Page 1 of 1
Fixed Effects or GMM-System? Panel Data Estimation GDP x Social Spending of Municipalities
Posted: Tue May 04, 2021 5:11 pm
by rafrods
Hello. I am doing a work on the impact of municipal social spending on municipal GDP for the period from 2002 to 2018, and I would like to know if the fixed effects model is sufficient, or if it is necessary to estimate also in GMM and compare the results. My dependent variable is the GDP per capita of the municipalities and my explanatory variables are some components of spending, classified as social (education, health, social security, social assistance, sanitation, housing), and some dummy's of time and electoral cycle.
Re: Fixed Effects or GMM-System? Panel Data Estimation GDP x Social Spending of Municipalities
Posted: Tue May 04, 2021 5:51 pm
by startz
Do you think the level of GDP affects the level of spending?
Re: Fixed Effects or GMM-System? Panel Data Estimation GDP x Social Spending of Municipalities
Posted: Wed May 05, 2021 1:42 pm
by rafrods
Yes, but GDP is my dependent variable (Y) and the types of speding are my X variables. Another question, are autocorrelation and heteroscedasticity also elements for using gmm?
Re: Fixed Effects or GMM-System? Panel Data Estimation GDP x Social Spending of Municipalities
Posted: Wed May 05, 2021 1:46 pm
by startz
If your Y variable can affect your X variable you should use GMM. GMM can also handle heteroskedasticity and autocorrelation.
Re: Fixed Effects or GMM-System? Panel Data Estimation GDP x Social Spending of Municipalities
Posted: Wed May 05, 2021 3:06 pm
by rafrods
Thank you!