Dummies time fixed effects

[Ana Vasconcelos]

Hello,
I have a dataset with several countries and quarterly values.
i' m running the following regression where i want to control for time (quarter) effects.
Xtrifreg Y X1 X2 X3, fe i(country_) q(10)
My doubt is that i don't understand the difference between the following issues to calculate the appropriate dummies:

1 - estimate the regression using only 4 dummies (one for each quarter). In that case i'll have: xtrifreg Y X1 X2 X3 dum1 dum2 dum3 dum4, fe i(country_) q(10)
2 - estimate the regression using several dummies for each quarter of each year. For example 1980q1, 1980q2, 1980q3, 1980q4, 1985q1, 1985q2 and so on. In that case I will have;

xtrifreg Y X1 X2 X3 dum1980q1 dum19080q2 dum1980q3 ...dum2013q1, fe i(country_) q(10).

Can you help me to understand the difference between these two approaches and which one is better to use?
thank you in advance.
best regards

[Clyde Schechter]

If you use just the four quarter indicators ("dummies") you are adjusting for seasonal variation. That is, you are taking into account seasonal differences in the outcome that recur in the same way every year. If you use indicators for 1980q1, 1980q2, 1980q3...2013q1, then you are adjusting for the possibility of a new, independent shock to the outcome in every quarter.

As to which one is better to use, that depends on the nature of the problem and the real-world patterns of fluctuation in the actual variables you are studying. As you don't identify what the Y variable even is, nobody can advise you. If you did say what Y is, somebody in your discipline who understands the science of that variable could help you. But at that point it is not a statistical question--it is a content-area science question.

[Ana Vasconcelos]

Hello Professor Clyde. Thank you for your kind answer. Now I understand the difference!

Best regards