Let's forget our past investments: simply executing Kolmogorov-Smirnov (KS) tests does not justify a hasty conclusion about distributional similarity.
We need to question the quality of reference data and the relevance of truncation thresholds used.
If the reference is outdated or the thresholds are arbitrary, the test may indicate a false similarity, hiding underlying systemic risks.
For example, comparing current financial data with a pre-2008 reference would ignore deep structural changes, such as the volatility of the Eurozone government bond market.
A realization requires robust assumptions, not just a process.