Week 04 4

1
Week 04- 4

• How Accrual Accounting Affects Loss Recognition
• Difficulties in Accounting For and in Pricing
  Credit Risk
• Textbooks Definition of the Equity Spread
• Leeway in Loan-Loss Reserving Different Ways to
  Benchmark a Breakeven Loan Rate
• Standalone Risk-Adjustment Methods
• Public-Policy Issues in LLR Decisions
• Charge-Offs and Workouts
• Deciding When a Bank Has Gone Bad?
• Hamilton National Bank Case

2

• Chapter 7 (p.45) defines R as the breakeven
  contract rate for zero value creation on a
  one-year loan.
• R' sets the discounted PV of the loans
  projected after-tax cash flow equal to the
  economic capital (K) allocated to support the
  risk of the loan.
• Because the answer is independent of the amount
  of loan, we may conveniently assume L 100
• K (1 t)x(R'x100) ix(100 K)K
• 1 COE

3
Finding R Example K 8 t 40 i 10
COD COE 15 Plugging data in gives 8
(.6)(Rx100) (.10)x(92) 8
1.15 Cross-Multiplying and starting to gather
terms in R' 9.2 8 (.6)x(Rx100) -
.6(9.2) (1.2 5.52) / 100 6.72/6 .6R R
11.2
4

• The Text defines the Equity Spread as the
  interest margin of R above i (Cost of Debt)
  that is necessary to allow risky loans to return
  the target COE to shareholders
• The calculation made in the previous slide gives
  an equity spread of 1.2. The spread can be
  thought of as
• Stated net of any operating expense entailed in
  making the loan and
• Impounding all loan-loss reserves implicitly in
  the capital allocation

5

• If the tax rate is halved to 20, the equity
  spread would narrow to 0.7 on this class of loan
• Similarly, if the loan were safer so that the
  capital allocated could be halved to 4, the
  equity spread would also narrow to 0.6
• If both t and K were halved, the breakeven spread
  would fall to 0.35
• (Table 7.1 on p. 47 of Text calculates R for
  many combinations of t and K).

6
P. 53 calculates R for a two-year 100 mil. loan
under the following conditions
What if there are two opportunities to default?

• t40
• Black-Box Equity allocation 6
• Cost of Equity 13.2
• Probability of Default in Year 1 0
• Probability of Default in Year 2 3
• Recovery in Case of Default 60
• i1 in Both Years 10
• Query Which two items are observables and which
  are merely projections?

7
K The sum of the PV of net cash flows in
possible outcomes that can be earned in the two
years.Query How many possible outcomes in Year
One? In Year Two?
8
Accounting Leeway Managers can lessen the
market-enforced equity spread by hiding risk from
stockholders and regulators. Opportunities exist
to reserve for and charge off problem loans in
deceptive ways

• Inaccurate provisioning can make reported
  earnings and capital appear larger or smaller
  than true economic income in a loans early life.
  Short-timers want to report larger current
  earnings incoming new team would prefer to put
  burdens on predecessors.
• Managerial Incentive Conflict in Discretionary
  Provisioning of Loan-Loss Reserves (LLR)
  relates to career enhancement and compensation
  arrangements
• To protect depositors, creditors, other
  contractual counterparties, and stockholders, LLR
  decisions are subject to review by
• 1. Board of Directors
• 2. Internal and external auditors
• 3. Bank regulators
• 4. The Securities Exchange Commission (SEC).

9
1998-2000 witnessed an open fight in the U.S.
over LLR auditing standards. The SEC, bank
regulators, and the American Institute of CPAs
(AICPA). CPAs, SEC were pushing in a different
direction from Bank Regulators.

• SEC proposed leeway to break LLR between
• a loss-accrual contra-revenue account that
  would take losses through current earnings
• a valuation account that could record
  fluctuations in the balance-sheet value of assets
  and liabilities directly against capital without
  passing the changes through current earnings
• i.e., without affecting a banks price-earnings
  ratio.

10
Debate Over Loan-Loss Reserving and Loss
Recognition

• Common Goal prudent, conservative, but not
  excessive LLR provisioning that does not mislead
  investors or regulators.
• SEC and AICPA Fear earnings management per se.
  Want strong proof of loss exposure and loss
  incurrence. Would emphasize
• For exposure, external and reproducible
  scoring-model credit risk ratings
• For loss events, past-due status and internal
  external rating downgrades.
• De-emphasize historical projections and
  correlations with economic factors.
• Differ from bank regulators, who are not always
  unwilling to countenance positive or negative
  hidden capital. Bank Regulators won the
  debate on grounds that bankers need
  prudent-man flexibility to avoid
  under-reserving.
• But claims that frequent internal individual-loan
  reviews are impractical are phoney. They presume
  the use of old information technologies and
  protect the interests of dishonest bankers and
  low-tech institutions.

11
Parable that Illustrates Value to Owners and
Regulators of the Transparency (accuracy plus
meaningfulness) created by Prompt Provisioning In
Sept. 2000, a convenience store clerk taped up
the stores security camera prior to emptying the
cash drawer and claiming that he was held up.
Critical flaw in his plan he used transparent
tape.
12
New Topic Understanding Loan-Loss Reserves (LLR)
and Allowances for Loan Losses (ALL)

• Assigning loss reserves when loans are made is
  called prompt provisioning
• Dedicated LLR are a contra-asset deducted
  promptly from a loans principal (and therefore
  NW) to get the book value of net loans (BVL).
• Chargeoffs When losses on uncollected loans are
  recognized, they are charged against the LLR
  until LLR is exhausted.
• Bathtub Analogy for LLR ALL is spigot
  Chargeoffs are the drain.
• Let us designate the level of LLR that insiders
  would understand to be a fair and accurate
  measure of expected loss exposure as LLR.

13

• TRUE NW (A - LLR) - Liabilities
• 1. Underprovisioning in any year overstates
  early earnings and cumulatively overstates NW by
  difference between LLR and LLR
• 2. Overreserving understates early earnings and
  hides NW in the amount LLR-LLR
• In practice, much leeway exists in how LLR, LLR,
  and chargeoffs may be calculated.
• No obligation exists for accountants to validate
  procedures statistically.

14
Ethical codes for Accountants feature safe-harbor
rules of thumb that let accountants earn fees
for helping managers duck their duty to provision
for credit risk realistically

• Decisions to switch between different rules do
  not have to be backed up by statistical evidence
  that switch increases transparency.
• Example Citigroups CFO announced in May, 2002
  that Citi had 5.4B in reserves for its consumer
  loan portfolio and 5.1B in reserves for its
  commercial portfolio as of March 31.
• Each figure was said to be at the upper end of
  the range of expected losses based on
  Citigroups nontransparent internal reserving
  guidelines.
• But Consumer loans would be under-reserved on a
  different LTM rule consumer reserves equalled
  roughly 90 of last-12-month losses, but the
  commercial loan reserves equalled 2.8x LTM losses
  in that lending line.

15
What do you see?
Are Loan-Loss Reserves Part of Accounting Net
Worth or Not?
16
Four Parameters Define a Shadow Price that Can
Loosely Benchmark an Individual Loans
Risk-Neutral LLR

• R lowest contract rate a competitive
  risk-neutral lender would dare to accept on loan
  (usually needs to bargain for more).
• RF risk-free interest rate
• p default probability (for convenience, assume
  either all or nothing will be received at
  maturity)
• s percentage of amount due at maturity that in
  the worst case bank can salvage or recover
  via collateral, covenants, and workouts (1s0).
  The obverse rate (1-s) is called loss given
  default or loss intensity.

17

• R solves a risk-adjustment model that treats loss
  of principal and promised interest symmetrically
• ps (1R) (1-p) (1R) 1RF
• In Words (expected salvage value) (expected
  timely repayments) (proceeds from riskless
  lending)
• The risk-premium benchmark R does not assure
  value creation. It just makes sure that expected
  returns from risky lending do not fall below
  returns from riskless lending.

18
In principle, LLR should express expected
shortfalls in contractual cash flows due on all
loans. In practice, leeway in choosing p, s,
RF, COD, and COE is routinely used to confuse
outsiders

• to improve current earnings and stock-based
  compensation
• to lighten the burden of regulatory capital
  requirements
• But underprovisioning in good times means that
  chargeoffs eat up LLR early in the default phase
  of loan cycles.
• Net chargeoffs at U.S. banks averaged .64 of
  loans in 1999-2000, but over 1.00 in 2001-02.

19

• R is not sensitive to the banks particular
  circumstances.
• To address the larger issue of whether value is
  created
• Allowances must be introduced for operating and
  capital costs (i.e., risk support) which the
  texts R? emphasizes or for net implicit interest
  from portfolio-diversification or relationship
  costs and benefits or for taxes

20
Gathering all Terms in R on the LHSR ps
(1-p) RF (1 s) p.

• Interpret R for case s 0.
• The coefficient of R represents the expected
  cash flow per each dollar of contractual interest
  promised. As long as s is less than one and p is
  greater than zero, the expected or effective loan
  rate will be less than the contractual rate.
• Righthand term adds to the risk-free rate the
  per-dollar expected measure of LGD loss given
  default.

21

• 1. Denominator states an expected return for each
  dollar lent. It inflates the numerator enough
  to restore the loans expected return to RF.
• 2. The Character of the risk adjustment becomes
  easier to interpret if we add plus and minus one
  to the righthand term

22

• is a benchmark
  risk adjustment multiplier
  applied to (1RF).
• It inflates (1 RF) to generate enough
  expected interest to offset expected
  nonpayments.
• RRF because s
• Base-Case Example p.10, s.5, and Rf.07
• step 1 Denominator .90 .05 .95
• step 2 1R 1.07/.95 1.1263

23

• Why can we interpret the benchmark interest
  spread R-RF on a hypothetical standalone loan as
  a default premium?
• Ans R-RF equals the expected contractual
  shortfall in the amount due at maturity per
  dollar of funds being lent out today.
• To exactly cover the expected per-dollar
  shortfall in loan proceeds at maturity in this
  simple case, LLR(1RF) must equal
• (R-RF).

24
Examples of how R falls with favorable movements
in p and s

• Case A (lower p) p.05, s.80, RF.05
• Case B (higher s) p.10, s.90, RF.05
• Compared to Case A, in Case B, higher salvage
  value exactly offsets higher default probability.

25

• Riskier Case C p.10, s.80, RF.05
• LLR Calculations

26
Estimating LGD or s

• SP launches loss estimation tool
• 8 April - Standard Poor's Risk Solutions has
  launched LossStats Model a tool for the
  estimation of loss-given default (LGD).
• The US version of the model is underpinned by a
  database of large corporates, with loss data from
  more than 2,500 defaulted obligations dating back
  to 1987. Inputs to the model include collateral
  type, debt position, the aggregate default rate
  and the industry default rate.
• Calculation of LGD enables banks to better
  understand their potential loss for a particular
  exposure in the event of a default. "Loss-given
  default estimates are a key requirement of Basel
  II, said Roy Taub, global head of Standard
  Poor's Risk Solutions. ?They are an essential
  component of a sound credit risk management
  process," he added. In addition to its use in
  bank internal ratings systems, the tool is aimed
  at securitisation specialists and investors. It
  can be used for exposures encompassing a variety
  of collateral, industries and subordinations of
  debt.

27
Taking a Portfolio View of LLR Creates Even More
Accounting Leeway

• Because value can be created or lost at the
  funding end, fair value of LLR is not independent
  of how loan is financed (or of the correlation of
  an individual loans riskiness with risks
  generated by other borrowers). (Equity
  allocation picks this up implicitly.)
• If loan has no adm. cost, offers no
  diversification or relationship benefits and is
  financed by equity, per-dollar reserve might be
  set as LLR

28
New Topic Public-Policy Concerns about
Inaccurate Reserve Provisioning.Problem GAAP
encourages banks to be deficient in breaking down
and managing the risk of performing loans.

• Choose too few risk categories (out of sight, out
  of mind)
• Seldom use internal or external risk ratings to
  guide
• profitability analysis (vs. contract rate)
• capital allocation
• allocation of workout effort

29
How do banks track hundreds of counterparties
with limited resources?

• 60 of credit managers review credit limits as
  needed
• Tend to focus on counterparties with lower
  credit quality or larger exposures
• Many counterparties are virtually ignored.

30
As Illustrated in the Hamilton Bank Case, Bankers
and Field Examiners Often Differ on Whether to
Reserve for a Loan Loss
Performing Borrower
31
Rosy LLR Provisioning is Typical in Booms

• This forces banks to restate past and/or current
  profits when large borrowers show or approach
  delinquency.
• Example Columbia Banking System Inc. in Wash.
  State revised its previous quarters earnings
  downward by 14.3 on 3-22-01 to reflect a
  deterioration in a single substantial credit
  relationship. Went from earning 32 per share
  to losing 1.
• Minicase in Notes Hamilton National Bank of
  Miami, FLA Dispute with OCC over need to write
  down Ecuadorian Loans before its closure in 2002.

32

• U.S. Bank Examiners define 4 categories of
  troubled loans
• Substandard loans have one or more well defined
  weaknesses that jeopardize full collection of
  that loan, and have a high probability of payment
  default.
• Doubtful loans have all the weaknesses inherent
  in those classified as substandard with the added
  characteristic that the weaknesses make
  collection or liquidation in full, on the basis
  of currently existing facts, conditions, and
  values, highly questionable and improbable.
• Loss loans are considered uncollectible and of
  such little value that their continuance as bank
  assets is not warranted. Any recovery is likely
  to occur only after lengthy recovery efforts such
  as litigation.
• Special Mention loans show distinct weaknesses,
  but collectibility still seems likely.
• Substandard, doubtful, and loss loans are
  collectively referred to as adversely classified
  assets.
• Resemble categories in weakly NFL injury
  reports probable, questionable, doubtful, out.
• Loss and out are the most-reliable categories.

33
Workout Personnel Use a Different Vocabulary than
Examiners

• A slow loan is one whose payments are
  noticeably or habitually in arrears (in practice,
  late or past due by 30 days or more). Bank
  Accountants accrue (I.e., credit) interest on
  these loans when earned rather than when payment
  is received.
• A nonperforming loan is one whose payments are
  overdue enough (90 to 180 days or more) to be
  deemed severely delinquent. Interest income on
  these loans is shifted from accrual to a cash
  basis. Interest can no longer be credited until
  it is actually received by the bank.
• An impaired loan is one whose principal value has
  come into serious question. FAS 114 defines a
  loan as impaired when, based on current
  information and events, the loan is judged less
  than fully collectable.

34
(No Transcript)
35
LLR Coverage of Bad Loans
American Banker Friday, May 30,2003
36
04
37
Difficult to Compare LLR ratios across banks.
But Loan Mix and Loan Quality may be similar at
most Money-Center Banks
38
In late 2003,
39
When Heads May Roll
loans
40
(No Transcript)
41
Minicase Underreserving for High-Risk Loans

• Rise and Fall of subprime autolender Mercury
  Finance in mid-1990s and subprime Credit-card
  lenders Providian Finance in 2001 Capital One
  in 2002.
• Story sold to analysts was that scoring
  technology could uncover misclassified low-risk
  borrowers
• Why should analysts Have Predicted these lenders
  Above-Normal Accounting Profits would Return to
  Normal Profitability?
• Growth Fallacy Reclassifications would
  eventually eliminate the alleged source of
  profits in long run
• Quality and transparency of industry accounting
  was suspect. Mercury Finance Gimmick Booked
  profits on loan sales in advance of their
  transfer to other parties via securitizations.
• Pools of subprime loans eventually fetched only a
  fraction of their book values.

42
Absurdity of Asking Investors to Assume
Possibility of Unlimited Prospecting
Sub-Primary Lending
43
New Topic Workouts of Bad Loans

• Ugly loan n a term used by regulators to describe
  a loan that is poorly collateralized and not
  priced according to risk.
• Haircut n what happens when the amount recouped
  by lenders is substantially less than the
  original loans value.
• Special asset n a euphemism for a defaulted loan.
• Mopes n pl. a bankrupt companys management.
  Usually considered by bankers to be incompetent
  or likely to be removed.

44
Ways of Dealing Promptly with Problem Loans

• 1. Watch List to Trigger Exit Strategies
• early identification of
• Leadership, profitability, and capitalization
  issues.
• Sentinel events court filings, etc.
• 2. Internal Workout Specialist(s)
• restructure loan rate, schedule, collateral
  package
• force strategic changes on borrower
• take possession of collateral
• 3. Secondary Market sell to bad loan
  specialists (Ereorg.com auction site) or partner
  lender
• 4. Pool and Securitize

45
Outsourcers are Willing to Ride to the Rescue
46
Opportunities to Sell Distressed Loans Have
Surged Since 1999

• Regulators require banks to account for impaired
  loans they plan to sell
• Such loans are supposed to be transferred into a
  held-for-sale account and written down
  regularly to cost or fair value, whichever is
  lower.
• Loans held in the banking book need not be
  marked down aggressively and usually arent.

47
Next Topic Deciding When a Bank Has Gone Bad
Peer Review Rating Reputations in Early 2003
48
WHAT HAPPENS TO AN INDIVIDUAL BANK WHEN RISKS EAT
UP ITS CAPITAL?

• 1. Regulatory Efforts both to stop and to correct
  the undercapitalization.
• 2. Customer Runs as Depositor Loss Mitigation and
  Market Discipline
• 3. Silent Runs vs. Open Runs (metaphors?)
• 4. Coping with Rational vs. Irrational Runs
• a. Assets Sales and Outside Credit Support
• b. Illiquidity vs. Insolvency
• c. Taxpayer Bailouts Explicit vs. Implicit

49
An External Regulator can Enhance the
Profitability of FS Production and Monitor
Capital Positions for Customers

• Convenience Standardization e.g., of checks
• Confidence Centralized Monitoring Safety Net
• Possible Subsidization Downside of Regulation

50

• Family-Feud Question of Top Reasons Customers
  Leave a Bank Longevity, Deal-Making, Fit
• Reasons Survey Says
• ? Weakening financial condition of bank
• ? Inefficient operations services
• ? Difficulty in negotiating credit terms and
  price
• ? Refusing to tailor credit arrangements
• ? Price for services too high
• ? Lack of personal attention
• -- Less Frequent Reasons
• ? Bank becomes too small
  (? 38)
• ? Too few cash management services
  (? 28)
• ? Lack of international capability
  (? 20)
• ? Bank becomes too big
  (? 18)

? 80

? 75

? 65
51
Hamilton Nat'l Bank Ended up with 10 to 13
Loss Rate

• Relatively Traditional Patterns of Loss
  Generation and Loss Concealment
• --High-growth concentration of risky loans
• --New Capital injected yearly over 1997-2000, but
  capital weakness concealed by overstating loan
  quality under-reserving for loan losses
• --CEO compensation extravagant.
• Aberration in Examiner assessments In 1991,
  cited poor underwriting practices in 1995,
  deficiencies in credit files in 1998, criticized
  country concentrations and surge in adversely
  classified assets. CAMELS composite mistakenly
  rose to 1 after 1997 IPO, but fell to 2 in 3-98,
  3 in 11-98, 4 in 12-99 and 5 in
• 11-01. failed 1-02.
• Efforts to reclassify bank as PCA-undercapitaliz
  ed begun in 3-01 delayed 3 mos. by Request for
  Hearing.

52
Fierce Management Resistance at Hamilton

• Board and management consistently failed both to
  eliminate violations and to comply with
  corrective orders re risk management.
• Management made aggressive and repeated use of
  its appeal rights disputed examiner findings and
  proposed writedowns and writeoffs with
  disinformation.
• Contested one OCC order in federal court.
• Irregular transactions used to mask losses.
• Claimed examiners were prejudiced against
  Hispanics

53
Lessons Drawn in IGs MATERIAL LOSS REPORT

• 1. Prior to 1999, OCC examinations and
  enforcement actions should have been more timely
  and more forceful.
• 2. Recommended specific changes in various
  policies and procedures
• 3. Without pointing out the need for supporting
  changes in the resistance rights that Hamilton
  obviously abused, OCC concurred in the findings
  and agreed to implement the recommendations.
• Since 1992, inspector general at each banking
  agency is required to review efficiency of
  closing process in large-loss cases.

54
COMMON FEATURES in IG Reports

• Strategic Sources of Failure
• Rapid and concentrated growth in a risky
  innovative activity or strategy
• Deficient Risk Management Severe
  Underdiversification of Risk accompanied by
  weaknesses in Verification and underwriting
  Procedures
• Aggressive or Fraudulent Accounting
• Not just Nonresponsive, but downright
  pugnacious management
• Assessment of Examiner and Enforcement
  Performance
• - Inadequate in some important respects
• Recommendations for Change
• Make Strategic Sources of Failure into Red
  Flags for extended follow-up examinations and
  strong enforcement actions.

55
Changing Financial Environment Expands the Range
of Loss Exposures and the Speed with which they
can be Expanded

• 1.Traditional Sources of Bank Losses
• a. Sour or Corrupt Loans
• b. Adverse Movements in Interest
• Rates or Currency Values
• c. Endgame Gambles for resurrection Funding
  Riskier Loans Expanding duration
  and currency imbalances
• 2. Nontraditional Sources of Bank Losses
• a. Residual obligations imbedded in securitized
  loan pools
• b. Basis risk in derivatives hedges
• c. Speculative (i.e., nonhedging) derivatives
  positions
• d. Lawsuits for violations of social-protection
  laws e.g., discriminatory treatment of
  classes of employees or customers
  weakness in safeguards against identity theft.

56
Runs on Many Banks in BankingSystems Crises
and Panics

• 1. Runs Spread to Many Banks Within or Across
  Countries List of Potential Triggering Events
  (on p. 40)
• 2. Banks Cannot Finance these Runs by Asset
  Sales at Fair Value or by Loans from Other
  Banks.
• 3. Fire-Sale Pressure deepens insolvencies.

57
Bank Insolvency 1975 - 2000
Systemic banking crises
Episodes of non-systemic banking crises
World Bank 2000
No crises
Insufficient information
58
Sidelight The Full Cost of Loan Refusals
Includes Resentments and Reprisals

• a. Customer-Relationship Risk Rejected Borrowers
  often feel victimized and move their future
  business elsewhere
• Major substitutes for business loans exist on
  borrowers side
• trade credit
• commercial paper
• junkbonds
• loans from friends and family
• b. Exposure to Lawsuits and Regulatory Sanctions
  Economic and legal meanings of credit
  discrimination differ
• c. BANKS AND LOAN OFFICERS SOMETIMES FACE
  PERSONAL RETRIBUTION

59
Retribution for Lending
60
Retribution for not Lending 1Shindlers
ListBanks blacklisted by Nextel CFO Steven
Shindler for refusing him in 3-98

• Long-Term Credit Bank of Japan
• Mitsubishi Trust and Banking
• Standard Bank London
• Sumitomo Bank
• U.S. Bancorp

• Bank of Montreal
• Bank of Tokyo-Mitsubishi
• Fuji Bank
• Key Bank
• Korea First Bank
• Nippon Bank

Source 1998 Nextel loan agreement
61
Retribution for not Lending 2

• Carmel, Ind. - A man apparently upset that his
  loan application had been denied opened fire on
  four bank employees yesterday, killing a woman.
  Penny Schmitt, 32, was shot several times and
  died yesterday morning after telling the suspect
  that his application had been turned down,
  authorities said. Three of her co-workers were
  injured, one critically. A massive manhunt for
  the suspect ended when he fatally shot himself as
  he crouched in a tree top with police closing in.
  (AP)
• News Item (May 1, 1998 Boston Globe)

62
Can Brokers Be Held Accountable for Bad Advice?
63
Last Two Slides are For Information Only

• Chap. 8 defines recovery as a dollar amount.
  Converting it into a proper fraction of principal
  implies that default always causes all promised
  interest to be lost and defines an s which is a
  fraction only of the loans principal. This
  simplifies both the numerator and the
  risk-adjustment multiplier.
• The modified risk-adjustment equation

ps (1 p)(1R) 1 RF (1R) (1 RF
ps)/(1 p)
64
Stylized Facts about Distribution of Loan Returns

• 1. Highly Asymmetric Left Tail of Losses is
  Long
• 2. Left Tail of Losses is Fatter at Extremes
  than the Right Tail
• 3. Nontransparency of Loan Value Makes it Hard
  to Hedge Loss Exposure and hard to Sell Loans,
  especially during Hard Economic Times.
• 4. All 3 Problems Intensify in Macro Downturns.