Are LEAPS black swans?

I tried two experiments with LEAPS a little while back (here and here). It's been a year since the end of the first experiment. I've now followed the same set of MFI stocks for an additional year, and can look at the change in stock and LEAP values for this additional time period. (Acually, it's not the exact same set of stocks; this depends on which stocks had LEAPS available). I'm really glad to have been able to get this extra data: just look at the last three months on the S&P in my previous post - it's been heading downwards for a while. In previous experiments with LEAPS, the market was moving upward fairly stongly. I think that this round of data provide a better picture of what may happen in using this strategy over a long period.

Of the 19 stocks in this data set, 10 were down (52%), while 13 (68%) of the options were in the red. The average change in stock value was +6.2%, +/- 50% with a median change of -0.7%. As past data might make us expect, the change in derivative value was much more variable, with a standard deviation of 157%, making the 18.5% mean increase in value less trustworthy. The median change in derivative value was -53%.

Overall, I think this is striking. Let's imagine that you could hold a portfolio with these 19 stocks. You'd have tripled the return by holding LEAPS rather than the underlying stocks.

What about if you only held a subset of the derivatives? Here I carried out a Monte Carlo simulation. It's small scale because I'm doing this by hand. (If anyone knows of freeware that will do these, please let me know.) I generated 65 portfolios of 5 LEAPS each. The average return was 21% +/- 66%; the median return was 24%. Of the 65 portfolios, 40 were in the black, and half of those were up by more than 50%. 25 of the portfolios were negative, and 12 of those were down by more than 50%.

Notice that one derivative is up by 22%, while five more are up by at least 100%. The remainder are all down. So any portfolio that's up is driven by these five derivatives.

I suspect that the randomization routine in Excel is skewed. If you look at the portfolios that have only one of the 5 derivatives that are up, the return is -13.7% +/- 26%, with a median of -13.8% out of 14 portfolios. Seems like these should have been in the majority, not the minority of the portfolios. I'm not sure that I trust the total results of the Monte Carlo simulation.

But let's go back to the aggregate result: +18.5%. This makes me think of Nassim Taleb's strategy for investing, as he described it in the Black Swan. He describes a barbell approach to risk. Most of the portfolio is put in extremely safe investments, like Treasury bonds. The rest is put in extremely risky investments like LEAPS. Say government bonds are 5% per year, and have 80% of the investment. Then the return is (80% of 5% and 20% of 18.5% for a sum of) 7.7%. That's better than either of the options of investing only in T-bills or only in the underlying stocks. Obviously it's not better than the return of only investing in LEAPS, but without the risk. What if none of the LEAPS had returned as much as they had?

I'm not sure. Maybe holding a portfolio of 19 LEAPS is sufficiently diversified that there's little risk involved. It would really take more data to understand what risks are involved in getting these returns.