There's gold in them thar transmission lines. Photographer: Chris Ratcliffe/Bloomberg
There's gold in them thar transmission lines. Photographer: Chris Ratcliffe/Bloomberg

Two things that you can say about the Federal Energy Regulatory Commission, as a market regulator, are:

  1. It produces the most incomprehensible prose of any market regulator, and
  2. Its markets have an unusual tendency to be gamed in embarrassing ways.

I suspect these facts are related. I posit that almost nobody, including at the FERC or its various regional power markets, can actually figure out how those markets' rules work. So they work badly. And while you have to be very smart to figure them out -- say, at the level of Blythe Masters, or electrical engineering Ph.D. Alan Chen -- once you have figured them out, they become comically easy to game. FERC builds markets with so many bells and whistles and buttons and valves that some of the buttons end up having no function but to dispense money. If you can find those buttons, what you do is just keep pressing them until the FERC notices and gets mad at you and starts scolding you incomprehensibly.

Here is a weird story about Powhatan Energy Fund, a "small Pennsylvania hedge fund," run by identical twin brothers Rich and Kevin Gates, who are delightfully pestering the FERC because it's thinking about suing them for market manipulation for finding one of those buttons.

The Gates twins are, or ought to be, legends of feistiness. Rich Gates got into Michael Lewis's "Flash Boys" by ripping himself off. When FERC sent Powhatan its preliminary findings that Powhatan's trades (and those by a related fund called Huntrise) "constituted manipulative trading in the PJM market," Powhatan sent back the following response:

Your preliminary findings make no sense. Should you choose to proceed with a public notice against Powhatan and/or Huntrise, please be advised that they will respond publicly and forcefully.

That's it -- that was the whole response. Amazingly, it seems to have worked: The preliminary findings are from August 2013, the reply is from October, and FERC still hasn't brought a public case against Powhatan. (Its "investigation is moving forward," though.) But Powhatan has responded publicly and forcefully anyway, apparently mostly for fun. ("Going public 'was an insurance policy against our personal and professional reputations,' said Kevin Gates.") And it's trying to prevent Norman Bay, the FERC enforcement director who oversaw the investigation, from being named chairman of the FERC.

The story is, lorem ipsum dolor sit amet, consectetur adipiscing elit, wait, no, OK, focus, we are going to do this. The story is that Powhatan let Chen invest some of its money in the electricity markets, and he found a money-dispensing button. In the assortment of charges and rebates and bonuses and other miscellanea in the PJM regional electricity market, there was a payment called a "transmission loss credit": PJM charged electric customers money to cover the cost of line losses in electric transmission, and then paid that money to the people who supplied that electricity. But the FERC-imposed rules at PJM collected that money under one economically rational formula, and distributed it under another, arbitrary formula.

The result was that, while normally the system charges you for trading electricity (because you have to pay for transmission lines), at some predictable times it instead paid you to trade electricity. So Chen traded tons of electricity at those times. And since just getting paid to trade is a good business -- much better than, like, trading -- Chen took steps to minimize his economic risk from all that trading, selling electricity from Point A to Point B, and then buying it back from Point B to Point Just a Little to the Left of A. The result was relatively safe profits from not doing much.

Powhatan explains that Chen's trades had some economic purpose beyond collecting those payments, but it doesn't put much weight on that. Instead its main argument is that "a market participant does not employ a fraudulent device, scheme or artifice when it rationally responds to economic incentives and risks created by the market, even when that market may have a flawed design." And it points out that FERC forced PJM to use this particular flawed design, over PJM's objections. FERC put the button onto its market; Powhatan just pressed it.

This may sound familiar if you remember that time that JPMorgan paid FERC $410 million to settle its own electricity manipulation case. While JPMorgan manipulated different markets -- CAISO, the California regional market, rather than PJM, the mid-Atlantic one -- and in different ways, the basic story was the same. Electricity market regulators built a market with huge gaping flaws. Smart traders found the flaws. And then they very openly and straightforwardly, with no deception or subterfuge, exploited the heck out of them.

Why did JPMorgan agree to pay $410 million as penance for these exploits, while Powhatan, which hasn't even been formally accused of manipulation, has decided that the best approach is a scorched-earth campaign against the likely next chairman of the FERC? A lot probably has to do with the fact that the Gateses run a small equity hedge fund that used Chen as basically an outside contractor to trade electricity. JPMorgan depends on good relations with all of its many regulators. Powhatan can afford to stand on its rights; no matter how much it annoys the FERC, the worst that can happen is it gets sued for manipulation. And you get the sense that the Gates twins will positively enjoy that.

The recent focus in market-structure regulation has been retroactively punishing big institutions for conduct that was clearly legal when they did it. The dynamics are fairly obvious: The regulators in charge of market structure are embarrassed to admit that their rules allowed what now looks like pretty goofy conduct, so they pretend that they didn't. The big institutions that want to remain in those regulators' good graces don't have much choice but to play along. Compromises are made.

But the rules of the power markets, and of the equity markets, were exploited by small players as well as big institutions. Those players have a lot less to lose from fighting the regulators, and a lot more to gain from standing on their rights. Plus, they seem to have a lot of fun doing it.

(Matt Levine writes about Wall Street and the financial world for Bloomberg View.)

  1. Look, this may just be my biases; I am used to, say, tax-law prose but less used to electric prose. But, man. Here is how FERC describes the issue here:

    PJM divided the surplus line losses collection by all eligible MWhs, without regard to the type of transmission service or difference in cost of that service. Thus, every MLSA-eligible MWh of transmission reservation received the same per-MWh allocation, even if the reservation was for non-firm point-to-point transmission service at $0.67 and the per-MWh allocation was more than that. Some market participants noticed this, realizing that in certain hours the MLSA appeared to be higher than the fixed costs of the transmission reservation plus the knowable range of other charges by PJM. If a trader could identify a node pair where changes in the price spread between the DAM and RTM consistently produced a de minimis profit or loss, or create wash-like or sham pairs of transactions with price spread changes that netted to a zero or near-zero profit or loss, that trader could design and use UTC transactions scheduled against transmission reservations as a vehicle to receive MLSA. The MLSA, in turn, consistently paid more than costs related to scheduling the UTC and transmission throughout certain hours of the day.
  2. From Felix Salmon's review:

    At one point, for instance, Lewis tells the story of Rich Gates, a mutual fund manager being front-run by HFTs. Gates "devised a test," writes Lewis, to see whether he was "getting ripped off by some unseen predator." The test involved placing two orders, a few seconds apart: The first would be an order to buy 1,000 shares of a certain thinly traded stock at $100.05, and then the second would be an order to sell 1,000 shares of exactly the same stock at $100.01. Gates "was dutifully shocked" when he discovered the results of his test: He ended up buying the stock at $100.05, selling it at $100.01, and losing 4 cents per share. "This," he thought, "obviously is not right."
  3. Powhatan explains that the charge was based on marginal loss:

    The credits themselves arise because the Commission has specified that transmission customers should pay transmission rates that charge for losses on a marginal, not average, basis. These marginal loss charges send a more efficient price signal for supply to locate closer to load, but they also create an over-recovery of costs system-wide. When deciding how to allocate this surplus, the Commission avoided allocating the credits to transmission customers in a way that would offset the marginal price signal created by paying marginal loss charges. And when the Commission decided to allocate transmission loss credits to financial traders engaging in up-to congestion transactions, it did so, in part, because that outcome would not undercut the price signal sent by paying marginal loss charges. This reasoning necessarily recognizes the obvious fact that receiving transmission loss credits will affect commercial decision-making, just like paying transmission loss charges.

    And FERC explains that the credits were allocated differently:

    Transmission line loss charges are a component of the per-MWh price of electricity in the PJM market. PJM uses the marginal loss method to calculate the charges to cover these line losses, which over-collects the cost of the losses. Pursuant to section 5.5 of the appendix to Attachment K of PJM's tariff, MWhs of successfully scheduled trades associated with paid-for transmission5 in a given hour receive a proportionate share of the surplus collected throughout the entire PJM market for the hour. This distribution is known as the Marginal Loss Surplus Allocation, or MLSA.
  4. Not that safe? FERC:

    An analysis of two sets of paired trades -- Mt. Storm-MISO/MISO-Greenland Gap and East Bend-MISO/MISO-Miami Fort 7 -- revealed that, for the hours in which Chen made these trades up through and including May 30, 2010, the trades yielded positive returns after accounting for both the LMP differentials and costs only 2% of the time. Only after accounting for the credit from the MLSA did the trades yield a more frequent positive return -- that is, in 54% of the hours.
  5. That is:

    Dr. Chen employed a "spread trading" strategy in which he hoped to hit it big if one of the legs of his Trades did not clear. Consistent with this strategy, he frequently entered into Trades which did not have the maximum congestion limit, thereby intentionally increasing the possibility that one of the legs would be rejected -- exposing Dr. Chen and Powhatan to a greater possibility of profit (as well as a corresponding greater risk of loss).
  6. See pages 10-12 here.

  7. Powhatan also makes the rather topical point that high-frequency trading in the equity markets is in large part driven by the desire to capture rebates, that "in many instances, it is the rebate itself that makes the trades economical," and that the Securities and Exchange Commission is fine with rebate-capture strategies. Since the FERC's market manipulation law comes from SEC law, the fact that low-risk rebate capture is just fine in the equity markets implies that it's fine in electric markets too. The fact that low-risk high-frequency equity trading is less popular now than it used to be is ... well, probably irrelevant to that argument, but maybe a little awkward.

  8. That's Pennsylvania-Jersey-Maryland, though it also includes a bunch of other states.

  9. The JPMorgan stuff really was pretty amazing. They found eight different buttons to push, of which my favorite was the one where they figured out how to charge people $999 per megawatt-hour for electricity from midnight to 2 a.m. (verus $12 market rates at that very off-peak time), because CAISO's system basically just lost its mind at midnight.

  10. That link is about the fairly off-the-run Blackrock survey thing, but you get the sense that Eric Schneiderman, who thinks that high-frequency trading is "insider trading 2.0," has more planned.

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.

To contact the author on this story:
Matthew S Levine at

To contact the editor on this story:
Zara Kessler at