PR for #257: Explore model with state by year-month fixed effects

[santiagohermo]

Can someone take a look at this code @diegogentilepassaro @gabrieleborg ? The goal of #257 was to produce a new table to test the difference between the state-time model and the baseline model. This led to a new folder in the repo: /fd_test/. We also explored some features of this model that led to some changes in /fd_robustness/ (see https://github.com/diegogentilepassaro/min_wage_rent/issues/257#issuecomment-1509863148).

Other things to do:

• [ ] Explore why the coefficients on the model with weights changed. --> To do in #268
• [x] Add make.py to /fd_test/ and review code.
• [x] Link table test_stateFE.tex create in ./analysis/ to the cover letter in ./paper_slides/reply and drop the version of the table that lives in ./paper_slides/input.

[gabrieleborg]

@santiagohermo, I am currently looking for a new apartment and will move soon so I might have very little time in the next 2 weeks. I will try to take a look at this next week though, and keep you posted on the progress.

[santiagohermo]

Thanks for letting me know @gabrieleborg! Maybe @diegogentilepassaro has time to look at this. No rush though as I already copied the table to include in the cover letter in #250.

[diegogentilepassaro]

Grande @santiagohermo! I am having a nightshift week at work, but will try to look at this ASAP :D

[santiagohermo]

Important thing to do here: Link table test_stateFE.tex to the cover letter in paper_slides/reply and drop the version of the table that lives in paper_slides/input.

[santiagohermo]

I think your idea was to re-run /analysis/ and check the models with weights @diegogentilepassaro. Can I help here?

[diegogentilepassaro]

Sorry for the delay here @santiagohermo! Running derived\estimation_samples now and when it finsihes will run analysis\fd_robustness and will make a commit so that we can take a look.

[santiagohermo]

I tried to run yesterday but my code broke! That's why you had problems Diego. Sorry for not warning. I'm trying to fix this now.

[santiagohermo]

Gracias por la charla querido @diegogentilepassaro! Avisame como va ese re-run. Si no se arreglan los coeficientes, podemos cerrar este issue y abrir uno nuevo para revisar esos weights antes de un nueva submission.

[diegogentilepassaro]

Ahí corrió todo querido @santiagohermo! Los weights dan diferentes a los que estaban commiteados:

[santiagohermo]

Gracias @diegogentilepassaro.

Algo en los counterfactuals cambió también, que raro che! Después le pego una mirada en más detalle a ver que onda.

El estilo de los plots cambió también :/. Sabes como resetear el estilo del stata en el server? Así volvemos al viejo formato.

[diegogentilepassaro]

Uhhh! No me di cuenta de esa. Qué te parece si corro todo base de nuevo tambien? Ahí vemos que da todo, y en función de eso decidimos.

Lo del default de los gráficos de STATA es así: set scheme s2color, permanently. Ese es el scheme default de stata hasta STATA 18. Ahora en el nuevo STATA es stcolor. Supongo que era uno de esos!

[santiagohermo]

Gracias @diegogentilepassaro! Estoy arreglando las figuras así puedo cerrar este issue. Voy a abrir otro para acordarnos de revisar los weights.