Whales & Prediction Markets
My thesis examines whether “whales,” or unusually large traders, influence prices in prediction markets. Using transaction-level data from Polymarket, I studied whether large trades led to meaningful price movements after the trade occurred. The project identified whale trades based on their share of total market volume and then measuring price changes across different time windows, including several hours and days after each trade. The main result was that whale trades did not have a statistically significant effect on future price changes. This suggests that prediction markets are relatively resilient to large individual trades and may efficiently aggregate information from many participants rather than being easily moved by a single trader. To execute the project, I used Python for data cleaning, transformation, and analysis, including working with APIs, handling large CSV files, and creating event-level datasets. I also applied econometric methods such as fixed-effects regressions, clustered standard errors, and event-study style analysis. The project required skills in data wrangling, statistical modeling, causal reasoning, financial market interpretation, and data visualization to communicate both the empirical results and their broader implications.