Return to Jeffrey Lin's home page
These data allow researchers to create their own historical data "normalized" to 2010 U.S. census tract boundaries. The basic idea is to weight historical census data, where the weights are determined by overlapping land area. A full description of our tract boundary normalization procedure can be found in the article cited below and Online Appendix B.
If you find the information on this page useful, please cite the following article.
Lee, Sanghoon and Jeffrey Lin. "Natural Amenities, Neighborhood Dynamics, and Persistence in the Spatial Distribution of Income,"Review of Economic Studies 85:1 (January 2018), 663-694 [doi:10.1093/restud/rdx018].
Census tract data from 191002010 can be accessed through NHGIS at https://www. nhgis.org. The NHGIS data selection tool allows filtering by geography--in this case census tract--and year.
Census microdata from 1880 can be accessed through IPUMS USA at https://usa.ipums.org/usa/. These individual data can then be aggregated using the enumeration district identifiers.
To normalize 1970–2000 data to 2010 census tract boundaries, we use the Longitudinal Tract Database (LTDB) (Logan, Xu, and Stults, 2014). The LTDB can be accessed at https://s4.ad.brown.edu/Projects/Diversity/Researcher/Bridging.htm.
To normalize 1880 and 1910-1960 data to 2010 census tract boundaries, we use weights determined by overlapping land area. Boundary files for 1910-1960 and 2010 census tracts can be accessed through NHGIS at https://www.nhgis.org/documentation/gis-data.
Boundary files for 1880 census tracts can be accessed through the Urban Transition Historical GIS (UTHGIS) project at https://s4.ad.brown.edu/Projects/UTP/index.htm
Each tract crosswalk file attached to this page (csv format, one for each historical census year) contains 3 columns.
A 2010 census tract identifier,
A historical census tract or enumeration district identifier,
Land area of overlap between the 2010 census tract and the historical census tract.
These files were created using ArcGIS’s "Intersect" tool, using the boundary files available from UTHGIS and NHGIS.
This read me file describes an example of how to use these crosswalk files.
Note that very small resolution errors due to misaligned boundaries in the map shapefiles from different Census years sometimes indicate overlap between tracts in different states. This issue affects less than 1 percent of tracts. Thanks to Casey Homan for identifying this error. 07 November 2019
Jonathan Schroeder has proposed an alternative interpolation of historical Census tract data called "cascading density weighting" (see pp. 57-69 and 74-82).
Last updated February 28, 2020
Home URL: http://jlin.org/
The views expressed on this site are those of the author and do not necessarily represent the views of the Federal Reserve Bank of Philadelphia or the Federal Reserve System.