What Counts as a Return to Education?

This series began with a question about women’s education that wages alone cannot answer: what changes in a woman’s career, family life, and children’s early circumstances when schooling shifts? The strongest conclusion from the California and German evidence is that education can return value through several channels at once, but the size, timing, and direction of those channels depend on the schooling margin and the institution around it. The reader’s task is therefore not to choose between “education raises wages” and “education changes everything,” but to identify which claim the evidence actually supports.

A return is a life-cycle portfolio

The U.S. study provides the clearest positive case. Women born just before California’s school-entry cutoff are more likely to complete some college. For women near that boundary, the associated career difference is visible in both earnings and employment, and it strengthens into mid-career rather than disappearing after an initial head start. The mid-career effect is about $783 in annual earnings and 0.7 percentage points in employment. Conditional earnings also rise, so the result is not only that a few additional women work; the pay associated with working improves as well.

That is the first ledger: career quantity. The second is career quality. By mid-career, the same local discontinuity is associated with more professional work, more jobs requiring some college but no degree, higher occupational prestige, more socially meaningful work, better-educated coworkers, and more teamwork. The study does not find a broad increase in hours, weeks, work pace, or scheduling burden. Some features move favorably but less precisely, and less independence offsets part of the amenity valuation. The defensible claim is not that education makes every job better. It is that the affected women move toward a bundle of occupations with a materially different mix of status, meaning, peers, and working conditions.

The third ledger is family timing and resources. The U.S. career advantage remains visible around a first birth; it is not erased when women become mothers. Fertility balance tests do not show that the cutoff changes who becomes a mother or the average timing of births in the baseline samples. This means the observed post-birth difference is not readily explained by one group selecting into motherhood at a different rate. The result also does not mean that education eliminates the motherhood penalty. It means that the education-related advantage persists through a transition that can impose a separate cost on both groups.

The fourth ledger is early child health. In the U.S. data, mid-career births to women just before the cutoff weigh about 9 grams more and occur about 0.45 days later, while early-career birth outcomes show no detectable discontinuity. Smoking during pregnancy falls, use of WIC and Medi-Cal falls, and the number of prenatal visits does not change. These outcomes move in the same stage of the life cycle as the maternal career gains. The alignment makes it reasonable to ask whether education-linked careers help create a more favorable prenatal environment, even though it does not identify one job attribute as the cause of an infant’s outcome.

The German evidence adds the fifth dimension: context can determine how the same broad schooling logic is expressed in earnings. In historical West Germany, being born after a school-entry cutoff generally delayed school entry and labor-market entry while increasing school starting age and, on average, educational attainment. Men experienced an early earnings loss and little detectable lifetime benefit. Women born after the cutoff had higher earnings through much of their careers, especially from the mid-twenties to around age 30, and the gender interaction favored women over men in lifetime and prime-age earnings. The favorable female result appears despite similar school-starting-age compliance and similar education responses.

The two studies therefore support a life-cycle portfolio rather than a single education premium. Schooling can alter the probability of working, the pay received while working, the occupations accessible to a woman, the attributes bundled with those occupations, the timing of family transitions, and early outcomes for children.