Sufi emphasized that the debt run-up resulted from a positive credit supply shock—not a sudden demand for more credit. Credit supply shocks occur when lenders decide to offer more loans based on “reasons unrelated to actual underlying performance of companies or the income of households.” Because these spikes occur when interest rates are low, Sufi rules out increased credit demand as a factor. If lenders were reluctant to expand credit supply, interest rates would have risen.
Measures of savings tend to show low savings rates at the time. Capital gains aren't included in savings rates, though. In the graph here, home values are inverted. Notice that savings moves up when home prices move down and vice versa. This is normally treated as a source of unsustainable consumption, but in this case, that is not necessarily true. (1) The gains were largely gains in Closed Access real estate, which are permanent gains until a sea change happens in urban governance, and are certainly permanent for households that sell and realize capital gains. (2) During this time, there appears to have been a significant amount of harvesting of real estate capital gains which were then transferred into savings instead of into consumption. This combination of factors explains why interest rates were low while savings was low and credit levels were spiking. Sufi is right that credit demand wasn't particularly high. This is further confirmed by the declining rate of homeownership at the height of the housing boom.
In a way, I think Sufi's comment here gets halfway to a contrarian point of view regarding credit and business cycles. There is an element of Scott Sumner's reasoning from a price change here. Let's call it reasoning from a credit quantity. As Sufi notes, credit levels could rise because of rising demand or rising supply. The rise in debt leading to a financial downturn is usually presented as if it is from demand. But, this seems wrong to me. Usually, it is associated with a rise in mortgage debt. In this recent case, mortgage debt was clearly associated with two factors:
1) Closed Access home values, which reflect an arbitrary limit on productivity, and thus would be associated with declining economic expectations.
2) Low real interest rates, which are usually also associated with declining sentiment.
Yet, these trends in debt are usually treated as bouts of risk taking and exuberance that are unsustainable - inevitably leading to a bust. Shouldn't we treat these periods of rising debt levels as the first signs of a downturn in sentiment, not as naïve exuberance?