Subject to:

• Risk constraints ~ Reg Cap, Econ Cap, CVAR
• Portfolio exposure growth constraints
• Segment exposure growth constraints
• Segment asset quality constraints
• Simplified version of Rockafellar and Uryasev CVAR optimization

All Data is Indicative Only 6

• Constraints avoid ‘corner’ solutions
• Constraint setting can consider LOB views on realizable business opportunities
• Available model setups:
• Generate efficient frontiers
• Total risk-weighted asset constraint
• Set freeze/decrease only segments
• Customer acquisition opportunities
• Pro forma asset quality constraints

Optimization Model - Model Setup

The Art of Optimization

All Data is Indicative Only 7

Optimization Model – Process Summary

Simulated Losses

Segment Level

Simulation Output Exposure Level

Optimizer

Revenue and Other Relationship Information

Portfolio File

Portfolio Characteristics Segment Level

Optimal Weights

Segment Level Associated Revenue/Risk

Constraints

Commercial, Consumer Portfolio Level, Segment Level Risk Limits Aggregation Map

$15 $20 $25 $30 $35 $40 $45 $15 $20 $25 $30 $35 $40 $45 Return Risk

14% 13% 18% 16% 13% 15% 19% 34% 29% 25% 17% 14% 14% 10%

Practical Optimal Portfolio Opportunities

Integrate forecasts with what the portfolio and the market give you over the chosen horizon

All Data is Indicative Only 8 Tactical Next 12 months Strategic Full range of portfolio changes

Recovery Mid-cycle Downturn

21%

Current

Optimization in Practice

Convergence with Business Plans

All Data is Indicative Only 9 Balance Growth: Optimization Strategy Balance Growth: LOB Strategy

1

Optimization growth more aggressive than LOB

2

LOB growth more aggressive than Optimization Optimization reductions more aggressive than LOB growth LOB reductions more aggressive than Optimization growth

3 4

LOB 4 LOB 5 LOB 2 LOB 9 LOB 8 LOB 10 LOB 7 LOB 11 LOB 1 LOB 3 LOB 6

• 20%
• 10%

0% 10% 20%

• 20%
• 10%

0% 10% 20%

1 2 3 4

Plans & Optimization Converge - Process

All Data is Indicative Only 10

• Business plan and
• ptimization are

recommending different growth and reduction

• pportunities

Two Separate Plans

• Collect business

plans at the most granular level possible

Collect Plans

• Evaluate and

compare business plan with both the tactical and strategic

• ptimization

Evaluate Plans

• Understand

business plan and adjust

• ptimization
• Adjust business

plan to be more

• ptimal

Drive to Consensus

• Business plan and
• ptimization

converge

• Optimal

benchmark provides

• pportunity to

track progress and adjust as conditions change

Integrated Plan

Iterate

Forecast Cycle Views – 1 to 3 year Horizon

All Data is Indicative Only 11

Baseline Consensus

Recession Recovery Mid-Cycle Downturn

Alternative Scenarios

• Stress
• Statistical
• Specified
• Market’s Greatest Hits
• Long Run Average

Determine Strategic Enterprise Asset Allocation

Depends on economic outlook. Retail 1 and Wholesale 2 cycle “swing plays”

All Data is Indicative Only 12

5% 10% 15% 20% 25% Retail 4 Retail 3 Retail 2 Retail 1 Wholesale 3 Wholesale 2 Wholesale 1

17% 17% 17% 12% 10% 14% 11% 21% 22% 18% 13% 21% 13% 14% 12%

Downturn Recovery Current % of Total Balance

‘Swing Play’ ‘Swing Play’

All Data is Indicative Only 13

Reduce RWA > 5% and Improve Portfolio REGROC, RAROC, or ROCVAR

15 20 25 30 25 30 35 40 45 Return Risk

= Efficiency (Baseline)

0.60 75% 0.72 0.73 0.72 0.70 0.68 94% 90% 88% 85% 95% = Optimized RWA (Baseline)

All Data is Indicative Only 14

15 20 25 30 35 40 45 50 10 15 20 25 30 35 40 Return ($B) Risk ($B) Strategic Horizon Tactical Horizon Optimal Efficiency = 1.75 – 2.5 Current Efficiency = 0.98 Optimal Efficiency = 0.93 – 1.2

Efficiency = Return/Risk Target Optimal Efficiency for 2012 is 1.1 by year-end

Determine highest value tactical plays

Plays

Deals and Portfolio Performance

Operate within Risk Appetite Standards, AQ Standards, and Limit Guardrails

All Data is Indicative Only 15

= Deal = Concentration

Enterprise Portfolio

Looking Ahead

All Data is Indicative Only 16

• Counterparty
• Liquidity Ratios and Interaction with Credit Risk
• Full Balance Sheet Optimization