Spgenerate with panel data in Stata15

[Marco Giuliani]

I have a perfectly balanced dataset with 25 countries and 10 years correctly xtset and spset in Stata15 for a spatial regression analysis.
I created a W connectivity matrix and, in order to save thm for other datasets and works, I would like to create some lagged variables with Spgenerate.

spmatrix create contiguity W if year==2007, normalize(row) replace

There are no specifications for panel data in the manual (like the ones for producing the matrix by using an if) but when I try to create the lagged variable

 spgenerate Wgdp =W*gdp

Stata returns me the following msg:

_IDs in weighting matrix W do not match _IDs in estimation sample
There are places in W not in estimation sample and places in estimation sample not in W.

I believe it has something to do with the panel structure, but I'm not at all sure, and I cannot figure out something different or a solution.

NB Running a spxtregress with those same data does not produce any error, and Stata is able to estimate the model.

Any suggestions?

Thanks a lot

Marco Giuliani

PS I tried with diverse matrices, including idistance, and even the type of normalization chosen does not make any difference

[Di Liu (StataCorp)]

-spgenerate- is not designed to generate spatial-lag for panel data. However, I think there is a work around for your problem. Basically, you can use -if- clause in -spgenerate- to generate a spatial lag for each time period.

For example,

generate Wy = .

foreach t in 1 2 3 {
capture drop tmp
spgenerate tmp = W*y if time == `t'
replace Wy = tmp if time == `t'
}