Hacker News Reader

Comment by elliekelly | Hacker News Reader

This bug had serious real-world consequences for real people: https://www.reddit.com/r/tradeXIV/comments/7vi6oa/xiv_after_...

View on HN

I don't think these discussions are about the bug, but rather about being wiped out by what the vix did that day.

Which leads me to what I think is the real fuck up. Not only the index was also live during after hours, but the fixing used as the basis for the formula for the next day was set at a time after trading hours. I am not 100% sure but I don't think this was the case when the index was initially published.

What that meant was that to replicate the index, the market-makers of the XIV and other short-VIX ETNs had to hedge themselves in after-trading hours market, i.e. in a much less liquid market. And that's when the short-squeeze happened. Now would that short-squeeze have happened during trading hours when there was more liquidity? Possibly, you would need to ask a market participant. But at the minimum it made the problem a lot worse. And I am surprised no one got any heat for that.

Just want to point out that a "short squeeze" is not possible with these volatility ETNs. They are collateralized debt instruments for which there is an unlimited supply. The underlying for these products is /VX futures, which also have an unlimited supply. A short squeeze happens when short positions cannot be covered due to limited availability of the underlying.

XIV collapse was caused by an unprecedented spike in the VIX index.

A short squeeze also happens when you have forced buyers (desperate to close their short VIX futures position) with a limited supply of sellers. Which is exactly what happened that day. It’s not the VIX index itself that caused the XIV to be wiped out.

They would just have to pay the premium that short sellers were demanding. I see your point, but calling it a short squeeze seems odd to me.

"Volpocalypse" was indeed an unprecedented spike in IV, which the futures market was reacting to. Seems pretty different to me than the classic short squeeze scenarios of GME or Volkswagen.

no. it was caused by spike of vix futures.

And why do you think the VX futures spiked?

The XIV collapse happened on the largest single day percent change in VIX history. It’s a bit silly to say it was the only the futures but not the literal index they track.

Do you have any reason to support your insult that it’s ‘a bit silly’?

Because those of us who have extensive knowledge of volatility products (I was a specialist in VIX options at the CBOE, the first to make markets in VXX, and one of the largest traders in volatility ETF options over the past decade), believe that the spike in the VIX index was caused by arbitrageurs reacting to a spike in the VIX futures. And VIX futures were simply reacting to the order flow caused by the daily rebalancing of leveraged ETFs.

What research have you done to show evidence otherwise?

Matt Levine's Bloomberg column discussed this contemporaneously (https://www.bloomberg.com/opinion/articles/2018-02-14/is-vix...). For Credit Suisse (issuer of XIV), hedging their exposure to XIV allegedly required buying volatility futures when the price increased. With limited liquidity, this moved the futures prices themselves. Credit Suisse has been sued (https://www.institutionalinvestor.com/article/b1rlv9r030854q...).

This collective price response wasn't a short squeeze, but instead it was closer to a gamma squeeze. It didn't rely on finite supply of any security, just limited liquidity and trade sizes large enough to move the market price.

Omg. You obviously have never traded a VIX future. There are not infinite counter parties.

had vix futures stopped trading at 4:00 instead of 4:15, all would have been fine. The extra 15 minutes is what killed them due to path dependency. The XIV and SVXY reset at 4:15 instead of 4:00, so once the critical termination threshold was crossed, it was game over. Indeed, after 4:15 futures prices began to fall, but it did not matter and XIV was doomed even if vix went to 0.

lesson: path dependency is a @##!

This is a timing issue with the trading. ETFs stop trading at 4pm. And the index for the ETF is calculated at the future close time of 415pm. Eastern Time.

Sure, there's after market hours trading for ETFs but liquidity is probably a lot less after 4pm.

Nonetheless, the failure to calculate a live index after 4pm is probably close to irrelevant for the average investor here.

That subreddit reminds me I've been meaning to follow up on https://www.bogleheads.org/forum/viewtopic.php?t=288192 (outperforming the market by holding UPRO/TMF). Wish that person were still disclosing their returns over time.

Not sure if they outperform on a risk adjusted basis but both UPRO and TMF are 3x levered (55%/45% weight)

A more fair comparison would be like 33% UPRO and 66% TMF. This would get you a portfolio that is similar to SPY, but with some leveraged treasuries on top.

Regardless, bonds have done really well due to the drop in interest rates over the last 10 years and I doubt they can go any lower. Bonds will probably perform a lot worse in the future, unless the fed wants to try out negative rates (unlikely).

Also you don't need them to disclose their returns, you can construct your own version: https://www.portfoliovisualizer.com/backtest-portfolio

Edit: constructing the portfolio, it does seem to outperform SPY on a risk adjusted basis, with a Sharpe of 1.25 from 2010-2021. However, my caveat about bonds still stands.

Agreed, long Treasurys has a low reward, high risk setup. Particularly if you believe the inflation narrative.

The risk/reward in isolation isn’t useful, because the market has priced in and bid them up so to their favorable inverse correlation with equities the past 40 years. Assets are priced due to their usefulness in the risk/reward for a portfolio as a whole.

Who buys long treasuries other than the Fed?

Pretty much every nation that has a trade surplus with the USA, worse it's also possible that a trade surplus between two non USA countries can result in the purchase of treasuries if they transact in USD.

I say "worse", because this money can only enter the US economy by making the government run a deficit, if it does not do this then unavoidable unemployment is the consequence.

They’re projecting future returns and correlations based on past 50 years (the ones we have easily available financial data) on constantly lowering interest rates. If asset inflation already pumped up both long term treasuries and stock assets (which pumped up 3x levered UPRO and TMF), then a follow on interest rate increase will equivalently destroy a stock asset and long term treasury portfolio, triply so for those. They think they’ve found an edge in the market (lol like major financial players don’t know about long term treasury and stock asset inverse correlation the past 40 years), but in reality they’ve just ignored the potential for substantial interest rate increases, because it’s not in their data, and they’re pocketing the extra on top until it blows up (they don’t realize it will).

No. The connection you're making is not correct. Since the GFC we were in a very low vol environment (2008 to 2017) so people shorted vol, picking up pennies, for years... Via XIV. The size of the fund caused it to blow up on one vol spike day. It's built into the mechanics of the fund. And has nothing to do with the fact that the index calculator failed to provide a realtime price after hours.

its not a bug, it was a undisclosed "feature"

Reply on news.ycombinator.com