Energy assets often operate for 30 to 60 years, far beyond the reach of any reliable forecast. Decision analysis helps planners replace a single projection with structured scenarios and probability ranges, test which choices hold up across many futures, and show regulators and boards how a plan behaves when assumptions break.
Few industries commit capital over longer horizons than energy. A plant approved today may still be running when its planners have retired, and almost every assumption behind it will have shifted along the way.
Why Do Energy Planning Horizons Defeat Ordinary Forecasting?
A combined cycle gas plant is usually designed for 30 years or more of operation, transmission lines for longer, and nuclear plants in the United States receive initial operating licenses of 40 years from the Nuclear Regulatory Commission, with renewals of 20 years after that. No forecast of fuel prices, demand, or policy holds its accuracy over periods like these, and planners know it.
The record makes the point plainly. The U.S. Energy Information Administration reports that the annual average Henry Hub natural gas price more than halved between 2022 and 2023. Over a longer stretch, IRENA estimates that the cost of electricity from new utility-scale solar fell by roughly 90% between 2010 and 2023, a decline few long-range plans written in 2010 came close to predicting. A plan that depends on a single projection of either variable bets on a number that will almost certainly be wrong.
This is where decision analysis earns its place. Rather than asking what the future will be, it asks which choice performs acceptably across the range of futures that could reasonably occur. Planners use a mix of tools for that work. Production cost and capacity expansion models such as PLEXOS from Energy Exemplar simulate power systems in detail, spreadsheet add-ins such as @RISK add probability ranges to financial models, and visual modeling environments such as Analytica let teams build decision models that carry uncertainty through every calculation. Each suits a different part of the problem, and many planning teams combine them.
Which Variables Dominate Long-Term Energy Outcomes?
Energy models can include hundreds of inputs, but a handful usually account for most of the variation in results. Sensitivity analysis identifies them for a specific plan, and in most studies, the same groups appear near the top.
Fuel and Carbon Prices
Natural gas, coal, and oil prices move with global markets, weather, and geopolitics, and their volatility feeds directly into the economics of thermal generation. Where carbon pricing exists or may be introduced, its future level can entirely change the ranking of options. A plan that looks cheapest without a carbon price can become the most expensive with one.
Demand Growth and Its Shape
Electricity demand was broadly flat in many advanced economies for over a decade, and planners in those markets grew used to modest growth assumptions. Electrification of transport and heating, new industrial load, and data center construction have since pushed several utilities to sharply revise their forecasts upward. Demand timing matters as much as the total, since peak hours determine how much capacity a system must hold.
Technology Costs and Policy
Capital costs for solar, wind, batteries, and other technologies have fallen faster than most projections expected, while costs for some large thermal and nuclear projects have run above estimates. Policy adds another layer, since subsidies, mandates, and permitting rules can change within a single electoral cycle. These variables are hard to assign probabilities to, which is one reason energy planners rely heavily on scenarios.
How Are Scenario Sets Built and Stress-Tested?
Scenario planning has deep roots in the energy sector. Shell’s planning group, led by Pierre Wack, used scenarios in the early 1970s to prepare the company for a sharp change in oil prices, and the 1973 oil crisis made the approach famous. The International Energy Agency’s annual World Energy Outlook continues the tradition, publishing several scenarios each year, including its Stated Policies and Net Zero Emissions by 2050 cases, rather than a single forecast.
A useful scenario set for a specific plan usually follows a clear sequence:
- Identify the few variables that most affect the decision, using sensitivity analysis on an initial model.
- Combine them into a small number of internally consistent futures, such as high demand with high gas prices, or rapid electrification with falling storage costs.
- Assign probability ranges to the inputs within each scenario where data supports it, so each future has its own spread of outcomes.
- Run every candidate plan through every scenario and record cost, reliability, and emissions results.
- Look for the plan with acceptable results across all scenarios, rather than the best result in one.
The last step is the heart of the method. Approaches such as Robust Decision Making, developed at the RAND Corporation, go further by searching systematically for the conditions under which a plan fails, then asking whether those conditions are plausible. Real options analysis adds the value of waiting, since a plan that keeps future choices open, through phased construction or flexible contracts, often outperforms one that commits everything at the start.
What Do Regulators and Boards Want to See?
Utility regulators in many U.S. states require integrated resource plans that cover 20 years or more, and these filings increasingly include scenario and risk analysis rather than a single preferred path. Boards approving large investments ask similar questions in different languages: what could go wrong, how badly, and what the company would do about it.
Both audiences respond best to analysis that is transparent about its assumptions. A filing or board paper that shows the range of costs for each option across scenarios, names the variables behind most of the spread, and explains why the preferred plan holds up gives decision-makers something they can defend. A single projected cost per megawatt hour, no matter how carefully calculated, gives them nothing to say when the future diverges. The same principle holds for any long commitment where timing and returns are uncertain, even outside energy: this site’s EB-5 planning roadmap for investors and families makes a similar case for mapping stages, deadlines, and risks before money changes hands.
The most credible plans also say what would trigger a course change. Committing to review a decision if gas prices stay above a stated level for two years, or if demand growth exceeds a set rate, shows that the plan was built with uncertainty in mind. It also gives regulators and boards a clear point at which to revisit the decision, which is far easier to defend than a plan that only changes after it has already gone wrong.
