Question about manipulation of the local data matrix in the BAR example

[jagman88]

Hi Paul,

Thanks so much for making this code available! I'm working with the R version of your code, but I have a question about the way you manipulate the "local" data matrix, as in the code block from your example below:

BAR_local %>%
    select(-sh_ind_) %>%
    mutate(ind = str_glue("t{year}_init_sh_ind_{ind}")) %>%
    spread(ind, init_sh_ind_, fill = 0) %>%
    print() -> BAR_local2

Why do you fill the init_sh_ind_ variable with zeros for the non-initial year entries? So your initial shares come from the year 1980, but you rework the data matrix so that the initial share is set equal to zero in all non-1980 years. Why are the shares not set equal to the initial shares in all years? Afterall, this is how the original data matrix is constructed.

Thanks for the help, and thanks again for the code! Very helpful!

[jjchern]

Maybe I will let Paul expound on the design decision later, but the effect of the code is to generate the interactions between year dummies and the initial industry shares. For example, if I limit to industry no. 100, the code produces the following:

library(bartik.weight)

BAR_local %>%
    arrange(czone, ind, year) %>% 
    select(-sh_ind_) %>%
    mutate(ind = str_glue("t{year}_init_sh_ind_{ind}")) %>%
    spread(ind, init_sh_ind_, fill = 0) %>%
    select(czone, year, contains("ind_100"))

#> # A tibble: 2,166 x 5
#> czone year      t1980_init_sh_ind_100 t1990_init_sh_ind_100 t2000_init_sh_ind_100
#> <dbl> <dbl+lbl>                 <dbl>                 <dbl>                 <dbl>
#> 1  100. 1980                    0.00276               0.                    0.     
#> 2  100. 1990                    0.                    0.00276               0.     
#> 3  100. 2000                    0.                    0.                    0.00276
#> 4  200. 1980                    0.00290               0.                    0.     
#> 5  200. 1990                    0.                    0.00290               0.     
#> 6  200. 2000                    0.                    0.                    0.00290
#> 7  301. 1980                    0.00111               0.                    0.     
#> 8  301. 1990                    0.                    0.00111               0.     
#> 9  301. 2000                    0.                    0.                    0.00111
#> 10  302. 1980                    0.00388               0.                    0.     
#> # ... with 2,156 more rows

The corresponding Stata code is about here: https://github.com/paulgp/bartik-weight/blob/master/code/example_BAR.do#L25-L62.

These variables are being used later as a vector of IVs (https://github.com/paulgp/bartik-weight/blob/master/code/example_BAR.do#L85):

bartik_weight, z(t*_`ind_stub'*)    weightstub(nat1980_empl_ind_*) x(`x') y(`y')  controls(`controls') weight_var(`weight')  absorb(czone) by(year)