Marco Bertoni, Giorgio Brunello, and Lorenzo Cappellari, "Who Benefits
from Privileged Peers? Evidence from Siblings in Schools", Journal of
Applied Econometrics, Vol. 37, No. 7, 2020, pp. 893-916.

The data used in this analysis are drawn from Danish administrative
registers and are confidential. However, our access is not unique, and
others can gain similar access by following a procedure described by
Statistics Denmark at:

http://www.dst.dk/en/TilSalg/Forskningsservice.aspx.

Researchers need to submit a written application to Statistics
Denmark. The application should include a detailed research proposal
describing the goals and methods of the project, a detailed list of
variables, and the selection criteria to be used.

Once received, applications must be approved by the Danish Data
Protection Agency in order to ensure that data are processed in a
manner that protects the confidentiality of registered individuals.
Conditional on this approval, Statistics Denmark will then determine
which data one may obtain in accordance with the research plan. All
processing of individual data takes place on servers located at
Statistics Denmark via secure remote terminal access.

The analysis was executed using stata MP version 14.

Our attached do-files replicate all the analysis reported in the
tables and figures of the paper (file analysis_1.do) and the appendix
(file appendix_analysis_2.do).

The ado file "reghdfe" is required to replicate the analysis.

The dataset required to replicate the analysis
(data_replication_JAE.dta) was generated by sourcing information from
various registers at Statistics Denmark These were made available to
us through the collaboration with researchers at VIVE-Copenhagen Most
of the variables are computed directly from the registers.

We use population registers to obtain demographic information and
trace linkages with mothers and fathers of each individual in the set
of cohorts that we consider (1958-1975). We reconstruct families as
sets of individuals born from the same mother and father, and
consequently reconstruct birth order and spacing.

We use school registers to measure completed education, college major,
and find the school in which each individual was enrolled on October
31st of the year when he/she turned 15, and reconstruct peer groups as
sets of individuals belonging to the same birth cohort and attending
the same school on October 31st of the year when they turned 15. Type
of school is also reported.

We use information on individual addresses obtained from the central
person register to find the parish in which each individual was
resident at age 15, and reconstruct neighbours as sets of individuals
born in the same year and residing in the same parish at age 15. The
same register is used to construct region of residence at age 15 and
31 and track mobility between age 15 and 31.

We use individual tax declarations from 1989 to 2015 to compute
pre-tax annual labor earnings -- or total income from labor -- in
Danish kronas at 2012 prices, including self-employment earnings.

We use firm registers to find firms where individuals and their
parents were employed in the year when individuals were aged 31. We
use this data to constuct a dyadic dataset to assess whether each
individual was working in the same firm of a peer or the same firm of
a parent of a peer at age 31. We then collapse it back to 1
observation per individual indicating whether "any" peer or parent
peer was a co-worker.

We also use information on immigration background and occupation at
the age of 31 from dedicated registers.

We construct peer and neighbourhood composition as described in the
data section of the paper.

We especially thank Paul Bingley for his continuous help with the data.

Please address any enquiry about the replication material to Marco
Bertoni [AT] marco.bertoni@unipd.it