Codes and Data for "Technological Development and the Employment of the Old: Evidence from China"

Description

With rapid population aging around the world, delaying retirement is considered as a practical way to deal with this problem by many countries, but rapid technological development may affect the employment of the old. Existing literature estimates the age-heterogeneous effect of technological development on employment but draw mixed conclusions. In this paper, we utilize China Family Panel Studies (CFPS) data spanning from 2010 to 2018 and examine the effect of technological development on the employment of workers of different ages. Using a Bartik-style instrument variable (IV) to deal with the endogeneity, we find that the effect of technological development on employment decreases with age, especially for the low socio-economic status group. Since old workers are more likely to lose their job due to technological development and the effect is heterogeneous, we further explore the effect of technological development on inequality and its effect on the mental welfare. We cannot share any CFPS data publicly due to the confidentiality requirements of CFPS. But CFPS data is publicly available and easy to request, CFPS data from 2010 to 2018 can be requested at https://www.isss.pku.edu.cn/cfps/. We share all other city-level and county-level variables used in our paper. We also share all codes used in the empirical part. We also add a readme file which briefly introduce our data and code shared here.

Steps to reproduce

“cfps_data_clean_code.do” is the code used to clean CFPS data from 2010 to 2018 separately. Since we clean CFPS data by year and most code are quite similar, we only explain the meaning of codes when firstly used instead of explaining them all. “summary_statistics_code.do” is the code for summary statistics results (Table 1 and Table 2 in the paper). “regression_code.do” is the code for all our regression results in the empirical part (Table 3 to Table 13, and Figure 1 in the paper). “2000_ratio.dta” is the ratio of local technology development to the national level in the year 2000, which is used to construct our Bartik IV. Dta files named by “county_covariate” are the county level covariates used in our regression models. Dta files named by “tech_census” are the city-level or county-level technology development, which is measured by the percentage of workers in high-tech industries and calculated using the census data. Dta files named by “number_obs_county” are the number of observation reporting their industry information in 2000, 2010 and 2015 census data respectively. We cannot share CFPS data here due to the confidentiality requirements of CFPS. But CFPS data is publicly available and easy to request, CFPS data from 2010 to 2018 can be requested at https://www.isss.pku.edu.cn/cfps/.

Institutions

Categories

Funding

Licence