Research

Working Papers

Conditioning or Orientation? Shrinkage Targets for Characteristic-Managed Portfolios (September 2026). When estimating the covariance of characteristic-managed portfolios, I try shrinking toward the covariance of the characteristics themselves. Out of sample, it has the highest Sharpe ratio of the 23 methods and benchmarks I compare, but it is statistically tied with a plain scaled-identity target. A decomposition shows that most of the gain comes from regularization in general, not from the information in the characteristics. covariance shrinkagehigh-dimensional statisticsportfolio optimizationpermutation tests
Returns on Presale Housing Markets in India (2025). Measures returns on housing bought before construction in Bangalore, NCR, Chennai, and Pune. Nominal returns are 2 to 5% a year, earned mostly right after launch, with no extra return for holding until completion. Written while I was a research associate at IIM Bangalore. real estaterepeat-sales indicesemerging markets

IIM Bangalore, with Venkatesh Panchapagesan (2024 to 2025). Pricing presale real estate.
Harvard Business School, with Victoria Ivashina (2023 to 2024). Pension portfolio allocation and the cost of illiquidity in alternative assets.
Harvard Business School, with Josh Lerner (2023 to 2024). Innovation and acquisitions among financial firms, using patent data and synthetic control.
Chicago Booth, with Dacheng Xiu (2022 to 2023). ML methods for pricing errors in housing and consumer credit; default prediction models.
Indian Statistical Institute, with Tanujit Chakraborty (summer 2019). Ensemble ML methods.

UCLA Anderson

Chicago Booth