By Leslie McCall, Stone Center Senior Scholar
Spotlight on Data
In this blog post*, Leslie McCall considers whether adjusting for family size eliminates economic benefits for families headed by couples compared to those headed by single people.
In a previous post, we showed that the economic benefits of having a spouse or partner relative to the economic benefits of being single were both large and unequally distributed across the earnings distribution.1 Among top-decile earners, for instance, the gap in family incomes between single and partnered individuals was considerably smaller than among earners at the bottom or in the middle of the earnings distribution. In effect, high earners are more insulated from the need to pool income with a partner to buffer against financial risks or to reach the top of the family income ladder.
We also showed that the economic benefits of being partnered were still much greater for women than for men in 2010. Using our new measure of self-reliance, however, we were able to identify which groups of men do, in fact, benefit financially from partnership (middle-to-low earners).
But how do the higher costs of living in a family with more than one adult, and possibly with children as well, affect the so-called benefits of partnership? Do these benefits vanish when family size is taken into consideration? Of course, single people also have children, so this might complicate matters.
Alternatively, do partners contribute enough income to cover their additional consumption needs, particularly if we take economies of scale into account (e.g., the rent of a two-bedroom apartment is not twice the rent of a one-bedroom apartment, all else equal)? Are the middle-to-low-earning men we identified in the previous post still benefiting from their partners’ earnings after adjusting for family size?
[Figure 1]
We find that, for women, the benefits of partnership are in fact smaller after adjusting for family size; nonetheless, the basic story is much the same.
Women who are partnered are still ranked considerably higher in the family income distribution than women who are single, despite falling in the same decile of earnings from employment, after adjusting for family size. At the same time, using our preferred method of family-size adjustment2, the average gap in family income ranks is smaller: In 1970, it is 32 percentage points after adjusting (see Figure 1), whereas it is 45 points before adjusting (see figure from prior blog post). In 2010, it is 21 points after adjusting, whereas it is 32 points before adjusting. An alternative family-size-adjustment method (see footnote 2) produces gaps closer to the levels found in the unadjusted data (a 39-point and 29-point gap in 1970 and 2010, respectively; full results not shown).
In sum, the economic benefits for women of sharing income with a male partner clearly exceed the costs of a larger family, even though these benefits (relative to being single) shrink once those extra costs are taken into consideration.
[Figure 2]
For men, the economic benefits are much less pronounced once we account for the costs of a larger family.
In our first post, we used unadjusted data to show that the widest gaps in family income ranks between partnered and single men are in the lower-middle of the earnings distribution. Such men benefit financially from partnership — that is, from their female partners’ earnings — more than men with high earnings do, and these benefits increased over time. But does this pattern still hold when family size is accounted for? Our preferred adjustment method shows that the benefits continue to exist in 2010 but only to a rather small degree. In 1970, this adjustment eliminates the benefits almost completely.
Indeed, as was once commonly assumed, partnership represents a significant net loss, financially speaking, for most men. We call that net loss an “income penalty.” In 1970, single men took home substantially more income than partnered men who earned the same amount from employment, after adjusting for family size (see Figure 2). Only zero-earning men, who relied entirely on earnings from partners or other sources, were exempted from this pattern.
But times have changed: by 2010, the income penalty for partnered men (after adjusting for family size) diminished sharply, and became a benefit or a wash for those men in the bottom half of the earnings distribution.
In essence, due to the increased contributions of female spouses and partners from 1970 to 2010, middle-to-low earning men in couples were positioned on par with, or even above, their single-male counterparts in the 2010 family income distribution. Among higher earners, single men still out-ranked partnered men in the 2010 family income distribution, though this gap is much smaller than in 1970. (As with women, an alternative adjustment method produces results similar to those found in the unadjusted data.)
So, yes, family size can reverse the economic benefits of partnership, turning them into penalties instead. But, historically, this was true only for men, and, in 2010, it was true only for men in the top half of the earnings distribution. Gender differences in the financial benefits of partnership are still large (20 percentage points on average in 2010), but single men with low earnings now experience the same kind of income penalty that all single women — though especially low-earning single women — face relative to partnered women with the same earnings from employment. In other words, today both low-earning men and women benefit economically from having a partner’s earnings to rely on.
In future posts we will look at (1) trends in partnership; (2) trends in the similarity of earnings levels among partnered men and women; and (3) trends in both of these broken down for men and women at the bottom, middle, and top of the distribution, revealing further patterns of inequality that have been hard to discern using the methods and measures of prior research.
Derek Burk and Deirdre Bloome are coauthors on this research project and blog post. George Lolashvili, a student in the Quantitative Methods in the Social Sciences Master’s program at The Graduate Center, contributed to this blog post.
*Minor edits to the data referenced in this blog post were made on November 23, 2021.
Related Research: Economic Self-Reliance and Gender Inequality Between U.S. Men and Women, 1970–2010
Footnotes:
1 As discussed in the previous post, living with a spouse or partner can significantly raise an individual’s family income beyond what they earn in the workplace, and thus raise their rank in the family income distribution. We call this raise the “economic benefit” of living with a partner. As also noted in the previous post, all partnership-based disparities in family income rank pertain to individuals who are otherwise similar in their earnings from employment. (Also, recall from the previous post that at this time our data examine only female/male couples.)
2 We use a routine adjustment for family size, which is to divide income and/or earnings by the square root of the number of family members, including children. The square root in the denominator allows for economies of scale obtained by purchasing goods collectively rather than individually (e.g., as in the example of rent that we discussed above; likewise, think of durable goods, such as a refrigerator, which are purchased only once per family, irrespective of family size).
Our preferred method adjusts only total family income by family size (i.e., divides total family income by the square root of family size) and not earnings from employment. That is, the y-axis is adjusted but not the x-axis. This approach allows individuals with similar earnings to be categorized in the same earnings deciles, regardless of how large their families are, so we can hold constant reported earnings in our contrast between family income ranks among single and partnered individuals.
However, it is possible that partnership status affects earnings levels, in which case partnership and related dynamics, such as fertility, ought to factor into the earnings decile that someone falls into. For instance, maybe a partnered individual exerts more effort than a single individual in order to support dependent family members. If this is the case, we might want to categorize this person according to their average level of effort (rather than their extra effort possibly due to partnership), which would mean categorizing them into a lower decile by dividing their earnings (and not only their total family income) by the square root of family size. This is the alternative adjustment method that we describe above: dividing both earnings and total family income by family size before establishing the rankings. We do not disagree with the logic of this adjustment, but for the current exercise prefer an adjustment that holds absolute earnings levels constant in a straightforward manner.