Short Premium Research Dissection (Part 22)

I left off with discussion of our author’s differentiation between high-risk and limited-risk strategies.

Her final point in comparing the two:

     > 5. Significantly lower margin requirements

For me, this is the most intriguing observation and one that raises questions. Apples-to-apples comparison of ROI should be normalized for margin use percentage (MUP).* In this table from Part 21, she states a fivefold difference in MUP between high- and limited-risk strategies. Position size for the latter could be doubled or tripled and still carry a lower MUP. In this case, total returns would significantly outpace high risk.

The issue of margin requirement (MR—also known as BPR and first mentioned in these last two paragraphs) is now very relevant. MR is directly proportional to MUP. In Part 15, I discussed her absurdity in mentioning MR without explaining the calculation behind it. This is necessary to understand the 25% vs. 5% difference. According to Tasty Trade:

     > We are able to define Undefined Risk by the
     > amount of margin that a brokerage firms
     > [sic] requires… this is normally the loss…
     > [due to] a 2 standard deviation move in the
     > underlying… the broker… will hold this
     > amount of capital as margin.

As described in the third paragraph here, PM is similar. The underlying price at 2 SD OTM moves with the market along a curved risk graph, which means PMR is dynamic. Also dynamic is the size of a 2 SD move, which is proportional to underlying price. The move gets smaller as underlying price decreases to ultimately put downward pressure on PMR as the market falls.

While MR calculation for the high-risk strategy should be dynamic on multiple fronts, the limited-risk MR is capped. I am therefore confused how our author arrived at the fixed fivefold difference between the two strategies. I would be a proponent of determining a maximum difference to cover most instances, but even if I give her the benefit of the doubt and assume this to be the case, I need explanation as to why fivefold would be it.

Aside from normalizing returns for MUP, I am also interested in normalizing for [M]DD (see last four paragraphs here). Because the concept suffers from future leak (see footnote), I would make conservative use of it. For example, given a fivefold DD difference I would consider doubling or tripling position size to compare total return.

If a 2-3x position size results in a proportional DD increase, then I wouldn’t do it. DD difference between the two strategies are roughly 30-50% (see Part 21 table). That would be overwhelmed by a 100%-200% change.

When I talk about normalizing for MR and MDD, CAGR/MDD seems to be a comprehensive metric. If the numbers are correct, [unknown sample size aside] then CAGR/MDD is ~10% better for the limited-risk strategy.

The drastic difference in MUP relates back to the author’s first point (Part 21). If MR is calculated correctly, then the question is whether I would implement a high-risk strategy that lost 13.3% in 2008 knowing it could have been far worse had the market crash been more severe.

I will continue next time.

* Suppose you and I put on the same trade for $30,000. If your trade takes up 1% of the account
   and mine takes up 50%, then you can be much more brave with regard to holding and adjusting.