Replace Net Metering With Batteries

We should end net metering, including for existing solar installations, and compensate people with a one-time subsidy for battery purchase. I'm going to give an argument from grid efficiency, which I expect to be the main consideration for most people, though the benefit that makes me enthusiastic about this change is actually increasing societal resilience.

Net metering is a common form of solar subsidy where you only pay for your "net" usage: the difference between how much you consume and how much you produce. At first glance this doesn't even seem like a subsidy: if you take 800 kWh and put back 800 kWh, then did you really use any? But it's not like a bank account: the kWh you put back are usually much less useful than the kWh you used.

Say I started a solar farm, putting out a lot of panels somewhere out in the less populated part of the state (MA), and sold the power to the grid. Averaging over the year, the electricity market might pay me $0.05/kWh. On the other hand, when the panels on my house send power back to the grid I get $0.32/kWh. [1] There are several factors that pull these apart, but I think the most illuminating one is how the value of electricity varies over time.

The $0.05/kWh that the solar farm might receive in direct market compensation is an average. It's a market-based system: when supply is high relative to demand you don't make much, and vice versa. In the summer, you might see lows of ~$0.02/kWh in the middle of the night (low power usage) or middle of the day (lots of solar), and highs of $0.08/kWh in the late afternoons and early evenings (solar diminishing; lots of AC).

In the winter the mismatch between what solar can supply and when power is demanded is even more stark, perhaps a high of ~$0.20/kWh in the mornings and evenings when solar isn't producing. As people install more solar and heat pumps, this supply-demand delta will continue shifting towards these times when solar isn't producing: on a cold winter morning the sun isn't up yet, but the heat pumps are working very hard.

Which is a long way of saying that if I send kWh to the grid when it's convenient for me (lots of sun) and draw kWh from the grid when it's convenient for me (no sun) the kWh I send are significantly less valuable to others than the kWh I draw. Then add in the large cost of maintaining the grid, and it's really very strange that my electric bill treats them the same. More than strange: when I described this system to a UK friend who has thought a lot about power, they assessed net metering as "completely insane". In MA, ratepayers are spending somewhere in the $150M to $300M range annnually [3] subsidizing households with solar.

So how did we get here? Net metering started out as a very simple technical solution. In 1978, after the oil shocks, congress passed PURPA. It required utilities to compensate based on (what today would be) the market value of their production:

the cost to the electric utility of the electric energy which, but for the purchase from such cogenerator or small power producer, such utility would generate or purchase from another source.

Residential solar installations back then were rare and small, and this number was hard to calculate. Collecting the information you'd need to get to the actual number would have been very hard, while letting the meter run in reverse was very easy, so net metering came about through technological expedience. When solar was a tiny part of the overall generation mix the overall effect was tiny. [2]

Over time it became practical to use other metering systems, but solar advocates fought to keep net metering: it's unusually politically acceptable for the scale of the subsidy, and really gets solar installed. But at the cost of making power more expensive for everyone else. Many states have stopped allowing new net metering customers, but discontinuing net metering for existing installs is much more controversial: people bought expensive systems or signed long-term leases under the assumption that net metering would continue.

Technology has changed a lot in other ways since the 1970s, and a big one is that batteries are also far cheaper. You can charge at times of low demand, and discharge a few hours later when demand is higher. When you can't do net metering, residential rooftop solar is often still worth it as long as you also install batteries. Instead of using the grid as a giant battery, drawing and exporting kWh as needed, you do it with an actual battery. This doesn't fully solve the incentives problem, because the right to draw as many kWh as you want whenever you want it is underpriced, but it does help.

The other advantage of batteries, which is the big reason I'm interested in this, is a battery is the expensive part of making a system that produces power when the grid is down. Regular residential grid-tied solar is useless in power outages: it shuts down and produces nothing. If people install batteries, however, making the house operate as an "island" during a blackout is standard.

I think people in places where the grid has been reliable are massively underrating the benefit of having power during blackouts. Living in Somerville it's been decades since we had an outage long enough to even spoil food in fridges. [4] If the power grid maintained this level of reliability, backup power would resolve an inconvenience at best. Looking at other countries, however, grids have become unreliable through natural disasters, war, and state mismanagement. And looking forward, I'm especially concerned about how the recklessly rapid pace of AI development increases the risk of all kinds of instability. Electricity is so useful for so many things that I see a lot of value in a distributed and resilient power system that can continue to make even small amounts of power available in many places if the grid goes down.

My proposal is that we end net metering, and instead of counting exported energy 1:1 against later consumption it's compensated based on the utility's "avoided cost". This is a lot like what CA did with NEM 3.0, and they saw large increases in battery installations. Unlike CA, where they allowed in existing installs to continue to use net metering for up to 20y, I propose we end it for everyone but partially buy out the subsidy with a credit you can use towards island-capable battery systems.

How big that credit should be is not something I have strong feelings about, but I expect it would be very controversial because there are big winners and losers here depending on the shape of the policy. At one end of the spectrum you could size the payment to attempt to fully compensate owners for the net present value of their foregone subsidy [5]; at the other you give them a token amount that's just sufficient to get many of them to install a battery. My big question here is whether there's enough of a constituency for any point along the spectrum that this could actually become law. This is unfortunately not a free lunch: while the batteries do save money, they don't save enough to pay for themselves and someone, whether solar owners or general ratepayers, would need to pay the bill.

(While any subsidy would apply to us, since we have solar without batteries, I think the benefits of batteries here are large enough that we're planning to go ahead and install them regardless, so we wouldn't qualify for a subsidy.)

[1] Both of these numbers exclude state incentives for solar production, beyond net metering. At maybe $0.04/kWh these help much more for solar farms than for net metering installs, but they're not enough to appreciably change the net metering picture. All numbers for MA since I live here.

[2] Solar growing in the mix is a big part of why the marginal exported solar kWh today isn't that valuable. A while ago, in sunny places, people would use a lot of AC when the sun was shining, which meant solar production was reasonably well timed. But today there's so much solar going to the grid already that power when there's no sun is disproportionately valuable.

[3] Very roughly, MA has ~1.5 GW of residential solar, producing ~1.7 TWh/y. A little under half of residential production is typically self-consumed, so figure 0.9 TWh/y in exports. These are credited at retail (my bill is $0.32/kWh) but the value to the grid is more like $0.06/kWh on average (wholesale energy, plus a little for avoiding line losses and capacity increases). This comes to $234M, but with wide error bars.

[4] I don't remember this happening and tried to look it up, but didn't find much. Even the Northeast Blackout of 2003 probably wouldn't have qualified, since power in most places was restored in 2-6hr, plus it didn't affect this part of MA.

[5] A fully "make-whole" payment would need to be sized to the net present value of the delta between the value of the current system over the remainder of a 20y operation window, and the value they'd get from the battery (less its purchase price). Penciling this out, if someone is averaging 15 kWh/day at a marginal cost of $0.32/kWh and has 5kW of solar on their roof, and received permission to operate 5y ago, the net present value of their future net metering credits, less avoided cost compensation, over the 15y remainder of the 20y window, would be ~10k. A battery (let's say 13 kWh) would regain ~$6k of that from increased self-consumption and another ~$5k from battery-operation incentives (ConnectedSolutions in MA; assuming drops to ~0 after 5y). This means you about break even (~+$1k) until you get into the cost of the battery and installation. Which is unfortunately a lot more than $1k; I see quotes for $16k, though this doesn't fully reflect how much improvements in battery tech should be bringing the price down.

Comment via: facebook, mastodon, bluesky