Hourly wage for those who are not paid hourly


I’m using the CPS.

I would like to calculate hourly wage for persons who are not paid hourly. There is hourwage, but it is populated for people who are paid hourly only. I’d like to estimate it for all the other working people.

First, I should acknowledge that I am limited to the Outgoing Rotation Groups (ORG), as most of the relevant variables are limited to that set. Also, I am aware of the need to use earnwt.

I figure that I should calculate:

Note that uhrsworkt is for all civilian workers of sufficient age, whereas earnweek is limited to the Outgoing Rotation Groups. I hope it is not a problem to blend variables from these different sets.

Another measure of work hours is uhrsworkorg (in ORG), but it is only for persons who are paid hourly, so it is not useful in this situation.

Another option for the denominator is ahrsworkt. I’m not sure if that is appropriate or better than uhrsworkt.

When I make this calculation, I get some oddly large values, as high as 961. This occurs for low values in the denominator, such as 1, 2, or 3. The high value of 961 is for a person who allegedly earns $2884 for 3 hours of work.

I also note that, for those who are paid hourly, my estimate can be compared with hourwage. they often agree, but sometimes not. And there are some high outliers as well in this set. Sometimes I’d get better agreement or a more reasonable value if I would use uhrsworkorg or ahrsworkt in the denominator.

So can anyone advise me as to whether I’m on the right track. Is my calculation appropriate? Should I suppress outliers? And by what criterion? Or should I adjust the hours?

Note that my goal is that, once I have an estimation, I would combine it with hourwage, to get a value for all workers.

Thanks to anyone who can help with this question.
–David K

I’d suggest searching in the literature for what other organizations or authors have done, in terms of both definitions and especially outlier cutoffs. One example: Economic Policy Institute gives details on their website for how they deal with outliers in CPS wage data (link).

Your approach makes sense to me. I would suggest using UHRSWORKT instead of AHRSWORKT because the “usual” numbers are more comparable to each other.

The mismatches you describe might have something to do with tips or commissions, or with people who have multiple jobs with different wages.

Hello Matthew,

Thank you for that reply; this may be very helpful. And sorry for the delay.in my response.

Thanks again,