This is a very simple chart. It takes the change in GDP and divides it by the change in Debt. What it shows is how much productivity is gained by infusing $1 of debt into our debt backed money system.

Described that way, I am not sure that chart reflects a causal relation (that is, a "marginal productivity of debt").

Then a funny thing happened along the way. Macroeconomic DEBT SATURATION occurred causing a phase transition with our debt relationship. This is because total income can no longer support total debt. In the third quarter of 2009 each dollar of debt added produced NEGATIVE 15 cents of productivity, and at the end of 2009, each dollar of new debt now SUBTRACTS 45 cents from GDP!

That is what I don't think follows from the data as presented.

What has happened recently is that US GDP has contracted while debt has expanded. The interesting thing is that in all previous recessions economic contraction was accompanied by deleveraging, given that the ratio of increments of GDP and debt was always positive.

This time we have increasing leverage while the economy slows down: that is, the system as a whole is refusing to deleverage in a recession. If the economy were reducing its stock of debt, the ration of increments of GDP to debt would be positive.

From a data presentation point of view, I'd PN that the vertical axis should not be labelled in dollars, but without units, since it is "per $1 debt".

The brainless should not be in banking -- Willem Buiter