July 28, 2008

As frequent readers of this piece know, we have always been proponents of assessing CTA and trading system performance on risk adjusted return ratios like the Sterling, Sortino, etc instead of just looking at total return or YTD numbers; as those ratios allow one to view returns in relation to risk.

A return of 30%, for example, is nice, but at what cost? If two programs both have 30% returns, but one has twice as much volatility to get those returns; the lower volatility program would have a Sharpe ratio twice that of the higher volatility program. The amount of volatility and whether a return is associated with large drawdowns need to always be considered side by side with any return number.

But there are several risks which dont show up in the normal risk adjusted ratios (Sharpe, Sterling, Sortino, etc). For example, there is the bus risk of one man/woman CTA shops which rely on a single person making all of the decisions, entering trades, etc. What if they get hit by a bus? Who is managing the positions then? You wont see that risk in the Sharpe ratio. There are also hidden risks such as the regulatory risks which surfaced with the talk earlier this month about Congress forbidding speculation on energy futures.

But not all hidden risks are outside of the statistics. A more tangible hidden risk can be found in the available statistics for CTAs with a little digging. That one missing ingredient in assessing risk adjusted performance ratios is how much margin is being used by the manager to achieve his or her risk adjusted performance. This metric is called the margin to equity ratio.

Lets first review how an advisor’s minimum investment amount is arrived at. Minimum investments for CTAs can be split into three distinct levels, specified as

1. the technical minimum amount needed to actually place the trades on the exchanges (the margin requirement)

2. the amount for an investor to withstand any eventual drawdown of the investment

3. the amount to make the percentage returns fit into generally accepted levels

1. Technical Amount. The amount technically needed to place trades is what the exchanges and clearing firms refer to as the margin requirement. Any account which wishes to trade a futures contract on a regulated futures exchange like the Chicago Mercantile Exchange must first have enough money in the account to cover the performance bond requirement of the exchange (the margin)

2. Drawdown Amount. The second part of the minimum investment amount — the amount an investor needs to withstand any eventual drawdown — is another technical level of sorts, on that we must have at least that amount in order to stay above zero. If the investment has the possibility of losing $150,000, for example, in the normal course of operation — than an investor better have at least that amount in order to proceed. If they didn’t, they would have to get out of the investment during the normal ups and downs of the investment.

3. Window Dressing Amount. The third part of the minimum investment amount — the amount needed to make the percentages appealing to potential investors, or window dressing amount — is simply a subjective amount the advisor computes in order for the average returns and risk of his or her program to come out nicely, for lack of a better term.

Margin to Equity Ratio :

Our concern for this article is the first level of a CTAs minimum investment amount above the technical amount, or margin amount. And while the levels of margin used by a CTA will vary from CTA to CTA depending on the type of strategy they employ, markets traded, frequency of trading, and hold period, etc. their minimum investment amounts do not tend to vary along the same scale.

So, it is not uncommon to have 5 CTAs each with $100,000 minimum investment amounts who each use different levels of margin in their trading. One may only trade a single market on a very quick time scale (out by the end of the day) and thus use only $5,000 in margin (or 5% of the minimum investment) on average. Another may be a diversified program with positions in multiple markets across multiple sectors, resulting in $30,000 in margin (or 30% of the minimum investment amount).

Luckily for us, this is a widely reported number for most CTAs, and one that is available on the detail page for each CTA program on the Attain Capital website. This number suffers from a slight self reporting bias, as it is a reported number by each CTA, not the actual statistic generated off of past data; but in our experience the reported number is close enough to the average amount used by the CTA to be meaningful.

The number reported by the CTAs is called the margin to equity ratio, and it is simply the dollar amount of margin they use for their base level accounts (on average) divided by the minimum investment amount. So, a CTA with a $100,000 minimum who has 10 positions which require $25,000 in margin on average can be said to have a 25% margin to equity ratio (often listed as M/E).

So what is a good margin to equity (M/E) ratio? Does it matter if one CTAs is higher/lower than anothers?

Unlike other ratios whose values dont mean anything standing by themselves, only on a relative basis when comparing programs the M/E ratio can be used both as a relative measure to compare programs, and as an absolute measure to give a clue as to future risk.

Comparing CTA returns using M/E :

When using the M/E ratio to compare programs, it all boils down to the technical amount needed to trade a CTA, the margin amount. Because 1/3 to 2/3 of a CTA investment is merely window dressing (the amount to make the reported returns look nice), we can effectively ignore that amount when trying to get down to the nuts and bolts of how the manager actually performed.

If two CTAs with $100,000 minimums both return 20% in a year, or $20,000 in profits if one did so using 10% margin (investing $10K) and the other 30% margin (investing $30K), the return on the actual money used, the returns on margin used are actually 200% and 66%. We can all see that 200% is greater than 66%.

Another way to look at it is to put the return onto the total amount of capital being controlled. Take two fictitious managers who use S&P 500 futures, for example. The S&P 500 futures contract is worth $250 * the index price, and requires just $20,000 or so in margin to trade. With a current price of around 1250, trading a single S&P 500 futures contract means you control $312,500 through that contract (1250*$250)

When looking at it in those terms, and considering two fictitious managers who each require a $100,000 minimum investment, who both trade S&P futures, and who both made 20% last year, but one of which uses 3 times the margin to equity (thus 3 times the number of contracts), we can see that one manager is making $20K on $312K, and the other is making $20K on $937.5K.

The manager who can make the money on less capital invested is more attractive for several reasons. First, because they appear to be more skilled doing more with less. Second, the use of less capital for the same return is more efficient. If the manager only needs $10K instead of $30K in order to make your expected profit, the extra $20K can be put to work elsewhere. Finally, a lower margin use means less capital is at risk in the market; which should lead to lower volatility and lower drawdowns over time. (over the short term, managers with high M/E ratios especially option sellers — may not show their true colors and have low drawdowns and volatility despite the M/E telling us they are at risk of higher numbers)

Using M/E in absolute terms

But what if youre not into notional funding, and dont care that one manager is making money more efficiently than another. Many have said — if they are both making $20,000 in actual cash, and you put in $100K, so are making 20% or $20K in both instances, whats the big deal with the margin to equity ratio? The issue is risk, and more specifically the hidden risk mentioned earlier that M/E can give clues to.

Unlike a Sharpe ratio of 1.65, for example, which tells us very little standing on its own; the M/E ratio can tell us quite a bit about a program when standing alone. I know that a 35% M/E on a $100K account, for example; is going to mean I will have to have at least $35,000 in free capital to post as margin for that program.

But the most important thing the M/E ratio can tell us are clues as to the potential future losses in store for the program its associated with. It is always important to remember that any statistics calculated using past performance are descriptive, not predictive. That is, they are merely showing us what did happen, and are not going to tell us what will happen.

Now, the M/E ratio is not predictive either; but it is not necessarily from past data like other stats. It is telling us the amount of risk (in terms of margin) the manager will be targeting in the future. And when considering that the exchanges set the margin numbers to equal roughly the amount of money a position could lose in a few days time, we can surmise that a really bad streak for a CTA where many if not all of its positions are losing simultaneously would equal that margin usage.

So a CTAs M/E number can give us a quick back of the napkin look at what a new max DD may look like, for example. If a program has had just a 5% past Max DD, yet trades using a M/E of 35% — one should not be too surprised to see an intramonth Max DD of 20% to 30% in the future. Or if we see a program such as Rosetta which has a 30%+ DD, but a M/E of just 7%, we can ascertain that they have deleveraged their program since suffering the 30%+ DD.

In short, a higher margin to equity implies more risk as the manager is controlling more money, thus by definition has more money at risk. A CTA whose returns especially option sellers with shorter track records are very consistent with low drawdowns, but who has a high M/E ratio of 50% or higher, has probably just not had its bad streak yet, versus unlikely to ever have one.

In conclusion, with all else held equal it is better to invest in a CTA with a lower margin to equity ratio. But it is not that easy all of the time. The M/E should really be used as a last tie breaker of sorts for two or more similar programs you are considering. If two trend followers you are looking at have similar risk/reward characteristics, choose the one with the lower M/E ratio.

You still have decisions about strategy types, non correlation, desired returns and what not to factor in. And as we do with the Sharpe, Sterling, et al the M/E should also be used in conjunction with returns, as theres no use choosing a program with a low M/E if it doesnt have the returns youre looking for.

The table below does just that. We calculated the compound ROR divided by the M/E for the 35 CTA programs tracked at Attain, and ranked the Top 10 by that metric showing the top 10 CTAs at Attain who give the most bang for the buck (the margin buck, that is).

IMPORTANT RISK DISCLOSURE