TODAY’S STUDY: THE SMART WAY TO DO SOLAR IN HAWAII
Efficient Design of Net Metering Agreements in Hawaii and Beyond
Coffman, et. al., July 20, 2015 (University of Hawaii)
In Hawaii, like most U.S. states, households installing rooftop solar photovoltaic (PV) systems receive special pricing under net-metering agreements. These agreements allow households with rooftop solar to buy and sell electricity at the retail rate, effectively using the larger grid to store surplus generation from their panels during sunny times and return it when the sun isn’t shining. If a household generates more electricity than it consumes over the course of a month, it obtains a credit that rolls over for use in future months. Net generation supplied to the grid in excess of that consumed over the course of a full year is forfeited to the utility. Net metering agreements often include a monthly fee to support billing, transmission and operation of the grid. On Oahu, the customer charge is $9, and additional monthly fees can bring the minimum bill up to about $17. Households who have installed enough rooftop solar such that they are “net zero” currently pay at most $200/year for grid connection and load-shifting services.
A growing concern is that the utility has many costs besides the fuel used in electricity generation, and most of these “fixed costs” are lumped in with perkilowatt hour (kWh) charges. As a result, under current net metering agreements, when a solar customer provides their own power, they don’t pay the fixedcost component for each kWh they produce. Under a revenue-decoupling rule, those costs are shifted to households and businesses without rooftop solar. As less power is sold in Hawaii, fixed costs per kWh are rising fast (Figure 1). Most of the decrease in power sales is due to gains in efficiency, but some of it is due to installations of solar PV. Residential customers now pay roughly $0.17/kWh for fixed costs. For a household consuming 15 kWh per day—which is about average— that amounts to $76.50 per month. After the drop in oil prices earlier this year, well over half the utility’s revenue from residential customers goes toward fixed costs.
A longer-term concern, particularly in Hawaii with its high electricity rates, is that an inefficient pricing system could encourage many households and businesses to install stand-alone systems, unplug from the grid, and further raise costs for everyone else.
So what would a better net metering agreement look like? There is no single answer to this question, but there are some basic economic concepts that ought to guide policy. A perfectly efficient scheme--one that minimizes waste--would set real-time per-kWh prices equal to the “marginal” generation cost, and allow anyone to buy or sell as much power as they want at this price. Marginal cost is the incremental cost of power production—the cost of generating one more kWh. This cost can vary a lot depending on total demand and the amount of renewable power, among other things, so ideal prices would vary over the course of each day, week, season and year. This variation is likely to become especially pronounced as the variable supply from renewable sources becomes more prominent. Fixed costs could be handled a few ways, but would not necessarily be included in the per-kWh price today. Implementing true marginal-cost pricing would solve problems with net metering agreements and make achieving the State’s renewable energy goals much less costly overall. It would also entail a number of challenges.
Here we sketch out a set of long-term solutions based on marginal-cost pricing as the primary platform. We also offer some near-term suggestions for interim policy. But first, we briefly review some of the benefits and challenges of distributed solar, and explain why net metering agreements were originally set up as they were…
The Long-Term Solution
Here we propose three simple ideas to achieve a more efficient and equitable electricity system that is particularly relevant for systems with a lot of variable renewable supply, like rooftop solar. I. To improve efficiency, all households and businesses should have an option to buy or sell as much electricity as they like at a price equal to the marginal cost of electricity production. II. Fixed cost of grid construction and maintenance (power plants, transmission cables, transformers, etc.) not covered by marginal-cost rates should be spread as widely as possible in order to limit grid defection and improve fairness. III. Subsidies for solar power, to the extent that they are necessary to achieve renewable energy goals, should be paid via tax credits rather than cross-subsidies among electricity customers, in order to aid transparency, avoid regressiveness, and further inhibit inefficient grid defection.
These simple ideas are easier to write down than to implement. In particular, marginal-cost pricing could imply prices that vary a lot from day to day and hour to hour. As a practical matter, such pricing would require smart meters that could measure electricity use on a short time step, and smart meters are costly. There would also be practical challenges surrounding basic measurement and regulation of what constitutes true marginal cost, which may include not just incremental generation costs, but adjustments for critical peaks that push the limit of what the whole system can supply. The marginal cost of electricity can even vary by location in power systems with limited transmission capacity.
Interim Solutions with Customer Choice
In our view, it would be best to offer a tariff to all customer classes with hourly prices that reflect the continuous variation in supply and demand of electricity. This balance will evolve over the coming years as renewable energy grows. Some customers will be able to reduce their bill by curbing electricity use during times of high marginal cost (high loads with low renewable power production) and increasing it during times of low marginal cost (low loads and/or high renewable power production). These customers will choose the hourly pricing tariff, and in so doing they will reduce their own costs and the system’s costs.
Many kinds of electricity loads are relatively easy to shift between hours of the day. Examples include water heating, water pumping, pre-cooling of commercial buildings, and electric vehicle charging. We expect this option to be attractive to many customers. Their private gains from shifting loads toward renewable supply would then significantly improve integration of renewable energy with the existing system.
Moreover, variable pricing would open a door to innovative ideas for storing electricity or otherwise shifting loads over time. Anyone with a good idea for shifting demand toward renewable supply could profit from it, and simultaneously improve the whole electrical system. In the future, we expect demand response to changing prices to be automated by a panoply of smart machines that a growing share of customers will use, because it will be economical to do so.
We also recommend that a flat-rate tariff be offered that is based on the average marginal cost for the overall power system. Customers selecting this tariff would pay a flat price per kWh that is adjusted monthly based on the average marginal cost. This tariff offers a less volatile option for customers who fear a “bill surprise” from hourly pricing. Although the calculation would differ somewhat from current billing, we expect the total bill would not look much different from current pricing for the average customer, although there may be some adjustment between residential and larger scale customers.
It would also be possible to offer a time-of-use (TOU) tariff, with flat rates each month that differ between peak, shoulder and off-peak hours. The rates for each slot would be the average marginal cost during these periods. This tariff would provide an intermediate option for customers willing and able to shift loads, but who prefer a more predictable size and timing of rate changes.
All three tariffs are based on the system’s marginal cost of power production; they do not include other costs in the per-kWh rate. Customers with solar therefore do not avoid paying any costs when they produce renewable power. It is possible that marginalcost pricing would recover enough revenues to cover both fixed costs as well as generation costs, but this is not necessarily the case. We address the tricky question of residual fixed costs (or profits) below.
Although a flat-rate marginal-cost tariff does not provide an incentive for customers to shift use from current peak-load times to high-renewable times, it serves as an important backstop for risk-averse customers. In order to protect this option, it may be necessary to move high-solar customers off of this tariff and onto one of the time-varying tariffs once daytime marginal costs become significantly lower than nighttime marginal costs. Otherwise, by producing power at low-value times and consuming it at highvalue times, solar customers would raise rates for other customers on this tariff, effectively obtaining a crosssubsidy from them.
Residual Fixed Costs or Profit
In the long run, marginal-cost pricing could generate more or less revenue than total costs. Marginal cost is typically above each hour’s average generation cost, since the utility draws from a mix of low-cost and high-cost generators, and marginal cost is based on the highest-cost generator currently operating. So revenue from marginal-cost pricing will exceed fuel and other variable costs, leaving a potentially large sum to cover grid management and capital costs. It is not clear whether this extra revenue will be greater or less than fixed costs.
Given the reductions in demand that we have seen in recent years, and the further reductions in peak demand that we would expect under marginal-cost pricing, it seems likely that revenue from marginal-cost pricing will fall somewhat short of total system costs. A large component of these costs are labor and other grid management costs, which tend to increase with wages, and are likely to increase further with more complex grid management. The practice of keeping most power plants running at all times also contributes to a possible revenue shortfall, since it raises fuel costs (the large amounts of fuel needed to run plants at minimum load), while artificially reducing marginal costs. At the same time, kilowatt hours sold are declining.
The best way to deal with these residual fixed costs is not completely clear. To keep the systems as efficient as possible, residual fixed costs probably should be recovered through some combination of fixed monthly fees or general tax revenues. Options include a flat charge on every customer’s bill, demand charges based on peak use from the grid, a small adder or percentage markup on per-kWh rates, or possibly a one-time Legislature-funded markdown of sunk costs. When choosing among these options or combining them, it will be necessary to find the right balance between prompting inefficient grid defection (via high base charges or demand charges) or producing unintended cross-subsidies of PV customers (by including residual costs in the kWh rate).
To the extent that fixed costs are not sunk costs, but rather fundamental to management of the grid itself—like the utility’s labor force—it makes sense to pass these costs on to customers in one form or another. If these costs, unlike sunk costs, cause grid defection, then it suggests grid defection is truly economical.
If the power company continues to use older generation and transmission assets for which capital costs have already been completely recovered, marginal-cost pricing could produce surplus revenue, reducing or eliminating residual fixed costs. Although it seems unlikely, in the case that marginal-cost pricing brings excess profits to the utility, the excess profits could be taxed to supplement the State’s general resources or otherwise used to offset income or excise taxes.
Demand Response: The Least Cost Option
Besides traditional cycling of power plants, there are just two ways to deal with the variability challenges of renewable energy: store it or shift it. Batteries can store energy over hours or days. Or we could build dams and reservoirs that could store energy by pumping water uphill during surplus times and later use it to make hydroelectric power. These would require huge capital outlays and have significant social and environmental implications.
Alternatively, we can try to shift demand toward the sun and wind. Shifting demand to cope with variable renewable supply is highly attractive because it doesn’t require major capital outlays. It only requires rate restructuring, a restructuring that would enhance efficiency and lower costs regardless of our renewable energy goals. Those who choose to have hourly marginal-cost pricing would require smart meters, but these costs are modest and they would be targeted at those most willing to shift loads. Those on time-of-use tariffs could get by with simpler meters.
Although it remains unclear how much demand could shift in response to variable marginal cost pricing, and whether large-scale storage will ultimately be prudent, it makes sense to start by taking advantage of the most flexible, least cost option.