Optimal Capital Allocation Strategy

1
Optimal Capital Allocation Strategy
Eric Falkenstein Moodys downloadable at
www.efalken.com ICBI Risk Management
Conference Geneva, 11/30/99
2

• Big Picture Behind Enterprise-Wide Capital
  Allocation
• Main Objectives
• Performance Measurement Within the Firm
• pricing
• incentive compensation
• exit/entry strategies
• securitizations
• planning
• Risk Management
• Informs top-down capital strategy

3
Why Economic Equity?
Without a true equity allocation within the
corporation, Net Income, ROA and ROE information
is ambiguous What we generally call profits,
the money left to service equity, is usually not
profits at all. Until a business returns a
profit that is greater than the cost of capital,
it operates at a loss. Peter Drucker
4
Internal Profitability Measure EVATM
Economic Profit (EP) Economic Value Added(
EVATM) Net Income after Capital Charge( NIACC)
After Tax Profit - Capital Charge Yield on
Asset () 8.00 loan 100 debt 100 Funding Cost
(-) -6.25 cash 6 equity 6 Option Cost
(-) -0.25 Provision (-) -0.25 Net
Spread 1.25 Fees () 0.50 Non-interest
Expense (-) -0.50 Equity Credit
() 0.30 Pre-Tax Net Income 1.55 Tax
35 (-) -0.54 Net Income 1.01 Capital
Charge (-) -0.75 New Idea!! (cap) (cost
of cap) Economic Value Added (EVA) 0.26
Assets
Liab
5
What About Status Quo Equity Levels?

• Regulatory risk-weightings do not recognize
  different loss rates for different loan types.
  Aaa and B rated obligors should not both get 8
  capital.
• Maintaining the minimum ratios does not even
  fully satisfy regulators they require
  comparability to peer bank capital levels for
  their ratings.

6
Risk Management Impact of Economic Capital
Allocation

• risk capital is a number that allows
  apples-to-apples comparison of most risks
• encourages empirical validation
• moves risk management from an audit function to a
  real part of the business, as output affects
  pricing models, incentive compensation, etc.

7
Top Down Target Considerations

• Avoid regulatory constraints

• Avoid costs of future
• financial distress

• Realize Tax Benefits

• Keep ability to make
• acquisitions

• Maximize EPS in most
• likely and current scenario

Highly Capitalized
Thinly Capitalized
Leverage
8
Top Down Rule of Thumb

• Regulatory Minimums for well-capitalized banks
  (6.0)
• capital for potential acquisitions (0.5
  Assets)
• cushion for bad times (1.5 Assets)
• 8.0 Tier 1 ratio
• Or look at peer equity levels for target debt
  rating, stay near median
• ignore if earnings are poor (EPS target is 1
  priority)

9
Reconciling Regulatory and Economic Capital

• Well-capitalized minimums Tier 1 - 6.00
  Tier 1 Leverage - 5.00 Total - 10.00
• When you cant reconcile to these (EPS woes), use
  adequate minimums
• Tier 1 - 4.00 Tier 1 Leverage -
  4.00 Total - 8.00
• Ways of Reconciling
• Use economic or regulatory regardless
• Economic grossed up/down to top-down target which
  looks at regulatory and peer requirements

10
Gross Up/Down Bottom-up Capital Allocations To
The Top-down Optimal Capital Target
Example
1.3B
Goodwill
1.2B
200
Goodwill
Other Intangible Asset Risk
200
110
Other Intangible Asset Risk
100
Operational Risk
220
Operational Risk
200
Interest Rate/ Market Risk
220
Interest Rate/ Market Risk
200
550
Credit Risk
500
Credit Risk
Bottom-Up Allocation (First-Cut)
Optimal Capitalization Level (Final Capital
Allocation)
11
Basic Idea Underlying the Economic Approach to
Equity
Earnings Cushion, sometimes mentioned
Probability of Loss
Risk Coverage Level 99.97Aa
Reserve
Capital
Potential Unexpected Losses Against Which It
Would be too Expensive to Hold Capital
Zero Losses
Potential Unexpected Losses for Which Capital
Should be Held
Expected Level of Loss
0
Loss Rate
100
12
Basic Equation of Capital Allocation

• This sort of language can give the impression
  capital is a messy but straightforward technical
  problem

13
In practice these only represent intuition for
how to approach the problem

• Graph is ambiguous about flows and stocks. If
  mark-to-market is used on commercial loans, what
  about consumer? If cash flow, what about future
  cash flows?
• Cushion available for unexpected losses book
  capital, cash flow and franchise value (i.e.,
  rents from new business not currently on books,
  which is positively related to access to outside
  funding)
• Marginal addition to firm-wide capital is the
  real target, very difficult to calculate and
  often very low (much LOB volatility is
  diversifiable within the firm)
• At the 99.95 level, different reasonable
  assumptions can lead to very different capital
  levels--too much precision counterproductive

14
Literature is Ahead of Practice, as Usual

• The problem is to construct a system of
  information, accounting, economic indices and
  stimuli which permit local decision-making organs
  to valuate the advantage of their decisions from
  the point of view of the whole (firm)objective
  just like capital allocation!
• The hard thing in a model realization is to
  receive and often to construct necessary data
  which in many cases have considerable errors and
  sometimes are completely absent, since none
  needed them previously. quite an
  understatement
• Lenoid V. Kantorovich, U.S.S.R.
• Economics Nobel acceptance speech, 1975
• Academics often speak as if the problem is solved
  when in fact it has simply been identified

15
Basis Point Capital Attribution by Credit Grade
Survey Results

• Aa A Ba B
• Low 21 35 50 185
• Avg. 78 132 196 444
• High 120 240 364 700
• Demonstrates that capital still in Beta testing
• source First Manhattan Consulting Group, RMA
  Winning the Credit Cycle Game A road Map for
  Adding Shareholder Value Through Credit Portfolio
  Management

16
How Not to Allocate Capital

• Precise portfolio algorithms just a small part of
  the process (Credit Metrics, Credit Risk, etc.)
• Allocating different LOBs with identical expected
  losses different capital is a tough sell
  internally, though a common implication
• In practice applied to mark-to-market or large
  corporate portfolios, which is a minority of most
  banks business
• In practice one jury-rigs models to generate the
  capital implication previously thgought
  appropriate, and then present it as if this
  independently validates your capital assessment!
• Monoline comparables (e.g. MBNA for credit card)
• The LOB in question is always materially
  different. Stand-alones liquidity constrained
  and not diversified.
• Any number that comes from a black box
• You need statistical backup, anecdotes and a
  transparent explanation to withstand LOB counter
  arguments

17
How to Allocate Capital

• 1) Bucket exposures into homogeneous risk
  groupings
• e.g., Product x Score x Collateral
• 2) Apply capital at the lowest level (e.g., to
  grade/tenor buckets within Media lending).
• - use loss curves on consumer to forecast
  lifetime losses
• - map commercial loans into agency ratings, use
  historical volatility of public bond defaults and
  internal experience to calculate expected losses
• - capital is usually defined as a function of
  expected lifetime losses, by product type
• - allocate other capital on case-by-case basis
  (intangibles, fixed assets, etc.)
• 3) Add up across LOBs, compare to top-down target
• 4) Gross up/down to top-down target
• 5) Redo by LOB if capital greatly at odds with
  current pricing

18
Risk Bucketing First Steps Major Categories

• Credit Risk
• Commercial Lending
• On-Balance-Sheet Loans and Lease
• Unused Loan Commitments and Letters of Credit
• Consumer Lending
• Direct
• Indirect
• Residential Mortgage
• Securitizations

• Interest Rate and Market Risk
• Interest Rate Risk (ALCO)
• Market Risk (Trading)
• Equity/Mezzanine Investments
• Operating Risk
• Deposits
• Fiduciary Services
• Non-Interest Expense
• Other
• Intangible Asset Risk
• Investments in Subsidiaries
• Fixed Assets
• Goodwill

19
Commercial Loan Products Lines

• Commercial loans
• Commercial Real Estate
• Mortgage
• Construction
• Corporate Commercial Loans
• Large Corporate
• Asset-Based Lending
• Other Specialty (e.g., Media, Health Care)
• Community Commercial Loans
• Middle Market
• Private Banking (MM lending secured by wealthy
  customer assets)
• Consumer Bank Commercial Loans
• Floor Plan (auto/marine/RV)
• Small Business

20
Example of a bottom-up capital attribution

21
Typical Consumer Breakdown

• Direct
• Closed-End Loans
• Auto
• Marine
• RV
• Mobile Home
• Second Mortgage
• Home Improvement
• High-LTV Home Equity
• Personal
• Open-End Loans and Lines
• Overdraft Protection
• Unsecured Lines of Credit
• Home Equity
• High-LTV Home Equity

• Indirect
• Auto
• Auto Leases
• Marine/RV
• Education Loans
• Credit Card
• Subprime Home Equity
• Subprime Auto
• Manufactured Housing

22
Example of Consumer Risk Buckets
23
Different Buckets for Different Lines of Business
24
Next Step Turn Loss information into Capital,
usually through a function like the following
25
Present Data to LOB with Validation and Outline
of Bottom-Up Algorithm for Maximum Effectiveness
26
Capital for Fixed Assets
Thought Experiment traditional CRE limit is a
Loan-to-value 70--implies 30 equity on
property. This makes sense until one looks at
this from a top-down perspective if top-down
target is 8, fixed assets are going to get close
to 8.
27
Intangibles
For accountability purposes 100 seems optimal,
while for cushion purposes something lower is
probably appropriate.
28
Operational Risk Categories
Fiduciary Asset number comes from study done by
First Manhattan for regulators, while noninterest
expense number makes sure everyones
covered. While important, operation risk is
almost impossible to measure by definition
solution pay valuable lip service, allocate for
noninterest expense, and ignore.
29
Deposits
Is this really necessary? Either way capital
doesnt affect the high average and marginal ROE
for this segment. Usually between 0.25 and
1.5
30
Market Risk
Balance Sheet or ALCO risk much higher than
trading risk, though not marked to market ALCO
risk usually 10 to 20 of book capital Use VaR in
conjunction with loss limits for trading books
31
Capital is not a Panacea

• Capital doesnt eliminate disagreements, it
  only makes them more explicit. People who expect
  capital to allow them to optimize the balance
  sheet from the top-down havent gotten their
  fingers dirty.
• Capital is a cost of doing business, and is as
  important as other costs. Refining other
  non-interest expenses, expected losses and
  
  
  
  
  funds transfer pricing is still most useful.

32
Biggest Blind Spots

• New businesses (artificially low current loss
  rates)
• Tax arbitrage business (e.g., cross-border
  leases)
• Business Cycle Uncertainties
• cross-sectional patterns much more stable (N gtgtT)
• small business, high LTV home equity,
  equity-collateral asset-lending aggressively
  increasing volume--recessionary performance
  unknown

33
Consultants are Key

• How consultants add value knowing the
  appropriate risk buckets, benchmarks for capital
  factors, how to construct the appropriate
  management information system, and general
  project management

34
Quick ways to test depth of capital impact within
the firm

• If no concrete answers to the following, assume
  they have not yet solved the problem, which in
  some cases implies capital is still only an
  academic exercise within the firm
• How do you reconcile economic and regulatory
  requirements?
• What portion of your capital allocation is
  allocated to credit?
• Do you have historical loss rates and forecasts
  by internal grade and LOB?
• How is capital integrated with
• Planning?
• Profitability reporting?
• Pricing models?
• Incentive compensation?

35
Conclusions

• Capital is most relevant to internal
  profitability measurements, not top-down capital
  management
• Validated, homogenous risk bucketing within the
  firm is the key
• Most firms overstate the extent capital affect
  real decisions within the firm