Intermediate bonds in a PP

[Lonestar]

For some time I've been fascinated with the prospect of using an Intermediate Term Treasury fund or 5 yr Treasuries bonds in place of both cash/LTBonds.  Seems it would just be more simple, and back testing shows long term results very close.  Personally, I've been using Fidelity Short Term Treasury for cash and a combination of TLT and 30yr. treasuries for LTB.

Because of the duration of the long bond or even the 20 yr TLT I'm really surprised substituting  intermediates backtest as well as they do.  Seems like when you shorten the overall duration (average) for the bond cash/bond portion of the PP down to 5 or 6 years the returns would suffer in comparison to the average of the cash/LTB returns.

The major benefit right now in favor of intermediates would be less loss should rates increase.  The major risk appears that intermediates would not offer the low correlations that the LTBonds would provide for the PP.

Your thoughts?? 

[AdamA]

For some time I've been fascinated with the prospect of using an Intermediate Term Treasury fund or 5 yr Treasuries bonds in place of both cash/LTBonds.  Seems it would just be more simple, and back testing shows long term results very close.  Personally, I've been using Fidelity Short Term Treasury for cash and a combination of TLT and 30yr. treasuries for LTB.

Because of the duration of the long bond or even the 20 yr TLT I'm really surprised substituting  intermediates backtest as well as they do.  Seems like when you shorten the overall duration (average) for the bond cash/bond portion of the PP down to 5 or 6 years the returns would suffer in comparison to the average of the cash/LTB returns.

The major benefit right now in favor of intermediates would be less loss should rates increase.  The major risk appears that intermediates would not offer the low correlations that the LTBonds would provide for the PP.

Your thoughts??

Some people have talked about using a fund like IEF for the cash/bond portion (50% of the portfolio).

[melveyr]

If you want to nerd out over it, do some googling about "bullet vs barbell" bond portfolios. Assuming the same duration, the only real difference is the convexity of the two bond portfolios. The barbell has more convexity, meaning interest rate changes are more favorable to you than under a bullet portfolio. However, historically you have had to pay for the privilege of the extra convexity with slightly lower returns.

The Treasury bond market is a never ending hole of learning. It is by far my favorite market 

[Gosso]

I prefer intermediate bonds over the barbell.  Although I personally use a hybrid of the two:

40% 1-5 year CD ladder
10% LTTs

This provides an average duration of around 5 years.  For me this minimizes costs and maximizes yield.  I don't support the thinking that CDs are less safe than T-Bills.

Also, I'm looking at this from a Canadian perspective, so there might be better options for a US investor.  Building a 1-10 year bond ladder would be good as well, but unfortunately I have to pay a $25 commission for each bond trade.

[Lonestar]

Good thoughts.  I can't see any way that an Inter Treas fund or 5yr treasury bonds are going to give you the anywhere near the duration of 50/50 barbell of cash/30 yr bonds.

But, that seems like the beauty of the bullet approach with Intermediates in a rising interest rate market.  Lower volatility with the same overall long term returns in your PP.  I gotta be missing something here. 

[Pointedstick]

Has anyone considered using their VP to switch in and out of long bonds according to interest rates?  You could switch to 30-year bonds once interest rates reach n, then sell and move into a 5-year ladder once they fall to (say) n-3. I believe someone once discussed doing a similar thing with gold.

[Gosso]

Good thoughts.  I can't see any way that an Inter Treas fund or 5yr treasury bonds are going to give you the anywhere near the duration of 50/50 barbell of cash/30 yr bonds.

But, that seems like the beauty of the bullet approach with Intermediates in a rising interest rate market.  Lower volatility with the same overall long term returns in your PP.  I gotta be missing something here.

Keep in mind that you cannot buy only 5 YR treasuries, unless you sell them at the end of the year and buy new ones, which seems a bit impractical.  You'd be better off building a 1-10 year bond ladder; this will act almost the same as holding only 5 YR treasuries (same average duration).  Or just go with IEF, maybe even IEI which is roughly equivalent to the 5 YR Treasury from Simba's spreadsheet.

[KevinW]

I always like to minimize moving parts and transactions, so the idea of merging cash and bonds into a 10-year bond ladder is appealing. Over the course of a year, the only two events in your entire cash+bond allocation could be buying one 10-year bond and allowing another to mature. Elegant and tax-efficient.

But, on the other hand, I like how the PP incorporates and makes use of emergency cash. And I'm concerned that an intermediate ladder may not carry the portfolio adequately in a really horrific deflationary environment.

[Gosso]

Another downside to the bond ladder is deciding what to do with new funds during the accumulation phase; it would get a bit top heavy if new funds were only added to the new 10 year bonds.  Unless you have access to commission free trading, in which case you can just spread the funds evenly along the ladder.  However, the ETF's IEF and IEI would solve this problem.

But, on the other hand, I like how the PP incorporates and makes use of emergency cash.

I keep my emergency cash in a high interest savings account which I still count towards the "bond" portion of the PP.  It lowers the overall duration a bit, but I'm not too concerned about it.

And I'm concerned that an intermediate ladder may not carry the portfolio adequately in a really horrific deflationary environment.

Long term bonds only protect from the initial shock of deflation, since interest rates will fall.  Once interest rates have plateaued then there isn't much difference between LTTs and STTs...except that LTTs will fall if rates begin to increase.

Intermediate bonds will also increase in price during a deflation (falling interest rates).  The overall deflation protection from a PP with intermediate bonds or a STT/LTT barbell is roughly equivalent, as long as the average durations are equal.  It's true the barbell will give you more convexity, however the intermediate bonds provide a higher yield to make up for it...at least that's the theory.

It's hard to go wrong with either the barbell or intermediate bonds.

[melveyr]

But, that seems like the beauty of the bullet approach with Intermediates in a rising interest rate market.  Lower volatility with the same overall long term returns in your PP.  I gotta be missing something here.

That's not quite right. If you are comparing barbells and bullets you do so with constant duration. That would give them comparable volatility. Additionally the convexity of the barbell would mean it would outperform in a rising rate environment.

Bullets outperform barbells if interest rate stay the same, but barbells outperform bullets if interest rates change either direction.

The differences are very subtle, and TBH not likely to be material in the long run for a PP 

[Lonestar]

Good thoughts.  I can't see any way that an Inter Treas fund or 5yr treasury bonds are going to give you the anywhere near the duration of 50/50 barbell of cash/30 yr bonds.

But, that seems like the beauty of the bullet approach with Intermediates in a rising interest rate market.  Lower volatility with the same overall long term returns in your PP.  I gotta be missing something here.

Keep in mind that you cannot buy only 5 YR treasuries, unless you sell them at the end of the year and buy new ones, which seems a bit impractical.  You'd be better off building a 1-10 year bond ladder; this will act almost the same as holding only 5 YR treasuries (same average duration).  Or just go with IEF, maybe even IEI which is roughly equivalent to the 5 YR Treasury from Simba's spreadsheet.

Actually you can buy treasuries, new and on the secondary market, from the Fidelity Bond Desk at no charge plus a very low bid/ask spread.  It's very feasible to buy a 5yr bond, sell it at the end of a year and buy another one.  Just got to be sure you are dealing with a brokerage firm that doesn't charge too much.  Or just as easy to do a 10yr ladder this way but if you are tracking current value it's just more paper work dealing with that many bonds.

As mentioned, it is a question what do you do with semi-annual interest payments?  If doing a ladder that means you have got bonds that will spin off years of interest every six months.  One solution would be to have an intermediate bond fund to sweep the interest payments into.  Another solution would be to sweep interest payments into a MM fund and add to the principal of the next purchased bond. 

[iwealth]

What if you have very limited tax advantaged space? Does a barbell make more sense in that you'd put long duration bonds into tax advantaged and shorter duration in taxable? I'm in this boat and I'm considering doing something as long as possible like EDV in my tax advantaged space and balancing it out with SHY or IEI to reach my target duration.

Is this train of thought correct?

[melveyr]

What if you have very limited tax advantaged space? Does a barbell make more sense in that you'd put long duration bonds into tax advantaged and shorter duration in taxable? I'm in this boat and I'm considering doing something as long as possible like EDV in my tax advantaged space and balancing it out with SHY or IEI to reach my target duration.

Is this train of thought correct?

I think you are correct that the barbell gives one greater flexibility to maximize tax advantaged accounts with the higher yielding bond.

[MachineGhost]

Has anyone considered using their VP to switch in and out of long bonds according to interest rates?  You could switch to 30-year bonds once interest rates reach n, then sell and move into a 5-year ladder once they fall to (say) n-3. I believe someone once discussed doing a similar thing with gold.

But, the time to switch into bonds was back in 1980 and if you rolled over Zeros every year, you would have made 60-fold, trouncing both stocks and gold.  Anything less lengthy would be short-term timing!

[buddtholomew]

Take a moment to read the article Melveyr linked above as there are more yield related variables to consider when selecting either a bullet or barbell approach. Also, I find that any modifications to the PP that momentarily provide mental comfort (e.g. TLT to IEF) could ultimately disrupt the balance of the portfolio and result in a negative outcome. If we were certain that rates would rise, then shortening duration is the obvious choice. The question is "do you feel lucky"?

[Pointedstick]

If we were certain that rates would rise, then shortening duration is the obvious choice. The question is "do you feel lucky"?

That's a great point. And people have been predicting rising rates for years. I bet most of those people never predicted that 30-year yields would drop below 3%. Who knows what the future may hold?

[dualstow]

Desert, there were people talking about doing this two years ago, and they would have royally screwed themselves.
They would not have had a permanent portfolio. Even those who live in countries that don't offer a 30-Year bond buy a 10-Year in lieu of both long bonds and cash.

[iwealth]

I fail to see how buying an intermediate bond fund with duration equal to a cash/LTT barbell will have any long term negative effects on the PP. If anything putting 25% in cash and 25% in LTT is the "simple" choice (mentally) because it promotes the idea that there are 2 separate economic conditions being addressed as opposed to 1. This is one area where backtesting does work to prove a point because the difference is just mathematical semantics.

[Lonestar]

I fail to see how buying an intermediate bond fund with duration equal to a cash/LTT barbell will have any long term negative effects on the PP. If anything putting 25% in cash and 25% in LTT is the "simple" choice (mentally) because it promotes the idea that there are 2 separate economic conditions being addressed as opposed to 1. This is one area where backtesting does work to prove a point because the difference is just mathematical semantics.

What's strange is that intermediate term treasury funds DON'T have durations equal to that of the average of cash and LTT's.  Depending on the fund, most have durations from 5yr to 6.5yr.  Why does long term backtesting show them to be as effective as the cash/LTT combination?

[iwealth]

I fail to see how buying an intermediate bond fund with duration equal to a cash/LTT barbell will have any long term negative effects on the PP. If anything putting 25% in cash and 25% in LTT is the "simple" choice (mentally) because it promotes the idea that there are 2 separate economic conditions being addressed as opposed to 1. This is one area where backtesting does work to prove a point because the difference is just mathematical semantics.

What's strange is that intermediate term treasury funds DON'T have durations equal to that of the average of cash and LTT's.  Depending on the fund, most have durations from 5yr to 6.5yr.  Why does long term backtesting show them to be as effective as the cash/LTT combination?

TLH has an effective duration of 9.97 years.
TLT at 16.82, SHY at 1.81 or an average of 9.31 years.

It's pretty close.

But yes even 5-year bonds work as well in a backtest. Must be because of the extra yield.

http://www.longtermreturns.com/p/histor ... 5__Gold_25

[buddtholomew]

I fail to see how buying an intermediate bond fund with duration equal to a cash/LTT barbell will have any long term negative effects on the PP. If anything putting 25% in cash and 25% in LTT is the "simple" choice (mentally) because it promotes the idea that there are 2 separate economic conditions being addressed as opposed to 1. This is one area where backtesting does work to prove a point because the difference is just mathematical semantics.

What was the drawdown in 2008 for PP investors who invested in ITT rather than the LTT/STT combination?

[Pointedstick]

I fail to see how buying an intermediate bond fund with duration equal to a cash/LTT barbell will have any long term negative effects on the PP. If anything putting 25% in cash and 25% in LTT is the "simple" choice (mentally) because it promotes the idea that there are 2 separate economic conditions being addressed as opposed to 1. This is one area where backtesting does work to prove a point because the difference is just mathematical semantics.

What was the drawdown in 2008 for PP investors who invested in ITT rather than the LTT/STT combination?

It would have actually been up a little bit that year, same as a 4x25 PP:

[iwealth]

Here are the effective durations of the ishares treasury funds:

TLT: 16.82
TLH: 9.97
IEF: 7.41
IEI: 4.39
SHY: 1.81
SHV: 0.39
TLT/SHY combined: 9.32

Whipped up the following scenarios w/ etfreplay.com, the numbers are %'s, PP means traditional TLT/SHY mix, the others assume holding 50% of the bond fund indicated::

For 2008 only:



PortfolioReturnVolatility


PP1.010.5


TLH1.111.0


IEF0.010.7


IEI-2.510.4



From January 1, 2008 through January 31, 2013:



PortfolioReturnVolatility


PP42.68.4


TLH50.98.8


IEF48.28.3


IEI41.88.0



The traditional PP (TLT/SHY) provided good protection in 2008. But overall is didn't fare as well as a portfolio with 50% TLH or IEF from January 1, 2008 through yesterday. Actually the PP return over this period was very close to holding 50% of the IEI fund which is even less sensitive to an interest rate increase...quite a bit less actually, with a 4.39 year duration as opposed to 9.32 for the traditional PP.

Now these portfolios were NOT rebalanced. I don't subscribe to ETFreplay so I wasn't able to do that. Not sure how much of a difference it would have made. But regardless, if one believes interest rates are closer to their lows right now than their highs, and you can get similar performance by overall holding shorter duration bonds...

[Peak2Trough]

What was the drawdown in 2008 for PP investors who invested in ITT rather than the LTT/STT combination?

There was a mid-year drawdown of 15.8% if you bought 10 year Treasuries instead of cash+LTT.  The cash and LTT drawdown was 14.3%.

[sophie]

Maybe these findings aren't surprising.  TLH and TLT have durations that are not so terribly different (16 vs 10 years, approximately).  Since the classic HB PP has a long bond duration of 25 years on average, it's probably more accurate to use EDV (duration 26 years) instead of TLT.  Repeating the backtest on etfreplay between 1/29/08 and 2/1/13, I get:

25% VTI, 25% GLD, 50% TLH:
CAGR +8.4%
Sharpe 0.75
Max drawdown -15.39 %

in comparison,
25% VTI, 25% GLD, 25% SHV, 25% EDV:
CAGR +7.4%
Sharpe 0.59
Max drawdown -13.61 %

Because both portfolios would have rebalanced sometime in 2009, I wouldn't pay much attention to the CAGR.  It's the difference in max drawdown that is noteworthy.