Aggregate labor market data for 46 commuting zones in Norway 1999-2013
Description
These data sets are used in the paper “Local Labor Demand and Participation in Social Insurance Programs” by Andersen, Markussen and Røed (2019) to be published in Labour Economics. The data sets are based on high quality individual level labor market data from Norwegian administrative registers. In order to make the data available to the public, the individual level data has been aggregated by year and commuting zone, resulting in a total of 690 observations (covering 46 regions over the period 1999-2013). Andersen, Markussen and Røed (2019) use a shift-share instrumental variables strategy to explore the effect of labor market conditions on entrance to different states, incl. unemployment and temporary disability. The dataset may be useful for anyone interested in investigating the dynamics of the Norwegian labour market in recent years. The material may also be useful for educational purposes, e.g. to illustrate the potential uses of shift-share instruments. The data material consists of two types of datasets, namely stock- and entry-data (see the below section on Experimental Design, Materials, and Methods for a description of the data types). The data is aggregated by base year (1999-2013) and region of residence in the base year (46 commuting zones) resulting in a total of 690 observations. The number of individuals in each cell is given by the variable n, and all other variables represent cell-averages. The only exceptions are the instrumental variables (z_nar and z_nar_i) and the regional employment rates (emp_0 and emp). These variables are also specific to the base year-region cell, but are based on the overall population in the age range 18-58 (the instrument is described more in detail in the paper and the Data-in-brief article).
Files
Steps to reproduce
Because the empirical analysis is based on Norwegian (encrypted) administrative data obtained under a special arrangement with Statistics Norway, we are not in the position to make the micro data directly available for other researchers. The data are, however, accessible through Statistics Norway, provided that the user abides by the confidentiality regulations set by Statistics Norway and the Norwegian Data Inspectorate (Datatilsynet), as well as national legislation. We will of course make available all code used to create our analyses samples from the original data along with programs used to generate results from those data extracts. We will also cooperate in efforts to replicate our results, including assisting in seeking permission from Statistics Norway to access the original data.