90% T-Bills/10% Berserk and Other Intriguing Portfolios

Even though I don't think it is the ideal strategy for the vast majority of investors, I still find Nassim Taleb's idea of 90% t-bills and 10% going berserk (that is my word, not his) to be intriguing on what I guess is an academic level.

If you knew what this year's Dry Ships (DRYS) - which went from $17 to $79 last year - and Suntech Power (STP) - which went from $34 to $80 last year - were going to be, and you put 5% into each, and put the rest into some combo of t-bills and euros, and could repeat, you'd be in pretty good shape.

There are various new fixed-income and, more recently, currency products listing, that make the 90% part easier with regard to foreign diversification. (The WisdomTree products that were filed a while ago would go a long way to filling some gaps if they actually list.)

If we move away from the specific 90/10 and think more about getting alpha from a smaller portion of the portfolio, while the rest is more conservatively allocated, here is one thing that a few people have asked about that very likely will not work: putting 50% into a double long ETF and the rest into treasuries.

The idea is that the double long ETF would replicate the market plus the treasury interest, so the two added together would beat the market. There are a couple of OEFs that do this with SPX futures and treasuries, and I believe these funds don't succeed regularly, but someone please correct me if I have that wrong.

The problem with the double long plus treasuries concept is that the double long fund attempts to capture twice the move of the S&P 500 on a daily basis (and even then it is not perfect) not over longer periods of time. Over the next three months (just a random time period, you could pick any other) the double long fund could lag twice the SPX, do better than twice the SPX or come in right on the nose. There is no way to predict, because it depends on the sequence of daily moves in the SPX, which can't be predicted.

Another issue is that it will never be 50/50 after the first day the idea is implemented. From day to day it would vary slightly and, depending on how the market does, it could vary a lot over time. To maintain something close to the intended 50/50, there would need to be a fair bit of rebalancing, which would probably be expensive, and maybe would wear out the calculator. Even then, it seems like you'd always be a day or two behind.

I am curious enough about the concept to keep exploring, but I would rule this specific idea out.

Comments

[User 155760:]

This is not a new idea, though I haven't seen it commerically applied. See Zvi Bodie paper @ prc.wharton.upenn.edu/...

2008 Feb 25 04:01 PM

[Chris Ng:]

I came across this type of portfolio when I was experimenting with options. Apparently some positive academic studies were done with a portfolio of 90% Tbills and 10% leap options. The maximum loss in any year would be around 6% or so, assuming you can earn 4% with the Tbills. So it is quite conservative.

This upside is of course unlimited.

2008 Feb 25 04:09 PM

[Roger Nusbaum:]

yes the concept has been around for a while. the academic aspect is interesting to me in that if, as some believe we are in for a period of below normal returns for equities (maybe this is been going on already if you read Hussman) then taking the risk of the equity market without the prospect of normal equity returns may not be worth it.

this could lead to a traditional 70/30 being 50/50 or something like a 35/35 and the last 30 going into alternative stuff. for now this is just a theoretical exploration.

2008 Feb 25 04:32 PM

[locke:]

I have heard several people say it is a bid idea to hold DDM for many years. I did some simulations, and while DDM did not double the long-term performance of DIA, it certainly would have outperformed DIA in all but the most poorly-timed scenarios.

Here's my question. If I'm retiring in twenty years, why would holding DDM for that time be worse than holding DIA?

2008 Feb 26 12:59 AM

[Roger Nusbaum:]

are you talking the same dollars? investing an entire portfolio (for discussion sake) in a double long fund is not much different than being on margin.

2008 Feb 26 11:14 AM

[Morrison:]

It's different than being on margin, as the risk is far less (no margin calls), and the there is no interest charged on the borrowed money. You pay the fund expense fee (generally 1.5% / year), but thats it.

I'd be interested in responses to Locke's question regarding holding a double long fund to an extended period vs the underlying index (eg. DDM vs DIA or SSO vs SPY).

2008 Feb 26 09:05 PM

[granger:]

I am anxiously awaiting those new Widsom Tree cash funds from other countries. That will fill a key hole in many portfolios.

2008 Feb 27 07:08 AM

[Augustus:]

There is no avoiding the interest costs on the leverage. You may not pay it directly from your acount, the fund pays it.

The double leverage concept works pretty will in the event of a market that is going in the same direction in small increments. It does not work so well in a situation of high volatility, no matter which direction that eventually turns out to be. Someone had a post on here looking at the tracking error of FXI and FXP. Pretty good example since there have been some large FXI swings to really drive the point home.

Suppose the market declines 10%. Unleveraged, you will have 90 cents left. Double leveraged you will have 80 cents. Now, the market goes up 11%. Unleveraged you will be even. Double leveraged, your 80 cents will gain 22% to then be 97.6 cents. So you have underperformed by about 2.6%.

2008 Feb 27 07:36 AM

[Augustus:]

BTW, one method of getting the leverage that is often overlooked is using fairly deep in the money LEAPS. Look at the option premiums and compare to financing costs. Particularly compare to what a modest retail account is paying at the brokerage firm for margin debt. Sure, it is ultimately a directional bet. However there is some loss limit and possibly a opportunity to use the remaining funds more profitably or for hedging and diversification.

2008 Feb 27 07:43 AM