Trading volume, range-based volatility and downside risk in the AI equity rally: evidence from NVIDIA
Abstract
This study examines the interaction between trading activity, volatility based on price ranges, and downside risk during the recent rally in artificial intelligence-related stocks, using 81 daily observations of NVIDIA from January 9, 2024, to May 3, 2024. The empirical design combines logarithmic returns with the Parkinson and Garman-Klass volatility estimators, calculated from daily opening, high, low, and closing prices. Market stress is identified endogenously through the quartiles of volatility based on price ranges, while downside risk is defined by the lower 5% tail of the distribution of daily returns. The sample covers an exceptionally information-rich period in the semiconductor markets, including the acceleration in NVIDIA’s stock price appreciation around its earnings announcements in early 2024. Robust ordinary least squares (OLS) estimates, quantile regressions, and a logit model for the lower tail are used to assess whether trading volume contains incremental information about price dispersion. The results show a positive relationship between trading volume and volatility, although the strength of this relationship varies across the volatility distribution. The evidence also reveals a substantial downside risk, despite the strong upward trend in prices, highlighting that periods of high returns can coexist with marked intraday uncertainty. The results contribute to the recent literature on market microstructure by demonstrating how trading intensity, volatility based on price ranges, and downside risk interact during a period of strong, concentrated growth in the technology market.