Can I use slag specification for incorporating a spatial lag of the dependent variable rather than specifying it while running the regression using spml. What would be the difference if I incorporated one of the three methods, and which one would be incorrect?
Case 1
Model1 = y ~ x_1 + x_2
Result1 = spml(Model1, data, listw = W, model = "pooling", lag = T)
Case 1 is conventionally what we follow, but what if Case 2:
Model2 = y ~ slag(WY, listw = W) + x_1 + x_2
Result2 = spml(Model1, data, listw = W, model = "pooling", lag = F)
And just in case I would want to add multiple spatial lag matrices, what if I club the two methods in that case, which brings us to Case 3
Model3 = y ~ slag(WY, listw = W_a) + x_1 + x_2
Result3 = spml(Model1, data, listw = W_b, model = "pooling", lag = T)
I have tried running all three cases but the difference does not seem to be very significant when the comes to the values of the coefficients but the p- values do change.
Supriyo Mondal is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.