Air Pollution and Infant Mortality in the Developing World



Air Pollution and Infant Mortality in the Developing World

Rema Hanna Faculty ResearcherRema Hanna, Associate Professor of Public Policy, Harvard Kennedy School
Paper TitleDoes the Effect of Pollution on Infant Mortality Differ between Developing and Developed Countries? Evidence from Mexico City
CoauthorsEva Arceo, CIDE; Paulina Oliva, University of California, Santa Barbara

It’s long been known that air pollution poses a major public health threat. In fact, the Organisation for Economic Co-operation and Development (OECD) estimates that, worldwide, almost 1.5 million people die each year from exposure to particulate matter in the atmosphere — far more than die from malaria or from drinking unclean water. With air pollution still growing, more than 3.5 million people may succumb to particulate exposure in 2050, the OECD claims. The biggest risk lies in developing countries, where pollution levels can be orders of magnitude higher than those in developed countries.

Researchers have nevertheless had difficulty determining how air pollution affects infant mortality in developing countries. There are several reasons: Data on pollution and infant mortality are often poor, and results from more technologically advanced nations may not be directly translatable to developing countries due to so-called "nonlinear" effects — meaning that small rises in pollution may have greater consequences in one setting than in another. "If we use the results from developed countries, we might get the wrong answer," notes Kennedy School Associate Professor Rema Hanna.

Hanna and her colleagues found a major metropolis that is well suited for studying the connection between air pollution and infant mortality in the developing world — Mexico City. One of the world’s largest urban centers, the Mexican capital provides an almost ideal site for this kind of analysis because of the availability of robust pollution data drawn from more than two dozen monitoring stations, along with high-quality data from the country’s Ministry of Health.

Pollution levels vary widely across the city, however, and poor people are more likely than rich people to live in highly polluted areas. To disentangle pollution effects from socioeconomic factors, the researchers made use of a natural phenomenon in Mexico City — the frequent occurrence of thermal inversions, during which a layer of warm air settles over a layer of cold air, trapping pollutants below. These inversions affect the metropolitan area indiscriminately, unrelated to where a person lives or his or her income. Elevated pollution levels during an inversion raise the risk of respiratory infection and subsequent death. In fact, thousands of people died in London in 1952 during such an event, which was subsequently dubbed "London’s killer fog."

Hanna’s team found that the number of inversions in a given week provides a good measure of overall pollution levels that week: Each additional inversion, the researchers discovered, caused a 3.5 percent increase in particulate matter and a 5.4 percent increase in carbon monoxide. The team compared the weekly pollution intensity, as reflected by the number of inversions, with the number of respiratory-related deaths among infants and found a strong correlation.

Furthermore, they observed that an increase in carbon monoxide concentrations of one part per billion led to more infant deaths in Mexico City than in the United States — an example of a nonlinear effect. Increases in particulate matter also exacted a health toll, but there was no difference — nor any hint of nonlinearity — between the Mexico City and U.S. data on that measure.

The explanation for carbon monoxide’s accentuated impact during inversions, Hanna suggests, may lie in the fact that inversions typically occur during morning rush hours; 98 percent of the carbon monoxide released in Mexico City comes from motor vehicles, whereas only 36 percent of particulates come from vehicles. This is just speculation, she insists, because the study was not designed to address that question.

But the investigation did show that findings from the developed world cannot always be applied to the developing world. Specifically, an increase in atmospheric carbon monoxide in Mexico City had a greater (and more lethal) health impact than a comparable increase in the less-polluted United States. As a next step, Hanna says, "we’d like to replicate these results in a different setting, taking advantage of another natural feature (like inversions) to isolate the effects of air pollution on infants, our most vulnerable citizens."

—by Steve Nadis

Print print | Email email

“If we use the results from developed countries, we might get the wrong answer.”