Scribd
Upload
72 views3 pages

Arch Garch Model

This document discusses analyzing the relationship between two time series (fchi and n100) using ARCH/GARCH and cointegration models. It provides steps to check for heteroskedasticity using ARCH models, estimate a GARCH model, test the series for unit roots and stationary, check for cointegration using the Johnson test, estimate a VAR model to determine the optimal lag length, then estimate a VECM if cointegration is found.

Uploaded by

Anees Ur Rehman
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
72 views3 pages

Arch Garch Model

This document discusses analyzing the relationship between two time series (fchi and n100) using ARCH/GARCH and cointegration models. It provides steps to check for heteroskedasticity using ARCH models, estimate a GARCH model, test the series for unit roots and stationary, check for cointegration using the Johnson test, estimate a VAR model to determine the optimal lag length, then estimate a VECM if cointegration is found.

Uploaded by

Anees Ur Rehman
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Arch garch model:

For first series (fchi)

Copy two series …genr rf=log(fchi/fchi(-1))

Dusri series ka b same retrun lena hy

Quick >estimate equation>rf c (-1)

View>residual diagnostic > hetrosakadicity > select ARch from the list > if value of r square is significant
hetro exist and garch model will imply.

Estimate > equation > (rf c rf(-1)) > model selection men sy arch model select krna hy. Arch M me sy
variance select krna hy

Go to proc > residual series > > create ordinary series > (rename series to en or ef

Estimate > write equation > add en series (rf c rf(-1) ef(-1)

Proc > residual series > create standardized Series >renam as (utm)

Proc > make garch variance > creat garch01 series

No for UTV

Genr utv= (utv^2)/garch01

Now for the 2nd series

Genr rn=log(n100/n100(-1))

Quick > Estimate equation > rn c rn(-1)

View > residual diagnostic > heteroskadacity > Arch

Estimate equation select model arch model

Proc > residual series > make ordinary series > rename as en

Estimate > write equation > rn c rn(-1) en(-1)

Now spill over checking of fchi on n100

Estimate equation > rn c rn(-1) en(-1) utm > (variance) > variance regerssor men utv likhna hy

Result ko interpret krna hy last py


Co integration:

Genr lf=log(fchi)

Genr ln=log(n100)

Plot lf ln

Copy past graphs in word file

Double click on ln series > view > unit root test > ADF test > leve > trend and intercept assumptions

Agr trend ki assumption significant hy to trend ki assumption len gy nai to drop kr den gy

View > unit root test > level > intercept

If still insignificant again test run kren gy

View > unit root test > first difference > intercept

Now significant hy to series stationary hy

Or interpret kren gy k LN is stationary at integerated level 1

Ab second series ko b aisy hi stationary kren gy

Jab dono series stationary ho jaen to

Quick > group statistic > Johnson test > enter ln lf both series

Selecti summarize at number 6 py,

Ab 1st and 5th ko chor k baki 3 ko check kren agr us men sary zero hon to koi b test use kia ja skta hy

Quick > estimate VAR > click unrestricted VAR > write lf ln in series

View > lag structure > lag length criteria > table men AIC ki row men jis value k sath * ho ga wo hmari lag
length ho gi.

Quik > group stat > jonson > lag length men 1 k sath 8 dal den jo hmari lag value i

Ab result me sy agr con integeration exist krti hy to to next step kren gy nai to yahn tak stop kr den gy

Agr cointegeration exist karti hy to:


Quick > estimate VAR > Vector error correction model > endgous men ln lf or lag length criteria men lag
lag ki length enter kren gy

You might also like