T5.Residential-Mortgage-Backed-Securities

[EIA]

Hi David,

I am going through BT Veronesi questions.

I came across this question on pg 10

105.3. Barry the analyst calculated the effective duration of a pass-through MBS as 5.3 years. His effective duration is based on re-pricing the MBS with a yield shock of 50 basis points; i.e., current yield plus and minus 50 bps. However, Barry's manager observes that Barry did not vary the prepayment (PSA) assumption when re-pricing under either the higher/lower yield scenarios. His manager argues that Barry should vary the PSA assumption as he varies the interest rate input. If Barry varies the PSA assumption as instructed by his manager, which of the following is true?
a) The accurate duration will be lower than 5.3 years
b) The accurate duration will be higher than 5.3 years
c) It does not matter, neither duration nor convexity will be impacted
d) Duration is approximately unchanged at 5.3 years but convexity will increase

In your solution, you selected A but I don't think this answer is correct. It would have selected C because Barry has calculated effective duration and the increase in PSA or decrease in PSA does not affect the effective duration calculation except if is affecting interest rate movement.

Please check Veronesi pg 300 section 8.3.3.

BR

EIA

 

[David Harper CFA FRM]

Hi BR,

This question is a deliberate attempt to query an understanding of Veronesi 8.3.3. The directional impact is maybe harder to follow (see http://forum.bionicturtle.com/threa...onvexity-of-pass-through-mbs.5256/#post-16876 )

... however, it's important to understand that PSA impacts effective duration. In fact, the need to use effective duration (which re-prices with rate shocks, as opposed to an analytical duration) to employ a varying PSA assumption in order to capture the negative convexity. Put another way, Veronesi uses effective duration because it is the only way to get a duration which incorporates PSA.

Without the PSA assumption change, like any bond, a lower rate of (R - y ) will gives a higher MBS bond price of (P + x). The point of 8.3.3 is: if the rate goes lower, prepayments will increase, so we use a higher PSA assumption, which lowers the price of the MBS, which alters the duration and the convexity (i.e., it is the change in PSA assumption that allows us to simulate the negative convexity of the MBS). I hope that helps to explain, thanks,