What encoding does the Census use in TIGER data?

Question

What encoding does the Census distribute TIGER data in? The example on page 88 of PostGIS in Action, Second Edition seems to indicate latin1.

shp2pgsql -s 4269 -g geom_4269 -I -W "latin1"
➥ "tl_2012_states" staging.tl_2012_states |
➥ psql -h localhost -p 5432 -d postgis_in_action -U postgres

Answer 1

According to this FAQ entry,

• TIGER/Line Files and Shapefiles through 2014
• Cartographic Boundary Files – Shapefiles through 2014
• Partnership Shapefiles – through 2014

Were all delivered in ISO-8859-1 (latin1), and

• TIGER/Line Files and Shapefiles 2015*
• Cartographic Boundary Files – Shapefiles 2015*
• Partnership Shapefiles – 2015*

While 2015 and later is UTF-8.

Note:
*: This data and all future dates