This problem is based on the Hansen and King (1996) data set and the replication by McCoskey and Selden (1998). Hansen and King (1996) studied the stationarity properties of real per capita health care expenditures (HCE) and real per capita gross domestic product (GDP) for 20 OECD countries over the period 1960-1987. All variables were expressed in logarithms. This was done on a country-by-country basis. Their conclusion was that these variables are nonstationary and inference based on regressions relating HCE to GDP are misleading and spurious. McCoskey and Selden (1998) challenged this finding by applying the IPS panel unit root test to this data finding the hypothesis of panel unit roots is rejected when individual effects are included, but not rejecting this hypothesis when individual effects and individual time trends are included. Using EViews, replicate the following results:

(a) Perform the panel unit root tests on HCE and GDP when individual effects are included.

(b) Check the sensitivity of these results when both individual effects and individual linear trends are included.

(c) Perform the Pedroni test including individual effects and individual linear trends in the deterministic variables.

(d) Perform the Fisher panel Johansen cointegration trace and maximum eigenvalue tests including individual effects and individual linear trends in the deterministic variables.