Arima + panel data

[Vernon Mila]

Hi there,

I am trying to run an ARIMA model on financial time series data.

I have time series data for 5 equity sectors on (e.g.) retained earnings vs. total assets going back to 1970s monthly. I would now like to be able to able to design an ARIMA model across the sectors (hence panel data).

Any ideas on how i can get started on Stata would be much appreciated. Do i model each sector individually -- what if i wanted to design a model across all 5 sectors, such that ONE arima specification would hold across these sectors.

I would additionally also like to run some tests on the mean reversion tendencies of these time series model -- can anyone suggest if running ADF on each of the series or if running gurst exponent might do the job. Also does stata have the capability to do Cointegrated Augmented Dickey Fuller (CADF) ? Thank you!
Thank you again ,and any advise would be much appreciated.

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Where you want to refer to a paper, please give us a complete reference - almost all of us will not be in your area.

You'll also probably do better if you ask one question rather than throwing out several.

If you took the time to Google "Cointegrated Augmented Dickey Fuller Stata" , you'd find Stata 15 includes a variety of ADF tests. I don't know about ARIMA particularly, but some of Stata's time series modeling are set up for one time series. Obviously, you can easily do ARIMA sector by sector with if statements. You might be able to test parameter equality across the models - I don't know if Hausman or suest work with ARIMA output.

[Vernon Mila]

Thanks Phil! Much appreciated -- i took your advise . Posted two new questions separately on VAR and REG -- if you might have any thoughts on those, please let me know

https://www.statalist.org/forums/for...ncluding-lag-0

https://www.statalist.org/forums/for...al-lags-in-var