Dear all,

I am using x_ols to estimate spatial standard errors of multiple models. Now, I'd like to plot the coefficients and SEs using coefplot, however I can't manage to get coefplot to display the spatial standard errors - it always uses the regular ones. I know that x_ols creates a matrix cov_dep that contains the covariance matrix and coefplot has se() and v() options to provide other standard errors, however it always asks for e() elements, which the cov_dep matrix isn't. Is there a simple solution to get coefplot to accept the spatial standard errors from a covariance matrix? I am using Stata 16.1 on Windows 10.

Here an example code (displaying regular standard errors, __000006 is the name of the stored variable of interest, ch):

Code:
x_ols x y cutoff1 cutoff2 sw_index const ch, xreg(2) coord(2)
est sto sw
x_ols x y cutoff1 cutoff2 total_nr const ch, xreg(2) coord(2)
est sto tot

coefplot sw tot, keep(__000006) xline(0, lcolor(black) mlabel(cond(@pval<.01, "***", cond(@pval<.05, "**", cond(@pval<.1, "*", ""))))
An an example output of x_ols:

Code:
     Results for Cross Sectional OLS corrected for Spatial Dependence

                                   number of observations=  822
 Dependent Variable= sw_index

variable    ols estimates   White s.e.   s.e. corrected for spatial dependence
--------    -------------   ----------   -------------------------------------
const       3.1523342       .02389286    .02146005
ch          -.05491425      .02544105    .02326226

Thank you for your help.