Merging a IPUMS with a different county-level data



I am using 5-year 2011 ACS data. I am trying to merge other external data sets to the ACS using “state + county fips” as unique identifier. However, there are many observations (45%) in ACS that have “0” for county fip code. What does this zero mean for county fip? Why is this the case? Should I drop them?

Thank you in advance


In years since 1950, counties are not directly identifiable in public use microdata, which is why many observations will have a zero. The reply from @JeffBloem on this forum post does a good job of explaining why and provides some guidance for alternatives.

Whether or not you should drop the observations depends on your analysis, but it would certainly eliminate a large portion of your data.