I love these types of stories. Pull out a big Wall Street pot and mix in equal parts of leverage, some win at all cost trading and an esoteric financial instrument. Kick in a dash of regulatory incompetence and you’ve got yourself one delightful tale about how someone inevitably blew up their book and caused some turmoil in the broader market.
We see these messes pop up every so often and they usually play out the same way: (i) someone has a “brilliant” trade that almost always includes derivatives or a structured product; (ii) they get caught flat-footed as the market turns against them; (iii) they try and trade their way out; (iv) other market participants catch wind of their troubles; and (v) the book blows up and regulators step in. The only thing that really changes is the asset class. Well gather round, folks: today we’re talking about power markets.
I’m not sure if the general population understands how wholesale power markets even work, so I suppose a quick intro couldn’t hurt. As a result of deregulation in the 90s and early 00s, two distinct market structures developed: (i) traditional, regulated markets in which a single utility company is responsible for generating and transmitting power to both wholesale and retail end users and (ii) RTO/ISO markets where private companies own the generation plants and transmission lines of a particular region. The generation companies participate in auctions, run by the RTO/ISO, for the ability to sell wholesale power into the market based on the energy efficiency of their plants. Once they win the auction, they’ll pay transmission companies a toll to use the grid to deliver the power. In layman’s terms, power generation companies say:
“Hey – we can provide xx megawatts of power at this price because we have a brand new, natural gas powered plant. The other guy who burns coal can’t come close to our price so we should win the bid.”
And that’s exactly what happens. The company with the new, more efficient plant is awarded the contract and is now obligated to provide a fixed amount of power into the system, in the future, at a certain price – done and done. In a vacuum the system works perfectly – but the world doesn’t operate in a vacuum. Sometimes power demand is higher than normal (like during a heatwave) so the system becomes congested, leading to higher tolls for the power plants and eroding their profits in the process. Well here in America we like profits, so the RTO/ISO markets decided to develop a financial product to hedge against fluctuations in the tolls, which is where our friends at GreenHat Energy come in.
I don’t want to get into all the nuance of what happened with these guys but here’s the skinny. The two traders, Andrew Kittell and John Bartholomew, took a look at historical congestion trends in the PJM market (the largest RTO/ISO in the country) and decided not only would that historical pattern continue to play out, it would be exacerbated into the future. They then went out and amassed a huge portfolio of hedges that would payout were their thesis to come to fruition (mind you these guys weren’t hedging exposure, they amassed a naked, speculative portfolio of power derivatives). The problem for these two rocket scientists was that they didn’t factor grid equipment upgrades / repairs into their congestion models. The aforementioned upgrades reduced congestion rates materially over the last few years, slowly eroding the value of their portfolio until 2018 when the floodgates opened.
According to RTO Insider, the derivative portfolio represented a hedge on 890 million MWh of generation – an absolutely massive financial liability for a fund of their size. The value of the portfolio continued to deteriorate and is now facing estimated losses of $145M… and counting. It doesn’t even matter for GreenHat, either – they’re already in default and thus the losses will be absorbed by PJM and it’s participants (including retail customers). So how were these two guys able to use so much leverage to put together, and subsequently blow up, a portfolio and pass the losses on to everyone else? Put quite simply the problem was poor market structure and lax regulatory oversight.
There were two key weaknesses in the way PJM ran the market that brought GreenHat where it is today. The first is that market participants don’t pay for the contracts in full until settlement which is normally a few years out. GreenHat was able to operate on 100% margin and, as long as their trade was profitable at settlement, never had to worry about paying out on losses (kids, hubris is a terrible risk management strategy). Now is the point where you should be asking yourself about collateral requirements:
“Well, sure – they didn’t need to pay out on the contracts yet, but of course they had to post collateral as the portfolio fluctuated in value. Right?”
Wrong, actually! The collateral requirements are determined on a portfolio-level basis, not a position-level one. This past April, PJM changed it’s credit policy for market participants which would have resulted in a $60M collateral call for the GreenHat portfolio. However, market participants were given thirteen months to comply with the changes and rework their trading strategies. GreenHat used this window to amass enough new, profitable positions on 100% margin to negate the $60M collateral call coming down the pipe for their losing positions. Lipstick, meet pig. The window dressing worked for a bit until those positions moved against them as well, sewing the seeds of GreenHat’s demise and sending PJM and FERC scrambling to stop the bleeding. Regulators and market participants are currently kicking around different ideas on how to best equitably absorb at least $145M in losses.
I’m not exactly sure these guys are two of the most prudent fiduciaries out there. However, while I wouldn’t sign up to be a LP in their next venture, I have to point out that they operated within the framework that PJM and FERC had laid out for them. This will obviously force PJM and the rest of the RTO/ISO markets to restructure their credit policies going forward, hopefully they can figure one out that works this time.