Authors: Till Raphael Saenger, Ethan B. Kapstein, Ronnie Sircar (all Princeton University)
Title: Estimating the collapse of Afghanistan’s economy using nightlights data
Journal: PLoS ONE

The Taliban’s August 2021 takeover of Afghanistan led to a significant downturn in the country's economy. But up to now, no one has known exactly how significant. Official data are both scarce and unreliable, and the Taliban has not released any information regarding Afghanistan’s gross domestic product.

The World Bank estimated a drop in the country’s gross domestic product of around one-third, while the United Nations Development Program stated that Afghanistan had suffered an “economic implosion.” These organizations relied largely on survey data as the basis of their claims.

A group of Princeton researchers, including Ethan Kapstein, executive director of Princeton SPIA’s Empirical Studies of Conflict Project, who has conducted research in Afghanistan, set about trying to gain a more accurate assessment by using nighttime light (NTL) emissions as a proxy for economic activity.

They analyzed satellite images of NTL in Afghanistan from 2020 to 2022, comparing them to NTL from neighboring countries, including Pakistan and Iran, that did not undergo regime change during that time and thus served as a control group.

“Our methodology stresses the importance of evaluating the change associated with the Taliban take-over relative to the counterfactual growth in the absence of the government’s collapse,” the researchers write, “rather than simply focusing on the economic level immediately before the takeover.”

Kapstein and his colleagues estimated that the drop in Afghanistan’s GDP was approximately 16%, well below the U.N.’s and World Bank’s figures but still firmly indicative of a devastating economic downturn.

“[T]here is little doubt that Afghanistan has suffered a major economic shock following the Taliban takeover, likely due to reductions in international aid and Western military spending,” they write. “More specifically, our work serves to isolate the economic shock associated with the Taliban takeover, shifting the country from a positive growth trend to a deep recession.”

That Afghanistan’s economy has suffered under Taliban rule is unsurprising. The larger implication of Kapstein and his colleagues’ work is the way they measured the extent of the suffering, using NTL to assess changes in GDP.

Afghanistan is hardly the only country where traditional economic data is hard to come by, and with democratic institutions under increasing threat, “the number of cases of such data-scarce regions may grow,” the authors argue, “creating a need for innovative approaches to data gathering and analysis.”

“Further,” they write, “researchers could potentially use this methodology to explore such questions as the effects of Western sanctions on the Russian or Iranian economies. In these settings, the use of relatively high-frequency data such as nightlights enables researchers to get closer to real-time analysis.”

The researchers note other potential novel ways to measure economic changes, including cellphone call data and social media activity.

“These new data sources, alongside the synthetic control methodology, promise more accurate and faster evaluations of regional shocks, providing researchers and policymakers with a powerful tool in support of economic analysis and potential interventions,” they conclude.