No, I don't. I think it's well-correlated, but that correlation varies and people's estimation of it is likely subject to a severe Dunning-Kruger effect - most people likely consider themselves to be above-average judges of character.
Now, as for causality, the more social and economic opportunities you withhold from ex-felons - never allowing them to vote, continually expanding the scope declaration requirements, etc. - the more attractive an option recidivism into criminal activity becomes, in proportion to the degree to which you discount their likely future contribution to society. Put another way, if there's no benefits on offer for going and staying on the straight-and-narrow, then why bother?
The belief you're expressing is known as the 'fundamental attribution error' - a tendency to overestimate the explanatory power of inherent characteristics and discount the role of circumstances. This is no wiser than taking the view that everyone is fundamentally good and that they're mindlessly conditioned by social factors.