Conditional Return Correlation

From the Q-Bank, the “conditional return correlations” can best be defined as correlation that depends upon market volatility.

It says that the observation that correlations increase during financial crises is inconsistent with the assumption that correlation and standard deviation remain stable over time. One solution to this problem is run mean variance analysis and optimize the portfolio over two sets of data: one set assuming normal conditions of higher returns with lower correlations and volatilities plus another set assuming lower returns with higher correlations and volatilities. In other words return and correlation are conditional on volatility.

Why lower correlations and volatilities would imply higher return and vice versa? Is it becase it is less affected by financial crisis? Thanks.

In my eyes cause in Financial Crises higher correlation bring in house higher volatility coming from the financially impacted countries and this might jeoparidise your expected returns?

instead of a stable / static approach where you get one point estimate i.e. correlation and volatilities you try to include a more realistic scenario with the conditional correlation model. Essentially the “conditional correlation” is an improved dynamic model to serve your mean variance analysis embedding also the assumption with higher correlations and volatilities.