My Experiment in LEAPs

I wondered a while back whether buying LEAPs would be a way to leverage the MFI for greater gains. On August 8, I recorded the 38 MFI stocks that had options available. For each stock, I recorded the close price, as well as the close price of Jan. '07, '08 and '09 LEAPs with a strike price just below the close price of the security. Now, 4 months later, using the close as of December 8, I have determined the change in security and derivative values. LEAP values that expire in Jan. '08 and '09 changed in close accordance with stock price. (Click on the figure, and all figures, to expand them.) As the Jan. '07 expiry date approached, the LEAPs changed in price, in many cases dramatically. Of the 32 stocks that had options expiring in Jan. '07, the mean ± SD was 0.51 ± 1.05. This is a huge gain in that period of time, but much greater variability. The median change in value was 12%.

Here's another way of looking at it: What is the fraction of LEAPs that changed by more than 10%? Figure 2 shows the distribution of LEAP by amount gained, as well as the gain for each category of LEAP. I think it is valid to then calculate an expected return: multiply the fraction in each category by the return of the category for an expected return in each category. By summing the expected returns of each category, you get the overall expected return, which is 17%.

Another way to do this, and probably the most accurate, is with a bootstrap method. Randomly generate portfolios of 5 LEAPs many times, and the average return of the portfolios is the expected return of this strategy. Using this strategy, the mean ± SD is 59.8% ± 41.4%, based on 50 portfolios. Only 3 of the 50 portfolios resulted in a loss, averaging -11% ± 6.5%. By comparison, 11 of the 50 porfolios ended up more than doubling, with average returns of 115% ± 17%.

Is there a correlation to Market Cap? Piotroski F-Score? There doesn't seem to be a correlation between either and returns. The distribution of returns by F-Score is pretty broadly distributed; there aren't really enough samples at most F-Scores to say. Similarly, if there is an effect of market cap, it is slight: the Pearson's coefficient of the curve fit suggests that market cap explains only ~2.5% of the variation - really not enough to be interesting for further study.

There are definite caveats to this. It's one sample of stocks, from a period when the general market is doing exceptionally well. This would really need more thorough backtesting to have a better idea whether using LEAPs of MFI stocks improve returns. But, this data suggests that there may be an advantage to buying LEAPs of MFI stocks as compared to the stocks themselves.