Simply put: the Kolmogorov-Smirnov test measures similarity, but its significance is never unconditional.
The simplest is to recognize that a statistical test is a tool, not an intrinsic truth.
If empirical data come from a heavily manipulated market, like an economy with strict capital controls, the observed "similarity" is just an artifact.
For example, tests on Chinese sovereign debt market data do not prove a natural distribution, but the impact of PBOC interventions.
The parsimony requires not attributing causality where there is only correlation.