I am currently working on a Panel data regression in R on stock return of Firms listed in the S&P 500 index related to public health variables and news attention to COVID-19. I have filtered out for those firms only present in the Health Care sector and I computed a variable return
that is defined as the (closingprice_t - closingprice_t-1 / closingprice_t-1) * 100 %. My model is listed as follows:
return ~ newsattention + Holiday + day_of_week + Cases + Hospitalizations + Rnumber + Deaths + Policy_stringency_index + mobility_residential + mobility_transit_stations + mobility_retail_and_recreation + Sent_interp
When I run the following regressions in R, I get exactly the same estimates with all three methods?:
pooledOLS <- plm(return ~ newsattention + Holiday + day_of_week + Cases + Hospitalizations + Rnumber + Deaths + Policy_stringency_index + mobility_residential + mobility_transit_stations + mobility_retail_and_recreation + Sent_interp, data = pfirms_hc, method = "pooling")
fe1 <- plm(return ~ newsattention + Holiday + day_of_week + Cases + Hospitalizations + Rnumber + Deaths + Policy_stringency_index + mobility_residential + mobility_transit_stations + mobility_retail_and_recreation + Sent_interp, data = firms_hc, index = c("name", "date_real"), model = "within")
re <- plm(return ~ newsattention + Holiday + day_of_week + Cases + Hospitalizations + Rnumber + Deaths + Policy_stringency_index + mobility_residential + mobility_transit_stations + mobility_retail_and_recreation, data = pfirms_hc, model = "random")
I don't know why these results all turn out the same as they should be different among all three models... Can someone please help me out!!!
Aucun commentaire:
Enregistrer un commentaire