Utility Transition Hub^TM

Fuel Adjustment Cost Sharing Savings Model

Methodology

Definitions

• Expected fuel costs: Estimated fuel expenditures for a given term. Expected fuel costs can be derived from either forecasted or historical values.
• Actual fuel costs: Actual fuel expenditures in the given year.
• Sharing rate: The percentage of deviation between the expected and the actual fuel cost that is shared between the utility and its customers during the true-up.
• Deadband: A predefined range around expected fuel costs within which no cost sharing occurs. Fuel cost deviations inside the deadband are fully absorbed by the utility.
• Pass-through threshold (optional): A predefined deviation level beyond which fuel costs are fully passed through to customers, resulting in a 100 percent true-up.
• Plant-Specific Price Adjustment (PSPA): A plant-level price differential that captures persistent differences between the benchmark Henry Hub price and the actual price paid at a specific plant, reflecting transportation costs, location, and contractual factors.
• Expected PSPA: A forward-looking estimate of the plant-specific price adjustment, calculated as the median of each plant's historical PSPA values from the prior calendar year.
• Heat input: Total fuel energy consumed by a power plant during a given period, measured in MMBTU.
• Gross generation: The total electricity produced by a power plant before accounting for station service or losses, measured in kilowatt-hours (kWh).
• Heat rate: A measure of plant efficiency, defined as heat input divided by gross generation. Heat rate is used to estimate missing operational data and is assumed to remain relatively stable over short time horizons.
• Revenue share: The proportion of a utility's total electricity sales revenue attributable to a specific state in a given month. Revenue shares are used to allocate plant-level fuel cost savings to state-specific customers when utilities operate across multiple states.
• Customer savings: The portion of fuel cost reductions passed through to customers under a fuel cost-sharing mechanism, calculated as the difference between expected and actual fuel costs multiplied by the sharing rate and adjusted for state-specific revenue shares.

Methodology

The FACSS model estimates how customer fuel costs would have changed under a hypothetical fuel cost sharing policy. It compares expected fuel costs to actual fuel costs at power plants, applies the selected sharing rate to the difference, and aggregates results to the state level. The model uses five years of historical data to provide an estimate of expected fuel costs. At the beginning of each year, expected fuel costs are set based on the following formula:

\[\text{Expected Fuel Costs} = \text{Fuel Price}_E \times \text{Heat Input}_E\]

Where:

• \(\text{Fuel Costs}_E\) ($) = Expected Fuel Costs
• \(\text{Fuel Price}_E\) ($/MMBTU) = Expected Fuel Price
• \(\text{Heat Input}_E\) (MMBTU) = Expected Fuel Consumption

At the end of the year, based on actual operation costs, we have:

\[\text{Fuel Costs}_A = \text{Fuel Price}_A \times \text{Heat Input}_A\]

Where:

• \(\text{Fuel Costs}_A\) ($) = Actual Fuel Costs
• \(\text{Fuel Price}_A\) ($/MMBTU) = Actual Fuel Costs
• \(\text{Heat Input}_A\) (MMBTU) = Actual Fuel Consumption

To calculate heat rate, we use:

\[\text{Heat Rate}_{\text{Power Plant}} = \frac{\text{Heat Input}}{\text{Gross Generation}}\]

If either heat input or gross generation data is missing for a given period, we can substitute the most recent available value or the annual average, based on EIA 923's Schedule 3. The heat rate is useful for interpreting plant efficiency and estimating missing values where needed.

Some power plants serve customers in multiple states, but power plant-level data does not reveal each power plant's end user location. To allocate fuel cost savings accurately by state, we apply a proportional method to separate out a plant's operations across states. Specifically, we use the ratio of a utility's electricity sales revenue in the target state to its total electricity sales revenue for the given month. This ratio is then applied to the plant's total generation data to estimate the share attributable to each state.

\[\text{Revenue Share} = \frac{\text{Monthly Sales Revenue in Relevant State}}{\text{Total Monthly Sales Revenue}}\]

Actual fuel prices at plant are sourced from EIA 923 fuel receipt records. The model computes a weighted average:

\[\text{Avg Fuel Price} = \frac{\sum(\text{Quantity} \times \text{Price})}{\sum \text{Quantity}}\]

Real Henry Hub prices are imported using EIA historical spot price data.

Expected Henry Hub prices are fetched from NYMEX futures as of the last work day of December of the prior year (e.g., December 2023 for 2024 values).

Forecasted gas prices at the plant level are typically proprietary and not publicly available. Therefore, we assume that the actual plant gas prices incurred at individual plants (which we collect using actual fuel receipts) are the Henry Hub price, plus a price adjustment that covers additional plant-specific factors, which we call Plant Specific Price Adjustment (PSPA). To convert benchmark prices (Henry Hub) into plant-level expected prices, the model derives a PSPA as the differential between actual plant prices and Henry Hub prices:

\[\text{Plant Fuel Price} = \text{Fuel Price at Henry Hub} + \text{PSPA}\]

We first calculate the historical monthly difference between the plant-level price and the Henry Hub price. We then use the median of the previous year's PSPA values to estimate a forward-looking adjustment for each plant.

The PSPA differential represents the cost (or savings) associated with transporting gas and relationships between supply and demand. For example, if historical data shows a consistent $1 difference between the hub and plant prices, we assume a $1 adder for all future months. We use the projected Henry Hub price from the relevant prior period to estimate the forecasted fuel price. In other words, we benchmark Henry Hub's price to each plant by adding a PSPA to account for the difference between the Henry Hub price and the actual price paid at specific plants. We derive the PSPA based on historical data and apply it to the forecasted hub price. In this model, we use the projected Henry Hub price as of the last work day of December of the previous year; i.e., for calendar year 2024, we use the Henry Hub future price forecasted as of the last work day of December 2023.

With the expected PSPA, we can calculate the expected fuel price forecasted by utilities, which in turn can lead to the expected fuel cost:

\[\text{Expected Fuel Cost} = \text{Expected Fuel Price} \times \text{Heat Input}\]

Rather than simply aggregate all the fuel receipt as plant's monthly fuel cost, we use the following to calculate it:

\[\text{Actual Fuel Cost} = \text{Actual Price} \times \text{Heat Input}\]

Given that there are some confidential fuel receipts and that utilities may use the remaining natural gas, simply aggregating fuel receipts cannot fully reflect the monthly fuel cost. The use of consistent heat input for both calculations reflects the assumption that generation is unaffected by price shifts.

We then calculate monthly savings for each plant and each month using:

\[\text{Monthly Savings} = (\text{Expected Fuel Cost} - \text{Actual Fuel Cost}) \times \text{Sharing Rate}\]

The savings are then pro-rated using the state specific revenue share to ensure only the state-specific portion is counted. All plant-level savings are summed up to provide the total estimated savings for that state's customers.

We calculate the margin (or loss) of the year's fuel costs, using:

\[\text{Fuel Costs Savings} = (\text{Fuel Costs}_E - \text{Fuel Costs}_A) \times \text{Sharing Rate %}\]

We then aggregate to calculate each state's savings from each plant's savings:

\[\text{State Savings} = \sum \text{Plant Savings}\]

The FACSS model assumes that estimated generation equals actual generation. We also assume a single plant has no buyer power but accepts the natural gas price of their state. The formula becomes:

\[\text{Fuel Costs Savings} = \text{Fuel Price Difference} \times \text{Heat Input} \times \text{Sharing Rate %}\]

\[\text{State Savings} = \sum(\text{Fuel Price Difference} \times \text{Plant Heat Input} \times \text{Sharing Rate %})\]

The model includes an optional cap on annual total statewide customer savings, set at a percentage of the prior year's total retail electric sales.

Dollar amounts for each year are inflation-adjusted, rounded to the nearest million, and reported in 2025 dollars.

Data Sources

• EIA-861 (Annual Electric Power Industry Report)
• EIA-861M (Monthly Electric Power Industry Report)
• EIA-860 (Annual Electric Generator Report)
• EIA-923 (Power Plant Operations Report)