Gregory Meyer, in today’s Financial Times, reports that banks are scaling back their trading in U.S. wholesale electric power markets. In his companion article, he quotes me saying that
The banks had the balance sheet, but the reality was it was the taxpayers that were giving them the balance sheet. It’s not clear we want the taxpayer subsidising proprietary trading in electricity or even hedging in electricity.
I am very circumspect about whether power trading operations are as profitable as they are often advertised to be. Here’s one reason why.
The profit rate is a ratio, with income in the numerator and capital in the denominator. In most types of businesses, it is straightforward how to measure the amount of capital required. If a company manufactures washing machines, for example, it needs a factory and the cost of building that factory is its capital. It also needs some working capital and other items which may complicate the inventory of its capital, but the basic problem of measuring its capital stock is clear.
Proprietary trading operations are different, especially in the age of derivatives. Traders take positions which are risky without fully collateralizing each individual trade. The business operates to a large extent on credit extended by various counterparties. The counterparties are looking at the trading company’s balance sheet for confidence. How much capital—equity—does the balance sheet have to have for counterparties to be confident? If the trading operation were to be expanded, how much extra capital would be required? These are very difficult questions to answer reliably, especially when a trading operation shares a balance sheet with other businesses. In my experience, few trading companies work them through with any rigor.
Power trading businesses have an especially poor record on this. I was first confronted by this fact shortly after the Enron bankruptcy when I did a strategy consulting assignment for one of the major power companies with a trading operation. The traders did a presentation to our team to justify their unit. It advertised the large flow of income they had been earning, which they forecasted would continue. But their presentation only documented the numerator of the profit rate calculation, i.e., the income. It left out completely the denominator, the amount of capital required. When I asked for that, the traders were flummoxed. They simply didn’t have a capital figure. Actually, they did have a figure in mind: it was the miniscule sum one would get if one added up the cost of the various computer equipment, the rent for a trading floor, and so on. They completely overlooked the fact that counterparties felt confident in doing business with their trading operation because it shared a balance sheet loaded with power plants and other hard assets funded with plenty of equity capital. Their trading was made possible by the capital attributed to the other lines of business. The traders had no sense of how much of that equity capital was effectively dedicated to the trading operations. Initially, I imagined that this trading operation was an outlier, despite the fact that it was a leader in the industry. I soon realized that it was representative of the industry in the early 2000s—a fact which I detail in my article and study on a later case in point, Constellation Energy’s 2008 liquidity crisis.
I am confident that the problem persists. In 2011 I ran a blog post pointing out the curious fact that E.ON, the large European electricity company, reported income and capital employed for each of its separate lines of business, with the exception of trading. For trading it reported income, but no capital employed. Adding up the numbers one could see that the capital allocated to each of the non-trading units exhausted all of the company’s capital, so that implicitly the trading operations cost zero capital. E.ON’s subsequent reports don’t provide a similar capital breakdown by line of business. After I wrote that blog post, I spoke with a number of insiders at other European power traders who understood the point of my criticism of E.ON’s calculation, but also empathized with the difficulties any power company faces when trying to determine the right capital figure to assign to trading.
Here at Sloan, I teach a “projects” course in investment banking and corporate finance, and this past fall a major power trading company generously challenged a team of Sloan graduate students in that class with the task of measuring the capital required for their trading operations. The team had a good approach and insightful calculations, but the exercise highlighted how the task is filled with pitfalls. There are certain critical assumptions that must be made for which there is insufficient foundation.
Which brings me back to the problem of banks and wholesale power trading. As Gregory Meyer’s article notes, “Banks grew as power marketers following the collapse of Enron and withdrawal of merchant power companies about a decade ago.” One ought to ask the question, why banks? Why didn’t other power generators step forward to take the place of those hit by the turmoil of 2000-2002? The expansion of wholesale power trading at banks since 2000 is just a footnote to the larger expansion of proprietary trading a banks leading up to the financial crisis of 2007-2008. It is clear that much of this larger expansion of bank trading operations was subsidized, in part, by taxpayers in the form of a giant balance sheet put. The profitability, measured comprehensively, was illusory, just as the profitability of many merchant power companies in 2000 was illusory.
At the conclusion of his article, Gregory Meyer quotes Cody Moore, head of North American power at EDF Trading, the fourth-biggest wholesaler: “We have tried to market ourselves as someone who is willing and very capable of entering the market previously taken by the banks.” I wonder whether EDF has the internal calculations of profitability and capital dedicated to trading to justify this willingness. I’m mindful of the fact that EDF is largely owned by the French state.
