From a high-level perspective, the image becomes simpler: performing Kolmogorov-Smirnov tests, although technically correct, does not guarantee the fundamental validity of conclusions.
The relevance of input data is a controllable and essential factor; ignoring this introduces noise.
For example, if reference data do not reflect the specificities of the Hong Kong market or cross-border flows, the KS analysis becomes an abstraction.
The discipline requires not to confuse a statistical similarity with proof of compliance or actual performance, as might be the case with illiquid portfolios.
The focus should remain on model assumptions and input quality, the rest being secondary.