This excel file demonstrates estimation of vector autoregression on multiple time series. It takes time series and lags as input, and calculates matrix of coefficients and fitted values using regression estimates. It is based on paper Vector Autoregression
Following are steps required to perform VAR:Step 1
Enter time series in sheet "timeseries". The observations are laid out in rows and each column represents a variable.
For example :
value in A1 stores observation of variable1 at t=0
value in A2 stores observation of variable1 at t=1
value in A3 stores observation of variable1 at t=2
value in B1 stores observation of variable2 at t=0
value in B2 stores observation of variable2 at t=1
value in B3 stores observation of variable2 at t=2
Enter no. of lags in sheet "Sheet1".Step 3
Press Start Calculation button in sheet "Sheet1".Step 4
Observe result of estimated coefficients in sheet "betavals". Rows represent coefficients for each equation and each column represents a different equationStep 5
Observe result of fitted values in sheet "fittedvals", which can be compared to values in "inputvals". Ideally they should match!
The class VectorAR in this file is used to perform vector autoregression on multiple time series. For usage of this class, please refer example code of CommandButton1_Click