By Ben Balint-Kurti and John Martin
Labor markets across North America are facing profound shifts. The industrial makeup of the United States and Mexico is constantly changing, and in some cases is leaving regions behind. To respond to these regional shocks, we need tools to capture what is happening at the local level. These ideas were at the center of the panel "New Developments in the Labor Markets of Mexico and the US", organized by Reimagining the Economy at Harvard Kennedy School, ITAM University, and Banco de Mexico.
Our panel featured Daniel Chiquiar, Professor of Economics at ITAM; Gordon Hanson, Faculty Co-Director of Reimagining the Economy at Harvard Kennedy School; and Lorenzo Aldeco, from the Directorate of Price Analysis and Regional Economics at Banco de Mexico.
Here are some of the key takeaways:
Constructing data according to economic geography allows for insights that aren’t otherwise possible.
Aldeco announced the launch of the Database for Local Labor Market Analysis in Mexico, a data tool developed by the Banco de Mexico. This tool constructs data on local labor markets, which are defined as regions of multiple different municipalities grouped into a single economic unit. The definition is constructed by identifying municipalities which are strongly linked in worker commuting and are thus similarly impacted during economic shocks. This data can be used in many ways, but one application Aldeco chose to explore was general labor market trends across Mexico. He observed how while employment increased from 1990-2020, the skill premium – or the ratio of skilled wages to unskilled wages – steadily fell across all regions. This implies a steady increase of skilled labor supply and a general shift toward a more educated workforce.
Professor Chiquiar elaborated on how these local measures allow one to analyze the effects of a policy change on a specific area. Aggregate measures, such as the national unemployment rate, can obscure a policy’s varying effects on different specific local labor markets – for instance, if unemployment increased in half of the local labor markets and decreased in the other half, the aggregate might incorrectly show that the policy had no effect. Also, because areas are now grouped in terms of economic activity, the data is better aligned to the types of questions economists want to ask with a focus on causal effects.
NAFTA had real – but not transformative – positive effects for Mexican workers.
Prof. Daniel Chiquiar used the new Local Labor Market measures to reevaluate prior estimations of the effect of trade liberalization and the 2018 U.S.-China trade war on the Mexican economy. First he presented new research that claimed that NAFTA led to a decrease in the skill premium – the average difference in wages for a “skilled” or formally educated worker versus an “unskilled” one – in Mexican labor markets that were close to the U.S.-Mexico border. Distance to the border is important because the data show that close regions export more.
The new data also allowed him to find that NAFTA decreased the wage gap between men and women: in combination with the previous finding, this meant that unskilled women benefited the most from NAFTA in Mexico.
These positive effects need to be contextualized within Mexico’s major decline national aggregate productivity and income after NAFTA’s passage. How could this have occurred if NAFTA was good for Mexico’s economy? Chiquiar answered that productivity has been decreasing in Mexico for decades, and that the only time it was not was directly post-NAFTA. While NAFTA did help, then, it was still unable to stem a long-run tide of decreasing economic performance in Mexico. Only by doing analysis at a local level using the Database measures is it apparent that NAFTA helped certain areas in Mexico (at least relative to others) – a fact obscured by looking only at national or state-level aggregates.
Multiple trade shocks from China reshaped Mexican labor markets.
Trade shocks caused by China’s entry into globalized trade and the U.S.-China Trade War both also affected the Mexican economy – in ways both similar and different to the U.S. Using differences in the products certain local labor markets make, Chiquiar identified how imports affect local economies and showed that the China Shock caused wages to decline in Mexico (places that used to make products that are now made in China saw the largest declines). The effect was especially pronounced for so-called “unskilled” workers. The China Shock’s effect in the Mexican context, then, was largely parallel to Autor, Dorn, and Hanson’s finding for the U.S. – that increased import competition from China decreased wages in places in the United States that produced those products.
In contrast, Chiquiar showed that the U.S.-China trade war increased manufacturing in certain sectors of the Mexican economy. At the time of the trade war, some theorized that when the U.S. made it more difficult for U.S. firms to manufacture in China, this might lead to a “nearshoring” effect where U.S. firms move their manufacturing to Mexico instead. Chiquiar’s analysis supported this theory: Mexican manufacturing increased most in the regions that were most exposed to the U.S.-China trade war. Similarly to the China Shock analysis, exposure to the trade war was measured in terms of whether the regions produced specific products that were often manufactured in China and/or affected by tariffs.
Development strategies for Mexico must be broad-based and come from within: taking advantage of foreign demand alone is not sufficient.
To conclude the talk, Chiquiar offered some policy prescriptions based on his empirical findings. Contrary to optimistic predictions that NAFTA would transform the Mexican economy – allowing it to transition to making high-tech goods and “move up the value chain” – NAFTA didn’t dramatically change the structure of Mexico’s economy, apart from its effects on border states that increased their shares of export-oriented manufacturing. To this point, Chiquiar noted that even if demand for high-tech goods from Mexico increased under NAFTA, this didn’t necessarily lead to an increase in skilled, high value-added employment in Mexico. This is because Mexico often acts only as the final stage of assembly of high-tech products; the “skilled” part of the production process may still happen in places like China or South Korea, with the intermediate products only sent to Mexico for final, relatively “unskilled” assembly. Prof. Hanson agreed that in the U.S. high-tech manufacturing also doesn’t necessarily imply high-skill jobs, noting that under President Biden’s tech agenda more investment went towards training manufacturing workers than scientists. With this in mind, Chiquiar argued that Mexico had to commit to using broad policy like education to develop high value-added jobs from within, rather than rely on trade liberalization and nearshoring to bring it about.
