Household Income (Without Individual) Variable

USA
1

1st Question:
I would like to include a household income variable (without accounting for the individual wages) as a control in my OLS regression. My y variable of the regression is ‘INCWAGE’.

I noted that the household income variable ‘HHINCOME’ in the ACS includes individual wages; How can I then proxy for the household income excluding the individual’s wages, given my y variable as ‘INCWAGE’?

2nd Question:
My analysis is on the individual-level, therefore I have applied weight using the svyset [pweight=perwt] and svy commands on STATA. Since I have included household income as my control, is it correct to only use the PERWT in my weighting?

2

My understanding of your first question is that you would like to measure household income minus wage and salary income using ACS data from IPUMS USA. Please let me know if this is not correct. HHINCOME is the sum of all household members’ individual incomes, as reported in INCTOT. INCTOT includes income from many different sources, which you can view in the comparability tab of the variable. You can recalculate a version of HHINCOME that does not include wage and salary income in two steps.

  1. Calculate a new version of INCTOT for each individual respondent that does not include wage and salary income (INCWAGE). To do this, sum the values of the other income variables that make up INCTOT: INCBUS00, INCSS, INCWELFR, INCSUPP, INCINVST, INCRETIR, and INCOTHER. Be sure to replace top codes with missings so as to not include them as dollar amounts.
  2. Sum the values of the new version of INCTOT for all individuals within the same household. Households are uniquely identified by a combination of SERIAL and SAMPLE. This will give you an estimate of household income that does not include the wage and salary income of any household member.

To answer your second question, if your analysis is at the person level, you should use PERWT (and use replicate weights to calculate empirically derived standard errors when using ACS data). While household income describes a household, the variable is applied to all person records and is therefore also an individual level variable. You do not need to use HHWT when including household income in your analysis, unless the analysis is at the household level. An example of a household level analysis would be estimating the number of households with income within a certain range.