Why not zero coupon bonds?

[KevinW]

The title says it all: why not use long-duration zero coupon Treasury bonds ("STRIPS") for the bond allocation?

I have a hazy recollection that Browne said in one of the radio shows that he analyzed this issue at great length, and concluded that zero-coupon bonds were too volatile.  That explanation is unsatisfying since usually we want to maximize volatility.

Has anyone figured this out?

[craigr]

Zeroes are very volatile and for taxable investors they have imputed income problems so they are expensive to hold. Part of the reason I like the nominal bonds is they are giving me a constant income stream which I use for living expenses and rebalancing. Zeroes only pay out at maturity.

However Tex has recommended a strategy of using the Zero bond ETF from Vanguard (Ticker: EDV) to bolster the relatively low bond exposure in the Permanent Portfolio Mutual Fund. This could be an interesting approach if you can shelter the bonds from taxes.  

[KevinW]

The tax treatment is a good point.  I am a tax-sheltered investor so I forgot about that.  The convenience factor of the cash coupon payments is valid, too, but that seems less compelling to an accumulator such as myself.

I was reminded of zeroes because I was revisiting the possibility of using Vanguard funds following their announcement of waiving brokerage fees on their own ETFs.  So I was wondering about using EDV for the bond component.

If I set aside the issues of tax treatment and the conveniences of coupon payments, is the only strike against zeroes their high volatility relative to nominal treasuries?  If so, is there an argument for why this heightened volatility is bad thing in a Permanent Portfolio?

I don't mean to be contrary, but I'm following cragr's advice to "Be Ruthless!" 

[craigr]

Kevin,

I never have taken a good look at zeroes once I found out they were mostly unsuitable for taxable investors. I did however have concerns about how volatile they were and how the lack of a coupon could affect the overall portfolio returns.

User "cdgoldin" on the Bogleheads forum addressed this particular topic with a previous writing from Harry Browne when he reflected back on this issue in 1992. Previously he thought zeroes were a good idea, but later decided that the negatives outweighed the positives. I quote from the thread here:

http://www.bogleheads.org/forum/viewtop ... ht=#187859

Although Harry Browne initially suggested the use of zero coupon bonds for [part of] the long term treasury bond allocation, he later recanted ("Why Zero Coupon Bonds Aren't What They Seem", October 29. 1992, Harry Browne's Special Reports, Issue 154 Page 15).

He points out that "zeros provide extra power during periods of falling interest rates, but they provide no extra leverage when yields are steady! At such times, a zero will increase in price at a rate roughly equivalent to the interest you would have earned on Treasury bonds. So the smaller investment in Zeros will lag behind the return you would have obtained with a full budget for conventional T-bonds."

He further states, "As we've seen, there's no way to know how much volatility zeros will add to a bond investment. When interest rates fell in 1985, the gains in zeros were roughly twice those of conventional bonds of similar maturities. But the next time interest rates drop, zeros may show more -- or less -- leverage."

Another distinct disadvantage to zeros is that the imputed coupon interest is taxable each year, even though it is not received. Thus, zeros are more appropriate as an IRA investment, where the income tax is deferred -- but then you lose the advantage of capital gains treatment of profits.

After presenting a great deal of detail in regard to the actual vs. theoretical performance of T-bonds and zeros, he concludes, "I don't think it's a good idea to use zero coupon bonds for the Permanent Portfolio -- except for a few investors in special circumstances. Zeros are attractive for someone whose wealth is so tied up in illiquid assets that only a small part is available for diversification and balance. Zeros are an imperfect substitute for Treasury bonds, but for such an investor they can help to achieve a degree of safety....The additional leverage of zeros might be useful for a Variable Portfolio speculation...Benham Target Maturity Trusts are a valid substitute for zeros...and more convenient to work with.

[KevinW]

Thanks for digging that up!

I'd like to try to understand the "zero will increase in price at a rate roughly equivalent to the interest you would have earned on Treasury bonds" part better.  I noticed you plugged the "Investment Strategy in an Uncertain World" compendium of Harry Browne's Special Reports over on Bogleheads.  Do you (or anyone else) happen to know if Issue 154 is included in that collection?

[craigr]

Thanks for digging that up!

I'd like to try to understand the "zero will increase in price at a rate roughly equivalent to the interest you would have earned on Treasury bonds" part better.  I noticed you plugged the "Investment Strategy in an Uncertain World" compendium of Harry Browne's Special Reports over on Bogleheads.  Do you (or anyone else) happen to know if Issue 154 is included in that collection?

I don't think it's included. I just searched the entire PDF and that article is not present. This is a small collection of articles (sixteen articles making up almost 300 pages!). The articles reiterate many things Browne states in his radio show and books and some things not there. It is a collection also of his thinking and, as he admits, how it evolved over time. One of the things I liked about Browne is he was always willing to go back and re-evaluate everything. You can see that in these writings how he came to settle on and refine his investment strategy.

Here is the PDF to which Kevin is referring. I definitely think it was worth buying:

http://www.trendsaction.com/product.php ... 1274330449

[MediumTex]

A full 25% allocation to zero coupon treasurys might provide TOO much volatility.  I like zeroes myself, but not for the whole LT treasury allocation.  I might consider something like 50% EDV and 50% VUSTX, TLT or individual bonds. 

OTOH, just using TLT or individual bonds for the full LT bond piece works just fine.

I like the extra punch zeroes provide.  When there is a spike in bond prices like we saw in 2008, it's nice to be able to participate in it fully, which zeroes allow you to do.

I'm a bit outside the PP recipe with this approach, though, so please do your own due diligence in exploring the best approach for your situation.

[jhigh]

For a Vanguard only (no fee) approximation of TLT, would a 50/50 mix of VUSTX and EDV be a good starting point?  


EDIT: This may be a moot question - I just looked at the expense ratios for VUSTX and TLT.  TLT is low enough compared to VUSTX that it would completely erase any brokerage/trading fees.

[rhymenocerous]

Sorry to resurrect an old thread, but this has been something I've been thinking about recently.  Does anyone have any updated thoughts on purchasing zero coupon bonds directly instead of 30 year treasuries for the tax-advantaged investor?  I'm talking about holding 25% zeros instead of 25% LTTs.  Craigr and Medium Tex, is this something you touch on in the bond portion of your upcoming book?

[AdamA]

Sorry to resurrect an old thread, but this has been something I've been thinking about recently.  Does anyone have any updated thoughts on purchasing zero coupon bonds directly instead of 30 year treasuries for the tax-advantaged investor?  I'm talking about holding 25% zeros instead of 25% LTTs.  Craigr and Medium Tex, is this something you touch on in the bond portion of your upcoming book?

I think in the grand scheme of things it wouldn't matter that much.  The biggest problem is the bid/ask spread bc zeroes are thinly traded. 

[sophie]

I've also been somewhat confused by the discussion on this issue.

He points out that "zeros provide extra power during periods of falling interest rates, but they provide no extra leverage when yields are steady! At such times, a zero will increase in price at a rate roughly equivalent to the interest you would have earned on Treasury bonds. So the smaller investment in Zeros will lag behind the return you would have obtained with a full budget for conventional T-bonds."

I must be missing something here, but I would guess that when yields are steady, nominal bonds won't "provide extra leverage" either.  The only mildly convincing arguments I've heard is the high bid/ask spread on zeros, and lack of timely income in a taxable account.  EDV takes care of the latter problem, and effectively hides the former.

I imagine the PP as a pigpen with 4 feeding troughs.  When the pigs run from one trough (e.g. stocks), they run to another trough (e.g. bonds).  When stocks go down, though, TLT and EDV both increase, but only EDV has enough punch to cover the losses.  It's really a shame HB is not around to comment on EDV, but I'm increasingly wondering if it wouldn't be preferable to TLT.

[Lone Wolf]

It's really a shame HB is not around to comment on EDV, but I'm increasingly wondering if it wouldn't be preferable to TLT.

I feel like Browne talked about zeroes in his excellent "Why the Best-Laid Investment Plans Usually Go Wrong" but I unfortunately don't have my copy handy.  I can check on this tonight and I'll post something up if this book contained any of his thoughts on the issue.

Edit: I should also mention that "Best-Laid Investment Plans" was written before the 1992 article that craigr cited above.  Therefore, I'd say that the 1992 advice supersedes whatever I'm going to find when I check the book tonight.

[hoost]

It's really a shame HB is not around to comment on EDV, but I'm increasingly wondering if it wouldn't be preferable to TLT.

I feel like Browne talked about zeroes in his excellent "Why the Best-Laid Investment Plans Usually Go Wrong" but I unfortunately don't have my copy handy.  I can check on this tonight and I'll post something up if this book contained any of his thoughts on the issue.

I believe he said something like using 1/3 of the allocation you normally would, but I also don't have my copy handy.  He definitely mentioned it, though.

[Lone Wolf]

It's really a shame HB is not around to comment on EDV, but I'm increasingly wondering if it wouldn't be preferable to TLT.

I feel like Browne talked about zeroes in his excellent "Why the Best-Laid Investment Plans Usually Go Wrong" but I unfortunately don't have my copy handy.  I can check on this tonight and I'll post something up if this book contained any of his thoughts on the issue.

I believe he said something like using 1/3 of the allocation you normally would, but I also don't have my copy handy.  He definitely mentioned it, though.

Hoost, you are pretty much right on.  I checked the book and Browne's discussion of zeroes is pretty thorough.  His rule of thumb was to consider them to be three times more volatile than a conventional bond.

Browne's advice was to stick with the conventional bonds when possible, as these would provide enough volatility to protect against a deflation.  He recommended zeroes only for those who were stuck in situations where they were not sufficiently liquid to build a proper 25% LT bond allocation.  For these people, he thought that zeroes would allow them to do more with less.

He also expressed great concerns about bid-ask spreads, which tended to be large.  I don't invest in zeroes, so I've no idea whether this applies to EDV and modern-day STRIPS.

[sophie]

I agree with Clive...per google finance, beta for EDV for the past year is -0.5, compared to -0.29 for TLT.

Also here's the results of simulations I ran comparing TLT with EDV, in the available time range (2/1/2008 - 6/30/2012). 

PP with TLT:  CAGR 8.55, 2 rebalances
PP with EDV: CAGR 10.92, 3 rebalances (extra one in late 2011)

This is using data from Yahoo Finance (eyeballed for accuracy), 15/35 rebalancing, SHY for cash, and reinvestment of dividends (since I used adjusted close data).  I haven't quite figured out how to calculate beta or sharpe yet, but there was minimal volatility difference between the two portfolios.

I don't think we can know for sure what HB would have thought of EDV.  Zeros were not likely a palatable option at the time, because of the need to pay taxes on phantom income, so he may not have considered them seriously.  These simulations suggest that anyone who wants to try EDV at least won't get hurt.  I'm not sure I'd want to do more than dip my toe in the water, though, until we've seen what happens during a period with big increases in bond yields.

[Greg]

I agree with Clive...per google finance, beta for EDV for the past year is -0.5, compared to -0.29 for TLT.

Also here's the results of simulations I ran comparing TLT with EDV, in the available time range (2/1/2008 - 6/30/2012). 

PP with TLT:  CAGR 8.55, 2 rebalances
PP with EDV: CAGR 10.92, 3 rebalances (extra one in late 2011)

This is using data from Yahoo Finance (eyeballed for accuracy), 15/35 rebalancing, SHY for cash, and reinvestment of dividends (since I used adjusted close data).  I haven't quite figured out how to calculate beta or sharpe yet, but there was minimal volatility difference between the two portfolios.

I don't think we can know for sure what HB would have thought of EDV.  Zeros were not likely a palatable option at the time, because of the need to pay taxes on phantom income, so he may not have considered them seriously.  These simulations suggest that anyone who wants to try EDV at least won't get hurt.  I'm not sure I'd want to do more than dip my toe in the water, though, until we've seen what happens during a period with big increases in bond yields.

I dipped my toe in a bit because since the IRA can only be a max of $5000, I was only able to buy 4 Treasury Bonds. With the balance left I bought 4 shares of EDV so I've got about $75 left in there that is just cash.

[MediumTex]

When reviewing EDV's performance in recent years, make sure to note the large distributions it has made.

[Greg]

When reviewing EDV's performance in recent years, make sure to note the large distributions it has made.

Would distributions really matter though if it is in a tax-deferred account though? The distributions would lower (potentially) the ETF price but with a tax-deferred account I would think it wouldn't matter. Is my understanding sound for this?

[sophie]

Quote from: 1NV35T0R
Quote from: MediumTex on Today at 08:38:27 AM
When reviewing EDV's performance in recent years, make sure to note the large distributions it has made.

Would distributions really matter though if it is in a tax-deferred account though? The distributions would lower (potentially) the ETF price but with a tax-deferred account I would think it wouldn't matter. Is my understanding sound for this?

The simulations used adjusted close, which takes distributions and dividends into account - but no deductions were made to allow for taxes, if that was your point.  You're of course correct, EDV is probably the least tax-efficient and should be in tax-advantaged accounts if at all possible.

[clacy]

I do hold a fair amount of BTTRX and EDV in my IRA for my portion of LTT.  At some point, if it appears bonds are headed for a significant period of rising rates, I may switch away from Zeros all together however.  I guess that's market timing, but I have some ability to catch major macro trends it seems.