3

I am trying to find unemployment rate data for the US that is:

  • at 5 digit zip code level
  • monthly, or quarterly, or even annually, for 2008-2012
  • free

That seems to be impossible as far as I can tell. It looks like the lowest level of aggregation is county and annual (the county data at http://www.bls.gov/lau/#tables). However, you can get monthly data from http://www.melissadata.com/lookups/bls.asp, which cites the BLS as source, which leads me to believe that it is available. Could someone point me where it can be downloaded?

Second, I will also need to aggregate my outcome data from zip to county FIPS code. Since zips change all the time and counties stay stable, is there a dataset that will allow me to translate from zip at some point in time to county?

EDIT: I was able to locate several option for crosswalk files between zip codes and counties. The most up-to-date versions were the USPS ones. The city state product links each 5 digit zip to the main county (the one that has the majority of the deliveries). The zip+4 product has a one-to-one mapping between them which is more precise, but also more expensive. These are subscription products that need to be updated, and are either monthly, bimonthly or quarterly.

The best free version is the HUD one. It has quarterly 5 digit zip to the main county using an allocation method based on residential addresses and postal vacancy data rather than by area or by population. It uses 2000 census geography. Another one is from the Missouri Census Data Center. Judging by the date it is from September of 2010.

Improve this question

2 Answers 2

Reset to default
3

I don't think BLS unemployment data is available by ZIP. The MelissaData source does not actually report unemployment by ZIP code, it tells you the unemployment of the county in which the entered ZIP code is located.

For example, a search for 02451, a small ZIP code in Waltham, MA (population 60k) returns 826k people in the labor force. This number must be for Middlesex County, in which Waltham is located, which has a total population of 1.51m.

An alternative source of high spatial resolution employment data is the American Community Survey. It appears that employment data are available by Census Tract (which is a much better geographic basis than ZIP code tabulation areas), but only for the 5-year estimates. If you need 1-year estimates, you'll have to go by county, MSA/CSA, or place.

Improve this answer
2
  • I agree with you completely. I think I was not too clear above. I realize that the Melissa data is at the county level. Its advantage is the monthly breakout. The BLS county tables I linked above are only annual. I just can't seem to find the monthly county data any where, and I am also not sure how to add up my zip level data to countries so that I can match the two.
    – dimitriy
    Commented Jul 17, 2012 at 4:39
  • It appears that the full database of county-level unemployment data are in text-based databases in this FTP folder. Instructions on how to read the data are in this text file.
    – dmahr
    Commented Jul 17, 2012 at 11:47
2

Regarding the correlation between zips and counties.

The zip code boundaries are generally stable, assuming the post office doesn't close a ton of offices like some people think they need to do. The big problem is that many zip code areas cross political boundaries. Some small cities can be entirely within a larger zip code. While it is possible to match the center of the code, or the post office location to a political jurisdiction there is no way to determine how to divide the data among the multiple jurisdictions, unless you have data for areas like a census block.

Improve this answer
3
  • 1
    As an addendum, it should be noted that there are many ZIP codes which do not have a ZIP code tabulation area (ZCTA)--they are individual points for dedicated PO boxes, large businesses, and warehouses that receive lots of mail. Moreover, ZCTAs are not created by the Postal Service--they are generalized representations created by the Census Bureau.
    – dmahr
    Commented Jul 17, 2012 at 14:14
  • There seem to be fair bit of changes, especially once you go down to zip+4. There also appear to be several ways to approximate the zip to county match using area, population, or residences/businesses. See my edit above.
    – dimitriy
    Commented Jul 17, 2012 at 18:43
  • As mhoran_psprep points out, ZIP codes don't relate to county boundaries at all so trying to use county data to give ZIP code rates is inherently inaccurate.
    – Sean
    Commented Jul 18, 2012 at 0:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.