Wind/Solar Expansion Will Require Perpetual Subsidies
- January 22, 2018
- 1494 views
- Wind/solar advocates point to continued cost reductions due to technological learning.
- Wind/solar opponents point to continued value declines due to intermittency.
- It turns out that these two effects cancel out fairly evenly.
- Wind and solar will thus remain as subsidy-dependent as they are today.
There can be no doubt that wind and solar power will be important players in the energy system of the future. Over the past decade or so, these sources have grown almost as fast as nuclear power did in the seventies (see below). Since 2010, wind and solar have achieved an almost perfectly linear expansion of about 5.5% of global electricity production per decade (2.3% of global primary energy per decade).
Although wind and solar are settled as important energy players, the magnitude of their contribution to the future energy system is a topic of vigorous debate. The advocate camp points to the continued cost declines of these technologies, often claiming that wind/solar power will soon achieve competitiveness without subsidies, spelling the end of conventional power sources. The following graphs from IRENA for wind and solar illustrate this argument.
In the opposing camp, people point to the variable and non-dispatchable nature of wind and solar power steadily reducing their value as market share increases. New wind or solar capacity will tend to generate power at about the same time as existing generators, thus creating an oversupply and reducing the value of all wind/solar power in the system. This effect is illustrated below showing that 15% market share will reduce the market value to just over 80% of the average for wind and just over 60% for solar.
This article will estimate how these two competing effects will play out over coming years.
Wind and solar value declines
The first thing to clarify in this study is that the value declines illustrated in the above figure represent the entire installed base. When looking at the expansion of wind and solar power in a race between cost and value, it is best to consider the marginal value of new generating capacity. Marginal value implies that existing (more expensive) capacity retains its initial value, while new (cheaper) capacity absorbs the value decline it causes to existing capacity. Mathematically stated, the integral under the marginal value curve (orange area below) must equal the area of a rectangle under the average value curve (blue shaded area below).
Naturally, the marginal value curve declines more rapidly than the average value curve. This larger value decline is more appropriate for use during the expansion period of the typical deployment S-curve where capacity installations greatly outweigh retirements. However, when the S-curve flattens out (installations mostly compensate for retirements), the average market value becomes the appropriate measure. Since wind and solar will expand for the next couple of decades, the marginal value curve will be explored first in this study.
The marginal values of wind and solar calculated from the original data are shown below. As outlined above, the marginal value curves were determined by ensuring that the integral of the marginal value curve equals the area of a rectangle below the average value curve for all the data points.
Cost vs. marginal value
Although estimates on cost declines of wind and solar power are scattered widely, the average tends to be about 10% per capacity doubling for wind and 20% per capacity doubling for solar (see this review for example). These numbers will be used in this study.
Calculations of cost and value will be made up to 15% wind or 15% solar power. Note that this represents an optimistically high value factor because simultaneous deployment of wind and solar will further reduce value by increasing the number of non-dispatchable generators in the system.
Based on data from the BP statistical review, we will start the study with wind and solar respective global market shares of 4% and 1.4%, and respective capacities of 450 GW and 264 GW. We will assume starting installation costs of $1500/kW for both wind and solar, while future wind and solar capacity are assigned capacity factors of 27% and 18% respectively. Operating life of wind and solar plants are assumed to be 25 years and 30 years respectively with no degradation. A 6% discount rate is employed and O&M costs are set to 2% of capital investment per year. Finally, we assume an average electricity wholesale value of $50/MWh and a doubling of global electricity demand by the time that wind and solar reach 15% market share each.
It is clear from the figure above that a persistent gap of about $20/MWh is maintained between cost and value as wind and solar capacity is expanded. This means that the cost declines experienced by wind and solar power are almost exactly cancelled out by value declines. When looking at the ratio of cost to value plotted below, the attractiveness of wind and solar power actually declines with increasing deployment.
Cost vs. average value
If wind and solar can double their constant expansion rate over the past seven years to about 10% combined market share per decade, we could end up with about 15% wind and 15% solar by 2050. By that time, the deployment S-curve may well start to flatten out as replacements of old capacity start to account for a large portion of new builds.
As discussed earlier, this implies that we should start to use average value instead of marginal value for wind/solar valuation. The marginal value graph given above is therefore repeated below for average value. Clearly, the gap between cost and value is significantly smaller in this case, but the gap still persists.
As more capacity is deployed to replace retired capacity and facilitate a continued expansion of global electricity demand, costs will continue to decline. However, given that the installed base will be 10-20 times larger at this point than it is today (about 3300 GW of wind and 4700 GW of solar), another cumulative doubling of total installed capacity will take several decades. Further cost declines beyond this point will therefore be very slow.
This analysis has shown that the subsidization requirements of onshore wind and solar PV will remain largely unchanged over coming decades. Cost reductions due to learning are cancelled out by value reductions due to the variable and non-dispatchable nature of these electricity sources.
Naturally, there are many ways to counter the value reduction of wind and solar power, but all of these methods impose large costs. Energy storage easily gets the most press, but, as outlined in an earlier article, the cost of these technologies needs to fall by at least an order of magnitude before they can start to challenge dispatchable power plants.
It is also important to point out that an increase in CO2 price will further reduce the value of wind and solar (below) because this will require low-emission backup power plants. These plants will have higher capital costs, thus increasing the cost related to the under-utilization of capital to accommodate wind and solar. A total ban on nuclear and CCS increases the market value, but total system costs and CO2 emissions will increase by about 25% and 150% respectively in this scenario (source).
We should also keep in mind that a large-scale shift in decarbonization strategy from wind/solar technology-forcing to true technology neutrality will strongly reduce the value of existing wind/solar generators (similar to the high CO2 price scenario in the figure above). If such a shift is done at a late stage when lots of wind and solar capacity is already deployed, the economic consequences could be severe. Wind and solar technology-forcing therefore remains a costly and risky pathway to combat climate change.
As always, I should emphasize that there are many places on Earth where wind and solar power can be deployed with significantly higher capacity factors than those assumed in this study. Higher capacity factors both decrease cost and increase value, strongly increasing attractiveness. Wind and solar will therefore certainly remain important (albeit not dominant) players even if we finally manage to replace inefficient technology-forcing policies with a more rational technology-neutral framework. Given the recommendations of climate science, implementation of such a technology-neutral approach is a matter of great urgency.