Risk-Management Principles for the Utility CEO

Board coordination is the key.
Fortnightly Magazine - June 2004

Board coordination is the key.

Many utility CEOs are happy to pass off risk-management policy to the CFO and the head of the trading desk. After all, with deregulation and re-regulation, collapsing spark spreads, hypersensitive rating agencies, and nervous investors, there is enough to worry about. So what's the problem? If the financial guys control and report the risks and profits and losses (P&L) within risk tolerances, why should the CEO be concerned about risk management?

The answer, in many cases, is that the board and CEO have not given specific strategic directives to those responsible for risk management. As a result, often they are disappointed when the trading results do not reflect the goals of the organization. On one hand, a trading group may show tremendous profits, while the board complains about the discount the equity analysts have applied to the company as a result of the trading risks. On the other hand, a trading group may appear to lose money on a stand-alone basis, while in fact the group had reduced corporate risk significantly. Many trading shops find themselves in this type of situation: They are urged to turn a profit, but are charged with hedging risk. The only way they win is if their trades both produce profit and reduce risk-an unrealistic expectation for any trading group.

This problem is exacerbated by trading reports that emphasize stand-alone P&L rather than the effect of trading on an integrated risk profile. Therefore, the CEO must set clear directives for trading off profit and risk that are interpreted and executed by the trading groups. A practical process for accomplishing this follows.

Principle #1: All risk management policy follows from the board's agreement with key principles.

Establish a set of key principles upon which the board and senior management agree. The debate about these principles highlights possible inconsistencies, the resolution of which leads to consistent risk-management policy. For example, it is common that companies do not want to supply collateral to a trading operation. However, if they want the trading operation to execute financial hedges, collateral is unavoidable. The company may, of course, decide if it is willing to provide collateral or not. However, the decision not to provide collateral must be accompanied by a prohibition against hedging using financial instruments, which might require posting collateral.

The board frequently hopes to achieve multiple objectives, higher profit, lower risk, fewer headlines, and the like. From a policy point of view, it is inappropriate to charge traders with the responsibility for multiple objectives. One cannot simultaneously, for example, maximize profits and minimize costs, unless revenues are fixed. By the same token, no one can both maximize trading profits and minimize enterprise risk. Yet it is not uncommon to find multiple trading objectives in a typical utility annual report or in a risk management policy document.

Principle #2: Management sets clear definitions, benchmarks, and objectives.

How often have utility CEOs said, "We don't speculate"? The statement is counterfactual in every company with a hedging or trading function. In fact, there is an element of speculation in every hedging decision: how much to hedge, when to hedge it, whether to leg-in to the hedge or execute all at once, what maturity to use, whether to use options or forwards, and whether to modify the hedge over time. These are all speculative elements that creep into hedging decisions. Words such as "speculate" must be clearly defined in order to carefully distinguish between permitted and unpermitted activities. Speculation is neither inherently good nor evil. Speculation is nothing but another business activity of the firm. If it earns more over time than it costs to support the business, including capital and risk charges, it should be maintained as a business. If not, it should be reduced to a minimum.

Also, the term "hedging," Merton Miller once said, is like an old coin whose face has been rubbed down by overuse. Hedging has one meaning for accountants, and another for economists. For accountants, a fixed-price fuel purchase is a hedge. For economists, unless the power is hedged simultaneously, it is likely not a hedge. Once again, the use of the word "hedge" in a hedging policy must be clearly defined to have any meaning at all.

Finally, risk reduction for its own sake has limited benefit. Yes, we probably want to stay out of bankruptcy. Beyond that, shareholders may prefer that we take risks when the returns justify it. Although this appears contradictory on the surface, it is not. Risk reduction may be sanctioned up to some minimum level, and subject to higher hurdles thereafter. This is not impossible to implement. One need only specify a two-part set of risk management principles consistent with this two-part objective.

Principle #3: Benchmarks are both specific and achievable.

What is a benchmark? In the case of a trading or spec book, one popular example is that one "must beat Treasuries," , earn a higher return on cash employed than earned on U.S. Treasury bills. A second popular example is to earn a higher return than Treasuries after risk charges. A third example is to beat the risk-adjusted return of other activities of the firm.

Each of these benchmarks may reflect alternative ways to use cash capital and risk capital that are specific and achievable. For example, the first benchmark is achievable by shutting down the trading operation and putting the cash in money-market funds. The second example is not specific, but does set a minimum standard that is achievable by doing nothing. The third is achievable only if the capital can be redeployed to earn a return for the firm elsewhere. If this is not possible, it is not an achievable benchmark.

EES North America

It is unfortunate that many management teams will direct the trading group to do only hedges and make money for the firm. Another popular directive is stay between 25 percent and 50 percent hedged, but do whatever you want in between those ranges to make money. The first is not achievable, since there is no trade that is guaranteed to make money for the firm, let alone do so while hedging. The second is ambiguous. Should the trader be measured on risk reduction, or P&L? Either way, management will be disappointed if both goals are not reached.

The second directive, offering a range of hedges, may be achievable, but introduces tremendous ambiguity and is therefore always inappropriate. Should the trading group be compared to a 25 percent hedge to see if they made money, or a 50 percent hedge at year-end? The "better of" benchmark is unachievable by definition. No money manager would agree to such a benchmark. The trading group is guaranteed to fail relative to at least one of these hedging endpoints, unless they trade their positions actively. Even if they trade actively and beat both benchmarks, they are starting from an underwater position. It is possible that they could be fantastic traders, but this benchmark could rate their value-added as negative.

Also, how does management deal with the possible illiquidity of the markets where they sell power? If a 25-percent hedge is achievable only at a significant cost, should the traders execute the hedge regardless of the price? Of course not. We need some pricing discipline in addition to a risk-management discipline. In this case, the benchmark is not specific and probably not achievable. Traders may be compared to an impossible alternative as a result. The board may be disappointed regardless of the outcome.

Principle #4: Each trading book is allowed only one benchmark and one objective.

Multiple objectives from trading must be separated into different trading books. For example, if you both want to hedge and to profit from speculation, you probably should have a hedging book and a spec book. Recognizing that the agencies hate terms like "spec book," you might want to call this by another name.

Notwithstanding the books' designations, each book may have only one benchmark. The hedge book, for example, should take as a benchmark the unhedged asset position, perhaps net of any marketing contracts. This is specific and achievable by hedging nothing financially. The goal of the hedge book should be either to reduce risk by a certain amount or percentage relative to the native asset position, or to reduce risk only when the expected cost of doing so does not exceed, say, 25 percent.

The first criterion is weak because it allows traders to execute any trade in the name of hedging, regardless of the price. Nevertheless, it is defensible if the alternative is a costly bankruptcy. The second criterion suggests that at year-end, one compares the native position to the combined position (native position plus hedges). The resulting portfolio performance is measured as its incremental P&L together with a risk rebate for reduction of risk. This allows the hedge to underperform the unhedged position, but only by a predetermined amount. One cannot ever look at the hedge book in isolation, since it shows results without reference to the underlying asset position. The proposed benchmark is specific and achievable, and sets a clear performance metric. It compares hedging performance to the alternative of doing nothing, and rewards traders for both profitability and risk reduction.

The spec book, unlike the hedge book, must be evaluated on a standalone basis. In this case, positive P&L alone is not enough to justify this business activity. The P&L must exceed risk and capital charges to justify continuing or expanding the activity. There are separate controls required to protect against a multitude of risks on this book, but these largely can be considered implementation details subject to an overall tolerance to bear speculative risk that does not disrupt the firm's capital plan.

Principle #5: Risk management objectives are linked to corporate financial objectives.

Many utilities link their trading risk limits to risk tolerance and choose their counterparty limits according to their credit ratings. This offers the appearance of trading risk control, but at the expense of missing an overall corporate objective. For example, if the corporate objective is to conserve enough internal cash to get through the planned capital expenditure cycle, then the trading controls should be chosen in concert with this corporate goal.

If the corporate goal is to maximize expected earnings before interest, taxes, depreciation, and amortization (EBITDA), subject to a minimum level of EBITDA, then the trading group should be measured on its contribution to expected EBITDA and worst-case EBITDA. If the corporate goal is to minimize the probability of insufficient cash to make fixed payments, then the trading goals should be expressed in terms of worst-case working capital requirements. In both cases, the link between corporate goals and the trading measures of risk causes the trading group to align with the rest of the corporation in achieving a common goal.

Principle #6: Scarce resources, like risk capital, credit capacity, collateral, and working capital, must come at a cost.

Many firms are short on personnel relative to their opportunities and can logically conclude that other things being equal, their allocation of people to an activity should correspond with the highest profitability per person. In fact, every scarce resource of the firm should be allocated to its best possible use. Furthermore, scarce resources should come with a price to ensure that managers use these resources wisely.

So why is risk capacity any different? If the company has a limited capacity to bear risk, then to induce traders not to squander this scarce resource, management must put a cost on it. Almost all banks now include risk charges for traders using a RAROC1 adjustment, along with many energy companies. Risk charges ensure that traders take risk efficiently. For example, if a trader goes long on gas by purchasing forwards or calls, if he is otherwise indifferent, he will choose calls to minimize his cost of risk capital, because they minimize risk to the downside.

Similarly, if the company is concerned about working capital usage, it can put a cost on it to induce traders to make trades with lower cash requirements or to seek cash funding of trading mark-to-market outside the firm's balance sheet. If the company is concerned about counterparty credit-risk capacity-not an unreasonable concern in today's markets-there should be a charge on counterparty risk to give traders the incentive to insist on collateral and mitigate counterparty risk where possible.

These charges have to be carefully thought out to prevent gamesmanship by traders. For example, a flat counterparty credit charge will cause traders to migrate toward counterparties with higher average risk than the fixed charge would indicate. Different games can be played with value-at-risk and other market risk statistics. For example, traders will search for positions, like spreads among very similar commodities, which have no VAR impact but nevertheless carry great risk.

Principle #7: The chief risk officer must be empowered and independent.

If you don't have a chief risk officer (CRO), stop stalling! Appoint someone at a high level of the executive hierarchy to this function, even if it is not full-time. The stakes are high, and the cost of failure is great.

The CRO must have complete reports both for himself or herself, and for presentation to management and the board. These include P&L reports, position reports (delta-equivalent and options separately), value-at-risk reports, liquidation value reports, stress testing results, counterparty risk reports, collateral usage forecast reports, working capital forecast reports, policy exception reports, and error reports. If one of your books is a hedge book, then the reports must integrate the underlying assets with the hedges and report on that basis.

Management sets the rules of the game through the policy/procedures documents. The control function referees the game, but the traders will optimize within those guidelines. Unfortunately, the CRO often is asked to be both the head referee and the coach of the game.2 Poor coaching results from poorly specified and communicated policies. Both poor coaching and poor refereeing can lead to unexpected and unwanted trading performance.

Principle #8: Most things are possible in risk management, except the recovery of past losses. Act now.

This is the saddest moment in a risk consultant's life. They always get the phone call after the big loss occurs, as if they knew how to get it back. Almost anything is possible in risk management policy. Setting policies and controls in line with corporate objectives, measuring results, and rewarding performance appropriately-these are stock in trade for the risk management consulting profession. But recovery of past losses is not possible.

The utility industry has seen disaster after disaster, but many CEOs have not yet focused on trading risk management as a strategic objective. The 1998 power spikes in the U.S. Midwest, the collapse of Enron and coincident disappearance of market liquidity, the agency loss of faith, debt-holder pullback, and re-regulation-if you have been fortunate enough to survive these calamities, are you bracing for the next one, or will you plan to set up your organization to better withstand these disasters in the future?

Perhaps uncharacteristically, the responsibility for integrated risk management now lies squarely on the shoulders of the utility CEO. Fortunately, for a small investment of time, the CEO can establish agreed risk management principles at the board level to ensure that the trading activities of the firm best reflect its overall strategy and corporate direction.


  1. RAROC = Risk-adjusted return on capital. The actual RAROC is computed by dividing P&L net of interest charges by value-at-risk from the trading activity. Typical one-year RAROC hurdles are near 25 percent p.a.
  2. See Shimko, "Derivatives in the Board Room," Public Utilities Fortnightly, June 1, 2002.

Articles found on this page are available to subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.