i have small panel data set
my topic is crowding-in and crowding-out effects of public investment on private investment is emerging Asian economies
i have 6 countries and yearly data for 15 years
one IV (Public Investment) and DV (Private investment) also have 8 control variables
which model do you suggest that better for my data
either i go for panel unit root test than ARDL Model or i go for Pooled OLS and FE, RE, and 2sls?
I recommend you to increase your number of observations, may take quarterly data which can be useful for any statistical test.
I would recommend linear regression models, Gaussian distribution or some advance models e.g., von Misses Fisher & Bingham distributions on the unit hyper sphere (https://arxiv.org/pdf/1605.00316.pdf).
Good Book----http://www.econ.brown.edu/Faculty/Tony_Lancaster/papers/book_selection.pdf
Also you can follow the research group of the University of Sydney (https://sydney.edu.au/research/opportunities/opportunities/2301), doing the same research as you do!
- Badges
- Science topic