Natural Gas Forecast Methodology and Model

A long-term forecast of natural gas prices is a key component of the Power Plan. But what is meant by a forecast of natural gas prices? Pricing at national hubs, WECC wide hubs, Northwest regional hubs, at power plant take off points, at city gate delivery points? 

At the Council, we have developed a modified Delphi Methodology to aid in the creation of a price forecast. The Delphi Method of Forecasting was developed in the 1950’s at the RAND Corporation. The method is based on the idea that a forecast derived from a group of experts is more accurate than a forecast from an individual. Our forecast methodology integrates national and regional outlooks of gas prices from our Natural Gas Advisory Committee (NGAC) with quantitative analysis and modeling to forecast monthly prices for each of these pricing points.  A new forecasting model that incorporates this methodology was developed for the 2021 Power Plan - GasModNorthwest

The Council’s Natural Gas Advisory Committee meeting in June of 2019 focused on natural gas prices for the plan. The NGAC is comprised of a cultivated base of natural gas industry professionals and stakeholders with diverse viewpoints on the present and future directions of gas in our region. Prior to meeting, the members were surveyed for their organization’s outlook of natural gas prices, including the most like likely pricing future, as well as their high and low pricing scenarios. The outlooks were kept anonymous and the group results were compiled and discussed at the meeting.      

Now, let’s turn to the natural gas price forecast model. There are 4 key calculation steps in the model:

1. Set the medium, high and low annual natural gas price forecast for the Henry Hub pricing point
2. Move the Henry Hub annual forecast to an annual hub price forecast for all the Western US hubs, including the key Northwest pricing hubs Sumas, AECO and Opal based on historic quantitative relationships
3. Move the annual forecast for each hub to a monthly price forecast based on history seasonal patterns
4. Qualitative adjustments to incorporate non-quantitative outlooks and discussion

The modeling begins by calculating a price forecast for the US national pricing point - Henry Hub. Each member’s outlook of future gas prices is a direct input to the model. Member views are then blended with forecasts from four other price forecast point sources to form three data sets: medium, high and low prices:

1. California Energy Commission (CEC 2019) 2019 Natural Gas Price Forecast
2. EIA Annual Energy Outlook (AEO 2019) 2019 Natural Gas Price Forecast
3. Global Insight 2019 Natural Gas Price Forecast
4. Northwest Power & Conservation (NWPCC 2017) 2017 Natural Gas Price Forecast

The three annual price forecasts (Medium, High, Low) for Henry Hub are calculated as follows:

• Medium Forecast – median of the medium price data set
• High Forecast – 3rd Quartile of the high price data set
• Low Forecast – 1st Quartile of the low price data set

Additionally, a 3-year supply and demand cycle curve is introduced to slightly modify the forecast time series throughout the 20 year forecast horizon. This modification was introduced based on feedback from the NGAC. 

The second step is to move the annual Henry Hub price forecast to the 17 key Western US and Canadian hubs. Quantitative fits are calculated from historic differentials between the hubs and the primary pricing point Henry Hub. Note, the Northwest hubs are part of the WECC region.

Forecast Price Hubs in Western US and Canada

The third step in the model is to calculate a monthly price forecast for Henry Hub, and the 17 Western hubs. A historic monthly pricing data set dating back to 2010 is used to compute a unique monthly pricing shape for each hub. The ratio of the hub monthly price to its annual price is computed and a 3-month centered smoothing equation is applied to develop a unique and consistent shape for each hub. Each hub’s monthly shape is applied to its annual forecast to produce a 20-year monthly price forecast.

Monthly Gas Hub Price Shape

Finally, qualitative judgements may be used to make slight adjustments for issues not reflected in quantitative analysis. Additional adjustments to reflect very near-term pricing are also worked into the forecast.