Series Filter
A conditional model for estimating expected risk and correlations.
Last updated
A conditional model for estimating expected risk and correlations.
Last updated
The Series Filter model partitions the full historical asset sample conditioned on the time series properties of a single reference asset within the portfolio.
For example, a client can use this risk model to test the hypothesis that the correlation between an asset pair increases when the S&P 500 has a negative return.
The investor must first specify which asset is the reference asset. Second, the investor must specify the threshold, or filter, to apply. The WPA allows users to specify filters in return or standard deviation units.
Using the model parameters, the WPA estimates risk and correlations based on the sub-sample of asset returns.
