1

Volatility Modeling
Using Daily Data

Elements of
Financial Risk Management
Chapter 4
Peter Christoffersen

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 2

Overview
• In this Chapter, we will proceed with the univariate
models in two steps.
• The first step is to establish a forecasting model for
dynamic portfolio variance and to introduce methods
for evaluating the performance of these forecasts.
• The second step is to consider ways to model
nonnormal aspects of the portfolio return - aspects that
are not captured by the dynamic variance.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 3

Overview
• We proceed as follows:
– We start with the simple variance forecasting and
the RiskMetrics variance model.
– We introduce the GARCH variance model and
compare it with the RiskMetrics model.
– We estimate the GARCH parameters using the
quasi-maximum likelihood method.
– We suggest extensions to the basic model
– We discuss various methods for evaluating the
volatility forecasting models.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 4

Simple Variance Forecasting

• We define the daily asset log-return, , using the
daily closing price, ,as

• can refer to an individual asset return or a

portfolio return.
• Based on findings of Chapter 1, we assume for
short horizons the mean value of is zero.
• Furthermore, we assume that the innovation to
asset return is normally distributed, i.e.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 5

Simple Variance Forecasting

• Where i.i.d. N(0,1) stands for “independently and
identically normally distributed with mean equal to
zero and variance equal to 1.”
• Note that the normality assumption is not realistic.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 6

Simple Variance Forecasting

• Variance, as measured by squared returns, exhibits
strong autocorrelation
• If the recent period was one of high variance, then
tomorrow is likely to be a high-variance day as
well.
• Tomorrow’s variance is given by the simple average
of the most recent m observations :

• However model puts equal weights on the past m

observations yielding unwarranted results
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 7
Figure 4.1:
Squared S&P 500 Returns with Moving Average Variance
Estimated on past 25 observations. 2008-2009

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 8

Simple Variance Forecasting

• In RiskMetrics system, the weights on past squared
returns decline exponentially as we move backward
in time.
• JP Morgan’s RiskMetrics variance model or the
exponential smoother is given by:

• Separating from the sum the squared return for ,

where , we get

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 9

Simple Variance Forecasting

• Applying the exponential smoothing definition
again we can write today’s variance, , as

• So that tomorrow’s variance can be written

• The RiskMetrics model’s forecast for tomorrow’s

volatility can thus be seen as weighted average of
today’s volatility and today’s squared return.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 10

Advantages of RiskMetrics
• It tracks variance changes in a way which is
broadly consistent with observed returns. Recent
returns matter more for tomorrow’s variance than
distant returns.
• It contains only one unknown parameter.
• When estimating  on a large number of assets,
Riskmetrics found that the estimates were quite
similar across assets and they therefore simply set
 0.94 . for every asset for daily
variance forecasting.
• In this case, no estimation is necessary, which is a
huge advantage in large portfolios.
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 11

Advantages of RiskMetrics
• Relatively little data needs to be stored in order to
calculate tomorrow’s variance.
• The weight on today’s squared returns is
(1   ) 0.06 and the weight is exponentially
decaying to on (1 the )100
99 thlag of squared return. After
0.000131
including 100 lags of squared returns the cumulated
weight is
100
(1   )   1 0.998
 1
• We only need about 100 daily lags of returns in order
to calculate tomorrow’s variance 2
,  t 1 .
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 12

Advantages of RiskMetrics
• Despite these advantages, RiskMetrics does have
certain shortcomings which motivates us to consider
slightly more elaborate models.
• For example, it does not allow for a leverage effect
and it also provides counterfactual longer-horizon
forecasts.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 13

The GARCH Variance Model

• This model can capture important features of returns data
and are flexible enough to accommodate specific aspects
of individual assets.
• The downside of the following models is that they require
nonlinear parameter estimation
• The simplest generalized autoregressive conditional
hetroskedasticity (GARCH) model of dynamic variance
can be written as,

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 14

The GARCH Variance Model

• The RiskMetrics model can be viewed as a special
case of the simple GARCH model where
 1   ,   , s.t.    1,  0
• However there is an important difference : We can
define the unconditional, or long-run average,
variance to be

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 15

The GARCH Variance Model

• If    1 as is the case in RiskMetrics, then the long-
run variance is not well-defined in that model.
• Thus an important quirk of the RiskMetrics model is that
it ignores the fact that the long-run average variance
tends to be relative stable over time.
• The GARCH model implicitly relies on 
• By solving for  in the long-run variance equation and
substitute it into the dynamic variance equation, we get :

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 16

The GARCH Variance Model

• Thus tomorrow’s variance is a weighted average
of the long-run variance, today’s squared return
and today’s variance.
• Ignoring the long-run variance is more important
for longer-horizon forecasting than for forecasting
simply one-day ahead
• A key advantage of GARCH models for risk
management is that the one-day forecast of
variance ,  2
is1|tgiven directly by the model as
t

 t21

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 17

The GARCH Variance Model

• Consider forecasting the variance of the daily
return k days ahead; the expected value of future
variance at horizon k is

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 18

The GARCH Variance Model

• The conditional expectation refers to taking the
expectation using all the information available at the
end of day t, which includes the squared return on day t
itself.
•     is the persistence.
• A high persistence -     close to 1- implies that
shocks that push variance away from its longrun
average will persist for a long time
• Similar calculations for RiskMetrics model reveal

• as  and  is undefined

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 19

The GARCH Variance Model

• Thus, persistence in this model is 1, which implies that
a shock to variance persists forever
• An increase in variance will push up the variance
forecast by an identical amount for all future forecast
horizons.
• RiskMetrics model ignores the long-run variance when
forecasting.
• If  is close to one, then the two models might
yield similar predictions for short horizons, k, but their
longer horizon implications are very different.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 20

The GARCH Variance Model

• If today is a low-variance day then RiskMetrics model
predicts that all future days will be low variance.
• The GARCH model assumes that eventually in the future
variance will revert to the average value
• The forecast of variance of K-day cumulative returns

• We assume that returns have zero autocorrelation, then the

variance is simply

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 21

The GARCH Variance Model

• So, in the RiskMetrics model we get

• But in GARCH model, we get

• If the RiskMetrics and GARCH model have

 2
identical t 1 , and if,  2
t 1   2

then the
GARCH variance forecast will be higher than the
RiskMetrics forecast.
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 22

The GARCH Variance Model

• Assuming Riskmetrics model, if the data looks more like
GARCH will give risk managers a false sense of the
calmness of the market in the future, when the market is
calm today and t 1  2
  2
.
• Fig.4.2 illustrates this crucial point.
• We plot  t 1:t  K K for K 1,2,...,250
2
for both the
RiskMetrics and the GARCH model starting from a low
2t+1 and setting  0.05 and  0.90.
• The long run variance in the figure is = 0.000140

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 23
Figure 4.2:Variance Forecast for 1-250 Days
Cumulative Returns

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 24

The GARCH Variance Model

• An inconvenience shared by the two models is that the
multi-period distribution is unknown even if the one-day
ahead distribution is assumed to be normal

• Thus while it is easy to forecast longer-horizon variance

in these models, it is not as easy to forecast the entire
conditional distribution.

• This issue will be further analyzed in Chap. 8

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 25

Maximum likelihood Estimation

• GARCH model contain a number of unknown
parameters that must be estimated.

• The conditional variance is an unobserved

variable, which must be implicitly estimated along
with the parameters of the model.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 Standard Maximum Likelihood
26

Estimation
• MLE can be used to find parameter values
• Recall the assumption that
Rt  t zt , with zt ~ i.i.d .N 0,1
• The assumption of i.i.d. normality implies that the
probability, or the likelihood, lt, of Rt is
1  Rt2 
lt  exp  
 2 t 
2
2 t2
• Thus the joint likelihood of our entire sample is

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 Standard Maximum Likelihood
27

Estimation
• Choose parameters  ,  ,... to maximize
the joint log likelihood of our observed sample

• First term in the likelihood function is a constant and so

independent of the parameters of the models.
• We can therefore equally well optimize

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 Standard Maximum Likelihood
28

Estimation

• The MLE approach has the desirable property that as

the sample size, T, goes to infinity the parameter
estimates converge to their true values.
• MLE gives the smallest variance for the estimates.
• In reality we don’t have infinite history of past data
• We may also have structural breaks.
• A good general rule of thumb is to use 1,000 daily
observations when estimating GARCH.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 Quasi Maximum Likelihood 29

Estimation
• One may argue that the MLEs rely on the conditional
normal distribution assumption which we argued in
chapter 1 is false.
• A key result in econometrics says that even if the
conditional distribution is not normal, MLE will yield
estimates of the mean and variance parameters which
converge to the true parameters as the sample gets
infinitely large, as long as mean and variance functions are
properly specified.
• This establishes the quasi maximum likelihood estimation
or QMLE, referring to the use of normal-MLE estimation
even when the normal distribution assumption is false .
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 30

Quasi Maximum Likelihood Estimation

• The QMLE estimates will in general be less

precise than those from MLE.
• Thus we trade off theoretical asymptotic
parameter efficiency for practicality.
• A simple trick than can be used in estimations is
variance targeting.
• Recall the simple GARCH model

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 Quasi Maximum Likelihood
31

Estimation
• Thus instead of estimating by MLE, we simply set the
long-run variance,  ,equal to the sample variance

• Variance targeting imposes the long-run variance on

the GARCH mode and reduces the number of
parameters to be estimated in the model

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 32

An Example
• Fig. 4.3 shows the S&P500 squared returns from
Fig. 4.1 but with an estimated GARCH variance
superimposed

• Using numerical optimization of the likelihood

function, the optimal parameters imply the
following variance dynamics:

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 33

An Example
• The persistence of variance in this model is
+=0.999, which is only slightly lower than in
RiskMetrics where it is 1
• However, even if small, this difference will have
consequences for the variance forecasts for
horizons beyond one day
• Furthermore, this very simple GARCH model may
be misspecified driving the persistence close to one
• So we consider more flexible models next

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 34

Figure 4.3: Squared S&P 500 Returns with GARCH

Variance Parameters Are Estimated Using QMLE

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 35

The Leverage Effect

• A negative return increases variance by more than
a positive return of the same magnitude
• This is referred to as the leverage effect
• We modify the GARCH models so that the weight
given to the return depends on whether it is
positive or negative, as follows:

• which is sometimes referred to as the NGARCH

(Nonlinear GARCH) model
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 36

The Leverage Effect

• The persistence of variance in this model is
and the long-run variance is:

• Another way of capturing the leverage effect is to define

an indicator variable, It, to take on the value 1 if day t’s
return is negative and zero otherwise

• The variance dynamics can now be specified as

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 37

The Leverage Effect

• Thus,  > 0 will capture the leverage effect.
• This is referred to as the GJR-GARCH model.
• A different model that also captures the leverage is the
exponential GARCH model or EGARCH
ln  t21    Rt   Rt  E Rt   ln  t2

which displays the usual leverage effect if < 0

• EGARCH model - Advantage : the log.specification
ensures a positive variance
• Disadvantage : future expected variance beyond one
period cannot be calculated analytically.
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 38

More General News Impact Functions

• Variance news impact function, NIF is the relationship in
which today’s shock to return, zt, impacts tomorrow’s
variance 2t+1
• In general we can write

• In the simple GARCH model we have

NIF (zt) = z2t
• so that the NIF is a symmetric parabola that takes the
minimum value 0 when zt is zero
• In the NGARCH model with leverage we have

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 39

More General News Impact Functions

• so that the NIF is still a parabola but now with the
minimum value zero when zt = 
• A very general NIF can be defined by

• The simple GARCH model is nested when

, , and .
• The NGARCH model with leverage is nested
when , and .

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 40

Figure 4.4: News Impact Function

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 41

More General Dynamics

• A simple GARCH model GARCH(1,1) relies on
only one lag of returns squared and one lag of
variance.
• Higher order dynamics is made possible through
GARCH(p,q) which allows for longer lags as
follows:

• The disadvantage of this more generalized models

is that the parameters are not easily interpretable.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 42

More General Dynamics

• The component GARCH structure helps to interpret
the parameters easily
• Using we can rewrite the
GARCH(1,1) model as

• In the component GARCH model the long-run

variance, s2, is allowed to be time varying and
captured by the long-run variance factor vt+1:

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 43

More General Dynamics

• Note that the dynamic long-term variance, vt+1, itself has
a GARCH(1,1) structure.
• Thus, a component GARCH model is a GARCH(1,1)
model around another GARCH(1,1) model.
• The component model can potentially capture
autocorrelation patterns in variance
• The component model can be rewritten as a GARCH(2,2)
model as

where the parameters in the GARCH(2,2) are functions of

the parameters in the component GARCH model
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 44

More General Dynamics

• The component GARCH structure has the advantage that
it is easier to interpret its parameters and therefore easier
to come up with good starting values for the parameters
than in the GARCH(2,2) model
• In the component model + capture the persistence
of the short-run variance component and vv capture
the persistence in the long-run variance component.
• The GARCH(2,2) dynamic parameters 
have no such straightforward interpretation

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 45

Explanatory Variables
• In dynamic models of daily variance, we need to
account for days with no trading activity
• Days that follow a weekend or a holiday have
higher variance than average days
• As these days are perfectly predictable, we need to
include them in the variance model
• So, we can model this by:

• where ITt+1 takes on the value 1 if date t+1 is a

Monday, for example
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 46

Explanatory Variables
• In general, we can write the GARCH variance forecasting
model as follows:

• where Xt denotes variables known at the end of day t

• As the variance is always a positive number, the GARCH
model should always generates a positive variance
forecast
• In the above model, positivity of h(Xt) along with positive
 and  will ensure positivity of 2t+1
• We can write

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 47

Generalizing the Low Frequency Dynamics

• Volatility often spikes up for a few days and then

quickly reverts back down to normal levels.
• Such quickly reverting spikes make volatility appear
noisy and thus difficult to capture by explanatory
variables.
• Explanatory variables are important for capturing
longer-term trends in variance, which need to be
modeled separately so as to not be contaminated by the
daily spikes.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 48

Generalizing the Low Frequency Dynamics

• In order to capture low-frequency changes in
volatility we generalize the simple GARCH(1,1)
model to the following multiplicative structure

• The Spline-GARCH model captures low frequency

dynamics in variance via the t+1 process, and
higher-frequency dynamics in variance via the gt+1
process.
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 49

Generalizing the Low Frequency Dynamics

• Low-frequency variance is kept positive via the

exponential function

• The low frequency variance has a log linear time-trend

captured by 1 and a quadratic time-trend starting at
time t0 and captured by 2

• The low-frequency variance is also driven by the

explanatory variables in the vector Xt.
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 50

Generalizing the Low Frequency Dynamics

• The long-run variance in the Spline-GARCH model is
captured by the low-frequency process

• We can generalize the quadratic trend by allowing for

many, say l, quadratic pieces, each starting at different
time points and each with different slope parameters:

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 51

Estimation of Extended Models

• GARCH family of models can all be estimated using
the same quasi MLE technique used for the simple
GARCH(1,1) model.
• The model parameters can be estimated by maximizing
the nontrivial part of the log likelihood

• The variance path, 2t , is a function of the

parameters to be estimated.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 52

Model Comparison using LR Tests

• Basic GARCH model can be extended by adding
parameters and explanatory variables
• The likelihood ratio test provides a simple way to judge
if the added parameter(s) are significant in the statistical
sense.
• Consider two different models with likelihood values L0
and L1, respectively.
• Assume that model 0 is a special case of model 1
• In this case we can compare the two models via the
likelihood ratio statistic

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 53

Model Comparison using LR Tests

• The LR statistic will be a positive number because model
1 contains model 0 as a special case and so model 1 will
always fit the data better

• The LR statistic tells us if the improvement offered by

model 1 over model 0 is statistically significant

• It can be shown that the LR statistic will have a chi-

squared distribution under the null hypothesis that the
added parameters in model 1 are insignificant.
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 54

Model Comparison using LR Tests

• If only one parameter is added then the degree of freedom
in the chi-squared distribution will be 1

• A good rule of thumb is that if the log-likelihood of model

1 is 3 to 4 points higher than that of model 0 then the
added parameter in model 1 is significant

• The degrees of freedom in the chi-squared test is equal to

the number of parameters added in model 1

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 55

Diagnostic Check on Autocorrelations

• In Chapter 1 we saw that the raw return auto-correlations
didn’t display any systematic patterns
• The squared return autocorrelations is positive for short
lags and decreases as the lag order increases
• We use variance modelling to construct 2t which has the
property that standardized squared returns, R2t / 2t have
no systematic autocorrelation patterns
• The red line in Figure 4.5, show the autocorrelation of
R2t / 2t from the GARCH model with leverage for the
S&P 500 returns along with their standard error bands.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 56

Diagnostic Check on Autocorrelations

• The standard errors are calculated simply as,
where T is the number of observations in the sample.
• Autocorrelation is shown along with plus/minus two
standard error bands around zero, which mean
horizontal lines at and
• These Bartlett standard error bands give the range in
which the autocorrelations would fall roughly 95% of
the time if the true but unknown autocorrelations of
R2t / 2t were all zero.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 57
Figure 4.5: Autocorrelation: Squared Returns and
Squared Returns over Variance

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 58
Volatility Forecast Evaluating Using
Regression
• A variance model can be evaluated based on simple
regressions where squared returns in the forecast period,
t+1, are regressed on the forecast from the variance
model, as in

• A good variance forecast should be unbiased, that is,

have an intercept b0 = 0, and be efficient, that is, have a
slope, b1 = 1.
• Note that so that the squared return is an
unbiased proxy for true variance.
Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen
 59
Volatility Forecast Evaluating Using
Regression
• But the variance of the proxy is

• where  is the kurtosis of the innovation

• Due to the high degree of noise in the squared returns,
the regression R2 will be very low, typically around 5%
to 10%
• The conclusion is that the proxy for true but unobserved
variance is simply very inaccurate.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 60

The Volatility Forecast Loss Function

• The ordinary least squares estimation of a linear
regression chooses the parameter values that minimize the
mean squared error in the regression
• The regression-based approach to volatility forecast
evaluation therefore implies a quadratic volatility forecast
loss function
• A correct volatility forecasting model should have b0 = 0
and b1 = 1 as discussed earlier
• Loss function to compare volatility models is therefore

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 61

The Volatility Forecast Loss Function

• In order to evaluate volatility forecasts allowing for
asymmetric loss, the following function can be used
instead of MSE

• QLIKE loss function depends on the relative volatility

forecast error, , rather than on the absolute error,
. ; which is the key ingredient in MSE
• The QLIKE loss function will always penalize more
heavily volatility forecasts that underestimate volatility

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 62
Figure 4.6: Volatility Loss Function

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen

 63

Summary
• In this Chapter we have
– Discussed the simple variance forecasting and the
RiskMetrics variance model.
– Introduced the GARCH variance model and
compare it with the RiskMetrics model.
– Estimated the GARCH parameters using the quasi-
maximum likelihood method.
– Suggested extensions to the basic model
– Discussed various methods for evaluating the
volatility forecasting models.

Elements of Financial Risk Management Second Edition © 2012 by Peter Christoffersen