correcting for autocorrelation
Posted: Tue Jul 31, 2012 3:16 pm
I am running a panel data regression on the effects of energy use per $1000 unit gdp, electricity consumption per capita and carbon emissions on the human development index and gdp per capita rates for 18 countries. My results show a random effects model is preferred to a fixed effects model, as per the hausmann test results. However due to a low durbin watson statistic and high correlation in the descrtive statistics, there exists auto correlation between the hdi and electricity consumption. How can I go about correcting for auto correlation in a random effects model? checking for causality. By stacking the data and running a granger causality test between hdi and electricity consumption?