After a brief chat with the local Gods (Patrick and Andrew), I thought I would host a Q&A for anybody interested in the life of a hedge-fund volatility trader. Brief background: * Undergrad: abroad so by definition a non-name school :) probably can be described as a double major in math and CS * Grad School: PhD in something or other quantitative, lot’s of math and lot’s of coding * Prequel to PMery: 15 years as an exotic derivatives market-maker, departed at director-level. * Summary: I am not very smart, but I can lift heavy things. I guess post your questions and I'll answer what I can.
Trending Content - Hedge Fund Forum
Career Resources
Can you please explain in layman's terms what's been going on with vol products in last several days?
A whole bunch of retail investors thought that selling volatility was a great idea. After all, they make money every day, what's not to like?
Initially they were shorting VXX (which is an ETF that buys a weighted mix of 1st and 2nd VIX futures). Eventually CS came out with XIV and, a bit later, Proshares issued SVXY. Both of these are (or, in case of XIV, was) "inverse volatility ETFs". Essentially, the ETPs are replicating the opposite position of what VXX has - a weighted basket of 1st and 2nd future with average maturity of 30 days. The mechanics are such that the fund manager holds cash and shorts futures to achieve the exposure. Since we literally had zero vol last year, these things became wildly popular, with multiple yards of NAV. More importantly, the futures positions that they held constituted a meaningful portion of the total open interest.
The way rebalance process works, the fund has to sell VIX futures when VIX futures are down and buy VIX futures when VIX futures are up. I guess you can see what's wrong with that picture. On Monday, VIX futures spiked intra-day causing what can be described as "event" in these ETPs. Essentially, if the weighted average between the two futures spikes above a 100%, the fund could lose more than the cash cushion they are holding. To protect themself from such an event (especially for XIV, which has the rebalance process contractually defined), they built in a trigger allowing them to redeem the note should they intra-day NAV cross a pretty low threshold. That's what happened intraday and, since they had to cover, they drove the futures even further. That awesome spike in the front two futures at the close? That was them.
The end result - from Friday, the two funds lost well over three billion. XIV is getting delisted. A bunch of people were long these things on margin, so they owe their brokers money. In short, BOHICA.
For example, if I think IBM is overpriced and want to short it. I go to some old lady that has it in her brokerage account and borrow the stock (well, IRL i got to her broker). He lends me the stock, I pay him something for that service - now I can sell IBM in the market and deliver it to the buyer.
Now, imagine a situation where there are a lot of people that want to borrow stock and not enough lenders (or, for some reasons, it's been restricted). Well, the stock is not going to sell-off as much as it "should". That's what people call "hard to borrow" - it causes some really interesting things in the market, like short squeezes.
The above suggest that first of all, most of the VIX trading was non-algorithmic, it was rebalancing by a bunch of risk managers (ETPs, option dealers etc). It also suggest that there definitely is an argument for the tail wagging the dog. Was there a bit of programmatic selling in S&P? Of course, people have algos that trade relative SPX/VIX movement, trade momentum, etc. However, I can't see why you would say that it was primarily driven by non-humans.
No opinion on that, to be honest. I have to point out, however, that * record highs and record lows in vol over last three years is probably false, especially if long-term history is taken into account. We certainly saw higher realized vol during the financial crisis and I S&P has realized lower volatility through the years. * statements like "Nth fastest decline" are a bit misleading for a variety of reasons. First of all, what is the exact definition? Also, liquidity and the rates of information transfer in the markets are so much higher today. To have a correct approach you should probably normalize this decline in the fraction of annual volume terms or something like that (e.g. % decline over % ADV).PS. I am not trying to argue (as common on the internet) to prove myself right and to prove you wrong. This is an interesting question that has some bearing on many things we do in the market, so both of us should be interested in actually coming up with the most likely hypothesis.