Bowdoin Economist Uses Pollution Satellite Data to Assess China's GDP

“So we’re flying in the dark,” Assistant Professor of Economics Stephen Morris said during a recent talk he gave to faculty about his current research. This inaccurate data frustrates China’s national leaders who need truthful numbers to design sound policy and funnel investment to needy areas, Morris explained.

The issue is that the party leaders of China’s 31 provinces, autonomous regions, and municipalities have an incentive to exaggerate their economic achievements, according to experts. If they can impress their superiors with the productivity of their regions, they have a better chance of climbing the political ladder. Some not only misreport GDP but also inflate the amount of electricity generated in their region.

The problems extend beyond China’s domestic politics. The International Monetary Fund also needs good economic output data, because it must assess the credit worthiness of the nations it is investing in, according to Morris.

Morris’ research looks for ways to correct for such false reporting. He and his research partner, Junjie Zhang of Duke Kunshan University, have sought out alternative methods to determine what’s actually happening on the ground in China. The puzzle has forced them to get creative — and to use tools from outside the usual economist’s toolbox.

For example, they turned to satellite instruments flying in the troposphere to see whether atmospheric measurements could validate, or invalidate, numbers reported by regional leaders. In particular, they analyzed public NASA satellite data of nitrogen dioxide emissions coming from China. NO2 is produced by the combustion of fossil fuels and can damage the ozone and cause acid rain. “Even if we can’t measure electricity generation, we can measure things that are correlated with it quite well,” Morris said. “Using these satellite instruments allows us to look at these specific regions and say, ‘It’s that region where the data looks suspicious’.”

To turn data on NO2 emissions into a reliable gauge of GDP, the economists must come up with a formulation that allows them to make a one-to-one comparison between the two. In other words, they have to correct for the fact that one molecule of NO2 does not exactly equal one dollar of GDP. “You have to convert the signal to a proxy, and scale it by some amount to get something comparable with output or income,” Morris explained.

In a working paper that is currently under review, Morris and Zhang zeroed in on data from two recent recessions as useful case studies, including the great recession of 2007-2009. During these two years, China’s aggregate reported GDP and regionally reported electricity production rose. This indicates, in a suspicious way, that as the world economy slowed, a healthy amount of production still appeared to be chugging along in China.

However, when you look at nitrogen dioxide emissions coming out of China during the same time period, another scenario emerges. During the 2007-2009 recession, China’s nitrogen dioxide emissions dipped by 7.8 percent, contradicting the reported 22 percent increase in China’s GDP.

Morris and Zhang also found they could use the NO2 data to look more closely at the regions within China that are most likely reporting faulty GDP numbers. While much of the data coming out of Shanghai, Beijing, and several other regions looked accurate, data from others, including the province of Chongqing, seemed off. “In Chongqing, we see a different story,” Morris said.

In his paper he and Zhang write, “In the midst of the downturn in 2008, the Chongqing model, brought forth by the city’s then Party secretary Bo Xilai, was hailed by many as an exemplary economic initiative.” Yet, as Xilai touted a 65 percent GDP growth year to year, the region’s NO2 emissions did not increase at a level consistent with that growth.

Xilai was eventually deposed from power and sentenced, in 2013, to life in prison for corruption. In addition, three northeast provinces, Liaoning, Jilin, and Heilongjian, also look questionable in light of Morris and Zhang’s economic analysis. Their finding was reinforced by a report by China’s official news agency citing rampant fabrication of data in these regions.

“Our hypothesis is that satellite data collected by NASA is useful for validating data, even from this side of the planet,” Morris said.