xtevent and trend(# saveoverlay) error(198)

[Francesca Rossi]

Hi All, I am estimating an event study plot, and I want to generate a 3 graphs: one with the "standard" xtevent, the second one with the overlay extrapolation line, and the third one with the extrapolated trend. This is an example of the dataset:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(year id perc_women treatment)
2014  2 .16666667 0
2015  2 .16666667 0
2016  2 .16666667 0
2017  2 .16666667 0
2018  2 .16666667 1
2019  2 .16666667 1
2020  2 .16666667 1
2021  2        .5 1
2022  2        .5 1
2023  2        .5 1
2014  4  .3333333 0
2015  4  .3333333 0
2016  4  .3333333 0
2017  4  .3333333 0
2018  4 .16666667 1
2019  4 .16666667 1
2020  4 .16666667 1
2021  4         0 1
2022  4         0 1
2023  4         0 1
2014  5         0 0
2015  5         0 0
2016  5         0 0
2017  5         0 1
2018  5         0 1
2019  5         0 1
2020  5         0 1
2021  5         0 1
2022  5         0 1
2023  5  .3333333 1
2014  6 .16666667 0
2015  6 .16666667 0
2016  6         0 0
2017  6  .3333333 1
2018  6  .3333333 1
2019  6  .3333333 1
2020  6 .16666667 1
2021  6 .16666667 1
2022  6 .16666667 1
2023  6 .16666667 1
2014  8         0 0
2015  8         0 0
2016  8         0 0
2017  8  .3333333 1
2018  8  .3333333 1
2019  8  .3333333 1
2020  8  .3333333 1
2021  8  .3333333 1
2022  8  .3333333 1
2023  8        .5 1
2014  9         0 0
2015  9         0 0
2016  9         0 0
2017  9  .3333333 1
2018  9  .3333333 1
2019  9  .3333333 1
2020  9  .3333333 1
2021  9  .3333333 1
2022  9  .3333333 1
2023  9  .3333333 1
2014 11 .16666667 0
2015 11 .16666667 0
2016 11 .16666667 0
2017 11 .16666667 0
2018 11        .5 1
2019 11        .5 1
2020 11        .5 1
2021 11         0 1
2022 11         0 1
2023 11         0 1
2014 12         0 0
2015 12         0 0
2016 12         0 0
2017 12         0 0
2018 12         0 0
2019 12 .16666667 1
2020 12 .16666667 1
2021 12 .16666667 1
2022 12 .16666667 1
2023 12 .16666667 1
2014 14         0 0
2015 14         0 0
2016 14         0 0
2017 14         0 0
2018 14         0 0
2019 14         0 1
2020 14         0 1
2021 14         0 1
2022 14         0 1
2023 14         0 1
2014 15  .3333333 0
2015 15        .5 0
2016 15        .5 0
2017 15        .5 0
2018 15        .5 1
2019 15        .5 1
2020 15        .5 1
2021 15  .3333333 1
2022 15  .3333333 1
2023 15  .3333333 1
end

where perc_women is my outcome variable (percentage of elected women), treatment is my policyvariable, and year is the time variable. I first set the panel, which is strongly balanced. Then I run this code:

xtevent perc_women, policyvar(treatment) window(4) impute(stag) trend(-4, method(gmm) saveoverlay)
xteventplot, overlay(trend)

but I get this error: "matrix operators that return matrices not allowed in this context"

same if I do not specify gmm, as the default is actually the gmm. Differently, when I run the same code by using method(ols) it works and I get the plot I need. By reading online it seems there could be mispelling errors or similar, but honestly I cannot see what it is happening here.

Could anyone help? Thank you very much

Francesca

[Jorge Eduardo Perez Perez]

Hi Francesca.

Can you try this again with xtevent 3.1.0 from SSC or Github? This may have been a bug that has been fixed in the latest version. I was able to execute this using xtevent 3.1.0 without issues.

JPP