January 16, 2011

Economist: Levelized cost is poor way to value intermittent renewables

By Steve Muller
One big question for the renewable energy industry is just how a project should be valued. Unlike a traditional power plant, a 100-MW wind farm will rarely produce 100 MW. Moreover, a 100-MW wind farm with a 40% capacity factor may be worth significantly more than the same plant with a 30% capacity factor — even if the project costs more to build.
One answer, of late, has been to use the levelized cost of energy. That, advocates say, incorporates all the costs of the project to calculate the cost of each kilowatt delivered — providing developers and investors with a solid understanding of a project's value.
Economist Paul Joskow, however, said even that is inadequate.
Using levelized cost to compare dispatchable generation with intermittent renewable generation is "seriously flawed" because the approach fails to take into account the fact that hourly wholesale electricity market prices vary significantly over the course of the day and year, Joskow said in a Massachusetts Institute of Technology Center for Energy and Environmental Policy Research discussion paper.
Intermittent renewable generation sources such as wind and solar have low capacity factors compared to traditional thermal baseload generation plants, and their generation may not occur at times when electricity is most valuable.
Therefore "[t]he extension and use of levelized cost comparisons to intermittent generation has been a mistake and tends to implicitly overvalue intermittent generating technologies compared to dispatchable alternatives," Joskow concluded.
That has important implications for the effectiveness of renewable energy subsidies and procurement policies and for determining the most cost-effective policies for reducing CO2 emissions, he added.
Joskow pointed out that most of the current research on renewable generation focuses on the infrastructure and operational challenges of reliably integrating new intermittent generation sources.
It is just as important, he argued, to "properly measure the economic value of additional investments in intermittent generating technologies compared to dispatchable generating technologies." Only then can costs and benefits of renewable power subsidies and mandates be properly evaluated, he said.
Levelized cost is a measure of the total life-cycle costs (capital and operating) of a generating unit per MWh supplied. It originally was developed for use in a regulated utility environment.
Joskow noted that many analysts believe that some renewables, such as wind, are already competitive with conventional generation on a levelized cost basis, and that others soon will be.
While levelized cost is an accepted methodology for comparing conventional dispatchable generating technologies, Joskow believes it is not particularly useful for comparing renewable generating technologies with conventional ones.
The flaw is that the approach does not factor in two important elements: Electricity prices vary widely throughout the day, and intermittent generating technologies (specifically wind and solar) have fixed production profiles that may not match electricity demand and prices.
Using simplified examples including peak and off-peak prices, capacity factors and output profiles, Joskow shows that although wind or solar generation can have the same levelized cost as dispatchable thermal generation, the value of the renewable generation is less.
Therefore, renewable portfolio standard programs that require selection of suppliers based on lowest levelized cost per MWh are "likely to fail to lead to the selection of the highest value generating offers and increase the burden placed on consumers who are forced to pay for the above market costs of the associated power supply contracts," he said.
The bidding framework is "likely to distort the kinds of projects that developers of intermittent technologies seek to bring forward," he continued.
It tends to undervalue solar, which generates electricity during the day when prices are higher, and overvalue wind, which typically generates much of its electricity at night when prices are lower.
Joskow also pointed out that since the renewable generator is being paid as if it were dispatchable, it has little financial incentive to incorporate energy storage to actually achieve dispatchability.
Instead of the levelized cost approach to comparing different generating technologies, Joskow favors a methodology that integrates differences in production profiles and the associated variations in the market value of electricity with traditional life-cycle costs.
Joskow, president of the Alfred P. Sloan Foundation, was formerly a professor of economics at MIT.

2 comments:

  1. I find this article to be much in line with what I have learned while studying intermittent resources in terms of my work in Energy Policy. In wind studies released by the NYISO along with others, it is clear that the capacity factor for wind is low and is generally not there when the system is peaking (when it is needed most). I also wonder in this "levelized cost" if it is counting the additional regulation service needed by the power grid to manage the variable output of wind power. In December 2010, the Federal Energy Regulatory Commission released a proposed rule making that would enable tranmission providers to recover the cost of the additional regulation service from the wind generators themselves (based on cost-causation principles). Currently, this is paid for by loads and is often not built into the equation of "total cost" of wind power. If this continues as is, it seems that this may hurt the development other renewables such as solar which may have more potential.

    ReplyDelete
  2. The takeaway for me is that bad measurements lead to skewed development incentives and, therefore, suboptimal policy.

    "[Levelized cost] tends to undervalue solar, which generates electricity during the day when prices are higher, and overvalue wind, which typically generates much of its electricity at night when prices are lower."

    ReplyDelete