Yingying Dong, "Regression Discontinuity Applications with Rounding
Errors in the Running Variable", Journal of Applied Econometrics,
Vol. 30, No. 3, 2015, pp. 422-446.

This paper uses data from the Urban Household Survey (UHS), which are
collected and compiled by the Department of Urban Social and Economic
Survey at the National Bureau of Statistics of China. The data are the
property of the National Bureau of Statistics of China, who do not
permit open access, so the data cannot be provided here. Enquiries
concerning access to the data can be made to the National Bureau of
Statistics of China (http://www.stats.gov.cn/) through phone
(010-68520066), fax (010-68576320), or email (info@gj.stats.cn).

The UHS is an on-going annual survey of urban households in China,
containing detailed information on demographics, education,
employment, and earnings of each household member, along with
household level income, expenditures, and consumption. The data go
back to 1988, though standardized questionnaires were adopted only
since 1992.

This paper uses the UHS data from 1997-2006 in five provinces and one
city. The five provinces are Liaoning, Zhejiang, Guangdong Shanxi, and
Sichuan, and the one city is Beijing. The sample consists of all male
household heads who were either labor force participants or retirees
during 1997-2006 and who were between 45 and 75 years old during the
sample period. The variables used are household food consumption
(deflated by its region-specific CPI to be in 1996 constant RMB),
household head's self-reported retirement status, household head's
age, household head's education in three categories (college or above,
high school, or the default category of less than high school), and
survey year. The household food consumption comes from the Urban
Household Consumption Expenditure Questionnaire, while all the
household head information comes from the Household Member Basic
Information Questionnaire.

Yingying Dong
yyd@uci.edu