Significant mega pool creation seen in the past few months has reduced the TBA deliverable supply, creating short squeezes in the sector.
In May, $27 billion in new FNMA 5 Megas were created. This followed four consecutive months of strong Mega creation in the coupon that altogether reached close to $35 billion. In fact, Merrill Lynch analysts reported that production in FNMA 5s could not keep pace even after the $9 billion issued so far in May, and $11 billion in April. Merrill researchers believe originators will eventually be able to make enough FNMA 5s for the coupon to underperform, adding that it is starting to make sense for originators to produce high coupon 5s while retaining excess servicing.
Art Frank, head of mortgage research at Nomura Securities, said banks that have large positions in specific coupons would prefer holding them in Mega pools rather than owning hundreds of smaller pools bought in TBA trades "for the convenience and the lower custody costs."
However, the increase in Mega pool creation for FNMA 5 - the predominant coupon in these pools - has taken billions of dollars of deliverable bonds away from the TBA market, creating short squeezes in TBAs. Increasing mega pool supply also has the 5% coupon rolling with less floating supply for this coupon. With more of these securities found in Megas, "TBA FNMA 5s were bid higher as the shorts were squeezed," Frank added.
Separately, Lehman Brothers researchers reported that of the entire post-2003 FNMA 5 production, almost 80% is in Mega pools or CMOs. The post-2003 pools total $110 billion, $11 billion of which are in CMOs while $72 billion are in mega pools. To the extent that investors not rolling their bonds are holding these Mega pools, just $27 billion of post-2003 TBA floaters remain in the coupon. A significant percentage of the coupon is sitting at the 20 WALA bucket. At 20 WALA, Lehman analysts said the cumulative free float is roughly $67 billion, and there is no reason for the deliverable to extend beyond 20 WALA of seasoning. If the new Mega pools created remain out of float, Lehman expects there to be almost negligible payups on seasoned 5s, although Lehman believes the coupon remains overvalued. The float issues in FNMA 5s have been due mostly to the strong Mega pool issuance in the coupon.
However, Mega pool creation has not been limited to the FNMA 5% coupon alone. From 1996 through early 2001, 30-year conventional Mega pools were being created at a roughly $5 billion monthly rate. By 2005, this rate had quadrupled to almost $20 billion per month. Prior to 2001, gross issuance was, not surprisingly, much less. Lehman's data showed that in 2005, Mega pools have totaled almost 80% of gross issuance, compared with 20% to 40% from 1996 to 2003.
Mega pool creation of this magnitude has advantages and disadvantages. For instance, when investors not rolling their bonds hold these Mega pools liquidity becomes an issue, analysts said, which at the margin, could also result in mortgage basis widening. On the other hand, a larger Mega pool share brings diversification benefits. Additionally, a Mega pool's prepayment volatility is far lower than individual pools making up the Mega. This improved convexity is, in turn, positive for the TBA market and should cause tightening.
The homogeneity arising from an increased Mega share limits uncertainty over TBA deliverable characteristics, Lehman analysts said. Lehman believes the increased homogeneity is positive for those who view the TBA market primarily as a way to add mortgage exposure or to hedge mortgage-specific risks in their portfolios. However, with alpha generation sources currently few and far between, increasing homogeneity is a negative for mortgage portfolios, added Lehman analysts. Those hoping to generate returns betting on TBA deliverable changes, TBA financing levels or by cherry-picking pools from TBA deliveries, will find these opportunities increasingly scarce with rising homogeneity.
(c) 2005 Asset Securitization Report and SourceMedia, Inc. All Rights Reserved.