GNMA behavior in sharply rising environment

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.

GNMA behavior in sharply rising environment

Postby timunderwood9 » Fri Dec 31, 2010 6:18 pm

Does anyone know how to predict the behavior of the GNMA fund if interest rates rise say 2-3 points? Given the sensitivity of effective duration to the rate of refinancing the duration number itself is irrelevant. If my math is right the duration on a 30 year mortgage that isn't repaid is around 19, but I'm not sure if that is relevant.
timunderwood9
 
Posts: 4
Joined: Sat Dec 11, 2010 11:11 pm

Re: GNMA behavior in sharply rising environment

Postby fishnskiguy » Fri Dec 31, 2010 7:22 pm

timunderwood9 wrote:Does anyone know how to predict the behavior of the GNMA fund if interest rates rise say 2-3 points? Given the sensitivity of effective duration to the rate of refinancing the duration number itself is irrelevant. If my math is right the duration on a 30 year mortgage that isn't repaid is around 19, but I'm not sure if that is relevant.


Not sure just what exactly you want to know. That said, if interest rates go up 2-3 points then:

Duration of the fund will go up
NAV of the fund will go down (percentage wise) by approximately 2-3 multiplied by the new duration.

BTW, I don't see how you came up with the duration of a thirty year mortgage being 19. The duration depends on the interest rate on the mortgage, which you did not specify.

Chris
Trident D-5 SLBM- "When you care enough to send the very best."
User avatar
fishnskiguy
Moderator
 
Posts: 2533
Joined: Tue Feb 27, 2007 2:27 pm
Location: Eagle, CO

Postby livesoft » Fri Dec 31, 2010 7:54 pm

I don't think anyone knows how to predict the behavior of the GNMA fund no matter what they tell you.

In past threads, I think folks (nisiprius?) has plotted the GNMA fund during times of rising interest rates. Not much happened as I recall. What does that mean? I don't know.
livesoft
 
Posts: 32503
Joined: Thu Mar 01, 2007 9:00 pm

Postby SpringMan » Fri Dec 31, 2010 8:18 pm

FWIW, Morningstar rates the interest rate sensitivity as limited for Vanguard's GNMA fund. This is the same rating as most short term bond funds. M* interest rate sensitivity ratings are limited, moderate, and extensive. This could be solely because their current effective duration is 1.9 years. The fund owns mortgages maturing in a wide variety of years, I don't see the relevance of one maturing in 30 years. Total Bond Index is rated as moderate, so I don't think interest rate sensitivity of GNMA will be any worse rated at limited. Duration is a not as important with MBS bonds because of negative convexity explained in other posts. Many experts/authors have posted they don't like MBS bonds. Other posters swear by Vanguard's GNMA fund, I guess it is a matter of opinion.
Best Wishes, SpringMan
User avatar
SpringMan
 
Posts: 4196
Joined: Wed Mar 21, 2007 12:32 pm
Location: Michigan

Postby SpringMan » Fri Dec 31, 2010 8:20 pm

Duplicate post deleted.
Best Wishes, SpringMan
User avatar
SpringMan
 
Posts: 4196
Joined: Wed Mar 21, 2007 12:32 pm
Location: Michigan

Re: GNMA behavior in sharply rising environment

Postby grok87 » Fri Dec 31, 2010 9:02 pm

timunderwood9 wrote:Does anyone know how to predict the behavior of the GNMA fund if interest rates rise say 2-3 points? Given the sensitivity of effective duration to the rate of refinancing the duration number itself is irrelevant. If my math is right the duration on a 30 year mortgage that isn't repaid is around 19, but I'm not sure if that is relevant.

1980 would be a good year to look at. Inflation clocked in at about 13%. Interest rates rose 3% points. Treasuries returned 5.6%. GNMAs returned 0.75%. So in that rising rate environment GNMAs underperformed treasuries by about 5% points.

Now some folks will look at the 1980 year and say GNMAs did fine in this rising rate environment ("they didn't lose money etc.) To me this is flawed thinking because it looks at nominal returns not real returns. The real story of 1980 is that treasuries had a real return of -7.4% and GNMAs had a real return of -12.3%. The S&P 500 had a nominal return of 31.9%, or a real return of 18.9%. So while a 30/70 stock/treasury portfolio eked out a small real gain of 0.5% in 1980, a 30/70 stock/gnma portfolio lost about 3% in real terms.
cheers,
Last edited by grok87 on Fri Dec 31, 2010 10:02 pm, edited 2 times in total.
grok, CFA
"You can only convince people who think they can benefit from being convinced."- Taleb
grok87
 
Posts: 6433
Joined: Tue Feb 27, 2007 10:00 pm

Postby Adrian Nenu » Fri Dec 31, 2010 9:59 pm

GNMAs will get hammered in rising interest rate environment because refinancings dry up and investors get stuck holding lower yield bonds. These bonds' prices also take the hit because their yields are not competitive with newly issued mortgages with higher yields. So GNMA investors get hit from both directions - yield and NAV losses.

Avoid GNMAs and stick with short duration, high credit quality bond funds because interest rates are much more likely to go up than down. CDs are another good option.

Adrian
anenu@tampabay.rr.com
User avatar
Adrian Nenu
 
Posts: 5229
Joined: Thu Apr 12, 2007 7:27 pm

Postby dave.d » Fri Dec 31, 2010 10:08 pm

I recall gaining an understanding from somewhere, years ago, that GNMA's actually LIKE an environment of gradually (say slowly) rising rates -- that is to say, perform well, or relatively well -- somehow because rising rates tend to discourage homeowners from refinancing.

That's too vague to be of any real help, but I think it's consistent with grok's observation about 1980: any bond fund that had even a nominally positive return in a year with a 3% rise in rates did relatively well. I would assume that intermediate and longer treasuries got absolutely slaughtered in such a year, even in nominal terms, and much worse in real terms. I am just barely old enough to remember those days, and recall them just well enough that I'm sure those years play a psychological role in my current allergy to bonds.

I've never felt like I've understood GNMA's well enough to invest in the specific fund, but perhaps this or some other thread out there will help. With interest rates so low, anything that can survive rising rates better than ordinary nominal bonds may be worth a look.
Value-based allocation: recently 48% stocks @S&P 1550 and real bond yields approx. -1% (targets again comparable to past, house purchase taken from overall portfolio)
User avatar
dave.d
 
Posts: 935
Joined: Mon Mar 19, 2007 11:30 pm
Location: Richmond, VA

Postby grok87 » Fri Dec 31, 2010 10:25 pm

dave.d wrote:I recall gaining an understanding from somewhere, years ago, that GNMA's actually LIKE an environment of gradually (say slowly) rising rates -- that is to say, perform well, or relatively well -- somehow because rising rates tend to discourage homeowners from refinancing.

That's too vague to be of any real help, but I think it's consistent with grok's observation about 1980: any bond fund that had even a nominally positive return in a year with a 3% rise in rates did relatively well. I would assume that intermediate and longer treasuries got absolutely slaughtered in such a year, even in nominal terms, and much worse in real terms. I am just barely old enough to remember those days, and recall them just well enough that I'm sure those years play a psychological role in my current allergy to bonds.

I've never felt like I've understood GNMA's well enough to invest in the specific fund, but perhaps this or some other thread out there will help. With interest rates so low, anything that can survive rising rates better than ordinary nominal bonds may be worth a look.

I would recommend credit union FDIC insured CDs with easy early withdrawal penalties. Pen Fed currently has a 7 year CD yielding 3.25%. That's a 55 bp spread over the 7 year treasury. Plus you get a cheap put option to get out at par if interest rates rise- can get out and just lose 1 years interest or 3.25%. If interest rates rise 300 bps then the 7 year treasury will lose about 20%. So you will get 80 cents on the dollar if you cash out. With the PenFed CDs you can cash out at about 97 cents on the dollar and take advantage of the higher rates immediately.
See this thread for example
viewtopic.php?t=44081&highlight=
cheers,
grok, CFA
"You can only convince people who think they can benefit from being convinced."- Taleb
grok87
 
Posts: 6433
Joined: Tue Feb 27, 2007 10:00 pm

Postby Jack » Fri Dec 31, 2010 10:51 pm

dave.d wrote:I recall gaining an understanding from somewhere, years ago, that GNMA's actually LIKE an environment of gradually (say slowly) rising rates -- that is to say, perform well, or relatively well -- somehow because rising rates tend to discourage homeowners from refinancing.

Actually, you have that backwards. Rising rates are detrimental to GNMA returns. As you say, when rates go up, homeowners keep their low interest loans longer. But as a bondholder, that is the worst thing that can happen to you because this effectively increases the duration of your bond fund.

Remember that the decrease in value of a bond is the product of the interest rate rise and the duration. So you get the double-whammy of rising rates and increasing duration at the same time. With other types of bonds, the duration is fixed so you get what you expect. But the average duration of GNMAs increases with rising rates. Not a good thing.

This is because of the unique nature of GNMAs in which the borrower can choose to terminate the loan if rates go down (decreases duration) and choose to keep the loan if rates go up (increase duration). The duration for GNMAs is estimated by the fact that the average loan is only kept for about seven years because people move, refinance etc. This duration increases when rates go up. This unique property is called negative convexity (or concavity).
Jack
 
Posts: 3168
Joined: Tue Feb 27, 2007 3:24 am

Postby fishnskiguy » Fri Dec 31, 2010 11:09 pm

Jack wrote:
dave.d wrote:I recall gaining an understanding from somewhere, years ago, that GNMA's actually LIKE an environment of gradually (say slowly) rising rates -- that is to say, perform well, or relatively well -- somehow because rising rates tend to discourage homeowners from refinancing.

Actually, you have that backwards. Rising rates are detrimental to GNMA returns. As you say, when rates go up, homeowners keep their low interest loans longer. But as a bondholder, that is the worst thing that can happen to you because this effectively increases the duration of your bond fund.

Remember that the decrease in value of a bond is the product of the interest rate rise and the duration. So you get the double-whammy of rising rates and increasing duration at the same time. With other types of bonds, the duration is fixed so you get what you expect. But the average duration of GNMAs increases with rising rates. Not a good thing.

This is because of the unique nature of GNMAs in which the borrower can choose to terminate the loan if rates go down (decreases duration) and choose to keep the loan if rates go up (increase duration). The duration for GNMAs is estimated by the fact that the average loan is only kept for about seven years because people move, refinance etc. This duration increases when rates go up. This unique property is called negative convexity (or concavity).


Sorry, but I disagree. In a slowly rising rate environment people will buy and sell houses at about the rate they would in a zero rate environment. People buy and sell due to marriage, divorce, death, corporate and government location transfers, etc. If rates are rising slowly these sales will still go forward, and the GNMA bond fund holder will continue to be rewarded for taking the negative convexity risk.

It is unexpected rapid rise in rates that hammer GNMA funds worse than other bond funds.

Chris
Trident D-5 SLBM- "When you care enough to send the very best."
User avatar
fishnskiguy
Moderator
 
Posts: 2533
Joined: Tue Feb 27, 2007 2:27 pm
Location: Eagle, CO

Re: GNMA behavior in sharply rising environment

Postby market timer » Fri Dec 31, 2010 11:46 pm

timunderwood9 wrote:If my math is right the duration on a 30 year mortgage that isn't repaid is around 19, but I'm not sure if that is relevant.


Your duration is incorrect, it is considerably shorter. Your calculation does not account for declining principal balance. The maximum possible duration is 15 years, and that is when rates are 0%.
User avatar
market timer
 
Posts: 5032
Joined: Tue Aug 21, 2007 2:42 am

Postby alec » Sat Jan 01, 2011 12:04 am

Go to M* and look at the chart for VFIIX during the early 80s. Ouch,
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" - Upton Sinclair
User avatar
alec
 
Posts: 2228
Joined: Fri Mar 02, 2007 3:15 pm

Postby fishnskiguy » Sat Jan 01, 2011 12:10 am

alec wrote:Go to M* and look at the chart for VFIIX during the early 80s. Ouch,


Had not inflation back then been a big deal (big IF for sure) it would have been no big deal.

Chris
Trident D-5 SLBM- "When you care enough to send the very best."
User avatar
fishnskiguy
Moderator
 
Posts: 2533
Joined: Tue Feb 27, 2007 2:27 pm
Location: Eagle, CO

Postby zaplunken » Sat Jan 01, 2011 12:53 am

SpringMan wrote:This could be solely because their current effective duration is 1.9 years.


GNMA duration seems to change quickly, not too long ago (a week, 2 weeks?) the duration was 1.9 years but today I see it is 2.5 years.

When rates rise don't new mortgages get added to the GNMA fund? I don't see why people think that rising rates are so bad for GNMA. I'm asking a question, I don't know but I am curious.
User avatar
zaplunken
 
Posts: 685
Joined: Tue Jul 01, 2008 10:07 am

Postby alec » Sat Jan 01, 2011 4:00 pm

zaplunken wrote:
When rates rise don't new mortgages get added to the GNMA fund? I don't see why people think that rising rates are so bad for GNMA. I'm asking a question, I don't know but I am curious.


If think you have it backwards. When rates fall, homeowners refinance and pay off the old higher rate mortgage. This has the effect of shortening the maturity/durations of the GNMA securities. This is called prepayment risk.

When rates rise, much less people refinance and the maturity/duration of the GNMA securities lengthens - extension risk.
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" - Upton Sinclair
User avatar
alec
 
Posts: 2228
Joined: Fri Mar 02, 2007 3:15 pm

Postby Jack » Sat Jan 01, 2011 4:31 pm

zaplunken wrote:
SpringMan wrote:This could be solely because their current effective duration is 1.9 years.


GNMA duration seems to change quickly, not too long ago (a week, 2 weeks?) the duration was 1.9 years but today I see it is 2.5 years.

When rates rise don't new mortgages get added to the GNMA fund? I don't see why people think that rising rates are so bad for GNMA. I'm asking a question, I don't know but I am curious.

Another way of looking at it is that as rates rise, people keep their low rate mortgages longer. This means that, as a bond holder, not only are you stuck receiving lower rates but you are stuck with them longer than you previously expected.
Jack
 
Posts: 3168
Joined: Tue Feb 27, 2007 3:24 am

Postby zaplunken » Sat Jan 01, 2011 4:53 pm

Thanks guys I get what you said, I knew that. But it seems you are ignoring that as rates rise people still buy houses and get mortgages. So some of these new mortgages must be bought and become part of the GNMA fund, this was my point. Again I don't know but it seems reasonable that some new mortgages get into the fund even tho refis stop when rates rise.
User avatar
zaplunken
 
Posts: 685
Joined: Tue Jul 01, 2008 10:07 am

Postby Ed 2 » Sat Jan 01, 2011 4:59 pm

Adrian Nenu wrote:Avoid GNMAs and stick with short duration, high credit quality bond funds because interest rates are much more likely to go up than down. CDs are another good option.

Adrian
anenu@tampabay.rr.com


How about I Bonds? Are they better than CD's? I don't have CD's but enough in I Bonds. What do you think?
"The fund industry doesn't have a lot of heroes, but he (Bogle) is one of them," Russ Kinnel
Ed 2
 
Posts: 1508
Joined: Sat May 15, 2010 10:34 am

Postby SpringMan » Sat Jan 01, 2011 5:09 pm

Adrian Nenu wrote:GNMAs will get hammered in rising interest rate environment because refinancings dry up and investors get stuck holding lower yield bonds. These bonds' prices also take the hit because their yields are not competitive with newly issued mortgages with higher yields. So GNMA investors get hit from both directions - yield and NAV losses.

Avoid GNMAs and stick with short duration, high credit quality bond funds because interest rates are much more likely to go up than down. CDs are another good option.

Adrian
anenu@tampabay.rr.com

Adrian,
Last I heard you were 100% in equities. If so, you are not leading by example :wink: . Your advice is good however.
Best Wishes, SpringMan
User avatar
SpringMan
 
Posts: 4196
Joined: Wed Mar 21, 2007 12:32 pm
Location: Michigan

Postby grok87 » Sat Jan 01, 2011 7:44 pm

Ed 2 wrote:
Adrian Nenu wrote:Avoid GNMAs and stick with short duration, high credit quality bond funds because interest rates are much more likely to go up than down. CDs are another good option.

Adrian
anenu@tampabay.rr.com


How about I Bonds? Are they better than CD's? I don't have CD's but enough in I Bonds. What do you think?

I think I bonds and cds are both preferable to gnmas. But you have to be careful with cds. See these threads

viewtopic.php?t=65724&highlight=

viewtopic.php?t=44081&highlight=

cheers,
grok, CFA
"You can only convince people who think they can benefit from being convinced."- Taleb
grok87
 
Posts: 6433
Joined: Tue Feb 27, 2007 10:00 pm

Postby Valuethinker » Sat Jan 01, 2011 9:37 pm

zaplunken wrote:Thanks guys I get what you said, I knew that. But it seems you are ignoring that as rates rise people still buy houses and get mortgages. So some of these new mortgages must be bought and become part of the GNMA fund, this was my point. Again I don't know but it seems reasonable that some new mortgages get into the fund even tho refis stop when rates rise.


New mortgages enter into the fund via purchases of *new* MBS by the GNMA Fund fund manager.

But those MBS will be at new interest rates. *existing* MBS, which hold *existing* mortgages, are where the problem lies with extension risk.

Right now repayment risk is minimal because bank lending conditions mean that even though long term interest rates are at record lows, only a small percentage of borrowers can refinance. That's the credit crunch at work.

Extension risk however applies to all mortgage holders.

The clincher for me was a graph I saw in a CFA presentation: the average GNMA repays something like 90% over 8 years? But the average mortgage in that MBS has a term of 30 years. So 90% of those mortgages are being refinanced in an 8 year period.

If homeowners stop doing that, and hold to maturity, then that's a huge swing.
Valuethinker
 
Posts: 24377
Joined: Fri May 11, 2007 12:07 pm

Postby Opponent Process » Sat Jan 01, 2011 9:48 pm

fishnskiguy wrote:
alec wrote:Go to M* and look at the chart for VFIIX during the early 80s. Ouch,


Had not inflation back then been a big deal (big IF for sure) it would have been no big deal.

Chris


I looked at it's no big deal anyway (the M* chart isn't showing GNMA underperformance in 1980 so far as I can tell, and it's one year anyway). I'd love if someone could show some actual evidence for (significant, I'm a scientist) GNMA underperformance (over maybe more than one year) and not just theory. I don't care either way, just curious. never seen any evidence. was it in the 70s?
30/30/20/20
US/International/Bonds/TIPS
Average Age=37
User avatar
Opponent Process
 
Posts: 5159
Joined: Tue Sep 18, 2007 10:19 pm

Postby Cody » Sat Jan 01, 2011 10:02 pm

Cd's in taxable at the Credit Union? Real return? Inflation?
Seems like a triple whammy.

Cody
Cody
 
Posts: 634
Joined: Sun Dec 02, 2007 10:19 am
Location: Stillwater, Mn

Postby Adrian Nenu » Sat Jan 01, 2011 10:09 pm

Adrian,
Last I heard you were 100% in equities. If so, you are not leading by example . Your advice is good however.


Thanks! I still hold 100% equities in roughly global index allocations but if I decide to invest some money in bonds, I'll go with short term bond funds, TIPS, CDs and stable value funds. The problem is that my 457 plan does not offer a short term bond fund. To minimize GNMA exposure, I'd go with the TIPS bond fund, stable value and total bond market index in equal parts.

Adrian
anenu@tampabay.rr.com
User avatar
Adrian Nenu
 
Posts: 5229
Joined: Thu Apr 12, 2007 7:27 pm

Postby Opponent Process » Sat Jan 01, 2011 10:10 pm

livesoft wrote:I don't think anyone knows how to predict the behavior of the GNMA fund no matter what they tell you.

In past threads, I think folks (nisiprius?) has plotted the GNMA fund during times of rising interest rates. Not much happened as I recall. What does that mean? I don't know.


yes but good men around here pontificate endlessly on how only stupid people would dare own GNMAs, that the negative convexity would form a gravitational singularity and suck the rest of your portfolio value down into it.

besides, "not much happened" pretty much sums up any bond fund, but I digress.

why people hatin' on GNMAs? is it just because intelligent people fall so much in love with a beautiful and symmetrical theory that they forget about including any evidence? or is this just another case of people being hypersensitive to bond fund losses?
30/30/20/20
US/International/Bonds/TIPS
Average Age=37
User avatar
Opponent Process
 
Posts: 5159
Joined: Tue Sep 18, 2007 10:19 pm

Postby Adrian Nenu » Sat Jan 01, 2011 10:26 pm

I looked at it's no big deal anyway (the M* chart isn't showing GNMA underperformance in 1980 so far as I can tell, and it's one year anyway). I'd love if someone could show some actual evidence for (significant, I'm a scientist) GNMA underperformance (over maybe more than one year) and not just theory. I don't care either way, just curious. never seen any evidence. was it in the 70s?


Interest rates went up in 1999 and Vanguard GNMA lost 5.65% in NAV value but it was covered by the 6.42% income and ended the year with a positive return of .78%.

https://personal.vanguard.com/us/funds/ ... t=tab%3A1a

GNMA yield is a bit over 3% right now, bond yields being at historical lows and likely to go up not down. The puny yield will not cover NAV losses this time. GNMA duration is rising making it more sensitive to interest rate changes. At least with other bond funds investors have some control over duration risk in a rising interest rate environment therefore the prudent thing to do is to go with short duration bond funds.

To some investors a 5% - 10% loss in the bond allocation is not a big deal but I think the bond allocation is not the place to take risk in. There is already plenty of risk in the equity allocation.

Adrian
anenu@tampabay.rr.com
User avatar
Adrian Nenu
 
Posts: 5229
Joined: Thu Apr 12, 2007 7:27 pm

Postby Opponent Process » Sat Jan 01, 2011 11:34 pm

Adrian Nenu wrote:
I looked at it's no big deal anyway (the M* chart isn't showing GNMA underperformance in 1980 so far as I can tell, and it's one year anyway). I'd love if someone could show some actual evidence for (significant, I'm a scientist) GNMA underperformance (over maybe more than one year) and not just theory. I don't care either way, just curious. never seen any evidence. was it in the 70s?


Interest rates went up in 1999 and Vanguard GNMA lost 5.65% in NAV value but it was covered by the 6.42% income and ended the year with a positive return of .78%.

https://personal.vanguard.com/us/funds/ ... t=tab%3A1a


and in 1999 Vanguard Total Bond Market Index Fund Investor Shares (VBMFX) lost 6.77% in NAV value but it was (almost) covered by the 6.02% income and ended the year with a negative return of .76%.

https://personal.vanguard.com/us/funds/ ... t=tab%3A1a

virtually identical in my book. certainly not a relative "hammering" of GNMAs.

the theory is a worthy one. it just doesn't seem to have much evidence backing it in relation to the average bond mutual fund investor.
30/30/20/20
US/International/Bonds/TIPS
Average Age=37
User avatar
Opponent Process
 
Posts: 5159
Joined: Tue Sep 18, 2007 10:19 pm

Postby timunderwood9 » Sun Jan 02, 2011 1:18 am

Your duration is incorrect, it is considerably shorter. Your calculation does not account for declining principal balance. The maximum possible duration is 15 years, and that is when rates are 0%.


Thankyou, I suppose that is what happens when calculating something without double checking the formulas, and not thinking about what the number would actually mean :P, oh well.

Given that the average maturity in the fund is probably a bit over 25 years (I looked at the listing of portfolio holdings, as of last quarter) though the actual maximum duration is probably around 12 years, but it will almost certainly be much shorter than that.

If investors expect almost everything to be paid off within ten years the fund will be treated as though it has less interest rate risk than the intermediate fund.

Which explains why its actual growth path has historically been less volatile than the intermediate treasury fund, which didn't make sense to me until now.

Actually I've been pushing my mother to purchase low penalty CDs, not GNMA :) However I was confused about how GNMA worked.
timunderwood9
 
Posts: 4
Joined: Sat Dec 11, 2010 11:11 pm

Postby ftobin » Sun Jan 02, 2011 5:14 am

Of course, the question is, are GNMAs priced accordingly to accommodate the risk from the negative convexity?
ftobin
 
Posts: 851
Joined: Fri Mar 20, 2009 4:28 pm

Postby grok87 » Sun Jan 02, 2011 9:20 am

ftobin wrote:Of course, the question is, are GNMAs priced accordingly to accommodate the risk from the negative convexity?

In my opinion, no.
Look at 10 year returns (using vanguard funds) avg annual:
GNMA: 5.82%
Intermed Treasuries 6.08%

plus you can buy treasuries yourself and skip the 0.25% fund expense ratio. THat would put intermed treasuries at 6.33% or 0.51% better than GNMAs.

People get seduced by the higher yield of GNMAs and forget that higher yield does not equate to higher return.

Stick with treasuries and tips and take your risk on the equity side (either through a slightly higher than usual equity allocation, or tilts to small and value).
cheers,
grok, CFA
"You can only convince people who think they can benefit from being convinced."- Taleb
grok87
 
Posts: 6433
Joined: Tue Feb 27, 2007 10:00 pm

Postby Valuethinker » Sun Jan 02, 2011 9:28 am

ftobin wrote:Of course, the question is, are GNMAs priced accordingly to accommodate the risk from the negative convexity?


Yes but only with an Option Adjusted Spread (OAS) reflecting the market's estimate of future interest rate volatility *at that time*.

We know that VOL and therefore OAS forecasts swing hugely.

If you look historically, I would have expected GNMA funds to outperform Intermediate US Treasury Funds by c. 1% pa (at least 0.5%) over the very long run, reflecting that higher risk.

They have not.
Valuethinker
 
Posts: 24377
Joined: Fri May 11, 2007 12:07 pm

Postby dm200 » Sun Jan 02, 2011 11:16 am

I see a lot of discussion about the pros and cons of GNMA holdings for individual investors and the merits of taking risks in equities not in bonds, like GNMAs.

However, I have not seem much discussion about organizational investors who, because of restrictions on investments, can not hold equities at all and the merits of GNMA holdings.
User avatar
dm200
 
Posts: 6565
Joined: Mon Feb 26, 2007 3:21 pm
Location: Washington DC area

keep this in mind

Postby larryswedroe » Sun Jan 02, 2011 11:32 am

In prior periods of rising rates, like late 70s, we had GNMAs with much higher yields. So much of the losses were covered by high yields. Not the case now.

And risks are much higher because we are starting from much lower levels of rates simply because that means they can rise much more. You can see that in just how quickly the duration of GNMA fund has increased in just couple of months. This is why you cannot measure GNMA risk, only estimated it, because you cannot know the cash flows as they change with rates.

Also keep in mind it matters how assets MIX. And if we get rising inflation, then GNMAs could get clobbered at exactly the same time stocks get hit (they are negatively correlated with inflation as are GNMAs).


Yes you get a small risk premium but as I have said many times--if it was such a good idea why don't people advise writing puts and calls on the Treasury bonds they own (or bank CDs) and taking in those premiums? No one I know recommends it, yet the same people recommend buying GNMAs when the two are equivalent strategies.
larryswedroe
 
Posts: 11683
Joined: Thu Feb 22, 2007 9:28 am
Location: St Louis MO

Postby Opponent Process » Sun Jan 02, 2011 11:47 am

grok87 wrote:Look at 10 year returns (using vanguard funds) avg annual:
GNMA: 5.82%
Intermed Treasuries 6.08%

plus you can buy treasuries yourself and skip the 0.25% fund expense ratio. THat would put intermed treasuries at 6.33% or 0.51% better than GNMAs.


you don't think that's random noise? you think that's a real, significant effect, a guaranteed pattern going forward? I'm having a hard time understanding what kinds of statistical analyses are used in financial fields and what is considered significant. this just seems like noise from a single decade at that.
30/30/20/20
US/International/Bonds/TIPS
Average Age=37
User avatar
Opponent Process
 
Posts: 5159
Joined: Tue Sep 18, 2007 10:19 pm

Postby tibbitts » Sun Jan 02, 2011 1:06 pm

Adrian Nenu wrote:
I looked at it's no big deal anyway (the M* chart isn't showing GNMA underperformance in 1980 so far as I can tell, and it's one year anyway). I'd love if someone could show some actual evidence for (significant, I'm a scientist) GNMA underperformance (over maybe more than one year) and not just theory. I don't care either way, just curious. never seen any evidence. was it in the 70s?


Interest rates went up in 1999 and Vanguard GNMA lost 5.65% in NAV value but it was covered by the 6.42% income and ended the year with a positive return of .78%.

https://personal.vanguard.com/us/funds/ ... t=tab%3A1a

GNMA yield is a bit over 3% right now, bond yields being at historical lows and likely to go up not down. The puny yield will not cover NAV losses this time. GNMA duration is rising making it more sensitive to interest rate changes. At least with other bond funds investors have some control over duration risk in a rising interest rate environment therefore the prudent thing to do is to go with short duration bond funds.

To some investors a 5% - 10% loss in the bond allocation is not a big deal but I think the bond allocation is not the place to take risk in. There is already plenty of risk in the equity allocation.

Adrian
anenu@tampabay.rr.com

That's all true, but to compensate for not taking bond risk, you have to increase equity allocation. But you say (for most people) never increase equities beyond 50%. So you're basically limiting 50% of a portfolio to 1%-ish returns, at today's rates. That's tough to take for most people.

Paul
tibbitts
 
Posts: 4778
Joined: Tue Feb 27, 2007 7:50 pm

Re:

Postby Fallible » Mon Jan 30, 2012 1:17 pm

livesoft wrote:I don't think anyone knows how to predict the behavior of the GNMA fund no matter what they tell you.

In past threads, I think folks (nisiprius?) has plotted the GNMA fund during times of rising interest rates. Not much happened as I recall. What does that mean? I don't know.


I agree about predicting GNMA, which I have. Here's a link to a previous GNMA thread that I think you may be referring to. It's excellent:

viewtopic.php?t=72593
“Knowing yourself is the beginning of all wisdom.” ~Aristotle
User avatar
Fallible
 
Posts: 3709
Joined: Fri Nov 27, 2009 5:44 pm

Re:

Postby Bongleur » Tue Jan 31, 2012 12:18 am

dm200 wrote:I see a lot of discussion about the pros and cons of GNMA holdings for individual investors and the merits of taking risks in equities not in bonds, like GNMAs.

However, I have not seem much discussion about organizational investors who, because of restrictions on investments, can not hold equities at all and the merits of GNMA holdings.


Do institutions own more or less GNMA than the market %? Is the institution : individual ratio for GNMA different than for other bond classes? If the pros own less, then probably GNMAs have uncompensated risks.
Seeking Iso-Elasticity.
Tax Loss Harvesting is an Asset Class.
A well-planned presentation creates a sense of urgency. If the prospect fails to act now, he will risk a loss of some sort.
Bongleur
 
Posts: 2024
Joined: Fri Dec 03, 2010 11:36 am


Return to Investing - Theory, News & General

Who is online

Users browsing this forum: berntson, bigb, conlius, CP4641, etarini, Exabot [Bot], Gary, Jazz56, John3754, kenschmidt, madbrain, plannerman, Rosco, tobyberkman and 91 guests