IND1950 error for auto employment

I think there may be an error in how IND is being coded to IND1950 for 1970.

There are two IND codes in 1970 that are included in IND1950 auto industry code (IND1950 = 376):

IND == 219
and
IND == 267

The second one (IND == 267) is in Professional and photographic equipment, and watches. I don’t think this makes any sense. The first one (IND== 219) is clearly defined as the auto industry.

The including of IND == 267 seems like an error. And the impact is to imply that auto employment doubled from 1960 to 1970.

Thanks for your inquiry. We are looking into this, but we likely won’t have a response until sometime next week.

In the 1970 Industry and Occupation classification system, the Census has a group of industry allocation categories which report codes from Census forms that were returned without an industry entry and were allocated by Census among the major industry groups during computer processing. Thanks to your inquiry, we have identified that our 1970 IND codes page has these allocated codes displayed in a manner that is not clear. This code (267==Manufacturing, durable goods–allocated) should actually refer to the supergroup of “Manufacturing - Durable Goods” and not the subgroup “Professional and photographic equipment, and watches.” I have made a note with the IPUMS USA team to separate the allocated codes into their own subgroups so that this is more clear.

The decision to include code 267 (from IND in 1970 samples) with 376 (Motor vehicles and motor vehicle equipment) in IND1950 is due to the crosswalk translation system used to create IND1950. Motor vehicles were the largest component of durable goods manufacturing in 1970, so that was the most likely factor to code these as such. 1970 is the only sample that has these “allocated” industry categories, and I have asked the IPUMS USA team to take a look at how we integrate them into IND1950. We classify these 1970 allocated industry codes as N/A in IND1990.

I have a few suggestions should you wish to take an alternative approach.

  • First, you could consider using the CPS. The ASEC samples are available for the entirety of the 1970s and the BMS for the tail end of the decade. The variables IND, IND1950 and IND1990 are all present in the CPS. IND1950 in the CPS does not use the IND code 267 for IND1950==376. IND1950==376 only reports the IND code 219 for the 1970s ASECs.
  • Second, if you wish to use IPUMS USA data, you could allocate a proportion of those with IND = 267 to 376 based on either:
    • The fraction of the workforce in durable goods manufacturing (codes 107-259) that works in auto manufacturing
    • Use some other allocative process based on the distribution of auto manufacturing (e.g., geographic based allocation).