Efficiency savings make no difference to aggregated demand so that is irrelevant. See Jevon's Paradox, which absolutely applies to electricity.

So you're saying people will install two water heaters if they buy a more efficient one?

Not sure this is the elastic droid that paradox is looking for.

No, that's not what I'm saying. Jevon's Paradox has a good Wikipedia page that describes the problem. Efficiency improvements don't reduce consumption in aggregate, because the lower demand causes lower prices that then unlock new use cases that create more demand that pushes up usage again.

As the wikipedia page points out, this assumes a highly price elastic good.

Household energy use is not such a good, because most uses have fixed utility.

Very few people are going to cool their house down to 10C because the price of cooling to 15C decreased.

Nor is their overall energy use budget-capped. (Outside of low income AC and/or heating use cases)

You started by talking about electricity usage, which is fungible and elastic. There's no dedicated grid for household electricity so Jevons Paradox certainly applies.

> Nor is their overall energy use budget-capped. (Outside of low income AC and/or heating use cases)

Requiring more power plants doesn’t mean that power will be more expensive. Power will only be more expensive if we get more demand and less supply. If supply (power plants) increases linearly with demand there won’t be a price difference.

Power doesn't work as simply as that.

Example: as a regulated utility it will often cost more per unit if underused vs optimal generating supply (which has already been passed through into rates)

Additionally, the generating source heavily influences ultimate cost.