cellardoor1

Berkshire Hathaway's Peculiar Volatility Numbers

1.from 2007 10-K: options expire between 2019-2027; “[t]hese puts had original terms of either 15 or 20 years and were struck at the market.” If you go back through the financials, he’s been selling these for maybe 4 years or so, though more volume in recent years.

b)from most recent 10-Q:

-mildly interesting is the fact that he now comments "Berkshire’s equity index put option contracts are European style options and at inception had durations of 15-20 years." This seems to suggest that sometime in the last 9 months, he wrote some puts that had duration of something _other_ than exactly 15 or 20 years

-another calibration point: "[T]he weighted average remaining life of [BRK's put option] contracts was approximately 13.5 years"

2.Re: your revision for long-dated vols from Andrew Clavell above – can you shed some additional light on the actual timespan/index/source of data?

Since I can’t say for sure what he’s referring to – as far as I know (off random Bloomberg searching), some of the longest-dated, exchange traded (non-OTC) options out there are probably December 2017 expiry eurostoxx index puts. Note that these are still too shortdated for BRK’s average put option. (Also, apply some transformation to the long-dated vols stated below to adjust for differences in short term implied vols between eurostoxx and S&P-based options.)

The thing is, what’re these things (or Clavell’s reference securities) actually transacting at? Assuming that no one’s actually transacting (as is so often the case in super-illiquid options markets), or at best, transacting in tiny size when someone crosses the mid, it means that the implied vol that far out is at best a hack of average (highest bid, lowest offer). You might argue that this is the case for any market – but when we’re talking about stuff this far out, even small potential asymmetries in bid vs offer prices hugely skew the mid – and if I had to guess, the supply demand balance between sellers and buyers of super-long-dated vol has gotta be in favor of the sellers (BRK aside).

Also, off Bloomberg for the 3000 strike dec2017 eurostoxx puts, all I can view is implied volatility using best price, which may be equally bad/even worse(“Measure of the volatility of the underlying security. The implied volatility is determined by using the underlying security's last trade price and the option best price currently existing in the market, rather than using historical data on the price changes of the underlying security.”) – but note that this is currently at 26.43%.

Let’s get back to the concept of actual arms length transactions pricing being the best determinant of implied vols (rather than mid, etc) -- the thing is, BRK has continued to successfully sell these puts during 2008 – and obviously in ridiculous size compared to the total market for such contracts (tho only a tiny amount compared to their total put sales over the last 5 years or so). To the degree that they’re probably still willing to sell at “low” (22-23%) vols, and no one else has been able to transact such things, they have pretty good justification for the pricing to be wherever they sold at.

Not that I think BRK is doing this to manipulate prices, but insofar as few people have the wherewithal (or insanely advantageous margin agreements) convince buyers to accept such an OTC contract for 20 years – where the buyer is exposed to the seller’s credit risk, rather than an exchange’s, and not get paid _any_ margin no matter how much money the seller owes – BRK can basically set market price via actual transactions, wherever they want (as long as they don’t price too aggressively and still find people willing to take on the other side).

3.Also, pretend there are exchange traded puts of the same lengthy duration as the ones that BRK sold. Comparing them to the one’s sold by BRK (not through an exchange), side by side, one would _expect_ the price of the ones BRK sold to be much lower.*

After all, you’re buying puts from someone who never posts collateral (short of downgrade in credit rating), no matter how much money s/he loses, rather than removing counterparty risk through exchange-based purchases. The fair economic value for such puts _should_ be lower. And, since implied vols are mechanically generated from price and other inputs, the implied vol of the OTC option sold by BRK _should_ be lower than that of the identical exchange-traded contract.

*-unless you think, perhaps rightly, that there may be less counterparty risk from BRK than the options clearinghouse/exchange – even after adjusting for mortality vs govt bailout likelihoods

4.The less tongue-in-cheek point here is that all those puts that buffett sold have embedded exposure to BRK credit risk (because of the collateral posting terms).

Insofar as people _think_ that BRK is more likely to blow up (check out the numerous articles on BRK CDS spreads blowing out) -- guess what -- it's consistent for those puts to get marked down in value.

5.Here's something that's maybe a bit more valid (or at least more interestingly incorrect):

if you think that we're about to enter a long-term deflationary cycle, since these puts are struck in nominal terms, the expected amount that BRK could be on hook for in 15 years might be much larger than current 10-15 year bond yields seem to imply. For that matter, the whole BRK business model of snagging cheap float up front starts looking a whole lot worse...