1
$\begingroup$

Suppose I have a time series of stock growth and one of gdp growth and education over the years.

Can I try to explain stock using gdp and education by running an OLS or would I be mistaken from a formal (not content) point of view?

$\endgroup$
3
$\begingroup$

I'm not sure what you mean by formal but if you have intuition into why gdp or your education metric should be related to the growth of stocks, then by all means, build a model that tries to find the relationship between them.

OLS is one way to find a relationship between gdp and education relative to stock growth. There would be nothing wrong with that concept in general.

That said, there are many things you could do wrong in setting up your model. You may want to share more in order to get more help.

$\endgroup$
  • $\begingroup$ I just wanted to know that I could use OLS for time series and I guess you have said I can: <<OLS is one way to find a relationship between gdp and education relative to stock growth. There would be nothing wrong with that concept in general.>>. Do I need all time series to be stationary? $\endgroup$ – user25954 Jan 6 '17 at 9:17
  • $\begingroup$ You don't "need" to have them all be stationary in the sense that nothing stops you from running the model. Is it better if they are stationary, yes. But sometimes it's impractical. $\endgroup$ – piRSquared Jan 6 '17 at 9:22
  • $\begingroup$ Because I have read here on stackexchange that if they are not stationary it is garbage... the problem is that I am struggling because if I make them stationary I have horrible results $\endgroup$ – user25954 Jan 6 '17 at 9:26
  • $\begingroup$ Theres a very fuzzy line between what's practical and what's theoretically ideal. You have to get to know your data and develop intuition and theory on how and why it should behave in certain ways. Then you test your theories. this site provides some good information as well. If you are just going to poke and prod and hope you find something useful, then you have to be careful not to abuse the theoretical assumptions. $\endgroup$ – piRSquared Jan 6 '17 at 9:34
  • $\begingroup$ But I have the economic interpretation... but how can I justify the fact that I am running OLS with non stationary data? $\endgroup$ – user25954 Jan 6 '17 at 10:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.