California paid utilities the same revenue whether or not they sold more power

In the 1970s California wanted utilities to sell less energy — efficiency programs, inverted block rates — while utility profits were still tied to kilowatt-hour sales. Every saved kilowatt was lost revenue, so the state was asking companies to saw off the branch they sat on. The CPUC first fixed gas in 1978 with the Supply Adjustment Mechanism, then by 1982 extended the Electric Revenue Adjustment Mechanism (ERAM) to all investor-owned electric utilities.

The mechanism is bookkeeping, not subsidy: regulators set an authorized revenue level, actual sales-based revenue went into a balancing account, and over- or under-collections were refunded or recovered through the next year's rates. The utility still had to control costs to earn its authorized return — but weather, conservation, and efficiency programs stopped being threats to its income.

The result made California's utilities willing partners in the most aggressive efficiency push in the US: high-efficiency appliances and building standards followed, and pre-1998 programs delivered roughly $1.5 billion in net resource benefits to customers. The EPA later found ERAM reduced both rate risk for customers and revenue risk for utilities.

True up revenue, not sales: sever profit from kilowatt-hours

Traditional ratemaking paid utilities per kilowatt-hour, so efficiency directly destroyed their fixed-cost recovery.

A balancing account removes volume risk without removing cost discipline — profit still depends on managing expenses.

It ended contentious fights over sales forecasts in rate cases: forecast errors simply washed out.

It let regulators pair high marginal prices (encouraging customer efficiency) with utility financial stability.

Institutions do what their revenue formula tells them to. If you want a seller to help buyers consume less, first make the seller's income independent of volume.

California dropped ERAM in 1996 during electricity restructuring, judging it incompatible with retail competition — and brought back revenue decoupling in 2004 after competition failed to deliver. The model spread: by the 2010s roughly half of US states had some form of gas or electric decoupling, and the design became a standard precondition for utility-run efficiency programs.