Streetwise Professor

April 20, 2020


Today was one of the most epochal days in the history of oil trading, which is saying something. The front month May contract–which expires tomorrow (21 April, 2020)–(a) settled at a negative price of -$37.63, (b) declined $55.90 from the previous day’s settlement, and (c) exhibited a trading range of $58.17, which is about 3.5x Friday’s settlement price.

Even these eye-popping numbers don’t tell the full story. The last traded price was -$13.10. Note that the settlement price is based on the volume weighted average price during the last two minutes of trading, so an average price of $-37.63 in the last two minutes and a last traded price of -$13.10 means prices moved more that $25/bbl in these minutes.

And we’re not done yet! As I write at 1841 CDT, the price is up to -$5.00–an increase of $32.63.

That there’s what they call volatility, folks.

To put these numbers in perspective, the largest trading range on the any day of the last three trading days of CL contracts from 2000-2019 is $26.65, a day around the time of the financial crisis and the aftermath of Hurricane Ike (October 2008 contract): here’s what I wrote about that event. The median intra-day range on the last three days is $1.6, the mean is $1.5, and the standard deviation is $1.47. So we are around a 40 standard deviation event here.

The largest daily price change during the last 3 trading days is $16.37, with a median of $.73 and a standard deviation of $1.45.

The calendar spread is also extreme, settling at $58.06 between the June and the May. Meaning that if you had storage, you could get paid to take delivery, sell it forward, and lock in that $58.06 (net of what the storage costs you). I guess you could call that the megacontango.

All in all, a historically unprecedented day.

The proximate cause of these wild gyrations, and unprecedented negative prices is, of course, the collapse of demand and the looming exhaustion of storage space, including at the delivery point of Cushing, OK. But although this is a necessary condition for today’s events, it is not a full explanation.

The storage issue has been known for weeks, and discussed intensely. It had been priced in to a considerable degree: contango was already at a historically high level. What information about the availability of storage arrived between Friday and today? Unlikely to be anything that could cause such chaotic price movements.

The likely cause is the difficulty of liquidating about 100,000 open contracts (100 million barrels!) in such extreme technical conditions. It is plausible, and indeed likely, that strategic behavior–perhaps rising to the level of manipulation–is the major cause of how prices moved today against the background of conditions that were widely known on Friday.

Let me start out by noting that something similar, though not as extreme, occurred during the demand collapse and associated flooding of storage during the Financial Crisis. As I documented here, the expiries of the January, February, and March 2009 WTI contracts saw what were then historically unprecedented price collapses. So did other US grades of oil. Here’s a picture from the linked document:

The big downward spikes in the front month-back month spreads correspond with the days around expiry.

How does strategic behavior/market power/manipulation play into this? The model of short manipulation in my 1996 book (only $169 paperback–buy two!) and 1993 J. of Business article formalizes the argument, but the intuition is fairly straightforward. Manipulation exploits frictions and bottlenecks. (My article/book refer to “frictional manipulations.”). There is now a huge friction/bottleneck in Cushing–constrained storage. This bottleneck makes the demand curve for crude at Cushing extremely inelastic, and means that the movement of even small excess quantities of oil into that location will cause prices to decline dramatically.

In these conditions, a trader, or a group of traders with modest-sized short positions can exercise market power by delivering even a small amount of oil over and above the quantity that should flow to Cushing. This drives down the price and allows the trader or traders to cover his (their) position(s) at artificially low prices.

In this situation, the storage bottleneck is the gasoline, the exercise of the market power is the match. With 100,000,000 barrels of open long positions needing to liquidate, given the storage constraint, the resulting conflagration can be epic.

This is, at this stage, a hypothesis. It is a possible explanation of the beyond extreme movements observed today. Under the circumstances, it is a very plausible explanation, and one that deserves scrutiny. And given the amount of money that changed hands today (~$6 billion on a mark-to-market basis) I’m sure that it will get it.

The only parallel I can think of is the onion market in 1955, when the movement of a couple of superfluous carloads of onions into Chicago, and delivery thereof against futures, caused the price to crash below the cost of the bags that they had to be delivered in. There was no demand for the onions (being perishable, and people eat only so many hot dogs), so many of the excess plants ended up getting dumped into the Chicago River. (Which, in 1955, probably improved the water quality.) (Another irony being that Chicago means “stinky onions” in the Miami-Illinois Indian language.)

In 1955, demand was inelastic because onions are perishable (i.e., they can’t be stored). In a way, the lack of storage space makes oil perishable. Even if that analogy isn’t perfect, the economics are the same: an economic constraint (the non-storability of, or the lack of storage space) leads to extremely inelastic demand that makes short market power manipulation possible.

Tomorrow is the last trading day for CLK20. Strap it up! It’s going to be a wild ride.

Print Friendly, PDF & Email


  1. Thank you. Found this from Paul Krugman’s tweet.

    Comment by RMellan — April 21, 2020 @ 9:12 am

  2. Craig, how would you include the impact of the movement of traded volumes from the closing period screen trading to TAS and MOC orders? My sense is that the movement of large players away from the “risk” of execution matching the settle and to the “safety” of a TAS or MOC order that matches up with the settle transfers that price impact power to a declining number of traders with volume to actually execute as individual trades during the settle. I would love to see the ratio of the volume that determined the settlement price versus the volume that prices at the settlement price.

    Comment by Thomas Lord — April 21, 2020 @ 3:30 pm

  3. Wow! Just wow, Krugman recommends SWP. End of times, indeed.

    Comment by The Pilot — April 21, 2020 @ 3:31 pm

  4. the last traded price is at Globex close, 5pm ET, not around settlement time (2:30pm ET) AFAIK

    Comment by br blbo — April 21, 2020 @ 4:53 pm

  5. Don’t tell me, ‘Cushing’ in the local native American dialect means ‘stinky dumped shale oil all over the ground’.

    Comment by Ex-Global Super-Regulator on Lunch Break — April 22, 2020 @ 6:32 am

  6. I am a trader who was active on that day, and I got hit by it (I was long the QM May cash-settled contract). I can assure you that there was LOTS of manipulation going on.

    I’ll give you one example… Once the contract went negative, Interactive Brokers traders were trapped because their trading platform wasn’t supporting negative prices (I got a “price doesn’t conform” error with and negative price). Essentially, we were therefore all trapped. Knowing this… all the major players had to do was move the marker deeper into negative on super-low volume and all of the people who were “stuck” for the open interest, i.e. Me and by broker IB would be on the hook. I could easily have covered at -$2-3, and I would have. But, I couldn’t be cause IB didn’t support negative prices, and presumably neither could many other a trader.

    That alone is robbery, and it’s just one way in which the markets were being manipulated that day.

    Comment by Darrin W — April 22, 2020 @ 10:58 am

  7. @Darrin W. IB reported that it lost $88 million because losses exceeded customer equity in many accounts.

    Given what you say about the IB platform, there might be some legal issues here.

    Comment by cpirrong — April 22, 2020 @ 6:05 pm

  8. @Ex-Global Super-Regulator on Lunch Break. Nah. Cushing means “godforsaken place that’s not good for anything but storing oil.”

    I tell my students: “I’ve been to Cushing, so you don’t have to.”

    Comment by cpirrong — April 22, 2020 @ 6:15 pm

  9. Hi Craig, great read. What do you expect for June contract deliveries? ~20/bbl? Also, as we all know, the demand is stunted with some presumptions that it will recover in the summer, but I think a lot of employers will continue to allow work from home, school will not resume, so we won’t see demand recover for quite some time. Anticipate any domino effects of this?

    Comment by Jared — April 24, 2020 @ 2:52 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress