How does pollution affect the migration and productivity of workers in China?

Alongside rapid growth and industrialization over the past 30 years, China has seen two significant trends: a rise in air pollution and an increase in migration. In an article forthcoming in the American Economic Journal: Applied Economics, EGC affiliate Mushfiq Mobarak and coauthors explore how these trends are connected. Using multiple data sources and measurement techniques, the research team reveals a consistent pattern in which workers choose to migrate away from polluted cities. Then, armed with a spatial model fitted to this data, they quantify the productivity and welfare consequences of this pollution-influenced migration.  

The modeling allows them to identify and quantify a new type of economic cost imposed by pollution: by forcing workers to move away from cities where they would otherwise be more productive, pollution undermines aggregate macroeconomic productivity. Strikingly, this new channel of productivity loss is estimated to be as large as the direct health effects of pollution that had been previously documented. The study also reveals why workers often remain in some cities even when they could earn higher wages and receive better services in others. 

Over the course of China’s industrialization, the country’s pollution levels – measured by the airborne concentration of PM2.5 (particulate matter less than 2.5µm in diameter) – has risen sharply, particularly due to the construction of large coal-fired thermal plants to power the country’s manufacturing industry. Because of this, pollution levels also vary significantly between regions. 

The rate of urbanization and rural-urban migration within China has also increased rapidly in recent decades. But migration trends have not been what they expected: researchers are finding that many Chinese workers – including skilled workers – prefer to work in relatively smaller cities, even if wages are lower. Macroeconomists have puzzled over spatial wage and labor gaps like this within countries, because such differences imply an imperfection in the labor market and a misallocation of labor resources.  

Exploring this complex phenomenon is difficult: residents of a city experiencing even no change in pollution may still see other reasons to migrate, for example if changes in the job market affect the wage rate or quantity of workers and job-seekers. Indeed, pollution and production often go hand-in-hand: pollution adversely affects worker health, while higher levels of production and more workers in a city themselves contribute to pollution. 

To build confidence in their observed causal effect of pollution on migration, Mobarak and coauthors Gaurav Khanna of University of California, San Diego, Ran Song of the National University of Singapore, and Wenquan Liang of Jinan University examined this same relationship using a variety of data sources ranging from census data to panel surveys, verifying their results through numerous robustness checks, and even measuring pollution itself using two different methods – only to reveal the same, consistent pattern of pollution and migration in each. 

They found that this migration negatively impacts the productivity of skilled and unskilled workers alike. Mobarak explained in an EGC interview, “These skilled workers are leaving [polluted] places where they are relatively rare, and going to clean places where they are relatively more abundant.” Meanwhile, unskilled workers tend to stay put, he explained, partly due to China’s hukou policy of complex residency restrictions, which allows skilled workers to internally migrate more than their unskilled counterparts.  

Because of this asymmetry, skilled workers are choosing to leave polluted cities where their skills become scarcer – places where they would be more impactful and productive. And because skilled and unskilled workers complement each other, for example, with skilled employees assisting or managing workers with lower skill levels, the productivity of unskilled workers also declines as their skilled counterparts leave polluted cities.  

To quantify this reduction in productivity, the researchers constructed a model that found that equalizing pollution between a high- and low-pollution city would reduce the skilled worker wage gap by 14%. In a country where employers are reportedly offering up to 20% of wages in relocation assistance, the model clearly reflects the real-world economic costs of pollution. 

“We've documented through this paper a new cost of pollution,” Mobarak said. Existing studies have shown that pollution makes workers less healthy and therefore less productive, and quantified the cost of pollution to workers’ health and productivity. Using their new model, Mobarak and coauthors found that the productivity loss from pollution-induced migration is just as large as from direct health impacts of pollution. “We always knew the economic costs of pollution were bad, but this tells you it’s twice as bad because this new channel is just as large.” 

The authors’ model points to significant distributional inequalities in these costs as well, with hukou policy restrictions undermining unskilled workers’ ability to migrate away from polluted areas – these workers bear the brunt of both air pollution costs and productivity losses in their cities. Reducing policy barriers to migration, in conjunction with pollution control, could therefore result in improved productivity as well as a more equitable distribution of economic benefits in the workforce. 

Research Summary by Peter Zhang.