The idea that income inequality might be tempered by “home production”—with low-income families cooking meals at home rather than eating out; mowing the lawn instead of hiring a service—is both intuitively and emotionally appealing. If families can’t earn big bucks, goes the notion, at least they can provide for themselves and thereby soften the blow of poverty due to low wages and unemployment. The story aligns well with America’s ethic of self-sufficient individualism.
But is that comforting picture a reality, or a myth?
Recent research by Minneapolis Fed economists suggests the latter. In “Inferring Inequality with Home Production” (WP 746, also NBER 24166), Job Boerma and Loukas Karabarbounis find that home production—a factor often ignored by economists—“amplifies … differences among households, meaning that inequality is larger than we thought.” Their mathematical model with both market production (that is, jobs) and household production generates greater inequality than a model that incorporates only market production.
Home production—a factor often ignored by economists—“amplifies … differences among households, meaning that inequality is larger than we thought.”
“Our result is surprising,” the economists write. “One could expect that home production tends to compress welfare differences that originate in the market.” But the assumption underlying this expectation, that households without large paychecks ramp up home production to compensate, is unfounded.
In fact, differences in home productivity among U.S. households are three times greater than wage dispersion, they find, and time that households put into home production doesn’t vary sufficiently to make up the difference. The idea that people without high-paying jobs have lots of time on their hands is a myth as well. “Thus, there is little scope for home production to offset differences that originate in the market sector. Rather, home production amplifies these differences.”
The research also bears on tax policy, say the authors. An optimal tax system—one that balances proper incentives to induce work and sufficient redistribution to address needs of the less well off—“should feature more progressivity” once home production is taken into account.
The economists start with a model—two, in fact. The first is a standard benchmark with households that vary with respect to productivity, preferences, time use and expenditures. The second is identical in all ways but one: It includes not only hours spent in market production (time on the job), but also time spent in home production (buying groceries, cooking, cleaning), which studies show to be half as much as time at work.
They then refine the models with real-world data between 1996 and 2015 on consumption spending, time at jobs, wages and time on home chores, pulling numbers from the Consumer Expenditure Survey and the American Time Use Survey plus additional information from the Current Population Survey.
Since their goal is to gauge inequality once time on home chores is accounted for, and to see how that estimate differs from one derived by conventional analysis that ignores household reality, they need to do a careful apples-to-apples comparison. For this, they use a special technique: the “observational equivalence theorem.”
An optimal tax system—one that balances proper incentives to induce work and sufficient redistribution to address needs of the less well off—“should feature more progressivity” once home production is taken into account.
In this context, the theorem says that if two models generate identical results on time at work, wages and consumption expenditures, then any other differences they generate (inequality, for example) can only be the driven by structural factors such as home production. An analogy: Two car models, identical except for windshield tilt, generate different air resistance results when tested in a wind tunnel. The difference in air resistance is thus due to tilt.
Results and implications
The models do indeed generate identical results on work hours, wages and consumption spending, matching U.S. data in all respects. And inequality? The economists look at four measures of inequality; for each, the model with household production generates greater inequality than the model without.
What drives this result? The economists argue that it’s due to the reality, seen in data, that home production hours are not correlated with market consumption or productivity, so home production does not offset income dispersion from working outside the home. “Instead, home production exacerbates inequality [due to] the large dispersion in home productivity,” they write. In fact, there is a slight positive correlation between market expenditures and value of home production.
The unexpected finding that income inequality is amplified, not mitigated, by home production results in a clear policy implication. A tax policy seeking to achieve the optimal balance of equity and work incentives should be more progressive, conclude Boerma and Karabarbounis. By their calculation, a household earning $200,000 annually would pay an average tax rate of 17 percent under a no-home-production model, but 29 percent when home production is included—reflecting that U.S. household inequality is considerably greater when flesh-and-blood households are part of the equation.