Hello,
I am looking at logreturns of two different stock market indices in two countries and seeing if the weather (Cloud, temp, humidity, rain) affects the returns in two countries (two different indices). I am using daily data from 1993-2017.
Therefore I have cross sectional timeseries data. I will be using a GARCH regression, but some of the questions below aren't related to only garch (see q1 and 2 and 8).
One country's data is on top of the others and I identified it as cross sectional by using:
bys country: gen time=_n
xtset country time
I have a few questions I would really appreciate help with
1) Firstly, what does it mean to include a lagged variable, how would I choose it?
2) How do i know whether to include a squared variable and again, how do i choose it?
3) How do I test for how many GARCH lags to use?
4) How do i look at volatility in a garch model?
5) How can I include interactive factor(dummy) variables in the regression?
6) Do i test the errors and residuals, and if so how?
7) how do i test for serial correlation and what does this even show?
8) as i have two regressions per country (pre shock and post shock), how would I show these results on one table?
Please help, I would appreciate any help I am so lost and stuck. If it helps i can send you my data/do file somehow, but otherwise just instructions on which commands is something I would be eternally grateful for.
THANK YOU SO MUCH
I am so desperate lol x
0 Response to Time Series Regression
Post a Comment