FRM Fun 24, MBS Prepayments (P1 or P2)

[David Harper CFA FRM]

P1 or P2

Today Bloomberg published a brief piece "Mortgage Prepayment Rate Reaches Highest Level Since 2005." I like how it gives color to our FRM topic on prepayment as a key risk factor in an MBS; and I like how it illustrates that prepayment can have several causes (financial versus non-financial) and influences (e.g., so-called burnout effect).

In the FRM, prepayment is firstly a call option, held by the borrower (mortgagor), which attaches to the loan (embedded option) and creates the negative convexity at low yields (btw, does it also contribute to negative convexity at high yields?).

About this Bloomberg piece, I have two questions:

1. Is this statement (in the piece) true? "Prepayment speeds also reflect borrower defaults and debt retired in home sales, which increased in August to a two-year high as the housing market showed signs of recovery."
2. Is this statement (in the piece) true? "Mortgage bond investors monitor prepayment rates as they influence returns. Bondholders risk losses when buying debt for more than 100 cents on the dollar as the value can be erased when homeowners take out new mortgages too quickly to repay existing debt. With debt trading below face value, returns increase when repayments accelerate."

(fwiw, i am not asking because i am 100% or even 90% certain of the answers. Rather, these two sections merely "give me pause" and I will need to consult my references to satisfy my doubts)

 

[trabala38]

Hi David,

Here are my answers:

1. Yes: higher prepayment speeds means mortgagors are able to refinance quicker, before a default occur. We can deduce that borrower default rate decreases as a consequence. Also, by prepaying their debt, the obvious consequence is that debt is retired, and that, probably, the home owner (=mortgagor) is able to lock-in some positive equity due the home price appreciation. As home-owners have more equity, they can re-finance their home with a lower debt amount since they have more equity.

2. No. This seems erroneous to me.
The risk of losses is not directly linked to the bond price (i.e. above or below the par). Of course, as prepayment rates increases, the value of the bond/tranches will decrease since less interests will be "paid" to investor (although it is 100% clear to me which tranches are impacted the most by the prepayment. I mean by that, that I am not certain of "how" the principal prepayment is re-distributed to the different tranches). Prepayment rate is likely to be linked to interest decrease (as explained by the article) and if interest rates decrease , this should increase the bond price, although some price compression, especially for the bond trading fore more than 100 cents on the dollar, could occur due to the negative convexity at lower yields.

Cheers,

trabala38

 

[ShaktiRathore]

Regarding part 1..I agree with trabala almost..Prepayment speed when goes higher than borrowers of mortgages pay all the mortgage due to low interest rates. They the borrowers sell the real estate at high price and pay the mortgage from the proceedings and keep the remaining sum with them, they than refinance the mortgage to buy the house at lower interest rate. The debt in this way retired in home sales increases. In this way borrowers already pays the mortgage outstanding so that there is no default on their part. As home prices keeps on increasing the capacity to pay by the mortgage holders increases so that their default rate decreases but what will happen when home price decreases?Because as capacity to repay decreases of homeowners(mortgage holders) the prepayment speed decreases thus the borrowers default rate increases but also the debt retired in home sales also decreases. So there is a negative relation between the prepayment speed and borrowers default rate and debt retired in home sales.

Regarding part 2...Statement"Bondholders risk losses when buying debt for more than 100 cents on the dollar as the value can be erased when homeowners take out new mortgages too quickly to repay existing debt." is correct for me. Due to negative convexity at low yields(high prepayment rates) the mortgage bond value remains almost constant or increases very slightly so the overall gain is eroded as compared to when when the prepayment rates are relatively lower due to relatively high yields at which the bond price increases more relatively so bondholders fear is justified that they can suffer losses due to increase in prepayment speed. Statement "With debt trading below face value, returns increase when repayments accelerate" also seems correct. Because for debt trading below face value the price shows positive convexity so that as prepayment rates increases(yield decreases) there is increase in the price of the bond.So returns accelerate when prepayment rates increases.

thanks

 

[RandomDaniel]

Just to throw my ideas out:

1. Yes regarding the defaults and home sales driving prepayments higher, but this doesn't point to a market recovery if your prepayment rates are going up because tenants are being kicked out of their homes.

2. These bonds trading above face value is probably a result of the current low interest rate environment and would erode in value if paid back too quicky although I'm not too sure I understand the below / above face value context here.

 

[David Harper CFA FRM]

Thanks for your thoughts. Both of these statements gave me pause (I had my doubts on the first read) because:

1. I was unsure what the author meant by "prepayment speeds also reflect borrower defaults" since defaults are, by definition, not prepayments. Veronesi makes the basic point: "Note that prepayment risk is fundamentally different from default risk. While in the case of default on a security the bank does not receive its money back, in the case of prepayment the bank actually does receive its money back".
   
   However, Veronesi goes on to say something I had not considered (p 294 of FRM assigned Chapter 8): "Interestingly, homeowner's defaults actually have an impact on the timing of prepayments in the case of agency RMBS, because if a homeowner defaults on his or her mortgage payment, the agency must step in a pay back the mortgage amount to the RMBS investors, thereby triggering prepayment." ... I GUESS (?) this is what Fabozzi means by a sentence that previously vexed me:" at its simplest, principal is returned to investors when borrowers default on their loans." (I think this Fabozzi statement is too loose, shouldn't he clarify this to refer only to agency MBS?).
   In any case, add this to trabala's excellent, provocative point!
   
   
2. On reflection, I agree with Shakti and I think the second statement is true. What gave me pause was the last sentence: "With debt trading below face value, returns increase when repayments accelerate."
   
   But I temporarily forgot about the dynamics of the high-rate (right-hand) side of the MBS P/Y curve. Our (my!) preoccupation is with negative convexity at low yields, but I think Shakti's reference to the right-hand side as "positive convexity" is spot-on; i.e., from the investor perspective, who has the short position in the call option, disadvantageous negative convexity as low yield is matched by advantageous positive convexity at high yield. However, the causes are entirely different. At low yields, prepayments are largely finance-rational exercise of the option; at high yields, the additional positive convexity (and note: interestingly, the additional convexity at high yields, due to the embedded option, is on top of the positive convexity already in the underlying bond P/Y curve) is largely due to turnover that is financially sub-optimal; it does not make financial sense for the homeowner to exercise an out-of-the-money option, but it still can be exercised for non-financial reasons. I hope this was fun to think about for you, too