Where Economists Meet Since 1968
The State of the US Residential Mortgage Market
A Summary of Remarks by
Senior Managing Director, Amherst Securities
April 22, 2010
Laurie Goodman opened her presentation with a summary of the positive aspects of the housing market: low prices and low mortgage rates have led to levels of affordability not seen in decades, but argues that when you dig below the surface, there is a huge overhang of delinquencies and foreclosures that are a risk to the market. In her talk, she will discuss her estimates as well as review the mortgage modification process, examining what is and is not effective.
Ms. Goodman makes the case, if stronger policy measures are not taken, an additional 12 million units could be liquidated. This is more than one out of every 5 mortgaged homes. The 12 million units include properties currently in some stage of the foreclosure process, properties with delinquent mortgages that are unlikely to be cured, and properties owned by underwater borrowers who are at high risk of default and have a low likelihood of being cured.
The size of the problem –Loans that are Already delinquent
There are 56 million housing units in the US with a mortgage. Using MBA default/delinquency rates and Amherst calculated liquidation probabilities, she estimated that 7.18 million of the 7.85 million properties in delinquency or default will ultimately be liquidated. Mortgage modifications may reduce the number of properties to be liquidated somewhat, but redefault rates on modified mortgages are very high—a high estimate for mortgages cured by modification would be 1.1 million units.
The driver of the housing overhang is rising delinquencies and falling cure rates on the delinquent loans. Ms. Goodman showed evidence that borrowers with higher combined loan to value ratios are more likely to be in delinquency. She argued that equity is the single most important determinant of delinquency/default and correspondingly, successful modifications will require principal reductions. She pointed out that equity is important for all types of loans, in fact, prime borrowers without equity seem to transition at a higher rate from a 30 day delinquency to a 60+ delinquency situation than do subprime borrowers without equity, perhaps as a result of strategic defaults.
Ms. Goodman also noted that the period of time from delinquency to foreclosure or REO sale has risen from about 15 months in March 2008 to 19 months in two years. This varies from state to state. Judicial states are much slower than non-judicial state and owner-occupied properties tend to liquidate more slowly than investor properties. In total, among recently liquidated properties, 22 percent were in the process, for more than 2 years. There was a question about the variance among servicers. Ms. Goodman reported that while the time to liquidation may vary, the slowdown in liquidation is pervasive among all servicers, in part due to modification rules and in some cases judicial backlog.
The result of the above lags is that the non-performing-to-foreclosure and foreclosure-to-REO roll rates are down at the same time that REO-to-liquidation roll rate is higher. This means that the inventory of REO properties is low—under half of what is was at the peak—but the mass of properties to worry about is a step further back in the process and has swelled by a greater amount than the REO properties have declined. This is the overhang that Ms. Goodman argues many have missed.
The problem in the mortgage market as a whole
The cure rates in the above analysis were taken from the private label securitized world which is admittedly an adversely selected group, but some of the lessons hold when examining the entire mortgage market. In the examination of the entire market, Ms. Goodman focused on default transitions and prepayment speeds. A prepayment is good for lenders because it removes loans out of the pool, eliminating default risk. However, because prepayments are coming disproportionally from loans with equity and defaults come from loans without equity, the remaining pool of loans is worsened over time.
The expanded market examination covered non-private-label securitized loans (non-PLS), which includes Fannie, Freddie, FHA, VA, and bank-held loans (see slide 13). Ms. Goodman categorized loans that were 60+ days delinquent as nonperforming and those that were 60+ days delinquent and are now performing or 30 days delinquent as reperforming loans. She then examined the rate at which reperforming loans and always performing loans with various loan to value ratios transition into default or prepayment. Again, she found that equity matters, default rates are higher and prepayment rates are lower for loans with higher loan to value ratios.
Taking what she learned from this examination Ms. Goodman arrives at the estimate of 12 million properties headed for liquidation by assuming that all non performing loans are liquidated, 90 percent of reperforming loans are liquidated, 75 percent of performing loans with LTV > 120 are liquidated, 25 percent performing loans with LTV =100 – 120 are liquidated, and none of the performing loans with LTV <= 100 are liquidated. These liquidation rates are somewhat more generous than strictly applying the results of her analysis of the market which suggest a liquidation total of about 21 million properties. The estimated 12 million properties on the way to liquidation represent slightly more than one of every five homes with a mortgage.
What causes default? Unemployment vs zero equity
Ms. Goodman briefly examined the trigger for default—looking at unemployment and equity. She argued that negative equity is the primary driver and unemployment is a trigger, but unemployment alone is not sufficient to cause a default.
Credit availability and housing demand
Ms. Goodman then transitioned into credit availability, calling it a necessary component for fixing the demand-side of the market just as mortgage modification is a necessary component for fixing the supply-side of the market. Foreclosures have limited the buyer base, and the tax credit has pulled a lot of demand forward. Ms. Goodman showed that in the housing boom, mortgage backed securities issuance was not coming from Fannie/Freddie, but rather from securities issued by Alt-A and Subprime lenders. Toward the end of 2007 the mortgage markets began to shut and mortgage backed securities issuance shifted back to agency lenders. In 2008 and 2009 agency lending was more than 95 percent of the market. There is almost no securitization in the private label residential mortgage market currently, only resecuritization of existing deals.
Without securitization, there is still some mortgage origination, but at a much lower level than in previous years. For example, 2009 prime jumbo originations are estimated to be $92 billion compared to $570 billion in 2005. The net result is bifurcated housing markets where we see much larger inventory in the high-end and other non-government backed sectors of the market. Further, as a result of tightening credit standards by the GSEs, there are few outlets for borrowing by borrowers with lower credit. FHA is the only source of credit for these borrowers. However, these low-credit buyers are pushing up FHA delinquency rates. Ms. Goodman also argues that the limited availability of mortgage credit has pulled down the MBA purchase applications index.
Mortgage modifications and the supply of distressed property
As for the supply-side, modifications in the non-agency market—a $1.5 trillion market—are not working as well as they could. Reperforming loans redefaulted at a rate of about 8.5 percent last month. While there are variations by loan type, the redefault rate for reperforming loans is 8 – 10 percent regardless of loan type or vintage.
What do redefault rates and the HAMP program mean for modification success? HAMP modifications have reduced front-end debt to income ratios from 45 percent to 31 percent, but have not reduced overall loan balances, and therefore do not help the negative equity problem, Previous modifications that were “HAMP-like,” those that had a reduction in front-end debt-to-income ratios from 30-40%, redefaulted at a rate of 65 percent in the year after modification. Other evidence in favor of principal reduction is the observation that when banks hold loans in their portfolio, they tend to do principal reduction, and bank portfolio loans redefault at a lower rate. It would be ideal to see a relationship between bank portfolio loans that had received a principal reduction, but that information is not available.
How big would the benefit be if principal reductions were to happen? Looking at outcomes among similar borrowers with full and low documentation, reducing combined loan to value ratios from 140-150 to 110-120 leads to a reduction in transition rate from 83 to 55 percent. Ms. Goodman acknowledges that principal reductions will not eliminate the problem, but can go a long way toward mitigating it.
One of the biggest obstacles to mortgage modifications is the prevalence of second liens and the mismatch of incentives between servicers and holders of first liens. Second liens are prevalent across loan vintages; in the private label securities universe the percent of seconds range from 40 to 57 percent across vintages. Second liens are of two types: approximately half were simultaneously obtained with the original loan; the remainder are “subsequent”, that is, taken out later. . Notably, prime borrowers are the most likely to have second liens whereas subprime borrowers are the least likely. The problem with second liens is that they raise the combined loan to value ratio which increases the likelihood of default on both the first and second liens.
Another problem with second liens, is the fact that the largest mortgage servicers—the top 4 commercial banks service 56 percent of loans—also own second liens—$440 billion of $1 trillion overall—and therefore are disincentivized to modify loans by reducing principal. Requiring write-downs on second liens is critical, Ms. Goodman argues.
In summary, Ms. Goodman argues that the upcoming liquidation of 12 million properties will have consequences for the housing market. She has two proposals to clean it up 1) do successful modifications—those that reduce the principal balance, 2) stimulate demand to clear off the inventory of liquidated properties by setting up an FHA-type program for investor buyers.
Laurie Goodman is Senior Managing Director at Amherst Securities where she is responsible for strategy and business development. Previously she has worked for UBS, Citicorp, Goldman Sachs, Merrill Lynch, and the New York Fed. She earned an undergraduate degree in mathematics from the University of Pennsylvania and her Ph.D. in Economics from Stanford University. Last year she was selected to the fixed income analyst hall of fame.
Rapporteur: Danielle Hale
National Economists Club
P.O. Box 19281
Washington, DC 20036