2. Import your data: File --> Open File
Let's use tutorial data for now: go to File --> Open File from Tutorial Data files --> FUTURESDATA
In the middle grey area you'll see a list of all imported variables.
3. By clicking on "Variables" button, you can view list of all the variables imported. Close.
4. Click on "Data" button to go through each datapoint. Close.
5. Next, go to menu bar and click on «Volatility Modelling» and select the 2nd option («Multivariate GARCH…»).
6. Define the period of study: from 02-Jan-95 to 31-Dec-04.
7. In the command field, provide list of variables by typing names with spaces, or click on "Variables" button and double-click on the names of variables. RUN.
8. In the 1st section "Decay Factor for the Variance" select "Unrestricted" and un-tick "Same for all assets".
9. In the 2nd section "Decay Factor for the Covariance" choose "Unrestricted".
10. Under the 3rd section of "Distributional Assumptions" select "Multivariate Normal" for now.
11. Keep "Rolling historical volatility" and "Rolling Volatility Window" as by default.
12. To estimate all the parameters of the model, click on RUN. For this data it will go through 51 iterations before giving the output.
13. Copy the output. Open MS Word, adjust page margins and paste. You can reduce the font size for better fit. Save file.
14. Go back to Microfit and close output window.
15. Select the 2nd option ("List/Plot/Save...") in "Multivariate GARCH Post Estimation Menu". OK
16. To plot graphs, choose the 2nd option ("Plot conditional..."). OK
17. Select estimates of conditional volatilities. OK. Here you can see the dynamics of volatilities of each variable. Close.
18. Again the 2nd option. Select estimates of conditional correlations. OK
19. Alternatively, you can export all the underlying estimates of volatilities and correlations into CSV file for later use in MS Excel (plotting graphs).
20. Next, we going to estimate results assuming Multivariate t-distribution.
21. Quit from the previous window. Select option 0 ("Re-specify MGARCH..."). OK
22. Now, change Distributional Assumption to «Multivariate t». RUN
23. After 21 iterations (with this data) you'll get the output.
24. Copy the output and paste into MS Word.
Done, alhamdulillah! The rest is just formatting and interpretation.
Watch video tutorial: HERE
