Please cite Adachi, Kawaguchi, and Saito (2022) "Robots and Employment: Evidence from Japan, 1978-2017" when using our data.

After preparing data, please open aks_robots.Rproj and run codes/MASTER_replicate.R. This master code automatically calls a series of sub-codes under codes/sub/ to generate figures and tables in output_final folder.

data/by-industry_by-application_1978-2017_labelled_prices.csv is the main data file that summarizes the number and value of robot shipments by industry, application, and year. It contains the following variables:

• code_union: industry code
• code_union_app: application code
• year: calendar year
• ind_union: industry name in Japanese
• ind_union_eng: industry name in English
• name_union: application name in Japanese
• app_code_name_eng: application name in English
• quantity: quantity
• sales.m: sales in current million JPY
• quantity_t: quantity in thousand units
• sales_b: sales in current billion JPY
• unitval_m: unit value in current million JPY (sales.m/quantity)
• q_indagg: quantity aggregated at the industry-year level
• pq_indagg: sales aggregated at the industry-year level
• q_indagg_loo: leave-one-application-out quantity aggregated at the industry-year level
• pq_indagg_loo: leave-one-application-out sales aggregated at the industry-year level
• p_indagg_loo: leave-one-application-out price (unit value) aggregated at the industry-year level

In addition to these, we have included the following datasets under data/ that are called in the replication files.

• by-industry_by-application_1978-2017.csv: Raw Japanese robot adoption data file between 1978 and 2017.
• by-industry_by-application_1978-2017_labelled_prices_5year.csv: 5-year aggregated robot data file.
• JARA_ait.csv: Robot data file ready for main analysis.
• JARA_it_by_a_quality.csv: Robot data with quality adjustment variables.
• price_output_export.csv: Industrial output price, output, export data from the JIP database.
• bojprice/raw/revision/nme_R031.24549.20210330072035.02_utf.csv: The Bank of Japan price series.
• by-industry_by-type_1974-2000_noGHI_noNonManuf.csv: Robot data file containing early periods 1974-1977.
• by-application-by-type_1982-1991_labelled.csv: Robot data file classified by robot types.
• alpha.csv: Robot application expenditure share data across industries.
• oecd/ALFS_SUMTAB_23042021012546383.csv: OECD annual employment data.
• oecd/china/Employment.xls: Chinese employment data available at the National Bureau of Statistics China.
• jip/jip_import.csv: Industrial import statistics obtained from the JIP database.
• jip/jip_capial.csv: Industrial capital stock statistics obtained from the JIP database.
• industry_codes_within_jara_v2_to_essCompatibleJARA2002.xlsx: Industry codes concordance.
• comtrade_847950.csv: The Comtrade data of industrial robots (HS code = 847950).
• exrate/AEXJPUS.csv: The exchange rate trend taken from the Federal Reserve Economic Data (FRED).

Furthermore, the following datasets based on proprietary data are needed to run the codes.

• JARA_ESS_it_by_a.csv: Data that match the above JARA data and employment statistics. Employment statistics are available from the Employment Status Survey (ESS).
• JARA_ESS_ct.csv, ESS_cit_d.csv, ESS_cit.csv: Commuting zone level data of robots and employment. The employment data are based on the ESS.
• ifr/industry_000 - All Industries.xlsx: Robot operational stock data from the International Federation of Robotics (IFR).
• soba/MP_by_industry.csv: Multinational production (MP) data compiled from the Basic Survey on Overseas Business Activities (BSOBA).

ESS and BSOBA are government statistics and available with the permission of Japan Statistics Bureau and Ministry of Economy, Trade and Industry (METI). There is a portal website called miripo that summarizes the use of these statistics.

If you have any problems or questions, please report issues.