Companies should adopt a far more robust metric.
Andy Dunn advises energy merchants in North America, Australia, and Europe on earnings ar risk and other risk-management topics.
Market risk remains one of the most significant issues for gas and power merchants. The SEC requires disclosure of market risks in a company’s annual filings. However, the allowable metrics fail to communicate the type of information an investor actually can use to gain an understanding of the market risk embedded in a company’s business.
An “earnings-based” market-risk metric is an alternative to the existing “value-based” risk metric. This article provides a simple example that demonstrates how the metric works for an energy merchant that controls generation and wants to hedge that exposure using gas and power swaps.
One common “value-based” risk metric is value at risk, or VaR, which is well defined and reasonably understood in the power and gas sector. In fact most investor-owned utilities use VaR to satisfy the SEC’s market-risk disclosure rules. VaR is defined as the maximum reduction of value that could be experienced from the impact of a set of market risks for a specfic holding period given a selected confidence level. Generally, holding periods are one to five days, while the confidence level is 95 to 99 percent.
VaR generally is considered an effective market risk measure for financial institutions that primarily follow mark-to-market (MtM) accounting and trade in liquid equity, fixed-income, and foreign exchange rate markets. On the other hand, energy companies have a more complicated business and financial reporting model. VaR is a much less effective metric for them.
While most energy merchants have some type of trading and risk-management activity that requires mark-to-market accounting, the core of their business includes signficant asset and customer-driven, load-serving activities that require accrual-based earnings recognition. While these activities often are more risky than the trading and risk-management acitivity, they usually are not included in the VaR metric.
Energy companies should adopt a far more robust market-risk metric that captures both the impact from trading and risk management, and the earnings from their accrual-based business. Earnings at risk (EaR), profit at risk, and cash flow at risk are all references often quoted in public disclosures and the press. However, there does not seem to be a common standard around which these metrics are based.
One definition asserts that EaR is the maximum shortfall of earnings, relative to a specific target, that could be experienced because of the impact of a set of risks for a specific period given a selected confidence level.
• Earnings are forecasted using business rules that are consistent with the organization’s financial or management reporting model. Accrual and MtM processes are incorporated on the appropriate positions so that forecasted earnings are consistent with what is reported ultimately in the financial statements.
• Specific targets vary. Some use the expected outcome from a large number of potential scenarios. Others used the forecast from a business plan or required earnings targets stated in debt covenants.
• Risks typically include forward and spot price risks as well as operational risks embedded in the business (e.g., forced outages in generation plants).
• Specific period covers the reporting horizon, or a business cycle. Typically one to five years.
• Confidence levels capture the majority of the outcomes and typically are set at 95 or 99 percent.
The term “earnings” has no well-accepted standard definition as to what is included from a company’s income statement. It could be gross revenues, income from operations, or even net income. “Profit” is another common term without an industry standard definition. Earnings-based risk measures should be labeled in a way that is consistent with financial reporting.
Table 1 provides a good example of an income statement for a large energy merchant. Constellation Energy reports a number of operating revenues and expenses. Some of these revenue and expense categories are considered market- risk sensitive or, “stochastic” with respect to market prices, while others are not.
EaR models consider the distribution of potential market prices that would generate the stochastic revenue and expense items across the reporting period. However, EaR by itself is difficult to interpret. If a company were simply to report its EaR was $350 million, investors would not know what that means. However, if a company were to report that it expected its business would generate income from operations of $1,100 million next year and its EaR95% around that forecast was $350 million, then an analyst could understand the expected operating income and the uncertainty around that forecast.
Earnings-based risk metrics need to be measured against a specific earnings number. Therefore, the appropriate non-stochastic revenues and costs must be added into the metric to generate one of the common earnings metrics used in business today:
• Income from Operations;
• Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA);
• Earnings Before Interest and Taxes (EBIT); or
• Net Income.
Management should decide which earnings metric is most used by the organization’s stakeholders, and the related risk metric should be reported with it:
• OpINC@R – based on operating income;
• EBITDA@R – based on EBITDA;
• EBIT@R – based on EBIT; or
• NI@R - based on net income.
Cash Flow at Risk
Cash flow at risk (CFaR) resembles EaR in many ways. It relies on the same time series estimates of prices, assets, customer load, and derivative instrument pay-offs. However, CFaR focuses on the actual cash flows that occur during the process, while EaR follows the earnings recognition rules of the relevant accounting methods used. Many of the components that make up the standard earnings functions (e.g., MtM, depreciation, amortization, etc.) are treated differently for CFaR.
For example, derivative traders are subject to MtM accounting record changes in their portfolio’s value each month prior to settlement. For the purposes of CFaR, all transaction gains or losses are accumulated and reported as part of this calculation. No interim changes in value are accumulated and reported as part of this calculation. Furthermore, a robust CFaR metric would include cap-ex, corporate borrowing changes, and capital structure changes to track the risk on the overall cash position change within the organization.
For the purposes of this article, we will focus on the elements of earnings and cash flow that are similar. The following sections use an example of a simple gas and power portfolio to better define EaR and contrast it with VaR.
Consider a 1,000-MW tolling contract on a natural-gas-fired generation plant in ERCOT that covers the plant’s operations from February 2006 through the end of June 2006. For each hour in the period, net earnings are determined by the following formula:
Net Earnings = Plant Capacity * Max(0,Pee – (Heat Rate * Png ) – VOM)
To generate a range of possible power and gas price scenarios, Monte Carlo simulation is used to generate 1,000 different time series of daily forward curves and hourly spot prices across the entire measurement period. These prices are then used in the net earnings equation above to generate a distribution of possible net earnings. These results also are conditional on various operational events (random plant break downs) and constraints (start up, shut down, ramp rates, etc.). The distribution of results, set out in Figure 1, is used to generate an expected net earnings amount of $14.63 million and an EaR of $10.75 million. In other words, the toll expects to generate $14.63 million in net operating income. However, there is a 5 percent probability that net earnings could be less than $3.88 million ($14.63 million - $3.88 million = $10.75 million).
Suppose the organization that holds the toll believes $10.75 million of EaR exceeds its risk appetite. The group could reduce this risk using a set of power and natural-gas swaps. Table 2 summarizes the expected output from the plant, the fuel consumed and a set of hedge volumes used in this example. Note that the hedge is only designed to cover the volatile months of May and June, and that the amounts hedged are considerably less than the total volume expected. These volumes reflect the fact that there is a band of uncertainty around these volume expectations. The optimal hedge volume considers this uncertainty and optimizes the amount of market risk reduction possible using the available hedge instruments.
Table 3 illustrates the importance of using the appropriate market-risk metric. The hedge contracts reduce the market risk of the combined portfolio by $4.4 million using EaR. Under the traditional VaR metric, these contracts eliminate only $0.4 million in risk. The difference in these amounts is attributed to the assumed holding period for VaR (1 day) vs. EaR (6 months). In addition, gas plants can take advantage of spikes in hourly spot prices. Some of the risk embedded in this tolling contract is generated from these hourly prices. EaR captures this risk. VaR captures only the risk in the monthly forwards.
The risk metric used depends on the merchant’s primary activities and its portfolio of assets and positions. In some cases, both VaR and EaR should be used. Table 4 contrasts the two standard metrics and provides some insight into the benefits that EaR offers. As systems and standards are developed around this metric, one would expect that it eventually will become the risk measure of choice in the industry.