© 2024 Arizent. All rights reserved.

Barclays examines the causes of fast GNMA discount speeds

Baclays Capital MBS analysts used the new delinquency information released by Ginnie Mae to examine the connection between delinquencies and prepayment speeds.

Analysts said in a recent report that although the new data has brought to light the direct impact of delinquencies on speeds, it has not resolved the long-standing controversy over why discount GNMA speeds are fast.

"The new disclosures reveal that foreclosures and delinquency buyouts have played a minor role in discount speeds over the two months for which data is available [data were released in December and January, reflecting prepayments in October and November]," Barclays analysts wrote. "At the same time, they also indicate exceptionally high delinquency rates that at first glance seem at odds with the buyout/foreclosure data."

The delinquency factor

The Barclays report looks at correlations between delinquencies and speeds in analysts' search for evidence that pools with high delinquencies prepay swiftly, even with foreclosure and buyout contributions being small. Though analysts found a weak connection between delinquencies and speeds, they found a strong correlation between home price appreciation and GNMA speeds.

Analysts explained that fast discount speeds on GNMAs could be explained by two theories.

The first is the "equity growth" hypothesis, which says that home price appreciation gives FHA homeowners the chance to bypass mortgage insurance premiums, thus causing significant FHA-to-conventional refinancings on mortgages that would otherwise be out of the money. The second theory - dubbed the "subprime borrower" hypothesis - says GNMA borrowers have deteriorated to subprime levels, and that fast speeds are a natural result of high defaults. Barclays analysts said that although both mechanisms should contribute to speeds in principle, they are leaning toward the equity growth mechanism.

"The answer is vital to the fundamental valuation of GNMAs because the equity growth hypothesis implies an extreme vulnerability to a housing slowdown or a sharp sell-off in rates, whereas the subprime borrower hypothesis implies relative immunity to either," analysts wrote.

They added that although some investors think fundamental valuations have little bearing on GNMA prices since overseas demand has resulted in a technically driven market, Barclays analysts believe that exposure to a U.S. housing slowdown is of considerable concern to overseas investors and that prepay characteristics are an important element of long-term performance.

Analyzing each theory

To examine the factors driving speeds, analysts looked at the last two months of pool level disclosures provided by Ginnie Mae. Their analysis involved dividing all the GNMA 5 and 5.5 outstanding pools for 2003 to 2005 vintages into groups based upon levels of delinquencies and home price appreciation. They also aggregated pools into 1% buckets by their 30-plus-day delinquency rate and observed how three-month CPRs depended on the levels of delinquencies.

Based on the subprime theory, analysts expected pool speeds to show a strong dependence on the level of delinquencies. The results of the study seemed clearly at odds with the subprime hypothesis. The data presented shows no apparent signs that pools having elevated levels of delinquencies generated higher levels of prepayments.

The results of the analysis are strongly supportive of the equity growth hypothesis as home price growth seems to be the dominant factor behind fast discount speeds, with analysts writing that, "we find the numerical dependence of GNMA pool speeds relative to home price appreciation significant, and capable of explaining the speed differential between discount Ginnie Mae and Fannie Mae collateral."

According to the data they presented, every 1% increase in home price appreciation leads to an approximate 1% rise in GNMA pools speeds.

Analysts stated that although these results are only based on two months of disclosure data from Ginnie Mae, the analysis seems conclusive. GNMA's pool level delinquency disclosures show that buyouts and foreclosures only explain 0.5 to 1.5 CPR on 5 and 5.5 2003 cohorts. Aside from this, analysts have shown that high levels of delinquencies are statistically unassociated with higher levels of prepayments, further disproving the subprime hypothesis. Analysts expect that, as more GNMA delinquency disclosures become available, the data would be more strongly supportive of the equity growth theory. Regardless of market perception, their analysis suggests that buysiders relying on delinquency-boosted speeds in a sell-off might find themselves with more extension risk then they had expected.

"With home price appreciation expected to be moderate during the next year, it logically follows that GNMA discount speeds should slow in sympathy," analysts wrote. " A sell-off would compound the problem."

Though market technicals are the primary element driving GNMA prices, analysts said the outsized exposure of discount GNMAs to a slowdown in housing as well as higher rates could result in a long-term market repricing.

(c) 2006 Asset Securitization Report and SourceMedia, Inc. All Rights Reserved.

http://www.asreport.com http://www.sourcemedia.com

For reprint and licensing requests for this article, click here.
MORE FROM ASSET SECURITIZATION REPORT