1- For the following variables compute the first 5 autocorrelations: gdp, ln(JAPAN IP), UNRATE, TSpread, and oil.
2- For these 5 variables, run 5 AR(1) models and based on your regression results, state what you can conclude about stationarity of these variables.
3- Take first difference of those series which are nonstationary. Did it make those series station- ary? Check by running AR(1) models of those first differenced series.
4- Prompt Stata to find out what lag orders you should be using for the following series: gdp, oil and ∆ln(JAPAN IP). Let pmax = 10. If you get conflicting answers for AIC and BIC, report the lag order as indicated by AIC.
5- Run Dickey Fuller test on all 5 variables in (1) to check whether there is unit root or not. Let p = 10.
6- Run Dickey Fuller test on all 5 variables after taking first differences to check whether there is unit root or not. Let p = 10.
7- Run the following ADL(2,2) model: ∆gdpt = β0 + β1∆gdpt−1 + β2∆gdpt−2 + δ1∆UNRATEt−1 + ∆UNRATEt−2
8- Test whether ∆UNRATEt granger cause ∆gdpt.