Scribd
Upload
36 views4 pages

Financial Risk Measurement and Management: Method Ii: The Riskmetrics Approach

This document discusses the RiskMetrics approach for measuring and forecasting financial risk. It describes how RiskMetrics uses an exponentially weighted moving average to assign decreasing weights over time to past covariance and correlation values, allowing the forecasts to adapt more quickly to changing market conditions. The document provides an example of how to calculate value at risk (VaR) for a portfolio using the RiskMetrics methodology and compares the resulting VaR forecasts to actual portfolio returns.

Uploaded by

Pratik Jain
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
36 views4 pages

Financial Risk Measurement and Management: Method Ii: The Riskmetrics Approach

This document discusses the RiskMetrics approach for measuring and forecasting financial risk. It describes how RiskMetrics uses an exponentially weighted moving average to assign decreasing weights over time to past covariance and correlation values, allowing the forecasts to adapt more quickly to changing market conditions. The document provides an example of how to calculate value at risk (VaR) for a portfolio using the RiskMetrics methodology and compares the resulting VaR forecasts to actual portfolio returns.

Uploaded by

Pratik Jain
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

12/2/2020

Financial Risk Measurement and Management


Session 11

Dilip Kumar 1

Method II: The RiskMetrics Approach


If we use a rolling window (say m-day) to forecast variance, then
1
= −

If we assume the mean term to be zero, then = ∑


How do we estimate the covariance of two returns?
1
, = , − , −

Assuming means to be zero, it can be written as:


1
, = , ,

Dilip Kumar 2

1
12/2/2020

Method II: The RiskMetrics Approach


What we are implicitly doing is assign equal weights to all terms,
irrespective of whether they happened yesterday or 252 days before!
As should be evident , this shows up as the “staleness” in the covariance forecast
1
, = , ,

Instead of assigning equal weights, we can assign exponentially


decreasing weights as below (like in the RiskMetrics methodology for
variance forecast)

, = 1−  , ,

The above expression can be simplified to


, = , + 1− , ,
Dilip Kumar 3

Method II: The RiskMetrics Approach


As on 3 Jan 2005, you have the following portfolio:
Infosys SBI
Spot Price 264.94 61.86
Units 100 428
The position has been chosen so that both stocks have roughly the
same weight (0.5002, 0.4998) in the portfolio.
Evaluate the 1-day 95% VaR for the portfolio using RiskMetrics
like forecast for covariance; assume  = 0.94.

Dilip Kumar 4

2
12/2/2020

Method II: The RiskMetrics Approach


Step-by-step approach for calculating the VaR of a portfolio
For each stock, calculate the daily log returns for the entire period
For each stock, create a series of RiskMetrics variance forecast
Next, create a series of RiskMetrics covariance forecast
Calculate the volatility forecast for the portfolio using the expression
= + +2
Estimate VaR for portfolio using
, = 1.645 × , × ,
Estimate the breach percentage.

Dilip Kumar 5

Method II: The RiskMetrics Approach


Before we examine the performance of VaR, let us examine how well the
correlation adapts itself to changing market conditions
1400 1

1200 0.8
0.6
1000
0.4
Correlation

800
Price

0.2
600
0
400
-0.2
200 -0.4
0 -0.6
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
SBI Infosys 252D Correlation RM Correlation

Dilip Kumar 6

3
12/2/2020

Method II: The RiskMetrics Approach


Below is a plot of VaR and actual P&L
20000
15000
10000
5000
0
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
-5000
-10000
-15000
-20000
RM VaR Actual PnL

There were total 154 breaches , accounting for a healthy 5.3%


breaches.
Dilip Kumar 7

Method II: The RiskMetrics Approach


Before drawing the curtains on this approach, I would like you to
observe the following graph
1 50.000%
0.8
%age Reduction in VaR

40.000%
0.6
Correlation

0.4 30.000%
0.2
0 20.000%

-0.22006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
10.000%
-0.4
-0.6 0.000%
Correlation %age Reduction in VaR
The graph is self-explanatory; what can you infer from the graph?
What does reduction in VaR refer to?
Dilip Kumar 8

You might also like