I imagine the number of people who would do it if they theoretically knew they had no chance of getting caught is different than the number of people who actually do it. I don't disagree with your conclusion about how many people do, but knowing how many people would lie, cheat, steal, or murder their way to wealth but don't due to sufficient deterrent is useful knowledge in how to structure a society.

To be clear, I'm not making any claims about whether this is a large proportion or not, because I have absolutely no idea (and I have doubts this would even be possible to calculate with even a remote degree of confidence purely via philosophical discussion). If anything, some sort of study that provides evidence that this number is lower than expected would be a strong argument against typical "tough on crime" policies that are often popular with people who express concern about human nature in this regard.

Agreed; an equally flawed assertion.

In my view we have some unflinching morals, some more flexible ones, and some you don't adhere to at all, and which is which tends to differ between people.

I personally don't believe in non-religious ontological good because of this aspect of human nature.