GRM - Day 1.2

Regulation, Basel II, and
Solvency II
Chapter 11
Risk Management and Financial Institutions 2e, Chapter 11, Copyright © John C. Hull 2009
History of Bank Regulation
l Pre-1988
l 1988: BIS Accord (Basel I)
l 1996: Amendment to BIS Accord
l 1999: Basel II first proposed
l 2003: the New Basel Capital Accord
l 2004: tentative Agreement
l 2006: Basel II
l 2009: Basel III proposal
The Model used by Regulators
(Figure 11.1, page 235)
Expected X% Worst
Loss Case Loss
Required
Capital
Loss over time

horizon
0 1 2 3 4
Pre-1988
l Banks were regulated using balance sheet measures
such as the ratio of capital to assets
l Definitions and required ratios varied from country to
country
l Enforcement of regulations varied from country to
country
l Bank leverage increased in 1980s
l Off-balance sheet derivatives trading increased
l LDC debt was a major problem
l Basel Committee on Bank Supervision set up
1988: BIS Accord (page 223)
l The assets:capital ratio must be less than
20. Assets includes off-balance sheet
items that are direct credit substitutes such
as letters of credit and guarantees
l Cooke Ratio: Capital must be 8% of risk
weighted amount. At least 50% of capital
must be Tier 1.
Example of Simple Bank Balance
Sheet: End 2012 (Table 2.2, page 24)
Assets Liabilities
Cash 5 Deposits 90
Marketable Securities 10 Subord L.T. Debt 5
Loans 80 Equity Capital 5
Fixed Assets 5
Total 100 Total 100
Types of Capital (page 225-226)
l Tier 1 Capital: common equity, non-

cumulative perpetual preferred shares
l Tier 2 Capital: cumulative preferred
stock, certain types of 99-year debentures,
subordinated debt with an original life of
more than 5 years
Risk-Weighted Capital
l A risk weight is applied to each on-balance- sheet
asset according to its risk (e.g. 0% to cash and govt
bonds; 20% to claims on OECD banks; 50% to
residential mortgages; 100% to corporate loans,
corporate bonds, etc.)
l For each off-balance-sheet item we first calculate a
credit equivalent amount and then apply a risk weight
l Risk weighted amount (RWA) consists of
l sum of risk weight times asset amount for on-balance sheet
items
l Sum of risk weight times credit equivalent amount for off-
balance sheet items
Credit Equivalent Amount
l The credit equivalent amount is calculated
as the current replacement cost (if
positive) plus an add on factor
l The add on amount varies from instrument
to instrument (e.g. 0.5% for a 1-5 year
swap; 5.0% for a 1-5 year foreign currency
swap)
Add-on Factors (% of Principal)
Table 11.2, page 225
Remaining Interest Exch Rate Equity Precious Other

Maturity (yrs) rate and Gold Metals Commodities
except gold
<1 0.0 1.0 6.0 7.0 10.0
1 to 5 0.5 5.0 8.0 7.0 12.0
>5 1.5 7.5 10.0 6.0 15.0
Example: A $100 million swap with 3 years to maturity worth $5 million

would have a credit equivalent amount of $5.5 million
The Math
N M
RWA = å wi Li + å w C j *
j
i =1 j =1
On-balance sheet Off-balance sheet items:

items: principal credit equivalent
times risk weight amount times risk
weight
For a derivative Cj = max(Vj,0) + ajLj where Vj is

value, Lj is principal and aj is add-on factor
G-30 Policy Recommendations
(page 226-227)
l Influential publication from derivatives

dealers, end users, academics,
accountants, and lawyers
l 20 recommendations published in 1993
Netting (page 227-228)
l Netting refers to a clause in derivatives
contracts that states that if a company
defaults on one contract it must default on
all contracts
l In 1995 the 1988 accord was modified to
allow banks to reduce their credit
equivalent totals when bilateral netting
agreements were in place
Netting Calculations
l Without netting exposure is
N
å max(V ,0)
j =1
j
l With netting exposure is

æ N ö
maxçç åV j ,0 ÷÷
è j =1 ø
l NRR =
Exposure with Netting
Exposure without Netting
Netting Calculations continued
l Credit equivalent amount modified from
N
å[max(V ,0) + a L ]
j =1
j j j
l To
N N
max(åV j ,0) + å a j L j (0.4 + 0.6 ´ NRR )
j =1 j =1
1996 Amendment (page 229-231)
l Implemented in 1998
l Requires banks to measure and hold
capital for market risk for all instruments in
the trading book including those off
balance sheet (This is in addition to the
BIS Accord credit risk capital)
l It also outlined a Standardized approach,
internal model based approach (IMA)
1996 Amendment (page 229-231)
l The VaR Calculations reflects movements
in interest rates, exchange rates, stock
index and commodity prices but not
company specific risks. Movement in stock
prices or credit spread is captured by
Specific Risk Charges (SRC).
The Market Risk Capital
l The capital requirement is k ´ VaR + SRC
l Max(VaRt-1, mc X VaRavg )+ SRC
l Where k is a multiplicative factor chosen
by regulators (at least 3), VaR is the 99%
10-day value at risk, and SRC is the
specific risk charge for idiosyncratic risk
related to specific companies
l If only one-day VaR is measured then use mc X
3.16
Basel II
l Implemented in 2007
l Three pillars
l New minimum capital requirements for credit
and operational risk
l Supervisory review: more thorough and
uniform
l Market discipline: more disclosure
New Capital Requirements
l Risk weights based on either external
credit rating (standardized approach) or a
bank’s own internal credit ratings (IRB
approach)
l Recognition of credit risk mitigants
l Separate capital charge for operational
risk
USA vs European Implementation
l In US Basel II applies only to large
international banks
l Small regional banks required to
implement “Basel 1A’’ (similar to Basel I),
rather than Basel II
l European Union requires Basel II to be
implemented by securities companies as
well as all banks
Standardized Approach, Table 11.3, page 233
Bank and corporations treated similarly (unlike Basel I)
Rating AAA A+ to BBB+ BB+ to B+ to Below Unrate

to A- to BB- B- B- d
AA- BBB-
Country 0% 20% 50% 100% 100% 150% 100%
Banks 20% 50% 50% 100% 100% 150% 50%
Corporates 20% 50% 100% 100% 150% 150% 100%
IRB Approach for corporate, banks and sovereign
exposures
l Basel II provides a formula for translating PD

(probability of default), LGD (loss given default),
EAD (exposure at default), and M (effective
maturity) into a risk weight
l Under the Advanced IRB approach banks
estimate PD, LGD, EAD, and M
l Under the Foundation IRB approach banks
estimate only PD and the Basel II guidelines
determine the other variables for the formula
Key Model (Gaussian Copula)
• The 99.9% worst case default rate is
é N -1 ( PD) + r ´ N -1 (0.999) ù
WCDR = N ê ú
êë 1- r úû
Numerical Results for WCDR
Table 11.4, page 236
PD=0.1% PD=0.5% PD=1% PD=1.5% PD=2%

r=0.0 0.1% 0.5% 1.0% 1.5% 2.0%
r=0.2 2.8% 9.1% 14.6% 18.9% 22.6%
r=0.4 7.1% 21.1% 31.6% 39.0% 44.9%
r=0.6 13.5% 38.7% 54.2% 63.8% 70.5%
r=0.8 23.3% 66.3% 83.6% 90.8% 94.4%
Dependence of r on PD
l For corporate, sovereign and bank
exposure
1 - e -50´PD é 1 - e -50´PD ù -50´ PD
r = 0.12 ´ -50
+ 0.24 ´ ê1 - -50 ú = 0.12[1 + e ]
1- e ë 1- e û
PD 0.1% 0.5% 1.0% 1.5% 2.0%
WCDR 3.4% 9.8% 14.0% 16.9% 19.0%
(For small firms r is reduced)

Capital Requirements
1 + ( M - 2.5) ´ b
Capital = EAD ´ LGD ´ (WCDR - PD ) ´
1 - 1. 5 ´ b
where M is the effective maturity and
b = [0.11852 - 0.05478 ´ ln( PD )]2
The risk - weighted assets are 12.5 times the Capital

so that Capital = 8% of RWA
Retail Exposures
Capital = EAD ´ LGD ´ (WCDR - PD )

For residential mortgages r = 0.15
For revolving retail exposures r = 0.04
For other retail exposures
1 - e -35´ PD é 1 - e -35´PD ù
r = 0.03 ´ -35
+ 0.16 ´ ê1 - -35 ú
1- e ë 1 - e û
= 0.03 + 0.13e -35´PD
There is no distinction between Foundation and Advanced IRB approaches.
Banks estimate PD, LGD, and EAD in both cases
Credit Risk Mitigants
l Credit risk mitigants (CRMs) include
collateral, guarantees, netting, the use of
credit derivatives, etc
l The benefits of CRMs increase as a bank
moves from the standardized approach to
the foundation IRB approach to the
advanced IRB approach
Adjustments for Collateral
l Two approaches
l Simple approach: risk weight of counterparty
replaced by risk weight of collateral
l Comprehensive approach: exposure adjusted
upwards to allow to possible increases; value
of collateral adjusted downward to allow for
possible decreases; new exposure equals
excess of adjusted exposure over adjusted
collateral; counterparty risk weight applied to
the new exposure
Guarantees
l Traditionally the Basel Committee has used the credit

substitution approach (where the credit rating of the
guarantor is substituted for that of the borrower)
l However this overstates the credit risk because both
the guarantor and the borrower must default for money
to be lost
l Alternative proposed by Basel Committee: capital
equals the capital required without the guarantee
multiplied by 0.15+160×PDg where PDg is probability of
default of guarantor
Operational Risk Capital
l Basic Indicator Approach: 15% of gross
income
l Standardized Approach: different
multiplicative factor for gross income
arising from each business line
l Internal Measurement Approach: assess
99.9% worst case loss over one year.
Supervisory Review Changes
l Similar amount of thoroughness in

different countries
l Local regulators can adjust parameters to
suit local conditions
l Importance of early intervention stressed
Market Discipline
l Banks will be required to disclose
l Scope and application of Basel framework
l Nature of capital held
l Regulatory capital requirements
l Nature of institution’s risk exposures
Possible Revisions to Basel II
l Incremental risk charge (credit items in
trading book treated in the same way as if
they were in banking book)
l Stressed VaR (takes account of
movements in market variables during a
one-year period of significant losses in
calculating market risk capital)
l Movement away from self-regulation
Solvency II
l Similar three pillars to Basel II

l Pillar I specifies the minimum capital requirement
(MCR) and solvency capital requirement (SCR)
l If capital falls below SCR the insurance company
must submit a plan for bringing it back up to SCR.
l If capital; drops below MCR supervisors are likely to
prevent the insurance company from taking new
business
Solvency II continued
l Internal models vs standardized approach
l One year 99.5% confidence for internal models
l Capital charge for investment risk, underwriting risk, and
operational risk
l Three types of capital
Volatility
Chapter 9
Risk Management and Financial Institutions 2e, Chapter 9, Copyright © John C. Hull 2009 38
Definition of Volatility
l Suppose that Si is the value of a variable on
day i. The volatility per day is the standard
deviation of ln(Si /Si-1)
l Normally days when markets are closed are
ignored in volatility calculations (see Business
Snapshot 9.1, page 177)
l The volatility per year is 252 times the daily
volatility
l Variance rate is the square of volatility
l Realized vs. Implied Vs. Conditional Volatility
Implied Volatilities
l Of the variables needed to price an option
the one that cannot be observed directly is
volatility
l We can therefore imply volatilities from
market prices and vice versa
VIX Index: A Measure of the Implied
Volatility of the S&P 500 (Figure 9.1, page
178)
Global Risk Management
Are Daily Changes in Exchange Rates
Normally Distributed? Table 9.2, page 181
Real World (%) Normal Model (%)

>1 SD 25.04 31.73
>2SD 5.27 4.55
>3SD 1.34 0.27
>4SD 0.29 0.01
>5SD 0.08 0.00
>6SD 0.03 0.00
Heavy Tails
l Daily exchange rate changes are not normally
distributed
l The distribution has heavier tails than the normal
distribution
l It is more peaked than the normal distribution
l This means that small changes and large
changes are more likely than the normal
distribution would suggest
l Many market variables have this property,
known as excess kurtosis
Normal and Heavy-Tailed
Distribution
Alternatives to Normal Distributions:
The Power Law (See page 182)
Prob(v > x) = Kx-a
This seems to fit the behavior of the

returns on many market variables better
than the normal distribution
Log-Log Test for Exchange Rate
Data
Standard Approach to Estimating
Volatility
l Define sn as the volatility per day between
day n-1 and day n, as estimated at end of day
n-1
l Define Si as the value of market variable at
end of day i
l Define ui= ln(Si/Si-1)
m
1
s n2 = å
m - 1 i =1
( un -i - u ) 2
1 m
u = å un -i
m i =1
Simplifications Usually Made in
Risk Management
l Define ui as (Si−Si-1)/Si-1
l Assume that the mean value of ui is zero
l Replace m-1 by m
This gives
1 m 2
s = åi =1 un -i
2
n
m
Weighting Scheme
Instead of assigning equal weights to the
observations we can set
s = å i =1 a i u
2 m 2
n n -i
where
m
åa
i =1
i =1
ARCH(m) Model
In an ARCH(m) model we also assign
some weight to the long-run variance rate,
VL:
s = gVL + åi =1 a i u n2-i
2 m
n
where
m
g + å ai = 1
i =1
EWMA Model (page 186)
l In an exponentially weighted moving

average model, the weights assigned to
the u2 decline exponentially as we move
back through time
l This leads to
s = ls
2
n
2
n -1 + (1 - l)u 2
n -1
Attractions of EWMA
l Relatively little data needs to be stored
l We need only remember the current
estimate of the variance rate and the most
recent observation on the market variable
l Tracks volatility changes
l RiskMetrics uses l = 0.94 for daily
volatility forecasting
GARCH (1,1), page 188
In GARCH (1,1) we assign some weight to
the long-run average variance rate
s = gVL + au
2
n
2
n -1 + bs 2
n -1
Since weights must sum to 1

g + a + b =1
GARCH (1,1) continued
Setting w = gVL the GARCH (1,1) model is
s = w + au
2
n
2
n -1 + bs 2
n -1
and
w
VL =
1- a - b
Example
l Suppose
s = 0.000002 + 013
2
n . u 2
n -1 + 0.86s 2
n -1
l The long-run variance rate is 0.0002 so

that the long-run volatility per day is 1.4%
Example continued
l Suppose that the current estimate of the
volatility is 1.6% per day and the most
recent percentage change in the market
variable is 1%.
l The new variance rate is
0.000002 + 013
. ´ 0.0001 + 0.86 ´ 0.000256 = 0.00023336
The new volatility is 1.53% per day
GARCH (p,q)
p q
s = w + å ai u
2
n
2
n -i + å b js 2
n- j
i =1 j =1
Other Models
l Many other GARCH models have been
proposed
l For example, we can design a GARCH
models so that the weight given to ui2
depends on whether ui is positive or
negative
Variance Targeting
l One way of implementing GARCH(1,1)
that increases stability is by using variance
targeting
l We set the long-run average volatility
equal to the sample variance
l Only two other parameters then have to be
estimated
Maximum Likelihood Methods
l In maximum likelihood methods we

choose parameters that maximize the
likelihood of the observations occurring
Example 1 (page 190)
l We observe that a certain event happens
one time in ten trials. What is our estimate
of the proportion of the time, p, that it
happens?
l The probability of the outcome is
p(1 - p) 9
l We maximize this to obtain a maximum

likelihood estimate: p=0.1
Example 2 (page 190-191)
Estimate the variance of observations from
a normal distribution with mean zero
é 1 n
æ - ui2 ö ù
Maximize: Õ ê expç ÷ú
i =1 ë 2 pv è 2v ø û
n
é ui2 ù
or: å ê- ln(v ) - v ú
i =1 ë û
1 n 2
This gives: v = å ui
n i =1
Application to GARCH (1,1)
We choose parameters that maximize
n
é u ù 2
å ê- ln(vi ) - ú
i =1 ë vi û
i
Calculations for Yen Exchange
Rate Data (Table 9.4, page 192)
Day Si ui vi =si2 -ln vi-ui2/vi

1 0.007728
2 0.007779 0.006599
3 0.007746 -0.004242 0.00004355 9.6283
4 0.007816 0.009037 0.00004198 8.1329
5 0.007837 0.002687 0.00004455 9.8568
….
2423 0.008495 0.000144 0.00008417 9.3824
22063.5833
Daily Volatility of Yen: 1988-1997
Forecasting Future Volatility
(Equation 9.14, page 195)
A few lines of algebra shows that

E[s ] = VL + (a + b) (s - VL )
2
n +t
t 2
n
To estimate the volatility for an option

lasting T days we must integrate this from
0 to T
Forecasting Future Volatility cont
The volatility per year for an option

lasting T days is
ì 1- e - aT
ü
s(T ) = 252íVL + [V (0) - VL ]ý
î aT þ
where
1
a = ln
a +b
Volatility Term Structures
l The GARCH (1,1) model allows us to

predict volatility term structures changes
l When s(0) changes by Ds(0), GARCH
(1,1) predicts that s(T) changes by
- aT
1- e s(0)
Ds(0)
aT s(T )
Risk Management
and Financial Returns
Elements of Financial Risk Management
Chapter 1
Peter F. Christoffersen
Elements of Financial Risk

Management Second Edition ©
2012 by Peter Christoffersen
Overview
lBecome familiar with the range of risks facing
corporations, and how to measure and manage these
risks
l Become familiar with the salient features of
speculative asset returns
l Define risks – Market, Credit, Operational, Liquidity
l Certain facts about asset returns
l Define VaR
l Understand the current academic and practitioner
literature

Why should firms manage risk?
l Classic portfolio theory: Investors can eliminate firm-
specific risk by diversifying holdings to include many
different assets
l Investors should hold a combination of the risk-free
asset and the market portfolio.
l Firms should not waste resources on risk
management, as investors do not care about firm-
specific risk.
l Modigliani-Miller: The value of a firm is independent
of its risk structure.
l Firms should simply maximize expected profits
regardless of the risk entailed.
l Bankruptcy: The real costs of company
reorganization or shut-down will reduce the current
valuation of the firm. Risk management can increase
the value of a firm by reducing the probability of
default.
l Taxes: Risk management can help reduce taxes by
reducing the volatility of earnings.

l Capital structure and the cost of capital: a major
source of corporate default is the inability to service
debt. Proper risk management may allow the firm to
expand more aggressively through debt financing.
l Employee Compensation: due to their implicit
investment in firm-specific human capital, key
employees often have a large and unhedged exposure
to the risk of the firm they work for.

Evidence on RM practices
l Large firms tend to manage risk more actively than
small firms, which is perhaps surprising as small
firms are generally viewed to be more risky.
l However smaller firms may have limited access to

derivatives markets and furthermore lack staff with
risk management skills.

Does RM improve firm
performance?
l The overall answer to this question appears to be YES.
l Analysis of the risk management practices in the gold
mining industry found that share prices were less
sensitive to gold price movements after risk
management.
l Similarly, in the natural gas industry, better risk
management has been found to result in less variable
stock prices.
l A study also found that RM in a wide group of firms
led to a reduced exposure to interest rate and exchange
rate movements.
A brief taxonomy of risks
l Market Risk: the risk to a financial portfolio from
movements in market prices such as equity prices,
foreign exchange rates, interest rates and commodity
prices.
l In financial sector firms market risk should be
managed. e.g. option trading desk.
l In nonfinancial firms market risk should perhaps be
eliminated.

l Liquidity risk: The particular risk from conducting
transactions in markets with low liquidity as
evidenced in low trading volume, and large bid-ask
spreads.
l Under such conditions, the attempt to sell assets may
push prices lower and assets may have to be sold at
prices below their fundamental values or within a
time frame longer than expected.
l Traditionally liquidity risk was given scant attention
in RM, but the events in the fall 1998 sharply
increased the attention devoted to liquidity risk.
l Operational risk: the risk of loss due to physical
catastrophe, technical failure and human error in the
operation of a firm, including fraud, failure of
management and process errors.
l Operational risk-“op risk”-should be mitigated and
ideally eliminated in any firm as the exposure to it
offers very little return (the short-term cost savings of
being careless for example).

l Credit risk: the risk that a counter-party may become
less likely to fulfill its obligations in part or in full on
the agreed upon date.
l Thus credit risk consists not only of the risk that a
counterparty completely defaults on its obligations,
but also that it only pays in part and/or after the
agreed upon date.
l The nature of commercial banks has traditionally
been to take on large amounts of credit risk through
their loan portfolios.
l Today, banks spend much effort to carefully manage
their credit risk exposure.
l Nonbank financials as well as nonfinancial
corporations might instead want to completely
eliminate credit risk as it is not a part of their core
business.
l However, many kinds of credit risks are not readily
hedged in financial markets and corporations are
often forced to take on credit risk exposure which
they would rather be without.
l Business risk: the risk that changes in variables of a
business plan will destroy that plan’s viability,
including quantifiable risks such as business cycle
and demand equation risk, and non-quantifiable risks
such as changes in competitive behavior or
technology.
l Business risk is sometimes simply defined as the
types of risks which are integral part of the core
business of the firm and which should therefore
simply be taken on.
Asset returns definitions
l The daily simple rate of return from the closing prices
of the asset:
• The daily continuously compounded or log return on

an asset is

• The two returns are fairly similar
• The approximation holds because ln(x) ≈ x−1 when x

is close to 1
• Let Ni be the number of units held in asset i and let

VPF;t be the value of the portfolio on day t so that

l Then the portfolio rate of return is
• where w = NiSi,t/VPF,t is the portfolio weight in asset i

i
• Most assets have a lower bound of zero on the price

• Log returns are more convenient for preserving this
lower bound in the risk model because an arbitrarily
large negative log return tomorrow will still imply a
positive price at the end of tomorrow.

l Tomorrow’s price when using log returns is
St+1 = exp(Rt+1)St
l where exp(•) denotes the exponential function
l If instead we use the rate of return definition then
tomorrow’s closing price is
St+1 = (1+rt+1)St
l Here St+1 could go negative unless the assumed
distribution of tomorrow’s return, rt+1, is bounded
below by −1

l With log return definition, we can easily calculate the
compounded return at the K−day horizon as the sum
of the daily returns:

Stylized facts of asset returns
l We can consider the following list of so-called
stylized facts which apply to most stochastic returns.
l Each of these facts will be discussed in detail in the
first part of the book.
l We will use daily returns on the S&P500 from
01/01/2001 to 12/31/2010 to illustrate each of the
features.

Stylized fact 1
l Daily returns have very little autocorrelation. We can
write
• Returns are almost impossible to predict from their

own past.
• Fig 1.1 shows the correlation of daily S&P500 returns
with returns lagged from one to 100 days.
• We will take this as evidence that the conditional
mean is roughly constant.

Figure 1.1
Autocorrelation of Daily S&P 500 Returns
Jan 1, 2001 - Dec 31, 2010
0,15
Autocorrelation of Daily Returns
0,10
0,05
0,00
-0,05
-0,10
-0,15
0 10 20 30 40 50 60 70 80 90 100
Lag Order
Stylized fact 2
l The unconditional distribution of daily returns have
fatter tail than the normal distribution.
l Fig.1.2 shows a histogram of the daily S&P500 return
data with the normal distribution imposed.
l Notice how the histogram has longer and fatter tails,
in particular in the left side, and how it is more
peaked around zero than the normal distribution.
l Fatter tails mean a higher probability of large losses
than the normal distribution would suggest.
Figure 1.2
Histogram of Daily S&P 500 Returns and the Normal
Distribution
Jan 1, 2001 - Dec 31, 2010
0,3000
0,2500
Normal Frequency
Probability Distribution
0,2000
0,1500
0,1000
0,0500
0,0000
-0,07 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0,00 0,01 0,02 0,03 0,04 0,05 0,06 0,07

Daily Return
Stylized fact 3
l The stock market exhibits occasional, very large
drops but not equally large up-moves.
l Consequently the return distribution is asymmetric or
negatively skewed. This is clear from Figure 1.2 as
well.
l Other markets such as that for foreign exchange tend
to show less evidence of skewness.

Stylized fact 4
l The standard deviation of returns completely
dominates the mean of returns at short horizons such as
daily.
l It is typically not possible to statistically reject a zero
mean return.
l Our S&P 500 data have a daily mean of 0.0056% and a
daily standard deviation of 1.3771%.

Stylized fact 5
l Variance measured for example by squared returns,
displays positive correlation with its own past.
l This is most evident at short horizons such as daily or
weekly.
l Fig 1.3 shows the autocorrelation in squared returns
for the S&P500 data, that is
• Models which can capture this variance dependence

will be presented in Chapters 4 & 5
Figure 1.3
Autocorrelation of Squared Daily S&P 500 Returns
Jan 1, 2010 - Dec 31, 2010
0,45
Autocorrelation of Squared Returns
0,40
0,35
0,30
0,25
0,20
0,15
0,10
0,05
0,00
0 10 20 30 40 50 60 70 80 90 100
Lag Order
Stylized fact 6
l Equity and equity indices display negative correlation
between variance and returns.
l This often termed the leveraged effect, arising from
the fact that a drop in stock price will increase the
leverage of the firm as long as debt stays constant.
l This increase in leverage might explain the increase
variance associated with the price drop. We will
model the leverage effect in Chapters 4 and 5.

Stylized fact 7
l Correlation between assets appears to be time varying.
l Importantly, the correlation between assets appear to
increase in highly volatile down-markets and extremely
so during market crashes.
l We will model this important phenomenon in Chapter 7

Stylized fact 8
l Even after standardizing returns by a time-varying
volatility measure, they still have fatter than normal
tails.
l We will refer to this as evidence of conditional non-
normality.
l It will be modeled in Chapters 6 and 9.

Stylized fact 9
l As the return-horizon increases, the unconditional
return distribution changes and looks increasingly
like the normal distribution.
l Issues related to risk management across horizons
will be discussed in Chapter 8.

A generic model of asset returns
l Based on the above of stylized facts our model of
individual asset returns will take the generic form
• The conditional mean return is thus mt+1 and the

conditional variance
• The random variable zt+1 is an innovation term,

which we assume is identically and independently
distributed (i.i.d.) as D(0,1).
JP Morgan’s RiskMetrics model for dynamic volatility
• The volatility for tomorrow, time t+1, is computed at the

end of today, time t, using the following simple updating
rule:
• On the first day of the sample, t = 0, the volatility can

be set to the sample variance of the historical data
available.

From asset returns to portfolio
returns
• The value of a portfolio with n assets at time t is the
weighted average of the asset prices using the
current holdings of each asset as weights:
• The return on the portfolio between day t+1 and day t

is then defined as when
using arithmetic returns
• When using log returns return on the portfolio is:

Introducting the VaR risk
measure
l Value-at-Risk - What loss is such that it will only be
exceeded p·100% of the time in the next K trading
days?
l VaR is often defined in dollars, denoted by $VaR
l $VaR loss is implicitly defined from the probability

of getting an even larger loss as in

Introducing the VaR risk measure
l Note by definition that (1−p)100% of the time, the
$Loss will be smaller than the VaR.
l Also note that for this course we will use VaR based
on log returns defined as

Introducing the VaR risk
measure
l Now we are (1−p)100% confident that we will get a
return better than −VaR.
l It is much easier to gauge the magnitude of VaR
when it is written in return terms
l Knowing that the $VaR of a portfolio is $500,000
does not mean much unless we know the value of the
portfolio
l The two VaRs are related as follows:

l
measure
Suppose our portfolio consists of just one security
l For example an S&P 500 index fund
l Now we can use the Risk-Metrics model to provide the VaR
for the portfolio.
l Let VaRPt+1 denote the p .100% VaR for the 1-day ahead
return, and assume that returns are normally distributed with
zero mean and standard deviation sPF,t+1. Then:

measure
l F(z) calculates the probability of being below the
number z
l F-1P= F-1(P) instead calculates the number such that
p.100% of the probability mass is below F-1P
l Taking F-1(*) on both sides of the preceding equation
yields the VaR as

l If we let p = 0.01 then we get F-1P= F-10.01= » -2.33
l If we assume the standard deviation forecast, sPF,t+1

for tomorrow’s return is 2.5% then:

measure
l -1 F P is always negative for p < 0.5
l The negative sign in front of the VaR formula is

needed because VaR is defined as a positive number
l Here VaR is interpreted such that there is a 1%

chance of losing more than 5.825% of the portfolio
value today.

l If the value of the portfolio today is $2 million, the

$VaR would simply be
l For the next figure, note that we assume K = 1 and

p = 0.01
Figure 1.4
Value at Risk (VaR) from the Normal Distribution
Return Probability Distribution
18
16
14
Return Distribution
12
10
1-day 1% VaR = 0.05825

2
0
-0,08 -0,05 -0,03 0,00 0,03 0,05 0,08
Portfolio Return
Figure 1.4
1
0,9
0,8
Cumulative Distribution
0,7
0,6
0,5
0,4
0,3
0,2 1-day 1% VaR= 0.05825
0,1
0
-0,08 -0,05 -0,03 0,00 0,03 0,05 0,08
Portfolio Return
measure
l Consider a portfolio whose value consists of 40
shares in Microsoft (MS) and 50 shares in GE.
l To calculate VaR for the portfolio, collect historical
share price data for MS and GE and construct the
historical portfolio pseudo returns

l The stock prices include accrued dividends and other
distributions
l Constructing a time series of past portfolio pseudo

returns enables us to generate a portfolio volatility
series using for example the RiskMetrics approach
where

measure
l We can now directly model the volatility of the portfolio
return, RPF,t+1, call it sPF,t+1, and then calculate the
VaR for the portfolio as
• We assume that the portfolio returns are normally

distributed

Figure 1.5
1-day, RiskMetrics 1% VaR in S&P500 Portfolio
Jan 1, 2001 - Dec 31, 2010
0,140
0,120
0,100
0,080
VaR
0,060
0,040
0,020
0,000
Sep 01
Sep 02
Sep 03
Sep 04
Sep 05
Sep 06
Sep 07
Sep 08
Sep 09
Sep 10
Jan 01
Jan 02
Jan 03
Jan 04
Jan 05
Jan 06
Jan 07
Jan 08
Jan 09
Jan 10
Mai 01
Mai 02
Mai 03
Mai 04
Mai 05
Mai 06
Mai 07
Mai 08
Mai 09
Mai 10
Drawbacks
l ofignored
Extreme losses are VaR - The VaR number only tells
us that 1% of the time we will get a return below the
reported VaR number, but it says nothing about what
will happen in those 1% worst cases.
l VaR assumes that the portfolio is constant across the
next K days, which is unrealistic in many cases when K
is larger than a day or a week.
l Finally, it may not be clear how K and p should be
chosen.

Risk Management
and Financial Returns
Elements of Financial Risk Management
Chapter 1
Peter F. Christoffersen

Overview
lBecome familiar with the range of risks facing
corporations, and how to measure and manage these
risks
l Become familiar with the salient features of
speculative asset returns
l Define risks – Market, Credit, Operational, Liquidity
l Certain facts about asset returns
l Define VaR
l Understand the current academic and practitioner
literature

l Classic portfolio theory: Investors can eliminate firm-
specific risk by diversifying holdings to include many
different assets
l Investors should hold a combination of the risk-free
asset and the market portfolio.
l Firms should not waste resources on risk
management, as investors do not care about firm-
specific risk.
l Modigliani-Miller: The value of a firm is independent
of its risk structure.
l Firms should simply maximize expected profits
regardless of the risk entailed.
l Bankruptcy: The real costs of company
reorganization or shut-down will reduce the current
valuation of the firm. Risk management can increase
the value of a firm by reducing the probability of
default.
l Taxes: Risk management can help reduce taxes by
reducing the volatility of earnings.

l Capital structure and the cost of capital: a major
source of corporate default is the inability to service
debt. Proper risk management may allow the firm to
expand more aggressively through debt financing.
l Employee Compensation: due to their implicit
investment in firm-specific human capital, key
employees often have a large and unhedged exposure
to the risk of the firm they work for.

Evidence on RM practices
l Large firms tend to manage risk more actively than
small firms, which is perhaps surprising as small
firms are generally viewed to be more risky.
l However smaller firms may have limited access to

derivatives markets and furthermore lack staff with
risk management skills.

Does RM improve firm
performance?
l The overall answer to this question appears to be YES.
l Analysis of the risk management practices in the gold
mining industry found that share prices were less
sensitive to gold price movements after risk
management.
l Similarly, in the natural gas industry, better risk
management has been found to result in less variable
stock prices.
l A study also found that RM in a wide group of firms
led to a reduced exposure to interest rate and exchange
rate movements.
l Market Risk: the risk to a financial portfolio from
movements in market prices such as equity prices,
foreign exchange rates, interest rates and commodity
prices.
l In financial sector firms market risk should be
managed. e.g. option trading desk.
l In nonfinancial firms market risk should perhaps be
eliminated.

l Liquidity risk: The particular risk from conducting
transactions in markets with low liquidity as
evidenced in low trading volume, and large bid-ask
spreads.
l Under such conditions, the attempt to sell assets may
push prices lower and assets may have to be sold at
prices below their fundamental values or within a
time frame longer than expected.
l Traditionally liquidity risk was given scant attention
in RM, but the events in the fall 1998 sharply
increased the attention devoted to liquidity risk.
l Operational risk: the risk of loss due to physical
catastrophe, technical failure and human error in the
operation of a firm, including fraud, failure of
management and process errors.
l Operational risk-“op risk”-should be mitigated and
ideally eliminated in any firm as the exposure to it
offers very little return (the short-term cost savings of
being careless for example).

l Credit risk: the risk that a counter-party may become
less likely to fulfill its obligations in part or in full on
the agreed upon date.
l Thus credit risk consists not only of the risk that a
counterparty completely defaults on its obligations,
but also that it only pays in part and/or after the
agreed upon date.
l The nature of commercial banks has traditionally
been to take on large amounts of credit risk through
their loan portfolios.
l Today, banks spend much effort to carefully manage
their credit risk exposure.
l Nonbank financials as well as nonfinancial
corporations might instead want to completely
eliminate credit risk as it is not a part of their core
business.
l However, many kinds of credit risks are not readily
hedged in financial markets and corporations are
often forced to take on credit risk exposure which
they would rather be without.
l Business risk: the risk that changes in variables of a
business plan will destroy that plan’s viability,
including quantifiable risks such as business cycle
and demand equation risk, and non-quantifiable risks
such as changes in competitive behavior or
technology.
l Business risk is sometimes simply defined as the
types of risks which are integral part of the core
business of the firm and which should therefore
simply be taken on.
l The daily simple rate of return from the closing prices
of the asset:
• The daily continuously compounded or log return on

an asset is

• The two returns are fairly similar
• The approximation holds because ln(x) ≈ x−1 when x

is close to 1
• Let Ni be the number of units held in asset i and let

VPF;t be the value of the portfolio on day t so that

l Then the portfolio rate of return is
• where w = NiSi,t/VPF,t is the portfolio weight in asset i

i
• Most assets have a lower bound of zero on the price

• Log returns are more convenient for preserving this
lower bound in the risk model because an arbitrarily
large negative log return tomorrow will still imply a
positive price at the end of tomorrow.

l Tomorrow’s price when using log returns is
St+1 = exp(Rt+1)St
l where exp(•) denotes the exponential function
l If instead we use the rate of return definition then
tomorrow’s closing price is
St+1 = (1+rt+1)St
l Here St+1 could go negative unless the assumed
distribution of tomorrow’s return, rt+1, is bounded
below by −1

l With log return definition, we can easily calculate the
compounded return at the K−day horizon as the sum
of the daily returns:

Stylized facts of asset returns
l We can consider the following list of so-called
stylized facts which apply to most stochastic returns.
l Each of these facts will be discussed in detail in the
first part of the book.
l We will use daily returns on the S&P500 from
01/01/2001 to 12/31/2010 to illustrate each of the
features.

Stylized fact 1
l Daily returns have very little autocorrelation. We can
write
• Returns are almost impossible to predict from their

own past.
• Fig 1.1 shows the correlation of daily S&P500 returns
with returns lagged from one to 100 days.
• We will take this as evidence that the conditional
mean is roughly constant.

Figure 1.1
Autocorrelation of Daily S&P 500 Returns
Jan 1, 2001 - Dec 31, 2010
0,15
Autocorrelation of Daily Returns
0,10
0,05
0,00
-0,05
-0,10
-0,15
0 10 20 30 40 50 60 70 80 90 100
Lag Order
Stylized fact 2
l The unconditional distribution of daily returns have
fatter tail than the normal distribution.
l Fig.1.2 shows a histogram of the daily S&P500 return
data with the normal distribution imposed.
l Notice how the histogram has longer and fatter tails,
in particular in the left side, and how it is more
peaked around zero than the normal distribution.
l Fatter tails mean a higher probability of large losses
than the normal distribution would suggest.
Figure 1.2
Histogram of Daily S&P 500 Returns and the Normal
Distribution
Jan 1, 2001 - Dec 31, 2010
0,3000
0,2500
Normal Frequency
Probability Distribution
0,2000
0,1500
0,1000
0,0500
0,0000
-0,07 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0,00 0,01 0,02 0,03 0,04 0,05 0,06 0,07

Daily Return
Stylized fact 3
l The stock market exhibits occasional, very large
drops but not equally large up-moves.
l Consequently the return distribution is asymmetric or
negatively skewed. This is clear from Figure 1.2 as
well.
l Other markets such as that for foreign exchange tend
to show less evidence of skewness.

Stylized fact 4
l The standard deviation of returns completely
dominates the mean of returns at short horizons such as
daily.
l It is typically not possible to statistically reject a zero
mean return.
l Our S&P 500 data have a daily mean of 0.0056% and a
daily standard deviation of 1.3771%.

Stylized fact 5
l Variance measured for example by squared returns,
displays positive correlation with its own past.
l This is most evident at short horizons such as daily or
weekly.
l Fig 1.3 shows the autocorrelation in squared returns
for the S&P500 data, that is
• Models which can capture this variance dependence

will be presented in Chapters 4 & 5
Figure 1.3
Autocorrelation of Squared Daily S&P 500 Returns
Jan 1, 2010 - Dec 31, 2010
0,45
Autocorrelation of Squared Returns
0,40
0,35
0,30
0,25
0,20
0,15
0,10
0,05
0,00
0 10 20 30 40 50 60 70 80 90 100
Lag Order
Stylized fact 6
l Equity and equity indices display negative correlation
between variance and returns.
l This often termed the leveraged effect, arising from
the fact that a drop in stock price will increase the
leverage of the firm as long as debt stays constant.
l This increase in leverage might explain the increase
variance associated with the price drop. We will
model the leverage effect in Chapters 4 and 5.

Stylized fact 7
l Correlation between assets appears to be time varying.
l Importantly, the correlation between assets appear to
increase in highly volatile down-markets and extremely
so during market crashes.
l We will model this important phenomenon in Chapter 7

Stylized fact 8
l Even after standardizing returns by a time-varying
volatility measure, they still have fatter than normal
tails.
l We will refer to this as evidence of conditional non-
normality.
l It will be modeled in Chapters 6 and 9.

Stylized fact 9
l As the return-horizon increases, the unconditional
return distribution changes and looks increasingly
like the normal distribution.
l Issues related to risk management across horizons
will be discussed in Chapter 8.

l Based on the above of stylized facts our model of
individual asset returns will take the generic form
• The conditional mean return is thus mt+1 and the

conditional variance
• The random variable zt+1 is an innovation term,

which we assume is identically and independently
distributed (i.i.d.) as D(0,1).
JP Morgan’s RiskMetrics model for dynamic volatility
• The volatility for tomorrow, time t+1, is computed at the

end of today, time t, using the following simple updating
rule:
• On the first day of the sample, t = 0, the volatility can

be set to the sample variance of the historical data
available.

From asset returns to portfolio
returns
• The value of a portfolio with n assets at time t is the
weighted average of the asset prices using the
current holdings of each asset as weights:
• The return on the portfolio between day t+1 and day t

is then defined as when
using arithmetic returns
• When using log returns return on the portfolio is:

Introducting the VaR risk
measure
l Value-at-Risk - What loss is such that it will only be
exceeded p·100% of the time in the next K trading
days?
l VaR is often defined in dollars, denoted by $VaR
l $VaR loss is implicitly defined from the probability

of getting an even larger loss as in

l Note by definition that (1−p)100% of the time, the
$Loss will be smaller than the VaR.
l Also note that for this course we will use VaR based
on log returns defined as

measure
l Now we are (1−p)100% confident that we will get a
return better than −VaR.
l It is much easier to gauge the magnitude of VaR
when it is written in return terms
l Knowing that the $VaR of a portfolio is $500,000
does not mean much unless we know the value of the
portfolio
l The two VaRs are related as follows:

l
measure
Suppose our portfolio consists of just one security
l For example an S&P 500 index fund
l Now we can use the Risk-Metrics model to provide the VaR
for the portfolio.
l Let VaRPt+1 denote the p .100% VaR for the 1-day ahead
return, and assume that returns are normally distributed with
zero mean and standard deviation sPF,t+1. Then:

measure
l F(z) calculates the probability of being below the
number z
l F-1P= F-1(P) instead calculates the number such that
p.100% of the probability mass is below F-1P
l Taking F-1(*) on both sides of the preceding equation
yields the VaR as

l If we let p = 0.01 then we get F-1P= F-10.01= » -2.33
l If we assume the standard deviation forecast, sPF,t+1

for tomorrow’s return is 2.5% then:

measure
l -1 F P is always negative for p < 0.5
l The negative sign in front of the VaR formula is

needed because VaR is defined as a positive number
l Here VaR is interpreted such that there is a 1%

chance of losing more than 5.825% of the portfolio
value today.

l If the value of the portfolio today is $2 million, the

$VaR would simply be
l For the next figure, note that we assume K = 1 and

p = 0.01
Figure 1.4
18
16
14
Return Distribution
12
10
1-day 1% VaR = 0.05825

2
0
-0,08 -0,05 -0,03 0,00 0,03 0,05 0,08
Portfolio Return
Figure 1.4
1
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2 1-day 1% VaR= 0.05825
0,1
0
-0,08 -0,05 -0,03 0,00 0,03 0,05 0,08
Portfolio Return
measure
l Consider a portfolio whose value consists of 40
shares in Microsoft (MS) and 50 shares in GE.
l To calculate VaR for the portfolio, collect historical
share price data for MS and GE and construct the
historical portfolio pseudo returns

l The stock prices include accrued dividends and other
distributions
l Constructing a time series of past portfolio pseudo

returns enables us to generate a portfolio volatility
series using for example the RiskMetrics approach
where

measure
l We can now directly model the volatility of the portfolio
return, RPF,t+1, call it sPF,t+1, and then calculate the
VaR for the portfolio as
• We assume that the portfolio returns are normally

distributed

Figure 1.5
1-day, RiskMetrics 1% VaR in S&P500 Portfolio
Jan 1, 2001 - Dec 31, 2010
0,140
0,120
0,100
0,080
VaR
0,060
0,040
0,020
0,000
Sep 01
Sep 02
Sep 03
Sep 04
Sep 05
Sep 06
Sep 07
Sep 08
Sep 09
Sep 10
Jan 01
Jan 02
Jan 03
Jan 04
Jan 05
Jan 06
Jan 07
Jan 08
Jan 09
Jan 10
Mai 01
Mai 02
Mai 03
Mai 04
Mai 05
Mai 06
Mai 07
Mai 08
Mai 09
Mai 10
Drawbacks
l ofignored
Extreme losses are VaR - The VaR number only tells
us that 1% of the time we will get a return below the
reported VaR number, but it says nothing about what
will happen in those 1% worst cases.
l VaR assumes that the portfolio is constant across the
next K days, which is unrealistic in many cases when K
is larger than a day or a week.
l Finally, it may not be clear how K and p should be
chosen.

Historical Simulation,
Value-at-Risk,
and Expected Shortfall
Elements of
Financial Risk Management
Chapter 2
Peter Christoffersen
168
Elements of Financial
Risk Management
Overview
Objectives
l Introduce the most commonly used method
for computing VaR, namely Historical
Simulation and discuss the pros and cons
of this method.
l Discuss the pros and cons of the V aR risk

measure
Consider the Expected Shortfall, ES,

l of Financial
169Elements
Risk Management
Chapter is organized as follows:
l Introduction of the historical simulation (HS)
method and its pros and cons.
l Introduction of the weighted historical
simulation (WHS). We then compare HS
and WHS during the 1987 crash.
l Comparison of the performance of HS and
RiskMetrics during the 2008-2009 financial
crisis.
l Then we simulate artificial return data and
assess the HS VaR on this data.
170 Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
l Compare the VaR risk measure with ES.
Defining Historical Simulation
l Let today be day t. Consider a portfolio of
n assets. If we today own Ni,t units or shares
of asset i then the value of the portfolio
today is
• We use today’s portfolio holdings but historical

asset prices to compute yesterday’s pseudo portfolio
value as
171Elements of Financial
Risk Management
l This is a pseudo value because the units of
each asset held typically changes over time.
The pseudo log return can now be defined
as
• Consider the availability of a past sequence of m
daily hypothetical portfolio returns, calculated
using past prices of the underlying assets of the
portfolio, but using today’s portfolio weights, call
it {RPF,t+1-t}mt=1
Risk Management
l Distribution of RPF,t+1 is captured by the
histogram of {RPF,t+1-t}mt=1
l The VaR with coverage rate, p is calculated as
100pth percentile of the sequence of past
portfolio returns.
• Sort the returns in {RPF,t+1-t}mt=1 in ascending order

• Choose VaRPt+1 such that only 100p% of the
observations are smaller than the VaRPt+1
• Use linear interpolation to calculate the exact VaR
number.
Risk Management
Pros
Pros
and Cons of HS
l the ease with which it is implemented.
l its model-free nature.
Cons
l It is very easy to implement. No numerical
optimization has to be performed.
l It is model-free. It does not rely on any
particular parametric model such as a
RiskMetrics
Risk Management
model.
Issues with model free nature of HS
How large should m be?
l If m is too large, then the most recent
observations will carry very little weight, and
the VaR will tend to look very smooth over
time.
l If m is too small, then the sample may not
include enough large losses to enable the risk
manager to calculate VaR with any precision.
l To calculate 1% VaRs with any degree of
precision
for the next 10 days, HS technique
Riskneeds
Managementa large m value
Figure 2.1:
VaRs from HS with 250 and 1,000 Return Days
Jul 1, 2008 - Dec 31, 2010
0,0900
0,0800
Historical Simulation VaR
0,0700
0,0600
0,0500
0,0400
HS-VaR(250)
HS-VaR(1000)
0,0300
0,0200
Jun 08 Jul 08 Sep 08 Okt 08 Dez 08 Feb 09 Mär 09 Mai 09 Jul 09 Aug 09 Okt 09 Dez 09 Jan 10
Risk Management
Weighted Historical Simulation
l WHS relieves the tension in the choice of m
l It assigns relatively more weight to the most recent
observations and relatively less weight to the
returns further in the past
l It is implemented as follows –
l Sample of m past hypothetical returns, {RPF,t+1-
} m
t t=1 is assigned probability weights
declining exponentially through the past as follows
Risk Management
Weighted Historical Simulation
l Today’s observation is assigned the weight h1 =
(1- h) / (1- hm)
l ht goes to zero as t gets large, and that the weights
ht for t = 1,2,...,m sum to 1
l Typical value for h is between 0.95 and 0.99
l The observations along with their assigned
weights are sorted in ascending order.
l The 100p% VaR is calculated by accumulating
the weights of the ascending returns until
100p% is reached.
Risk Management
Pros and Cons of WHS
Pros
l Once h is chosen, WHS does not require

estimation and becomes easy to implement
l It’s weighting function builds dynamics into
the WHS technique
l The weighting function also makes the choice
of m somewhat less crucial.
l
179 WHS
Elements responds quickly to large losses
of Financial
Risk Management
Pros and Cons of WHS
Cons
lNo guidance is given on how to choose η
lEffect on the weighting scheme of positive

versus negative past returns
lIf we are short the market, a market crash has
no impact on our VaR. WHS does not respond to
large gains
lthe multiday VaR requires a large amount of
past
180Elementsdaily
of Financialreturn data, which is not easy to
Risk Management
Advantages of Risk Metrics model
l It can pick up the increase in market variance

from the crash regardless of whether the crash
meant a gain or a loss
l In this model, returns are squared and losses
and gains are treated as having the same
impact on tomorrow’s variance and therefore
on the portfolio risk.
Risk Management
Figure 2.2 A:
Historical Simulation VaR and Daily Losses from
Long S&P500 Position, October 1987
25%
Return VaR (HS)

20%
15%
HS VaR and Loss
10%
5%
0%
-5%
-10%
-15%
01-Okt 06-Okt 11-Okt 16-Okt 21-Okt 26-Okt 31-Okt
Loss Date
Risk Management
Figure 2.2 B:
Short S&P500 Position, October 1987
25%
Return VaR (WHS)

20%
15%
HS VaR and Loss
10%
5%
0%
-5%
-10%
-15%
Loss Date
Risk Management
Figure 2.3 A:
Short S&P500 Position, October 1987
15%
Return VaR (HS)

10%
5%
HS VaR and Loss
0%
-5%
-10%
-15%
-20%
-25%
Loss Date
Risk Management
Figure 2.3 B:
Weighted Historical Simulation VaR and Daily Losses
from Short S&P500 Position, October 1987
15%
Return VaR (WHS)

10%
5%
WHS VaR and Loss
0%
-5%
-10%
-15%
-20%
-25%
Loss Date
Risk Management
Evidence from the 2008-2009
Crisis
l We consider the daily closing prices for a
total return index of the S&P 500 starting in
July 2008 and ending in December 2009.
l The index lost almost half its value between

July 2008 and the market bottom in March
2009.
l The recovery in the index starting in March

2009 continued
through the end of 2009.
Risk Management
Figure 2.4: S&P 500 Total Return Index:
2008-2009 Crisis Period
2.200
S&P500
2.000
1.800
S&P500-Closing Price
1.600
1.400
1.200
1.000
Jul 08 Sep 08 Nov 08 Jan 09 Mär 09 Mai 09 Jul 09 Sep 09 Nov 09
Risk Management
Evidence from the 2008-2009 Crisis
l The 10-day 1% HS VaR is computed from the
1-day VaR by simply multiplying it by
• Alternative to HS is the RiskMetrics variance

model
• 10-day, 1% VaR computed from the Risk- Metrics
model is as follows:
Risk Management
l where the variance dynamics are driven by
Difference between the HS and the RM VaRs

•The HS VaR rises much more slowly as the crisis gets
underway in the fall of 2008
•The HS VaR stays at its highest point for almost a
year during which the volatility in the market has
declined considerably
•HS VaR will detect the brewing crisis quite slowly
and will enforce excessive caution after volatility
drops in the market
Risk Management
Figure 2.5: 10-day, 1% VaR from Historical
Simulation and RiskMetrics During the
2008-2009 Crisis Period
40%
RiskMetrics HS
35%
30%
Value at Risk (VaR)
25%
20%
15%
10%
5%
0%
Jul 08 Aug 08 Sep 08 Okt 08 Nov 08 Dez 08 Jan 09 Feb 09Mär 09 Apr 09 Mai 09 Jun 09 Jul 09 Aug 09 Sep 09 Okt 09 Nov 09 Dez 09
Risk Management
l The units in figure above refer to the least
percent of capital that would be lost over the
next 10 days in the 1% worst outcomes.
l Let’s put some dollar figures on this effect

l Assume that each day a trader has a 10-day,
1% dollar VaR limit of $100,000
l Thus each day he is therefore allowed to
invest
Risk Management
l Let’s assume that the trader each day invests
the maximum amount possible in the S&P 500
• The daily P/L is computed as
Risk Management
Figure 2.6: Cumulative P/L from Traders with
HS and RM VaRs 150.000
P/L from RM VaR

100.000
P/L from HS VaR
50.000
0
Cumulati ve P/L
-50.000
-100.000
-150.000
-200.000
-250.000
-300.000
-350.000
-400.000
Jul Jul Aug Aug Sep Okt Okt Nov Dez Dez Jan Feb Mär Mär Apr Apr Mai Jun Jun Jul Aug Aug Sep Okt Okt Nov Dez Dez
08 08 08 08 08 08 08 08 08 08 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09 09
Risk Management
Performance difference between HS and RM
VaRs
l The RM trader will lose less in the fall of 2008

and earn much more in 2009.
l The HS trader takes more losses in the fall of
2008 and is not allowed to invest sufficiently in
the market in 2009
l The HS VaR reacts too slowly to increases in
volatility
194Elements of Financial as well as to decreases in volatility.
Risk Management
The Probability of Breaching the HS VaR
l Assume that the S&P 500 market returns are

generated by a time series process with
dynamic volatility and normal innovations
l Assume that innovation to S&P 500 returns
each day is drawn from the normal
distribution with mean zero and variance
equal to
l We can write:
Risk Management
l Simulate 1,250 return observations from above
equation
l Starting on day 251, compute each day the 1-
day, 1% VaR using Historical Simulation
l Compute the true probability that we will
observe a loss larger than the HS VaR we have
computed
l This is the probability of a VaR breach
Risk Management
Figure 2.7: Actual Probability of Loosing More
than the 1% HS VaR When Returns Have
Dynamics Variance
0,18
0,16
0,14
Probability of VaR Breach
0,12
0,10
0,08
0,06
0,04
0,02
0,00
1 101 201 301 401 501 601 701 801 901
Risk Management
l where is the cumulative density function
for a standard normal random variable.
l If the HS VaR model had been accurate then
this plot should show a roughly flat line at
1%
l Here we see numbers as high as 16% and
numbers very close to 0%
l The HS VaR will overestimate risk when true
market volatility is low, which will generate a
low probability of a VaR breach
HS will underestimate risk when true
lManagement
Risk
VaR with Extreme Coverage Rates
l The tail of the portfolio return distribution
conveys information about the future losses.
l Reporting the entire tail of the return

distribution corresponds to reporting VaRs for
many different coverage rates
l Here p ranges from 0.01% to 2.5% in

increments
When using HS with a 250-day sample it is not
Risk Management
l
Figure 2.8: Relative Difference between Non-
Normal (Excess Kurtosis=3) and Normal VaR
80
70
60
Relative VaR Difference
50
40
30
20
10
0
0 0,005 0,01 0,015 0,02 0,025
VaR Coverage Rate, p
Risk Management
VaR with Extreme Coverage Rates
l Note that (from the above figure) as p gets
close to zero the nonnormal VaR gets much
larger than the normal VaR
l When p = 0.025 there is almost no difference

between the two VaRs even though the
underlying distributions are different
Risk Management
Expected Shortfall
l VaR is concerned only with the percentage of
losses that exceed the VaR and not the magnitude
of these losses.
l Expected Shortfall (ES), or TailVaR accounts for
the magnitude of large losses as well as their
probability of occurring
l Mathematically ES is defined as
Risk Management
Expected Shortfall
l The negative signs in front of the expectation
and the VaR are needed because the ES and
the VaR are defined as positive numbers
l The ES tells us the expected value of
tomorrow’s loss, conditional on it being worse
than the VaR
l The Expected Shortfall computes the average
of the tail outcomes weighted by their
probabilities
l ES tells us the expected loss given that we
actually get a loss from the 1% tail
Risk Management
Expected Shortfall
l To compute ES we need the distribution of a
normal variable conditional on it being below
the VaR
l The truncated standard normal distribution is

defined from the standard normal distribution
as
Risk Management
Expected Shortfall
l f(•) denotes the density function and F(•)
the cumulative density function of the
standard normal distribution
l In the normal distribution case ES can be
derived as
Risk Management
Expected Shortfall
l In the normal case we know that
• Thus, we have
• The relative difference between ES and VaR is
Risk Management
Expected Shortfall
l For example, when p =0.01, we have
and the relative difference is then
• In the normal case, as p gets close to zero, the ratio of

the ES to the VaR goes to 1
• From the below figure, the blue line shows that when
excess kurtosis is zero, the relative difference between
the ES and VaR is 15%
Risk Management
Expected Shortfall
l The blue line also shows that for moderately
large values of excess kurtosis, the relative
difference between ES and VaR is above 30%
l From the figure, the relative difference
between VaR and ES is larger when p is larger
and thus further from zero
l When p is close to zero VaR and ES will both
capture the fat tails in the distribution
l When p is far from zero, only the ES will
capture the fat tails in the return distribution
Risk Management
Figure 2.9: ES vs VaR as a Function of Kurtosis
50
p=1% p=5%
45
40
35
(ES - VaR ) / ES
30
25
20
15
10
0
0 1 2 3 4 5 6
Excess Kurtosis
Risk Management
Summary
l VaR is the most popular risk measure in use

l HS is the most often used methodology to
compute VaR
l VaR has some shortcomings and using HS to
compute VaR has serious problems as well
l We need to use risk measures that capture the
degree of fatness in the tail of the return
distribution
l
210 Weof need
Elements Financial risk models that properly account for
Risk Management
The VaR Measure
Chapter 8
The Question Being Asked in VaR
“What loss level is such that we are X%

confident it will not be exceeded in N
business days?”
Value at Risk (VaR)
l “VaR measures the worst expected loss over a given time interval
l under normal market conditions at a given confidence level.”
l - Jorion (1997)
l
l “Value at Risk is an estimate, with a given degree of confidence, of how
much one can lose from one’s portfolio over a given time horizon.”
l - Wilmott (1998)
l “Value-at-Risk or VAR is a dollar measure of the minimum loss that would

be expected over a period of time with a given probability”
l - Chance(1998)
l 95% confidence level VaR Û 5% probability minimum loss

l (over given horizon)
l max. loss with 95% confidence Û min. loss with 5% probability
l (for given time interval)
Asset price standard deviation
l Assume lognormal returns: Let dS/S ~ lognormal(µ,s)
l where s is annualized return volatility (standard deviation)
l The standard deviation of the asset price (S) over a period t is:
l S (s t.5) = S ( sÖ t )
l For example, let
l S = $ 100.00
l s = 40%
l t = 1 week = 1/52
l then the weekly standard deviation (s.d.) of the price is
l weekly s.d. = 100 (0.40) (.192).5 = 40(0.139) = $ 5.55
l similarly,
l daily s.d. = 100(0.40)(1/252).5 = 40(0.063) = $ 2.52
l monthly s.d. = 0.40(0.40)(1/12).5 = 40(0.289) = $ 11.55
Confidence levels and the inverse distribution function
`
Let VaR = - S (sÖt) N (1 - confidence level)
(for long position in underlying asset)
`
where N (x%) = inverse cumulative distribution function
for the standard normal
N`(x%) = number of standard deviations Run

Standard normal distribution
from the mean such that
the probability of obtaining
a lower outcome is x%
example:
desired confidence level is 95%,
then N`(1-.95) = N`(5%) = - 1.65 -1.65 -1 0 1 1.65
In other words, N(-1.65) = 0.05 (5%) 5% probability of return lower than

so N`(0.05) = -1.65 1.65 standard deviations
below the mean
VaR and Regulatory Capital
l Regulators base the capital they require

banks to keep on VaR
l The market-risk capital is k times the 10-
day 99% VaR where k is at least 3.0
l Under Basel II, capital for credit risk and
operational risk is based on a one-year
99.9% VaR
Advantages of VaR
l It captures an important aspect of risk
in a single number
l It is easy to understand
l It asks the simple question: “How bad can
things get?”
Example 8.1 (page 159)
l The gain from a portfolio during six month
is normally distributed with mean $2
million and standard deviation $10 million
l The 1% point of the distribution of gains is
2−2.33×10 or − $21.3 million
l The VaR for the portfolio with a six month
time horizon and a 99% confidence level is
$21.3 million.
l All outcomes between a loss of $50 million
and a gain of $50 million are equally likely
for a one-year project
l The VaR for a one-year time horizon and a
99% confidence level is $49 million
Examples 8.3 and 8.4 (page 160)
l A one-year project has a 98% chance of
leading to a gain of $2 million, a 1.5%
chance of a loss of $4 million, and a 0.5%
chance of a loss of $10 million
l The VaR with a 99% confidence level is $4
million
l What if the confidence level is 99.9%?
l What if it is 99.5%?
Cumulative Loss Distribution for
Examples 8.3 and 8.4 (Figure 8.3, page 160)
VaR vs. Expected Shortfall
l VaR is the loss level that will not be

exceeded with a specified probability
l Expected shortfall is the expected loss
given that the loss is greater than the VaR
level (also called C-VaR and Tail Loss)
l Two portfolios with the same VaR can
have very different expected shortfalls
Distributions with the Same VaR but
Different Expected Shortfalls
VaR
VaR
Coherent Risk Measures (page 162)
l Define a coherent risk measure as the amount of
cash that has to be added to a portfolio to make its
risk acceptable
l Properties of coherent risk measure
l If one portfolio always produces a worse outcome than
another its risk measure should be greater
l If we add an amount of cash K to a portfolio its risk measure
should go down by K
l Changing the size of a portfolio by l should result in the risk
measure being multiplied by l
l The risk measures for two portfolios after they have been
merged should be no greater than the sum of their risk
measures before they were merged
VaR vs Expected Shortfall
l VaR satisfies the first three conditions but
not the fourth one
l Expected shortfall satisfies all four
conditions.
Example 8.5 and 8.7
l Each of two independent projects has a probability 0.98
of a loss of $1 million and 0.02 probability of a loss of
$10 million
l What is the 97.5% VaR for each project?
l What is the 97.5% expected shortfall for each project?
l What is the 97.5% VaR for the portfolio?
l What is the 97.5% expected shortfall for the portfolio?
Examples 8.6 and 8.8
l Two $10 million one-year loans each has a 1.25%
chance of defaulting. All recoveries between 0 and 100%
are equally likely. If there is no default the loan leads to a
profit of $0.2 million. If one loan defaults it is certain that
the other one will not default.
l What is the 99% VaR and expected shortfall of each
project
l What is the 99% VaR and expected shortfall for the
portfolio
Spectral Risk Measures
l A spectral risk measure assigns weights to
quantiles of the loss distribution
l VaR assigns all weight to Xth quantile of
the loss distribution
l Expected shortfall assigns equal weight to
all quantiles greater than the Xth quantile
l For a coherent risk measure weights must
be a non-decreasing function of the
quantiles
Normal Distribution Assumption
l The simplest assumption is that daily
gains/losses are normally distributed and
independent with mean zero
l It is then easy to calculate VaR from the
standard deviation (1-day VaR=2.33s)
l The T-day VaR equals T times the one-day
VaR
l Regulators allow banks to calculate the 10 day
VaR as 10 times the one-day VaR
Independence Assumption in VaR
Calculations (Equation 8.3, page 166)
l When daily changes in a portfolio are
identically distributed and independent the
variance over T days is T times the
variance over one day
l When there is autocorrelation equal to r
the multiplier is increased from T to
T + 2(T - 1)r + 2(T - 2)r 2 + 2(T - 3)r 3 + ! 2r T -1
Impact of Autocorrelation: Ratio of N-day
VaR to 1-day VaR (Table 8.1, page 204)
T=1 T=2 T=5 T=10 T=50 T=250
r=0 1.0 1.41 2.24 3.16 7.07 15.81
r=0.05 1.0 1.45 2.33 3.31 7.43 16.62
r=0.1 1.0 1.48 2.42 3.46 7.80 17.47
r=0.2 1.0 1.55 2.62 3.79 8.62 19.35
Choice of VaR Parameters
l Time horizon should depend on how quickly
portfolio can be unwound. Bank regulators in
effect use 1-day for market risk and 1-year for
credit/operational risk. Fund managers often
use one month
l Confidence level depends on objectives.
Regulators use 99% for market risk and 99.9%
for credit/operational risk.
l A bank wanting to maintain a AA credit rating will
often use confidence levels as high as 99.97%
for internal calculations.
VaR Measures for a Portfolio where an
amount xi is invested in the ith component
of the portfolio (page 168-169)
l Marginal VaR: ¶VaR
¶xi
l Incremental VaR: Incremental effect of

the ith component on VaR
¶VaR
l Component VaR: xi
¶xi
Properties of Component VaR
l The component VaR is approximately the
same as the incremental VaR
l The total VaR is the sum of the component
VaR’s (Euler’s theorem)
l The component VaR therefore provides a
sensible way of allocating VaR to different
activities
Back-testing (page 169-171)
l Back-testing a VaR calculation methodology involves
looking at how often exceptions (loss > VaR) occur
l Alternatives: a) compare VaR with actual change in
portfolio value and b) compare VaR with change in
portfolio value assuming no change in portfolio
composition
l Suppose that the theoretical probability of an exception
is p (=1−X). The probability of m or more exceptions in n
days is
n
n!
å
k = m k!( n - k )!
p k
(1 - p ) n-k
Bunching
l Bunching occurs when exceptions are not
evenly spread throughout the back testing
period
l Statistical tests for bunching have been
developed by Christoffersen (See page 171)
Correlations and
Copulas
Chapter 11
Correlation and Covariance
l The coefficient of correlation between two
variables V1 and V2 is defined as
E (V1V2 ) - E (V1 ) E (V2 )

SD (V1 ) SD (V2 )
l The covariance is
E(V1V2)−E(V1 )E(V2)
Independence
l V1 and V2 are independent if the
knowledge of one does not affect the
probability distribution for the other
f (V2 V1 = x ) = f (V2 )
where f(.) denotes the probability density
function
Independence is Not the Same as
Zero Correlation
l Suppose V1 = –1, 0, or +1 (equally
likely)
l If V1 = -1 or V1 = +1 then V2 = 1
l If V1 = 0 then V2 = 0
V2 is clearly dependent on V1 (and vice
versa) but the coefficient of correlation
is zero
Types of Dependence (Figure 11.1, page 235)
E(Y) E(Y)
X X
(a) (b)
E(Y)
(c)
Monitoring Correlation Between
Two Variables X and Y
Define xi=(Xi−Xi-1)/Xi-1 and yi=(Yi−Yi-1)/Yi-1
Also
varx,n: daily variance of X calculated on day n-1
vary,n: daily variance of Y calculated on day n-1
covn: covariance calculated on day n-1
The correlation is
cov n
varx ,n vary ,n
Covariance
l The covariance on day n is
E(xnyn)−E(xn)E(yn)
l It is usually approximated as E(xnyn)
Monitoring Correlation continued
EWMA:
cov n = l cov n -1 + (1 - l) xn -1 yn -1
GARCH(1,1)
cov n = w + axn -1 yn -1 + b cov n -1
Positive Finite Definite Condition
A variance-covariance matrix, W, is
internally consistent if the positive semi-
definite condition
wTWw ≥ 0
holds for all vectors w
Example
The variance covariance matrix
æ 1 0 0.9ö
ç ÷
ç 0 1 0.9÷
ç ÷
è 0.9 0.9 1ø
is not internally consistent
V1 and V2 Bivariate Normal
l Conditional on the value of V1, V2 is normal with
mean
V1 - µ1
µ 2 + rs 2
s1
and standard deviation s 2 1 - r 2 where µ1,, µ2, s1,

and s2 are the unconditional means and SDs of
V1 and V2 and r is the coefficient of correlation
between V1 and V2
Multivariate Normal Distribution
l Fairly easy to handle

l A variance-covariance matrix defines
the variances of and correlations
between variables
l To be internally consistent a variance-
covariance matrix must be positive
semidefinite
Generating Random Samples for
Monte Carlo Simulation (pages 239-240)
l =NORMSINV(RAND()) gives a random

sample from a normal distribution in
Excel
l For a multivariate normal distribution a
method known as Cholesky’s
decomposition can be used to generate
random samples
Factor Models (page 240)
l When there are N variables, Vi (i = 1,
2,..N), in a multivariate normal distribution
there are N(N−1)/2 correlations
l We can reduce the number of correlation
parameters that have to be estimated with
a factor model
One-Factor Model continued
l If Ui have standard normal distributions
we can set
U i = ai F + 1 - ai2 Z i
where the common factor F and the
idiosyncratic component Zi have
independent standard normal
distributions
l Correlation between Ui and Uj is ai aj
Gaussian Copula Models:
Creating a correlation structure for variables that are not
normally distributed
l Suppose we wish to define a correlation structure between

two variable V1 and V2 that do not have normal distributions
l We transform the variable V1 to a new variable U1 that has a
standard normal distribution on a “percentile-to-percentile”
basis.
l We transform the variable V2 to a new variable U2 that has a
standard normal distribution on a “percentile-to-percentile”
basis.
l U1 and U2 are assumed to have a bivariate normal
distribution
Risk Management and Financial Institutions 3e, Chapter 11, 252

Copyright © John C. Hull 2012
The Correlation Structure Between the V’s is
Defined by that Between the U’s
-0.2 0 0.2 0.4 0.6 0.8 1 1.2 -0.2 0 0.2 0.4 0.6 0.8 1 1.2
V1 V2
One-to-one
mappings
-6 -4 -2 0 2 4 6 -6 -4 -2 0 2 4 6
U2
U1
Correlation
Assumption
Example (page 241)
V1 V2
V1 Mapping to U1
V1 Percentile U1
0.2 20 -0.84
0.4 55 0.13
0.6 80 0.84
0.8 95 1.64
V2 Mapping to U2
V2 Percentile U2
0.2 8 −1.41
0.4 32 −0.47
0.6 68 0.47
0.8 92 1.41
Example of Calculation of Joint
l Probability that V1 and V2 are both less

than 0.2 is the probability that U1 < −0.84
and U2 < −1.41
l When copula correlation is 0.5 this is
M( −0.84, −1.41, 0.5) = 0.043
where M is the cumulative distribution
function for the bivariate normal
distribution
Other Copulas
l Instead of a bivariate normal distribution
for U1 and U2 we can assume any other
joint distribution
l One possibility is the bivariate Student t
distribution
5000 Random Samples from the
Bivariate Normal
0
-5 -4 -3 -2 -1 0 1 2 3 4 5
-1
-2
-3
-4
-5
5000 Random Samples from the
Bivariate Student t
10
0
-10 -5 0 5 10
-5
-10
Multivariate Gaussian Copula
l We can similarly define a correlation
structure between V1, V2,…Vn
l We transform each variable Vi to a new
variable Ui that has a standard normal
distribution on a “percentile-to-percentile”
basis.
l The U’s are assumed to have a
multivariate normal distribution
Factor Copula Model
In a factor copula model the correlation
structure between the U’s is generated by
assuming one or more factors.
Credit Default Correlation
l The credit default correlation between two
companies is a measure of their tendency
to default at about the same time
l Default correlation is important in risk
management when analyzing the benefits
of credit risk diversification
l It is also important in the valuation of some
credit derivatives
Model for Loan Portfolio
l We map the time to default for company i, Ti, to a
new variable Ui and assume
U i = aF + 1- a Z i 2
l Where F and the Zi have independent standard

normal distributions
l The copula correlation is r=a2
Vasicek’s Model for Loan Portfolio
l Suppose the Bank has a large no. of Loans where PD is 1% but will
change in each year. Suppose Ti is the time firm defaults. All firms has
the same Cum. Prob. Dist, for the time to default Qi.
l The Gaussian copula method can be used to define correlation structure
between Ti’s.
l We map the time to default for company i, Ti, to a new variable Ui N(0,1)
and assume
U i = aF + 1- a Z i 2
l Where F and the Zi have independent standard normal distributions

l The copula correlation is r=a2
Vasicek’s Model for Loan Portfolio
l
Analysis
l To analyze the model we
l Calculate the probability that, conditional on the value
of F, Ui is less than some value U
l This is the same as the probability that Ti is less that T
where T and U are the same percentiles of their
distributions
This leads to
é N -1 [PD] - r F ù
Prob(Ti < T F ) = N ê ú
êë 1- r úû
where PD is the probability of default in time T
The Model continued
l The X% worst case value of F is N-1(X)
l The worst case default rate during time T with a
confidence level of X is therefore
æ N -1[Q(T )] + r N -1 ( X ) ö
WCDR(T,X) = N ç ÷
ç 1- r ÷
è ø
l The VaR for this time horizon and confidence limit
is
VaR (T , X ) = L ´ LGD ´ WCDR (T , X )
where L is loan principal and LGD is loss given default
Gordy’s Result
l In a large portfolio of M loans where each
loan is small in relation to the size of the
portfolio it is approximately true that
M
VaR (T , X ) = å Li ´ LGDi ´ WCDR i (T , X )
i =1
Estimating PD and r
l We can use data on default rates in
conjunction with maximum likelihood
methods
l The probability density function for the
default rate is
ì é ü
1 - r N (DR ) - N (PD) ù ï
2
1- r ï 1 ê -1 æ -1 -1
ö
g (DR ) = expí ( N (DR )) - çç
2
÷ úý
÷ ú
r ïî 2 êë è r ø û ïþ
Value at Risk
Chapter 9
The Question Being Asked in VaR
“What loss level is such that we are X%

confident it will not be exceeded in N
business days?”
Value at Risk - Definitions
l VaR measures the worst expected loss over a given time interval
l under normal market conditions at a given confidence level.”
l - Jorion (1997)
l
l “Value at Risk is an estimate, with a given degree of confidence, of how
much one can lose from one’s portfolio over a given time horizon.”
l - Wilmott (1998)
l “Value-at-Risk or VAR is a dollar measure of the minimum loss that would

be expected over a period of time with a given probability”
l - Chance(1998)
l 95% confidence level VaR Û 5% probability minimum loss

l (over given horizon)
l max. loss with 95% confidence Û min.loss with 5% probability
l (for given time interval)
VaR and Regulatory Capital
l Regulators base the capital they require

banks to keep on VaR
l The market-risk capital is k times the 10-
day 99% VaR where k is at least 3.0
l Under Basel II, capital for credit risk and
operational risk is based on a one-year
99.9% VaR
Advantages of VaR
l It captures an important aspect of risk
in a single number
l It is easy to understand
l It asks the simple question: “How bad can
things get?”
l The gain from a portfolio during six month
is normally distributed with mean $2
million and standard deviation $10 million
l The 1% point of the distribution of gains is
2−2.33×10 or − $21.3 million
l The VaR for the portfolio with a six month
time horizon and a 99% confidence level is
$21.3 million.
l All outcomes between a loss of $50 million
and a gain of $50 million are equally likely
for a one-year project
l The VaR for a one-year time horizon and a
99% confidence level is $49 million
Examples 9.3 and 9.4 (page 186)
l A one-year project has a 98% chance of
leading to a gain of $2 million, a 1.5%
chance of a loss of $4 million, and a 0.5%
chance of a loss of $10 million
l The VaR with a 99% confidence level is $4
million
l What if the confidence level is 99.9%?
l What if it is 99.5%?
Cumulative Loss Distribution for
Examples 9.3 and 9.4 (Figure 9.3, page 186)
VaR vs. Expected Shortfall
l VaR is the loss level that will not be

exceeded with a specified probability
l Expected shortfall is the expected loss
given that the loss is greater than the VaR
level (also called C-VaR and Tail Loss)
l Two portfolios with the same VaR can
have very different expected shortfalls
Distributions with the Same VaR but
Different Expected Shortfalls
VaR
VaR
Coherent Risk Measures (page 188)
l Define a coherent risk measure as the amount of
cash that has to be added to a portfolio to make its
risk acceptable
l Properties of coherent risk measure
l If one portfolio always produces a worse outcome than
another its risk measure should be greater
l If we add an amount of cash K to a portfolio its risk measure
should go down by K
l Changing the size of a portfolio by l should result in the risk
measure being multiplied by l
l The risk measures for two portfolios after they have been
merged should be no greater than the sum of their risk
measures before they were merged
VaR vs Expected Shortfall
l VaR satisfies the first three conditions but
not the fourth one
l Expected shortfall satisfies all four
conditions.
Example 9.5 and 9.7
l Each of two independent projects has a probability 0.98
of a loss of $1 million and 0.02 probability of a loss of
$10 million
l What is the 97.5% VaR for each project?
l What is the 97.5% expected shortfall for each project?
l What is the 97.5% VaR for the portfolio?
l What is the 97.5% expected shortfall for the portfolio?
Examples 9.6 and 9.8
l A bank has two $10 million one-year loans. Possible outcomes are
as follows
Outcome Probability
Neither Loan Defaults 97.5%
Loan 1 defaults, loan 2 does not default 1.25%
Loan 2 defaults, loan 1 does not default 1.25%
Both loans default 0.00%
l If a default occurs, losses between 0% and 100% are equally likely.

If a loan does not default, a profit of 0.2 million is made.
l What is the 99% VaR and expected shortfall of each project
l What is the 99% VaR and expected shortfall for the portfolio
Spectral Risk Measures
l A spectral risk measure assigns weights to
quantiles of the loss distribution
l VaR assigns all weight to Xth quantile of
the loss distribution
l Expected shortfall assigns equal weight to
all quantiles greater than the Xth quantile
l For a coherent risk measure weights must
be a non-decreasing function of the
quantiles
Normal Distribution Assumption
l The simplest assumption is that daily
gains/losses are normally distributed and
independent with mean zero
l It is then easy to calculate VaR from the
standard deviation (1-day VaR=2.33s)
l The T-day VaR equals T times the one-day
VaR
Independence Assumption in VaR
Calculations (Equation 9.3, page 193)
l When daily changes in a portfolio are
identically distributed and independent the
variance over T days is T times the
variance over one day
l When there is autocorrelation equal to r
the multiplier is increased from T to
T + 2(T - 1)r + 2(T - 2)r 2 + 2(T - 3)r 3 + ! 2r T -1
Impact of Autocorrelation: Ratio of T-day
VaR to 1-day VaR (Table 9.1, page 193)
T=1 T=2 T=5 T=10 T=50 T=250
r=0 1.0 1.41 2.24 3.16 7.07 15.81
r=0.05 1.0 1.45 2.33 3.31 7.43 16.62
r=0.1 1.0 1.48 2.42 3.46 7.80 17.47
r=0.2 1.0 1.55 2.62 3.79 8.62 19.35
Choice of VaR Parameters
l Time horizon should depend on how quickly
portfolio can be unwound. Bank regulators in
effect use 1-day for market risk and 1-year for
credit/operational risk. Fund managers often
use one month
l Confidence level depends on objectives.
Regulators use 99% for market risk and 99.9%
for credit/operational risk.
l A bank wanting to maintain a AA credit rating
might use confidence levels as high as 99.97%
for internal calculations.
VaR Measures for a Portfolio where an
amount xi is invested in the ith component
of the portfolio (page 195-196)
l Marginal VaR: ¶VaR
¶xi
l Incremental VaR: Incremental effect of

the ith component on VaR
¶VaR
l Component VaR: xi
¶xi
Properties of Component VaR
l The component VaR is approximately the
same as the incremental VaR
l The total VaR is the sum of the component
VaR’s (Euler’s theorem)
l The component VaR therefore provides a
sensible way of allocating VaR to different
activities
Aggregating VaRs
An approximate approach that seems to works
well is
VaR total = åå VaR VaR r

i j
i j ij
where VaRi is the VaR for the ith segment,

VaRtotal is the total VaR, and rij is the coefficient
of correlation between losses from the ith and
jth segments
Back-testing (page 197-200)
l Back-testing a VaR calculation methodology involves
looking at how often exceptions (loss > VaR) occur
l Alternatives: a) compare VaR with actual change in
portfolio value and b) compare VaR with change in
portfolio value assuming no change in portfolio
composition
l Suppose that the theoretical probability of an exception
is p (=1−X). The probability of m or more exceptions in n
days is
n
n!
å
k = m k!( n - k )!
p k
(1 - p ) n-k
Bunching
l Bunching occurs when exceptions are not
evenly spread throughout the back testing
period
l Statistical tests for bunching have been
developed by Christoffersen (See page 171)
Chapter 10
Properties of Stock
Options
Options, Futures, and Other Derivatives, 8th Edition, 296

Copyright © John C. Hull 2012
Notation
c: European call option C: American call option
price price
p: European put option P: American put option
price price
S0: Stock price today ST: Stock price at option
maturity
K: Strike price
D: PV of dividends paid
T: Life of option
during life of option
s: Volatility of stock
price r Risk-free rate for
maturity T with cont.
comp.
Options, Futures, and Other Derivatives, 8th Edition, Copyright © John C. Hull 2012 297
Effect of Variables on Option
Pricing (Table 10.1, page 215)
Variable c p C P
S0 + − + −
K − + − +
T ? ? + +
s + + + +
r + − + −
D − + − +
American vs European Options
An American option is worth at least as
much as the corresponding European
option
C³c
P³p
Calls: An Arbitrage Opportunity?
l Suppose that
c=3 S0 = 20
T=1 r = 10%
K = 18 D=0
l Is there an arbitrage opportunity? 3.71

Lower and upper Bound for European Call
Option Prices; No Dividends (Equation 10.4, page
220)
c ³ S0 –Ke -rT
c < S0 p <Ke -rT
Puts: An Arbitrage Opportunity?
l Suppose that
p= 1 S0 = 37
T = 0.5 r =5%
K = 40 D =0
l Is there an arbitrage opportunity?

1.79
Lower Bound for European Put
Prices; No Dividends
p ³ Ke -rT–S0
Put-Call Parity: No Dividends
l Consider the following 2 portfolios:

l Portfolio A: European call on a stock +
zero-coupon bond that pays K at time T
l Portfolio C: European put on the stock +
the stock
Values of Portfolios
ST > K ST < K
Portfolio A Call option ST − K 0
Zero-coupon bond K K
Total ST K
Portfolio C Put Option 0 K− ST
Share ST ST
Total ST K
The Put-Call Parity Result (Equation
10.6, page 222)
l Both are worth max(ST , K ) at the maturity

of the options
l They must therefore be worth the same
today. This means that
c + Ke -rT = p + S0
Arbitrage Opportunities
l Suppose that
c= 3 S0= 31
T = 0.25 r = 10%
K =30 D=0
l What are the arbitrage

possibilities when
p = 2.25 ?
p=1?
Early Exercise
l Usually there is some chance that an

American option will be exercised early
l An exception is an American call on a

non-dividend paying stock
l This should never be exercised early
An Extreme Situation
l For an American call option:

S0 = 100; T = 0.25; K = 60; D = 0
Should you exercise immediately?
l What should you do if
l You want to hold the stock for the next 3
months?
l You do not feel that the stock is worth holding
for the next 3 months?
Reasons For Not Exercising a
Call Early (No Dividends)
l No income is sacrificed
l You delay paying the strike price
l Holding the call provides insurance
against stock price falling below strike
price
Bounds for European or American
Call Options (No Dividends)
Should Puts Be Exercised
Early ?
Are there any advantages to

exercising an American put when
S0 = 60; T = 0.25; r=10%

K = 100; D = 0
Bounds for European and American
Put Options (No Dividends)
The Impact of Dividends on Lower
Bounds to Option Prices
(Equations 10.8 and 10.9, page 229)
- rT
c ³ S 0 - D - Ke
- rT
p ³ D + Ke - S0
Extensions of Put-Call Parity
l American options; D = 0
S0 − K < C − P < S0 − Ke−rT
Equation 10.7 p. 224
l European options; D > 0

c + D + Ke −rT = p + S0
l American options; D > 0

S0 − D − K < C − P < S0 − Ke −rT

GRM - Day 1.2

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

GRM - Day 1.2

Uploaded by

Copyright:

Available Formats

Regulation, Basel II, and

Loss over time

l Tier 1 Capital: common equity, non-

Remaining Interest Exch Rate Equity Precious Other

Example: A $100 million swap with 3 years to maturity worth $5 million

On-balance sheet Off-balance sheet items:

For a derivative Cj = max(Vj,0) + ajLj where Vj is

l Influential publication from derivatives

l With netting exposure is

Bank and corporations treated similarly (unlike Basel I)

Rating AAA A+ to BBB+ BB+ to B+ to Below Unrate

Country 0% 20% 50% 100% 100% 150% 100%

Banks 20% 50% 50% 100% 100% 150% 50%

Corporates 20% 50% 100% 100% 150% 150% 100%

l Basel II provides a formula for translating PD

• The 99.9% worst case default rate is

PD=0.1% PD=0.5% PD=1% PD=1.5% PD=2%

(For small firms r is reduced)

The risk - weighted assets are 12.5 times the Capital

Capital = EAD ´ LGD ´ (WCDR - PD )

l Traditionally the Basel Committee has used the credit

l Similar amount of thoroughness in

l Similar three pillars to Basel II

Real World (%) Normal Model (%)

Prob(v > x) = Kx-a

This seems to fit the behavior of the

l In an exponentially weighted moving

Since weights must sum to 1

l The long-run variance rate is 0.0002 so

l In maximum likelihood methods we

l We maximize this to obtain a maximum

Day Si ui vi =si2 -ln vi-ui2/vi

A few lines of algebra shows that

To estimate the volatility for an option

The volatility per year for an option

l The GARCH (1,1) model allows us to

Elements of Financial Risk

Elements of Financial Risk

Elements of Financial Risk

Elements of Financial Risk

l However smaller firms may have limited access to

Elements of Financial Risk

Elements of Financial Risk

Elements of Financial Risk

• The daily continuously compounded or log return on

Elements of Financial Risk

• The approximation holds because ln(x) ≈ x−1 when x

• Let Ni be the number of units held in asset i and let

Elements of Financial Risk

• where w = NiSi,t/VPF,t is the portfolio weight in asset i

• Most assets have a lower bound of zero on the price

Elements of Financial Risk

Elements of Financial Risk

Elements of Financial Risk

Elements of Financial Risk

• Returns are almost impossible to predict from their

Elements of Financial Risk

Elements of Financial Risk

Elements of Financial Risk

Elements of Financial Risk

• Models which can capture this variance dependence

Elements of Financial Risk

Elements of Financial Risk

Elements of Financial Risk