Insecurity and Business Displacement in Afghanistan

• Ethan B. Kapstein (Princeton University), Sylvan Herskowitz (World Bank), Joshua E. Blumenstock (University of California, Berkeley), Tarek Ghani (Washington University in St. Louis), Thomas L. Scherer (University of California, San Diego), and Ott Toomet (University of Washington)
• Insecurity and Firm Displacement: Evidence from Afghan Corporate Phone Records
• Journal: American Economic Journal: Economic Policy Vol. 18, No. 1, February 2026

During the civil war in Afghanistan, Taliban-linked violence disrupted economic life, with local businesses struggling to maintain economic growth amid persistent disruption and fragility. Yet tracking how firms responded to the insecurity presented a challenge to researchers because of the limited amount of data available. In new research published in the American Economic Journal, Ethan Kapstein, a lecturer and executive director of the Empirical Studies of Conflict Project, and his coauthors addressed this gap by developing a new way to measure business displacement.

The researchers introduced what they described as “a new approach to measuring the geographic footprint of private firms using corporate mobile phone metadata.” Specifically, they relied on call detail records (CDR) — the automatic logs created when a corporate mobile phone places a call. Although CDR does not capture the content of conversations, it records the time of the call and the cell tower used, allowing the researchers to approximate the caller’s location. By tracking the geographic presence of employees affiliated with each company across cell tower districts before and after major attacks, Kapstein and his colleagues mapped the geographic presence of these firms and observed how it shifted in response to violent outbreaks.

The team analyzed CDR from one of Afghanistan’s largest mobile operators, covering 2,292 private firms, over 125,000 associated phone numbers, and 217 million corporate calls across 173 geographical districts from April 2013 through December 2016. This dataset enabled the authors to examine how firms adjusted their presence immediately after major attacks and in the months that followed.

The research team identified several key findings about how firms respond to major outbreaks of violence.

First, violent attacks caused a shock that reduced local business presence instantaneously after such an event. After an instance of violence, firms immediately reduced their presence in the affected district by around 3-5% in the month right after the attack.

Second, the likelihood of businesses entering the area decreased, while efforts to leave the area rose. In the same month after a violent attack, companies that were in the district are 5-23% more likely to exit, while ones that are not are 7-16% less likely to enter.

Third, firms responded most strongly to violence in the area that researchers concluded are their “headquarters” district, which researchers identified as the district where the majority of a firm’s employees are based on the phone data.

Fourth, the paper finds that there was a “spillover effect” that emerged in the months after the initial attack in surrounding regions. These attacks in provincial capitals may spread to nearby rural districts while producing stronger displacement in more secure settings.

Lastly, the team concluded that using their CDR approach can help researchers gain virtual access to areas that are hard or unsafe to do so, notably in conflict zones.

“Private sector development has been a key element in many bilateral and multilateral foreign assistance programs,” Kapstein said. “But measuring firm activity in conflict zones and understanding how firms respond to violent attacks is essential if these programs are to be effective. Specifically, it is crucial to address the security environment that the private sector faces and to ensure that firms are able to operate safely if they are to grow.”