Estimating Volatilities and Correlations, Chapter 15
Download
Report
Transcript Estimating Volatilities and Correlations, Chapter 15
Estimating
Volatilities and
Correlations
Chapter 21
2015/7/17
1
Standard Approach to
Estimating Volatility
Define sn as the volatility per day on day n, as
estimated at end of day n-1
Define Si as the value of market variable at
end of day i
Define ui= ln(Si/Si-1), an unbiased estimate is
m
1
2
s n2
(
u
u
)
m 1 i 1 n i
1 m
u un i
m i 1
2015/7/17
2
Simplifications Usually Made
Define ui as (Si-Si-1)/Si-1
Assume that the mean value of ui is zero
Replace m-1 by m
This gives
1 m 2
s i 1 un i
m
2
n
2015/7/17
3
Weighting Scheme
Instead of assigning equal weights to the
observations we can set
s i 1 i u
m
2
n
2
n i
where
m
i 1
2015/7/17
i
1
4
ARCH(m) Model
In an ARCH(m) model we also assign some
weight to the long-run variance rate, V:
s V i 1 i u
m
2
n
2
n i
where
m
i 1
i 1
2015/7/17
5
EWMA Model
In an exponentially weighted moving
average model, the weights assigned to the
u2 decline exponentially as we move back
through time.
2015/7/17
6
Especially, if
i 1 i
1 1 1 2
s (1 )(u
2
n
2
n 1
u
1, 1 (1 )
2
n2
u
2
2
n 3
)
s n21 (1 )(un2 2 un23 2un2 4 )
s s
2
n
2
n 1
(1 )u
2
n 1
m
s n2 (1 ) i 1un2i ms 02
i 1
2015/7/17
7
Attractions of EWMA
Relatively little data needs to be stored
We need only remember the current
estimate of the variance rate and the most
recent observation on the market variable
Tracks volatility changes
JP Morgan use = 0.94 for daily volatility
forecasting
2015/7/17
8
GARCH (1,1)
In GARCH (1,1) we assign some weight to
the long-run average variance rate
s V s
2
n
2
n 1
bu
2
n 1
Since weights must sum to 1
b 1
2015/7/17
9
GARCH (1,1) continued
Setting w V the GARCH (1,1) model is
s w s
and
2
n
2
n 1
bu
2
n 1
w
V
1 b
2015/7/17
10
Example
Suppose
s 0.000002 013
. u
2
n
2
n 1
086
. s
2
n 1
the long-run variance rate is 0.0002 so that
the long-run volatility per day is 1.4%
2015/7/17
11
Example continued
Suppose that the current estimate of the
volatility is 1.6% per day and the most
recent proportional change in the market
variable is 1%.
The new variance rate is
0000002
.
013
. 00001
.
086
. 0000256
.
000023336
.
The new volatility is 1.53% per day
2015/7/17
12
GARCH (p,q)
p
q
i 1
j 1
s n2 w i un2i b j s n2 j
2015/7/17
13
Mean Reversion
The GARCH(1,1) is equivalent to a model
where the variance V follows the
stochastic process
dV a(VL V )dt Vdz
GARCH(1,1) incorporates mean reversion,
EWMA does not. When w is negative,
GARCH(1,1) is not stable, and we should
use EWMA.
2015/7/17
14
Other Models
We can design GARCH models so that the
weight given to ui2 depends on whether ui is
positive or negative.
2015/7/17
15
Maximum Likelihood Methods
In maximum likelihood methods we choose
parameters that maximize the likelihood of
the observations occurring
2015/7/17
16
Example 1
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
The probability of the outcome is
9
p(1 p)
We maximize this to obtain a maximum
likelihood estimate: p=0.1
2015/7/17
17
Example 2(Suppose the variance is
constant)
Estimate the variance of observations from a
normal distribution with mean zero
1
ui2
exp
2v
i 1 2 v
n
Maximize:
ui2
ln(v ) v
i 1
1 n 2
v ui
n i 1
n
or:
This gives:
2015/7/17
18
Application to GARCH
We choose parameters that maximize
u
ln(vi )
vi
i 1
n
2015/7/17
2
i
19
Excel Application (Table 21.1, page 477)
Start with trial values of w, , and b
Update variances
Calculate
ui2
ln( vi )
vi
i 1
m
Use solver to search for values of w, , and b that
maximize this objective function
Important note: set up spreadsheet so that you are
searching for three numbers that are the same
order of magnitude (See page 478)
2015/7/17
20
Variance Targeting
One way of implementing GARCH(1,1) that
increases stability is by using variance
targeting
We set the long-run average volatility equal
to the sample variance
Only two other parameters then have to be
estimated.
2015/7/17
21
How Good is the Model?
We compare the autocorrelation of the ui’s
with the autocorrelation of the ui/si
The Ljung-Box statistic tests for
autocorrelation
2015/7/17
22
Forecasting Future Volatility
A few lines of algebra shows that
E[s n2k ] V ( b ) k (s n2 V )
The variance rate for an option expiring on
day m is
1 m1
2
E
s
n k
m k 0
2015/7/17
23
Forecasting Future Volatility
continued (equation 19.4, page 473)
Def ine
1
a ln
b
The volatility per annum f or a T - day option is
1 e aT
V (0) VL
252VL
aT
2015/7/17
24
Volatility Term Structures (Table 21.4)
The GARCH (1,1) suggests that, when calculating vega,
we should shift the long maturity volatilities less than the
short maturity volatilities
Impact of 1% change in instantaneous volatility for
Japanese yen example:
Option Life
(days)
10
30
50
100
500
Volatility
increase (%)
0.84
0.61
0.46
0.27
0.06
2015/7/17
25
Correlations
Define ui=(Ui-Ui-1)/Ui-1 and vi=(Vi-Vi-1)/Vi-1
Also
su,n: daily vol of U calculated on day n-1
sv,n: daily vol of V calculated on day n-1
covn: covariance calculated on day n-1
2015/7/17
26
Correlations continued
Under GARCH (1,1)
covn = w + un-1vn-1+b covn-1
2015/7/17
27
Positive semi-definite Condition
A variance-covariance matrix, W, is internally
consistent if the positive semi-definite
condition
w Ww 0
T
for all vectors w is satisfied.
2015/7/17
28
Example
The variance covariance matrix
1
0
0.9
0
1
0.9
0.9
0.9
1
is not internally consistent
2015/7/17
29