The Bond Market Association recently held a conference call with MBS dealers and originators to discuss the creation of a hybrid ARM TBA program. After previous attempts have proved unsuccessful, analysts from dealerships said that investors would benefit from the improved price transparency and ability to purchase ARMs resulting from the development of a TBA market, while others are not hopeful that a liquid and reliable program could actually come to fruition.
"From the point of view of traders that need to hedge, a liquid TBA contract would serve as an ideal hedging vehicle," analysts from Countrywide Securities said. "However, we are not optimistic that a TBA product can be created with the necessary liquidity and reliability to make it useful to hedgers, i.e., to make traders comfortable in shorting it."
Countrywide analysts noted that hybrid ARM production has gone through some rapid changes recently and the product is now made up of several variables including duration of the fixed-rate period, the index utilized, the cap structure, etc. Thus, determining which of these attributes should be considered deliverable will necessarily exclude a significant amount of the production from TBA trades, potentially hurting the sector's liquidity. This is particularly true when it comes to amortization, considering that much of the production is currently in IO loans.
Pushing a TBA ARM product is partly aimed at providing originators with superior execution while causing lenders to scale back production of non-deliverable collateral in favor of loans that can be delivered. This implies that the TBA would need to trade considerably richer versus specified hybrid pools to make non-deliverable products uneconomical. However, Countrywide does not expect this to actually happen as many ARM products are too popular with borrowers to simply disappear. Also, analysts believe that eliminating products is contrary to the overreaching trend of offering consumers more alternatives. Although execution is still a major driver of production dynamics, consumer tastes and needs are currently playing a bigger role than they have in the past. Additionally, analysts said creating a specific TBA hybrid ARM is also against the secondary market trend of trading on favorable loan-level attributes.
Countrywide noted the possibility of a market squeeze resulting from the lack of liquidity in TBA ARMs. For instance, if production of TBA deliverable ARM pools is sufficient to promote liquidity for the TBA after being introduced, traders may start using it as a hedge. However, a production shift could result in deliverable pools not being made, causing liquidity to dry up and traders shorting the TBA to get squeezed.
Analysts cited the case of Ginnie Mae ARMs in the mid-1990s that were often shorted as a hedge. A later drop in production caused traders that were short to be squeezed, while traders exploiting the squeeze made the situation worse. Squeezed coupons thus started trading two to three points in price over fair value and traders short a TBA for a particular month sometimes covered their positions by hunting down pools issued that month The large losses resulting from extreme short squeezes caused some dealers to get out of MBS. Another case cited was that of the Fannie Mae L.A. program, an ARM program that had specific pooling requirements involving the creation of fixed-coupon pools in quarter coupon increments. Fannie Mae L.A. production - which peaked at 5% of total Fannie ARMs in April 2004 - has remained at one to two percent of production since August and is used by just two large issuers.
"Therefore while we understand and appreciate the desire to create a successful TBA market, we don't have a great deal of optimism that the current effort will be successful, as the hybrid product does not lend itself to the degree of standardization necessary," Countywide analysts wrote.
Deutsche Bank Securities MBS analyst Victoria Averbukh in a recent report said that it is important to create a TBA program allowing enough flexibility in TBA pools so as not to disrupt existing market practices. Averbukh noted that Deutsche Bank supports the creation of a broadly defined contract, while others favor tighter delivery stipulations, adding that a narrowly defined contract would probably not be viable as both originators and the dealer community would not be able to commit going forward.
Deutsche Bank also agrees with the focus on 5/1 hybrids - currently making up at least 50% of the issuance - has a better chance of being produced in a flatter curve scenario, adding that 5/1s should have the 5/2/5 cap structure as 5% initial caps have become more popular in the past two years, a trend likely to continue.
Averbukh also said that coupon variance among pools issued in each particular month is an important aspect of any ARM TBA market and limiting the deliverable coupons too much by only allowing fixed half-coupons for delivery, for example, would likely hurt the market's feasibility, agreeing that quarter-coupons are a better solution. However, either constraint would need originators to commit to a specific amount of loan issuance into TBA pools. "Originators should commit to originate enough TBA in any given month for the program to be successful," notes Averbukh. "That's why originators are very important to the process." Averbukh said that if originators find fixed coupons too constricting, a solution is to allow for variable coupons and adjust prices based on some type of conversion factor, something like a standardized U.S. Treasury futures contract. "This would create a broadly defined program while mitigating the risk of the hybrid ARM TBA market deterioration into the worst-to-deliver scenario," Averbukh said.
Other analysts disagree with the idea of quarter coupons for TBA delivery, as they believe this might result in the creation of smaller coupon buckets. "This would dilute the TBA standard of creating liquidity," said another MBS analyst.
In the report, Averbukh suggests that the underlying index used for this product should be the one-year Libor, mentioning that about 70% to 80% of Fannie Mae 5/1s and 50% to 60% of Freddie Mac 5/1s are indexed off one-year Libor and not one-year CMT. Averbukh also said that seasoned loans with more than a six-month WALA should not be allowed into pools while a six-month dispersion between resets on individual loans should be the maximum allowed in a pool.
(c) 2005 Asset Securitization Report and SourceMedia, Inc. All Rights Reserved.