Run the Risk

 

Jul 17 - Power Economics

Using risk and return and the relationship of these two important measures can lead to superior decisions about changes to any energy- based portfolio

THE US ENERGY MARKET has gone through a period of unprecedented turbulence over the last few years. The California crisis in 2000/ 2001 was followed by the events of 11 September, 2001, then the Enron collapse followed by a general flight from energy trading. These events have led to an excess of supply in the market at a time when a number of new plants are in the development phase.

Many energy companies are re-thinking their current strategies. The drive to build robust trading and marketing infrastructures have given way to a flight back to managing assets themselves and/or load servicing. The vision of a 'virtual energy portfolio' has given way to the traditional view of managing the hard assets and using risk management products primarily to take advantage of these assets. A portfolio of energy assets could include any combination of the energy value chain (production, processing, storage, transportation, generation and transmission). These asset can be thought of as a group of interrelated components that do not always fit seamlessly together. Add the demand side to the equation (retail and wholesale) and you increase the complexity of the relationships even more. For this reason, risk management and trading is the essential glue that can keep it together. If the market is trending back to bricks and mortar, a company's assets and demand side businesses are the bricks and risk management is the mortar.

Portfolio optimisation is a concept that is growing in popularity at many of the integrated energy companies today. Organisations are beginning to realise that their combination of assets, demand and risk management products are unique and can potentially add additional value because of these interrelationships. Isolating any single portfolio component can lead to decisions that result in unexpected changes in the risk and earnings profile of the organisation.

An example

A subset of a fully integrated energy company can be used to illustrate some very important points about asset analysis and portfolio optimisation. Consider PowerCo, a small utility with 2.4 GW of generation and 700 MW of load servicing commitments in Texas. Exhibit 1 indicates that the generation portfolio is a mix of base load and peaking assets while the retail business consists of both residential and small business in a number of congestion zones within ERCOT.

While the expected supply and peak demand levels are roughly equivalent, their volumetric shapes across time are driven by very different fundamentals. The shape of generation capacity is a function of market factors as well as operational constraints. Base load plants run most of the time but they are still at risk of being forced out of services for varying lengths of time. Transmission constraints can also affect the amount of generation that can actually reach the markets the energy company is committed to serve.

Exhibit 1: Composition of PowerCo's Portfolio

On the demand side, weather is often an important factor. Extreme temperatures in either direction often lead to increases in the demand in the market. Exhibit 2, below, illustrates the expected and worst case monthly volumes from this portfolio. Each plant and all of the retail customers were simulated individually given their specific operating profile. The expected or average volumes for the portfolio indicate that the generation capacity far exceeds the needs of the retail business. This imbalance grows to a significant level under the "worst case" scenario.

In addition to volumetric risk, generation and load exposures carry a significant amount of exposure to spot price risk for power and fuels. Absent any long term purchase and sale agreement or hedge contracts, generation must acquire fuel to run the plants and then they must sell the power into the open market. On the other side, load servicing entities must acquire power on the open market in order to serve their customers. The fact that generators are long power prices and load serving entities are short power exposure makes load a convenient hedge for generators.

Exhibit 2: Monthly Electricity Position (Expected and "Worst Case")

This can be better understood if all exposures within an organisation are evaluated from a risk/return perspective. The first step in this process is to examine each exposure on its own merit and understand how the risk/return profile compares to all other components in the portfolio. Exhibit 3, opposite, provides an example of all elements of PowerCo. In this exhibit, Expected Earnings is quantified on the vertical axis. This amount is determined via Monte Carlo simulation by taking the average of one thousand potential scenarios of price and volume outcomes for each plant and load exposure. The horizontal axis is the risk around the expectation measured by taking the mean of all scenarios less the 5 per cent worst case in the distribution.

A benchmark risk/return profile of one is often used to compare exposures of differing size. Exhibit 3 summarises the risk/return profiles of each load and generation exposure for PowerCo. The coal plant is the largest exposure and it is also the most attractive as it is the farthest away from the benchmark line.

Exhibit 3 PowerCo's Marginal Rick/Return Profiles of Each Asset

Absolute risk/return of each exposure independent of all other elements of a portfolio can be deceptive. Perhaps a more important metric is the incremental risk of a portfolio. Incremental risk can be defined as the amount of risk a new position will either add to or subtract from the overall portfolio that is already in place. The same positions above can be re-summarised in this manner.

Exhibit 4, below, shows the significant risk reducing impact of the load deals on the overall portfolio. For instance the load deals that serve residential customers in NCent and SCent ERCOT have $37 MM and $26MM in expected earnings AND provide $13 MM each in risk reduction on the overall portfolio.

Exploring the framework

This framework of considering volumetric balance and risk/return profiles allows for improved portfolio optimisation and bid price decision making. Suppose PowerCo has decided to consider bidding for 650 MW of load servicing primarily consisting of residential customers in the coastal region of ERCOT. This load following contract is expected to significantly reduce the imbalance between expected generation and demand. Exhibit 5, (see next page) demonstrates that this deal will significantly reduce the mis-match of supply and demand.

Exhibit 4: PowerCo's Incremental Rick/Return Profiles of Each Asset

In many markets in North America, bidding for the right to service load can be very competitive. In these situations, it is important to develop a bid price that reflects the risk of the transaction as well as that transaction's contribution to the overall risk profile of an organisation. This is typically done by considering a number of factors including current forward prices, load shape, ancillary service costs, regulatory capacity charges, transmission charges, and transmission and distribution losses. Many of these factors can be clearly quantified. However, prices and loads can be quite volatile and this uncertainty must be factored into the pricing of any load. Exhibit 6 summarises the distribution of potential costs required to service this load.

Based on a Monte Carlo simulation of load and price, we would expect this commitment would cost $124 MM in energy and other costs to fully serve. However, we are uncertain as to the actual load volume and the actual spot price. Therefore, the simulation indicates that the costs of this deal could be as high as $253 MM or double what we are expecting. Risk managers often site the 5 per cent worst case and a reasonable lower bound. In this case, there is a 5% chance this commitment could cost more than $170 MM. The difference between the expected loss of $124 MM and the 'worst case' loss of $ 170 MM is $46 MM. This amount is generally referred to as the Earnings at Risk or EaR of the1 portfolio.

Recommendations

If this were PowerCo's only investment, it would be prudent to submit a bid that would return enough of an expected profit to reasonably compensate the company for the "risk capital" taken on by this commitment, 'risk capital' is the amount of financial risk (from market, credit or operational factors) that exists within a contract. In this example, one can define the risk capital as the amount of potential loss on the agreement. If we expect this contract will cost $ 124 MM and there is a 5 per cent chance it could cost more than 170 MM, risk capital can be defined as the EaR ($46 MM). While this amount is not cash invested up front, it does represent an amount of value that could be lost by a company. Any investment in risk capital requires a reasonable return. Traditionally, risk capital in energy companies carry a 20-50 per cent return requirement. Therefore, in this example, risk capital of $46 million would require an expected profit between $9-23 MM depending on the organisation's hurdle rate on risk capital. For purposes of this example, assume the organisation requires a return on risk capital of 25 per cent. This will require a profit of $ 11.57 MM ($46 MM in risk x 25 per cent).

The bid price on this commitment would require $43.96 / Mwh to cover t\he expected costs. In other words, if PowcrCo were to bid $43.96/Mwh the expected value of this transaction would be zero. PowerCo would have to increase this bid by $4.22 to $48.18 in order to generate an expected profit of $11.57 MM. The risk distribution would remain the same. However, it would have shifted enough into profitability in order for the deal to compensate the company for the risk they are taking on.

Bidding risks

Competitive bidding in the US often precludes a bidder with such a large risk premium from winning any deals. One common approach is to determine the optimal hedge for a given opportunity in order to generate a lower risk profile that requires a much lower risk premium. Fixed volume, monthly forward contracts are generally available for hedging purposes. These contracts do not completely immunise the trade because of the volumetric uncertainty embedded in a load following transaction. However, it is possible to significantly reduce the risk on the deal.

Examples

One method is to determine the optimal hedge required to minimise the risk around this load commitment. By defining a set of fixed price forward contract, much of the market risk on this position can be eliminated.

By reducing the overall risk on this trade from $46 MM to $5.5 MM, PowerCo has reduced the risk adjusted margin requirement from $4.22 / MWh to $0.50 / MWh. The overall bid price can therefore he reduced from $48.18 / MWh to $44.46 / MWh. However, there is a significant flaw in this approach. If the load deal itself is being hedged, it will have no ability to hedge the existing portfolio's risk exposure. A better alternative would be to consider the risk reducing capability of this trade on the overall portfolio. This approach could potentially allow for bid pricing that is lower than current market prices.

The first case is the New Load Deal priced with a 25 per cent charge on the overall risk of the deal without consideration of any potential hedges or risk reduction value on the overall portfolio. This requires the $4.22/MWh increase in the bid price.

The second example illustrates how someone could potentially bid below market prices. In this case the deal is considered with the existing portfolio of exposures. The new load deal carries a negative incremental risk of ($30 MM). That is to say that it provides $30 MM is risk reduction on the overall portfolio. As such, a bidder might be tolerant of some level of expected losses on the individual trade in return for an overall reduction in their portfolios risk profile. If a 25 per cent RAROC 'benefit' were allowed to be incorporated into the bid price, the bidder could bid for this load at a level that is $2.65/MWh below current market prices. This would certainly increase the bidders likelihood of winning a deal in an auction process.

Exhibit 7: Impact of an optimal hedge portfolio on the 650 MW load deal

Exhibit 6: Overview of the 650 MW load servicing commitment prior to setting a bid price

Exhibit 5: Pro Forma monthly electricity position if 650 MW load deal is added to portfolio

The final two lines in the table show how a load deal that is subsequently hedged with an optimal level of forwards have little or no impact on the overall portfolio. As such, their bid price would be something around the current market curve when considering the trade by itself or in conjunction with the rest of the company's exposures.

Conclusion

Risk and return are the foundations for effective risk management at the individual trade level and across an entire portfolio of energy based investments. By using a simple example of a typical utility with generation and retail assets, it has been demonstrated that a focus on the individual asset can often lead to ineffective risk management activity. Simple risk and return measures allow any portfolio manager to quickly understand the impact of the addition of an asset or exposure. These techniques should be employed at any company, whether it is a 2.5 GW utility or an integrated energy company that controls a range of assets types. Risk management allows a user to optimise the risk profile of any aggregate portfolio. To focus on a subcomponent without considering the risk and return relationships of the whole organisation is, at best, a waste of effort. At its worst, it is potentially a risk adding exercise that could ultimately lead to significant losses.

By Andy Dunn

RISK CAPITAL

Biography

Andy Dunn is a partner at Risk Capital, an energy risk management consultancy based in New York. Andy joined the company in September, 2002, and works out of Denver. Andy previously worked as vice president of risk management services at c.Acumen. Prior to that, he spent 11 years at PricewatcrhouseCoopers, mainly working in risk management. Andy can be contacted via telephone on: (+1) 303 293 3057, or via email on: dunn@e-rcm.com

Copyright Wilmington Publishing Ltd. Jun 2004