Gyroscopic Investing

The general theory behind the PP is to buy 30 year bonds at auction, and sell them when they get down to about 20 years left, then buy new 30 year bonds.

On the cash side, you buy 1 year T-Bills, and when they mature, you rebuy new 1 year T-Bills.

What if there was a way that you never had to sell The Long Bonds on the secondary market and instead moved them to count towards the cash portion? There would be a few benefits:

1) Not having to deal with capital gains events if the long term bonds you are selling are within taxable accounts.
2) Not having to pay a fee to a broker to sell bonds on the secondary market, if your broker charges one.
3) Not having to deal with the loss due to the spread on bonds (although it's probably very very very small).
4) Reducing the risk that the federal government will discontinue long term bonds (again) in the future, as they have done in the past

Points 1 through 3 are probably worth very little to most people, since most can use Fidelity for free 2ndary market trades, and the spreads are small, and most have at least some tax-sheltered space to stick their bonds in. Point 4 is the big possible benefit, but it won't make sense until I explain the theoretical process.

MediumTex and I discussed in a previous post about how it's not necessary to have full liquidity in the Cash portion of our allocation. It's impossible that we would need access to all 100% of the cash portion to rebalance other portions, because mathematically it would mean the portfolio would be at $0. As we gradually move up, we might a sweet spot that says 40% of the Cash portion being perfectly liquid, is the sweet spot. The other 60% of cash can be in "deep storage" that provides a better yield, but still have the safe guaranteed properties of cash.

For example I-Bonds are backed by the Treasury, but cannot be redeemed within the first year. If you had 100% of your cash portion in IBonds in the first year of issue, you'd have 0% liquidity since you can't sell them yet. But since IBonds are yielding 3% above the rate of 1 year T-Bills, it would be stupid not to have at least some of your cash portion in them. We are sacrificing some liquidity for a higher yield, but it's liquidity that's unlikely to be needed.

Let's take this to the next level. In theory, a 30 year treasury bond is just as "safe" as a 1 year T-Bill, and would qualify for the cash portion, except that the safety requires one wait the full 30 years to redeem it at maturity. If you need to sell the bond before maturity, you are subject to the whims of the interest rate/bond market and the money isn't "safe" as cash in that regard.

However, if one were to put 99% of their cash portion into 1 year T-Bills, and 1% into a 30 year treasury bond, it's very unlikely that one would need to sell the 30 year bond at a loss, prior to maturity, to get access to cash, because the chances of needing access to the full 100% of cash is virtually zero, as explained above. We can push that percentage up a little bit - 10% long term bonds to 90% t-bills as the cash portion is probably OK too. How about 20% bonds/80% T-Bills? Probably good too.

Here's my theory. What if instead of selling 30 year bonds, from the long-term bond portion of your PP, when they hit 20 years remaining, you simply transfer ownership of that security into your cash portion, as part of the cash "deep storage."

If you're in your accumulation phase, then you are probably directing new contributions to the lowest asset. If using this bond-transfer strategy, you would simply re-classify the 20 year t-bond into cash, and then direct new money as needed into new 30 year bond issuance.

We can exploit this technique for some tax savings as well. Suppose the 30 year bond has hit the 20 year mark and HB would have you sell it. But since you bought it, interest rates have dropped, and the bond has appreciated 20%. If it's in a taxable account, you will have to pay capital gains tax on that 20% gain. If however, you don't sell it, and instead transfer it into your cash portion, you avoid this tax.

Now suppose it's the following year. That bond now has 19 years left, and interest rates have risen, and now the bond is at a 10% loss from when you bought it. It's now being considered as part of your deep storage Cash allocation. But, you have an opportunity to sell it from your taxable account, take an immediate tax benefit, and then just put the sale proceeds into your Treasury MMF or a new T-Bill, since it was supposed to be cash anyway.

The principles to follow of this theory would be:

1) Never let the Bonds within your Cash allocation exceed 60% of the cash portion. If it does, then sell the bonds that would have the lowest capital gains.

2) Once you can sell bonds from your deep storage Cash allocation, and achieve a tax deduction from the loss, do so.

3) If you are selling bonds, try to keep the bonds with the longest duration left, and try to sell the bonds with the shortest time to maturity.

The reason for Principle 3 explains my Benefit #4 above. Suppose the government discontinues issuing long term bonds. No more 30s, and only issuing 10s. But if you happen to have a bunch of 15 to 19 year bonds that are sitting in your Deep Storage cash allocation, you can re-purpose those into the long term portion (because it's the "best" that exists) and then replace them with T-Bills for cash.

I would classify this as an advanced PP strategy to perhaps squeeze out another few hundreths of a basis point of value if you were so inclined. I don't think it would make you rich, but it would be kind of cool and interesting to try.

Good idea....except that it defeats the whole purpose of holding short-term cash equivalent and near-cash instruments. The 30-year US Treasury bond is currently yielding around 2.95% . If 30-year yields rise (and I'm not even talking about Volcker-style rates; even a moderate rise would hurt, as I'm about to demonstrate) to around 5.20% (where they were at in mid-2006...which wasn't that long ago) then a 30-year bond after one year (i.e. it's now a 29-year bond) would show a capital loss of over 34%! The 2.95% interest would offset this somewhat but you'd still be facing a loss of over 31. Even if the bond only had 20 years left you'd still be looking at a loss of over 25% once you counted the interest paid to you.

If you had 10% of your "cash" tied up in 30-year T-bonds and the rate rose as above, you'd be looking at a loss in the cash section of your PP of 3.1% and an overall loss of a little less than a percent for the whole portfolio (assuming nothing else changed...which is a bad assumption; see next paragraph). If you had 20% of your cash as 30-year or near 30-year T-bonds and rates rose to 5.20% or thereabouts, you'd be facing a loss of nearly 7% in your cash portion!

The worst thing about all the above is that if rates rose that much in a short period of time we probably wouldn't be in a period of prosperity (good for stocks) or high inflation with negative real rates (good for gold) but in a year like 1969, 1981, or 1994 when cash was the only thing that provided any help to the PP at all.

Unless you KNOW that rates aren't going to rise any time soon (and remember what Harry Browne said about nothing ever turning out as expected in the investment world) and know exactly when they will rise, you are IMO playing with fire if you replace any cash or near-cash with LT bonds.

I know you are doing this to get a few extra basis points and to have LT bonds still left if they are discontinued. My advice:

A. If you really want to be sure that you have LT bonds (with plenty of years remaining) left if the government discontinues them, then simply buy 30-year LTTs every year and roll them over each year-and-a-day (so if you get any cap gains in a taxable account at least they will be LT cap gains...or put them in tax-sheltered space) instead of letting them wait until they have only 20 years left and then selling them.

B. If you just want higher yields for your cash instead of the near 0% that STTs are paying, set aside 20% or so of your "cash" allocation and buy I-bonds, long-term CDs (like Ally and PenFed) with small early redemption penalties, and high cash value CVLI, or put it into reward checking accounts.

D1984 wrote: If you had 10% of your "cash" tied up in 30-year T-bonds and the rate rose as above, you'd be looking at a loss in the cash section of your PP of 3.1% and an overall loss of a little less than a percent for the whole portfolio (assuming nothing else changed...which is a bad assumption; see next paragraph). If you had 20% of your cash as 30-year or near 30-year T-bonds and rates rose to 5.20% or thereabouts, you'd be facing a loss of nearly 7% in your cash portion!

Let me expand further. Do *not* use mark to market accounting. Value the Treasury Bonds at face value, regardless of the market value at the time you transition them into the cash portion. So even if the bonds are market valued at 20% above face value, add them to the cash portion at maturity face value.

You will not be selling the bond before it matures, so you will not have a loss. The point is that you only do this with a portion of your cash assets that you are highly unlikely to need to touch.

If your goal of the PP is to have money in 20 years, then it doesn't matter that the bond won't be valuable for 20 years if interest rates drop now.

In fact, if interest rates dropped, you could sell the bond, take an immediate tax loss, and rebuy a similar maturity bond to place in the cash portion (deep storage) so you are market neutral on that position but get to realize an immediate tax benefit.

You will not be selling the bond before it matures, so you will not have a loss. The point is that you only do this with a portion of your cash assets that you are highly unlikely to need to touch.

Correction....you will not have a REALIZED loss (since you didn't sell or mark-to-market). That doesn't change the fact that even if you don't sell, your bond is still worth a lot less if rates rise. Refusing to acknowledge losses doesn't make them go away (ask any underwater homeowner who is still thinks their house can be sold at the 2005 or 2006 price they bought it at).

Should one choose not to sell an asset then you are (somewhat) correct; one would not show a realized loss/gain in their PP. However, if they hold the bond in lieu of selling it when it is down, the "loss" will come from the drag on their portfolio that results from owning a debt instrument that pays less than 3% in an environment where even short-term cash and equivalents are paying 5% (or 6%....or 7%, or 8.5%...or anything higher than 3%). TANSTAAFL. Several years of that won't be good for the cash portion of your portfolio at all and could really hurt you if rates rise to double-digits again (although tax loss harvesting like you mentioned could mitigate that somewhat).

Bottom line is I still think it's too risky. If you want to speculate on rates not rising or on the government discontinuing the long bond, do it in the VP (or keep your LT bond maturity at the long end like I suggested, or hold a little of your LT portion in zeros instead of regular bonds).

If you still want to hold a few LTTs in the ST portion of your PP, you might at least want to wait until:

1). LTT rates rise from the lowest they've been (with a brief exception in late 2008) since shortly after WWII

2). The yield curve inverts or flattens somewhat

Additionally, the PP's weakest year (1981) was when rates were shooting up. ST treasuries are simply not as volatile as the other components, therefore they are the last component I would think about watering down. It's easy to think of cash as just "dry powder" that maintains its nominal value, but I think it can be much more than that in a PP because it is the only investment that you can be certain to perform well if rates rise. Other asset classes might do well, but you can be certain about your short bonds that are continually rolling over into higher yielding ones.

Clive often says that over the long term, STT rates are the same as LTT rates. Currently STT rates are much lower but if they were to rise to say 15% (it happens) and LTT rates to say 8%, then you would very much want to have your STT as much as anything so that you had the opertunity to buy those 8% yielding LTT rather than being lumbered with your 2011 vintage 3.5% LTT.

When someone has a 401k with no long term bond option, sometimes we suggest the "lemonade" portfolio: merge cash and bonds into a 50% allocation to intermediate-term bonds.  That sounds like a different way of describing your theory.  Instead of a separate 25% LTT and 25% T-bill allocation, put 50% in a ladder of individual 30-year treasuries.  In aggregate they average out to resemble intermediate treasuries, which resemble a barbell of short- and long-term bonds.

Is that accurate, or am I missing something?

Kevin,

It would be a similar thing, but not quite exactly what I'm going for.

A few posters above hit the nail on the head. I'm mistaken because I neglected to imagine the world in rising interest rate environments. I'm not old enough to have lived through that.

My whole investing career has been one downfall of interest rates over the last 15 years.