Hello, I am forecasting the interest rate using normal VAR models and Im gonna compare the results (or improve) to Bayesian VAR models (litterman/minnesota prior). The variables im using is the interest rate (levels), inflation (annual change), the spread between long and short interest rates, the output gap and a trade-weighted foreign interest rate variable (levels). When changing lambda1 and lambda2 to different values, the RMSE naturally varies from model to model. However, I'm not seeing any better results than the standard VAR model. The normal VAR model uses 2 lags (although Ive tested up to 5 lags, no autocorrelation is present by the way) and it is quarterly data from 1995 to 2015. Do you have any tips on increasing the performance of the Bayesian VAR model, or do you suspect Im doing some mistake?
Thanks in advance
Bayesian VAR forecasting
Moderators: EViews Gareth, EViews Moderator
Return to “Econometric Discussions”
Who is online
Users browsing this forum: Google [Bot] and 19 guests