I know it doesn't because the paper is linking certain properties it observes without proving the causation, and then tries mention them as if they're relevant while in reality there's no relevance shown.
(For example, an alternative plausible explanation could be a relation between corruption and wealth.)
> By analogy, suppose I have circles, squares and triangles in blue, green and orange. It is meaningful to group by shape. It may also be meaningful to observe that 90% of the triangles are blue. It does not follow that a random blue shape is a triangle.
Imagine shapes and colours can change, and there's no proof the change is related to the shapes or colours. Imagine the conclusion for the 'blue' is 'beautiful'. Does it then matter to mention that the blue ones happen to be 'beautiful' and 'mostly triangles'? Is it a relevant conclusion that the 'non-beautiful' are 'mostly not triangles' or 'that they are not beautiful because they're mostly not triangles'? The latter is what your quote implied, and its merely suggestive. You have no proof.
Because correlation is not causation. So why does the paper mention it? If you're not willing to prove the missing point, why do you quote it?
The paper divides economies into three buckets: market-based, bank-based and underdeveloped. It does this by looking at economic factors. It then looks at what other factors, Y, are similar between countries in each bucket. This doesn't work going backwards.
By analogy, suppose I have circles, squares and triangles in blue, green and orange. It is meaningful to group by shape. It may also be meaningful to observe that 90% of the triangles are blue. It does not follow that a random blue shape is a triangle.