i’m taking the BSAS afternoon session one vignette at a time b/c that’s all i can handle right now. in quant- there’s a question # 9 that asks about time series stuff. answer says “in general, any time series accurately described with a linear or log-linear trend model is not covariance stationary”. why is this?
I would say in general time series are not covariance stationary in finance.
I think the reason is the non-constant mean. If the linear and log - linear trend model fits perfectly with the data, then the mean will keep on increasing. Only exception will be linear line parallel to x-axis.
I think Kabhii’s reasoning is correct.
kabhii is right. The series will at least have a non-constant mean. Plus a log linear trend series will most likely have a non-constant variance.
bannisja Wrote: ------------------------------------------------------- > i’m taking the BSAS afternoon session one vignette > at a time b/c that’s all i can handle right now. > in quant- there’s a question # 9 that asks about > time series stuff. answer says “in general, any > time series accurately described with a linear or > log-linear trend model is not covariance > stationary”. > > why is this? a “trend” is never “stationary”