I am conducting a time series analysis with 11 macroeconomic variables, using the Stock Price Index (SPI) as the dependent variable and the remaining 10 variables as independent variables. Following standard practice, I have log-transformed the variables and confirmed that they are non-stationary at the raw level but become stationary at the first difference (I(1)).
Using the Johansen cointegration test with an optimal lag of 1, I found 5 cointegrating vectors. However, when I estimate the Vector Error Correction Model (VECM) in Stata with a rank of 5, I obtain multiple equations, whereas I am only interested in the relationship where SPI is the dependent variable and the other variables are independent.
My question is: what is the correct procedure to obtain the desired output? Should I estimate the VECM with a rank of 1, which would provide a single equation, or estimate the VECM with a rank of 5 and extract the coefficients from the first equation, which corresponds to the SPI-dependent relationship? The coefficients obtained from these two approaches differ, and I would appreciate guidance on the appropriate method to use.
Using the Johansen cointegration test with an optimal lag of 1, I found 5 cointegrating vectors. However, when I estimate the Vector Error Correction Model (VECM) in Stata with a rank of 5, I obtain multiple equations, whereas I am only interested in the relationship where SPI is the dependent variable and the other variables are independent.
My question is: what is the correct procedure to obtain the desired output? Should I estimate the VECM with a rank of 1, which would provide a single equation, or estimate the VECM with a rank of 5 and extract the coefficients from the first equation, which corresponds to the SPI-dependent relationship? The coefficients obtained from these two approaches differ, and I would appreciate guidance on the appropriate method to use.
