[R-sig-Geo] Error in impacts() after lagsarlm (spdep package)

[Roger Bivand]

On Wed, 13 Nov 2019, Denys Dukhovnov via R-sig-Geo wrote:

> I posted this question on StackOverflow a while back but received no answer. 

I maintain the spatialreg package to which I think you are referring. I 
never visit SO, both to save time and because the signal/noise ratio there 
is not encouraging (with exceptions where reproducible examples are used).

While your post was pain text (thanks), what you copied from SO got rather 
mangled. The spdep package does warn you that the functions you are using 
are deprecated and are actually to be called from spatialreg.

>
> I have 9,150 polygons in my data set. I was trying to run a spatial 
> autoregressive model (SAR) in spdep to test spatial dependence of my 
> outcome variable. After running the model, I wanted to examine the 
> direct/indirect impacts, but encountered an error that seems to have 
> something to do with the length of neighbors in the weights matrix not 
> being equal to n. I tried running the very same equation as SLX model 
> (Spatial Lag X), and impacts() worked fine, even though there were some 
> polygons in my set that had no neighbors.
>
>> # Defining queen contiguity neighbors for polyset and storing the 
>> matrix as list
>> q.nbrs <- poly2nb(polyset)
>> listweights <- nb2listw(q.nbrs, zero.policy = TRUE)
>
>> # Defining the model
>> model.equation <- TIME ~ A + B + C
>
>> # Run SAR model reg <- lagsarlm(model.equation, data = polyset, listw = 
>> listweights, zero.policy = TRUE)

You are completely unnecessarily using the "eigen" method. If your weights 
are symmetric, use Cholesky decomposition ("Matrix"), much faster, same 
output.

>
>> # Run impacts() to show direct/indirect impacts
>> impacts(reg, listw = listweights, zero.policy = TRUE)
>
>> Error in intImpacts(rho = rho, beta = beta, P = P, n = n, mu = mu, Sigma = Sigma,  :
>     length(listweights$neighbours) == n is not TRUE
>

Never, ever, do this. Did you read LeSage & Pace? Use traces, not the 
weights themselves. With the listw object, you need to invert an nxn 
matrix once in this case, 1+R times if you run Monte Carlo simulations.

If you provide a reproducible example using built-in data, I can try to 
provide a more informative error message.

> What am I doing wrong? I am running Windows 10 machine with R 3.5.3 with 
> the most up-to-date spdep package, if it helps.
>

R is at 3.6.1.

Hope this clarifies,

Roger

> Thank you very much.
>
> Regards,
> Denys Dukhovnov
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

-- 
Roger Bivand
Department of Economics, Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; e-mail: Roger.Bivand using nhh.no
https://orcid.org/0000-0003-2392-6140
https://scholar.google.no/citations?user=AWeghB0AAAAJ&hl=en