Why the Gains Favor Some Workers
The worker-level evidence from Viet Nam's robot-exposed districts explains how total employment can rise while formal work disappears for particular groups. This chapter identifies the tasks and workers behind the average result: who gains from complementarity with new production, who is exposed to routine-task substitution, and how informality or movement can hide the loss from the unemployment count. The aim is to read “more jobs” as a distributional outcome, not as a uniform improvement.
The distinction is not a minor qualification. A robot does not meet a representative worker. It changes the value of particular tasks, and firms then search for workers who fit the production system that remains. Education, occupation, age, gender, wage position, and the ability to move can all affect whether a worker reaches the expanding side of the market or absorbs the adjustment on the contracting side.
The first split is between complementary and routine work
The study finds that the positive employment effect is driven mainly by more-skilled workers, especially those with college education. Robot exposure increases high-skilled employment in both formal and informal work, with reported effects in the range of roughly 5–15%. This does not mean that every educated worker benefits or that education alone causes the gain. It indicates that the expanding production system demands capabilities that are more common among workers with higher levels of formal education, while the measured exposure is concentrated in industries where machines and skilled labor can operate together.
The contrast appears in low-skilled formal employment. In districts with greater robot exposure, formal wage employment among low-skilled workers falls by about 15%. The coexistence of these estimates is the core distributional fact: high-skilled employment can increase enough to lift the district total while low-skilled formal workers lose positions. The positive aggregate coefficient is therefore not evidence that the low-skilled workers were simply transferred into the new jobs.
Occupation makes the pattern more concrete. Displacement is concentrated among low-skilled formal workers in elementary occupations, plant and machine operator jobs, and craft and trade work. These categories contain many routine manual tasks—repeated movements, standardized handling, and predictable production steps—that are more directly exposed when industrial automation becomes cheaper and more capable. The evidence does not say that everyone in one occupational label performs the same tasks. It shows that the measured losses cluster where routine manual work is a larger part of the job.
This is why “low-skilled” should not be treated as a synonym for “unproductive.” Workers can be valuable to a production process and still be vulnerable if the particular tasks they perform can be standardized. Conversely, the gains associated with higher education may partly reflect access to complementary tasks, such as supervision, quality control, coordination, or technical adjustment. The study's education categories are proxies for these capabilities. They help identify the distribution of effects but do not reveal every worker's full skill set.
The results also distinguish formal and informal destinations. Low-skilled workers in routine manual occupations experience a significant increase in informality in more robot-exposed districts.