cleaning_and_tables_03222020_FINAL.txt

---------------------------------------------------------------------------------------------------
      name:  <unnamed>
       log:  XXX/jhr_output/cleaning_a
> nd_tables_03222020.txt
  log type:  text
 opened on:  22 Mar 2021, 11:25:11

.         global figdir XXX/figs_miles

.         global outsheet XXX/jhr_outp
> ut

. 
. 
. foreach var in 5 {
  2. if `to_`var'' {
  3.         use $datadir/initial_data
  4.         local yos = `var'
  5.         local group to_`var'
  6. }
  7. }

. 
. 
. compress
  variable filedt_s was float now int
  variable sum_combat was float now byte
  variable min_combat_filedt_s was float now int
  variable year was float now int
  variable pna_end_dt was float now int
  variable next_move_5 was int now byte
  variable next_move_8 was int now byte
  variable next_move_10 was int now byte
  variable next_move_12 was int now byte
  variable next_move_20 was int now byte
  variable msa was float now int
  variable state_fips was float now byte
  variable re was float now byte
  variable edgrp4 was float now byte
  variable metaread_old was float now int
  variable female was float now byte
  variable metaread was float now int
  variable statefip was float now byte
  variable age_marr was float now int
  variable date_first_kids was float now int
  variable date_first_notmarr was float now int
  variable spells_startdate was float now int
  variable spells_re_enlist was float now byte
  variable num_ssn3 was float now byte
  variable maxyos was float now byte
  variable new_terms was float now byte
  variable onetermonly was float now byte
  variable evermarr was float now byte
  variable everdissolve was float now byte
  variable everkids was float now byte
  variable maxpaygr was float now byte
  variable prom_early was float now byte
  variable state was str5 now str2
  variable msa_new was str6 now str4
  (310,818,960 bytes saved)

. 
. 
. drop _merge

. 
. *** merging in the county data on ammenities
. gen county_fips = countyfips
(464,136 missing values generated)

. destring county_fips, replace
county_fips: all characters numeric; replaced as long
(464136 missing values generated)

. merge m:1 county_fips using $datadir/county_natamen.dta

    Result                           # of obs.
    -----------------------------------------
    not matched                     1,049,493
        from master                 1,046,967  (_merge==1)
        from using                      2,526  (_merge==2)

    matched                         2,838,270  (_merge==3)
    -----------------------------------------

. 
. ren _merge merge_ammen

. 
. drop metaread

. ren metaread_old metaread

. 
. replace metaread = 8840 if state == "DC"
(2,354 real changes made)

. 
. merge m:1 metaread year using $datadir/median_2br_rent_allyrs
(note: variable year was int, now float to accommodate using data's values)

    Result                           # of obs.
    -----------------------------------------
    not matched                     1,759,829
        from master                 1,755,097  (_merge==1)
        from using                      4,732  (_merge==2)

    matched                         2,132,666  (_merge==3)
    -----------------------------------------

. ren _merge merge_2br

. 
. drop statefip

. 
.         gen statefip = .
(3,892,495 missing values generated)

. 
.         replace statefip = 1 if state == "AL"
(43,114 real changes made)

.         replace statefip = 2 if state == "AK"
(69,084 real changes made)

.         replace statefip = 4 if state == "AZ"
(41,230 real changes made)

.         replace statefip = 5 if state == "AR"
(369 real changes made)

.         replace statefip = 6 if state == "CA"
(54,014 real changes made)

.         replace statefip = 8 if state == "CO"
(118,248 real changes made)

.         replace statefip = 9 if state == "CT"
(66 real changes made)

.         replace statefip = 10 if state == "DE"
(11 real changes made)

.         replace statefip = 12 if state == "FL"
(10,295 real changes made)

.         replace statefip = 13 if state == "GA"
(425,574 real changes made)

.         replace statefip = 15 if state == "HI"
(128,075 real changes made)

.         replace statefip = 16 if state == "ID"
(11 real changes made)

.         replace statefip = 17 if state == "IL"
(1,165 real changes made)

.         replace statefip = 18 if state == "IN"
(4,270 real changes made)

.         replace statefip = 19 if state == "IA"
(169 real changes made)

.         replace statefip = 20 if state == "KS"
(91,804 real changes made)

.         replace statefip = 21 if state == "KY"
(271,893 real changes made)

.         replace statefip = 22 if state == "LA"
(62,334 real changes made)

.         replace statefip = 23 if state == "ME"
(330 real changes made)

.         replace statefip = 24 if state == "MD"
(53,172 real changes made)

.         replace statefip = 25 if state == "MA"
(2,623 real changes made)

.         replace statefip = 26 if state == "MI"
(651 real changes made)

.         replace statefip = 27 if state == "MN"
(323 real changes made)

.         replace statefip = 28 if state == "MS"
(364 real changes made)

.         replace statefip = 29 if state == "MO"
(121,860 real changes made)

.         replace statefip = 30 if state == "MT"
(15 real changes made)

.         replace statefip = 31 if state == "NE"
(51 real changes made)

.         replace statefip = 32 if state == "NV"
(49 real changes made)

.         replace statefip = 33 if state == "NH"
(23 real changes made)

.         replace statefip = 34 if state == "NJ"
(5,743 real changes made)

.         replace statefip = 35 if state == "NM"
(3,656 real changes made)

.         replace statefip = 36 if state == "NY"
(100,405 real changes made)

.         replace statefip = 37 if state == "NC"
(281,027 real changes made)

.         replace statefip = 38 if state == "ND"
(10 real changes made)

.         replace statefip = 39 if state == "OH"
(541 real changes made)

.         replace statefip = 40 if state == "OK"
(119,931 real changes made)

.         replace statefip = 41 if state == "OR"
(261 real changes made)

.         replace statefip = 42 if state == "PA"
(1,690 real changes made)

.         replace statefip = 44 if state == "RI"
(37 real changes made)

.         replace statefip = 45 if state == "SC"
(178,374 real changes made)

.         replace statefip = 46 if state == "SD"
(62 real changes made)

.         replace statefip = 47 if state == "TN"
(347 real changes made)

.         replace statefip = 48 if state == "TX"
(474,249 real changes made)

.         replace statefip = 49 if state == "UT"
(793 real changes made)

.         replace statefip = 50 if state == "VT"
(14 real changes made)

.         replace statefip = 51 if state == "VA"
(168,855 real changes made)

.         replace statefip = 53 if state == "WA"
(145,885 real changes made)

.         replace statefip = 54 if state == "WV"
(183 real changes made)

.         replace statefip = 55 if state == "WI"
(426 real changes made)

.         replace statefip = 56 if state == "WY"
(11 real changes made)

. 
. *ren state_fip statefip
. merge m:1 statefip year using $datadir/median_2br_rent_state_allyrs_2
(label statefip_lbl already defined)

    Result                           # of obs.
    -----------------------------------------
    not matched                       910,204
        from master                   909,961  (_merge==1)
        from using                        243  (_merge==2)

    matched                         2,982,534  (_merge==3)
    -----------------------------------------

. ren _merge merge_2br_state

. 
. replace md_rrentgrsi = md_rrentgrsi_state if md_rrentgrsi == . & merge_2br_state == 3
(868,028 real changes made)

. 
. 
. *** merge in share employment data using msa
.         drop shcon shman shgov shmil 

.         
.         merge m:1 msa year using $datadir/BEA_msa, keepusing(shcon shman shgov shmil shsvc)

    Result                           # of obs.
    -----------------------------------------
    not matched                     1,593,376
        from master                 1,580,679  (_merge==1)
        from using                     12,697  (_merge==2)

    matched                         2,312,059  (_merge==3)
    -----------------------------------------

.         foreach var in shcon shman shgov shmil shsvc {
  2.         ren `var' `var'msa
  3.         }

.         tab _merge

                 _merge |      Freq.     Percent        Cum.
------------------------+-----------------------------------
        master only (1) |  1,580,679       40.47       40.47
         using only (2) |     12,697        0.33       40.80
            matched (3) |  2,312,059       59.20      100.00
------------------------+-----------------------------------
                  Total |  3,905,435      100.00

.         drop if _merge == 2
(12,697 observations deleted)

.         ren _merge merge_msa2

.         sort state_fips year

.         merge m:1 state_fips year using $datadir/BEA_state, keepusing(shcon shman shgov shmil shs
> vc)

    Result                           # of obs.
    -----------------------------------------
    not matched                       944,285
        from master                   943,130  (_merge==1)
        from using                      1,155  (_merge==2)

    matched                         2,949,608  (_merge==3)
    -----------------------------------------

.         tab _merge

                 _merge |      Freq.     Percent        Cum.
------------------------+-----------------------------------
        master only (1) |    943,130       24.22       24.22
         using only (2) |      1,155        0.03       24.25
            matched (3) |  2,949,608       75.75      100.00
------------------------+-----------------------------------
                  Total |  3,893,893      100.00

. 
.         foreach var in shcon shman shgov shmil shsvc {
  2.         ren `var' `var'state
  3.         }

.         drop if _merge == 2
(1,155 observations deleted)

.         ren _merge merge_state_year2

.         
.         foreach var in shcon shman shgov shmil shsvc {
  2.         gen `var' = `var'msa if merge_msa2 == 3
  3.         replace `var' = `var'state if merge_state_year2 == 3 & (merge_msa2 ~= 3 | msa == . | `
> var'msa == .)
  4.         }
(1,580,693 missing values generated)
(657,953 real changes made)
(1,580,752 missing values generated)
(657,796 real changes made)
(1,580,679 missing values generated)
(658,000 real changes made)
(1,580,679 missing values generated)
(658,000 real changes made)
(1,580,679 missing values generated)
(658,000 real changes made)

. 
. *** creating the first arloc for someone - for clustering
.         sort id training filedt_s

.         bys id training: gen first = _n

.         replace first = 0 if first ~= 1
(3,142,410 real changes made)

. 
.         egen arlocnum = group(arloc)
(7,501 missing values generated)

.         gen tempx = arlocnum if first == 1 & training == 0
(3,514,368 missing values generated)

.         bys id: egen first_arlocnum = max(tempx) 
(7,501 missing values generated)

.         drop tempx

. 
. **** first_term variable
.         sort id filedt_s

.         by id: gen numterm = _n

.         gen tempterm6 = real(terms) if numterm == 1
(3,517,020 missing values generated)

.         
.         bys id: egen term_first = max(tempterm6)
(32,333 missing values generated)

.         
. drop tempterm6

. 
.         gen tempterm6 = terms == "6" & numterm == 1

.         
.         bys id: egen term6 = max(tempterm6)

. 
. forvalues x=3/4 {
  2.         gen tempterm`x' = terms == "`x'" & numterm == 1
  3.         
.         bys id: egen term`x' = max(tempterm`x')
  4.         }

. forvalues x=5/5 {
  2.         gen tempterm`x' = terms == "`x'" & numterm == 1
  3.         
.         bys id: egen terms`x' = max(tempterm`x')
  4.         }

. 
. ****** location variables
. 
. *** creating weighted average epops
. 
.         ** time under an epop if not in training
.         gen epoptime = filedt_s[_n+1] - filedt_s if training == 0 & id == id[_n+1] & training[_n+
> 1] == 0
(1,956,313 missing values generated)

.         gen epop_x_time = epop * epoptime if training == 0
(2,429,324 missing values generated)

.         sort id filedt_s

.         by id: gen sumtime = sum(epoptime) if epop ~= .
(922,679 missing values generated)

.         
.         sort id filedt_s

.         by id: gen sum_epop_x_time = sum(epop_x_time)

.         gen avg_epop = sum_epop_x_time/sumtime
(1,788,656 missing values generated)

.         replace avg_epop = . if filedt_s > td(31dec2010)
(11,888 real changes made, 11,888 to missing)

.         replace avg_epop = . if sum_epop_x_time == 0
(0 real changes made)

. 
. *** creating weighted average cell shares using the same method as epop
.         
.         sort id training filedt_s

.         
.         gen celltime = filedt_s[_n+1] - filedt_s if training == 0 & id == id[_n+1] & training[_n+
> 1] == 0
(1,667,995 missing values generated)

.         gen cell_x_time = cellsharei * celltime if training == 0
(2,264,592 missing values generated)

.         sort id filedt_s

.         by id: gen sumtime2 = sum(celltime) if cellsharei ~= .
(931,079 missing values generated)

. 
.         sort id filedt_s

.         by id: gen sum_cell_x_time = sum(cell_x_time)

.         gen avg_cellshare = sum_cell_x_time/sumtime2
(1,741,560 missing values generated)

.         replace avg_cellshare = . if sum_cell_x_time == 0
(0 real changes made)

.         
. *** creating weighted average cell poi using the same method as epop
.         
.         drop celltime cell_x_time sumtime2 sum_cell_x_time

.         sort id training filedt_s

.         
.         gen celltime = filedt_s[_n+1] - filedt_s if training == 0 & id == id[_n+1] & training[_n+
> 1] == 0
(1,667,995 missing values generated)

.         gen cell_x_time = cellpopi * celltime if training == 0 & merge_marriage_market_msa == 3
(2,678,151 missing values generated)

.         sort id filedt_s

.         by id: gen sumtime2 = sum(celltime) if cellpopi ~= .
(931,079 missing values generated)

. 
.         sort id filedt_s

.         by id: gen sum_cell_x_time = sum(cell_x_time)

.         gen avg_cellpop = sum_cell_x_time/sumtime2
(1,741,560 missing values generated)

.         replace avg_cellpop = . if sum_cell_x_time == 0
(471,750 real changes made, 471,750 to missing)

.         
.         gen not_in_msa = merge_marriage_market_msa ~= 3 & training == 0

.         bys id: egen ever_not_in_msa = max(not_in_msa)

.         
.         replace avg_cellpop = . if ever_not_in_msa == 1
(897,006 real changes made, 897,006 to missing)

.         
. *** creating weighted average 2br rent price
.         
.         drop celltime cell_x_time sumtime2 sum_cell_x_time

.         sort id training filedt_s

.         
.         gen celltime = filedt_s[_n+1] - filedt_s if training == 0 & id == id[_n+1] & training[_n+
> 1] == 0
(1,667,995 missing values generated)

.         gen cell_x_time = md_rrentgrsi * celltime if training == 0 & md_rrentgrsi ~= .
(2,249,167 missing values generated)

.         sort id filedt_s

.         by id: gen sumtime2 = sum(celltime) if md_rrentgrsi ~= .
(887,535 missing values generated)

. 
.         sort id filedt_s

.         by id: gen sum_cell_x_time = sum(cell_x_time)

.         gen avg_cell2br = sum_cell_x_time/sumtime2
(1,721,248 missing values generated)

.         replace avg_cell2br = . if sum_cell_x_time == 0
(0 real changes made)

.         
.         gen not_in_2br = md_rrentgrsi == . & training == 0

.         bys id: egen ever_not_in_2br = max(not_in_2br)

.         
.         replace avg_cell2br = . if ever_not_in_2br == 1 
(894,004 real changes made, 894,004 to missing)

. 
. *** creating weighted average ammenities
.         
.         drop celltime cell_x_time sumtime2 sum_cell_x_time

.         sort id training filedt_s

.         
.         gen celltime = filedt_s[_n+1] - filedt_s if training == 0 & id == id[_n+1] & training[_n+
> 1] == 0
(1,667,995 missing values generated)

.         gen cell_x_time = amenityscale * celltime if training == 0 & amenityscale ~= .
(2,371,684 missing values generated)

.         sort id filedt_s

.         by id: gen sumtime2 = sum(celltime) if amenityscale ~= .
(1,051,942 missing values generated)

. 
.         sort id filedt_s

.         by id: gen sum_cell_x_time = sum(cell_x_time)

.         gen avg_cell_amm = sum_cell_x_time/sumtime2
(1,888,119 missing values generated)

.         replace avg_cell_amm = . if sum_cell_x_time == 0
(4 real changes made, 4 to missing)

.         
.         gen not_in_amm = amenityscale == . & training == 0

.         bys id: egen ever_not_in_amm = max(not_in_amm)

.         
.         replace avg_cell_amm = . if ever_not_in_amm == 1
(981,396 real changes made, 981,396 to missing)

.         
. *** shcon shman shgov shmil shsvc
. foreach var in shcon shman shgov shmil shsvc {
  2.         drop celltime cell_x_time sumtime2 sum_cell_x_time
  3.         sort id training filedt_s
  4.         
.         gen celltime = filedt_s[_n+1] - filedt_s if training == 0 & id == id[_n+1] & training[_n+
> 1] == 0
  5.         gen cell_x_time = `var' * celltime if training == 0 & `var' ~= .
  6.         sort id filedt_s
  7.         by id: gen sumtime2 = sum(celltime) if `var' ~= .
  8. 
.         sort id filedt_s
  9.         by id: gen sum_cell_x_time = sum(cell_x_time)
 10.         gen avg_`var' = sum_cell_x_time/sumtime2
 11.         replace avg_`var' = . if sum_cell_x_time == 0
 12.         
.         gen not_in_`var' = `var' == . & training == 0
 13.         bys id: egen ever_not_in_`var' = max(not_in_`var')
 14.         replace avg_`var' = . if filedt_s > td(31dec2010)
 15. 
.         replace avg_`var' = . if ever_not_in_`var' == 1
 16. }
(1,667,995 missing values generated)
(2,253,592 missing values generated)
(922,740 missing values generated)
(1,751,705 missing values generated)
(0 real changes made)
(11,911 real changes made, 11,911 to missing)
(970,066 real changes made, 970,066 to missing)
(1,667,995 missing values generated)
(2,253,753 missing values generated)
(922,956 missing values generated)
(1,751,938 missing values generated)
(0 real changes made)
(11,908 real changes made, 11,908 to missing)
(970,137 real changes made, 970,137 to missing)
(1,667,995 missing values generated)
(2,253,557 missing values generated)
(922,679 missing values generated)
(1,751,659 missing values generated)
(0 real changes made)
(11,912 real changes made, 11,912 to missing)
(969,945 real changes made, 969,945 to missing)
(1,667,995 missing values generated)
(2,253,557 missing values generated)
(922,679 missing values generated)
(1,751,659 missing values generated)
(0 real changes made)
(11,912 real changes made, 11,912 to missing)
(969,945 real changes made, 969,945 to missing)
(1,667,995 missing values generated)
(2,253,557 missing values generated)
(922,679 missing values generated)
(1,751,659 missing values generated)
(0 real changes made)
(11,912 real changes made, 11,912 to missing)
(969,945 real changes made, 969,945 to missing)

. 
. 
. 
. *** going to create worse move if moving to a location with more than 1 standard deviation from t
> he previous.  
. 
. ** first need to create standard deviation of all possible locations
. 
. foreach var in epop amenityrank md_rrentgrsi cellsharei shcon shman shgov shmil shsvc {
  2.         egen tempz2_`var' = std(`var')
  3.         }
(922,679 missing values generated)
(1,051,942 missing values generated)
(887,535 missing values generated)
(931,079 missing values generated)
(922,740 missing values generated)
(922,956 missing values generated)
(922,679 missing values generated)
(922,679 missing values generated)
(922,679 missing values generated)

.         
. *** moving to a worse location in terms of cell size
.         sort id training filedt_s

.         gen worse_location = tempz2_cellsharei < (tempz2_cellsharei[_n-1]-1)  if training == 0 & 
> training[_n-1] == 0 & id == id[_n-1] & abroad ~= 1 & cellsharei[_n-1] ~= . & cellsharei ~= . & mo
> ve_not == 1
(3,751,778 missing values generated)

.         gen better_location = (tempz2_cellsharei[_n-1]+1) < tempz2_cellsharei  if training == 0 &
>  training[_n-1] == 0 & id == id[_n-1] & abroad ~= 1 & cellsharei ~= . & cellsharei[_n-1] ~= . & m
> ove_not == 1
(3,751,778 missing values generated)

.         
.         gen worse_epop = tempz2_epop < (tempz2_epop[_n-1]-1) if training == 0 & training[_n-1] ==
>  0 & id == id[_n-1] & abroad ~= 1 & epop[_n-1] ~= . & epop ~= . & move_not == 1
(3,752,945 missing values generated)

.         gen better_epop = (tempz2_epop[_n-1]+1) < (tempz2_epop)  if training == 0 & training[_n-1
> ] == 0 & id == id[_n-1] & abroad ~= 1 & epop[_n-1] ~= . & epop ~= . & move_not == 1
(3,752,945 missing values generated)

.         
.         
.         gen worse_amm = tempz2_amenityrank < (tempz2_amenityrank[_n-1]-1)  if training == 0 & tra
> ining[_n-1] == 0 & id == id[_n-1] & abroad ~= 1 & amenityrank[_n-1] ~= . & amenityrank ~= . & mov
> e_not == 1
(3,783,244 missing values generated)

.         gen better_amm = (tempz2_amenityrank[_n-1]+1) < (tempz2_amenityrank) if training == 0 & t
> raining[_n-1] == 0 & id == id[_n-1] & abroad ~= 1 & amenityrank[_n-1] ~= . & amenityrank ~= . & m
> ove_not == 1
(3,783,244 missing values generated)

. 
. foreach var in shcon shman shgov shmil shsvc {
  2.         gen worse_`var' = tempz2_`var' < (tempz2_`var'[_n-1]-1)  if training == 0 & training[_
> n-1] == 0 & id == id[_n-1] & abroad ~= 1 & `var'[_n-1] ~= . & `var' ~= . & move_not == 1
  3.         gen better_`var' = (tempz2_`var'[_n-1]+1) < (tempz2_`var') if training == 0 & training
> [_n-1] == 0 & id == id[_n-1] & abroad ~= 1 & `var'[_n-1] ~= . & `var' ~= . & move_not == 1
  4.         }
(3,752,961 missing values generated)
(3,752,961 missing values generated)
(3,753,005 missing values generated)
(3,753,005 missing values generated)
(3,752,945 missing values generated)
(3,752,945 missing values generated)
(3,752,945 missing values generated)
(3,752,945 missing values generated)
(3,752,945 missing values generated)
(3,752,945 missing values generated)

.         
.         gen more_exp_rent = tempz2_md_rrentgrsi > (tempz2_md_rrentgrsi[_n-1]+1) if move_not == 1 
> & training == 0 & training[_n-1] == 0 & id == id[_n-1]  & abroad ~= 1 & md_rrentgrsi[_n-1] ~= . &
>  md_rrentgrsi ~= . 
(3,750,151 missing values generated)

.         gen less_exp_rent = (tempz2_md_rrentgrsi[_n-1]+1) > (tempz2_md_rrentgrsi) if move_not == 
> 1 & training == 0 & training[_n-1] == 0 & id == id[_n-1] & abroad ~= 1 & md_rrentgrsi[_n-1] ~= . 
> & md_rrentgrsi ~= . 
(3,750,151 missing values generated)

. 
. foreach var in tempz2_cellsharei tempz2_epop tempz2_amenityrank tempz2_md_rrentgrsi tempz2_shcon 
> tempz2_shman tempz2_shgov tempz2_shmil tempz2_shsvc {
  2.         gen ch_`var' = `var' - `var'[_n-1] if training == 0 & training[_n-1] == 0 & id == id[_
> n-1] & abroad ~= 1 & `var'[_n-1] ~= . & `var' ~= . & move_not == 1
  3. }       
(3,751,778 missing values generated)
(3,752,945 missing values generated)
(3,783,244 missing values generated)
(3,750,151 missing values generated)
(3,752,961 missing values generated)
(3,753,005 missing values generated)
(3,752,945 missing values generated)
(3,752,945 missing values generated)
(3,752,945 missing values generated)

.         
.         
. *** creating variable for initial rank
.         bys id: egen mind = min(filedt_s)
(7,501 missing values generated)

.         gen temprank = real(substr(grade, 3, 1))
(7,501 missing values generated)

.         gen initial_rank_temp = temprank if filedt_s == mind
(3,514,368 missing values generated)

.         bys id: egen initial_rank = max(initial_rank_temp)
(7,501 missing values generated)

. 
.         
. *** variable for large posts
.         bys id arloc year: gen num_id_arloc_yr = _n

.         replace num_id_arloc_yr = . if num_id_arloc_yr ~= 1
(811,491 real changes made, 811,491 to missing)

.         bys arloc year: egen tot_arloc_year = total(num_id_arloc_yr)

.         bys arloc year: gen num_arloc_yr = _n

.         summ tot_arloc_year if num_arloc_yr == 1, de

                       tot_arloc_year
-------------------------------------------------------------
      Percentiles      Smallest
 1%            1              1
 5%            1              1
10%            1              1       Obs              11,164
25%            2              1       Sum of Wgt.      11,164

50%           11                      Mean           275.9985
                        Largest       Std. Dev.       1155.77
75%           83          16656
90%          386          18293       Variance        1335803
95%         1099          19432       Skewness       8.016278
99%         5675          20213       Kurtosis       84.09162

. 
.         gen large_post = tot_arloc_year >= r(p90)

.         
.         sort id training filedt_s

.         gen large_to_small = large_post == 0 & move_not == 1 & training == 0 & large_post[_n-1] =
> = 1 & id == id[_n-1]

.         
.         
. *** creating variable for initial year
.         gen initial_year = year(mind)
(7,501 missing values generated)

.         
. *** are people in divisions and bcts?
.         gen bct = div == 1 | bdereg == 1

. 
. *** married when enter  
.         gen temp3 = married == 1 & firstobs == 1

.         bys id: egen m_ent = max(temp3)

.         drop temp3

.         
. *** kids when enter
.         gen temp3 = kids == 1 & firstobs == 1

.         bys id: egen kids_ent = max(temp3)

.         drop temp3

.         
. *** married with less than a year in the Army
. 
.         gen less_yr = (filedt_s - startdate <= 365) & filedt_s > startdate

.         gen temp5 = less_yr == 1 & marst == "M" & m_ent == 0

.         bys id: egen mar_lyr = max(temp5)

.         drop temp5

. 
. *** age when enter
.         gen temp3 = age if firstobs == 1
(3,529,368 missing values generated)

.         bys id: egen age_when_enter = max(temp3)
(162,191 missing values generated)

.         drop temp3

. 
.         ren _merge merge_old1

. 
. ** figuring out age when first get married - merging in data from age_marr.do.  also creates vari
> able for first time not married and first time with kids.
. 
.         /*age_marr is a filedt_s*/
.         gen tempage = age_marr - birthmo
(819,159 missing values generated)

.         gen date_marr = age_marr
(816,956 missing values generated)

.         gen tempage2 = tempage/365.25
(819,159 missing values generated)

.         replace tempage2 = . if tempage2 < 17 | tempage2 > 55
(400 real changes made, 400 to missing)

.         drop age_marr tempage

.         ren tempage2 age_marr

. 
. ** timing between marriage and moves
.         gen marr_filedt_s = date_marr - filedt_s if move_not == 1
(3,575,660 missing values generated)

.         gen marr2 = abs(marr_filedt_s)
(3,575,660 missing values generated)

.         bys id: egen marr3 = min(abs(marr_filedt_s))
(1,434,490 missing values generated)

.         gen marr4 = marr_filedt_s if marr2 == marr3 
(3,667,290 missing values generated)

.         by id: egen marr5 = min(marr4)
(1,434,490 missing values generated)

.         by id: gen marr_num = _n

. *       histogram marr5 if marr_num == 1, title("Time Between Marriage and Move") frac xtitle("Ma
> rriage Date - Move Date")
. *       graph export $figdir/married_move_`gr'.ps, replace
. *       histogram marr5 if marr_num == 1 & female == 0, title("Time Since Marriage and Move -- ME
> N") frac
. *       graph export $figdir/married_move_`gr'_m.ps, replace
. *       histogram marr5 if marr_num == 1 & female == 1, title("Time Since Marriage and Move -- WO
> MEN") frac
. *       graph export $figdir/married_move_`gr'_f.ps, replace
.         drop marr_filedt_s marr2 marr3 marr4 marr_num

.         ren marr5 days_to_move

.         gen months_to_move = floor((days_to_move/30))
(1,434,490 missing values generated)

. 
. 
.         gen shot_gun = abs(months_to_move) <= 3

.         replace shot_gun = . if months_to_move == .
(1,434,490 real changes made, 1,434,490 to missing)

.         
.         gen m_to_m_l6 = abs(months_to_move) <= 6

.         replace m_to_m_l6 = . if m_ent == 1 | evermarr == 0 | months_to_move == .
(2,561,312 real changes made, 2,561,312 to missing)

.         
.         gen m_to_m_g6 = abs(months_to_move) > 6

.         replace m_to_m_g6 = . if  m_ent == 1 | evermarr == 0 | months_to_move == .
(2,561,312 real changes made, 2,561,312 to missing)

.         
. *** people who were ever abroad
.         bys id: egen everabroad = max(abroad)
(7,501 missing values generated)

.         
.         
.         *** first move
.         gen date_move = filedt_s if move_not == 1
(3,493,108 missing values generated)

.         bys id: egen min_date_move = min(date_move)
(803,906 missing values generated)

.         
. *** same as above, but conditioning on those who were not married 6 months before a move and did 
> not move again within 24 months. - just for FIRST moves
. 
. *** creating a histogram which shows when people's first moves are.  
. 
. gen time_since_start_30 = floor((min_date_move - startdate)/30)
(803,906 missing values generated)

. *histogram time_since_start_30 if filedt_s == min_date_move
. 
. *graph export $figdir/time_since_start_30.pdf, replace
. 
. preserve 

. keep time_since_start_30 filedt_s min_date_move term3

. save $datadir/app_fig_1, replace
file XXX/dta/app_fig_1.dta saved

. restore

. 
. /*
> *** Data for figure is in jhr_app_fig_1.do
> ** Appendix C Figure 1
> 
> histogram time_since_start_30 if filedt_s == min_date_move & term3 == 1
> 
> graph export $figdir/time_since_start_30_term3.pdf, replace
> tab time_since_start_30 if filedt_s == min_date_move & term3 == 1
> */
. 
. ***** creating yearly variables for new tables eight
. 
. gen yr_1 = filedt_s - startdate < 365 if startdate ~= . & filedt_s ~= .
(7,501 missing values generated)

. gen yr_2 = filedt_s - startdate < 730 & yr_1 == 0 if startdate ~= . & filedt_s ~= .
(7,501 missing values generated)

. gen yr_3 = filedt_s - startdate < 1095 & yr_1 == 0 & yr_2 == 0 if startdate ~= . & filedt_s ~= .
(7,501 missing values generated)

. gen yr_4 = filedt_s - startdate < 1460 & yr_1 == 0 & yr_2 == 0 & yr_3 == 0 if startdate ~= . & fi
> ledt_s ~= .
(7,501 missing values generated)

. gen yr_5 = filedt_s - startdate < 1825 & yr_1 == 0 & yr_2 == 0 & yr_3 == 0 & yr_4 == 0 if startda
> te ~= . & filedt_s ~= .
(7,501 missing values generated)

. 
. 
. bys id: egen lastdate = max(filedt_s)
(7,501 missing values generated)

. 
. egen arloc_unique = group(arloc)
(7,501 missing values generated)

. gen temp_first = arloc_unique if first == 1
(3,142,411 missing values generated)

. gen temp_last = arloc_unique if filedt_s == lastdate
(3,514,368 missing values generated)

. 
. 
. bys id: egen first_arloc = max(temp_first)
(7,501 missing values generated)

. bys id: egen last_arloc = max(temp_last)
(7,501 missing values generated)

. 
. drop temp_first temp_last lastdate

. 
. forvalues x=1/5 {
  2. bys id yr_`x': egen married_yr_`x' = max(married) if yr_`x' == 1
  3. bys id: egen ev_marr_yr_`x' = max(married_yr_`x')
  4. drop married_yr_`x'
  5. bys id yr_`x': egen move_yr_`x' = max(move_not) if yr_`x' == 1
  6. bys id: egen ev_move_yr_`x' = max(move_yr_`x')
  7. drop move_yr_`x'
  8. 
. foreach var in more_exp_rent less_exp_rent worse_location better_location worse_epop better_epop 
> worse_amm better_amm  {
  9. bys id yr_`x': egen `var'_yr_`x' = max(`var') if yr_`x' == 1
 10. bys id: egen ev_`var'_yr_`x' = max(`var'_yr_`x')
 11. drop `var'_yr_`x' 
 12. }
 13. foreach var in tempz2_cellsharei tempz2_epop tempz2_amenityrank tempz2_md_rrentgrsi tempz2_shc
> on tempz2_shman tempz2_shgov tempz2_shmil tempz2_shsvc {
 14. bys id yr_`x': egen ch_`var'_yr_`x' = max(ch_`var') if yr_`x' == 1
 15. bys id: egen ev_ch_`var'_yr_`x' = max(ch_`var'_yr_`x')
 16. drop ch_`var'_yr_`x' 
 17. }
 18. 
. bys id yr_`x': gen dist_less300_yr_`x' = distance_move_not < 300 & ev_move_yr_`x' == 1 & yr_`x' =
> = 1
 19. bys id: egen ev_dist_less300_yr_`x' = max(dist_less300_yr_`x')
 20. 
. bys id yr_`x': gen dist_300_1000_yr_`x' = distance_move_not >= 300 & distance_move_not < 1000 & e
> v_move_yr_`x' == 1 & yr_`x' == 1
 21. bys id: egen ev_dist_300_1000_yr_`x' = max(dist_300_1000_yr_`x')
 22. 
. bys id yr_`x': gen dist_1000_2000_yr_`x' = distance_move_not >= 1000 & distance_move_not < 2000 &
>  ev_move_yr_`x' == 1 & yr_`x' == 1
 23. bys id: egen ev_dist_1000_2000_yr_`x' = max(dist_1000_2000_yr_`x')
 24. 
. bys id yr_`x': gen dist_2000_3000_yr_`x' = distance_move_not >= 2000 & distance_move_not < 3000 &
>  ev_move_yr_`x' == 1 & yr_`x' == 1
 25. bys id: egen ev_dist_2000_3000_yr_`x' = max(dist_2000_3000_yr_`x')
 26. 
. bys id yr_`x': gen dist_gr_3000_yr_`x' = distance_move_not > 3000 & ev_move_yr_`x' == 1 & yr_`x' 
> == 1 & distance_move_not ~= .
 27. bys id: egen ev_dist_gr_3000_yr_`x' = max(dist_gr_3000_yr_`x')
 28. 
. }
(2,493,359 missing values generated)
(13,467 missing values generated)
(2,491,554 missing values generated)
(7,572 missing values generated)
(3,873,840 missing values generated)
(3,844,633 missing values generated)
(3,873,840 missing values generated)
(3,844,633 missing values generated)
(3,874,326 missing values generated)
(3,845,858 missing values generated)
(3,874,326 missing values generated)
(3,845,858 missing values generated)
(3,873,868 missing values generated)
(3,844,698 missing values generated)
(3,873,868 missing values generated)
(3,844,698 missing values generated)
(3,876,159 missing values generated)
(3,850,969 missing values generated)
(3,876,159 missing values generated)
(3,850,969 missing values generated)
(3,874,326 missing values generated)
(3,845,858 missing values generated)
(3,873,868 missing values generated)
(3,844,698 missing values generated)
(3,876,159 missing values generated)
(3,850,969 missing values generated)
(3,873,840 missing values generated)
(3,844,633 missing values generated)
(3,873,868 missing values generated)
(3,844,698 missing values generated)
(3,873,872 missing values generated)
(3,844,710 missing values generated)
(3,873,868 missing values generated)
(3,844,698 missing values generated)
(3,873,868 missing values generated)
(3,844,698 missing values generated)
(3,873,868 missing values generated)
(3,844,698 missing values generated)
(3,292,503 missing values generated)
(15,332 missing values generated)
(3,291,628 missing values generated)
(7,501 missing values generated)
(3,870,787 missing values generated)
(3,798,437 missing values generated)
(3,870,787 missing values generated)
(3,798,437 missing values generated)
(3,871,070 missing values generated)
(3,799,680 missing values generated)
(3,871,070 missing values generated)
(3,799,680 missing values generated)
(3,870,831 missing values generated)
(3,798,632 missing values generated)
(3,870,831 missing values generated)
(3,798,632 missing values generated)
(3,876,694 missing values generated)
(3,824,405 missing values generated)
(3,876,694 missing values generated)
(3,824,405 missing values generated)
(3,871,070 missing values generated)
(3,799,680 missing values generated)
(3,870,831 missing values generated)
(3,798,632 missing values generated)
(3,876,694 missing values generated)
(3,824,405 missing values generated)
(3,870,787 missing values generated)
(3,798,437 missing values generated)
(3,870,831 missing values generated)
(3,798,632 missing values generated)
(3,870,843 missing values generated)
(3,798,688 missing values generated)
(3,870,831 missing values generated)
(3,798,632 missing values generated)
(3,870,831 missing values generated)
(3,798,632 missing values generated)
(3,870,831 missing values generated)
(3,798,632 missing values generated)
(3,260,216 missing values generated)
(10,832 missing values generated)
(3,259,738 missing values generated)
(7,501 missing values generated)
(3,819,947 missing values generated)
(3,601,319 missing values generated)
(3,819,947 missing values generated)
(3,601,319 missing values generated)
(3,820,686 missing values generated)
(3,604,295 missing values generated)
(3,820,686 missing values generated)
(3,604,295 missing values generated)
(3,820,083 missing values generated)
(3,601,900 missing values generated)
(3,820,083 missing values generated)
(3,601,900 missing values generated)
(3,833,114 missing values generated)
(3,651,890 missing values generated)
(3,833,114 missing values generated)
(3,651,890 missing values generated)
(3,820,686 missing values generated)
(3,604,295 missing values generated)
(3,820,083 missing values generated)
(3,601,900 missing values generated)
(3,833,114 missing values generated)
(3,651,890 missing values generated)
(3,819,947 missing values generated)
(3,601,319 missing values generated)
(3,820,085 missing values generated)
(3,601,912 missing values generated)
(3,820,103 missing values generated)
(3,601,976 missing values generated)
(3,820,083 missing values generated)
(3,601,900 missing values generated)
(3,820,083 missing values generated)
(3,601,900 missing values generated)
(3,820,083 missing values generated)
(3,601,900 missing values generated)
(3,238,588 missing values generated)
(10,844 missing values generated)
(3,238,046 missing values generated)
(7,501 missing values generated)
(3,758,318 missing values generated)
(3,341,750 missing values generated)
(3,758,318 missing values generated)
(3,341,750 missing values generated)
(3,759,550 missing values generated)
(3,346,931 missing values generated)
(3,759,550 missing values generated)
(3,346,931 missing values generated)
(3,758,630 missing values generated)
(3,343,013 missing values generated)
(3,758,630 missing values generated)
(3,343,013 missing values generated)
(3,799,032 missing values generated)
(3,499,820 missing values generated)
(3,799,032 missing values generated)
(3,499,820 missing values generated)
(3,759,550 missing values generated)
(3,346,931 missing values generated)
(3,758,630 missing values generated)
(3,343,013 missing values generated)
(3,799,032 missing values generated)
(3,499,820 missing values generated)
(3,758,318 missing values generated)
(3,341,750 missing values generated)
(3,758,645 missing values generated)
(3,343,077 missing values generated)
(3,758,673 missing values generated)
(3,343,205 missing values generated)
(3,758,630 missing values generated)
(3,343,013 missing values generated)
(3,758,630 missing values generated)
(3,343,013 missing values generated)
(3,758,630 missing values generated)
(3,343,013 missing values generated)
(3,298,068 missing values generated)
(11,343 missing values generated)
(3,297,487 missing values generated)
(7,501 missing values generated)
(3,777,224 missing values generated)
(3,373,843 missing values generated)
(3,777,224 missing values generated)
(3,373,843 missing values generated)
(3,778,500 missing values generated)
(3,379,557 missing values generated)
(3,778,500 missing values generated)
(3,379,557 missing values generated)
(3,783,240 missing values generated)
(3,398,450 missing values generated)
(3,783,240 missing values generated)
(3,398,450 missing values generated)
(3,805,071 missing values generated)
(3,492,090 missing values generated)
(3,805,071 missing values generated)
(3,492,090 missing values generated)
(3,778,500 missing values generated)
(3,379,557 missing values generated)
(3,783,240 missing values generated)
(3,398,450 missing values generated)
(3,805,071 missing values generated)
(3,492,090 missing values generated)
(3,777,224 missing values generated)
(3,373,843 missing values generated)
(3,783,260 missing values generated)
(3,398,533 missing values generated)
(3,783,295 missing values generated)
(3,398,703 missing values generated)
(3,783,240 missing values generated)
(3,398,450 missing values generated)
(3,783,240 missing values generated)
(3,398,450 missing values generated)
(3,783,240 missing values generated)
(3,398,450 missing values generated)

. 
. *** creating arloc fixed effects for each year
.         forvalues x=1/5 {
  2.         sort id yr_`x' filedt_s training
  3.         by id yr_`x': gen yrnum = _n
  4.         gen arloc_yr_`x'_1 = arloc_unique if yr_`x' == 1 & training == 0 & yrnum == 1
  5.         gen arloc_yr_`x'_2 = arloc_unique if yr_`x' == 1 & training == 0 & yrnum == 2
  6.         gen arloc_yr_`x'_3 = arloc_unique if yr_`x' == 1 & training == 0 & yrnum == 3
  7.         bys id: egen ev_arloc_yr_`x'_1 = max(arloc_yr_`x'_1)
  8.         bys id: egen ev_arloc_yr_`x'_2 = max(arloc_yr_`x'_2)
  9.         bys id: egen ev_arloc_yr_`x'_3 = max(arloc_yr_`x'_3)
 10.         drop yrnum arloc_yr_`x'_*
 11.         }
(3,869,466 missing values generated)
(3,788,438 missing values generated)
(3,701,467 missing values generated)
(3,666,213 missing values generated)
(2,916,025 missing values generated)
(2,021,401 missing values generated)
(3,552,089 missing values generated)
(3,802,476 missing values generated)
(3,837,033 missing values generated)
(456,090 missing values generated)
(2,835,051 missing values generated)
(3,184,183 missing values generated)
(3,549,998 missing values generated)
(3,796,487 missing values generated)
(3,829,267 missing values generated)
(438,968 missing values generated)
(2,782,797 missing values generated)
(3,096,759 missing values generated)
(3,564,048 missing values generated)
(3,783,574 missing values generated)
(3,824,941 missing values generated)
(588,006 missing values generated)
(2,664,502 missing values generated)
(3,061,776 missing values generated)
(3,559,783 missing values generated)
(3,799,592 missing values generated)
(3,840,710 missing values generated)
(542,075 missing values generated)
(2,831,362 missing values generated)
(3,250,720 missing values generated)

. 
. *** only including if have data on the characteristics of the move
. forvalues x=1/5 {
  2. foreach var in more_exp_rent less_exp_rent worse_location better_location worse_epop better_ep
> op worse_amm better_amm ch_tempz2_cellsharei ch_tempz2_epop ch_tempz2_amenityrank ch_tempz2_md_rr
> entgrsi ch_tempz2_shcon ch_tempz2_shman ch_tempz2_shgov ch_tempz2_shmil ch_tempz2_shsvc {
  3. replace ev_`var'_yr_`x' = 0 if ev_move_yr_`x' == 0
  4. }
  5. }
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,801,788 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(3,139,345 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,797,554 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,529,749 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)
(2,685,548 real changes made)

. 
. *** time since start for first move.
.         gen temp4 = min_date_move - startdate if filedt_s == min_date_move
(3,609,983 missing values generated)

.         bys id: egen time_since_start = max(temp4)
(803,906 missing values generated)

. 
. *** creating a sample for the figures
.         
.         * 1) just first moves
.         gen figure_sample = 1 if move_not == 1 & filedt_s == min_date_move
(3,609,983 missing values generated)

.         
.         * 2) in military for at least 6 months before first move
.         bys id: egen firstdate = min(filedt_s)
(7,501 missing values generated)

.         replace figure_sample = 0 if min_date_move < firstdate + 180
(6,389 real changes made)

.         
.         *3) in military for at least 2 years after first move
.         bys id: egen lastdate = max(filedt_s)
(7,501 missing values generated)

.         replace figure_sample = 0 if lastdate - min_date_move < 730
(1,781,340 real changes made)

.         
.         *4) do not move for at least 2 years after first move
.         sort id move_not filedt_s

.         replace figure_sample = 0 if filedt_s[_n+1] - filedt_s < 730 & move_not == 1 & move_not[_
> n+1] == 1 & id == id[_n+1]
(48,851 real changes made)

.         
.         *5) not married 6 months before first move
.         replace figure_sample = 0 if date_marr - filedt_s < -180
(553,388 real changes made)

.         replace figure_sample = 0 if m_ent == 1
(95,035 real changes made)

.         
.         *6) not moving from or to an abroad location
.         sort id training filedt_s

.         gen conus_conus = abroad == 0 & abroad[_n-1] == 0 & training[_n-1] == 0

.         
.         gen abroad_conus = abroad == 0 & abroad[_n-1] == 1 & training[_n-1] == 0

.         
.         gen conus_abroad = abroad == 1 & abroad[_n-1] == 0 & training[_n-1] == 0 & (substr(arloc,
>  1, 2) == "GE" | substr(arloc, 1, 2) == "GM" | substr(arloc, 1, 2) == "IT")

.         
.         gen abroad_abroad = abroad == 1 & abroad[_n-1] == 1 & training[_n-1] == 0

.         
.         summ time_since_move_not if abroad_abroad == 1, de

                     time_since_move_not
-------------------------------------------------------------
      Percentiles      Smallest
 1%          181             28
 5%          183             28
10%          212             28       Obs              23,057
25%          304             31       Sum of Wgt.      23,057

50%          427                      Mean           540.6306
                        Largest       Std. Dev.      322.8262
75%          730           1764
90%         1035           1764       Variance       104216.8
95%         1216           1795       Skewness       1.139115
99%         1522           1796       Kurtosis       3.798293

.         summ time_since_move_not if abroad_conus == 1, de

                     time_since_move_not
-------------------------------------------------------------
      Percentiles      Smallest
 1%          212            181
 5%          303            181
10%          335            181       Obs             131,918
25%          366            181       Sum of Wgt.     131,918

50%          457                      Mean           622.6333
                        Largest       Std. Dev.      319.4824
75%          791           1795
90%         1126           1796       Variance         102069
95%         1247           1796       Skewness       .9622279
99%         1520           1796       Kurtosis        3.05449

.         
.         *** saving a file to then re-randomize the moves in another do file
.         preserve

.         gen floor_tenure = floor((filedt_s - startdate)/30)
(7,501 missing values generated)

.         bys id: egen temp_age = min(age)
(7,501 missing values generated)

.         gen entry_age = floor(temp_age)
(7,501 missing values generated)

.         drop temp_age

.         keep if figure_sample == 1 
(3,838,134 observations deleted)

.         keep id move_not filedt_s min_date_move firstdate lastdate date_marr entry_age age floor_
> tenure filedt_s startdate

.         save  $datadir/figre_sample_fake_1, replace
file XXX/dta/figre_sample_fake_1.dta s
> aved

.         restore

.         
. *** creating another sample of movers where there is a longer pre-lag and a shorter post lag
.         
.         * 1) just first moves
.         gen figure_sample2 = 1 if move_not == 1 & filedt_s == min_date_move
(3,609,983 missing values generated)

.         
.         * 2) in military for at least 2 years before first move
.         replace figure_sample2 = 0 if min_date_move < firstdate + 730
(811,810 real changes made)

.         
.         *3) in military for at least 6 months after first move
.         replace figure_sample2 = 0 if lastdate - min_date_move < 180
(461,055 real changes made)

.         
.         *4) do not move for at least 6 months after first move
.         sort id move_not filedt_s

.         replace figure_sample2 = 0 if filedt_s[_n+1] - filedt_s < 180 & move_not == 1 & move_not[
> _n+1] == 1 & id == id[_n+1]
(362 real changes made)

.         
.         *5) not married 2 years before first move
.         replace figure_sample2 = 0 if date_marr - filedt_s < -730
(399,493 real changes made)

.         replace figure_sample2 = 0 if m_ent == 1
(256,622 real changes made)

.         
.         *** creating an alternate sample for figure 1 that fits the restrictions below
.         gen figure_sample_alt = figure_sample
(1,353,131 missing values generated)

. 
.         bys id: egen ever_figure_sample = max(figure_sample)
(368,217 missing values generated)

.         bys id: egen ever_figure_sample2 = max(figure_sample2)
(486,617 missing values generated)

.         
.         *** saving another file to then re-randomize the moves in another do file
.         preserve

.         gen floor_tenure = floor((filedt_s - startdate)/30)
(7,501 missing values generated)

.         bys id: egen temp_age = min(age)
(7,501 missing values generated)

.         gen entry_age = floor(temp_age)
(7,501 missing values generated)

.         drop temp_age

. 
.         keep if figure_sample2 == 1 
(3,772,251 observations deleted)

.         keep id move_not filedt_s min_date_move firstdate lastdate date_marr entry_age age floor_
> tenure filedt_s startdate

.         save  $datadir/figre_sample_fake_2, replace
file XXX/dta/figre_sample_fake_2.dta s
> aved

.         restore 

.         
. *** Create a sample of people that can be used as a comparison.
.         * 1) create sample -- had been in the army for at least six months -- a date for each per
> son that occurs at the average time and age for people
. 
.         gen figure_sample_nomove = filedt_s == firstdate

.         
.         * 2) look similar to figure_sample (balance on age and time since enter // and other char
> acteristics)
.         gen tenure = filedt_s - startdate
(7,501 missing values generated)

.         summ age tenure female ged hsd hsg asc_smc college_pl female afqsc black hispanic other_r
> ace sum_combat if figure_sample == 1

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
         age |     54,557    22.43417    3.022576   18.04931    51.8137
      tenure |     54,604    660.9838    200.5696        181       1068
      female |     54,599     .159252    .3659143          0          1
         ged |     54,316    .0719493    .2584064          0          1
         hsd |     54,316    .0053759    .0731242          0          1
-------------+---------------------------------------------------------
         hsg |     54,316    .8412991    .3654003          0          1
     asc_smc |     54,316    .0608292     .239019          0          1
  college_pl |     54,316    .0205464    .1418614          0          1
      female |     54,599     .159252    .3659143          0          1
       afqsc |     53,814    58.67629    18.24822          1         99
-------------+---------------------------------------------------------
       black |     54,604    .2629478    .4402383          0          1
    hispanic |     54,604    .0917149    .2886257          0          1
  other_race |     54,604    .0651417    .2467781          0          1
  sum_combat |     54,604    1.385851    3.692495          0         25

.         summ tenure if figure_sample == 1, de

                           tenure
-------------------------------------------------------------
      Percentiles      Smallest
 1%          183            181
 5%          304            181
10%          426            181       Obs              54,604
25%          546            181       Sum of Wgt.      54,604

50%          639                      Mean           660.9838
                        Largest       Std. Dev.      200.5696
75%          822           1068
90%          943           1068       Variance       40228.14
95%          976           1068       Skewness      -.0671783
99%         1064           1068       Kurtosis       2.608012

.                 gen figure_nomove_date = firstdate + r(mean)
(7,501 missing values generated)

.                 replace figure_sample_alt = 0 if tenure < r(mean) - 365 | tenure > r(mean) + 365
(921,812 real changes made)

. 
.         gen temp_age =  (figure_nomove_date - birthmo)/365
(10,139 missing values generated)

.         summ age if figure_sample == 1, de

                             age
-------------------------------------------------------------
      Percentiles      Smallest
 1%     18.93151       18.04931
 5%     19.54795       18.06575
10%      19.8548       18.09863       Obs              54,557
25%     20.47671       18.11233       Sum of Wgt.      54,557

50%     21.52055                      Mean           22.43417
                        Largest       Std. Dev.      3.022576
75%     23.36986        46.3863
90%     26.18082       47.36164       Variance       9.135966
95%     28.52603       48.99178       Skewness       2.149237
99%     34.12055        51.8137       Kurtosis       9.459619

.                 replace figure_sample_nomove = 0 if temp_age < r(mean) - 1 | temp_age > r(mean) +
>  1
(291,433 real changes made)

.                 replace figure_sample_alt = 0 if age < r(mean) - 1 | age > r(mean) + 1
(356,633 real changes made)

.         summ age tenure female ged hsd hsg asc_smc college_pl female afqsc black hispanic other_r
> ace sum_combat if figure_sample_nomove == 1

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
         age |     94,438    20.49505    .5732246   19.62466   21.62192
      tenure |     94,438    .0209555    2.307712          0        365
      female |     94,406    .1296422    .3359111          0          1
         ged |     92,366    .1083191    .3107847          0          1
         hsd |     92,366    .0090834    .0948737          0          1
-------------+---------------------------------------------------------
         hsg |     92,366    .8229868    .3816819          0          1
     asc_smc |     92,366    .0588961    .2354315          0          1
  college_pl |     92,366    .0007145    .0267216          0          1
      female |     94,406    .1296422    .3359111          0          1
       afqsc |     93,454    57.79757    19.01584          2         99
-------------+---------------------------------------------------------
       black |     94,438    .2397975    .4269621          0          1
    hispanic |     94,438    .1115441    .3148064          0          1
  other_race |     94,438    .0510176    .2200348          0          1
  sum_combat |     94,438    .0004659    .0450875          0          8

. 
.         replace figure_sample_nomove = 0 if figure_nomove_date < firstdate + 180        
(0 real changes made)

. 
. 
. *3) do not move for at least 2 years after filedt_s
.         sort id  filedt_s

.         replace figure_sample_nomove = 0 if min_date_move - figure_nomove_date < 730 & min_date_m
> ove ~= .
(56,331 real changes made)

.         
. *4) not married 6 months before first move
.         replace figure_sample_nomove = 0 if date_marr - figure_nomove_date < -180
(10,308 real changes made)

.         replace figure_sample_nomove = 0 if m_ent == 1
(0 real changes made)

. 
. *5) in military for at least 2 years after filedt_s
.         replace figure_sample_nomove = 0 if lastdate - figure_nomove_date < 730
(0 real changes made)

. 
.         forvalues x=1/24 {
  2.         gen t_neg`x' = 0 if figure_sample == 1 | figure_sample_nomove == 1 | figure_sample_alt
>  == 1
  3.         gen t2_neg`x' = 0 if figure_sample2 == 1
  4.         
.         gen t_kids_neg`x' = 0 if figure_sample == 1 | figure_sample_nomove == 1 | figure_sample_a
> lt == 1
  5.         gen t_div_neg`x' = 0 if m_ent == 1 | figure_sample2 == 1
  6.         }
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,028,334 missing values generated)

.         forvalues x=0/25 {
  2.         gen t_pos`x' = 0 if figure_sample == 1 | figure_sample_nomove == 1 | figure_sample_alt
>  == 1 
  3.         gen t2_pos`x' = 0 if figure_sample2 == 1 
  4.         
.         gen t_kids_pos`x' = 0 if figure_sample == 1 | figure_sample_nomove == 1 | figure_sample_a
> lt == 1 
  5.         gen t_div_pos`x' = 0 if m_ent == 1 
  6.         }
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)
(3,810,335 missing values generated)
(3,772,251 missing values generated)
(3,810,335 missing values generated)
(3,148,821 missing values generated)

.         
.         *** 1 if married at each of these month intervals, 0 if not married at each of these mont
> h intervals
.         
.         forvalues x=1/24 {
  2.         local y = `x' -1
  3.         local z = `y' * 30
  4.         replace t_neg`x' = 1 if (figure_sample == 1 | figure_sample_alt == 1) & date_marr - fi
> ledt_s < -`z' & date_marr ~= .
  5.         replace t2_neg`x' = 1 if (figure_sample2 == 1 | figure_sample_alt == 1) & date_marr - 
> filedt_s < -`z' & date_marr ~= .
  6.         
.         replace t_neg`x' = 1 if (figure_sample_nomove == 1) & date_marr - figure_nomove_date < -`
> z' & date_marr ~= .
  7. 
.         
.         }
(4,016 real changes made)
(32,749 real changes made)
(1,840 real changes made)
(3,647 real changes made)
(31,663 real changes made)
(1,534 real changes made)
(2,997 real changes made)
(29,926 real changes made)
(1,221 real changes made)
(2,190 real changes made)
(27,766 real changes made)
(924 real changes made)
(1,399 real changes made)
(25,669 real changes made)
(613 real changes made)
(680 real changes made)
(23,843 real changes made)
(288 real changes made)
(0 real changes made)
(22,095 real changes made)
(0 real changes made)
(0 real changes made)
(20,450 real changes made)
(0 real changes made)
(0 real changes made)
(18,942 real changes made)
(0 real changes made)
(0 real changes made)
(17,544 real changes made)
(0 real changes made)
(0 real changes made)
(16,168 real changes made)
(0 real changes made)
(0 real changes made)
(14,929 real changes made)
(0 real changes made)
(0 real changes made)
(13,773 real changes made)
(0 real changes made)
(0 real changes made)
(12,665 real changes made)
(0 real changes made)
(0 real changes made)
(11,630 real changes made)
(0 real changes made)
(0 real changes made)
(10,594 real changes made)
(0 real changes made)
(0 real changes made)
(9,580 real changes made)
(0 real changes made)
(0 real changes made)
(8,438 real changes made)
(0 real changes made)
(0 real changes made)
(7,294 real changes made)
(0 real changes made)
(0 real changes made)
(6,153 real changes made)
(0 real changes made)
(0 real changes made)
(5,024 real changes made)
(0 real changes made)
(0 real changes made)
(3,867 real changes made)
(0 real changes made)
(0 real changes made)
(2,733 real changes made)
(0 real changes made)
(0 real changes made)
(1,624 real changes made)
(0 real changes made)

.         
.         forvalues x=1/26 {
  2.         local y = `x' - 1
  3.         local z = `y' * 30
  4.         replace t_pos`y' = 1 if (figure_sample == 1 | figure_sample_alt == 1 | figure_sample2 
> == 1) & date_marr - filedt_s < `z' & date_marr ~= .
  5.         replace t2_pos`y' = 1 if (figure_sample2 == 1) & date_marr - filedt_s < `z' & date_mar
> r ~= .
  6. 
.         replace t_pos`y' = 1 if (figure_sample_nomove == 1) & date_marr - figure_nomove_date < `z
> ' & date_marr ~= .
  7. 
. 
.         }
(34,334 real changes made)
(32,238 real changes made)
(1,840 real changes made)
(41,881 real changes made)
(38,366 real changes made)
(2,141 real changes made)
(43,877 real changes made)
(39,936 real changes made)
(2,479 real changes made)
(45,047 real changes made)
(40,810 real changes made)
(2,806 real changes made)
(46,151 real changes made)
(41,625 real changes made)
(3,126 real changes made)
(47,341 real changes made)
(42,522 real changes made)
(3,473 real changes made)
(48,470 real changes made)
(43,399 real changes made)
(3,837 real changes made)
(49,744 real changes made)
(44,366 real changes made)
(4,165 real changes made)
(51,014 real changes made)
(45,323 real changes made)
(4,505 real changes made)
(52,357 real changes made)
(46,360 real changes made)
(4,880 real changes made)
(53,719 real changes made)
(47,387 real changes made)
(5,211 real changes made)
(55,088 real changes made)
(48,411 real changes made)
(5,532 real changes made)
(56,539 real changes made)
(49,510 real changes made)
(5,873 real changes made)
(58,057 real changes made)
(50,697 real changes made)
(6,202 real changes made)
(59,526 real changes made)
(51,849 real changes made)
(6,547 real changes made)
(60,863 real changes made)
(52,852 real changes made)
(6,910 real changes made)
(62,167 real changes made)
(53,843 real changes made)
(7,249 real changes made)
(63,426 real changes made)
(54,774 real changes made)
(7,592 real changes made)
(64,679 real changes made)
(55,688 real changes made)
(7,908 real changes made)
(65,911 real changes made)
(56,580 real changes made)
(8,258 real changes made)
(67,204 real changes made)
(57,499 real changes made)
(8,575 real changes made)
(68,462 real changes made)
(58,377 real changes made)
(8,899 real changes made)
(69,747 real changes made)
(59,262 real changes made)
(9,259 real changes made)
(71,043 real changes made)
(60,144 real changes made)
(9,594 real changes made)
(72,254 real changes made)
(60,978 real changes made)
(9,944 real changes made)
(73,573 real changes made)
(61,869 real changes made)
(10,329 real changes made)

.         
.         *** 1 if have kids at each of these month intervals, 0 if no kids at each of these month 
> intervals
.         
.         forvalues x=1/7 {
  2.         local y = `x' -1
  3.         local z = `y' * 30
  4.         replace t_kids_neg`x' = 1 if (figure_sample == 1 | figure_sample2 == 1) & date_first_k
> ids - filedt_s < -`z' & date_first_kids ~= .
  5.         }
(16,929 real changes made)
(16,419 real changes made)
(15,544 real changes made)
(14,444 real changes made)
(13,411 real changes made)
(12,548 real changes made)
(11,742 real changes made)

.         
.         forvalues x=1/26 {
  2.         local y = `x' - 1
  3.         local z = `y' * 30
  4.         replace t_kids_pos`y' = 1 if (figure_sample == 1 | figure_sample2 == 1) & date_first_k
> ids - filedt_s < `z' & date_first_kids ~= .
  5.         }
(16,929 real changes made)
(19,893 real changes made)
(21,042 real changes made)
(21,781 real changes made)
(22,505 real changes made)
(23,289 real changes made)
(24,111 real changes made)
(25,038 real changes made)
(25,886 real changes made)
(26,696 real changes made)
(27,661 real changes made)
(28,592 real changes made)
(29,587 real changes made)
(30,662 real changes made)
(31,710 real changes made)
(32,700 real changes made)
(33,659 real changes made)
(34,631 real changes made)
(35,547 real changes made)
(36,517 real changes made)
(37,498 real changes made)
(38,493 real changes made)
(39,555 real changes made)
(40,588 real changes made)
(41,547 real changes made)
(42,620 real changes made)

.         
.         save $datadir/data_for_graph_miles, replace
file XXX/dta/data_for_graph_miles.dta 
> saved

.         
. ***** Figure 1 is now created in jhr_fig_1.do
. 
. save $datadir/figs_3a_3b, replace
file XXX/dta/figs_3a_3b.dta saved

. *** Figures 3A and 3B in jhr_figs_3a_3b.do
. 
. *** dual spouse
.         gen dualspouse = spouse_id ~= .

.         replace married = 1 if dualspouse == 1
(4,206 real changes made)

.         dis "summarizing dualspouse"
summarizing dualspouse

.         summ dualspouse

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
  dualspouse |  3,892,738    .0387856    .1930835          0          1

.         
. 
. 
. *** deployment variables
.         gen temp = sum_combat >0 & sum_combat ~= .  

.         bys id: egen everdepl = max(temp)

.         
.         
. *** moves pre and post deployment
. 
.         gen pre_move_not = move_not if filedt_s < min_combat_filedt_s
(965,481 missing values generated)

.         gen post_move_not = move_not if filedt_s > min_combat_filedt_s
(2,963,614 missing values generated)

.         
. *** married 3 months before first deployment
. 
.         gen diff_marr_combat = min_combat_filedt_s - date_marr
(1,413,865 missing values generated)

.         gen marr_3mo_1deploy = (diff_marr_combat > 0 & diff_marr_combat <= 92 & diff_marr_combat 
> ~= .)

.         replace marr_3mo_1deploy = . if evermarr == 0
(1,493,629 real changes made, 1,493,629 to missing)

. 
. *** date of first non-training move
. 
.         bys id move_not: egen temp_move = min(filedt_s)
(7,501 missing values generated)

.         replace temp_move = . if move_not == 0
(3,485,607 real changes made, 3,485,607 to missing)

.         bys id: egen first_move_dt_s = min(temp_move)
(803,906 missing values generated)

.                 
. 
.         dis "count number of people in sample, check 0"
count number of people in sample, check 0

.         count if num_id_check == 1
  378,370

. 
. 
.         
. *** people who were ever in a dual relationship
.         bys id: egen everdual = max(dualspouse)

.         replace evermarr = 1 if everdual == 1
(8,206 real changes made)

.         gen ndual = 1 if everdual == 0
(454,736 missing values generated)

.         replace ndual = 0 if everdual == 1
(454,736 real changes made)

. 
. *** for people who were ever abroad, don't include them in the epop / cell share data
.         replace avg_epop = . if everabroad == 1
(859,890 real changes made, 859,890 to missing)

.         
.         replace avg_cellshare = . if everabroad == 1
(886,301 real changes made, 886,301 to missing)

. 
. 
. *** creating year, job, rank interaction variables
.         bys initial_rank pmos initial_year: gen exper_br_year_flag=1 if _n==1
(3,877,036 missing values generated)

.         replace exper_br_year_flag=0 if exper_br_year_flag==.
(3,877,036 real changes made)

.         gen exper_br_year_fe=sum(exper_br_year_flag)

.         drop exper_br_year_flag

.         
. *** creating year, job, rank, gender interaction variables
.         bys initial_rank pmos initial_year sex: gen exper_br_year_flag_s=1 if _n==1
(3,868,259 missing values generated)

.         replace exper_br_year_flag_s=0 if exper_br_year_flag_s==.
(3,868,259 real changes made)

.         gen exper_br_year_sex_fe=sum(exper_br_year_flag_s)

.         drop exper_br_year_flag_s

. 
.         
. *****************************************************
. ** Creating a number of right hand side variables
. *****************************************************
. 
. ** move_yr - did you move in a particular year? 
.         bys id: egen totmoves  = total(move)

. *** non-training moves
.         bys id: egen totmoves_not = total(move_not)

.         replace move_not = 0 if last_not == 1 & distance_move_not == 0
(6,440 real changes made)

.         replace distance_move_not = 0 if move_not ~= 1
(1,373,666 real changes made)

.         bys id: egen tot_dist_not = total(distance_move_not)

.         replace tot_dist_not = . if everabroad == 1
(2,196,528 real changes made, 2,196,528 to missing)

.         
. *** total moves pre-deployment and post deployment
.         bys id: egen pre_totmoves_not = total(pre_move_not)

.         bys id: egen post_totmoves_not = total(post_move_not)

.         
.         
. *** total worse moves
.         bys id: egen tot_worse_not = max(worse_location)
(2,567,407 missing values generated)

.         bys id: egen tot_better_not = max(better_location)      
(2,567,407 missing values generated)

.         bys id: egen tot_worse_epop = max(worse_epop)
(2,574,551 missing values generated)

.         bys id: egen tot_better_epop = max(better_epop) 
(2,574,551 missing values generated)

.         bys id: egen tot_figure_sample = max(figure_sample)
(368,217 missing values generated)

.         bys id: egen tot_conus_abroad = max(conus_abroad)

.         bys id: egen tot_abroad_conus = max(abroad_conus)

.         bys id: egen tot_abroad_abroad = max(abroad_abroad)

.         bys id: egen tot_worse_amm = max(worse_amm)
(2,854,874 missing values generated)

.         bys id: egen tot_better_amm = max(better_amm)   
(2,854,874 missing values generated)

.         bys id: egen tot_more_exp_rent = max(more_exp_rent)
(2,553,897 missing values generated)

.         bys id: egen tot_less_exp_rent = max(less_exp_rent)
(2,553,897 missing values generated)

.         
. 
. 
. *** changing the totmoves to only include non-abroad moves --- just for Germandy and Italy where 
> it is most likely to be accomanied.  
. 
.         gen abroad_move = 0 if move_not == 1 & conus == 1
(3,623,538 missing values generated)

.         sort id move_not

.         replace abroad_move = 1 if move_not == 1 & conus == 0 & (substr(arloc, 1, 2) == "GE" | su
> bstr(arloc, 1, 2) == "GM" | substr(arloc, 1, 2) == "IT") & conus[_n-1] == 1 & id == id[_n-1]
(40,709 real changes made)

.         replace abroad_move = 1 if move_not == 1 & conus == 1 & (substr(arloc[_n-1], 1, 2) == "GE
> " | substr(arloc[_n-1], 1, 2) == "GM" | substr(arloc[_n-1], 1, 2) == "IT") & id == id[_n-1]
(37,034 real changes made)

.         
.         gen abroad_move_c_o = 0 if move_not == 1
(3,499,548 missing values generated)

.         replace abroad_move_c_o = 1 if move_not == 1 & conus == 0 & (substr(arloc, 1, 2) == "GE" 
> | substr(arloc, 1, 2) == "GM" | substr(arloc, 1, 2) == "IT") & conus[_n-1] == 1 & id == id[_n-1]
(40,709 real changes made)

.         
.         gen abroad_move_o_c = 0 if move_not == 1
(3,499,548 missing values generated)

.         replace abroad_move_o_c = 1 if move_not == 1 & conus == 1 & (substr(arloc[_n-1], 1, 2) ==
>  "GE" | substr(arloc[_n-1], 1, 2) == "GM" | substr(arloc[_n-1], 1, 2) == "IT") & id == id[_n-1]
(37,034 real changes made)

.         
.         gen abroad_to_abroad_move = 0 if move_not == 1
(3,499,548 missing values generated)

.         replace abroad_to_abroad_move = 1 if move_not == 1 & (substr(arloc, 1, 2) == "GE" | subst
> r(arloc, 1, 2) == "GM" | substr(arloc, 1, 2) == "IT")  & (substr(arloc[_n-1], 1, 2) == "GE" | sub
> str(arloc[_n-1], 1, 2) == "GM" | substr(arloc[_n-1], 1, 2) == "IT") & id == id[_n-1]
(18,483 real changes made)

.         
.         gen move_not_us = move_not
(7,501 missing values generated)

.         replace move_not_us = 0 if abroad_move == 1 | (conus == 0 & move_not == 1)      
(161,024 real changes made)

.         bys id: egen totmoves_not_us = total(move_not_us)       

.         ren totmoves_not totmoves_not_w_abroad

.         bys id: egen abroad_moves_not = total(abroad_move)

.         gen totmoves_not = totmoves_not_us

.         
.         bys id: egen abroad_moves_not_c_o = total(abroad_move_c_o)

.         bys id: egen abroad_moves_not_o_c = total(abroad_move_o_c)

.         bys id: egen abroad_to_abroad_moves = total(abroad_to_abroad_move)

. 
.         ** could be domestic moves within us or within italy/germany
.         label var totmoves_not "Domestic Moves"

. 
. *       histogram time_since_move if move == 1, percent
. *       graph export $figdir/time_since_move.ps, replace
. 
. ** total time spent in training
.         sort id filedt_s

.         gen time_in_training = filedt_s - filedt_s[_n-1] if training == 1 & id == id[_n-1] & file
> dt_s > filedt_s[_n-1]
(2,965,708 missing values generated)

.         bys id: egen min_train = min(filedt_s)
(7,501 missing values generated)

.         gen mintrain = (filedt_s == min_train & training == 1)

.         replace time_in_training = filedt_s - startdate if mintrain == 1
(355,094 real changes made)

.         bys id: egen tot_time_t = total(time_in_training)

.         drop min_train mintrain

. 
. ** time overseas - NEED TO CORRECT FOR OTHER PLACES WHERE FAMILY DOES NOT MOVE (BESIDES KOREA)
.         sort id training conus filedt_s

.         gen temp_overseas = filedt_s - filedt_s[_n-1] if conus[_n-1] == 0 & id == id[_n-1] & cnty
> [_n-1] ~= "KS" & training == 0 & filedt_s > filedt_s[_n-1]
(3,504,073 missing values generated)

.         bys id: egen time_overseas = total(temp_overseas)

.         drop temp_overseas

. 
. 
. *** urban versus rural
.         gen rural = msa == .

.         replace rural = . if zipcd == "" 
(107,575 real changes made, 107,575 to missing)

.         replace rural = . if conus == 0
(784,378 real changes made, 784,378 to missing)

.         sort id training rural filedt_s

.         gen temp_rural = filedt_s - filedt_s[_n-1] if rural[_n-1] == 1 & id == id[_n-1] & conus[_
> n-1] == 1 & training == 0 & filedt_s > filedt_s[_n-1]
(3,553,264 missing values generated)

.         bys id: egen time_rural = total(temp_rural)

.         drop temp_rural

. 
. *** time not in region 
.         gen inhomeregion = region == regionhor

.         sort id training inhomeregion filedt_s

.         gen temp_region = filedt_s - filedt_s[_n-1] if inhomeregion[_n-1] == 0 & id == id[_n-1] &
>  training == 0 & filedt_s > filedt_s[_n-1]
(2,339,636 missing values generated)

.         replace temp_region = 0 if inhomeregion[_n-1] == 1 & id == id[_n-1] & training == 0
(608,653 real changes made)

.         bys id: egen time_notreg = total(temp_region)

.         bys id: egen ever_home = min(temp_region)
(7,501 missing values generated)

.         replace ever_home = 1 if ever_home > 0
(2,103,660 real changes made)

.         replace time_notreg = . if regionhor == ""
(207,148 real changes made, 207,148 to missing)

.         drop temp_region

.         
.         drop ever_home

.         gen temp_ever_home = inhomeregion == 1 & training == 0

.         bys id: egen ever_home = max(temp_ever_home)

.         
. *** ever in home division
.         gen inhomedivision = division == divisionhor & training == 0

.         bys id: egen ever_hom_division = max(inhomedivision)

. 
.         sort id training inhomedivision filedt_s

.         gen temp_div = filedt_s - filedt_s[_n-1] if inhomedivision[_n-1] == 0 & id == id[_n-1] & 
> training == 0 & filedt_s > filedt_s[_n-1]
(2,013,013 missing values generated)

.         replace temp_div = 0 if inhomedivision[_n-1] == 1 & id == id[_n-1] & training == 0
(307,750 real changes made)

.         bys id: egen time_notdiv = total(temp_div)

.         replace time_notdiv = . if divisionhor == ""
(225,170 real changes made, 225,170 to missing)

.         
. *** ever in home state
.         gen inhomestate = state == statehor & training == 0

.         bys id: egen ever_hom_state = max(inhomestate)

. 
. ** number of years with a move
.         bys id year: egen move_in_year = max(move_not)
(7,501 missing values generated)

.         bys id year: gen num_temp = sum(move_in_year)

.         gen temp3 = (move_in_year == 1 & num_temp == 1)

.         bys id: egen tot_year_move = total(temp3)

.         drop move_in_year num_temp temp3

. 
. 
. ** moves in the last five years
.         bys id: egen lastobs2 = max(filedt_s)
(7,501 missing values generated)

.         gen temp5 = lastobs2 - 1826
(7,501 missing values generated)

.         gen last5 = (filedt_s >= temp5)

.         gen tempmove = (move_not == 1 & last5 == 1)

.         bys id: egen moves_5years = total(tempmove)

.         bys id: egen max_yos = max(yos)
(7,501 missing values generated)

.         replace moves_5years = . if max_yos < 4
(0 real changes made)

.         drop max_yos

.         drop  temp5 last5 tempmove

. 
. *** move in the last year
. 
.         gen temp1 = lastobs2 - 365
(7,501 missing values generated)

.         gen last1 = (filedt_s >= temp1)

.         gen tempmove = (move_not == 1 & last1 == 1)

.         bys id: egen moves_1year = total(tempmove)

.         drop lastobs2 temp1 last1 tempmove

. 
.         foreach var in msa region division {
  2.         bys id `var': gen num_`var' = _n
  3.         gen temp`var' = num_`var' == 1
  4.         bys id: egen tot`var' = total(temp`var')
  5.         drop temp`var' num_`var'
  6.         }

. 
. dis "count number of people in sample, check 2"
count number of people in sample, check 2

. count if num_id_check == 1
  378,370

. drop lastdate

. bys id: egen lastdate = max(filedt_s)
(7,501 missing values generated)

. 
.         /* creating a file to check some of the data */
.         egen newid = group(id)
(7,501 missing values generated)

.         sort newid filedt_s

. 
. *** number of kids
.         gen nbr_kids = nrdep - 1 if married == 1 & dualspouse ~= 1
(2,638,150 missing values generated)

.         replace nbr_kids = nrdep if married == 0 | dualspouse == 1      
(2,617,863 real changes made)

.         replace nbr_kids = nrdep if married == 0 & dualspouse == .
(0 real changes made)

.         
.         bys id: egen temp_kids = max(nbr_kids)
(7,729 missing values generated)

.         drop nbr_kids

.         ren temp_kids nbr_kids

. 
. ******
. ** keeping one observation
. ******
.         bys id: egen max = max(filedt_s)
(7,501 missing values generated)

.         keep if filedt_s == max
(3,506,867 observations deleted)

. 
. 
. ** some people are missing their marriage indicator for the last observation.  replacing with dee
> rs data in that case
.         replace married = 1 if married == . & spouse_deers == 1
(207 real changes made)

.         replace married = 0 if married == . & spouse_deers == 0
(235 real changes made)

.         
. 
. 
. *** moves per year variables
.         gen tot_mov_nt_yr = totmoves_not / `yos'

.         gen tot_moves_yr = totmoves / `yos'

.         gen tot_msa_yr = totmsa / `yos'

.         gen tot_region_yr = totregion/ `yos'

.         gen tot_division_yr = totdivision / `yos'

.         gen tot_lm_yr = totlongmove / `yos'
(7,501 missing values generated)

.         gen tot_sm_yr = totshortmove / `yos'
(7,501 missing values generated)

.         set emptycells drop, permanently
(set emptycells preference recorded)

. 
. 
. ** regressing their time in service before first move 
.         eststo clear

. 
.         ren tot_division_yr tot_div_yr

. 
.         dis "count number of people in sample, check 3"
count number of people in sample, check 3

.         count if num_id_check == 1
  0

. 
.         gen w_m = female == 0 & white == 1

.         foreach var in longestmove time_overseas time_1_2 mean_last2 time_notreg time_rural {
  2.         gen temp2 = `var' /365.25
  3.         drop `var'
  4.         ren temp2 `var'
  5.         }
(15,611 missing values generated)
(89,196 missing values generated)
(15,611 missing values generated)
(26,546 missing values generated)

. 
.         foreach var in  totpop {
  2.         gen temp2 = `var' / 100000
  3.         drop `var'
  4.         ren temp2 `var'
  5.         }
(82,015 missing values generated)

.         foreach var in 5 {
  2.         if `to_`var'' {
  3.         gen re_enlist_10 = maxyos >= 10
  4.         }
  5.         }

. 
. 
. 
.         gen male = sex == "M" if sex ~= ""
(7,502 missing values generated)

. 
. 
. 
. *** creating indicators for 1, 2, 3 moves, etc.  will use in the regressions
.         gen one_move = totmoves_not == 1

.         gen zero_one_move = totmoves_not_w_abroad <= 1

.         gen two_move = totmoves_not == 2

.         gen zero_two_move = totmoves_not_w_abroad <= 2

.         gen three_move = totmoves_not == 3

.         gen zero_three_move = totmoves_not_w_abroad <= 3

.         gen any_move = totmoves_not > 0 if totmoves_not ~= .

.         gen any_move_intl = totmoves_not_w_abroad > 0 if totmoves_not_w_abroad ~= .

. 
.         gen one_abroad_move = abroad_moves_not == 1

.         gen two_abroad_move = abroad_moves_not == 2

.         gen three_abroad_move = abroad_moves_not == 3

.         
.         gen one_abroad_moves_not_c_o = abroad_moves_not_c_o == 1

.         gen two_abroad_moves_not_c_o = abroad_moves_not_c_o == 2

.         gen three_abroad_moves_not_c_o = abroad_moves_not_c_o == 3

. 
.         gen one_abroad_moves_not_o_c = abroad_moves_not_o_c == 1

.         gen two_abroad_moves_not_o_c = abroad_moves_not_o_c == 2

.         gen three_abroad_moves_not_o_c = abroad_moves_not_o_c == 3      

.         
.         gen one_abroad_abroad_move = abroad_to_abroad_moves == 1

.         gen two_abroad_abroad_move = abroad_to_abroad_moves == 2

.         gen three_abroad_abroad_move = abroad_to_abroad_moves == 3

.         
. foreach var in tot_less_exp_rent tot_better_amm tot_better_epop tot_better_not tot_worse_not tot_
> worse_epop tot_worse_amm tot_more_exp_rent {
  2. replace `var' = . if one_move ~= 1
  3. }
(35,266 real changes made, 35,266 to missing)
(31,829 real changes made, 31,829 to missing)
(34,803 real changes made, 34,803 to missing)
(34,934 real changes made, 34,934 to missing)
(34,934 real changes made, 34,934 to missing)
(34,803 real changes made, 34,803 to missing)
(31,829 real changes made, 31,829 to missing)
(35,266 real changes made, 35,266 to missing)

.         
.         gen drop0_totmov = totmoves_not

.         replace drop0_totmov = . if totmoves_not == 0
(187,612 real changes made, 187,612 to missing)

. 
. /* creating a normalized version of epop for each of these outcomes below*/
. 
. *** creating some sub populations
.         gen all = 1

.         gen nm_ent = 1 if m_ent == 0
(73,540 missing values generated)

.         replace nm_ent = 0 if m_ent == 1
(73,540 real changes made)

. 
.         gen term5 = terms == "5"

.         
.         gen max_three = totmoves_not_w_abroad <= 3

.         
.         gen neverdepl = everdepl == 0

.         gen infantry = branch == "IN"

. 
. 
. *** interating home with moves
.         replace ever_home = . if time_notreg == .
(26,546 real changes made, 26,546 to missing)

.         summ ever_home

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
   ever_home |    359,325    .5059208    .4999656          0          1

. 
.         gen home_x_tot = ever_home*totmoves_not
(26,546 missing values generated)

. 
. *** creating distances in terms of 1,000s       
.         gen tot_dist_not_old = tot_dist_not
(192,953 missing values generated)

.         replace tot_dist_not = tot_dist_not_old / 1000
(95,939 real changes made)

. 
. *** normalizing avg_epop, time, and distance
. 
.                 qui areg re_enlist totmoves_not ged hsd hsg asc_smc college_pl female afqsc black
>  hispanic other_race sum_combat age age_sq if everabroad == 0, absorb(exper_br_year_fe)          
>       

. 
. 
.                 summ longestmove, de

                         longestmove
-------------------------------------------------------------
      Percentiles      Smallest
 1%     1.749487       .0848734
 5%     2.250513        .495551
10%     2.505133       .4982888       Obs             370,260
25%     3.080082       .4982888       Sum of Wgt.     370,260

50%     3.832991                      Mean            4.20584
                        Largest       Std. Dev.      1.687037
75%     4.999316       20.75017
90%     6.335387       20.75017       Variance       2.846093
95%       7.4141       20.91992       Skewness       1.601406
99%     9.998631       21.08419       Kurtosis       7.598997

.                 ren longestmove longestmove2

.                 gen longestmove = (longestmove - r(mean))/r(sd)
(15,611 missing values generated)

.                 gen long_avg = longestmove > 0

. 
.                 gen long_x_tot = longestmove*totmoves_not
(15,611 missing values generated)

. 
.                 summ avg_epop if e(sample) == 1, de

                          avg_epop
-------------------------------------------------------------
      Percentiles      Smallest
 1%     .3322973       .1854919
 5%     .3647434       .2217213
10%     .3810251       .2520448       Obs             161,940
25%     .4366765       .2839452       Sum of Wgt.     161,940

50%     .4835924                      Mean           .4769212
                        Largest       Std. Dev.      .0633358
75%     .5246888       .6238691
90%     .5545018       .6245306       Variance       .0040114
95%     .5679604       .6246693       Skewness      -.4344565
99%     .6000841        .629722       Kurtosis       2.590464

.                 ren avg_epop avg_epop2

.                 gen avg_epop = (avg_epop2 - r(mean))/r(sd)
(221,381 missing values generated)

.                 gen long_emp = avg_epop > 0

.                 replace long_emp = . if avg_epop == .
(221,381 real changes made, 221,381 to missing)

.                 gen emp_x_tot = avg_epop*totmoves_not
(221,381 missing values generated)

.                 
.                 summ avg_cellshare if e(sample) == 1, de

                        avg_cellshare
-------------------------------------------------------------
      Percentiles      Smallest
 1%     .0003651      -.0002691
 5%     .0007406      -.0000614
10%     .0014551      -.0000524       Obs             181,529
25%     .0092245       .0000208       Sum of Wgt.     181,529

50%     .0253726                      Mean           .0238804
                        Largest       Std. Dev.      .0156539
75%     .0366357       .0725837
90%     .0431491       .0753528       Variance        .000245
95%      .047176       .0873995       Skewness       .0121378
99%     .0556883       .0898028       Kurtosis       1.932064

.                 ren avg_cellshare avg_cellshare2

.                 gen avg_cellshare = (avg_cellshare - r(mean))/r(sd)
(203,045 missing values generated)

.                 
.                 gen cell_x_tot = avg_cellshare*totmoves_not
(203,045 missing values generated)

. 
.                 summ avg_cellpop if e(sample) == 1, de

                         avg_cellpop
-------------------------------------------------------------
      Percentiles      Smallest
 1%     88.14836      -63.05801
 5%     287.6448      -21.62916
10%     520.5052      -14.92359       Obs             117,171
25%     2942.208       6.348557       Sum of Wgt.     117,171

50%     5321.728                      Mean           8853.208
                        Largest       Std. Dev.       12697.1
75%     10030.33         214063
90%     17928.37       239901.7       Variance       1.61e+08
95%     29278.61       245833.6       Skewness         5.4391
99%     61272.13       255863.5       Kurtosis       51.45476

.                 ren avg_cellpop avg_cellpop2

.                 gen avg_cellpop = (avg_cellpop - r(mean))/r(sd)
(268,330 missing values generated)

.                 
.                 foreach var in avg_cell2br avg_cell_amm avg_shcon avg_shman avg_shgov avg_shmil a
> vg_shsvc {
  2.                 egen z2`var' = std(`var')
  3.                 }
(201,323 missing values generated)
(233,391 missing values generated)
(221,576 missing values generated)
(221,624 missing values generated)
(221,553 missing values generated)
(221,553 missing values generated)
(221,553 missing values generated)

.                 
.                 
.                 gen cellpop_x_tot = avg_cellpop*totmoves_not
(268,330 missing values generated)

.                 
.                 sum tot_dist_not if e(sample) == 1, de

                        tot_dist_not
-------------------------------------------------------------
      Percentiles      Smallest
 1%            0              0
 5%            0              0
10%            0              0       Obs             182,694
25%            0              0       Sum of Wgt.     182,694

50%     .1612215                      Mean           .6905291
                        Largest       Std. Dev.      1.013469
75%     1.030123       9.604156
90%     2.019274       10.66001       Variance        1.02712
95%     2.963264       11.33297       Skewness       2.018528
99%     3.839683       11.59207       Kurtosis       8.239119

.                 ren tot_dist_not tot_dist_not2

.                 gen tot_dist_not = (tot_dist_not2 - r(mean))/r(sd)
(192,953 missing values generated)

.                 
.                 summ tot_dist_not, de

                        tot_dist_not
-------------------------------------------------------------
      Percentiles      Smallest
 1%     -.681352       -.681352
 5%     -.681352       -.681352
10%     -.681352       -.681352       Obs             192,918
25%     -.681352       -.681352       Sum of Wgt.     192,918

50%     -.681352                      Mean          -.0266079
                        Largest       Std. Dev.      .9896237
75%     .2657203       9.836983
90%     1.279019       10.06043       Variance        .979355
95%     2.232284       10.50101       Skewness       2.077827
99%     3.086484       10.75666       Kurtosis       8.554193

.                 gen tot_dist_not_temp = (tot_dist_not - r(mean))/r(sd)
(192,953 missing values generated)

.                 
.                 gen tot_dist_gr_less300 = tot_dist_not_old <= 300 & tot_dist_not_old ~= 0 if tot_
> dist_not_old ~= . 
(192,953 missing values generated)

.                 gen tot_dist_gr_300_1000 = tot_dist_not_old > 300 & tot_dist_not_old <= 1000 if t
> ot_dist_not_old ~= .
(192,953 missing values generated)

.                 gen tot_dist_gr_1000_2000 = tot_dist_not_old > 1000 & tot_dist_not_old <= 2000 if
>  tot_dist_not_old ~= .
(192,953 missing values generated)

.                 gen tot_dist_gr_2000_3000 = tot_dist_not_old > 2000 & tot_dist_not_old <= 3000 if
>  tot_dist_not_old ~= .
(192,953 missing values generated)

.                 gen tot_dist_gr_great3000 = tot_dist_not_old > 3000 if tot_dist_not_old ~= .
(192,953 missing values generated)

. 
.                 
.                 gen dist_x_tot = totmoves_not*tot_dist_not
(192,953 missing values generated)

. 
. *** labeling variables 
.                 label var everdissolve "Dissolve Marriage"

.                 label var everkids "Have Children"

.                 label var ged "GED"

.                 label var hsd "High School Dropout"

.                 label var hsg "High School Graduate"

.                 label var term5 "Term Legnth = 5 Years"

.                 label var male "Male"

.                 label var female "Female"

.                 label var nm_ent "Not Married when Enter"

.                 label var m_ent "Married when Enter"

.                 label var mar_lyr "Married in 1st Year"

.                 label var ever_home "Ever in Home Region"

.                 label var home_x_tot "Ever Home x Moves"

.                 label var long_avg "Longer than Average Stay"

.                 label var long_x_tot "Longer x Moves"

.                 label var long_emp "Avg. Employment > Mean"

.                 label var emp_x_tot "Emp x Moves"

.                 label var avg_epop "Average Employment to Pop"

.                 label var longestmove "Longest Spell"

.                 label var time_notreg "Time not in Home Region"

.                 label var tot_dist_not "Total Distance of Moves"

.                 
.                 gen abroad_notGM_GE_IT = (conus == 0 & (substr(arloc, 1, 2) ~= "GM" & substr(arlo
> c, 1, 2) ~= "GE" & substr(arloc, 1, 2) ~= "IT"))

.                 bys id: egen ever_abroad_notGM_GE_IT = max(abroad_notGM_GE_IT)

. 
.                 
. ********
. *** dataset for creating all of the tables 
. ********
. 
. save $datadir/jhr_clean_for_tables, replace     
file XXX/dta/jhr_clean_for_tables.dta 
> saved

. 
. /* creating sample to use for Table 8 */
. 
.                 qui areg evermarr totmoves_not ged hsd hsg asc_smc college_pl female afqsc black 
> hispanic other_race sum_combat age age_sq if nm_ent == 1 & everabroad == 0, absorb(exper_br_year_
> fe)

.                 gen sample_evermarr_us = e(sample)

.                 gen sample_marr_nm_ent_us = e(sample)

.                 dis "summ evermarr by nbr moves, us only"
summ evermarr by nbr moves, us only

.                 tabstat evermarr if e(sample) == 1, by(totmoves_not) c(stats) stats(mean n)

Summary for variables: evermarr
     by categories of: totmoves_not (Domestic Moves)

totmoves_not |      mean         N
-------------+--------------------
           0 |  .4986717     71143
           1 |  .6028826     64734
           2 |  .6481765      7842
           3 |  .6659708       479
           4 |  .6981132        53
           5 |         1         3
-------------+--------------------
       Total |   .554203    144254
----------------------------------

.                 local demos_evermarr female ged hsd asc_smc college_pl afqsc black hispanic other
> _race sum_combat age age_sq

.                 
.                 gen marr_nm_ent = married if nm_ent == 1 
(81,041 missing values generated)

.                 local demos_marr_nm_ent female ged hsd asc_smc college_pl afqsc black hispanic ot
> her_race sum_combat age age_sq

.                 
.                 **** saving this sample to use for the Table 8 regressions with five obs per pers
> on
.                 preserve

.                 keep if sample_evermarr_us == 1 
(241,617 observations deleted)

.                 save $datadir/evermarr_us_sample, replace
file XXX/dta/evermarr_us_sample.dta sa
> ved

.                 restore

.         
. 
end of do-file