The Project at a Glance
The purpose of this project is to put in practice some analytical methodologies that I have learnt in the making of my Master's degree thesis, which you can freely download. At the end of every trading day of the New York Stock Exchange I collect the official high, low, open and close prices of the NYSE Composite Index and use them to forecast tomorrow's volatility. In particular, I estimate a model for the High-Low Range and one for the Realized Volatility. Further details on the models are given here. The realized volatility is a volatility estimator obtained as the sum of squared high frequency returns over a time interval. For instance we may calculate daily volatility by summing all the consecutive squared five-minutes returns over a day. The other proxy for the true underlying volatility is represented by the sum of squared high-low ranges over short intervals, that is the realized range. The two indicators make different use of the same available information set: realized volatility only considers prices sampled at discrete time intervals, while realized range processes the whole information before picking up the two prices thought to give the best indication of intra-interval variability. I am currently performing my forecasts on a daily frequency basis. In this setting, the realized volatility and realized range correspond respectively to the squared return and the squared high-low range. In the next future I will provide forecasts based on intra-daily frequencies, starting from tick-by-tick data. Now go to check the volatility forecasts!