Volatility Networks

Garg, Vaishnav (2022) Volatility Networks. Masters thesis, Indian Institute of Science Education and Research Kolkata.

Text (MS dissertation of Vaishnav Garg (17MS137)) 17MS137_Thesis_file.pdf - Submitted Version Restricted to Repository staff only Download (1MB)

Abstract

Volatility is a measure of degree of variation of stock prices over a time. We use three approaches to measure Volatility for 41 stocks from NIFTY50 index - Realized Volatility (RV), GARCH Volatility (GV) & Options Implied Volatility (IV). The first two are a backward looking measure of volatility whereas the third one is a forward looking measure. We want to inspect whether the IV measure contains some information uncaptured by GV or RV measure, which might help us gain more information about evolution of volatility over time. We also test our hypothesis that a volatility measure in its network is made up of two components - an aggregated market wide volatility component which is common to all stocks, & an idiosyncratic volatility component which is specific to a particular stock. Our primary analysis gives us a twofold result. First, a Granger-Causal analysis shows that IV significantly causes GV which confirms our hypothesis that IV contains some information uncaptured by GV. Second, a PCA analysis on residuals from regression of IV on GV gives us a huge 1st Principal component which rejects our hypothesis and says that there is some third influential factor apart from aggregated market factor that effects the volatility of a stock.

Actions (login required)

View Item