Multicollinearity and ARDL

[wajahat Ali]

Hi All,
Currently i am analyzing time series data of 41 years from 1971 to 2011. All the Variables of the study (C02 emissions, GDP, GDP square, Quasi money, trade openness, population and energy consumption, are integrated of order I(1) only one variable is I(0) therefore i applied ARDL Bound testing approach. When i checked the correlation matrix i found that all the variables are highly correlated (0.98, 0.95, 0.91, 0.81, 0.78). To combat with multicollinearity i applied the Ridge Regression and the VIF was less than 10 with a biasing constant value of 0.02. The question is that which estimation table i should replace with the Ridge regression estimated coefficient values and will my ARDL, long and short run estimates be valid.... Please comment..What should i do with multicollinearity..Dropping variables did not work in my case