VIX Curve Data

Hi all,

I was having a spirited debate with a friend of mine over the VIX curve and forward returns. His "rule" is that he never buys when the VIX curve is in backwardation (i.e., negative sloping VIX curve). His focus is short-term in nature for M1-M2. Based on my observations, buying during backwardation can result in very short-term losses but in the next 3-6 month period likely result in returns in excess of the mean.


Does anyone know the best way to pull VIX curve data? Specifically, the curve is always changing, so I'm effectively looking for spreads between M1-M2, M1-M3, ... M1-M8 by day. 

Then I'll simply map QQQ and SPY for corresponding dates and look at returns for which the M1-M2 spread is negative, so on and so forth.


CBOE posts VIX settlement prices and data online, but it all seems to be the 1 month forward contract at current time. I want to look back over the last 5-10years at the curve at each point in time.

 

It really depends on your view of not only where implied volatilities are headed but also vol of vol. Backwardation typically occurs in high volatility periods (e.g. right now VIX futures are in backwardation Jan 2022) in the near term months and less so in the back months. And that's why you might think it's a bad time to buy because vol is priced in high and hence the VIX options are priced at a premium. VIX is mean reverting but obviously the variance is very high - you can easily get burnt selling vol. I don't think excess returns are possible with a strategy that takes advantage of curve shape alone. You need to have some thesis or catalyst that will drive implied vol levels (specifically in SPX options - which is how VIX is calculated) [you should use SPX quotes over SPY or QQQ]. VIX options are listed for only 9 consecutive months along with some weeklys and the underlyings are the VIX futures. So trading VIX you are always playing short dated (<1 year maturities).

I've played around with VIX futures and options quotes data. If you have a TD Ameritrade account, you can use the API to get current levels and even execute trades with automatic trading strategies through there. I do it all in Python. Be careful with any strategies surrounding arbitrage as most of the time the spreads of these contracts eat into your profits. It is always better to use limit orders when you are long gamma and stop orders when you are short gamma.

Let me know if you have follow ups. Hope this helped!

 
Most Helpful

Super helpful / insightful - thanks for sharing.

I think ultimately what I was trying to look at is buying some other asset during a time for which the M1-M2 Vix curve is in backwardation. So, I maybe something as simple as going long QQQ or a more quantitative way to dip buy. Basically, each monthly expiration date will have it's own security. I think what I'd need boils down to the Daily settlement price for each forward futures contract, going back historically (i.e., the Jan 2022 VIX contract is M1 in December).

Interesting paper I found on this subject below:

https://www.mdpi.com/1911-8074/12/3/113/pdf

The idea is not necessarily to buy or sell vol specifically. I've heard of folks getting smoked doing that (some examples if anyone knows off top of the head would be great, although of course I can simply google). But maybe there's some sort of excess return on the "market" when the VIX curve is in backwardation. I haven't gotten into too much detail yet on this paper as I've gotten busy with work, but high level it looks like it's effectively what I'm trying to look at. If you had the M1, M2, ... M8 contract prices each day historically, you could then map out days for which M1>M2. Those days would be backwardation. This paper then appears to take it a step further and define the slope of the curve, which is awesome. How deep into backwardation is it? Now let's look at the slope stats for population N of slope data and look at S&P or Nasdaq returns for the lower quintile of that population N (30 days, 60 days, 90 days out).

 

I see, thanks for the clarification. Either way, if you are using VIX futures curve as a signal it means you have a view on implied volatility levels even if you are not trading vol (e.g. you're saying go long the dip so you effectively have bearish view on vol). Just because the curve is in backwardation doesn't mean it necessarily has to revert back to contango within the M1-M2 period. It's likely it can but not always.

I skimmed the paper quickly, seems like it uses some multivariate regression model as some shaping signal. Relying on technicals to trade could be a strategy but can also lack statistical significance. You need to backtest it first before using it in practice. The reason why regression is a horrible model for VIX is because VIX exhibits autocorrelation and more importantly VIX is highly nonlinear in nature (among other reasons). Effectively what you are trying to do is map a time series of nonlinear returns and shrink the dimensionality to one or several variables. Achieving alpha using a linear model for nonlinear model might work for a small sample of time with some luck but I'd be willing to bet consistent excess returns won't be achieved. A better approach might be to use a time series model (MA, AR, ARMA on the equity returns and GARCH or stochastic vol model on the VIX) or deep neural net if you are trying to go the statistical route. ML models will be sensitive to their hyperparameters and time series have their own issues in ANOVA testing.

Like I said, your best bet would be to have a catalyst in mind where you see some vol regime shift in the front months and take your position accordingly. Or use scenario analysis to determine how the curve might change. To add on, you can look at historical scenarios on how the curve went from inverted to un-inverted and vice-versa and identify similar themes. VIX futures tracks SPX implied vol, so ideally SPX or SPY or anything with high correlation (scaled up for vol differences) could be used to trade. Side note also would be not to start trading single names using a VIX signal as as the vol surface might look very different from the SPX term structure of implied vol used to calculate VIX.

 

Also, CBOE has the below historical prices which actually looks like the data I need.

https://www.cboe.com/us/futures/market_statistics/historical_data/

This would be a painful aggregation process of downloading all of these reports and lining them up. Surely there's a quicker way. I called CapIQ support to see if they have these specific securities which would obviously be great, but they don't. Don't have access to Bloomberg terminal anymore, but suppose it'd be on Bloomberg, although again, would have to pull each security manually.

Maybe it'd be best to download all the data, build a mini database somewhere (Microsoft Access?). Effectively the securities will push forward every month with a new curve, so there's probably still some manual work moving data around every month. But maybe something that just takes a few hours. Nonetheless, snowing in NY this weekend so maybe this will be a little project for myself.

Anyone can see where I'm ultimately going with this - do I BTFD?

 

Quia fuga modi quibusdam adipisci necessitatibus. Deserunt doloremque occaecati maxime modi. Iure adipisci id sed dolor nobis.

Ullam quisquam facere ut harum consequuntur. Molestias quia rerum culpa incidunt. Iste aut commodi non et omnis animi autem veniam.

Occaecati quas labore ipsa perspiciatis esse neque. Illo optio qui in eos ea enim repellat. Sit odio sunt harum ea qui sed distinctio maxime.

Minus est dolorem nesciunt fugiat porro et omnis. Et sequi maxime est non ipsa accusamus.

Career Advancement Opportunities

March 2024 Investment Banking

  • Jefferies & Company 02 99.4%
  • Goldman Sachs 19 98.8%
  • Harris Williams & Co. (++) 98.3%
  • Lazard Freres 02 97.7%
  • JPMorgan Chase 03 97.1%

Overall Employee Satisfaction

March 2024 Investment Banking

  • Harris Williams & Co. 18 99.4%
  • JPMorgan Chase 10 98.8%
  • Lazard Freres 05 98.3%
  • Morgan Stanley 07 97.7%
  • William Blair 03 97.1%

Professional Growth Opportunities

March 2024 Investment Banking

  • Lazard Freres 01 99.4%
  • Jefferies & Company 02 98.8%
  • Goldman Sachs 17 98.3%
  • Moelis & Company 07 97.7%
  • JPMorgan Chase 05 97.1%

Total Avg Compensation

March 2024 Investment Banking

  • Director/MD (5) $648
  • Vice President (19) $385
  • Associates (86) $261
  • 3rd+ Year Analyst (13) $181
  • Intern/Summer Associate (33) $170
  • 2nd Year Analyst (66) $168
  • 1st Year Analyst (202) $159
  • Intern/Summer Analyst (144) $101
notes
16 IB Interviews Notes

“... there’s no excuse to not take advantage of the resources out there available to you. Best value for your $ are the...”

Leaderboard

1
redever's picture
redever
99.2
2
Secyh62's picture
Secyh62
99.0
3
BankonBanking's picture
BankonBanking
99.0
4
Betsy Massar's picture
Betsy Massar
99.0
5
kanon's picture
kanon
98.9
6
dosk17's picture
dosk17
98.9
7
GameTheory's picture
GameTheory
98.9
8
CompBanker's picture
CompBanker
98.9
9
DrApeman's picture
DrApeman
98.9
10
Jamoldo's picture
Jamoldo
98.8
success
From 10 rejections to 1 dream investment banking internship

“... I believe it was the single biggest reason why I ended up with an offer...”