RSM318H1 Lecture Notes - Lecture 1: Risk-Free Interest Rate, Implied Volatility, Central Moment
Document Summary
Lower bound of -1 is problematic if assuming normal distribution. Max loss is 100% but there is no upper bound of potential gains. Stock splits and dividends require adjustment of pt because value of asset has changed. Return of risky asset above risk free rate. Length of "one period" depends on decision interval. More data will not improve estimate of expected return because only beginning and terminal values matter for overall return. Variance of financial assets are time-varying and quite predictable. Can use model such as garch(1,1) to model time-varying volatility. Estimation efficiency is improved by more frequent data sampling: volatility clusters: variances of returns tend not to be constant. High volatility today tends to lead to high volatility tomorrow. Variance and mean are not constant over time. Skewness and kurtosis: lottery-like features: negative returns most of time but small chance of high positive return, often used for testing normality assumption is the k-th central moment of rt.
