00:00:00
we are still on the Gaucho model and
00:00:02
series this is coach econometrics and it
00:00:04
is good to have you back in this video
00:00:07
I'll be estimating a guardian main
00:00:09
Moodle give you the intuition behind
00:00:13
where we estimates gotcha in main
00:00:15
modules but before I do so in my usual
00:00:17
way I will encourage you to please watch
00:00:19
these videos in sequential order please
00:00:22
do not skip any I really wants you to
00:00:25
understand how to estimate GARCH models
00:00:27
so let us get some intuition for
00:00:30
estimating a gosh M model remember that
00:00:34
risk-averse investors may require a
00:00:37
premium as a compensation for them to
00:00:40
hold a risky assets that premium is
00:00:43
clearly a function of the risk that is
00:00:47
the higher the risk the higher the
00:00:49
premium should be if the risk is now
00:00:52
captured by the volatility or by the
00:00:55
conditional variance then the
00:00:58
conditional variance may enter into the
00:01:00
conditional mean equation as specified
00:01:03
here so this is the conditional variance
00:01:05
in the main equation so the GOC M allows
00:01:12
the conditional mean to depend on its
00:01:15
own conditional variance its models a
00:01:18
time varying risk premium the same catch
00:01:22
em model can also use the standard
00:01:25
deviation of the series to capture the
00:01:27
risk so in the first one we have the
00:01:30
conditional variance and the second one
00:01:33
we have the standard deviation therefore
00:01:36
the GOC um PQ model can be generalized
00:01:41
towards you are seen on the screen so
00:01:44
the first two are the main equations for
00:01:47
the GOC M while this one this part this
00:01:50
last equation is the usual conditional
00:01:55
variance equation that we are used to
00:01:58
buy now so now let's proceed to give us
00:02:00
an estimate the Gotcha model using the
00:02:03
variance in the main equation then
00:02:06
secondly we use the standard deviation
00:02:08
and the main equation and we compare
00:02:11
your results I double
00:02:12
Kongo series I got a quick click on
00:02:15
estimates equation I list the series in
00:02:18
the usual form that we'll be using
00:02:21
I go to methods I change it to arch
00:02:24
please follow me and do likewise with
00:02:27
your data remember we are estimating a
00:02:31
catch
00:02:31
M model so we come to this box and we
00:02:34
change none we open it and we chain on
00:02:37
to variance so let's start with the
00:02:39
variance forced in the main equation so
00:02:42
we change this to variance we go to
00:02:45
options optimization method I've been
00:02:49
using a vyas legacy so I change this to
00:02:51
Eevee Isleta see I don't change any
00:02:53
other thing I go back to specification
00:02:56
everything looks fine here my sample
00:02:59
size is okay is the usual sample size I
00:03:02
click okay so here we can see the GOC m
00:03:05
model using the variance in the main
00:03:09
equation and the conditional variance is
00:03:12
captured by the gosh you can see here in
00:03:15
the main equation also do not confuse it
00:03:17
with the gosh -1 here they are not the
00:03:20
same
00:03:20
remember we estimating in catch a model
00:03:23
and what can you observe the coefficient
00:03:26
of the variance in the main equation is
00:03:29
not statistically significant but we can
00:03:33
say that by including it's the main
00:03:35
equation it has improved the
00:03:37
significance of the got charm in the
00:03:40
variance equation so this is what we can
00:03:42
contribute by using the variance in the
00:03:46
main equation probe value is 20 percents
00:03:49
or 21% clearly not significant by
00:03:53
including it has improved the Gaucho
00:03:56
term in the various equation now let's
00:04:00
use the standard deviation and see
00:04:03
whether we are going to have a different
00:04:05
outcome so we click on estimates the
00:04:07
modified variance now to standard
00:04:09
deviation we don't change anything we
00:04:12
click ok so here we have once you see a
00:04:15
square root gosh this is the standard
00:04:17
deviation so we can also see that the
00:04:20
standard deviation is clearly not
00:04:22
significant is 22.2%
00:04:25
it's not significant statistically by
00:04:29
including it has improved the Gaucho
00:04:31
term in the variance equation so the
00:04:37
findings is not different from what we
00:04:39
go from the DEA's so what do we conclude
00:04:42
as a financial analyst let's go back to
00:04:44
powerpoints for my explanation so this
00:04:47
is the specification I use to remember
00:04:49
to change the HM from non so variance
00:04:53
and my optimization method is a vyas
00:04:56
legacy so this is for the variance
00:04:58
equation and here is a result that we
00:05:02
got we saw that the conditional variance
00:05:05
term in the main equation is
00:05:07
statistically not significant so it's
00:05:09
one percent approximately here so what
00:05:12
do we conclude we can say that this
00:05:16
fixed premium is no significant to hedge
00:05:18
it gets holding a risky assets it's not
00:05:22
significant so we can say that the asset
00:05:25
in question may not be risky to hold so
00:05:30
as an investor if you are using the
00:05:32
variance to help you against holding the
00:05:35
risk you can clearly see that this asset
00:05:38
is not risky at all so you can hold it
00:05:41
the variance term which is a cash town
00:05:44
is clearly not significant and the main
00:05:47
equation again for the standard
00:05:49
deviation remember to modify the mbox to
00:05:52
reflect standard deviation I use a vyas
00:05:55
legacy as usual and our results also
00:05:57
shows that the standard deviation
00:05:59
coefficient is also not significant is
00:06:01
over 22 percents same conclusion we can
00:06:05
say that the risk premium is not
00:06:07
significant to hedge against holding
00:06:09
this risk and therefore we can say that
00:06:11
he asset is not risky to hold so this
00:06:14
video has summarized your basic
00:06:18
knowledge on how you can estimate in
00:06:20
kajam model using the conditional
00:06:22
variance on a min equation or using the
00:06:25
standard deviation in the main equation
00:06:27
check out the coefficient in the menu
00:06:30
question to see where a significant
00:06:32
enough to be a premium that will hurt
00:06:35
against
00:06:36
holding a risky asset in my situation
00:06:39
both coefficients are statistically not
00:06:42
significant so these are references are
00:06:44
readings to support what we have just
00:06:46
watched in relation to estimating a
00:06:48
gaucho a model kindly read of at least
00:06:51
one or two papers
00:06:52
it was strengthen your understanding of
00:06:54
GARCH models and video tutorials are
00:06:57
clearly not sufficient you have to read
00:06:59
thank you so much we have covered five
00:07:01
topics now as you can see on the screen
00:07:03
do not skip any keep yourself abreast of
00:07:06
GARCH models by watching these videos in
00:07:10
sequential order I am grateful to all
00:07:12
the comments I've received so far since
00:07:14
I began this gosh modeling series
00:07:16
they've been encouraging and thank you
00:07:17
for your questions for the queries for
00:07:19
the comments for the critics for the
00:07:22
uploads for the commendation I thank you
00:07:25
all so very much
00:07:26
continue to share my videos to your
00:07:28
students to your academic community as
00:07:31
many who are still afraid of
00:07:32
econometrics please tell them chronicle
00:07:35
matrix simplifies understanding thank
00:07:38
you for watching please don't go away
00:07:40
I'll be right back with the next video
00:07:42
which is on how to estimate a treasured
00:07:46
Gaucho or what you can call the gjr gosh
