A.I.M. Users Bulletin Board

[ls7550]

Hi Conrad

Most leveraged funds look to scale the DAILY price moves, so if the non leveraged (1x) rises 1% over a day, a 2x leveraged version might be expected to rise 2% that same day.

Double the daily average gain also carries double the volatility. Which compounds out over multiple days to being vastly different to the multi-day 1x gain/loss. Most of the reading about leveraged funds focus in on that multi-day difference between the 1x and 2x (or 3x or whatever).

If looked at from a different angle however, things are (IMO) a little clearer. Invest half the amount in a 2x and keep the other half in cash and that will compare quite closely to that of having invested 100% in the 1x.

A simple model is that it typically costs a leveraged fund the LIBOR rate to borrow and if you have $1000 and invest $500 in cash and $500 in a 2x leveraged fund, then the leveraged fund will borrow another $500 and invest the combined $1000 amount in the 1x. Since 1989 the LIBOR rate (cost of borrowing) has typically averaged much the same as 18 month treasury bond yields.

You invest $500 in 18 month treasury's, $500 in a 2x fund
The funds borrows $500 (at 18 month treasury yield) and invests $1000 in the 1x.
Overall that's little different to having invested all $1000 in the 1x directly.

If your 'cash' earns an average return greater than the 18 month treasury yield (you might perhaps for instance invest in a 5 year treasury ladder and average the 5 year treasury yield rate of return), then, assuming 5 year T's earn more than 18 month T's you win out overall. Since 1989, 5 year T's have averaged around 0.78% more than 18 month T's.

Here's an example of 50% SSO (2x SPY) and 50% SHY (2 year T) compared to 100% SPY (S&P500), total returns year to date



Over time however the weightings will drift from having been 50-50 2x and 2 year T, which will induce some tracking error. So you'd have to periodically rebalance back to 50-50 2x and cash if you wanted track the 1x. This chart for instance shows 50% in 2x gold, 50% in 2 year T compared to 100% in 1x Gold



The tracking error can be either good (relatively outpaces) or bad (relatively lags) with 50-50 chance of either (i.e. is neutral overall).

Why might you want to load 50% into a 2x and 50% into cash instead of 100% into the 1x ? Well there can be advantages. For one if the stock price declined more than 50% in a single day your maximum loss is limited to 50%. For another it can be beneficial to be both a borrower and a lender. In some periods, interest rates have been artificially suppressed whilst inflation has raged. In 1918/1919 I think it was, inflation raged at 18% whilst short term treasury's yielded much much less (lost out heavily in real (after inflation) terms), again a similar happened in the mid/late 1940's. Typically yields are artificially suppressed by government intervention - such as we're more recently seeing with Quantitative Easing. Short term treasury's are being held down at 0.5% or lower levels and if we had a year or two of high inflation then most investments lose out in real terms. $50 in a 2x, $50 in cash and with the 2x fund borrowing 100% of the amount you deposit with them ($50 borrowed), is more neutral relative to real yields. If real yields are highly negative (low yield on short term treasury's, high inflation), then the borrowed sums are eroded by inflation such that you lose less overall in real terms.

Another benefit of leveraged funds is that they can be used to tune your asset allocation. Consider for instance that I'd like to hold a Talmud type asset allocation of a third in business (stocks), a third in real assets (gold, oil, land ..etc) and keep a third in reserves (cash). A simple three fund version of that might be 33.3% Small Cap Value stocks, 33.3% Oil, and 33.4% cash.

If instead I invest 16.6% in 2x SCV, 16.6% in 2x oil and keep 66.6% in cash, then I have similar reward potential, but I've also added in 33.3% exposure to 'borrowed' money. If as a UK investor I buy US versions of 16.6% in 2x SCV and 16.6$ in 2x oil, then I've also added in a virtual 33.3% USD/GBP position. i.e. is more widely diversified and potentially less risky overall.

The Talmud suggested weighting a third each into business, land, reserves - but IMO an acceptable more modern day equivalent of that might include
stocks for the business,
gold, oil, land, agriculture, forestry - type assets for the 'land' (real assets)
Diversified across as many/few individual holdings as you deem to be appropriate.

This chart shows a three way SCV, Oil and CPI (inflation bond) blend since 1947, that up to and including 2008 generated a 6.8% real (after inflation) annualised return in a relatively consistent manner



In Robert Lichello's book he outlined the asset allocation for the Trust 1 (largest) Rockefeller investments back in the mid 1970's. Which broadly speaking was quite concentrated (most of the amounts in a few large investments) and loosely might have been considered as primarily being around a third in business (mostly IBM and Kodak), a third in Oil (mostly Exxon), and a third in Rockefeller centre (cash like investments perhaps).

Harry Browne's Permanent Portfolio might also be considered as a Talmud type blend, but where instead of 33% stocks, 33% land, instead those 33% allocations are split down 75-25 stock/cash type amounts (resulting in 25% stocks, 25% gold (Harry preferred gold as the 'Real' asset) and 50% cash (Harry preferred a short term and long term treasury barbell for 'cash')). Personally I believe that there's already enough cash reserves (33%) and low asset correlation between stocks and gold not to need to reduce down the 33% stock and gold amounts into 75/25 stock/cash proportions - which is also reflected in the PP's relatively lower rewards over the mid to longer term. The STT/LTT barbell can also be equally replaced with a 5 year T ladder whilst still producing similar overall mid to longer term rewards.

Regards. Clive.