MONASH

BUSINESS
SCHOOL

BFF3651 Week 8

Risk Management:
Market Risk
 Unit Learning Outcomes
• On successful completion of this unit, you should be able to:
– explain the role of treasury operations in an international
or a local bank
– describe how risk management processes work
– demonstrate the application of hedging techniques used in
banks' treasury operations
– apply critical thinking, problem solving and presentation
skills to individual and/or group activities dealing with
treasury management and demonstrate in an individual
summative assessment task the acquisition of a
comprehensive understanding of the topics covered by
BFF3651.

MONASH
BUSINESS
SCHOOL
 Resources
• Lecture note
• Saunders and Cornett’s Financial Institution
Management Chapter 15-Market Risk

MONASH
BUSINESS
SCHOOL
 Learning Objectives
• Our discussion will focus on
– how market risk arises and how it can threaten the
solvency of FIs.
– how to measure market risk based on RiskMetrics
model
– the limitation of RiskMetrics model and alternative
market risk measurement, such as back simulation
approach and Monte Carlo simulation approach.
– Measuring market risk using the Expected Shortfall
(ES) method

MONASH
BUSINESS
SCHOOL
 Learning Objectives
• Our discussion will focus on
– how market risk arises and how it can threaten the
solvency of FIs.
– how to measure market risk based on RiskMetrics
model
– the limitation of RiskMetrics model and alternative
market risk measurement, such as back simulation
approach and Monte Carlo simulation approach.
– Measuring market risk using the Expected Shortfall
(ES) method

MONASH
BUSINESS
SCHOOL
 Introduction: Market Risk
• The uncertainty of FI’s earnings
resulting from changes, mainly
extreme changes, in market
condition
– Related to other risks such as
interest rate risk, FX risk, and
credit risk
– Credit risk can be systematic or
firm-specific.
• Our focus: Systematic credit risk
– Normally used for trading book
– Normally focusing on short-
term

MONASH
BUSINESS
SCHOOL
 Introduction: Market Risk

• Trading book is different from banking book (investment portfolio) in terms of liquidity
and time horizon.
• Loan sale will be reported in trading book as well.
MONASH
BUSINESS
SCHOOL
 Introduction: Just for your interest
• Market risk: Threat to solvency of FIs
• Trading exposes banks to risks
• Nick Leeson’s bet failure demolished Barings Bank
http://news.bbc.co.uk/2/hi/business/375259.stm

• SocGen Losses: Jérôme Kerviel and the Fraud (2008)

• https://en.wikipedia.org/wiki/J%C3%A9r%C3%B4me_Kerviel

• A list of rogue traders in the history :

https://www.theguardian.com/business/2011/sep/15/who-are-
worst-rogue-traders

MONASH
BUSINESS
SCHOOL
 Market risk measurement: Why?

• Important in terms of:

• Management information
• Setting limits
• Resource allocation (risk/return tradeoff)
• Performance evaluation
• Regulation: capital requirement

MONASH
BUSINESS
SCHOOL
 Methods to measure risk exposure

• Generally concerned with estimated potential loss under

adverse market circumstances
• Intuitively, loss = dollar value of the position x
estimated return under adverse market circumstances
• So the focus is to estimate the potential return of the
investment under adverse market circumstances
• Major approaches of measurement:
• JPM RiskMetrics
• Historic or Back Simulation
• Monte Carlo Simulation
• Expected Shortfall (ES)

MONASH
BUSINESS
SCHOOL
 Learning Objectives
• Our discussion will focus on
– how market risk arises and how it can threaten the
solvency of FIs.
– how to measure market risk based on RiskMetrics
model
– the limitation of RiskMetrics model and alternative
market risk measurement, such as back simulation
approach and Monte Carlo simulation approach.
– Measuring market risk using the Expected Shortfall
(ES) method

MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model

• Essential idea
– To estimate a single dollar number that tells the
market risk exposure over the next days if those
days turn out to be extremely bad days
– How to preserve equity if market condition moves
adversely tomorrow
• Implementation of the idea
– To define the criteria for adverse market
circumstances, for example, the 5% worst
scenarios, i.e., those worst scenarios which could
happen with a low probability (like, 5% here).
– To assume a distribution of future returns, based
on which the potential return under the adverse
scenarios could be estimated.
MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model

• Daily earnings at risk = (dollar market value of the position) x

(price sensitivity of the position) x(potential adverse move in
yield)
• DEAR = dollar market value of position × price volatility
where,
price volatility = price sensitivity of position ×
potential adverse move in yield
• Interpretation of DEAR
– the actual loss could go beyond this number
– Exact meaning: In the adverse market scenarios, the loss
could be equal to or above DEAR, and such case could
happen with (let assume) 5% chance.
– Alternatively, DEAR is the minimum loss you should
expect under the adverse market scenarios (5% worst case).

MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model: Confidence Intervals

• If we assume that changes in the yield are normally

distributed, we can construct confidence intervals around the
projected DEAR
• Assuming normality, 90% of the area under normal distribution
to be found within ±1.65 standard deviations of the mean
(1.65sd)
– (5% of the extreme values remain in either tail of the
distribution)
• Assuming normality, 95% of the area under normal distribution
to be found within ±1.96 standard deviations of the mean
(1.96sd)
– (2.5% of the extreme values remain in either tail of the
distribution)
• Assuming normality, 98% of the area under normal distribution
to be found within ±2.33 standard deviations of the mean
(2.33sd)
– (1% of the extreme values remain in either tail of the
distribution)
MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model: Confidence Intervals

• Suppose, during the last year, the mean change in daily yields
on 7-year zero coupon bonds was 0%. Standard deviation was
10 basis points.
• You are long in bonds and thus, we will lose if yield increases.
• We assume, changes in yields are normally distributed and we
define bad yield changes such that there is only 1% chance that
the next day’s yield will move adversely.
• 98% of the area under normal distribution to be found within
±2.33 standard deviations of the mean (2.33sd)
– Over the last year, daily yields fluctuated by more than
23.3 basis points (2.33 sd =2.33 x 10) 2 percent of the
time.
– Adverse moves in yields are those that decrease the value of
the security and occurred 1 % of the time.

MONASH
BUSINESS
SCHOOL
The RiskMetrics Model: Confidence Intervals

MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model: Fixed Income security

• Suppose, during the last year, the mean change in daily yields
on 7-year zero coupon bonds was 0%. Standard deviation was
10 basis points. Face value = $1,631,483
• Dollar market value position = $1 m
• Current yield = 7.243% per year
• $1,631,483/(1.07243)^7 = $1m
Daily price volatility = price sensitivity to small change in
yield × adverse daily move in yield
=MD X adverse daily move in yield
=(D/1+R) x 2.33 x 0.001
=(7/1.7243) x 0.00233 = 0.01521
• DEAR = $1m x 0.01521 =$15,210
• If the 1 bad day in 100 occurs tomorrow (1%), the potential
daily loss in earnings on $1m position = $15,210

MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model: Fixed Income security

N-day market value at risk (VAR) = DEAR X

• 5-day market value at risk (VAR) = DEAR X

• Implications: Due market illiquidity, you need to hold bolds for

longer (let assume, 5 days) which will intensify your loss.

MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model: Foreign Exchange Contracts

• Trading position = €800,000

• Exchange rate = €0.8000/$1 = or $1.25/€
• Dollar value position = €800,000 x $1.25/€ =$1m
• Last year, the standard deviation in the spot exchange rate
was 56.5 bp.
• FI is interested in the adverse move that will not take more
than 1% of the time.
• 98% of the area under normal distribution to be found within
±2.33 standard deviations of the mean (2.33sd)
• FX volatility = 2.33 x 56.5 bp = 0.0131645
• DEAR = Dollar value position x FX volatility = 1 m x
0.0131645= $13,164.5

MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model: Equities

• Total risk = systematic risk + unsystematic risk

•
• Beta = systematic risk reflecting the co-movement
of the returns on a specific stocks with returns on
market portfolio
• = volatility of market portfolio
• = unsystematic risk> diversified away in a well
diversified portfolio

MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model: Equities

• Trading position = $1m

• Beta =1; over the last year, the standard deviation
of market index = 200bp
• FI is interested about 1% worst case loss.
• Stock market return volatility = Beta x 2.33sd=1 x
2.33x 200bp = 0.0466
• DEAR = Dollar value position x stock market
volatility =$1 x 0.0466 = $46,600

MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model: Portfolio aggregation

• Three assets case:

DEARp = [DEAR12 + DEAR22 + DEAR32 + 2r1.2 × DEAR1 ×
DEAR2 + 2r1,3 × DEAR1 × DEAR3 + 2r2,3 × DEAR2 × DEAR3]1/2
– DEAR of fixed income security (1) = $15,210
– DEAR of foreign currency (2) = $13,164
– DEAR of equities (3)= $46,600
– R1,2 = -0.2, r1,3 =0.4, r2,3 =0.1
– DEARp = [15,2102 + 13,1642 + 46,6002 + 2 ×(-.2) x 15,210 ×
13,164 + 2 × 0.4 x 15,210 × 46,600 + 2 × 0.1 x 13,164 ×
46,600]1/2 = $56,443
– Only if returns on component assets are perfectly correlated, ρi,j =
1 (and daily returns of each component asset follow a normal
distribution), then DEARs are additive, i.e., portfolio DEAR =
sum of individual asset positions’ DEARs.

MONASH
BUSINESS
SCHOOL
 The RiskMetrics Model
• Hedging:
– Well-diversified portfolio will reduce DEAR for equities as
well as the aggregate portfolio.
– Review the limits of trading if DEARs are very high and
thus, could be a threat to a solvency of FI
• Our discussions so far focus on the long positions, for example,
you are holding $1 million of shares, and thus the risk/loss
comes from negative returns (or returns on the left tail of the
distribution). If you are short selling a security, you are
concerned with the possibility that the security price will go
up.
• Sign of DEAR: DEAR refers to potential loss and it is
normally negative. So, in lots of scenario, we simply ignore the
negative sign because we just want to measure the magnitude
of the exposure.

MONASH
BUSINESS
SCHOOL
 Learning Objectives
• Our discussion will focus on
– how market risk arises and how it can threaten the
solvency of FIs.
– how to measure market risk based on RiskMetrics
model
– the limitation of RiskMetrics model and alternative
market risk measurement, such as back simulation
approach and Monte Carlo simulation approach.
– Measuring market risk using the Expected Shortfall
(ES) method

MONASH
BUSINESS
SCHOOL
 Limitation of RiskMetric method

• Assumptions about return distributions

• Have to make an assumption that asset returns are
normally distributed
• Problems with estimating mean and standard deviation of
returns
• For portfolios, problems with estimating correlation
coefficient

• Alternatives
• Historic (back) simulation
• Monte Carlo simulation

MONASH
BUSINESS
SCHOOL
 The Historic or Back Simulation Approach

• The essence of DEAR calculation: portfolio value * the

hypothetical return under adverse circumstances, for example, the
case for which the worse scenarios could have only 1% (let assume)
to occur.
• Historic or back simulation approach: get the adverse return from
the past actual returns. Rank adverse return from worst to best.
• For example, previous 500 days, 1% worst-case means 5th lowest
value out of 500. let assume, the 5th lowest value of portfolio is -
$92,210. We will say, there is 1% chance that you will lose $92,210
if tomorrow is a worst trading day.

MONASH
BUSINESS
SCHOOL
 The Historic or Back Simulation Approach

Advantages:
– Simplicity,
– No need for distribution assumption,
– No necessity to calculate correlations or standard deviations of
asset returns.
Weaknesses:
• How far should we go back in the history to estimate the return
under adverse market circumstances
• More observations used, more reliable is the estimate
• But, returns in the distant past may be irrelevant
• Hence a tradeoff between reliability and relevance of
estimates, and normally a lack of (relevant) historical data
• A common approach is to weight recent observations more
heavily and go further back.
• .
MONASH
BUSINESS
SCHOOL
 Monte Carlo Simulation

• To overcome problem of limited number of observations, one

possible solution is to simulate additional observations
• Employ historic covariance matrix and random number generator
to synthesize observations
• From the large number distribution generated by Monte Carlo
Simulation, we can calculate what is VAR if there is 1% chance
that tomorrow will be the worst trading day.

MONASH
BUSINESS
SCHOOL
 Learning Objectives
• Our discussion will focus on
– how market risk arises and how it can threaten the
solvency of FIs.
– how to measure market risk based on RiskMetrics
model
– the limitation of RiskMetrics model and alternative
market risk measurement, such as back simulation
approach and Monte Carlo simulation approach.
– Measuring market risk using the Expected Shortfall
(ES) method

MONASH
BUSINESS
SCHOOL
 VAR

• Recall the exact meaning of VAR:

• In the adverse market scenarios, the loss could be equal to or above
VAR, and such case could happen with 1% chance (if we use 1%
worst case as the criteria for adverse market circumstances).
• Under adverse market scenarios, the actual loss could go beyond
VAR.
• Hence VAR is the minimum loss you should expect under the
adverse market scenarios – it is just a point estimate to indicate
the lower boundary for the potential loss range under adverse
scenarios.
• VAR does not provide information about the potential size of the
loss that exceeds it. VAR ignores the pattern and the severity of
the losses in the extreme tail.
• How to develop a measure to take into account all possible losses in
the whole worst market scenarios?

MONASH
BUSINESS
SCHOOL
 Expected Shortfall (ES)

• ES is a measure of market risk that estimates the expected value of

losses beyond a given confidence level-the average of VARs beyond
a given confidence level.
• For a given confidence level, say 99 percent, we measure the area
under the probability distribution from 99th to 100th percentile.
• ES analyzes the size and likelihood of losses above certain
percentile (say 99th percentile) in a crisis period for a traded asset
and thus measure “tail-risk” more precisely.
• It is a risk measure that considers a more comprehensive set of
potential outcomes than VAR.
• Basel III proposes to replace VAR models with those based on
extreme value theory and Expected shortfall (ES).

MONASH
BUSINESS
SCHOOL
 Expected Shortfall (ES)

• Consider the following discrete probability distribution of payoffs of

two securities A and B, held in trading portfolio of an FI.

• Expected return is $80 m for both A and B.

• For 99% Confidence level, VAR(A)=VAR(B)=-$920m
• Yet, for 99% Confidence Level, ESA=-$920m
• However, ESB=0.25*(-$920m)+0.75*(-$1704m)= -$1508m
• While VAR is identical for both securities, the ES suggests that
security B has the potential to subject the FI to much greater
losses than security A.

MONASH
BUSINESS
SCHOOL
 Conclusion
• How market risk arises and how it can threaten the
solvency of FIs.
• How to measure market risk based on RiskMetrics
model
• The limitation of RiskMetrics model and alternative
market risk measurement, such as back simulation
approach and Monte Carlo simulation approach.
• Measuring market risk using the Expected Shortfall
(ES) method

MONASH
BUSINESS
SCHOOL