Work, Sectors, and the Shape of Mobility

Mexico’s 1993 compulsory-schooling reform changed the economic and geographic paths available to some adults, but not in the simple form of “more education, higher wages, more migration.” This chapter follows reform-induced schooling into employment, sectoral allocation, and movement through space. The evidence supports a more precise judgment: schooling helped some people enter or remain in work and shifted workers away from agriculture, while mobility changed in composition rather than moving in one direction. Current OECD wage and unemployment figures can provide context, but they cannot be substituted for the reform’s causal estimates.

Work changes in more than one dimension

The reform study cannot provide a clean causal earnings estimate. Census earnings data have high non-response, so the analysis focuses on outcomes that are more reliably observed: whether a person reports paid work, how many hours employed people work, and which sector employs them. That choice changes the question from “What wage did education buy?” to “How did education change participation and the allocation of labour?”

Employment gains were not evenly distributed. The largest reported increase appeared among people speaking an Indigenous language: an additional year of schooling raised the probability of employment by 5.5 percentage points from a baseline of 41 percent. Among the same group, weekly hours rose by about 1.16 from a baseline of 39.7 hours. No subgroup showed a reduction in employment. These estimates suggest that the schooling margin mattered most for people whose prior educational position was also associated with weaker labour-market integration.

That result should not be rewritten as a universal employment guarantee. The instrumental-variable estimate applies to compliers whose schooling rose because of the reform. It also measures a person’s employment status at the time of the census, not a complete career history, job quality, earnings path, or protection from unemployment. Still, it adds something a wage comparison cannot: it shows that education can change whether and how people participate in work even when a causal pay estimate is unavailable.

The clearest labour-market result concerns sectoral allocation. The probability of working in agriculture fell by about 1.2 percentage points overall. The decline was larger in small rural communities, at about 2.9 points from a baseline agriculture share of 34 percent, and among Indigenous workers, at about 2.8 points from a baseline of 39 percent. Some of the reduction was offset by movement into manufacturing and services, including a reported increase in health-sector employment.

Sector matters because a person can be working before and after a policy while doing fundamentally different economic work. A shift out of agriculture may indicate access to a more diversified set of occupations, particularly where farming had accounted for a large share of employment. It does not by itself prove that every new job was more productive, better paid, or more secure. The study interprets the pattern as occupational upgrading away from low-productivity agriculture, but the data do not provide a complete measure of productivity, wages, or job quality for every transition.

The subgroup pattern gives the sectoral result its meaning.