Since 2004, any mortgage having an annual percentage interest rate exceeding the rate on treasury securities of comparable maturity by a threshold (3% for first lien and 5% for subordinate lien) must be reported in the lender's annual Home Mortgage Disclosure Act (HMDA) data submission. All reported mortgages having annual percentage rates below the HMDA rate spread threshold will not include a rate spread value in the lender's annual HMDA data submission.
The HMDA data was examined from 2007 back to 2004 to determine the proportion of mortgages showing a reportable rate spread value (higher cost loans) to mortgages not having the rate spread value (lower cost loans). The regulators desired having rate spread values that exceeded a threshold because they believed that this enabled them to have greater monitoring of the sub-prime segment of the mortgage market while removing the reporting burden from mortgages believed to be seperate from the sub-prime segment of the mortgage market.
In the examination of the data for this post, we only included first lien conventional purchase and first lien conventional refinance mortgages on owner occupied 1-4 family dwellings. Normally lenders charged higher interest rates on non-owner occupied home loans and lenders also charged higher interest rates on manufactured housing. Mortgages with dollar conforming amounts were compared to each other and mortgages having jumbo dollar amounts were also compared. Lenders typically charge higher rates for mortgages exceeded the conforming loan limits. HMDA data provides very few data items to help compare mortgage loans of similar risks to investors but the criteria used to compare mortgages in this study help show relations of mortgages with similar risk factors.
The chart shows that in 2007 there was a substantial reduction in the proportion of rate spread reportable mortgages. From 2004 to 2006 there were rapidly rising proportions of mortgages having HMDA reportable rate spreads.
One surprise (at least too me) was that the proportion of rate spread reportable mortgages was always higher for the mortgages having loan amounts within the conforming loan limits as compared to the proportion of jumbo mortgages having rate spread reportable mortgages. We assumed that jumbo mortgage amounts added risks and typically carried higher rates. Given that jumbo mortgages carried higher rates, we might expect that to flow to higher rates of rate spread reportable mortgages for the jumbo segment. Not so!