Title: Securitization 301
1Securitization 301
- Dynamic Structuring Analysis
RR Consulting
2US Capital Markets, 1970-1980s
market risk
Securitization (á la 101 )
Derivatives
Corporate Finance
basis risk
credit risk
Securitization?
liquidity / credit risk
cash
synthetics
operational risk
3Securitization 101
- Benchmark Pool (an adaptation of the corporate
finance method) - Back-of-the-Envelope (liquidation) Analysis
(securitization) - Credit risk value is a function of CE and
expected losses - Prepayment risk to the extent it reduces CE
- Counterparty risk covers everything else
4US Capital Markets, 1990s
market risk
Securitization (á la 101
or 201)
Corporate Finance
basis risk
Rated, repackaged market risk
credit risk
liquidity / credit risk
cash
Derivatives
synthetics
operational risk
5Securitization 201
- Scenario-Driven Cash Flow Analysis
(securitization) - Credit risk value is a function of CE and loss
volatility prepayment risk embedded in the CF
model - Counterparty risk covers everything else
- Monte Carlo Cash Flow Analysis (securitization)
6US Capital Markets Now
market risk
Securitization (MC simulation)
Corporate Finance
basis risk
Liquidity/credit risk
cash
Derivatives
synthetics
operational risk
7Securitization 301
- Monte Carlo Cash Flow Analysis (securitization)
- Credit risk value is a function of CE and loss
volatility prepayment risk embedded in the CF
model - Servicer risk has operational and credit
dimensions - Liquidity risk was always there but is more
highlighted - Market risk also highlighted for both accounting
portfolio management reasons - Basis risk may be part of the cash flow analysis
- Counterparty risk do ratings really do the job?
- Option-Theoretic Valuation Framework
- Market risk price is the goal. Fair value is a
structural analysis prices are a random walk - Credit risk value is approximated through a
Merton default model for credit portfolios, via
a Gaussian copula - Servicer risk value is approximated through a
Merton default model - Liquidity risk addressed in a market sense
- Counterparty risk not quite on the radar screen.
8The Drivers of Dynamic Analysis
- Drivers of Change
- Economic efficiencies
- Labor market pressures
- Increased regulation
- Market Effects of Change
- Commoditization of Risk
- Competition of ideas
- Market convergence
9Technical Items in this Module
- The non-credit elements in the total analysis of
payment certainty liquidity, basis, market,
operational risk - The expanded set of performance metrics
volatility, correlation duration, convexity - The expanded set of solutions contingent claims
modeling Monte Carlo simulation Gaussian Copula - Competitor paradigms of credit analysis
- The credit derivatives market products,
vocabulary, metrics of credit default modeling
10Synthetic vs. Analytical Approaches
11Measures of Risk, by Domain
12Credit Risk
- Measures currently in use
- (1) Default
- an estimate of the probability that a borrower
will not repay all or a portion of a loan on time
(OTS) - an ISDA credit definition
- an empirical point-estimate taken from static
pool history - a random deviate from a distribution (or
guesstribution)
13Credit Risk (alt)
- (2) Loss
- an estimate of the shortfall on a financial
contractual amount due (originally signified
assets, now also signifies liabilities) after
recoveries are netted from defaults - an input into the IRB risk-weighting model to
produce a capital charge - an output of a Vasicek-type credit risk model
- a point-estimate taken from static pool history
- a statistical point-estimate on a logistic curve
- (3) Reduction of Yield difference between the
sample average yields in a Monte Carlo simulation
and a contractual or target yield.
14Discussion
- Rating agency ratings map all three types of
measure to the alphanumeric rating. They are by
no means interchangeable - They are unlike in their information efficiency
??IRR is fungible, can be compared to other
yields E(L) has more information than defaults
but it can be manipulated by changing the
recovery assumption Default-based analysis
over-states high frequency/low severity events
and understates low frequency/high severity
events. It is the furthest from the cash flow
analysis. - Each produces a different numeric and a different
rating
15Liquidity Risk
- The term specifies very different contexts
- The risk of a companys working capital becoming
insufficient to meet near term financial demands.
(Treasury Management Association of Canada) - The risk associated with transactions made in
illiquid markets. Such markets are characterized
by wide bid/offer spreads, lack of transparency
and large movements in price after a deal of any
size. (Federal Home Loan Bank of Dallas)
16Market Risk
- Risk associated with fluctuations in (asset)
prices (Minnesota Mutual) - The possibility that the price of a security will
change over time (David Gerster) - A random walk, or, equivalently, Geometric
Brownian motion - Most simply written
- where the first term signifies the expected rate
of change with respect to time and the second
term signifies deviations from the first term
that are normally distributed error terms. - Prices in equilibrium are assumed to move as
17Basis Risk
- A risk that the value of the financial
instrument does not move in line with the
underlying exposure. Generally, it refers to an
imperfect hedge where the matched risk-offsetting
positions are not in identical markets (Capital
Market Risk Advisers) - Generally presumed to be less risky than
outright market risk exposurebut data
granularity is important. When the markets stop
moving in tandem, the magnitude of risk is
outside expectation.
18Operational Risk
- According to 644 of International Convergence of
Capital Measurement and Capital Standards, known
as Basel II, operational risk is defined as the
risk of loss resulting from inadequate or failed
internal processes, people and systems, or from
external events. (Wikipedia) - Operational risk may be defined by what it does
not include market risk, credit risk, and
liquidity risk. (CMRA)
19How Well Do Servicer Ratings Benchmark
Operational Risk?
20Technical Items in this Module
- The non-credit elements in the total analysis of
payment certainty liquidity, basis, market,
operational risk - The expanded set of performance metrics
volatility, correlation duration, convexity - The expanded set of solutions contingent claims
modeling Monte Carlo simulation Gaussian Copula - Competitor paradigms of credit analysis
- The credit derivatives market products,
vocabulary, metrics of credit default modeling
21Definitions Volatility
- A measure of the fluctuation in the market price
of the underlying security. Mathematically,
volatility is the annualized standard deviation
of returns. (optiondigest.com) - If the average quarterly asset price volatility
is 25, annualized price volatility is - If the average one-year price volatility is 25,
daily price volatility is -
22Applications - Volatility
- Credit Risk used to contextualize the
microstructure of E(L) variability in structured
securities. Theoreticalnot substantiated by
empirical data in real applications. - Market Risk the exogenous input in a
Black-Scholes model for valuing contingent claims
on market risk exposures. - Basis Risk the exogenous input in a
Black-Scholes model for valuing contingent claims
on basis risk exposures.
23Definitions Correlation
- The word is used in two different senses
-
- If I hold two securities and one defaults, what
is the likelihood that the other will also
default? -
- Strictly speaking, this is not correlation but
conditional probability. It takes on a range of
values 0,1, reflecting only positive
correlation. - The common statistical measure of correlation is
the Pearson correlation coefficient, a number
with a range of -1,1, - This reflects diversification as well as
interdependence. It should not be confused with
causality, however. -
24Critical Applications - Correlation
- Credit Risk used to quantify the
interdependence of risk exposures in credit
portfolios and the impact on cash flow certainty
CDOs, credit basket trades. -
-
25Definitions Modified Duration
- Measures the sensitivity of bond prices to
changes in rate environment - As a first derivative of price with respect to
yield, it gives a rough indication of how much
price will rise (fall) for a small unit change - Begin with price
- Take the first derivative with respect to yields
- To normalize the output, divide the result by P.
- Although duration is approximately correct for
small changes, due to the non-linear relationship
between price and yield, it is not very accurate
for larger changes.
26Convexity
- Measures the sensitivity of price to changes in
rate environment - As the second derivative of price with respect to
yield, it shows the magnitude of sensitivity of
the change in price to the change in yield
27Modified Duration/Early Repayment
- When the call date is certain, Effective Duration
provides a linear adjustment to Modified Duration
that averages the asymmetrical price impact of
rising or falling rates - Effective Duration is not a good approximation
when the call date is uncertain. Prepayment
ability by the borrower (a call option) turns
cash flows that are fixed into a cash flow that
is itself a function of interest rates - ,
for a vector of cash flows, Ct(r). - The algebra of duration and convexity become more
complex with cash-flow dependency. The formula
for modified duration becomes
28Definition Gaussian Copula
29Definitions Recoveries
- The definition of recoveries is trivial
- 1-lgd (loss-given-default)
- The problem is one of data quality, or perhaps it
should be called data scrupulousness.
30Technical Items in this Module
- The non-credit elements in the total analysis of
payment certainty liquidity, basis, market,
operational risk - The expanded set of performance metrics
volatility, correlation duration, convexity - The expanded set of solutions contingent claims
modeling Monte Carlo simulation Gaussian Copula - Competitor paradigms of credit analysis
- The credit derivatives market products,
vocabulary, metrics of credit default modeling
31Impact of Prepayments on Value
- Some bonds, like MBS, have a tendency to prepay
in some interest rate environments. - The tapering off of interest (and principal)
cash flows only impairs their creditworthiness to
the extent it affects XS, but it has adverse
consequences for reinvestment or trading
activity. - I need a way to price a callable bond that
reflects the impact of prepayment risk.
32Price Sensitivity to Yield Change
How actual prices change
Price estimates
33Negative Convexity
34Interest vs. PPMT Cash Flows
35PACs and TACs
36Problem Valuing Rights of Ownership
- Rights of ownership (contingent claims) are not
the same as outright ownership. - Intuitively, the value of contingent claims is a
random variable that should rise when price
volatility increases and fall when
time-to-expiration amortizes. - I need a consistent method for pricing an
ownership right in the pre-ownership phase.
37Contingent Claims Valuation
- Single-most influential valuation concept in
modern finance. Sprenkel published the first
general approach in the 1960s, which did not rely
on risk neutrality. - Fischer Black and Myron Scholes published their
arguments for a closed form solution to the
problem of valuing contingent assets using the
heat diffusion equation. - Black-Scholes facilitates pricing uncertain cash
flows by transforming them into risk-neutral
equivalents through a process of continuous
re-hedging. The approach rests on certain
simplifying assumptions (next page, pls) - The fundamental insight underlying risk-neutral
pricing is the put-call parity condition, where S
asset price, P is the price of a put, C, is the
price of a call, and Ee-r(T-t) is the price of a
risk-free loan
38Black-Scholes Assumptions
- The risk-free rate, dividends and asset
volatility can be known over the life of the
exposure - The hedge costs are de minimus
- The asset trades continuously (short or long
positions are both possible) and it is divisible - The marketplace responds instantaneously to new
information (efficient market hypothesis) to form
a rational price deviations from the equilibrium
price are random
39Black-Scholes Modeling
- Critical Applications
- Market Risk the consensus fair value metric for
pricing futures, options and structured
derivative trades (swaps, collars, caps) in
organized and OTC exchanges. Aspects of the
underlying argument are actively used in
establishing and maintaining market risk-neutral
positions. Continuous trading is an operational
requirement. A central clearing and settlement
function is highly desirable from the standpoint
of credit risk elimination. - Credit Risk used in structural (Merton default)
models to establish an implied default risk of a
corporation. Fundamental insight is the
characterization of residual value as a call on
the company assets and the insolvency boundary as
a put on the company assets back to the lender. - Other applications (1) Borrowers who refinance
their mortgage loans before maturity are said to
be long a call option with respect to the loan,
which they can exercise if interest rates go down
(price goes up/call option is in the money). An
implied price for these securities can be worked
back to from a back-of-the-envelope calculation
on the value of the borrowers call. (2) Sellers
of default protection (CDS) are said to go long
the probability of corporate default on the
reference obligation of the firm and buyers of
default protection are said to be short the
probability of corporate default on the same. -
40Problem Process Modeling without a Closed Form
Solution
- Black-Scholes uses the heat-transfer equation to
describe the dissipation of errors. - What if there is no known analogue from physics
or engineering that I can use to model the
financial process? - I need a way to use what I know about the past
to condition my expectations on the future.
41Monte Carlo Simulation
- Multiple sampling from a real portfolio is
impossible. Hence the usefulness of sampling from
a theoretical universe. - If we could draw a suitably large number of
samples from the theoretical universe reflecting
the underwriting criteria of the loans in
question, we could perform parametric statistical
analysis on the samples, and use the results to
structure a transaction. - One method of simulation, the Inverse
Distribution Function Method (IDFM), can be
performed in spreadsheets using Excel functions,
or in Visual Basic for Excel. Assume an initial
cumulative loss distribution - Flipping coins on the y-axis using a random
number generator to find the cumulative frequency
of occurrence (the left-hand term in the equation
below) a corresponding loss is drawn (the
right-hand term). - Flipping many such coins to draw many will
eventually populate the original distribution ,
by the law of large numbers.
42Inverse Distribution Function Method
91 of the probability mass
43Monte Carlo Simulation
- Critical Applications
- Credit Risk used by some rating agencies to
rate asset-backed or mortgage-backed securities
or CDOs, to rate transactions. MC simulation
allows the impact of the microstructure of risk
on the payment certainty of structured securities
to be measured systematically with
probability-weighted scenarios. - Market Risk used in Option Adjusted Spread
(OAS) calculations. The difference between the
theoretical price of the MBS and what MBS
investors are willing pay can be evaluated in
cash flow terms. This is the bonds
option-adjusted spread or OAS. -
44OAS Modeling
- Simulates sequences of interest rate paths to
produce a set of cash flows and an average life,
for each security in the structure. Three main
building blocks - Interest rate model, used to generate a set of
rate paths that are inputs to the next block.
Rate paths need to be as long as the longest
maturity of any loan in the MBS pool. - Prepayment rate model using rate paths produced
in Step 1 to produce cash flows. Prepayment
models are conditional in the sense that they
attempt to predict prepayment rates given
interest rates and other driver variables,
instead of trying to predict these independent
variables themselves. - Cash flow model able to combine the prepayment
rates from Step 2 and compute the OAS spreads by
reference to market bond prices and the yield
curve. Schematically, the OAS methodology can be
visualized in the figure below.
Yield Curve (current coupon)
Rate Volatility Assumptions
Interest Rate Model (MC Scenarios)
Prepayment Model (PPMT Vector)
Cash Flow Model (PI)
OAS, Duration, Convexity
45Problem Sizing the Cash Flow Impact of
Correlation on Credit Portfolios
- I know how to calculate correlation
coefficients, but what kind of data should I use?
- I need a way to systematically stress a
portfolio of exposures to reflect the impact of
sectoral inter- and intra-dependence.
46Technical Content
- Non-credit elements in the total analysis of
payment certainty basis, market, operational
risk - Solutions and the expanded set of performance
metrics and methods volatility, correlation
duration, convexity contingent claims modeling
Monte Carlo simulation Gaussian Copula. - Competitor paradigms for credit analysis
- The credit derivatives market products,
vocabulary, metrics of credit default modeling
for buying selling pure default risk.
47Alternative Credit Paradigms
- Structural (Merton Default)
- Intensity (Hazard Rate) Modeling
48Technical Content
- The non-credit elements in the total analysis of
payment certainty basis, market, operational
risk. - Solutions and the expanded set of performance
metrics and methods volatility, correlation
duration, convexity contingent claims modeling
Monte Carlo simulation Gaussian Copula - Competitor paradigms.of credit analysis
- The credit derivatives market products,
vocabulary, metrics of credit default modeling
for buying selling pure default risk
49Credit Synthetics
- Are not securitizations under Reg AB
- Are said to facilitate separation of risk
management, funding roles - International Swaps Derivatives Association
(ISDA) provides transaction governance structure
contracts, confirmations, legal opinions, key
definitions, day count conventions, settlement
procedures - Basic valuation framework is cash-and-carry trade
- More sophisticated modeling alternatives
structural, intensity models
50Product Typology
51New Risks Come into Focus
- Swap replacement risk
- Swap settlement risk
- Physical delivery risk
- Cash-Synthetic basis risk
52Where do we go from here?
market risk
Securitization (MC simulation)
Corporate Finance
basis risk
Liquidity/credit risk
cash
Derivatives
synthetics
operational risk
53Hypothesis Inversion of the pre-1990 Market
Structure
market risk
Institutions
Securitization (MC simulation)
basis risk
Liquidity/credit risk
cash
Innovation, policy risk
Derivatives
synthetics
operational risk