/gpss_replication

GPSS Replication Files

Primary LanguageStata

Summary

This code replicates the figures and tables from Goldsmith-Pinkham, Sorkin and Swift (2019). The main file for rerunning the code can be run using master.do. The individual do-files are outlined below. The do-files use finalized datasets, which are constructed from various data sources, outlined below.

  • The canonical Bartik analysis (BAR) is replicated using data from IPUMS and uses cross-walks generously provided by David Dorn on his website.

  • The China shock analysis (ADH) is replicated using a combination of data sources:

    • the replication file from Autor, Dorn and Hanson (2013),
    • data generously provided by Borusyak, Hull and Jaravel (2019),
    • and data generously provided by Adao, Kolesar and Morales (2019).

  • The Card immigration analysis (CARD) is replicated using replication code provided by David Card from Card (2009) and data from ICPSR

Code process

The master.do file executes the following code:

  1. do make_BAR_table.do constructs Table 3 from the paper and uses input_BAR2.dta, the finalized Bartik analysis file. [NOTE: This code is slow due to bootstrapping.]
  2. make_rotemberg_summary_BAR.do constructs Table 1, Figure 1, and Appendix Figure A1. It uses input_BAR2.dta, the finalized Bartik analysis file.
  3. make_char_table_BAR.do constructs Table 2. It uses input_BAR2.dta, the finalized Bartik analysis file.
  4. do make_ADH_table.do constructs Table 6 from the paper and uses ADHdata_AKM.csv, Lshares.dta and shocks.dta. [NOTE: This code is slow due to bootstrapping.]
  5. make_rotemberg_summary_ADH.do constructs Table 4, Figure 3 and Appendix Figure A2. It uses uses ADHdata_AKM.csv, Lshares.dta and shocks.dta.
  6. make_pretrends_ADH.do makes Figure 2 and Appendix Figure A4. It uses workfile_china_preperiod.dta, ADHdata_AKM.csv, Lshares.dta and shocks.dta.
  7. make_char_table_ADH.do constructs Table 5. It uses uses ADHdata_AKM.csv, Lshares.dta and shocks.dta.
  8. make_CARD_table_hs.do and make_CARD_table_college.do make Table 9. They use input_card.dta.
  9. make_rotemberg_summary_CARD_hs.do and make_rotemberg_summary_CARD_college.do make Table 7, Figure 6 and Appendix Figure A3. They use input_card.dta.
  10. make_char_table_CARD.do makes Table 8. It uses input_card.dta.
  11. make_pretrends_CARD.do makes Figures 4 and 5. It uses input_card.dta.

Data construction for canonical Bartik

IPUMS data cannot be posted. However, the following steps below allow researchers to recreate input_BAR2.dta themselves.

The file is created using two do-files:

  1. create_bartik_data.do, which creates Characteristics_CZone.dta and shares_long_ind3_czone.dta, and takes nine inputs:
    1. IPUMS_data.dta
    2. IPUMS_ind1990.dta
    3. IPUMS_geo.dta
    4. IPUMS_bpl.dta
    5. cw_ctygrp1980_czone_corr.dta
    6. cw_puma1990_czone.dta
    7. cw_puma2000_czone.dta
    8. czone_list.dta
  2. make_input_bar.do, which creates input_BAR2.dta and takes two inputs:
    1. Characteristics_CZone.dta
    2. shares_long_ind3_czone.dta

These files are described in further detail below:

IPUMS_data.dta

Our large base dataset downloaded from IPUMS here: https://usa.ipums.org/usa/data.shtml Note that of the 2009-2011 ACS samples were pooled to form the 2010 sample.

Samples:

  1. 1980 5% state;
  2. 1990 5%;
  3. 2000 5%;
  4. 2009 ACS; 2010 ACS; 2011 ACS

Variables:

year; datanum; serial; hhwt; statefip; conspuma; cpuma0010; gq; ownershp; ownershpd; mortgage; mortgag2; rent; rentgrs; hhincome; foodstmp; valueh; nfams; nsubfam; ncouples; nmothers; nfathers; multgen; multgend; pernum; perwt; famsize; nchild; nchlt5; famunit; eldch; relate; related; sex; age; marst; birthyr; race; raced; hispan; hispand; ancestr1; ancestr1d; ancestr2; ancestr2d; citizen; yrsusa2; speakeng; racesing; racesingd; school; educ; educd; gradeatt; gradeattd; schltype; empstat; empstatd; labforce; occ; ind; classwkr ; classwkrd; wkswork2; uhrswork; wrklstwk; absent; looking; availble; wrkrecal; workedyr; inctot; ftotinc: incwage; incbus00; incss; incwelfr; incinvst; incretir; incsupp; incother; incearn; poverty; occscore; sei; hwsei; presgl; prent; erscor90; edscor90; npboss90; migrate5; migrate5d; migrate1; migrate1d; migplac5; migplac1; movedin; vetstat; vetstatd; pwstate2; trantime

IPUMS_ind1990.dta

An additional dataset of 1990 standardized industries to merge onto the main dataset, again downloaded here: https://usa.ipums.org/usa/data.shtml Note that in the ACS samples, 2009-2011 were pooled to form the 2010 sample. Merging with the main dataset occurred by matching year-serial-pernum.

Samples:

  1. 1980 5% state;
  2. 1990 5%;
  3. 2000 5%;
  4. 2009 ACS; 2010 ACS; 2011 ACS

Variables:

year; datanum; serial; hhwt; gq; pernum; perwt; ind1990

IPUMS_geo.dta

An additional dataset of geographies to merge onto the main dataset, again downloaded here: https://usa.ipums.org/usa/data.shtml

Samples:

  1. 1980 5% state;
  2. 1990 5%;
  3. 2000 5%;
  4. 2009 ACS; 2010 ACS; 2011 ACS

Variables:

year; datanum; serial; hhwt; gq; pernum; perwt; county; countyfips; cntygp98; puma

IPUMS_bpl.dta

An additional dataset of birthplace to merge onto the main dataset, again downloaded here: https://usa.ipums.org/usa/data.shtml

Samples:

  1. 1980 5% state;
  2. 1990 5%;
  3. 2000 5%;
  4. 2009 ACS; 2010 ACS; 2011 ACS

Variables:

year; datanum; serial; hhwt; gq; pernum; perwt; bpl

Data construction for Card (2009)

1980

  1. read80.do - reads the state-specific files of the 1980 5% extracts (available from ICPSR), does minimal data cleaning, merges all state-specific files. The output is all80.dta. Takes as input:

    i. Census of Population and Housing, 1980 [United States]: Public Use Microdata Sample (A Sample): 5-Percent Sample (ICPSR 8101). Download it here: https://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/8101/summary.

  2. read_all80.sas - creates all80.sas7bdat. Takes as input all80.dta.

  3. Run the scripts provided by Card.

    i. np2.sas - creates a working data set of wage-earners age 18+, with recodes, etc. This is np80.sas7bdat. These data are used to build wage outcomes. Takes as input all80.sas7bdat. *reads the code in smsarecode80.sas to re-code msa's.

    ii. allnp2.sas - creates a working data set of EVERYONE age 18+, with recodes, etc. This is supp80.sas7bdat. These data are used to build supply variables. Takes as input all80.sas7bdat. *reads the code in smsarecode80.sas to re-code msa's.

    iii. cell1.sas - creates a big summary of data by cell ==> bigcells.sas7bdat. Takes as input np80.sas7bdat.

    iv.t1.sas- creates a big summary of data by cell ==> allcells.sas7bdat. Takes as input supp80.sas7bdat.

    v. supply1.sas - gets supply measures ==> cellsupply.sas7bdat. Takes as input np80.sas7bdat.

    vi. imm1.sas - gets counts of immigrants by sending country in each city ==>ic_city.sas7bdat (IC is Card's classification of sending countries). Takes as input `supp80.sas7bdat.

    vii.indist.sas - gets fraction of workers in manufacturing by city. Takes as input np80.sas7bdat.

  4. Export some datasets to Stata:

    i. cell1_to_stata.sas - creates datasets on wages of immigrants and natives by education class. Exports them to Stata (1980_bigcells_new1.dta, 1980_bigcells_new2.dta, nw80.dta, iw80.dta, nw801.dta, nw802.dta, nw803.dta, nw804.dta, iw801.dta, iw802.dta, iw803.dta, iw804.dta). Takes as input bigcells.sas7bdat.

    ii. t1_to_stata.sas - creates 1980_allcells_new2.dta. Takes as input allcells.sas7bdat

    iii. indist_to_stata.sas - creates 1980_mfg.dta. Takes as input mfg.sas7bdat

1990

  1. read90.do - reads the state-specific files of the 1990 5% extracts (available from ICPSR), does minimal data cleaning, merges all state-specific files. The output is all90.dta. Takes as input:

    i. Census of Population and Housing, 1990 [United States]: Public Use Microdata Sample: 5-Percent Sample (ICPSR 9952). Download it here: https://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/9952.

  2. read_all90.sas - creates all90.sas7bdat. Takes as input all90.dta.

  3. Run the scripts provided by Card.

    i. np2.sas - creates a working data set of wage-earners age 18+, with recodes, etc. This is np90.sas7bdat. These data are used to build wage outcomes. Takes as input all90.sas7bdat. *reads the code in smsarecode90.sas to re-code msa's.

    ii. allnp2.sas - creates a working data set of EVERYONE age 18+, with recodes, etc. This is supp90.sas7bdat. These data are used to build supply variables. Takes as input all90.sas7bdat. *reads the code in smsarecode90.sas to re-code msa's.

    iii. cell1.sas - creates a big summary of data by cell ==> bigcells.sas7bdat. Takes as input np90.sas7bdat.

    iv. t1.sas - creates a big summary of data by cell ==> allcells.sas7bdat. Takes as input supp90.sas7bdat.

    v. supply1.sas - gets supply measures ==> cellsupply.sas7bdat. Takes as input np90.sas7bdat.

    vi. imm1.sas - gets counts of immigrants by sending country in each city ==>ic_city.sas7bdat (IC is Card's classification of sending countries). Takes as input `supp90.sas7bdat.

    vii. indist.sas - gets fraction of workers in manufacturing by city. Takes as input np90.sas7bdat.

  4. Export some datasets to Stata:

    i. cell1_to_stata.sas - creates datasets on wages of immigrants and natives by education class. Exports them to Stata (1990_bigcells_new1.dta, 1990_bigcells_new2.dta, nw90.dta, iw90.dta, nw901.dta, nw902.dta, nw903.dta, nw904.dta, iw901.dta, iw902.dta, iw903.dta, iw904.dta). Takes as input bigcells.sas7bdat.

    ii. t1_to_stata.sas - creates 1990_allcells_new2.dta. Takes as input allcells.sas7bdat

    iii. indist_to_stata.sas - creates 1990_mfg.dta. Takes as input mfg.sas7bdat

2000

  1. read2000.do - reads the state-specific files of the 2000 5% extracts (available from ICPSR), does minimal data cleaning, merges all state-specific files. The output is all2000.dta. Takes as input:

    i. Census of Population and Housing, 2000 [United States]: Public Use Microdata Sample: 5-Percent Sample (ICPSR 13568). Download it here: https://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/13568.

  2. read_all2000.sas - creates all2000.sas7bdat. Takes as input all2000.dta.

  3. Run the scripts provided by Card.

    i. np2.sas - creates a working data set of wage-earners age 18+, with recodes, etc. This is np2000.sas7bdat. These data are used to build wage outcomes. Takes as input all2000.sas7bdat.

    ii. allnp2.sas - creates a working data set of EVERYONE age 18+, with recodes, etc. This is supp2000.sas7bdat. These data are used to build supply variables. Takes as input all2000.sas7bdat.

    iii. cell1.sas - creates a big summary of data by cell ==> bigcells.sas7bdat. Takes as input np2000.sas7bdat.

    iv. t1.sas - creates a big summary of data by cell ==> allcells.sas7bdat. Takes as input supp2000.sas7bdat.

    v. supply1.sas - gets supply measures ==> cellsupply.sas7bdat. Takes as input np2000.sas7bdat.

    vi. imm3.sas - gets counts of immigrants by sending country in each city ==> ic_citynew.sas7bdat (IC is Card's classification of sending countries). Takes as input supp2000.sas7bdat.

    vii. imm2.sas - gets a count of immigrants present in 2000 by IC - this is used to construct the instrumental variable ==> byicnew.sas7bdat. Takes as input supp2000.sas7bdat.

    viii. inflow3.sas - constructs the supply push instrument by "education and experience cell" and city. This is newflows.sas7bdat. Takes as input ic_city.sas7bdat (output of imm1.sas' in 1980) and byicnew.sas7bdat(output ofimm2.sas` in 2000).

  4. Export some datasets to Stata:

    i. cell1_to_stata - creates datasets on wages of immigrants and natives by education class. Exports them to Stata (2000_bigcells_new1.dta, 2000_bigcells_new2.dta, nw.dta, iw.dta, nw.dta, nw.dta, nw.dta, nw.dta, iw.dta, iw.dta, iw.dta, iw.dta). Takes as input bigcells.sas7bdat.

    ii. t1_to_stata - creates 2000_allcells_new1.dta and 2000_allcells_new2.dta. Takes as input allcells.sas7bdat.

    iii. inflow3_to_stata - exports `newflows.sas7bdat' to dta.

Replicate Table 6 of Card (2009) and construct input dataset for Bartik analysis

  1. table6.do - replicates Table 6 of Card (2009) and constructs the dataset input_card.dta. Takes as input the Stata datasets exported from SAS (cited above) for 1980, 1990, and 2000.