Jobs Grew, Wages Did Not: The Local Labor-Market Arithmetic

The export response becomes economically meaningful only when we ask what happened to people in the municipalities that captured more U.S. demand. This chapter separates the labor-market outcomes in the Mexican study: total labor income, the number of employed people, average labor income, employment rates, informality, skill groups, migration, and services. The distinction matters because a municipality can gain jobs and formal coverage without recording a statistically significant increase in average pay. That is not a contradiction; it is a particular pattern of adjustment with particular distributional consequences.

Total labor income is not average pay

The headline estimate is that a 1 percent increase in municipal exposure to the U.S. tariffs on Chinese goods was associated with a 5.6 percent increase in total labor income. Total labor income adds together the labor income of people working in the municipality. It is therefore a measure of the size of the local labor-income pool, not a measure of what the average worker earns.

A local labor-income pool can grow through at least three routes. Existing workers may earn more. More residents may find jobs at roughly the existing average. Or both changes may happen together. The study estimates that the second route was the most visible. A 1 percent increase in exposure was associated with about a 6.3 percent increase in the total number of employed people, while average labor income did not change significantly.

The arithmetic is straightforward. If the number of people receiving labor income rises while the average amount per person stays broadly unchanged, the total rises. The result does not require a fall in pay, either. It says that the measured expansion in the income pool came mainly from adding workers rather than from lifting the average earnings of workers already counted. The distinction between the extensive margin—how many people work—and the intensive margin—how much each worker earns—is the key to reading the evidence.

This is also why “wages did not rise” needs a qualification. The study uses municipal aggregates and reports no statistically significant change in average labor income. It does not follow every incumbent worker's wage through time. Some workers may have received raises, some may have entered at lower or higher pay, and the aggregate average may have remained statistically indistinguishable from its earlier level. The evidence supports a statement about the average measure and the study period, not a claim that every worker's pay was frozen.

The distinction becomes sharper when we compare total employment with the employment rate. Total employment can rise because a municipality has more working people, including people who move in. An employment rate asks how large the employed group is relative to a relevant population. The paper finds no significant effect on the overall employment rate, but it does find a significant increase for semi-skilled workers, reported as 0.41 percentage points for a 1 percent increase in exposure. These outcomes are not redundant. One describes the number of jobs; the other describes employment relative to a population group.

The measures also have different implications for how a local gain is experienced. More jobs can help people who were previously outside paid work, while flat average income can disappoint workers who expected an export boom to improve bargaining power or pay. A fall in informality can bring benefits that do not appear in a wage average.