It is almost a given that the mortgage market will be smaller this year than last so discussions have turned to the availability of mortgage credit and whether it is too tight.  Mark Fleming, chief economist at CoreLogic, says that the question of availability has taken on more importance as the Qualified Mortgage (QM) standard is implemented and the industry has to determine the types of credit it will offer to borrowers.  He presents a different take on the subject in the current issue of the company's Marketpulse. 

Whether credit is too tight or too loose, he says, is an especially difficult question to answer because there is no single measure of its availability.  However there are a variety of metrics that collectively influence that access; credit worthiness, loan-to-value and debt-to-income ratios, the level of documentation, the availability of adjustable rate loans, and the share of purchase loans.

Fleming set out to measure the current limits on these measures by determining what constitutes a "normal" level of availability for each metric.  To do this he settled on the year preceding an announcement by the Federal Reserve in February 2004 encouraging the use of adjustable rate mortgages and its subsequent raising of the federal funds rate as baseline period for his demonstration and looked at averages for the metrics during that period.  For example, over that year the average credit score for all first lien purchase loans was 710.  In October 2013 it was 749.

While this is only a 5 percent difference, he points out that the share of loans with credit scores below 620 was 29 percent over the baseline period and was 0.3 percent in October 2013.  "It is much more insightful to compare the share of the riskiest subset of the entire measure's distribution to that same share before the housing bubble," he says.  "Credit availability in each metric represents the extent to which lenders originate loans to the riskier subset of the distribution."

Fleming constructed the figure below with each axis representing a different measure of credit availability and the inner (blue) hexagon crossing each axis at the value of 100 which represents the normal value for that metric based on the average during the year preceding the Fed's 2004 announcement.  The outer (green) hexagon represents the deviation from normal at the loosest point during the housing bubble.  Red represents the conditions in October 2013.


Going back to the example of credit scores, the share of first lien purchase mortgages with scores below 620 was 29 percent during that year and that is indexed to 100.  The October 2013 origination of mortgages with scores that low was virtually zero (0.3 percent). 

Fleming says that what is immediately apparent is the credit availability is tight for two important underwriting criteria, credit scores and low documentation levels, neither of which are being originated relative to either the height of the expansion of credit or the normalized level of availability.  Underwriting eligibility in the current market requires both good credit and the ability to fully document loans.  However the shares of high LTV and high-DTI lender are currently close to normal.  Both had expanded availability during the housing boom and high LTV lending remains modestly loose relative to normal while high DTI lending is modestly tight.  ARM loans are much more restricted than normal "as many subprime ARM loan products are no longer available."

Fleming says that looking at the riskiest subset of an entire measure's distribution and comparing it to the share in a more typical time gives more insight into whether credit is too loose or too tight.  Currently credit is tight for low credit score borrowers, those who can't or don't want to document their loans, and those who want an ARM product.  "For many, the choice to document or select an ARM product is not necessarily an impediment to credit availability," he says.  "However, for those with low credit scores there are fewer options."