Decompose OINCWAGE into second job and switching jobs

CPS
1

I know INCWAGE=INCLONGJ+OINCWAGE, and OINCWAGE is either outside work or due to job changes (per definition below). I would like to isolate the income earned from a second job. I have a few theories on how I could go about this.

  1. Those with OINCWAGE that switch jobs would face frictional unemployment, and so WKSWORK1 would be less than 52. However, if there are liquidity constraints an individual would wait to quit their job until they have a new one lined up. Furthermore, some individuals could work two jobs seasonally, which would be excluded if we eliminate those who worked less than 52 weeks.

  2. There is UHRSWORKT (hours work per week at all jobs) and UHRSWORK1 (hours work per week at main job). This could isolate someone who had a second job. But, there are some observations that showed individuals working 2 hours per week at a different job with zero in the OINCWAGE category. In theory, this would be volunteer work, but INCWAGE and INCTOT do not match. Therefore, it would be assumed that this is a second job that was coded incorrectly.

  3. Outside work might be captured by INCTOT-INCWAGE. However, some people that are unemployed have external income, which is accounted for by INCTOT. This could be those with external incomes, such as rental properties or survivor’s income.

Overall, should I estimate additional jobs income by INCTOT-INCWAGE? Or, INCTOT-INCWAGE if UHRSWORKT- UHRSWORK1>0? Or, continue to use OINCWAGE because very few of those coded this way switched jobs? Or, OINCWAGE+ (INCTOT-INCWAGE)? Or, is there a different variable that I should use?

By definition:

INCTOT indicates each respondent’s total pre-tax personal income or losses from all sources for the previous calendar year. Amounts are expressed as they were reported to the interviewer.

INCWAGE indicates each respondent’s total pre-tax wage and salary income–that is, money received as an employee–for the previous calendar year. Amounts are expressed as they were reported to the interviewer.

  • For ASEC samples 1988-onward, INCWAGE is derived from a Census recode variable. The topcoded components of INCWAGE are OINCWAGE and INCLONGJ. OINCWAGE is always a component of INCWAGE. When SRCEARN indicates that INCLONGJ is earned from wage and salary, INCLONGJ is an additional component of INCWAGE.

OINCWAGE reports the net amount (prior to deductions) that the respondent earned in wages and salary, other than the amount earned in the primary (longest-held) job, during the preceding calendar year. The amount earned (in wages, salary, or earnings from self-employment in a business or farm) from the longest held job during the preceding calendar year is reported in INCLONGJ.

  • Imagine that an individual worked at one job for 8 months, earning $10,000 in wages, and worked at two other jobs for the remaining 4 months, earning $12,000 in wages. The $12,000 in “other” wage and salary earnings (wages and salary from jobs other than the person’s primary, or longest-held, job) would be reported in OINCWAGE; the $10,000 in wages from the individual’s longest-held job would be reported in INCLONGJ; and the total amount of wage and salary income from the previous calendar year ($22,000) would be reported in INCWAGE.

  • Imagine, instead, that as reported in SRCEARN, an individual was self-employed in a business or farm for their primary job, and also worked one or more additional jobs for wages. In such a case, self-employment income from the primary job would be reported in INCLONGJ; all wage and salary earnings would be reported in OINCWAGE, and the value of OINCWAGE and INCWAGE would be the same.