When a parent’s job is displaced by robots, what happens to their children’s future? Policy debates about automation usually focus on today’s workers and who gains and who loses. But if technological change harms families’ incomes and stability, it may also shape the next generation’s opportunities.

In our recent study (Heyman and Olsson, 2026), we ask whether parental exposure to automation in the 1990s reduced children’s long-run income mobility, earnings, and educational attainment. Using Swedish data that follow parents and children over three decades, we find that this is the case: on average, children whose parents worked in robot-intensive industries and in occupations with more automatable tasks have lower income mobility and worse labour market and educational outcomes in adulthood. These effects are concentrated among low-income families, while children from better-off households are largely shielded. Existing work on robots and labour markets (e.g., Acemoglu and Restrepo, 2020; Dauth et al., 2021) has focused on direct effects on employment and wages. We show that the consequences reach across generations and can widen inequality in ways that persist long after the initial shock.

For parents, we observe who was working in 1990, in which industry and occupation, and how much they earned. We then track their children into adulthood and measure their earnings, employment, and education up to 30 years later. To see which parents were exposed to automation, we combine two measures: the industry spread of robots (using international data on robot installations) and how easily the tasks in each occupation can be performed by robots (using a measure that scores occupations by their similarity to tasks described in robot-related patents created by Webb (2020)). This approach captures both where robots expanded (industries) and which jobs were most exposed (occupations). Parents can therefore face different levels of exposure, even if they hold similar jobs, depending on how quickly their industry adopts robots.

A key challenge is separating the effect of automation from other differences between families. Parents in robot-heavy industries may differ in ways that also influence their children’s outcomes. To address this, we compare parents in similar robot-exposed occupations with similar earnings, education, age, and firm-related characteristics, but working in industries with different levels of robot adoption. We also use robot adoption in other countries to isolate the role of technology from industry-specific shocks, and we confirm the pattern using firm-level data on automation. Taken together, these approaches make it more likely that our results capture the impact of automation itself, rather than other underlying differences.

We find that when parents work in highly robot-intensive industries and automated occupations, their children do worse in the long run. In particular, income mobility falls, meaning children’s outcomes are more closely linked to their parents’ income. In contrast, for parents in less-exposed occupations, industry-level robot adoption has no clear effect on their children.

In addition, we find that children from low-income families are more adversely affected by automation. Figure 1 illustrates this pattern, showing how children’s outcomes vary across the parental income distribution, comparing children whose parents in highly robot-exposed occupations worked in high- versus low-robot-adopting industries. Children whose parents were at the bottom of the earnings distribution face a range of adverse outcomes: they earn less, are more likely to experience unemployment, obtain lower school grades, and are less likely to attend university. For children from high-income families in the same industries and occupations, we find no such effects.

Figure 1: How parental automation exposure affects children’s outcomes across the income distribution