Week 067

1
Week 06-7

• CREDIT RISK

• Topics
• Lending Process Duties of Due Diligence Include
  Calculating The Cost of Providing Risk Support
• Use of Scoring Software
• Credit-Risk Issues Pricing Quality in a
  Portfolio Context
• Book-Cooking Opportunities in Loan Accounting

2
1st Topic Duty of Due Diligence Lending is
Banks Principal Product-Management Chain and
Counterparties are Looking for Weak Links
3
Important Slide All forms of dealmaking have to
be funded in part by an appropriate allocation of
FSF capital. To Assure Due Diligence at the staff
level, a Loan-Review Committee Should Ask
Deal-Making Staff to Report Four Features of
Every Deal

• 1. What are the risks?
• 2. What are the costs of capital and loss
  reserves that must be allocated to cover the
  portfolio risks the loan entails.
• 3. What explicit and implicit returns does the
  proposed contract offer the firm for bearing the
  costs of supporting these risks?
• 4. Allowing for differences in risk, how does the
  risk-adjusted return line up with other deals
  that we are or might be making?

4
Modern FSFs Can Outsource Some of These Questions
5
Each Lender Employs Multiple Technologies of
Lending Deal Formats Must Adapt to the
Informational and Regulatory Environments in
Which FSF and the Borrower Operate

• Three Mutually Reinforcing Components Define the
  Technology Used in a Particular Lending Chain
• Screening Mechanisms
• Contract Structure (e.g., covenants, collateral
  rights, enhancements, amortization schedule,
  reporting requirements)
• Monitoring Strategy

6
LOAN COVENANTS
Definition Loan covenants are forms of implicit
interest that restrict a borrowers future
activities in ways designed to lessen conflicts
of interest between the borrower and lender. If
not waived, any violation of the covenant package
results in so-called technical default. 1.
Negative Covenants- restrict future production,
investment, or financing decisions
especially limitations on dividend payouts
and on future debt. 2. Affirmative
Covenants- impose contractual obligations
to submit financial statements and to
report other material issues.
7
Some Different Business Lending Technologies

• Financial-Statement Lending
• Relationship Lending
• Business Credit Scoring
• Asset-Based Lending (Collateralization Leasing)
• Trade Credit
• Factoring (Purchase of Receivables)
• Credit Insurance (enhancements)
• Classroom Exercise What are the Strengths and
  Weaknesses of each?

8
Many Business Loans and All Household Mortgage
and Auto Loans pledge property to lender as
security for a loan.
Asset-Based Lending

• What is a mortgage loan? Who is the Mortgagor?
  ANS. The borrower. Who is the Mortgagee?
• First Mortgages vs. Second Mortgages
• Lender must assess prospects of borrower default
  and the possible correlation of default events
  with changes in collateral value

9
In All Technologies, Loan Officers Must
Efficiently Collect Appropriate Information to
Make Optimally Four Decisions About Each Customer

• 1. How much to lend?
• In what form i.e., with what safeguards?
• Covenants
• Monitoring Rights
• Default Remedies
• 3. At what price? implicit explicit
  compensation
• 4. On what repayment schedule?

10
Vocabulary Lesson

• Holism is the belief that once an entity has
  existence its parts do not. The idea is that the
  parts stick together in an inseparable way (e.g.,
  life-force of a person vs. mortar in a brick
  wall).
• Test is reversibility.

11
Historically, lending was an holistic process.
Modern Lending Deconstructs the Steps Traversed
in Making a Loan

• Allows FSFs either to specialize in-house or to
  outsource the subset of risks and skills needed
  at each particular stage.
• 1. Applications Generation
• 2. Processing
• 3. Underwriting
• 4. Closing
• 5. Servicing/Collection
• 6. Insuring risk of shortfalls in payments due
• 7. Funding (temporary vs. permanent risk
  support)
• 8. Postloan monitoring and risk support or
  transfer

Origination
12
Unbundled Parts of Lending Process are
Automating, Digitizing, and Globalizing

• Outsourcing may slow some decisions intensifies
  Ethical Risk the problem of assuring due
  diligence is performed in individual functions
• with holistic loan-officer model, a continuous
  double-checking role is played by high-level
  committees who are subject to legal penalties for
  negligence and malfeasance. Good judgment
  comes from experience. Experience comes from
  exercising poor judgment error-learning.
• with outsourcing model, loan committee must rely
  on reputations, bonding agreements, and fraud
  negligence laws
• We explore these Issues in the last 2 Weeks of
  the course

13
Even Flexible Deal-Makingmust be Supported by
Due Diligence in Prospecting, Information
Gathering, and Analysis
14
First Link in Chain Generating Applications

• Referrals, Prospecting, and Prequalifications
  Computer cross-sell triggers
• Product Selection
• Application Completion
• Document Collection

15
Due Diligence in Information Gathering uses 3d-
Party sources

• Application Verification (must guard against
  identity theft false data)
• Credit Investigation
• Collateral Valuation
• Blue-Book Values for Autos
• Repeat-Sales Data Base (Automated Appraisal) vs.
  Custom Appraisal for Houses

16
Underwriting What Constitutes Due Diligence in
Analysis?

• Credit Analysis standards, guidelines vs. credit
  scoring
• Pricing
• Mortgage Insurance Decision
• Single payment vs. cancelable
• Commitment Issuance

17
Operational Links Closing and Postclosing
Activities

• Closing
• Document Preparation
• Compliance with Commitment Conditions
• Packaging and Delivery
• Post-Loan
• Postclosing Document Tracking
• Set up Servicing System
• Collections/Monitoring/Dunning

18
Final Link Funding Decision

• Temporary warehousing prior to choosing how to
  permanently finance the deal.
• Three main alternatives for permanent funding
• Own debt and capital (intermediation)
• Loan sales
• whole
• partial (syndication)
• Securitization (pooling pricing loan packages)

19
The Funding Decision Also Affects the Allocation
Across the Banks Counterparties of the Risks
That Are Left Unhedged

• Self-Insurance supporting with Loan-Loss
  reserves and Ownership Capital
• External Credit Enhancement (partial recourse vs.
  complete risk transfer to borrower or third
  parties)
• Collateral (puts some risk back on borrower)
• Recourse to borrower or corporate officers
  (ditto)
• Personal or Corporate Cosignors or Guarantors
• Private Mortgage Insurance (Mort. Guaranty Ins.
  Corp. GE Capital Mort. Ins. Corp. United
  Guaranty Corp.)

20
2nd Topic Automation of Due-Diligence and
Pricing Activity Computer Scoring

• Theme Scoring is Driving Automation of all links
  in the lending chain
• Value of scores depends on size of underlying
  sample and representativeness of its relevant
  subsample cells.
• Mines or Tortures Data to make them sing
  Uncover Recognizable Patterns and Convert them
  into Point scores that classify customers in
  terms of probability of some targeted form of
  behavior
• Credit scores can be fed directly into an
  implicit and explicit loan pricing matrix.
• Targeted behavior can be anything. AI Expert
  Systems can identify loan leads, slow payers,
  deadbeats, profitable customers, volatility of
  collateral value, etc.
• In use at all large U.S. banks most small ones.
  33 of small banks by early 2001.

21
Table 1Survey Results for Large U.S. Banks Using
Small Business Credit Scoring Data as of January
31, 1998
Source Frame, Srinivasan, and Woosley (2001)
22
When Automated Lending Works
23
Use of Scoring Presupposes and Shapes the
Collection and Verification of Databases

• Individual Application input
• External Credit Bureau Input (will score for and
  sell credit directly to customers)
• KnowX.com and Lexis-Nexus input data on arrests,
  scandals
• Internal Credit Information File (CIF) from data
  warehouse (FSF base of information by which it
  manages)
• Lenders must address customer and legal concerns
  about privacy and accuracy

24
Computer credit-scoring models objectify credit
standards. They input multiple proxies for
ageold Five Cs of Credit.
Intuitive Basis for Scoring

• The 5 Cs of Credit a checklist of informational
  items that track a customers unobservable
  repayment speed and repayment probability. Scores
  should be tracked both before and after making a
  loan. Why?
• Scores on proxy items can be used also to size
  and price a customers serviceable demand for
  borrowed funds.
• Increasingly, computer credit-scoring models
  re-estimate the rate outstanding loan portfolios
  should carry and business loan contracts reset
  the loan rate when and as a borrowers score
  changes.

25

• Character customers reputation for probity and
  fairness (past willingness to pay bills can be
  checked with credit agencies)
• Capacity projected future income of customer
  payment-coverage ratio.
• Capital strength of customers balance sheet
• Collateral any credit enhancement offered
  --consists of implicit and explicit guarantees,
  including right of recourse
• Conditions re economic cycles how changing
  economic environment affects the customers other
  Cs measures of vulnerability or fragility.
• Modern financial economists add 4 more Cs
  Regulatory Compliance Costs, Customer
  Relationships, Correlation, and Costs of a
  borrowers opportunities for hidden actions and
  hidden information disadvantage a lender.

26
Major External Vendors of Online Scoring Services

• Fair Isaac (Grandaddy of scoring)
• Fiserv
• Credit Bureaus
• Global national (Transunion, Experian, Equifax)
• Local
• Mark-It Partners credit database (shareholders
  are an evolving member of European and US
  megabanks) Partner banks supply price
  information.

27
About Fair Isaacs ScoresThe formula for the
Fair Isaac creditworthiness score deals only with
financial information about a borrower and
doesnt consider such factors as place of
residence, age, race, sex, or nationality.

• Factors in determining the credit score and the
  weight they are given

28
How the Fair Isaac Score Works

• Fair Isaac licenses its software to credit
  bureaus.
• Based on the credit information on file, the
  credit bureau uses Fair Isaacs formula to
  generate a credit score, also known as a FICO
  score. Scores are on a 900-point scale.
  Generally, a score of 640 or higher results in a
  mortgage on favorable terms subprime (prime, superprime ratings (720)
• Lenders acquire from a credit bureau a borrowers
  credit report and FICO score to evaluate the
  applicants creditworthiness and price loans.
• High Score High Probability of complete and
  timely performance by borrower.
• Fair Isaac traditionally limited the information
  passed to borrower. Now, loan applicants can
  purchase their scores and use experts to improve
  their score to a lenders threshold.

29
DISCUSSION QUESTIONS ON SCORINGPlease indicate
in one paragraph whether you agree or disagree
with the following statements and why

• 1. Credit-scoring software is making human loan
  officers obsolete. Software can identify several
  times as many potential losses as most
  institutions best human underwriters can.
• 2. Credit-scoring is a new and untested idea.
• 3. Scoring software is useful only in
  originating and pricing loans.
• 4. Once an institution switches to
  credit-scoring software, its approval rates
  usually decrease.

30
DISCUSSION QUESTIONS ON SCORING (continued)

• 5. Credit-scoring software can be used only on
  loan applicants that have a prior credit history.
• 6. Is scoring fair to immigrants and low-income
  households?
• 7. Mortgage-Loan automation can consolidate the
  many steps on mortgage lending into a single
  virtual back office that can bid on (but not
  seal) a deal in a matter of minutes.
• 8. It is good practice to explain and doctor
  credit scores for rejected customers.
• 9. It should be a source of pride to some
  bankers that they dont use credit scoring.

31
Course Theme Reshaping of Job Opportunities and
Branch Architecture by New Lending Technologies

• 1. Calling Officers (Salespersons or Drummers).
• 2. Credit Analysts
• 3. Loan Review Committee
• 4. Workout Specialists

32
Business-Loan Officers Going the Way of the Dodo?
MINICASE On Morphing of Firms Employee Skillsets
33
(No Transcript)
34
3rd Topic Credit-Risk Issues

• What is risk? The downside of a deal caused by
  its negatives.
• What is a loans credit risk? Obverse of Asset
  Quality
• Danger that an individual counterparty wont
  perform as promised in a contract a priceable
  chance of suffering default losses - e.g., in
  loans to customers who go bankrupt.
• Delinquent vs. truly nonperforming loans
• workouts on partial vs. complete defaults show
  some chargeoff against bank loan reserves or
  net worth

An obverse is the positive quality that can be
obtained from a negative or the negative quality
corresponding to a positive.
35
Risk Management ? Risk Avoidance. Ways an
institution can price, support, diversify or
transfer credit risk introduce the concept of
financial engineering.

• Develop and maintain an accurate and consistent
  risk-grading system.
• Establish a credit culture that reflects the risk
  appetite of your institution.
• Use a loan approval process, by committee or
  otherwise, that provides a formal, systematic
  review of total exposure to a borrower in
  different products.
• Implement credit scoring, though it need not be
  the sole factor in lending decisions.
• Establish a credit database with heavy emphasis
  on the collection and retention of risk ratings.
• Develop a process to review and control
  exceptions to credit policy.
• Introduce a pricing model, even a rudimentary
  one, to bring disciplined risk-based pricing into
  the underwriting and review process. Even
  better, use a pricing model, either internally or
  vendor purchased, to price your credits to
  reflect risk, relationship, and capital
  allocation.
• Manage your loans as portfolio investments.
  Quantify the return relative to the risk and
  manage portfolio diversification.

36
Managing Risk in Loan Origination

• Adage Every debt is paid, if not by the
  borrower, by the lender.
• How does a lender pay for borrower defaults?
• Lending officers may be disciplined for defaults
  or credit deterioration on loans they originate.
• Computer scoring cannot capture every negative
  Jürgen Schneider Warning Signal Wasteful excess
  shown in personal life by CEOs gilding an iron
  fence in 1994 was used by a smart lender to be an
  indication of bad character.

37
Recognizing Loan ScamstersTop Ten Warning
Signs (Paul Nadler, American Banker, March
1996)Beware of Customers who

• 10. Soft-soap you.
• 9. Seem too dumb to fool you.
• 8. Pride themselves on breaking rules.
• 7. Focus on what your firm will do in the event
  of delinquency.
• 6. Seem unusually charming.
• 5. Seem unusually optimistic.
• 4. Challenge your policy on overdrafts or on the
  use of uncollected funds.
• 3. Always want to meet you on your premises.
• 2. Whose buildings and equipment show signs of
  neglect.
• 1. Who appeal to your greed.
• 0. Seem unconcerned about payments Schedules
  and Interest Costs.

38
Concept of Pricing a Loan

• Concept of a loans price as an opportunity cost
  value of all implicit and explicit compensation
  received for extending this credit.
• Every contractual requirement is a potential
  burden. Nonprice terms is an oxymoron.
• Benchmark-plus Pricing In practice, rates are
  set as spreads above a low-risk benchmark
  interest rate.
• a. Libor
• b. Treasury yields
• c. Own CD or prime rate
• Etymology of benchmark known height of a
  prominent landmark or some kind

39
Why do the following items Constitute Implicit
Interest on a Business Loan?
Explicit Interest is Not the Full Price of a Loan

• 1. Compensating-Balance Requirements (a legal
  tying of use of deposit and loan products).
• Does it ever make sense to pay interest on ones
  own money? Credit repairs/ money laundering
• 2. Collateral Requirements.
• 3. Covenant Rights.
• 4. Monitoring requirements.
• 5. Coercive tie-in arrangements (mostly
  unwritten) though illegal, these are alleged to
  exist

40
The credit risk bankers must Price and Manage
is a Portfolio Concept

• Measures of this risk must identify and account
  for
• 1. Correlations in underlying risk factors
  that cause individual-customer default events
• 2. Correlations in the size of the losses driven
  by different events and risk factors
• 3. Last half of course will introduce the
  effects of hedging transactions that mitigate
  or transfer particular categories of loss
  exposures (hedging will be the focus of the
  middle weeks of this course).

41
GOOD AND BAD LOANS CAN BE SOLD
42
Memory Device A Financial-Company Portfolio
Garages a Fleet of Investment and Funding Vehicles

• With tangible long positions in, e.g.,
• Loans
• Marketable Securities
• Real Estate and Equipment
• With tangible short positions in
• Deposit-like accounts
• Debt
• Explicit commitments (insurance obligations
  securities held in street names)
• With numerous harder-to-monitor derivative and
  intangible positions

43
To Manage the Risk of Operating This Fleet, One
Must Understand the Dangers the Fleet Faces Its
Exposure to Deterioration in Net Value from
Various Kinds of Adversity

• Unexpected market moves.
• Model risk a source of hedging errors.
• Insufficient management oversight.
• Carrying too much risk relative to capital.
• Internal external fraud
• Counterparty lawsuits.
• Unsupportable Debts.

44
Risk Landscape for Banks
Systemic Risk
Institutional Risk

• Event Risk Market Risk
• (Devaluation Risk
  (FX., interest
• Large moves in Asset Prices) rate)

Liquidity Risk (Inability to unwind a position
without loss of value)
Credit Risk (Default counterparty Potential loss
due to change in credit quality)
Bank
Reputational Risk (environment)
Legal Risk (contracts are not documented
correctly or cannot be enforced)
Settlement Risk (not receiving funds)
E-business
Operational Risk (loss due to execution error)
45
Focus on Correlations is the Element in Portfolio
Analysis

• Definition Correlated items undergo mutual or
  reciprocal movements
• Why are correlations in default and collateral
  value relevant? ANS. In Bivariate Models of
  Repayment, Expected Loss Product of
  (Probability of Default) and (Loss Given
  Default). When collateral value falls whenever
  probability of default rises, loss exposure
  worsens on both counts.

46
4th Topic Introduction to the Discipline and Art
of Loan Accounting

• GAAP Allows Considerable Leeway for timing
  reported changes in economic value Example of
  deferred interest (i.e., negative amortization)
  on option ARM loans at Golden West
• TRANSPARENCY of Accounting Reports of Bank Income
  and Net Worth is deliberately weak (opacity)
• Book Value (usually par) and Market Value (PDV
  of future cash flows) of Loans can Diverge
  Greatly after a loan is made
• late payments
• changes in credit quality of borrowers
• changes in market risk premia
• changes in level of riskless rate

47
Parable can Illustrate Value to Owners and
Regulators of the Transparency (accuracy plus
meaningfulness) created by Prompt and Accurate
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.
48
Age-Old Conflict Between Regulators of FSFs and
Watchdogs Who Regulate Securities Markets and
Auditing Activity

• 1. FSF regulators adopt rules and enforcement
  systems to assure safe and sound operation of
  firms in their client industry.
• ? These regulators want to entertain
  forward-looking
• industrywide reasons to justify
  unallocated LLR.
• 2. Those whose set auditing standards are
  concerned with how to document occurrences of
  revenues and expenses in an objective and
  reproducible manner.
• ? Prefer to emphasize historical loss
  experience or observable changes in
  activities or skillsets of each individual
  client.

49

• On average, borrowers are slow. Why?
• Expect to extract leniency
• Many plan to move in and out of delinquency
• Exercise In Booking Accruals, Provisions, and
  Charges stresses two kinds of nontransparency
• Banks employ accrual accounting rather than cash
  accounting for yet-to-be-received receipts on
  slow, but performing loans. Nonperforming
  Loans are shifted to a nonaccrual status.
• They are obliged by rules to shift to cash
  accounting only for loans on which payments are
  so far overdue that they must be classified as
  nonperforming (threshold varies across
  countries 90 days in U.S.).

50
In U.S., Lightly Disciplined Judgments Shape
Loan-Loss Reserves (LLR) and Allowances for Loan
and Lease Losses (ALLL)

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

51
Examiner Criticism May Sometimes Force a Loan
into Nonaccrual Status

• A four-way categorization of criticized loans
  is used by examiners special mention,
  substandard, doubtful, loss.

52

• Definitions of Examiner 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 weekly NFL injury
  reports probable, questionable, doubtful, out.
• Loss and out are the most-reliable
  categories.
• 4. Regulators just postponed an effort to
  adopt a new nomenclature.

53
Workout Personnel Use a Different Vocabulary From
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.

54
Chart 1 Aversely Rated Credits Over Two Business
Cycles
55
Cross-Country Variation in the Timing and LLR
Impact of NPLs Status
Source World Bank
56
Exercise in Booking Accruals, Provisions, and
Chargeoffs
1. Eagle Bank lends 200,000 to BC Company on
January 1 at 7 percent simple interest to be paid
quarterly on the unpaid balance. Payments of
50,000 plus quarterly interest are due on March
30, June 30, September 29, and December 30. a.
Calculate the payments due at each payment date
(in thousands).
March 30 June 30 September 29 December 30
57
b. Suppose the first payment is made on time and
no further payments are received until the
following year. Assuming a 180-day threshold for
putting a slow loan into nonaccrual status, what
would the bank initially report as current
revenue on the loan in each quarter of the
current year ( thousands)?
1Q 2Q 3Q 4Q
3.5
2.625 (Credited but not received)
152.625(.07/4)2.671 (Credited but not received)
-5.296 (2Q 3Q Accruals must be reversed at year
end and 2.718 in unpaid interest added to the
outstanding balance
58
1Q 4Q
3.5 - 1 (in provisioned funds on Jan. 1) 2.5
-5.296 1 (in chargeoffs) -4.296
59
3. Suppose on February 1 of year 2 federal bank
examiners forced Eagle Bank to restructure the
loan and write off 10 percent of the principal
including unpaid interest to this date still
due on the loan against its net worth account.
Suppose also that the new contract structure
required the BC Company to make only a single
payment of 155,000 on June 15 of year two. How
would these events affect the banks income in
the first quarter and how would the timely
receipt of the settlement payment affect the
income in the second quarter of year two?
Year 2, 1Q -15.893 end-of-year
principal accrued interest 158.934 Year 2,
2Q 155 158.934 -
15.893 11.959
60
(No Transcript)