“Principal-only (PO) strips have negative key rate durations in the short and intermediate rates, which turn positive for longer (e.g., 10 year) rates. Interest-only (IO) strips start out with positive key rate durations which turn negative.” Can anyone help to explain why the key rate duration goes from negative to positive in PO, and vice versa in IO? Thanks!
my study pal for L3 is a bond PM at Pimco. on the weekend we spent hours with a white board going through all the ins and outs of P/Os and I/Os. But here’s the deal: IOs act like Bonds - ie similar duration patterns at the medium long end - ie rise in value when rates fall, etc - just like bonds. (Key: IO = Interest. Bonds pay Interest. They act like Bonds.) BUT at the short end they switch and do the reverse - ie value rise when rates rise) PO’s are the opposite of IO’s --> ie fall in value when rates fall. (Key: PO’s are opposite of IO’s) - But at the short end they also switch and do the reverse. So: IOs act similar to bonds (but switch at the short end). POs are the mirror opposite of IOs, and also switch at the short end If you ask me why, my brain will start to hurt. The net result of all this is that duration strategies will not work - you need to use key rate durations to cater for the different behaviour across the whole spectrum.
I believe schweser said this part is optional…isnt it? or am i mistaken? pls. confirm
LOS 32©, (d) and (e) are all about mortgage securites - yield curve risks - different behaviour relative to straight treasuries - cfai text goes into MBSs, PO’s, IO’s and their particular yield curve behaviours, use of key rate duration, 2-bond hedges, etc. the theme is to undertand the different behaviour of mortgage-backed securites - how the behave, how to hedge, etc I don’t use schweser so I can’t comment on that…
null&nuller, I think your explanation is just the opposite of Corrupted’s comment. Which one is correct??
yuhuang - you had me worried - you are right of course. Let me correct my first post: POs act like Bonds - ie similar duration patterns at the long end - ie rise in value when rates fall, etc - just like bonds. BUT at the short end they switch and do the reverse - ie value rise when rates rise) IO’s are the opposite of PO’s --> ie fall in value when rates fall. But at the short end they also switch and do the reverse. (had to check LOS 59(j) from L2 last year!) a thousand appologies…
I remember why IOs and POs act the way they do not sure about the short end POs is basically like a Tresury bond. you buy it at discount and pays the whole principal at maturity OR if called. Because decreases in interest rates lead to prepayments, there is a higer chance that POs will be called at par(loan will be prepaid) IOs - you buy a future cash flow. If the principal is prepaid - you are screwed. That is why as rates go down, the value of IOs go down. On the short end I think nobody would refinance because of the additional costs so for a short period to maturity IOs and POs loose their characteristics
Null&Nuller, you had it right the first time and it looks like Schweser is ass-backwards on these: PO’s respond VERY positively to decreasing interest rates on the short end because prepayments dramatically accelerate the repayment of principal. If the entire discount from par is repaid immediately, you win. IO’s are nothing but interest payments. The ideal situation for an IO holder is that the bond is NOT repaid at any point, and they end up receiving interest for the entire length of the loan. There is now principal involved, so the longer you receive interest the better. If the loan is repaid immediately, you get nothing. This is why IO’s move WITH interest rates at levels below the coupon rate. Once interest rates are sufficiently high, IO’s start to act like normal annuities (positive duration) because the chance of prepayment fall. From Fabozzi “CMO’s and Stripped MBS”: “If mortgage rates decline below the coupon rate, prepayments are expected to accelerate. This results in a deterioration of the expected CF for an IO. The net effect is typically a decline in the price of an IO.” “Thus we see an interesting characteristic of an IO: It’s price tends to move in the same direction as the change in mortgage rates. This effect occurs 1) when mtg rates fall below the coupon rate and 2) for some range of mtg rates above the coupon rate.” “When mtg rates decline below the coupon rate, prepayments are expected to speed up, accelerating payments to the PO holder. The result is that the price of a PO will increase when mtg rages decline.” “When mtg rates rise above the coupon rate, prepayments are expected to slow down. Coupled with a higher discount rate, the price of a PO will fall when mtg rates rise”.
Like flo-pop said…PO overall has positive duration. If rates fall, people re-finance and pay off their original mortgate. So, you get your money faster, which is worth more (time value of $). So rates fall, prices rise. My guess as to why it’s negative at the short end, is people are unlikely to re-fi within a short amount of time of taking out the mortgage in the first place. IOs are the opposite, negative duration. Rates fall and people re-fi, so you won’t be getting your interest payments as long as you would have…or at least as much.
I see… I was taking “short and intermediate rates” to mean low to moderate interest rates. I guess if they’re talking about short to intermediate maturity this makes some sense.
Great explanation McLeod81. Now it starts to make more sense. Thanks
And also, short to intermediate term interest rates don’t drive mortgage prepayments. Intermediate to long term rates do.
If you think it terms of prepayment risk: PO benefit from contraction risk (principal paid back as far as possible). This will happen from refi’s at lower rates. IO benefit from extension risk because you will receive interest for a longer period than was priced into the instrument.
I don’t recall seeing anything is Schweser about this. Is this required?
it’s in schweser, and los mentions it
Where is Schweser is this?
Re: PO vs IO Posted by: null&nuller (IP Logged) [hide posts from this user] Date: May 18, 2009 10:17AM LOS 32©, (d) and (e) are all about mortgage securites - yield curve risks - different behaviour relative to straight treasuries - cfai text goes into MBSs, PO’s, IO’s and their particular yield curve behaviours, use of key rate duration, 2-bond hedges, etc. the theme is to undertand the different behaviour of mortgage-backed securites - how the behave, how to hedge, etc I don’t use schweser so I can’t comment on that…
this where my heard hurts - Take a straight bond - rates rise --> price falls. So the duration (the yield/price relationship) is negative mathematically. But because this is normal behaviour for bonds, by convention the industry calls this “positive” duration. (it’s a bit like delta in Puts - it’s always expressed as positive number even tho’ it’s negative price/value relationship) so for example the key rate duration chart on cfai volume 4 pate 152 shows the bars are positive - showing positive duration as per the convention (despite mathematically being negative relationship) then on the chart on the next page shows key rate durations for IOs and POs. POs (for 10+ onward) have “positive” duration in the same way as straight bonds have “positive” duration by convention. IO’s have “negative” duration at 10+ onward in the sense that it is opposite to straight bonds and POs. that’s why my head hurts. “Positive” means “like straight bonds” (which are mathematicaly negatve yield/price relationsihp). “Negative” actually means “the opposite of positive bond duration convention” (so “Negative” actually is mathematically positive yield/price relationship) then on top of all that, the “duration” relationships of both POs and IOs switch at the short end (under about 10 years) appologies for giving anyone else a headache…
i sort of hate it when others ask/say this, but is this optional? i thought i remembered schweser saying “don’t get fussed on the details”, but NOT optional… but vague recollection. i certainly wouldn’t be white-boarding this, but to each their own.
I would say there’s a 99.7% chance that this won’t be on the exam.