The robustness of CIB data does not guarantee low NPAs, while non-robustness of the data may pose a risk in the future. Do not confuse correlation for causation.
Also this is not about trusting an analyst more than the promoter. Both have their vested interests and opposing views. The promoter says the public systems are robust, the analyst says the public systems are not. Upto us to choose whom to believe. The analyst presented his arguments well to cause suspicion on the robustness of the data. The promoter did not address the concerns for me to drop my suspicions. That was the point I was trying to put forward. Where is the proof that the data is robust aside from a correlation metric?
Problems arising from an incomplete database may or may not develop in the future. Who knows? NPAs are non-existent today and may or may not rise in the future.