May 2017

Why Fixed Charges for Fixed Costs are a Bad Idea

I recently filed testimony in a rate case in Ohio. The utility proposes to recover much of its demand-related distribution fixes costs through the customer charge - adding $10 to the monthly electricity bill of all residential customers. Here is a question and answer addressing the core issue.
You can read the entire testimony by following this link to the
Public Utilities Commission of Ohio website. The question appears at page 18 of the testimony.
The PUCO case number is
Q. Is the Company’s statement of justification adequate and reasonable?
A. The Company’s justification for its residential rate redesign is not adequate or reasonable. First, the Company appears to confuse fixed costs and sunk costs. Sunk costs do not vary with levels of usage; they are, by definition, not subject to change with usage of the associated asset. Once the money is spent to install a conductor of a certain size, that investment is fixed no matter how much, or how little, electricity is carried over it. Sunk costs are historical, or embedded. Given that usage of almost every asset impacts its useful life and the ultimate replacement costs for that asset, very few fixed cost investments involve truly sunk costs.
Fixed costs are costs, like sunk costs, that tend not to vary with level of use over the short term. Over the long term, fixed costs do change with the level of use. An increasing number of utilities are also recognizing, with so-called Non-Wires or Non-Transmission Alternatives projects, that some future fixed costs can be cost-effectively deferred or avoided in the mid- and short-term as well.
In the past, electric utilities did not worry about over-forecasting demand and incurring excessive demand-related fixed distribution costs. If the system was overbuilt, year-over-year growth in energy sales and accompanying demand quickly caught up with any over-building. As Warren Buffet commented in a letter to Berkshire Hathaway investors, “[h]istorically, the survival of a local electric company did not depend on its efficiency. In fact, a ’sloppy’ operation could do just fine financially.” In recent years, utilities have experienced decreasing sales growth, flat sales, and even negative sales growth. At the same time, demand has increased, loads have become peakier, and load factors have declined. Peakier system loads can be addressed in three ways: (1) aggressively pursuing peak reduction programs for all customers, (2) spending more on the system to meet peaks, and/or (3) implementing rate structures that immunize the utility from the consequences of increased demand-related fixed cost investment through non-bypassable rates that ensure utility revenues remain constant regardless of customer usage. The Company’s residential rate proposals focus on the rate redesign approach, with the likely result that they will have to spend more money on distribution system infrastructure.
It is understandable that the Company would try to fix its larger problems with rate restructuring, but it is not reasonable. If a utility company forecasts greater demand for energy than it ends up experiencing, it will have an overbuilt system and experience a situation where sunk fixed costs are potentially stranded—not subject to recovery under current rates. The economically efficient solution is good price signals that do not undermine the economics of demand response and energy efficiency, better forecasting, and a smarter grid that leverages the potential benefits of all manner of distributed energy resources. As explained previously in the section discussing impacts on energy efficiency and distributed generation, the Company residential rate proposals not only constitute the bad choice, they frustrate the good ones.
For example, if the utility forecasts that demand on a particular feeder will be heavy, it may install a larger, more expensive transformer. The money spent on that transformer is a historical or sunk cost. Since the money is for a transformer, the costs will be treated as a fixed cost, and allocated accordingly. If demand does not match the forecast, the utility will face problems recovering the cost of the too-large transformer through volumetric rates. Of course, if the utility is guaranteed recovery of the costs through fixed charges, it will have no incentive to improve the accuracy of its forecasts. Importantly, the size of the next transformer and associated cost is a fixed cost that can be impacted by customer demand in the future. Energy efficiency, demand response, and other factors can reduce the fixed cost requirements in the future, and perhaps even allow for the installation of smaller replacement equipment. These measures can also extend the useful life of the installed fixed cost assets. For these reasons, the price signal impacts of rate design can and do impact fixed costs on a going forward basis.
Second, even if demand and customer connection costs are the primary drivers of distribution costs, this does not compel or even justify the allocation of demand-related fixed costs to the customer charge. The Company offers no evidence to support the leap of logic that because demand-related fixed costs are, like customer connection costs, a driver of distribution costs, they should therefore be collected as a customer cost.
Third, the statement about price signals is illogical in the extreme. The Company assertion is that recovery of fixed costs through the volumetric energy is a false price signal because a change in usage cannot reduce demand-related costs. Again, the Company confuses fixed costs with sunk costs. It is widely accepted—and a strong justification for grid modernization investments—that customers can reduce the requirement for expensive infrastructure investments by reducing their usage during particular times of the day. These reductions arise as a result of reduction in system loading so to avoid upgrades, as well as reduction in wear and tear (temperature-related degradation) and resultant capital cost deferrals for replacement. Higher volumetric charges for on-peak usage can support demand response programs and energy storage deployment with similar results.