Abstract
A model of risk with multiple independent unconditional calendar and non-calendar variance components is used to explain time-varying returns. Digital signals represent finite stock return series. The random walk hypothesis is tested using digital signal processing methods. A stochastic additive market noise model measures total and idiosyncratic risks of return signals. With monthly returns and four year signals the white noise hypothesis is rejected using small sample signal processing methods. Large firms have four-year and one-year calendar risk. Mid-cap firms have two-year and six-month risk. Small firms display January-like risk. Unconditional calendar based risk appears coincident to calendar return anomalies.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call