Intermediate bonds in a PP

[Pointedstick]

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.

If we start from that rebalance--say, June 15, 2009, we get the following:

25% VTI, 25% GLD, 50% TLH

Code: Select all

CAGR +13.5% 
Sharpe 1.64 
Max drawdown -5.14 %

25% VTI, 25% GLD, 25% SHV, 25% EDV

Code: Select all

CAGR +12.9% 
Sharpe 1.42
Max drawdown -5.93 %

By comparison, during this time period, a portfolio of 100% stocks had a CAGR of 16.8%, Sharpe of 0.86, and max DD of -18.61%.

[Gosso]

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.

Hmmm, I guess you could.  I have never heard of anyone doing that before, but with no commissions and low spreads then it seems feasible, although I'm not sure how well this would work in a taxable account.

***

Another thing to keep in mind is during the 80's when interest rates were above 10%, the duration of a vanilla 30 year treasury was around 10 years, rather than the current 20 years.  This is due to the higher coupon payments; for more info see HERE.

This means that the PP was originally designed with a bond duration of around five years (even less if the LTTs were less than 30 years).  I'm not sure if duration or maturity is more important -- my hunch is duration since it provides a consistent volatility.

EDIT: Since duration is based on a percent change in bond price per a 1% change in rates, means that it makes sense that duration would be lower when rates are above 10%, since a move from 10% to 11% is of far less magnitude than a move from 2% to 3%.

[Peak2Trough]

I'm not convinced the differences in returns or volatility between cash+LTT and ITT is going to amount to much.

If anything, I think the ability to switch between the two provides an opportunity to realize losses on LTT without triggering a wash sale, while remaining invested in a way materially consistent with the goals of the portfolio.

[buddtholomew]

What about the inner workings of an STT/ITT approach enable this blend to outperform in a rising interest rate environment? Is the STT portion able to adjust sooner given its shorter duration? I suspect this is the case, but there is no guarantee that rates will rise uniformly across the yield curve.

[MachineGhost]

Another thing to keep in mind is during the 80's when interest rates were above 10%, the duration of a vanilla 30 year treasury was around 10 years, rather than the current 20 years.  This is due to the higher coupon payments; for more info see HERE.

That's an excellent observation.  It gives more weight to using zero-coupons.  It's not really interest we care about in the end, but total return.

[Gosso]

Another thing to keep in mind is during the 80's when interest rates were above 10%, the duration of a vanilla 30 year treasury was around 10 years, rather than the current 20 years.  This is due to the higher coupon payments; for more info see HERE.

That's an excellent observation.  It gives more weight to using zero-coupons.  It's not really interest we care about in the end, but total return.

You responded right before I was about to amend that post.  After thinking more about it, it makes sense that duration would be lower when rates are high, since a move from 10% to 11% is of far less magnitude than 2% to 3%.

This makes duration tricky since it is generally based on a 1% move in the bond yield.  So I like the idea of maintaining a constant maturity, while the duration will decrease as rates increase, or increase as rates decrease.

[BP]

I have been reading this thread with interest.

If one keeps the barbell approach (25% short term treasuries and 25% long treasury bonds), but shortens the maturity of the long bonds (10, 15, 20), does this give the best of both a bullet and barbell strategy or simply do nothing?

[melveyr]

I have been reading this thread with interest.

If one keeps the barbell approach (25% short term treasuries and 25% long treasury bonds), but shortens the maturity of the long bonds (10, 15, 20), does this give the best of both a bullet and barbell strategy or simply do nothing?

I think that is unrelated.

Bullets vs. barbells are only a discussion for when you are dealing with constant duration. Your proposal of shortening the duration sounds more like a strategic bet based on your thoughts on the direction of interest rates.

[KevinW]

The higher risk is compensated with higher returns.  As we can see in the drawdown numbers, however, the increased risk is real and may show up at the worst times.

This is my principal concern with switching to a bullet. I worry about what would happen in a *really* severe, scorched earth, will the economy survive? sort of deflation or tight-money recession. It's times like that when the purity of PP components really shines and I worry that a bullet would not work as well, even though the mathematical model of duration calculations says it should. Backtests over a relatively short span of years isn't enough to assuage me of this. I'd be curious to compare 25x4 to 25/25/50 through something like the US great depression, though.

[melveyr]

I agree melveyr. 

By the way, another very interesting point that arose in this thread is the fact that the duration of our 30 year bonds is much longer than in the old days of higher interest rates.  I'm now wondering if the PP would be more consistent if a more constant duration was held by modifying the LTT ladder.

Another thing you can do is simply weight based off volatility, the idea being to hold less of whatever is the most volatile and more of whatever is the least volatile. To do this you just do 1/vol for each component, and do the weightings based off of that.

It's not perfect because relative volatilities change but it is a nice reference point.

[buddtholomew]

I'm not sure what we can learn from the below calculation, but I am sure glad that I hold 25% of my portfolio in LTTs.

Hypothetical returns for a barbell bond portfolio of TLT/SHY (50K ea.) versus TLH (100K)

Performance on 2/4
SHY - .02%, $10
TLH - .72%, $720
TLT - 1.28%, $640

[iwealth]

Seems about right, TLH holds a slightly longer duration mix of bonds than a TLT/SHY barbell hence the slightly larger gain. That's a pretty simplistic way of looking at it but it's about right.

[buddtholomew]

That's as complex as I can make it 

[Gosso]

I'm not sure what we can learn from the below calculation, but I am sure glad that I hold 25% of my portfolio in LTTs.

Hypothetical returns for a barbell bond portfolio of TLT/SHY (50K ea.) versus TLH (100K)

Performance on 2/4
SHY - .02%, $10
TLH - .72%, $720
TLT - 1.28%, $640

Also, TLH has a yield-to-maturity of 2.43%, while TLT/SHY is at 1.69%.  IEF is at 1.68%.

[Mark Leavy]

I have really appreciated this discussion of the relative merits between a 'bullet' and 'barbell' strategy for holding the cash/long bond portion of the PP.

I'm now wondering how the dynamic of the 'bullet' approach changes if the bonds are held as a 10 year ladder with an average duration of 5 years - where all bonds are held to maturity, then repurchased as 10 year bonds to keep the ladder flowing.

Does holding to maturity provide any additional benefit?

[KevinW]

There is a small logistical benefit in holding bonds to maturity, since that kind of event happens automatically and does not incur any transaction fees or bid/ask spreads. So you only need to take action and pay transaction costs on the buy side, not on the sell/mature side. The fees and spreads on Treasuries are extremely small, though, so it's not very significant in practice.

[Mark Leavy]

The fees and spreads on Treasuries are extremely small, though, so it's not very significant in practice.

Thank you Kevin.  Much appreciated.
Mark

[Lonestar]

There is a small logistical benefit in holding bonds to maturity, since that kind of event happens automatically and does not incur any transaction fees or bid/ask spreads. So you only need to take action and pay transaction costs on the buy side, not on the sell/mature side. The fees and spreads on Treasuries are extremely small, though, so it's not very significant in practice.

I mentioned earlier in the thread that Fidelity sells treasuries with no fee and a small spread.  Just checked and 2/28/18 maturity has a bid/ask spread of $0.055 per $1000.  That's pretty cheap if one wanted to buy 5 year bonds and then sell and repurchase at the end of the year.  However, I see no benefit of doing that over a 10 year ladder, except you have less bonds to track if you are periodically updating your PP total.  I'm sure none of us ever look at what our PP is worth.........

By the way I'm not promoting Fidelity.  There are lots of good brokerages out there that will probably negotiate the same deal.  I've really enjoyed the input on this thread.  Interesting concepts.