Empirical features of returns 22
Simple and log returns cannot be distinguished in such graphs
Erratic (“
white noise
”), strongly oscillating behavior of returns around the
more or less constant mean (“
stationary process
” i.p. “
mean
reverting
”)
Variance / volatility (sta
ndard deviation) is no
t constant over time
(“
heteroskedasticity
”) i.p. we have
periods of different length with
approximately the same degree of variation (
“volatility clustering
”)
Single-period random cash flows: Stocks