Connectivity and Household Power: Fast Internet and Women's Empowerment in Egypt (with Reem Taha)
We examine how the expansion of high-speed internet affected women's household outcomes in Egypt. We exploit plausibly exogenous variation generated by the rollout of submarine fiber-optic cables and the spatial distribution of Egypt's terrestrial backbone network. We find that internet expansion increased women's exposure to emotional and sexual violence by 15.6 and 12.9 percentage points, respectively, while having little effect on physical violence. It also raised the likelihood that a woman was employed in the previous year by 1.4 percentage points. Participation in household decisions rose as well, by 5 to 25 percentage points, including decisions about her own health care and earnings. We find no evidence of effects on attitudes toward intimate partner violence, marriage-market sorting, or the timing of marriage and divorce. Taken together, the results suggest that high-speed internet reshaped women's position within the household. It increased their economic participation and decision-making authority, but also their exposure to violence, rather than altering partner characteristics or social attitudes. This pattern is consistent with gains in women's autonomy provoking conflict within the household rather than a broader shift in social norms.
Beyond Survival: The Paycheck Protection Program and New Business Formation (with Tannista Banerjee) (Draft on request)Â
This paper asks a simple question: did the Paycheck Protection Program (PPP) cause new businesses to form in 2020 and 2021? We use the full sample of PPP loans, matched to county-by-size cells, along with Homebase data on establishments and employment. We find no detectable effect. None of the 190 event-time coefficients we estimate is statistically distinguishable from zero. We can rule out any effect larger than about 1.6%. Given that PPP disbursed $790 billion across 11.5 million loans, this null result matters on its own: a program this large left no measurable mark on entrepreneurship.
Along the way, we also flag two design problems worth watching for in other studies. Sample-based denominators can create fake trends that still pass standard pre-trend tests. Also, treatment measures correlated with a confound at treatment start can produce the same false effect. The only way to catch it is a known-zero placebo.
Effects of Marketing Authorization Holder System: Evidence from China's Medical Device Industry (with Tannista Banerjee and Yiyu Xing) (Draft on request)
Disentangling the Solow Residual: What WFH and AI Diffusion Reveal About U.S. Productivity Growth (with Arun K. Srinivasan and Dr Kathleen Arano)
The Impact of Child Tax Credit on Household Health Recovery
DUTTA, S., BOSE, M. and RUNGTA, R., Discourse on Ambulatory Healthcare in Uttar Pradesh, India: A Study Based on NSS Data. BUSINESS, ECONOMICS AND SUSTAINABLE DEVELOPMENT, p.218