The Federal Reserve's recent report, showing large commercial bank passthrough holdings plunging by $63.5 billion to $367.3 billion from $430.8 billion in the week ending June 15, spurred a flurry of research reports to try and explain the numbers.
In a report released yesterday, Merrill Lynch analysts suggest that the bank selling was for the front end of the month, due to the standard PSA settlement being on June 13. Analysts added that the simplest explanation for the huge dip is that a large number of mortgages were actually sold into the market, making the drop consistent with the roughly $23 billion rise in dealer holdings that happened between May 18 and June 15. There is an inconsistency in commercial bank holdings declining more than $60 billion, while dealer inventories rose by merely $23 billion, which Merrill attributed to various factors including rolls, securities distribution, and even mega creation.
The significant bank holding drop has a number of adverse implications for the MBS market, making analysts negative on the sector. If the coupon most heavily sold was 5%, it may negatively impact both the 5 roll and the 5 fly. Although it could also benefit higher coupons, this could only be realized if the market backs up.
In a separate report, RBS Greenwich Capital analysts did not believe banks sold a large MBS portion recently, instead attributing the significant MBS holdings dip to reporting quirks or data reflecting staggered settlements of an up-in-coupon trade.
Since many large bank report their holdings to the Fed on a settlement basis, although the trade could have happened back in April or May, it may have only been reported in June at settlement. For example, a bank selling FNMA 5s for June and purchasing 5.5s for July would experience this kind of dip. This allows banks to take gains in 5s before the end of the quarter while lessening the size of their balance sheet without a big drop in MBS exposure.
