Research
Published Papers
We report on an experiment examining why default options impact behavior. By randomly assigning employees to different varieties of a salary-linked savings account, we find that default enrollment increases participation by 40 percentage points---an effect equivalent to providing a 50% matching incentive. We then use a series of experimental interventions to differentiate between explanations for the default effect, which we conclude is driven largely by present-biased preferences and the cognitive cost of thinking through different savings scenarios. Default assignment also changes employees' attitudes toward saving, and makes them more likely to actively decide to save after the study concludes.
Firms use relational contracts to support repeated trade. Do these informal agreements evolve in response to market conditions? In a market for ice, firms reestablish relationships on new terms when a prior agreement breaks down. Using transaction data, we show that ice retailers prioritize deliveries to loyal buyers -- fishing firms -- when supply from the monopolistic manufacturer is scarce. After an upstream competition shock increases supply, repeated trade lapses, threatening retailers' positions. Incumbent retailers establish a new agreement expanding trade credit to loyal buyers, which impedes new retailer entry. Upstream competition also increases downstream firms' productivity and lowers consumer fish prices.
We provide evidence that violence reduces the adoption and use of mobile money in three separate empirical settings in Afghanistan. First, we spatially merge nationwide administrative data on 96,000 violent events with the universe of mobile money transactions and find that users exposed to nearby violence reduce their mobile money account balances and conduct fewer transactions. Second, using high-frequency panel survey data from a field experiment, we find that subjects expecting violence are half as likely to respond to a randomized mobile money supply shock as those not expecting violence. Finally, analyzing financial survey data from nineteen of Afghanistan's 34 provinces, we find that individuals expecting violence hold more cash. Collectively, our evidence suggests that violence can impede the growth of formal financial systems.
Existing explanations for the well-established relationship between mining and conflict predominantly interpret violence near mines as conflict over territory or government. We provide evidence that competition between artisanal and industrial miners is also an important source of natural resources-related conflict, drawing on qualitative case studies at mining sites in the Democratic Republic of Congo and Zimbabwe, and a large-N analysis. For the latter, we use machine learning to estimate the feasibility of artisanal mining across the continent of Africa based on geological conditions. The impact of price shocks on violent conflict is over three times larger in locations with industrial mining where artisanal mining is feasible than in places with industrial mining unsuitable for artisanal mining. Our estimates suggest that 31 to 55% of the observed mining-conflict relationship is due to violent industrial-artisanal miner competition. This implies new avenues for conflict-mitigation as the clean energy transition increases demand for minerals.
Working Papers
“Insecurity and Firm Displacement: Evidence from Afghan Corporate Phone Records,” with Joshua Blumenstock, Sylvan Herskowitz, Ethan Kapstein, Thomas Scherer and Ott Toomet. (2nd R&R, AEJ: Economic Policy)
We provide empirical evidence on how insecurity affects firm behavior by linking data on deadly terrorist attacks in Afghanistan to geolocated data on corporate mobile phone activity. We first develop an approach to estimate the geographic footprint of firms from employee locations. Using these measures, our main analysis finds that violent shocks reduce local firm presence by both increasing firm exit and decreasing entry. Firms react most to violence in their ‘headquarters’ district. We further find suggestive evidence of persistence, stronger impacts in more secure districts, and spillovers, whereby attacks in provincial capitals reduce firm presence in surrounding rural districts.
Global hunger levels have set new records in each of the last three years, outpacing aid budgets. Most households experiencing food insecurity crises are now in fragile states, making it difficult to support vulnerable, hard-to-reach populations without interference from oppressive governments and non-state actors. Despite growing interest in using digital payments for crisis response, there is limited causal evidence on their efficacy. We show that digital payments can address basic needs during a humanitarian crisis. We conducted a randomized evaluation among very poor, mostly tech-illiterate, female-headed households in Afghanistan with digital transfers of $45 every two weeks for two months. Digital aid led to significant improvements in nutrition and in mental well-being. We find high usage rates, no evidence of diversion by the Taliban despite rigorous checks, and that 80% of recipients would prefer digital aid rather than pay a 2.5% fee to receive aid in cash. Conservative assumptions put the cost of delivery under 7 cents per dollar, which is 10 cents per dollar less than the World Food Program’s global figure for cash-based humanitarian assistance. These savings could help reduce hunger without additional resources. Asked to predict our findings, policymakers and experts underestimated beneficiaries’ ability to use digital transfers and overestimated the likelihood of diversion and the costs of delivery. Such misperceptions might impede the adoption of new technologies during crises. These results highlight the potential for digital aid to complement existing approaches in supporting vulnerable populations during crises.
Building state capacity is uniquely challenging in fragile states. We report results from a randomized evaluation of a major Afghan government initiative to increase capacity by modernizing its payroll. The reform, which required teachers to biometrically register and receive salary payments via mobile money, did little to reduce payments to non-existent “ghost” workers, but significantly reduced delays. The reform also improved educational outcomes and increased formal financial inclusion. The impacts were not immediate – highlighting the importance of long time-horizons – and were largest in urban areas. The results have implications for state-building and are potentially actionable for policymakers.