Streetwise Professor

April 21, 2020

WTI-WTF? Part II (of How Many???)

Filed under: Clearing,Commodities,Derivatives,Economics,Energy,Regulation — cpirrong @ 2:23 pm

Just another day at the Globex, folks. May WTI up a mere $49.88 on its last trading day at the time I write this paragraph, a while before the close. (Sorry, can’t calculate a percentage change . . . because the base number is negative!) That’s just sick. But at least it’s positive! ($12.25. No, $9.96. No . . .) (This reminds me of a story from Black Monday. My firm did a little index arb. We called the floor to get a price quote on the 19th. Our floor guy said “On this part of the pit it’s X. Over there it’s X+50. Over there it’s X-20. I have no fucking idea what the fucking price is.”)

But June has been crushed–down $7.35 (about 35 percent). Now the May-June spread is a mere $.83 contango. That makes as little sense as yesterday’s settling galactic contango (galactango!) of $57.06. (Note that June-July is trading at at $7.71 and July-August at $2.65.

I’m guessing that dynamic circuit breakers are impeding price movements, meaning that the prices we see are not necessarily market clearing prices at that instant.

A few follow-ons to yesterday’s post.

First, the modeling of the dynamics of a contract as it approaches expiration when the delivery supply/demand curve is inelastic, and some traders might have positions large enough to exploit those conditions to exercise market power, is extremely complicated. The only examples I am aware of are Cooper and Donaldson in the JFQA almost 30 years ago, and my paper in the Journal of Alternative Investments almost a decade ago.

Futures markets are (shockingly!) forward looking. Expectations and beliefs matter. There are coordination problems. If I believe everyone else on my side of the market is going to liquidate prior to expiration, I realize that the party on the other side of the contract will have no market power at expiration. So I should defer liquidating–which if everyone reasons the same way could lead to everyone getting caught in a long or sort manipulation at expiration. Or, if I believe everyone is going to stick it out to the end, I should get out earlier (which if everybody else does the same results in a stampede for the exits.)

In these situations, anything can happen, and the process of coordinating expectations and actions is likely to be chaotic. Cooper-Donaldson and Pirrong lay out some plausible stories (based on particular specifications of beliefs and the trading mechanism), but they are not the only stories. They mainly serve to highlight how game theoretic considerations can lead to very complex outcomes in situations with market power and inelasticity.

One thing that is sure is that these game theoretic considerations don’t matter much if the elasticities of delivery supply and demand are large. Then no individual can distort prices very much by delivering too much or taking delivery of too much. Then the coordination and expectations problems aren’t so relevant. However, when delivery supply or demand curves are very steep–as is the case in Cushing now due to the storage constraint–they become extremely relevant.

Perhaps one analogy is getting out of a theater. When there are many exits, there won’t be queues to get out and little chance of tragedy even if someone yells “fire.” If there is only one exit, however, hurried attempts of everyone to leave at once can lead to catastrophe. Moreover, perverse crowd dynamics occur in such situations. That’s where we were yesterday.

About 90 percent of open interest liquidated yesterday. That is why today is returning to some semblance of normality–the exit isn’t so crowded (because so many got trampled yesterday). But that begs the question of why the panicked rush yesterday? That’s where the game theoretic “anything could happen” answer is about the best we can do.

About that storage constraint. My post yesterday focused on someone with a large short futures position raising the specter of excessive deliveries by not liquidating that position, thereby triggering a cascade of descending offers until the short graciously accepted at a highly profitable price.

But there is another market power play possible here. A firm controlling storage could crash prices (and spreads) by withholding that capacity from the market. The most recent data from the EIA indicates about 55 mm bbl of oil storage at Cushing. That’s about 80 percent of nameplate capacity (also per EIA.). Due to operational constraints (e.g., need working space to move barrels in and out; can’t mix different grades in the same tank) that’s probably effectively full. Therefore, someone with ownership of a modest amount of space could withhold it drive up the spread. If that party had on a bull spread position . . .

Third, we are into Round Up the Usual Suspects mode:

And first in line is the US Oil ETF. There has been a lot of idiotic commentary about this. They were forced to take delivery! (Er, delivery notices aren’t possible before trading ends.) They were forced to dump huge numbers of contracts yesterday! (Er, they publish a regular roll schedule, and were out of the May a week before yesterday’s holocaust. They also report positions daily, and as of yesterday were 100 pct in the June.)

Not to say that USO can be implicated in hinky things going on in the June right now, but as for May–that dog don’t hunt.

Fourth–WTF, June WTI? Well, my best explanation is that the carnage in the May served to concentrate minds regarding June. No doubt risk managers, or risk systems, forced some longs out as the measured and perceived risk for June shot up yesterday. Others just decided that discretion was the better part of valor. The extremely unsettled positions no doubt impaired liquidity (i.e., just as some wanted to get out, others were constrained by risk limits formal or informal from getting in), leading to big price movements in response to these flows. If that’s a correct diagnosis, we should see something of a bounceback, but perhaps not too much given the perception (and reality) of an extremely asymmetric risk profile, with going into expiry short being a lot more dangerous than going into it long. (This is why expectations about future conditions at delivery can impact prices well before delivery.)

Fifth, on a personal note, in an illustration of the adage that the apple doesn’t fall far from the tree (and also of Merton’s Law of Multiples) my elder daughter Renee completely independently of me used “WTI WTF” in her daily market commentary yesterday. I’m so proud! She also raised the possibility of negative prices some time ago. Good call!

And I finish this just in time to bring you the final results. CLK goes off the board settling at $10.01, up a mere $47.64. CLM settles at $11.57, down -$8.86. The closing KM20 spread, $1.56.

Someday we’ll look back on this and . . . . Well, we’ll look back on it, anyways.

Print Friendly, PDF & Email

23 Comments »

  1. Hin Leong?

    Comment by SS — April 21, 2020 @ 3:27 pm

  2. Personally, I would look at who had open calendar spread positions and the ability to withhold storage in the prompt month. But that’s just me.

    Comment by Jim — April 21, 2020 @ 3:57 pm

  3. M/N spread volatility was nuts, going from -6 to -9.5 to -6 to -10.5 to -6.5.

    USO announced a midday change to m/n/q 40/55/5% from just m/n 80/20% … Lots of speculation around what USO’s NAV would have meant if prices went negative and if the exchange could intervene before then, since the holders have the implicit 0 strike put, prospectuses poured over. Let alone the implications of levered ETFs like UCO, SCO rebalancing end of day or the liquidation of the OIL etf.

    Interactive Brokers one loser, -88mm via K0, likely small hedge funds as what retail broker would allow retail to trade two days to expiry?

    Implied vol up from 100 to 205 yest to 320 today, although a % vol makes little sense in a Bachelier world.

    Comment by BluBle — April 21, 2020 @ 4:41 pm

  4. If it wasn’t so gut wrenching and disastrous for commodity houses and producers, I’d be enjoying the hell out of this from a purely academic standpoint. When’s an FCM going to implode, I’m on the short end of 10 days unless there is some magic money shoring up balance sheets from the FED via overnight repo. Ronnin was only the first…

    Comment by clf — April 21, 2020 @ 4:45 pm

  5. Where the rubber hits the road, I can now fill the tank of my vintage Beemer for about $90. Haven’t seen prices like this for about 20 years. Long may it continue!

    Comment by Ex-Global Super-Regulator on Lunch Break — April 22, 2020 @ 3:45 am

  6. Hi Craig,

    I am note sure the USO website is updated. It is written in their last report:
    “Commencing on April 21, 2020, because of extraordinary market conditions in the crude oil markets, including super contango, USO has invested in
    other permitted investments, as described below and in its prospectus. In particular, on April 21, 2020, USO invested in approximately 40% of its
    portfolio in crude oil futures contracts on the NYMEX and ICE Futures in the June contract, approximately 55% of its portfolio in crude oil futures
    contracts on the NYMEX and ICE Futures in the July contract and approximately 5% of its portfolio in crude oil futures contracts on the NYMEX
    and ICE Futures in the August contract, except when the front month contract is within two weeks of expiration, in which case the futures contracts
    held by USO will be rolled into the July contract, August contract and September contract. ”

    While on the webpage you linked to, there is nothing about August and the figures in their report:
    http://www.uscfinvestments.com/holdings/uso

    Comment by Etienne Borocco — April 22, 2020 @ 3:50 am

  7. This and the previous post are very interesting in the light of what is going on.

    At a personal level, I was in a dealing room during the 1987 market shutdown. LIBOR, a stalwart benchmark that would seem impossible to break, vanished in exactly the same way as you indicated in your own experience. Nobody knew, and hence flew blind. Consequently, I wonder if what happened isn’t a cascade effect?–particularly if prices lagged what was possible. I also agree that this is a “crowded exit”, something that I looked at with a colleague some years ago in the context of equity short sellers. Market microtheory would furher suggest that the depth of the market to act is potentially limited. As the futures here were very close to being quasi-spot, surely one needs to look both at their behaviour and the concommitant spot price behavior? Storage obviously is a key issue and your analysis does suggest a huge problem. But why didn’t market participants see it coming?

    Another point that suggests congestion or storage issues is that this is a purely futures driven issue for the WTI contract. As far as I can tell, while there is some contagion, Brent futures didn’t move in the same way.

    A final observation. It will be interesting to see the distribution of gains and losses for market participants caught up in this. It might help to untangle the causes.

    Keep up the insightful reporting!

    Comment by Peter Moles — April 22, 2020 @ 5:33 am

  8. A lesson I learned in the First crash of 1987 – MBS on April 1st when the current coupon GNM A opened down 2-3 points was that an OTC market closes when the dealers won’t answer the phone.

    Comment by Sotosy1 — April 22, 2020 @ 2:02 pm

  9. @Etienne. That doesn’t alter my point. They were out of the May. Historically they have always been in the nearby month. The change they made was to spread out positions in the 1st, 2d, and 3d months.

    It is plausible that these big shifts were partly responsible for the big decline in the CLM20 yesterday. I said as much in the post.

    Comment by cpirrong — April 22, 2020 @ 6:07 pm

  10. @BluBle–Crazy stuff.

    BTW, a little bird told me that the switch to Bachelier was something of a cluster.

    Comment by cpirrong — April 22, 2020 @ 6:09 pm

  11. @Sotosy1–We are old! I remember that day very, very well. I was working for an FCM and my main task was to design hedging strategies for banks with MBS portfolios. Long MBS/Short 30 year T-bonds. The basis exploded and my customers freaked.

    I tell that story in class, and say “That was a 3 scotch night, not a 1 scotch night.”

    There’s also the story about the Sollie trader who claimed he had the “perfect hedge” for mortgages. He got wrecked on April 1. Traders being the compassionate types they are, his colleagues had a gift waiting for him on his desk the next morning. It was a bonsai tree, with a little sign: “This is the only perfect hedge.”

    Comment by cpirrong — April 22, 2020 @ 6:13 pm

  12. @Peter Moles. Ah, another example of how liquidity evaporates when you need it most. And yes, there are definitely cascade effects–positive feedback loops (which always have negative consequences, it appears!).

    The storage problem was anticipated. What was not anticipated (but probably should have been) is the exploitation of that shortage on the 20th. Note that the contract went off the board on Tuesday at a price very similar to where it closed on Friday. The huge move on Monday was aberrant, and was clearly driven by the problem of closing out 100mm bbl of open interest when the threat of delivery of even a few million barrels (maybe less) would have catastrophic consequences.

    The CLK20 open interest at T-2 was substantially higher than in the previous four Ks. But available capacity to stop deliveries was far smaller. Not a good combination.

    Comment by cpirrong — April 22, 2020 @ 6:20 pm

  13. @Jim. Yes. Exactly what I was referring to when I mentioned that another way of exercising market power would be to withhold storage. A bear spread would be the way to monetize that.

    Comment by cpirrong — April 22, 2020 @ 6:23 pm

  14. too bad it’s hard to build new storage facilities…….and expensive too

    Comment by jeff — April 22, 2020 @ 9:01 pm

  15. Regarding the settlement price from Monday: do you have napkin based estimate on cost of hypothetical logistic operation involving trucking 2 million barrels from Cushing to Gulf Coast or East Coast and selling it there for one cent a barrel. Would it cost more than $35 per barrel?

    Comment by T — April 23, 2020 @ 1:07 am

  16. Perfesser, Sounds like Andy Stone. In my bundle of joy i had been moved to the parent bank while still being responsible for hedging, pricing and secondary marketing for the Mortgage bank. No power to set policy however. The idiot running the bank let the originators take overages. If the rate was 7.5 and no points, but they got the client to pay 1 point they kept it. This turned out sales force into bulls. They kept back apps until rates dropped and took the overage. No internal date controls at all despite my temper tantrums. I new they were holding back but when we set rates that Monday the morons showed up with 6x the normal weekend production. I just giggled! And got out my axe.

    Comment by Sotosy1 — April 23, 2020 @ 7:51 am

  17. Excellent posts, very interesting.

    “with going into expiry short being a lot more dangerous than going into it long.” – could you expand on this please? Would like to understand why?

    Comment by sness — April 23, 2020 @ 2:34 pm

  18. The solution is simple! Just put this guy in charge:

    Anders Åslund
    @anders_aslund

    Now Bloomberg claims that WTI had increased by $39.18 per barrel or by +104% to $1.55 per barrel. None of this makes any sense. Is the US able to organize an elementary oil market? Otherwise outsource it to more able people! This is serious.

    https://mobile.twitter.com/anders_aslund/status/1252430012835221507

    Comment by Ivan — April 23, 2020 @ 7:59 pm

  19. Thank you for your post and your analysis. I admit that I don’t understand everything, but I enjoy learning and trying to get what I can get out of it.

    Comment by Nadav — April 26, 2020 @ 6:55 am

  20. Hi Perfesser

    At some point, is the fund industry just going to shrug and decide WTI is not only no longer a proxy for global crude, but not even a good one for US crude?

    WTI went negative because Cushing: 500 miles inland, can hold about 3 days’ of US domestic oil demand i.e. prone to empty or hit tanktops on not much of a shock, and bottlenecked both in and out. That’s not an accurate picture of US oil demand – AFAIK the US continues to import North Sea crude.

    Comment by Green as Grass — April 27, 2020 @ 10:26 am

  21. Can you post a link to your daughters blog? A quick Google did not result in anything.

    Comment by Holmes Gwin — April 27, 2020 @ 11:11 pm

  22. @Holmes Gwin–Not a blog. She writes a daily market commentary for Tellurian. Thanks for your interest.

    Comment by cpirrong — April 30, 2020 @ 2:18 pm

  23. @Nadev–glad you find the posts informative.

    Comment by cpirrong — May 1, 2020 @ 12:02 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress