CV/Resume

I'm a PhD Candidate in Economics at Stanford University. I work on projects in applied microeconomics and machine learning.

My graduate research focuses on analyzing the effects of public policies. My current and past projects include investigating whether short-term rental bans achieve their policy goals, analyzing the impact of international environmental regulation on firm innovation, quantifying the impact of education policy on the diffusion of agricultural innovations, and understanding the impact of food-safety laws on mortality.

Research Papers

These projects investigate how regulatory interventions influence market outcomes and social welfare. Each paper constitutes ongoing research and represents work-in-progress.

Do Short-Term Rental Restrictions Achieve their Policy Goals?

In this paper, I investigate whether city-level short-term rental (STR) restrictions increase the supply of long-term housing and improve housing affordability.

To do this, I study eight U.S. cities that imposed restrictions on short-term rentals between 2021 and 2025. Using synthetic control methods, I estimate city-specific policy effects and document considerable variation in policy stringency and outcomes across jurisdictions. Next, I use difference-in-differences to provide a global estimate of the average effect of the policies I study. On average, the policies reduced the number of active listings by 10-15 percentage points and the number of active short-term rental listings by 20-40 percentage points, suggesting that many properties shift from providing short-term rental accommodations to medium-term rental accommodations lasting one month or more. I also find that secondary-unit listings declined more than primary-unit listings in most cities. However, my analysis does not yield conclusive evidence on housing affordability outcomes.

Spillover Effects of Short-Term Rental Regulations: Evidence from New York City

In this project, I investigate how New York City’s near-total ban on short-term rentals impacted nearby jurisdictions. Using Airbnb listing data and synthetic control methods, I estimate the policy’s impact on the supply of short- and medium-term rental listings and long-term rent prices in Jersey City and Newark. I find that Jersey City saw a 26 percentage point increase in active listings and a 39 percentage point increase in active STR listings, driven largely by growth in secondary-unit listings. Newark experienced a 35 percentage point rise in active listings and a 30 percentage point increase in active STR listings, with increases across both secondary-and primary-unit listings. I also find that Newark experienced post-treatment rent price growth that was 8 percentage points higher than that of the synthetic control unit. These results provide causal evidence of spatial spillovers and suggest that isolated STR policies may inadvertently shift housing pressure to adjacent jurisdictions.

Milk Pasteurization and Mortality with James Feigenbaum and Lauren Hoehn-Velasco

In this analysis, we study the role of milk pasteurization in the decline of typhoid fever in early 20th-century U.S. cities. Using an event-study design, we estimate the causal effect of city-level pasteurization ordinances on typhoid mortality and morbidity. We find that pasteurization reduced typhoid mortality and case rates by 28–30%, with estimated effects comparable in magnitude to those of water filtration. These results suggest that milk safety reforms, particularly pasteurization, played a critical and previously underappreciated role in public health improvements during this period.

Machine Learning Projects

Predicting Airbnb Prices using Supervised Machine Learning

Conducted data cleaning (data-type recasting, managing sparse classes, encoding categorical variables), feature engineering, and exploratory analysis
Set up pipelines to tune and train supervised machine learning model
Developed models using algorithms with embedded feature selection, including regularized regression techniques (Lasso, Ridge, Elastic Net) and tree-based ensemble methods (Random Forest, Gradient Boosting, XGBoost)
Tuned hyperparameters using cross-validation
Used performance metrics to choose best model, selecting the best based on R^2 and MAE
Results: Best model had an R^2 of 0.7 and MAE of $53 or 2/5 of an SD

Customer Segmentation using Cluster Analysis

Used K-means clustering to identify e-commerce customer segments based on past purchases
Conducted feature extraction and dimensionality reduction using Principal Component Analysis (PCA)

Other Contributions

The Economic Impacts of COVID-19: Evidence from a New Public Database Built Using Private Sector Data

The Quarterly Journal of Economics, Raj Chetty, John Friedman, Michael Stepner, and Opportunity Insights Team

Dealer Trading and Positioning in Floating Rate Notes

Liberty Street Economics, Michael Fleming and Amanda Wahlers

November 2024: This website is under construction.