Barclays' Power Traders Manipulated Markets To Build Their Trading Skills And Self-Confidence

Author:
Updated:
Original:

Yesterday the Federal Energy Regulatory Commission ordered Barclays and four of its traders to pay $488 million for manipulating energy prices by doing basically this:

  • (1) Buying electricity with medium-term swaps,
  • (2) Selling electricity with short-term forward contracts, and
  • (3) Buying electricity in the spot market.

And vice versa (switching buys and sells). The idea is that since the swaps in step (1) settled based on the spot price in step (3), Barclays can manipulate the value of its swaps arbitrarily high by just overpaying in the spot market. Like, buy 100 whateverowatts of swaps at $1 per WW, then buy 5 WWs physical, pushing up the price to $3/WW, and you've spent $15 (5 WW physical x $3/WW) to make $200 (100 WW swap x [$3 - $1]).

This is nonsensical on first principles, though that doesn't make it wrong; lots of true things are nonsensical on first principles; that's like a feature of first principles. When you lay it out like that, though, you can see the oddity, which is that FERC thinks step (3) (buying physical) pushed up the price, but thinks step (2) (selling like next-day electricity) did not push down the price. If this worked then you could replicate it in any market, like:

  • Buy a derivative betting that price of X will go up.
  • Sell a ton of X, having no effect whatsoever on its price.
  • Buy back a ton of X, pushing the price up.

Obviously if your trading on one side of the market has a huge effect on prices, and your trading on the other side has no effect, you can use that fact to manipulatively make money. It's a perpetual motion machine.1

Like I said, though, none of that makes it wrong: FERC and, to be fair, Barclays really seems to think that it's a feature of power markets that Barclays' trading in one direction affected prices and its trading in the other direction did not. You can read the FERC's order here and its press release here. If you want more background, which is unlikely, you can read the FERC's October order to show cause and proposed penalty here; we talked about it a bit at the time and I was as puzzled then as I am now. Here's Barclays' response to the October order, as well as someearliersubmissions. I cannot in good conscience recommend that you read any of this. Barclays is adamant that it intends to continue fighting this in court, which will be horrible.

The problem for Barclays is that its four traders here - Scott Connelly, Karen Levine,2 Daniel Brin and Ryan Smith - sent a lot of emails and IMs being all "let's manipulate prices," though unlike in the Libor cases the emails and IMs are fantastically boring.3E.g.:

In an email to her colleagues on January 31, 2007, Levine stated that she would be out of the office for a while and described her position—including her Balance of Month Financial Swap position in the SP to PV spread. She then stated that, “undefinedf we can keep the PV index up and the SP daily index down somehow that will be good to keep the BOM in.” OE Staff alleges that this statement “on its face is a request for her colleagues to trade Dailies to move the daily settlements for the PV index higher and the SP index lower to benefit her short position in the SP to PV spread, a BOM Financial Swap position which she had set forth in the same email.”

I will take their word for it that that is incriminating! All those letters, my lord.

Barclays response is not half-bad; the argument is more or less:

  • FERC weirdly cherry-picks the times that Barclays lost money on its physical trading and finds those manipulative, ignoring the broader pattern of Barclays' regular trading.4
  • Barclays physical traders were just doing their own thing, building a business, and learning how to trade power,5 not manipulating prices to help their swaps traders.
  • Their trading was pretty normal, without any of the indications (always hitting the bid, trading when liquidity was light, etc.) that they were trying to push up prices.
  • In fact, according to some analyses Barclays did, "without Barclays’ trades, the ICE index would have been only 0.18% (0.0018) lower in the alleged manipulation months," which really isn't that much.
  • Anyway they didn't have such a big swap position that it would make sense for them to manipulate the physical market.6
  • Basically all you've got is some emails, and those emails were just traders being stupid, you can't take them seriously, everyone likes to brag about how influential they are but that doesn't mean they are;7

I find that sort of compelling? It's a me-or-your-lying-eyes sort of question: if you believe in general background principles of market efficiency, then the FERC's theory seems pretty suspect, but if you believe the actual words Barclays' traders wrote, then they're in trouble. So Barclays' argument is that those words were wrong, and that its power traders weren't manipulating the market. It was just important to their self-esteem for them to think they were. Not sure it was worth the $1mm fines for each of them ($15mm for Connelly), but maybe? I'm sure feeling like energy-market big shots was a lot of fun for them.

Barclays, Traders Fined $487.9 Million by U.S. Regulator [Bloomberg]
Barclays to fight $453 million power fine in U.S. court [Reuters]
FERC Orders $453 Million in Penalties for Western Power Market Manipulation [FERC]
Order Assessing Civil Penalties [FERC]
Earlier Barclays / FERC things: Order to Show Cause, Barclays responses [FERC]
Earlier: Barclays Did Plenty Of Non-Libor Manipulating Too

1.You don't even need the derivative: just, like, start buying at $1, push the price up to $3, your average purchase price is like $2, sell at $3, you make $1 per whatever. That's not even really manipulation.

2.Incidentally your default assumption for any Levine should be "no relation." That is true here too. Same with Levins.

3.Are there exceptions? I mean, feh, there is this:

In an IM exchange between Connelly and one of his former colleagues, the former colleague describes that morning’s trading as “a shitshow” to which Connelly responded, “crazy—I love it.” He then went on to say that “your boy started crying this morning[.] [H]e sent me an [ICE] message—said he was calling ferc,” and then added, “lol.” In an email exchange later that day between Connelly and the same trader, Connelly’s former colleague asks, “you going to have fun with the index all month?” to which Connelly responds, “no—it isn’t going to affect much.”

That's not incriminating, it's just annoying. "Crazy—I love it." Also the FERC order has section heading "The Friday Burrito Incident," which raised my hopes, but it did not live up to its name:

Three days after the Commission notified Barclays that it was under investigation for its Western U.S. power trading, a newsletter called The Friday Burrito, distributed by a staff member of the Western Power Trading Forum, included an article noting the unusual patterns of physical trading. The article asked, “[w]hat the hell is going on out there?” and noted that “the worst thing possible would be one party trying to move the financial markets with large physical positions.” Before the next business day, Connelly wrote an email to the publisher of The Friday Burrito offering an innocuous explanation for the physical trading volumes, and urging him to “embrace the change . . . as opposed to being afraid of it.” Connelly consented to allow The Friday Burrito to publish his email, but requested that his and Barclays’ identities be kept anonymous.

4.This is sort of a weird defense because it sounds like "well it's okay because we only manipulated the market sometimes" but it's kind of right? Like the FERC treats as evidence of manipulation "the difference of Respondents’ trading behavior in the Manipulation Months versus the Respondents’ trading behavior in months where manipulation was not alleged to have occurred" but that sounds a lot like data mining. Like: if you sometimes make money, and sometimes lose money, and FERC chooses to look only at the money-losing months and say "you must have been losing money to manipulate prices, because otherwise why would you lose money so consistently?," then a reasonable answer is to say "we didn't lose money consistently, you jackasses; you just chose to look only at the money-losing months." That's not quite what happened but you can sympathize with Barclays' claim that

But OE’s cherry-picked “alleged manipulation months” include only 17 percent of the total months/markets covered by the alleged manipulation period. When one looks at the entire period under investigation, the alleged “pattern” disappears, and so does OE’s case.

5.From their 2011 submission:

Part of building a wholesale power business involves training junior traders to transact in various wholesale power markets. Like virtually every other power marketer, Barclays sought to train junior traders by having them trade instruments and products that created relatively lower risk for the company. That meant having them trade spot and short-term transactions that are subject to less market risk than longer term transactions.

They weren't manipulating markets, they were just practicing trading.

6.In their words:

Barclays lacked any before-the-fact knowledge or incentive on most days during the alleged relevant period. That is, in many instances a rational trader could not see, looking ahead, that the possible volume of day-ahead fixed-price trades, given the volume of Barclays’ financial swap position, would be sufficient to gain on the company’s financial swaps such that, assuming causation for the sake of argument, the gains on those swaps would exceed day-ahead power contract losses.

That's spelled out more in the 2011 submission but the numbers are mostly redacted. Basically they look at what kind of "ex-ante leverage" Barclays had (basically ratio of how much they'd gain on their swaps vs. lose on physical for each penny of price manipulation they did) and found that either (1) they didn't have any leverage (manipulation would lose them money on net) or (2) they did, but didn't trade any differently from when they didn't.

7.E.g.:

The other electronic IMs cited by OE largely consist of loose, ambiguous, and boastful statements that OE interprets as reflecting a belief by the trader that a trade could “prop up,” “support,” “defend,” or otherwise affect a market price, in some instances to support some other position. Typical of these is an IM in which a third-party broker asks Mr. Smith if he has sold his “index load” and then references a scene from the movie Braveheart: “its like that battle [scene] from Braveheart: hold . . . hold . . . unleash hell!!”

God.

Related