Naked Capitalism has the article Finance and Growth: The Direction of Causality.
Our findings also have an implication for the presumption after the start of the Great Recession that resuming credit flows was essential for economic recovery. In the Eurozone, where the recovery has lagged, efforts to create a greater supply of credit have been an important policy focus. Recapitalisation of banks, in particular, has been seen as crucial to the recovery of both the banking system and the economy. But where in the past the financial sector served a primarily wealth-preserving function, and where substantial wealth has now been destroyed, pushing more credit in the hope that this will generate growth could be counterproductive. This attempt to push more credit into the economy would be particularly harmful if growth potential is low. The risk is that zombie banks will be propped up, and the costs of cleaning them up and closing them down will only increase over time.
True conclusions, but a slightly misdirected search for the cause.
It seems to me that the key variable is the transfer of wealth from the bottom tiers of wealth holders to the upper tiers. All this analysis of credit to gdp ratio, and advanced economies versus developing economies, and pre 1990 versus post 1990, all point to the the possibility that transfer of wealth is the issue. Measuring the growth of the financial sector as this article does, we see scatter plots with dots all over the place. That means that there are huge exceptions to the rules being gleaned. As an electrical engineer, I would call this a large noise to signal ratio. If we plotted all of the growth data against the axis of wealth disparity we would get smaller noise and larger signal. Then we could talk about large signal to noise ratios, and fewer and smaller exceptions to the rules.