How to Reduce the Impact of Fossil Fuels on Electricity Prices

2022 August 23 Twitter Substack See all posts

An explanation of a Matt Ygelesias tweet.

Marginal Supply Sets Wholesale Electricity Cost

Matt Yglesias notes:

I think something that's not as widely understood as it should be is that even if you had a grid that's 85% renewables, swings in the price of whatever you're using for the last 15% (presumably gas) would still drive economy-wide energy prices. In other words, even if you decisively commit to decarbonizing the price of fossil fuels remains very economically relevant all throughout the transition and is something you're going to need to worry about. Especially if the transition means nobody wants to invest in it. In general, electrical grids are kind of weird and unintuitive so in any system it's going to be the price of your marginal power source (whether that's gas or hydrogen or pumped hydro or whatever else) that determines the wholesale electricity price.

Matt is correct because the electrical grid must balance perfectly at all times. In wholesale markets, the price that the supplier of the last kilowatt needs to participate is the spot price that everyone unhedged pays.

We see this in places like Texas. The cost of generation for solar panels, wind turbines, or the South Texas Nuclear Project doesn't change, but combined-cycle gas plants provide the last kilowatt most of the day. An increase in natural gas prices increases wholesale prices for all hours where gas is online.

Energy prices might be painful if we disincentivize natural gas production and transportation because gas is still setting our electricity prices many hours of the day.

Understanding Power and Energy

Power is a rate that we measure in watts. Power plants produce so many watts we usually use megawatts (a million watts) or gigawatts (a billion watts) to talk about power.

Energy is an amount. With electricity, we typically use watt-hours. A watt-hour is the energy expended in one hour at a one watt rate. The relationship is the same as speed and distance, where power is like speed and energy is distance.

Some types of power plants excel at providing energy. Solar, wind, nuclear, and geothermal fit this description. They can produce lots of energy cheaply, but they deliver power on their schedule, not yours. Increasing their flexibility is expensive.

Other power plants have a comparative advantage in delivering power. Natural gas, oil, and coal can be pricey on an energy basis but can ramp up and down more economically because fuel is a large portion of their cost. Even though batteries are called energy storage, they behave like power storage in markets. They are a substitute for fossil power plants to provide the last kilowatt.

Hydropower can be exceptional at both power and energy but is geographically limited.

Replacing fossil fuels creates abundant energy but does not create abundant power.

Limiting the Tyranny of the Last Kilowatt

Utilities charge almost all customers by average prices. The key to lowering the average is pushing fossil generation to zero in most hours.


Batteries are great for daily cycling. They can soak up extra electricity from the wind, sun, nuclear, or geothermal during periods of low demand and discharge when we all use more electricity in the evening.

In a renewables-dominated grid, batteries can reduce the number of times a gas power plant has to start. Instead of ramping every evening, it only runs on windless or cloudy days. More efficient combined-cycle plants can run more hours at steady rates to meet current demand and charge batteries for afternoon/evening peaks. More expensive simple-cycle turbines run only in emergencies.

Vehicle Charging

Electric cars use lots of energy but can be agnostic about power usage. We will likely have 10x as much battery capacity in vehicles as in grid storage. Incentivizing car owners to charge based on grid conditions is the low-hanging fruit.

Cars can soak up excess renewables or nuclear generation. They can stop charging if margins get tighter and avoid firing up a gas power plant. California's recent move to allow the submetering of electric cars is the future.

The Grid as Premium Service

Because electricity has historically been expensive and scarce, we use electricity for things we really value. Most people in the US will pay whatever price it takes to run their air conditioning. We are used to power and energy being relatively pricey.

Falling energy costs change the paradigm. Not only do some users need energy more than power, but new ideas will emerge to take advantage of cheap electricity while being power agnostic. One example is making fuels and materials from electricity with plants that only run when the sun shines.

A distribution company and the Pittsburgh airport are examples of users that need less power service than their utility requires them to buy. Because they mostly need electricity when it is sunny, solar and a few batteries meet most of their electricity needs.

Falling energy costs relative to power costs is one of the main reasons I think there could be a lot of defection from the grid. The grid is a premium service that provides power on demand. Many users do not need that level of service.

Focus on the Power

The key to keeping fossil fuels from spiking electricity costs is to manage how we use power. We can use batteries, vary our demand, or create microgrids. We can only realize the opportunity of falling energy costs if we change how we use power.

We also need to trust that the market will take advantage. An anti-carbon crusade that makes it hard to drill or build pipelines will only increase the pain of the transition. Eventually, we will reach a point where fuels derived from cheap energy can augment and cap the cost of delivering power, but we won't get there overnight.

There are very few technologies that can create abundant power. Cheaper batteries will help. Inexpensive, solid-state heat engines like thermoelectric generators (TEGs) are part of the solution. Tiny nuclear devices that use cheap TEGs or betavoltaics/thermovoltaics would be the final step. Centralized solutions will remain expensive by definition because of the coordination and low-utilization infrastructure they require.