No Slide Title

Download Report

Transcript No Slide Title

Optimal Capital Allocation
Strategy
Eric Falkenstein
Moody’s
downloadable at www.efalken.com
ICBI Risk Management Conference
Geneva, 11/30/99
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
2
Eric G. Falkenstein
11/8/99
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
3
Eric G. Falkenstein
11/8/99
Internal Profitability Measure: EVATM
Economic Profit (EP) =Economic Value Added( EVATM) =Net
Income after Capital Charge( NIACC) =After Tax Profit - Capital
Charge
Assets
4
loan 100
cash 6
Liab
Yield on Asset (+)
Funding Cost (-)
Option Cost (-)
Provision (-)
Net Spread
8.00%
-6.25%
-0.25
-0.25%
1.25%
debt 100
equity 6
Fees (+)
Non-interest Expense (-)
Equity Credit (+)
Pre-Tax Net Income
0.50%
-0.50%
0.30%
1.55%
Tax = 35% (-)
Net Income
Capital Charge (-)
Economic Value Added (EVA)
-0.54%
1.01%
-0.75% New Idea!! (cap) × (cost of cap)
0.26%
Eric G. Falkenstein
11/8/99
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.
5
Eric G. Falkenstein
11/8/99
Risk Management Impact of Economic
Capital Allocation
– risk capital is a number that allows apples-toapples 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.
6
Eric G. Falkenstein
11/8/99
Top Down Target Considerations
•Avoid regulatory
constraints
•Realize Tax Benefits
•Avoid costs of future
financial distress
•Maximize EPS in most
likely and current scenario
•Keep ability to make
acquisitions
Highly
Capitalized
Thinly
Capitalized
Leverage
7
Eric G. Falkenstein
11/8/99
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)
8
Eric G. Falkenstein
11/8/99
Reconciling Regulatory and Economic
Capital
• “Well-capitalized” minimums :
Tier 1 - 6.00%
Tier 1 Leverage - 5.00%
Total - 10.00%
• When you can’t 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
9
Eric G. Falkenstein
11/8/99
Gross Up/Down Bottom-up Capital
Allocations To The Top-down Optimal
Capital Target
Example
$1.3B
$1.2B
Goodwill
$200
Other Intangible
Asset Risk
$100
Operational Risk
$200
Interest Rate/
Market Risk
$200
Credit Risk
$500
Bottom-Up Allocation
(First-Cut)
10
$200
Goodwill*
$110
Other Intangible
Asset Risk
$220
Operational Risk
$220
Interest Rate/
Market Risk
$550
Credit Risk
Optimal Capitalization Level
(Final Capital Allocation)
Eric G. Falkenstein
11/8/99
Basic Idea Underlying the Economic Approach
to Equity
Risk Coverage Level
Loss
Probability of
Earnings
Cushion,
sometimes
mentioned
99.97%=Aa
Reserve
Zero
Losse
s
0%
11
Capital
Expected
Level of
Loss
Potential
“Unexpected”
Losses for Which
Capital Should be
Held
Loss
Rate
Potential “Unexpected”
Losses Against Which It
Would be too Expensive
to Hold Capital
100
%
Eric G. Falkenstein
11/8/99
Basic Equation of Capital
Allocation
p 
  
i
i
j
ij
j
This sort of language can give the
impression capital is a messy but
straightforward technical problem
12
Eric G. Falkenstein
11/8/99
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
13
Eric G. Falkenstein
11/8/99
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
14
Eric G. Falkenstein
11/8/99
Basis Point Capital
Attribution by Credit Grade:
Survey Results
Low
Avg.
High
Aa
21
78
120
A
35
132
240
Ba
50
196
364
B
185
444
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”
15
Eric G. Falkenstein
11/8/99
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 bank’s 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
16
Eric G. Falkenstein
11/8/99
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
17
Eric G. Falkenstein
11/8/99
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
18
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
Eric G. Falkenstein
11/8/99
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
19
Eric G. Falkenstein
11/8/99
Example of a bottom-up capital attribution
OUTSTANDINGS ($000)

Risk Remaining Tenor (Years)
Grade Demand
1
1
(1,697)
(20,193)
2
0
194,830
3
686
853,557
4
(675)
479,472
5
0
52,334
6
0
62,421
7
0
40,474
8
3,671
0
9
0
0
10
0
0
NA
240
226,006
Total $
2,224 $ 1,888,901 $
2
3
0
2,659
16,557
482
25,025
6,247
28,782
19,192
4,595
0
253
0
0
0
0
32
0
0
0
0
22,487 (110,271)
97,699 -$ 81,660 $
4
2,535
1,125
606
1,969
667
0
0
0
0
0
15,380
22,282
5
6
7
8
9
10
0
781
10,521
3,928
0
293
182,000
0
0
0
2,960
$ 200,482
0
0
4,093
0
53
0
0
0
0
0
19,371
$ 23,516
0
0
0
0
0
0
0
0
0
0
3,820
$ 3,820
0
2,080
172
0
0
0
0
0
0
0
8,675
$ 10,927
0
0
0
0
0
0
0
0
0
0
3,605
$ 3,605
0
0
0
0
196
0
0
0
0
0
29,688
$ 29,883
Total
(16,698)
215,856
900,906
532,667
57,844
62,966
222,474
3,703
0
0
221,961
$ 2,201,679
CAPITAL FACTORS (bps)
Risk Remaining Tenor (Years)
Grade Demand
1
1
13
13
2
20
20
3
45
45
4
181
181
5
271
271
6
500
500
7
1,000
1,000
8
1,900
1,900
9
3,300
3,300
10
4,500
4,500
NA
226
226
2
3
4
5
6
7
8
18
35
68
300
429
700
1,000
1,900
3,300
4,500
365
25
50
92
363
517
850
1,000
1,900
3,300
4,500
440
33
65
119
426
596
950
1,000
1,900
3,300
4,500
511
40
85
181
500
667
1,000
1,000
1,900
3,300
4,500
583
48
100
225
588
758
1,050
1,000
1,900
3,300
4,500
673
55
120
270
675
842
1,050
1,000
1,900
3,300
4,500
758
68
150
317
725
883
1,050
1,000
1,900
3,300
4,500
804
9
10
90
210
368
775
925
1,050
1,000
1,900
3,300
4,500
850
78
180
343
731
883
1,050
1,000
1,900
3,300
4,500
807
REQUIRED CAPITAL for OUTSTANDINGS ($000)
=
Risk Remaining Tenor (Years)
Grade
Demand
1
1
(2)
(25)
2
0
390
3
3
3,841
4
(12)
8,690
5
0
1,417
6
0
3,121
7
0
4,047
8
698
0
9
0
0
10
0
0
NA
5
5,109
Total $
692 $
26,590 $
2
3
0
7
58
2
171
57
863
696
197
0
18
0
0
0
0
6
0
0
0
0
820
(4,847)
2,127 -$ 4,079 $
4
8
7
7
84
40
0
0
0
0
0
786
932
$
5
6
0
7
190
196
0
29
18,200
0
0
0
173
18,795
0
0
92
0
4
0
0
0
0
0
1,304
1,400
$
7
$
8
0
0
0
0
0
0
0
0
0
0
290
290
$
9
0
31
5
0
0
0
0
0
0
0
698
734
$
10
0
0
0
0
0
0
0
0
0
0
291
291
$
0
0
0
0
18
0
0
0
0
0
2,523
2,542
Total
$
(12)
495
4,367
10,518
1,676
3,168
22,247
704
0
0
7,151
50,314
Typical Consumer
Breakdown
Direct
Indirect
Closed-End Loans
Auto
Marine
RV
Mobile Home
Second Mortgage
Home Improvement
High-LTV Home Equity
Personal
Auto
Auto Leases
Marine/RV
Education Loans
Credit Card
Subprime Home Equity
Subprime Auto
Manufactured Housing
Open-End Loans and Lines
Overdraft Protection
Unsecured Lines of Credit
Home Equity
High-LTV Home Equity
21
Eric G. Falkenstein
11/8/99
Example of Consumer Risk
Buckets
INDIRECT AUTO
Annualized Charge Off Rate
Bureau
Fico
22
No Score 350-574
575-634
635-669
670-900 Invalid
Total
No Score
2.34%
2.35%
2.35%
2.35%
2.35%
2.34%
2.34%
1-164
7.67%
10.71%
7.78%
5.34%
5.10%
7.67%
7.83%
165-179
6.33%
5.91%
3.71%
2.62%
1.57%
6.33%
5.27%
180-194
3.92%
3.97%
2.54%
1.68%
1.14%
3.92%
2.74%
195-224
3.16%
2.48%
2.42%
1.52%
0.80%
3.16%
1.92%
225-239
1.91%
1.70%
1.60%
0.72%
0.41%
1.91%
0.91%
240-400
0.59%
0.80%
0.60%
0.47%
0.15%
0.59%
0.27%
Total
2.62%
5.39%
2.70%
1.31%
0.31%
3.12%
1.46%
Eric G. Falkenstein
11/8/99
Different Buckets for Different
Lines of Business
AUTO LEASE
Annualized Charge Off Rate
23
Fico
Business
Consumer
Total
No Score
0.41%
0.31%
0.36%
<200
1.02%
0.92%
0.97%
200-224
0.75%
0.65%
0.70%
225-249
0.36%
0.26%
0.31%
250+
0.14%
0.04%
0.09%
Total
0.73%
0.63%
0.68%
Eric G. Falkenstein
11/8/99
Next Step: Turn Loss information
into Capital, usually through a
function like the following
24
Eric G. Falkenstein
11/8/99
Present Data to LOB with Validation and Outline
of Bottom-Up Algorithm for Maximum
Effectiveness
Bureau Score As of 6/97
No Score
$9.19
<550
$42.17
550-599
$34.56
600-649
$72.04
650-699
$29.83
700-749
$47.72
750+
$8.50
Total
$244.00
Gross Losses Gross Loss
Expected Gross
7/97-6/98
Rate (%) Current O/S
Loss Rate (%)
$0.06
0.67%
$47.59
1.34%
$2.29
5.43%
$82.48
4.94%
$0.81
2.35%
$92.34
3.03%
$1.16
1.62%
$90.98
2.42%
$0.21
0.70%
$68.95
1.32%
$0.52
1.08%
$71.53
0.58%
$0.02
0.28%
$82.69
0.22%
$5.08
2.08%
$536.56
Expected Gross Expected Net
Loss $
Losses
$0.64
$0.51
$4.07
$3.26
$2.80
$2.24
$2.20
$1.76
$0.91
$0.73
$0.41
$0.33
$0.18
$0.15
$8.98
Total Capital Rate
25
1.67%
Eric G. Falkenstein
11/8/99
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%.
Category
26
Capital
Factor (%)
Land and Buildings
8% to 30%
Capitalized Value of Operating Leases
8% to 30%
Vehicles and Aircraft
8% to 30%
Telecommunications, Data Processing, PC’s and Software
8% to 50%
Other Premises and equipment
8% to 50%
Eric G. Falkenstein
11/8/99
Intangibles
For accountability purposes 100% seems
optimal, while for cushion purposes
something lower is probably appropriate.
Goodwill
Other Intangibles
27
50% to 100%
Credit Card Acquisition Premium
25%ish
Deposit Acquisition Premium
25%ish
OMSRs
25%ish
All Other Intangibles
25%ish
Eric G. Falkenstein
11/8/99
Operational Risk Categories
Fiduciary Asset number comes from study done by
First Manhattan for regulators, while noninterest
expense number makes sure everyone’s covered.
While important, operation risk is almost impossible to
measure by definition; solution: pay valuable lip
service, allocate for noninterest expense, and ignore.
Category
Assets Under Management
Mutual Funds
Personal Trust: Disretionary
Personal Trust: Nondisretionary
Noninterest Expense (Proxy for Overall Operational Risk)
28
Capital Factor (%)
0.40%
0.09%
0.04%
3.29%
Eric G. Falkenstein
11/8/99
Deposits
Is this really necessary? Either
way capital doesn’t affect the high
average and marginal ROE for this
segment.
Usually between 0.25% and 1.5%
29
Eric G. Falkenstein
11/8/99
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
Category
99% 3 month VaR ($000)
Trust Securities Lending
$5,000
Balance Sheet Interest Rate Risk
$750,000
Capital
$5,000.00
$750,000
Market Risk (Trading)
Category
Capital Markets
Customer Derivatives
Foreign Exchange
30
Loss Limit ($000) 99% daily VaR ($000)
$1,000.00
$500.00
$640.00
$277.00
Capital
$928.00
$470.00
Eric G. Falkenstein
11/8/99
Capital is not a Panacea
• Capital doesn’t eliminate disagreements, it only
makes them more explicit. People who expect
capital to allow them to “optimize the balance
sheet” from the top-down haven’t 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.
31
Eric G. Falkenstein
11/8/99
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 >>T)
• small business, high LTV home equity, equity-collateral assetlending aggressively increasing volume--recessionary
performance unknown
32
Eric G. Falkenstein
11/8/99
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
33
Eric G. Falkenstein
11/8/99
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
–
–
–
–
34
Planning?
Profitability reporting?
Pricing models?
Incentive compensation?
Eric G. Falkenstein
11/8/99
Conclusions
• Capital is most relevant to internal
profitability measurements, not topdown capital management
• Validated, homogenous risk bucketing
within the firm is the key
• Most firms overstate the extent capital
affect real decisions within the firm
35
Eric G. Falkenstein
11/8/99