Crosswalk Issue

The relationship between 2000 and 2010 census tracts is complex and many-to-many, so there is often no single 2000 tract code corresponding to a 2010 tract. A 2010 tract may correspond to multiple 2000 tracts or only to a small part of a 2000 tract, etc.

For this application, I would recommend instead that you obtain 2000 poverty data for 2000 block group parts and allocate those data to 2010 tracts using the corresponding crosswalks. Then you could have 2000 poverty rates for 2010 tracts, corresponding directly to the 2010 tract codes in the LIHTC data.

Your plan to use 2005-2009 data poses a special challenge because the 2009 5-Year ACS Summary File provides data for 2000 block groups but not for 2000 block group parts, and NHGIS doesn’t supply a crosswalk for 2000 block groups. Also, according to ACS documentation, “in 19 counties from 8 different states, many of the census tracts and block groups used to tabulate and present the 2005-2009 ACS 5-year estimates are either those submitted to the Census Bureau for the 2010 Census, or a preliminary version of 2010 Census definitions.” As a result, even if you used NHGIS block crosswalks to interpolate the 2005-2009 block group data to 2010 tracts, there’d be additional complications in these 19 counties.

If it’s possible for you to use 2006-2010 ACS instead of 2005-2009 data, that would simplify matters, enabling you to match data directly to the 2010 FIPS codes in the LIHTC data.