However, doesn' the trend (red line) suggest something about the changing effect of debt on GDP? If not, can we think of a better way to get at a "marginal productivity of debt"? It feels like if we can find a way to calculate/graph it, it could be a powerful tool of analysis...

However, that particular chart does bear further examination. I'm just not sure there's a causal relation or a productivity to be found in it...

The brainless should not be in banking -- Willem Buiter