We demonstrate that high 1996 latent demand zip codes experience a dramatic relative reduction in mortgage denial rates during this time period. Simultaneous with the reduction in denial rates, high 1996 latent demand zip codes experience much larger increases in the debt to income ratio of accepted mortgage applications. The source of this increase in credit availability is disintermediation: the fraction of mortgages sold by originators in the secondary market experiences a sharp relative increase for high latent demand zip codes from 2001 to 2005. Finally, the interest spread between mortgages to low credit quality borrowers and high quality borrowers narrows to historical lows during this period. Taken together, these facts demonstrate a sharp relative increase in the supply of mortgages to high 1996 latent demand zip codes from 2001 to 2005.
They measure latent demand for mortgages as the percentage of mortgage applications turned down in that particular zip code in 1996. What they are saying is that in zip codes where lots of folks were turned down in 1996, you see lenders approving many more loans, and at lower risk spreads in 2001 through 2005, fueling the home price bubble. The lower risk spreads tells me that this was not predatory lending, but the opposite.
Keep this in mind when people say that better regulation could have prevented this problem. It sounds like what they are saying is that lenders charged exorbitant interest rates to hapless borrowers. In fact, lenders were guilty of charging borrowers too little for loans, as well as approving too many unqualified borrowers. If you think that alert regulators would have cracked down on lenders for providing too many home ownership opportunities to Americans, especially minorities, then you believe in a different political environment than what I remember.