This ain't normal

[solarbird]

ETA: The data is right, but is incomplete data. See below.

Okay, so, historically, most of oil traded for US consumption is not traded directly on the NYMEX; that's the spot market and while it does set the price, it's not the bulk of the oil - it's the tradable oil, and it's about 3%-5% of US consumption overall, which is to say, under 1M barrels/day.

That's historically, anyway. Check out trading volume over the last month and a half:


From Marketwatch.com's graphing tools

That's original except for my highlighter-coloured adds (two arrows, some obvious text). The last day's volume on that chart is actually higher than shown in this view, the chart isn't complete for today yet, because volume data lags a bit.

We're seeing trading volumes that exceed total US consumption and continuing trading volumes twenty times normal in the face of rising prices. Now I can think of several things this could mean, from day-traders swapping the same 1M barrels back and forth and back and forth (maybe this is one of the places those liquidity injections have gone) to demand destruction from contracted-delivery cancellation freeing more oil for the trading pits (but not that much, sorry - and particularly not in the face of rising prices) to China moving in to buy oil off the US spot market directly (in which case, where's that oil coming from? US inventories have been falling sharply but not that sharply) to significantly more dire interpretations. But really I just don't know.

But unless this data is wrong, something is going on. A one-day error showing a massive volume pop I'd attribute to data error. But I can't do that for this many days in a row. So... what the hell?

ETA: Since I've had a few people raise the "it's just the end of the year" response (two in IM, one below): please note that this is a two-year chart. This isn't normal end-of-year trading. For that, look at the tiny bip at the end of November '06 in this chart for comparison. That's how far down the rest of this chart has been compressed by this year's data. If the data presented here is correct, this is very, very different.

ETA 2: Y'know what's even weirder? Nobody else seems to chart this. At least, not the places I use, and not the places I found in Google that I can get to without being a subscriber. That means I can't validate (or invalidate) these numbers. It also means some other things. See comments (linked) below for details.

ETA 3: I got explorative with their URL parametre passing and was able to get data back to mid-2005, but no further. Nothing like this showed up EOY 2005, either - not even the little bump you get showing up EOY 2006. But that's throwing my own parametres at their graphing software so while I'm reasonably sure given everything that it's right, I'm less confident. Also, 26 December volume numbers are still being reported as 5 Mb, which is still crazytalk high, but if final is also a dramatic comedown from the previous month or so, which makes me ask okay, why? Is somebody on vacation, or is somebody done setting something up?

ETA 4: This does not entirely jibe with what I have read previously (including yesterday), but that could have been wrong or I could have misunderstood before. However, according to a commenter over on The Oil Drum, Marketwatch's charting, while accurate, only shows contracts expiring in a single month, which is to say in this case February, which makes their numbers look very strange. You have to have tools they don't offer to make volume numbers across all months appear in a chart. (And since theirs are the only tools I have that show it at all...) They also say that spot-market volume has gone up dramatically in recent times, but not anything like this chart would have you believe. The part that doesn't line up is that I've read (from multiple sources) that the vast majority of actual pricing follows the NYMEX but is not actually traded on the NYMEX - that's where I get that 3% to 5% number - so if these are the actual volumes on a monthly basis, that implies that most oil has actively moved to this market.

Comments

[technoshaman.livejournal.com]

The Chinese have been at economic war with us for decades. I don't have any proof, but I would suspect them of something like this.... next thing you know they'll be selling us gas, too. And once we take that hook, we're had.

[solarbird]

China isn't going to be selling anyone gasoline, or anything else oil-related. They're locking up contracts with Iran and anyplace they can deal with in central Asia for domestic use.

[solarbird]

For that matter, China is having problems with spot shortages, particularly of diesel. (This has been reported by many sources.) They're not in any position to export anything in fuels.

[pstratt.livejournal.com]

It's probably end of the year big oil company trading. There's a poster on diehards.org that works in a large oil company trading room. They continuously are looking for various trades geographically, time based or by type to enhance returns. So they may sell gas in one area and buy diesel in another if the prices are right.

There is also the big end of year rebalancing. If an oil company ends the year with less stock than they began they have to pay taxes on the difference. So the big rule is never have lower inventory at the end of the year.

Paul

[solarbird]

But I posted a two-year chart. Last year's end-of-year is lost in noise compared to this trend. See that bip at the end of November '06? That's last year's version of the phenomenon you describe, compressed down to almost a flat line.

[pstratt.livejournal.com]

I see what you're saying. It could be commodity futures contracts then. We are now getting beyond what I know. I do know they play around a lot with futures. Theoretically, as in I may be blowing smoke, futures contracts may count as "stock".

This is really common with commodity mutual funds where they use "collateralized commodity futures" (CCF). What the fund does is use treasuries or TIPS as collateral for buying commodity futures. That way the fund value is based on the commodity futures and it can pay a yield based on the treasury yield minus the futures cost.

In the case of an oil company they can lock in future oil prices and even use leverage if they want to make bets on prices. I don't think they use insane amounts (10x) of leverage like the Amaranth hedge fund that collapsed over natural gas futures.

If you look around on some commodity sites you might find an answer.

Paul

[solarbird]

Regardless of any of this, we're still talking 15-30 times normal volume, even year-to-year. Either something highly atypical is happening, or the data is wrong.

(If I understand this market correctly, and I may not, these are all technically futures contracts, with February delivery and expiration, but they're essentially universally exercised so can be counted directly as barrels from a trading count standpoint.)

And, really, nobody seems to be talking about this. I mean sure, somebody must, but I'm not seeing who right now. It's a financials matter so the oil sites aren't really into it, and I haven't seen anything on my financial sites. Some mild chat about rising oil prices and whether they'll hit "triple digits" next year, but I haven't seen any talk about volume.

[sianmink.livejournal.com]

Could be all the groups that usually invest in savings&loan jumping ship and buying oil instead. you wouldn't see that last year.

[solarbird]

I am told by day-traders that playing with shorts in the commodities is a great way to have your ass handed to you in the shape of a hat. So maybe, but... damn.

[sianmink.livejournal.com]

The fact that no news source I can find mentions this though, is terribly odd.

[solarbird]

I know. It's freakish. Marketwatch.com's charts are done by their subsidiary corporation bigcharts.com, who according to Google seem to be pretty well respected and appear to be established as a reliable data source for in-depth charting. (I have not seen a history of inaccuracy from them personally, either.) Which lends credence to the idea that this is actual data and not noise.

But nobody else lets you chart this stuff. (See comments further below.) Maybe nobody's paying attention.

[solarbird]

BTW, I played around with their URL parametres a bit, and can get data this way going back a little further, to mid-2005, and (again if accurate) then December 05 doesn't show this behaviour either. Or anything like it.

[maellenkleth]

Guesswork from prior knowledge of Asian supply situation, since will not have access to hard information until Friday: spot traders may be hedging against possibility of dedollarisation of crude oil at end of year. Quite concur with your analysis that Chinese are incapable of affecting American spot oil -- their shortages are real, ongoing and worsening.

[solarbird]

spot traders may be hedging against possibility of dedollarisation of crude oil at end of year
...that would take an assortment of people by surprise. But aside from the Saudi clerics thing, I'm not seeing a strong support for that.

The US dollar was down fairly sharply against the basket today, (77.25) but is still not really near recent lows, and came off its own daily low at end-of-session, and is, as they say, "testing the 100-day moving average." (The 50-day has been broken and is in fact rising.) So while technical trading is fortune-telling, it doesn't look like the currency traders are along for this ride.

(Plus, that's a hell of a short play, in the sense of, "where the hell are we going to get that oil for delivery if people actually exercise these?" The volumes are... freakish.)

[mattsnaps.livejournal.com]

If a speculator wanted to take a position in oil, is this the only game in town?

[solarbird]

It's not the only game, but it's certainly the shortest game. At roughly 15,000,000 bpd average at roughly $90/barrel average for 30 days, that's a US$40.5 billion dollar position, which means that someone is all-in, unless it's a sovereign wealth fund acting on the quiet. And you really have to ask exactly where anyone is gonna get that crude if they actually have to cough up the actual oil, because that kind of spare production capacity is simply not in the system. If someone is selling this kind of quantity of shorts, then that's the kind of bet you really, really, really shouldn't be making, because the destabilisation potential is awesomely high.

Y'know what else is weird? Nobody seems to chart this except Marketwatch. They kinda don't do it on purpose, I don't think - reason you can get to it is because they offer volume charts on everything because of their one-size-fits-all web programming and throwing whatever data they have into the standard forms and graphing engines. This means you end up with a lot of N/A coming up on volume charts ("This exchange does not provide volume data"), since many exchanges don't give you, say, interday volume information. But if they have the data, they'll throw it at you. CNN Money, CNBC, the Wall Street Journal, Barcharts, the New York Mercantile Exchange and so on - they don't let you chart this stuff because they have special more-appropriate-to-specific market charting pages for things like, well, commodities. And traditionally nobody cares much about volume in commodities trades, only price, and net demand/net supply, which in oil in particular is monitored very differently, since the spot markets were 300,000-400,000 bpd typical daily markets through 2006, and afaik not dramatically more in most of 2007. But now, if this data is right, maybe they are quite a bit busier now.

This is weird. The only reason I charted it was because I was charting something else oil-related and it came up, and I looked at it and went, "...well, that ain't normal." Is it even possible nobody has really noticed? That seems freakishly unlikely. But on the other hand, the Nymex is not the most transparent market in the world, and almost nobody even gives you this chart as an option. Maybe nobody's looking.

[measurethrice.livejournal.com]

If the spot market sets the price for the big volumes, perhaps the big suppliers have realized that they could drive the price up by trading the market up.
If they bought and sold large volumes at higher prices, they WOULD run the price up, and would sell tanker-loads at the higher price by buying and selling bucketfuls on the market.

The problem is that the colluding traders would need to buy and sell large volumes to make thier trading look like the real trading. The trading could not be so expensive that it would not be more than paid for by nearly doubling the spot price...

The only way that sort of manipulation would be visible is from the volume, and it's the only reasonable explanation for what you're showing. Once the market established a new trading range (the traders that saw the market was high and sold short all jumped out thier windows) they stopped playing with the market.
Remember Enron. There are no ethics in this business, and anything that makes money gets done, whether it's legal or not.

[solarbird]

See ETA4 above, an explanation has been proffered elsewhere.