About 90% of City Light's power comes from hydroelectric resources. Hydro power is an inexpensive energy source, which directly depends on hydro conditions (i.e., amount of rainfall and snowmelt) and, thus, is very unpredictable. Water can be stored behind dams, but only to a limited extent. In general, water delivered to dams in a year is converted to power. City Light provides power to meet its customers' demand. If it produces more than needed, it sells the surplus power in the wholesale market. Similarly, if it produces less, then it purchases additional power in the wholesale market. Due to hydro variability, City Light could have 50% more power than needed to meet its customers' needs, or have almost no surplus in a given year. This uncertainty in hydro conditions poses financial risk for City Light and is made worse by volatility in prices in the wholesale energy markets when it sells or purchases power.
Wholesale energy markets have been transformed during the last two decades. Historically, most power was bought or sold for delivery day-ahead or hour-ahead and sales were executed between utilities. Today it is a commonplace to buy or sell power by entering into a forward contract, which is a supply contract for future delivery of a fixed quantity of power at a predetermined price, time and location. The power price specified in the forward contract is called a forward price whereas the price at which power is traded on a specific day is called a spot price. Electricity is a very specific commodity. It cannot be stored, depends on weather and economic conditions, and does not have substitutes (i.e., it is price inelastic). All of these factors contribute to extreme volatilities in electricity spot prices. Forward contracts are used to hedge against unpredictable spot prices. In addition, counterparties no longer include only utilities. Market participants can also be companies that do not generate power but rather market it (e.g., Cargill).
City Light's forecast of net wholesale power sales focuses on uncertainty in three variables: (1) energy sales to City Light customers which affect amount of power available for wholesale sales, (2) energy generated by City Light's owned and contracted hydro resources, and (3) wholesale market energy prices. All three variables are modeled but the last two have the largest impact.
Hydro Generation Volume
City Light's hydro generation forecast is based on the 70-year data set published by the Northwest Power Pool, a voluntary organization comprised of major generating utilities serving the Northwestern U.S., British Columbia and Alberta. This data set is the accepted standard used by almost all utilities in the Pacific Northwest. In its financial planning forecasts City Light assumes normal water conditions, which is equivalent to the average of the 70-year history.
Electricity Price Forecast
In its wholesale revenue model, City Light uses the forecasted price of natural gas to derive the forecasted electricity price. Electricity price ($/MWh) is a product of natural gas price ($/MMBtu) and market heat rate (MMBtu/MWh). City Light uses forecasted natural gas prices and market heat rates provided by Platts, an independent service that reports forward market prices. In the most recent Rate Study the 2010 average annual gas price is assumed to be $5.34/MMBtu, which is the average annual forward price reported on August 19, 2009, the date when the study began. Coincidentally, this price is very close to the average gas price for the last 10 years ($5.28/MMBtu).
City Light's wholesale revenue forecasting process is a result of over 20 years of development effort, and is arguably among the most sophisticated in the region. It is shown in Figure A.
It begins with expected hydro conditions to forecast expected power generation and combines that with expected customers' demand for electricity to calculate expected power surplus/deficit for the coming 18+ months. The available power surplus/deficit is the difference between City Light's generation and customers' demand. Forecasted electricity prices (as described above) and surplus/deficit MWh are then multiplied to produce a wholesale revenue forecast. These financial results are used to compute financial and risk metrics.
Figure A: City Light's Wholesale Revenue Forecasting Process
Traditional forecasting techniques entail making assumptions about key inputs, and then combining them to create a best-estimate forecast. The problem with this approach is that it doesn't account for the possibility of the estimates being wrong-a very real risk.
To capture these risks, City Light employs a method of forecasting called "probabilistic" modeling. In a probabilistic forecast, 2,001 scenarios are run by varying hydro conditions, customers' demand and market prices. Natural gas price distribution is built by varying gas demand and holding its supply constant. The heat rate distribution is built by varying regional hydro resources and electricity demand. Electricity price distribution is calculated by multiplying natural gas price and heat rate for each of the 2,001 scenarios. The 2,001 scenarios that have different hydro conditions, customers' demand and market prices are then grouped into a forecast distribution. The forecast distribution shows the range of uncertainty, including the good and the bad outcomes and their probability of occurrence. The forecast distribution can be averaged to make a traditional best-guess forecast, called an "expected value." Figure B shows City Light's wholesale revenue forecast distribution for 2010. The expected 2010 wholesale revenue is $120 million, which is an average of 2,001 possible scenarios. We can see that in 2010 with very low probability, City Light might need to buy nearly $15 million worth of power, or with extremely low probability it might sell $400 million worth of power.
Figure B: City Light's 2010 Wholesale Revenue Forecast Distribution