Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Aggregating individual-level data to country-level data

    Dear Statalist Community,

    I am new to Stata and need guidance with two questions (please see below under dataex).

    I am looking at whether internet and phone access of an economy (developing economies) drives its female financial inclusion. I have three different dependent variables of financial inclusion (owning a bank account, saving, borrowing). These variables are binary and on the individual-level (survey micro-level data). Moreover, there are other important individual-level measures such female (used as an interaction term with the independent variables, and also as control variable in another model), education, and income quintile (both used as control variables). The independent variables, internet access and mobile subscriptions, and are on the country-level. Other control variables as for instance GDP are also on the country-level. Since the 'main' independent variables of interest are on the economy-level, I have left them constant throughout the variables in the dataset. Please see an extract of my dataset below:

    Code:
    * Example generated by -dataex-. For more info, type help dataex
    clear
    input long id str24 economy byte(t account_formal saved_formal borrowed_formal female educ inc_q) double(internet mobile)
      1 "Afghanistan" 1 1 . 0 0 1 3 5 45.81362616
      2 "Afghanistan" 1 0 . 0 0 1 3 5 45.81362616
      3 "Afghanistan" 1 0 . 0 1 1 1 5 45.81362616
      4 "Afghanistan" 1 . 0 0 0 2 1 5 45.81362616
      5 "Afghanistan" 1 0 0 0 0 2 5 5 45.81362616
      6 "Afghanistan" 1 0 . 0 0 1 2 5 45.81362616
      7 "Afghanistan" 1 0 . 0 1 1 3 5 45.81362616
      8 "Afghanistan" 1 0 0 0 1 1 4 5 45.81362616
      9 "Afghanistan" 1 0 0 0 1 2 4 5 45.81362616
     10 "Afghanistan" 1 0 . 0 1 1 4 5 45.81362616
     11 "Afghanistan" 1 0 . 0 0 1 5 5 45.81362616
     12 "Afghanistan" 1 0 . 0 0 1 2 5 45.81362616
     13 "Afghanistan" 1 0 . 0 0 1 1 5 45.81362616
     14 "Afghanistan" 1 0 . 0 0 1 4 5 45.81362616
     15 "Afghanistan" 1 . . 0 1 1 4 5 45.81362616
     16 "Afghanistan" 1 0 0 0 1 1 5 5 45.81362616
     17 "Afghanistan" 1 1 . 0 1 2 5 5 45.81362616
     18 "Afghanistan" 1 0 . 0 0 2 3 5 45.81362616
     19 "Afghanistan" 1 0 . 0 0 1 1 5 45.81362616
     20 "Afghanistan" 1 . . 0 0 2 5 5 45.81362616
     21 "Afghanistan" 1 0 . 0 1 2 5 5 45.81362616
     22 "Afghanistan" 1 0 0 0 0 1 4 5 45.81362616
     23 "Afghanistan" 1 0 . 0 0 1 3 5 45.81362616
     24 "Afghanistan" 1 0 . 0 1 1 3 5 45.81362616
     25 "Afghanistan" 1 0 0 0 1 1 4 5 45.81362616
     26 "Afghanistan" 1 0 . 0 1 1 5 5 45.81362616
     27 "Afghanistan" 1 0 . 0 1 1 4 5 45.81362616
     28 "Afghanistan" 1 0 . 0 1 1 4 5 45.81362616
     29 "Afghanistan" 1 . . 0 0 1 4 5 45.81362616
     30 "Afghanistan" 1 0 . 0 0 1 5 5 45.81362616
     31 "Afghanistan" 1 1 . 0 0 1 4 5 45.81362616
     32 "Afghanistan" 1 1 . 1 1 3 4 5 45.81362616
     33 "Afghanistan" 1 0 . 0 1 1 3 5 45.81362616
     34 "Afghanistan" 1 0 . 1 1 2 4 5 45.81362616
     35 "Afghanistan" 1 0 . 0 1 1 5 5 45.81362616
     36 "Afghanistan" 1 0 0 0 0 1 5 5 45.81362616
     37 "Afghanistan" 1 1 . 1 0 1 5 5 45.81362616
     38 "Afghanistan" 1 0 . 0 1 2 5 5 45.81362616
     39 "Afghanistan" 1 0 . 0 0 1 2 5 45.81362616
     40 "Afghanistan" 1 1 . 0 0 1 3 5 45.81362616
     41 "Afghanistan" 1 0 . 0 0 1 4 5 45.81362616
     42 "Afghanistan" 1 . . 0 1 1 1 5 45.81362616
     43 "Afghanistan" 1 0 . 0 1 2 4 5 45.81362616
     44 "Afghanistan" 1 0 0 0 0 1 4 5 45.81362616
     45 "Afghanistan" 1 0 . 0 1 1 4 5 45.81362616
     46 "Afghanistan" 1 0 . 0 0 1 5 5 45.81362616
     47 "Afghanistan" 1 0 . 0 1 1 4 5 45.81362616
     48 "Afghanistan" 1 0 . 0 1 1 1 5 45.81362616
     49 "Afghanistan" 1 0 . 0 0 1 4 5 45.81362616
     50 "Afghanistan" 1 1 . 0 0 1 4 5 45.81362616
     51 "Afghanistan" 1 0 . 0 0 2 5 5 45.81362616
     52 "Afghanistan" 1 0 . 0 0 2 2 5 45.81362616
     53 "Afghanistan" 1 0 . 0 1 1 2 5 45.81362616
     54 "Afghanistan" 1 0 . 1 0 1 4 5 45.81362616
     55 "Afghanistan" 1 0 . 0 1 1 4 5 45.81362616
     56 "Afghanistan" 1 0 . 0 1 2 3 5 45.81362616
     57 "Afghanistan" 1 0 . 0 1 1 4 5 45.81362616
     58 "Afghanistan" 1 1 . 0 0 2 3 5 45.81362616
     59 "Afghanistan" 1 0 . 0 0 1 4 5 45.81362616
     60 "Afghanistan" 1 0 . 0 1 1 4 5 45.81362616
     61 "Afghanistan" 1 0 . 0 1 1 3 5 45.81362616
     62 "Afghanistan" 1 0 . 0 0 1 3 5 45.81362616
     63 "Afghanistan" 1 0 . 0 1 1 2 5 45.81362616
     64 "Afghanistan" 1 1 . 0 0 2 3 5 45.81362616
     65 "Afghanistan" 1 0 . 0 0 1 4 5 45.81362616
     66 "Afghanistan" 1 0 . 0 1 1 2 5 45.81362616
     67 "Afghanistan" 1 0 . 0 1 1 2 5 45.81362616
     68 "Afghanistan" 1 0 0 0 1 2 4 5 45.81362616
     69 "Afghanistan" 1 0 . 0 1 1 4 5 45.81362616
     70 "Afghanistan" 1 0 . 1 0 1 5 5 45.81362616
     71 "Afghanistan" 1 0 . 0 0 1 3 5 45.81362616
     72 "Afghanistan" 1 0 . 0 1 1 5 5 45.81362616
     73 "Afghanistan" 1 0 . 0 1 1 3 5 45.81362616
     74 "Afghanistan" 1 0 . 0 1 1 3 5 45.81362616
     75 "Afghanistan" 1 0 0 0 0 1 5 5 45.81362616
     76 "Afghanistan" 1 0 . 0 1 1 3 5 45.81362616
     77 "Afghanistan" 1 0 . 0 1 2 5 5 45.81362616
     78 "Afghanistan" 1 0 . 0 0 1 3 5 45.81362616
     79 "Afghanistan" 1 0 . 0 0 1 1 5 45.81362616
     80 "Afghanistan" 1 0 . 0 0 1 5 5 45.81362616
     81 "Afghanistan" 1 0 0 0 1 2 5 5 45.81362616
     82 "Afghanistan" 1 0 . 1 0 1 5 5 45.81362616
     83 "Afghanistan" 1 0 . 0 0 2 3 5 45.81362616
     84 "Afghanistan" 1 1 1 0 0 2 3 5 45.81362616
     85 "Afghanistan" 1 0 . 0 0 2 4 5 45.81362616
     86 "Afghanistan" 1 0 . 0 1 1 4 5 45.81362616
     87 "Afghanistan" 1 0 . 0 0 2 5 5 45.81362616
     88 "Afghanistan" 1 0 . 0 1 1 2 5 45.81362616
     89 "Afghanistan" 1 0 . 0 1 1 2 5 45.81362616
     90 "Afghanistan" 1 1 0 0 0 2 5 5 45.81362616
     91 "Afghanistan" 1 0 0 0 1 2 5 5 45.81362616
     92 "Afghanistan" 1 0 . 1 0 2 3 5 45.81362616
     93 "Afghanistan" 1 1 . 1 0 1 5 5 45.81362616
     94 "Afghanistan" 1 0 . 0 1 1 2 5 45.81362616
     95 "Afghanistan" 1 0 . 0 1 2 5 5 45.81362616
     96 "Afghanistan" 1 0 . 0 1 1 5 5 45.81362616
     97 "Afghanistan" 1 0 . 0 0 1 1 5 45.81362616
     98 "Afghanistan" 1 0 . 0 1 1 3 5 45.81362616
     99 "Afghanistan" 1 0 . 0 1 1 2 5 45.81362616
    100 "Afghanistan" 1 0 . 0 1 1 1 5 45.81362616
    end

    My methodology is based on the methodology of the linked paper (p. 16-19) (Demirguc-Kunt, Asli; Klapper, Leora; Singer, Dorothe. 2013. Financial Inclusion and Legal Discrimination Against Women : Evidence from Developing Countries).

    I have two specific questions:

    1) Data aggregation: effect of internet/mobile on general financial inclusion
    Before analyzing whether internet/mobile have an influence on female financial inclusion, I want to test whether they have an effect on financial inclusion in general. Here my question is, do I have to aggregate the individual-level data (of the financial inclusion dependent variables) to the country-level? Or is it sufficient that I have kept internet/mobile constant throughout the individual-level financial inclusion measures in the dataset? And if you believe this should be aggregated, what command would you recommend in my scenario on Stata? Without aggregating the individual-level data, my do-file currently looks like the following:

    Code:
    probit account_formal internet i.female age i.educ i.inc_q pop_adult GDPpc UNEMP_ILO GOVEFF POLSTA ROL i.t i.neconomy
    margins, dydx (*) post
    estimates store herearemargins
    outreg2 [herearemargins] using model_results, replace excel dec(3)
    
    probit account_formal mobile i.female age i.educ i.inc_q pop_adult GDPpc UNEMP_ILO GOVEFF POLSTA ROL i.t i.neconomy
    margins, dydx (*) post
    estimates store herearemargins
    outreg2 [herearemargins] using model_results, replace excel dec(3)
    2)Data aggregation: effect of internet/mobile on general financial inclusion
    After having looked at whether internet/mobile has an effect on financial inclusion, I want to see if it has an effect on female financial inclusion. Here, I am using an interaction variable between female and internet or mobile. Again, would you aggregate the individual-level data of female to the country level? Until now, I am using the following do-file script:

    Code:
    probit account_formal c.internet##c.female age i.educ i.inc_q pop_adult GDPpc UNEMP_ILO GOVEFF POLSTA ROL i.t i.neconomy
    margins, dydx (*) post
    estimates store herearemargins
    outreg2 [herearemargins] using model_results, replace excel dec(4)
    
    probit account_formal c.mobile##c.female age i.educ i.inc_q pop_adult GDPpc UNEMP_ILO GOVEFF POLSTA ROL i.t i.neconomy
    margins, dydx (*) post
    estimates store herearemargins
    outreg2 [herearemargins] using model_results, replace excel dec(4)
    I would highly appreciate any advice or help on this. Please let me know if you need additional information.

    Many thanks and kind regards,
    Julie
Working...
X