Can mass transit make the places it connects safer for women? I study the staggered arrival of first Delhi Metro stations across police jurisdictions between 2016 and 2019. I construct a high-frequency dataset from more than 300,000 police reports, 88% geocoded, and estimate effects using staggered difference-in-differences. A jurisdiction's first station reduces reported harassment of women by 32% relative to the pre-opening mean among treated jurisdictions, with no comparable change in other crimes against women. Harassment falls within one kilometer of a station without evidence of displacement to the adjacent ring. The decline is concentrated in evening hours and neighborhoods with low baseline activity, while independent safety audits show improved lighting, visibility, and perceived safety. These findings identify a gendered place-based safety return to transit that conventional evaluations may overlook, with implications for where new lines and stations are built.

Can civic infrastructure shape political competition? We study public parks in Delhi, combining electoral returns with novel data on the RSS, a Hindu nationalist organization closely associated with the BJP. Using predetermined recreational land allocations from the 1962 Master Plan as an instrument for park access, we find that a one-standard-deviation increase raises the BJP-led alliances (BJP+) vote share by 1.92 percentage points and increases RSS presence. Yet this advantage is not fixed: across four Delhi elections from 2008 to 2020, the relationship between park access and BJP+ vote share is positive in three but weakens sharply in 2013, when the park-related advantage shifts toward the Aam Aadmi Party (AAP), which emerged from an anti-corruption movement that mobilized through public spaces. The findings suggest that civic infrastructure lowers the spatial costs of political organization, benefiting actors able to mobilize through it.

This paper examines how shared religious identity between deposed indigenous rulers and local populations shaped literacy in Colonial India. Using within-district differences across religious groups, we show that Hindu literacy is lower where Hindu rulers were displaced, while Muslim literacy is lower where Muslim rulers were displaced. These patterns do not appear in Princely States, where indigenous rulers retained authority, and are reinforced by instrumental variable strategies. The effects are stronger where rulers had historical ties with local populations, suggesting that removing co-religionist rulers reduced demand for British-led education, highlighting how identity and political displacement shaped the adoption of colonial institutions.