Fannie Mae去年12月一篇文章，清楚地解釋了美國Mortgage Rate的組成以及近20年歷經的幾個轉變，筆者覺得值得紀錄，讀完後衍生一些投資想法，也藉此記錄下來。
(1) 10年美債利率 +
(2) The primary-secondary mortgage spread +
(3) The secondary mortgage spread
10年美債殖利率在這個公式作為benchmark，而所謂的primary-secondary mortgage spread，則是30年房貸利率跟MBS的利差，根據文中的解釋，這個利差必須cover承作房貸的成本、二房會收取的保證費以及融資者(應該是商業銀行)所需的成本及利益，這邊的意思應該是指商業銀行承做房貸後，會把房貸出售給Fannie Mae或Freddie Mac這樣的機構(二房)，二房會將這些跟銀行買來的房貸以MBS的方式證券化再出售給其他投資人，並保證這些房貸如果違約，MBS投資人仍然可以收到還款。因此，對二房來說，房貸跟MBS中間的利差(=The primary-secondary mortgage spread)，就是做這個生意賺到的錢以及足以cover承擔房貸違約的風險貼水。
Most mortgages are securitized into an MBS for investors to purchase. The spread between the mortgage rate that is offered to borrowers and the rate on an MBS is known as the primary-secondary spread. This spread reflects the costs of originating a mortgage and includes servicing fees, guaranty fees, and other lender costs and profits.
一個值得注意的點，這一個房貸利率與MBS的利差，在Covid後的這五年處於近20-30年歷史高位，金融海嘯前平均大概只有0.5%，金融海嘯後到Covid爆發前大概1%，近五年大概1.1~1.2%，文中對於這個利差增加的原因說明原文如下： 金融海嘯後，承作房貸的成本變高，而且二房收取的保證費也增加，(筆者註： 銀行可能必須轉嫁承做的房貸成本以及二房的保證費給借款人，所以反映在房貸利率提高，因此也使The primary-secondary mortgage spread提高)。
This shift is due primarily to an increase in the costs of originating a mortgage after the Great Financial Crisis, as well as an increase in the guaranty fee charged by Fannie Mae and Freddie Mac.
接下來講到 The secondary mortgage spread 這個利差，原文如下，這個利差就是指MBS跟10年美債的利差，會有這個利差的原因，是因為MBS本身有(a)被提前還款風險以及(b)信用風險這兩個國債所沒有的風險，也就是(a)房貸借款人萬一中了樂透提前還掉房貸，導致投資人拿不到原本MBS預期要有的現金流的風險，以及(b)因為財務狀況變不好而無法繳房貸的風險。
The other component of the mortgage spread is the secondary spread, which is the difference between the rate on an MBS yield and the rate on a 10-year Treasury note. Relative to the 10-year Treasury note, MBS carry two additional risks:
(a) Prepayment risk: Unlike the 10-year Treasury, which has a guaranteed rate for the duration of the bond, mortgage borrowers may prepay before the end of their mortgage term to refinance, move, or pay off their mortgage early.
(b) Credit risk: Whereas the 10-year Treasury is considered a virtually risk-free investment, a mortgage borrower or the entity providing a guarantee to the MBS investor may defaultii.
文中一樣有探討這個MBS跟10年債利差的近30年趨勢，跟前面那個30年房貸與MBS利差一樣，Covid後的這五年平均值是近30年的高位，金融海嘯前次之，中間2010年代最低，而文中也清楚地解釋了原因，聯準會的QE/QT政策，包含了大量的MBS購入及賣出，在2010s年代，聯準會大量購買MBS，因為聯準會的MBS購買無須考量是否要有足夠的利差來彌補上述的兩個風險，因此利差縮小；反之，2022年後開始縮表，MBS的買家變成投資人，就會需要更多利差來彌補前述的兩個風險，因此QT時期的利差大於QE時期。
While changes in credit or prepayment risk may explain some of the volatility in the secondary spread during specific periods, the Federal Reserve's balance sheet actions are the main factor driving the difference in the secondary spread.
Following the Great Financial Crisis, the Federal Reserve began purchasing MBS and Treasury securities, a policy known as quantitative easing, to stabilize the financial system and drive interest rates, particularly mortgage rates, lower. The Fed is a non-economic buyer, which means it is not sensitive to the rate an MBS is paying. Therefore, when the Fed purchases MBS, it replaces private investors who are sensitive to the rate on MBS, which drives the secondary spread lower.
When the Fed does not purchase MBS to offset the ongoing runoff of MBS in its portfolio, a policy known as quantitative tightening, it gradually owns a smaller share of total MBS outstanding. As a result, private investors need to purchase that additional share, and as rate-sensitive buyers, they require a higher interest rate to do so. That's why the secondary spread is driven upward when private investors replace the Fed in the MBS market.
文章中是把房貸利率用下圖的方式建構，也就是本文一開始的公式，但是筆者覺得應該把(2)跟(3)互換一下位置，變成以下，其中(1)+(3)就是MBS利率，然後(2)就是MBS跟房貸的利差。
30年期Mortgage Rate =
(1) 10年美債利率 +
(3) The secondary mortgage spread +
(2) The primary-secondary mortgage spread
最後總結一下並補充一些思考，錯了歡迎指正。
(1) QE時代，因為聯準會大量買入MBS，secondary mortgage spread被扭曲式壓低；QT時代則是相反(聯準會賣MBS)，secondary mortgage spread被扭曲式拉高。
(2) The primary-secondary mortgage spread則是二房做生意賺的利差，近年處於高位。
(3) 萬一聯準會的下一步是買國債但是賣MBS，那會怎樣? The secondary mortgage spread仍會高位，因為聯準會是MBS賣方，但卻是10年美債買方，所以MBS利率下降的幅度會低於10年國債利率下降的幅度，房市不會過熱，財政利息支出仍可緩解。
(4) 如果(3)發生，對銀行保險等金融機構來說，資金成本降低(假設一定程度定錨10年國債收益率)但投資標的回報率(MBS)不變或增加，是偏向有益的環境。
(5) 但(3)若發生，對二房來說是福是禍? 筆者之後會再思考後留下紀錄。
