Debunking the Williams Hedge (Part 8)

Today I will offer some concluding, miscellaneous remarks about the Williams Hedge (WH), which I introduced here.

The course supports the WH with a very small sample size: Sep/Oct 2008, Jul 2011, and Aug 2015. Why do they not include Nov 2008 when the crash really accelerates? How does it fare through the May 2010 Flash Crash, when Karen Bruton said in an interview she was looking to get tickets to Mexico? One post-production rough market period still available for testing is Feb 2018.* This blog mini-series, of course, has been about the meltdown in March 2020.

As suggested in the fifth paragraph here, I believe there comes a time when the LP hedge should not be purchased at all. Cheap insurance with large potential payout in case of rare disaster makes some sense to me. Expensive insurance with same potential payout in case of rare disaster does not. I mentioned that purchase of the LP at the 800 strike anticipates a market crash of 65% after already crashing 30%. Do I want to spend money trying to protect against that? This is probably an individual decision having to do with risk tolerance.

I definitely think the course is incomplete without any of this being discussed. Backwardation seriously compromises the WH when followed by an increasing number of uncovered PCS tranches as shown in the bottom table of Part 6. To me, the only thing that makes sense about purchasing expensive insurance is margin control (technical term), which I addressed with the teenies last time. I would never expect the teenies to serve as a true hedge by offsetting realized losses.

One caveat to avoiding expensive insurance is to remain protected in case the market actually does go to zero. Being insured 24/7/365 would be marketable to a large segment of the population. I do not believe this is necessary for success, though. I believe “no risk, no reward” has merit and since I have to carry some risk, I might as well carry it in the tail of the tail.

I believe the WH fails in its stated attempt to get LP for free. We do this by financing the LP with cash raised from selling PCS. This works as long as the PCS remain covered. This may not work once the LP start to expire. If backwardation occurs and short-dated puts become relatively expensive, then the LP expire faster leaving more PCS exposed.

The WH is like a Ponzi scheme of sorts: money from one expiration is used to fund costs in another. In a Ponzi, money from new investors is used to cover redemption requests by older. This can continue until many investors come calling at once, in which case money from newer investors is insufficient to cover redemption demand causing the account to go bust.

* — I would expect the PCS not to show much of a drawdown since this was
       only ~10%, but I could be wrong because the volatility spike was huge.