Many people are asking how much of our power we can get from wind and solar. With ongoing double-digit growth rates, optimists are starting to get very excited. Pessimists are looking for flaws in the argument.  And pragmatists are making plans for accommodating wind and solar in the power mix.

Firmly in the optimist camp is Mark Jacobson, the Stanford professor who has been modeling a “wind-water-sun” system to see what it would take to run states and countries on just those three resources. 

In the pessimist camp would certainly be some big players in the fossil industry, including ExxonMobil CEO Rex Tillerson, who recently told shareholders that the company isn’t getting into renewables because “we choose not to lose money on purpose.”

And while the Breakthrough Institute often puts itself into the pessimist camp when it comes to renewables, a recent essay by Alex Trembath and Jesse Jenkins could be regarded as inching closer to the pragmatist camp.

In the essay, Trembath and Jenkins highlight findings from a report recently released from MIT (where Jenkins is a grad student) on the future of solar. The MIT report finds that even though solar has boundless technical potential and very significant economic potential, it may be limited by its impacts on electricity markets.

“It is worth noting that price reductions from solar PV production are systematically most significant during the same hours when solar generators deliver maximum output,” the report says. “As a consequence, higher levels of solar penetration lead to lower revenues per kW of installed solar capacity. For this reason, at any given per-kW installation cost of solar PV, there is a system-dependent threshold or limit beyond which adding further increments of PV capacity will not break even from a cost perspective.”

Trembath and Jenkins expand on this issue, saying that as the penetration of solar and wind reach a percentage of peak demand that is roughly equal to their capacity factors, they will experience periods where they produce 100 percent of the power demand on the grid.  So if solar capacity is equal to 15 percent or 20 percent of the peak demand on a grid, it will be able to produce all power needed at certain moments.

(Just to clarify, the system, they say, has to be the whole system, with no opportunity to export surplus power to a neighbor. Popular reports like this one about Denmark, where wind output recently hit 140 percent of the country’s demand, are misleading, since Denmark is part of the larger Nordic power pool and has transmission connections to Germany.)

The research highlights that large amounts of wind and solar will cause some peculiar effects in traditional electricity markets designed for dispatchable power plants. 

In a typical market, generators place bids to supply power during a future hour or day.  The bids are lined up by price, and selected in order (the “merit order”), from low to high, until demand is met.  That last winner sets the price for power for that hour, and all the winners get that price.

This fun merit order calculator from the University of Texas Energy Institute illustrates how it works.

Because wind and solar plants will generate regardless of price, they bid into the market at zero, and take whatever the clearing price is.

If there is enough wind and solar, they start pushing other bidders out of the market due to what is called the “merit order effect.”

But with large amounts of wind and solar, the system breaks down.  Imagine an hour when the whole system runs on wind and solar, all bidding zero -- the clearing price would be zero!  Nobody would get paid, including the wind and solar generators.

As Trembath and Jenkins put it, “Solar eats its own lunch.”

Of course, this is not a new finding for those that have explored the outer reaches of energy futures.  Andrew Mills and Ryan Wiser identified the problem in a 2012 report from the Lawrence Berkeley National Lab. The report found that while solar has high value at low levels, the value drops off as more solar comes on the system.  By the time it reaches 15 percent, it has half the value; at 30 percent, only a quarter.

Concentrating solar with storage fares better, keeping two-thirds of its value at 30 percent penetration.  Wind, since it gets most of its value from displacing energy instead of capacity, retains more value even at levels over 40 percent.

“More solar -- absent storage -- means less marginal value, in part because more solar doesn’t provide any more peak reduction benefit,” Ryan Wiser said in a recent interview. “At high levels of solar penetration, solar has already wiped out the daytime peak, shifting it to non-solar times, so the marginal value of the next kilowatt-hour of solar, at least for offsetting generating capacity, approaches zero.”

“We’re beginning to see this effect in California now,” he added. “Utilities are still interested in solar, but are also looking toward other kinds of renewables. At lower levels of penetration, solar was more attractive, even at higher prices than wind.” 

To simplify their studies, both MIT and Trembath and Jenkins predicate their assumptions on a “free competitive market” -- that is, a short-term energy-only market. To compensate for the merit order effect, they assume increasingly large -- and ultimately unaffordable -- subsidies would be necessary to make solar viable. 

But there is another, more viable solution.

“The spot market just reveals market fundamentals,” said Wiser. “The actual payments to solar generators is a policy question, whether they get paid through the market or through a feed-in tariff or a power-purchase agreement.”

This is clear in the short term, where wind and solar can simply be sold directly to buyers for their revenues, and stay out of the daily markets. There are plenty of buyers who don’t want to be exposed to the risks of spot markets.

But in the long term, as spot market prices fall due to large amounts of wind and solar energy, more buyers will be tempted to play in short-term markets. This will begin to drag down the value of long-term contracts. 

“All the markets will converge over time,” says Wiser. “Without a policy motivation, in the long term, the markets will tend to value solar like the spot market does.” 

In other words, if there is enough solar, the market will no longer pay for it, and the system will break down. No more investment in solar. 

This suggests a need for different market design for electricity, where wind and solar are paid by other means.  The German feed-in tariff is the classic example, where renewable generators are paid a fixed price through 20-year contracts, regardless of daily market conditions. Power-purchase agreements can be a similar tool, as long as buyers are willing (or obligated) to sign them at a viable price.

There would be two streams of payments -- one for wind and solar, the other for what is left over, a residuals or “net demand” market. In this residual market, dispatchable power sources, demand response, and storage would compete by bidding in the conventional way. Load shapes would look completely different.

A great irony, given Breakthrough Institute’s usual preferences, is that this market design failure is equally true for nuclear power. Due to their high investment risks, nuclear plants cannot be financed in a normal competitive market. Their long lead times and massive capital costs expose investors to the risk of changes to markets, prices and regulations, before the plant comes on-line. The back-end risks of equipment failure, catastrophe, waste disposal, and decommissioning add further red flags.

The only plants being built now are in competition-free zones, and require heroic measures such as loan guarantees, production tax credits, and most important of all, a captive customer base that is financing the plants through Construction Work in Progress (CWIP) charges.

If nuclear plants are to succeed, they will need to be sheltered from market forces through a separate payment mechanism. The U.K.’s feed-in tariff of 14.4 cents per kWh for 35 years for the Hinkley nuclear plant is an expression of this idea, but is proving very controversial in the EU, as other states argue that it provides excessive “state support.”

As a final point, both MIT and Trembath and Jenkins make a number of simplifying assumptions, notably that demand is a given. But in the long run, the shape of demand will change, as consumers are attracted to low-cost periods of abundant wind and solar power. Electric vehicles, heat pumps, power-to-gas, and the myriad forms of storage, from batteries to precooled homes, will emerge in response to nature-driven supply.

Mills and Wiser of LBNL did a follow up study in 2014 looking at ways to mitigate the declining marginal value of solar and wind. They found a wealth of options, including geographic diversity of wind siting, technological diversity (through simultaneous combinations of variable generation technologies), more flexible new conventional generation, bulk power storage, and shifts in demand subject to real-time pricing.

“If solar is going to be big, we will necessarily have a completely different temporal pattern of demand and prices,” said Wiser. “The potential decline in value is real, but it is also a clarion call for how and when and where we use energy.”