The flaws in the utilities' push for residential demand charges
Time-based rates offer far better incentives, argue James Tong and Jon Wellinghoff
The following is a guest post from James Tong and Jon Wellinghoff. Tong is the CEO of Advanced Grid Consulting and Wellinghoff is the Chief Policy Officer at SolarCity and a former FERC commissioner.
Previously, we wrote about how fixed charges represent a false fix to utilities’ revenue challenges, potentially hurting all customers as well as utility shareholders. While some have since backpedaled on fixed charges, others are doubling down and pushing for mandatory residential demand charges.
To fairly evaluate the merits of demand charges, we should understand the utilities’ conundrum. Utilities must continually invest billions of dollars to ensure enough capacity to meet the demands of all customers at all times. Utilities mostly recover these investment costs, which are largely fixed, through variable charges, i.e., flat rates on every kilowatt-hour (kWh) consumed.
Recovering fixed costs through variable charges historically wasn’t a problem, as strong growth in demand for electricity meant a strong revenue base to pay for those costs. But in recent decades, demand growth has been falling dramatically, which is eroding the revenue base (see Figure 1).
Compounding the problem, peak demand has been deviating from average demand (see Figure 2). This means that utilities are increasingly building and maintaining more capacity to meet peak demand but are, on average, getting less usage from it. Less usage means less kWh sales or revenues to pay for that expensive capacity. And the majority of any given utility’s distribution resource needs are projected to stem from capacity additions to meet peak usage on the system.
With declining revenues and rising costs, utilities must either raise electric rates (which can discourage consumption, thereby further eroding their revenue base) or find other ways to charge customers. So utilities have been pushing increased fixed charges and now demand charges.
The case for residential demand charges
Demand charges vary. The ones being proposed for residential customers typically involve monthly fees based on one’s highest average usage (measured in kilowatt or kW) over a certain time interval (e.g., 15 minutes) in a given billing period. Demand charges have long been used for commercial and industrial customers, who tend to be more sophisticated than residential customers about energy consumption and have far greater peak usage, often measured in megawatts. But given trends in costs and revenues, utilities are now seeking demand charges for residential customers, and most controversially, to make them mandatory or the default option for all.
Proponents justify demand charges on three grounds: transparency, equity, and efficiency. First, demand charges would send more transparent pricing signals of the enormous and largely fixed costs required to meet peak grid demand.
Second, they would assign costs more equitably to cost-causers and restore the “cost-causation principle”. Those who use energy more intensively (thus requiring more capacity) would pay more.
Finally, demand charges would incentivize more efficient consumption to lower peak demand, thereby lowering grid costs. Even more, demand charges would unlock new opportunities for technology providers – such as storage companies – and for customers to lower their energy bills.
A closer look at demand charges
For these utility arguments to be valid, two conditions unequivocally must be true: 1) residential customers must be able to understand demand charges, and 2) demand charges must reflect incremental costs of customer usage (that is, costs created from customers’ actions, such as turning on air conditioners, and not costs incurred prior to those actions).
But there is scant evidence that customers could ever understand the difference between kWh versus kW (even industry experts regularly confuse “energy” and “power”), much less respond to how each is priced. An incentive is meaningless when nobody can take advantage of it. In this regard, demand charges may be no better than fixed charges – and perhaps in some instances, even worse. Innocuous activities, like doing chores (vacuuming while running your washer and dryer) on a cool Saturday afternoon, may result in a significantly higher bill, even though such actions would create minimal grid costs.
This speaks to the bigger problem of proposed demand charges: they poorly reflect incremental costs to the grid. A single household’s peak usage is too small to cause significant system-wide costs. It is the aggregate peak usage that drives costs. These aggregate grid peaks correspond closely with time (see Figure 3) and poorly with individual demand peaks, which often occur outside of the system peak.
Furthermore, the only things that utilities size according to demand from individual residential customers are the final line transformers and connecting secondary lines. These costs are small relative to those of generation and transmission capacity. And most of these capacity costs are sunk. By definition, sunk costs cannot be incremental. Using the cost-causation principle to justify demand charges to pay for sunk costs makes no sense; future usage behavior does not cause costs that have been sunk .
This point is worth underscoring, for it exposes the contorted logic in advocating for demand charges as a means for recovering fixed costs. If fixed costs are truly as inescapable as many utilities want us to believe, where would the bill savings promised by utilities come from if a customer actually responded to demand charges?
There are only three possibilities: they result from shifting costs to other customers (which is what demand charges are supposed to prevent); they stem from being first overcharged by utilities (akin to marking up items before offering a fantastic sale); or the customer was never expected to respond to demand changes and the savings opportunities were mere window dressing for demand charges. Whatever the scenario, none can be characterized as transparent, equitable, or efficient.
There is an important caveat. Demand charges, when carefully structured and targeted for specific customer segments, can incentivize behavior that helps avoid or defer future (not past) capacity investments, which indeed can lower costs for all. But even here proponents must reconcile two inconveniences.
First, this argument assumes that fixed costs are not necessarily inescapable, but are indeed variable in the long-run – a position identical to that of advocates for distributed generation. However, proponents of demand charges often downplay or outright deny this, especially as it relates to the value of distributed solar. One notable demand-charge proponent recently criticized net metering, saying: “I would never ask consumers to pay for [reductions in grid costs] that theoretically might be provided at some unpredictable time.” Yet this is the same ask from demand-charge advocates.
Second, and even more inconvenient, there are better pricing alternatives that can improve cost-recovery, are more readily understandable to consumers, and have yet to be widely implemented – namely, time-based rates (TBRs), including time-of-use (TOU) rates, variable peak pricing (VPP), critical peak pricing (CPP), or real time pricing (RTP).
It’s about time
Compared to demand charges, TBRs are far more precise in recovering costs and more flexible in tailoring prices according to customer preferences as well as the needs of the grid .
TBRs can be simple, such as TOU or seasonal rates, or more granular and sophisticated, such as CPP or RTP (see Figure 4). This flexibility and precision will be essential as the grid become more distributed and as customers expect different benefits from the grid at different times.
More importantly, TBRs are far easier to understand. We regularly pay different prices at different times: peak minutes for cell phone services, matinee discounts for movies, higher airfares for peak travel time and dates, surge pricing for ride-sharing services, etc. Consumers are accustomed to paying more when the resources they want are scarce (and paying less when they are abundant). This is how most markets work.
Though regulation is supposed to replicate market forces, utility ratemaking has largely ignored the economic concept of scarcity (a concept so fundamental that it begins many an economic textbook). Flat electric rates that don’t vary with time convey to customers that resources are equally available (or scarce) at all times, which is hardly the case (again, see Figure 3). Inevitably, customers will consume and conserve energy at the wrong times. We should therefore not be surprised that too much energy is consumed during peak hours and too little at all other times to pay for the expensive peak capacity.
A better way to pay for that capacity is to charge customers more at times when it is in high demand. Paying a premium for premium times is far more understandable than paying demand charges . And clearer pricing signals will better enable customers to avoid peak times and use technologies, such as solar and storage, to shift consumption times, or even relieve the grid of congestion.
Demand charges may have the opposite effect. To make demand charges more acceptable to the public, utilities often propose accompanying decreases in kWh rates. However, lower rates would further distort retail prices from actual costs that vary with time (again, see Figure 3) . So even if individual peak usage may decrease, aggregate demand may actually increase during peak times. At the very least, lower rates will encourage overall energy consumption and counter efforts on energy efficiency and climate change.
Energy economists have long been advocating TBRs as a way to improve efficiency, flatten demand peaks, and lower grid costs. Furthermore, a recent analysis from the Lawrence Berkeley National Laboratory also suggests that TBR would improve fixed-cost recovery, increase solar integration, and help resolve differences over net metering.
We do not suggest that demand charges are categorically wrong. But they are wrong when they are broadly applied with little regards to the needs of the grid and the ability of customers to respond to them.
Moreover, demand charges alone cannot offer the same promise as TBRs, because they are less flexible and do not align with peak system demand. To be as effective, demand charges will have to be combined with some component of time. In any event, they will require the same advanced metering infrastructure as TBRs. So why not try TBRs first?
If utilities are sincere about reducing grid costs, integrating new smart grid technologies, or creating a customer-centric industry (as they are looking to do), they should be asking for TBRs, not demand charges. Why hit customers with confusing and largely unproven pricing schemes, when TBRs provide a far more comprehensible and effective mechanism (see here and here, for instance) at achieving the same policy goals?
 Many readers may object, arguing that customers’ peak usage indeed create incremental costs because utilities must plan enough capacity to meet everyone’s demand. This may be true for future capacity costs, but not capacity costs already incurred, i.e., sunk costs. Whether a customer uses 1kW or 5kW does not change the amount of sunk costs. Put it differently, if we were to assume future customer demand peaks cause sunk costs and charge customers accordingly, then any customer with lower-than-expected peaks could not be paying her fair share. Such a conclusion defies cost-causation principles, if not common sense.
 Some utilities have been pushing for demand charges that correspond with peak demand. This is essentially a variation of time-based pricing, the discussion of which is outside of the scope of this paper. While it has merits, it is very similar to, but less efficient than critical peak pricing, a pure time-based rate. For more information, see here.
 Clearly, we would want to protect vulnerable customers who cannot readily shift energy consumption for basic needs. But as we have written separately, the emergence of distributed energy resources is giving policymakers new options to empower customer to lower bills rather than rely on traditional methods of subsidizing rates.