E.ON, is a large European electricity and natural gas company. It manages its trading operations as a separate profit center. The traders are both responsible for optimizing E.ON’s own generation and other assets–sourcing inputs cheaply and maximizing the value of outputs–and also for running a proprietary trading book.
E.ON just announced its second quarter results and was very candid about the fact that the outlook for the trading unit this year is not good. Indeed, it’s forecasting a loss in trading for the full 2011 fiscal year.
So let’s look back at last year, FY2010, when the trading unit recorded a profit. The Annual Report details each unit’s return on capital employed (ROCE). Here’s what it shows:
The chart shows the reported ROCE for each of the separate units as well as the ROCE for the group as a whole. Not shown is the ROCE for the corporate center. But EON doesn’t report an ROCE for the trading unit. Why not?
The ROCE is calculated as the ratio of the Adjusted EBIT and the capital employed at each unit, and these inputs are reported for each unit. In addition, each unit is assigned a cost of capital, and the realized ROCE is compared against the cost of capital to determine the Value Added by each unit.
Except trading. In 2010, the trading unit showed an Adjusted EBIT of €1,196 million. But the Annual Report doesn’t show any value for the capital employed in trading, and no ROCE, no cost of capital and no Value Added figure. Instead, the Annual Report reads:
Due to the structural particularities of the trading business, Energy Trading’s ROCE and value added have very limited information value and are therefore not included here.
What’s that?
If the trading unit is being managed as a profit center, shouldn’t management be evaluating it based on profitability? Does it earn it’s cost of capital? Certainly any proprietary trading desk ought to be evaluated that way.
Trading units within end-users have a checkered history. A key problem has been evaluating performance. The traders all claim that they are adding enormous value. But often times one discovers that the trading units are not being charged appropriately for the capital that they employ. They lean on the rest of the company’s balance sheet without ever explicitly acknowledging how much capital they require. My paper on Constellation Energy’s crisis in 2008 discusses this general problem, and highlights this as a key factor in the crisis that hit many electricity company trading units after Enron’s collapse in 2001.
Does E.ON really know how much capital is required for the trading unit? What is the unit’s cost of capital, and how much value really is added? I wonder.
Interestingly, E.ON’s calculation of the Group ROCE seems to imply that it attributes zero capital to the trading unit. The Group Adjusted EBIT equals the sum of the Adjusted EBIT from all of the units, including the Adjusted EBIT from the trading unit. The Group capital employed equals the sum of the capital employed from all of the units, including implicitly a zero capital in the trading unit. The only way for the Group ROCE to make sense and for the separate unit ROCE’s to make sese, is for the capital employed in the trading unit to be zero.
Hopefully, E.ON’s 2011 Annual Report will provide a clearer presentation that allows investors to evaluate the profitability of the trading operations.
Betting the Business 
3 Comments
Be careful about thinking about the capital employed as limited to the explicit margin that is posted. Often times a company may be able to trade derivatives without posting any margin, if it has a strong balance sheet and its counterparties are willing to take its credit risk. But that doesn’t mean it hasn’t used up capital making the trade. When it leans on its balance sheet, it is using up scarce capital that could support a different trade or a different line of business. There is an opportunity cost. All bank trading desks know this. Banks impute to the desk a capital charge based on the risk of the desk’s positions. Companies like E.ON should be doing the same thing, and they could report in their Annual Report the imputed capital dedicated to the trading unit.
A related issue is that what the cost of using margin is; brokers and exchanges pay interest on margin, usually close to the “risk-free rate” (which is admittedly low now). For a company that keeps a certain amount of cash lying around anyway, the cost of holding some of it in interest-earning margin is 0, or close to it.
It could be the case, as generally with energy trading, they are employing derivatives in their hedging and prop trading books. Technically this would mean capital employed is equivalent to the margin requirement required by counterparties.
Their concern may be, and I only guess, that such a capital employed & ROCE figure would not readily translate into something comparable to the rest of the group. Since margin requirements can be very small and varying. So for example, a 100mio OTC contract may be margined at 20% and generate 5% return, giving a theoretical 25% return on capital employed.
Another issue is the appropriate proxy to cost of capital. This does not have a ready equivalent in a trading book but my personal predilection would be to use a risk capital figure. Well managed trading desks have independent risk managers who would charge a portfolio with a risk capital weight based on their own risk models, such as but not limited to VAR.
It certainly would be appropriate to compare risk capital weight to the margin capital employed even if neither can be compared to the rest of the book. Banks publish their risk capital figures so why not utilities?
Other questions would be: Do they mark to market their portfolio? Do they limit exposures against their internal risk capital figure (trading solely to the limit of your counterparty’s margin requirement is a classic trading desk blow up scenario)? Could they publish their risk capital model and does it incorporate liquidity and counterparty risk? And if not, why don’t they have one?
One Trackback
[…] trading work for […]