I'm doing my dissertation on the effect of different QE's episode on Malaysia's portfolio liabilities using the Rate Spread (US 10Years Treasury Constant Maturity Rate (minus) US 3Months Treasury Constant Maturity Rate) as the proxy for QE. My variables are only "push" variables (EFFR, VIX, US Real GDP and Rate Spread). I have constructed unit root test (ADF & PP = I(1) variables confirmed), lag length 3 satisfies no autocorrelation, Johansen test for cointegration (max rank 1 = only one cointegration equation). From here, what should I do if I want to pursue ARDL? I would want to pursue ARDL since it could let me estimates the coefficient of Rate Spread on Malaysia portfolio liabilities at different phases. I'm using STATA by the way, thus if you could lay out the steps in Stata's command, it would be very helpful and appreciated. I have included my analysis (so far) below.Thank you