On Energy Cost Trends (applying caution to the big talk of energy transformation)

“Comparing straight costs for disparate products or where there are intangibles involved (convenience of a gasoline engine versus low maintenance on an electric vehicle) can be tricky and allow for wild differences in estimated consumer demand for a product, even without disagreements over future cost trends.”

“Although costs have been dropping for the Cleantech products such as solar, wind power, and lithium-ion batteries, costs for a number of other long-advocated energies such as cellulosic ethanol have not…. The ultimate lesson is an old one:  skepticism should always be applied, especially to the more optimistic predictions.”

Any number of analytical and advocacy groups have pointed to plummeting costs for wind, solar and lithium-ion batteries to predict massive changes in the future energy industry, some more aggressive than others.  The UK-based think tank RethinkX, for example, expects a huge shift from private vehicle ownership to the use of Transportation As A Service, provided mostly by battery-powered autonomous (self-driving) vehicles. [1] Bloomberg New Energy Finance’s New argues that continuing cost improvements will make electric vehicles cheaper than conventional vehicles by 2025, causing demand to soar. [2]

Certainly, cost is the most important and typically dominant factor. But it is not the only thing. Despite its cheapness, for example, few in this country heat with coal. And certainly there are those who are willing to pay extra to reduce their carbon footprint. And perhaps most importantly, ‘cheaper’ renewables might be rejected because of intermittency (wind, solar) or inferior reliability (hydro, as in a bad water year)?

The point? Comparing straight costs for disparate products or where there are intangibles involved (convenience of a gasoline engine vs. low maintenance on an electric vehicle) can be tricky and allow for wild differences in estimated consumer demand for a product, even without disagreements over future cost trends.

Cost Prediction

Predicting costs from a startup is particularly difficult vs. calculating costs from a commercialized product. But even with the latter, beware of Peak Technology predictions associated with Peak Oil where either falling costs are ignored or overly optimistic projections are made.

In one classic case, Richard Lugar and James Woolsey argued in a 1999 Foreign Affairs article against “optimists” about petroleum who make “assumptions about new drilling technologies that may accelerate production but are unlikely to expand reserves.”[3] On the other end of the spectrum, Michael Liebreich of Bloomberg New Energy Foundation says that he has “raised a reliable laugh by comparing their past clean energy forecasts with historical out-turns.” [4]

What are Costs?

But one regularly confounding effect is the measurement differences between analyses, which tends to be much greater for renewables than conventional power plants.  A coal boiler or gas turbine will have roughly the same output everywhere, but solar and wind output varies enormously depending on local conditions.  Further, some studies, such as M.I.T.’s The Future of Solar Energy,[5] indicate the impact of subsidies on apparent costs, but most do not. Advocates especially tend to downplay their role. And the costs of back-up power for renewables is frequently not included.

Thus, as Williams and Hittinger note,[6] estimates of learning cost trends range from -3 to +33 % (with a doubling of capacity), apparently due to definitional and measurement discrepancies. Similarly, a comparison of some estimates of the cost trends for lithium-ion batteries found they ranged from 4.4% to 36% per year decline from 2010 to 2015.[7]

And too many analysts fail to differentiate between cost trends according to the type of progress that is occurring. There is an enormous difference between cost trends where gradual engineering improvements are occurring, and those where a significant breakthrough is required.

For comparison, consider the progress in wind turbines versus the growth of hydraulic fracturing of shale oil and gas. Turbines are better designed and use advanced materials compared to those during the first modern boom era beginning in the late 1970s (which serves as a cautionary tale about rushing to deploy an immature technology.)  Progress has been steady. For fracking, however, the productivity per well jumped by a factor of hundred or more while well costs only tripled; improvements continue but the initial change was enormous.

A good illustration is the way Curtis and Romm described, in 1996, wind power costs that “could” hit 3 cents per kwh by 2020 (about 5 cents in 2018 dollars), which proved roughly correct, but their prediction that electricity from gasified biomass could reach 4.5 cents per kwh in a decade (2006; bout 6 cent in 2018$), was far off the mark, given estimates of 8 to 15 cents per kwh in 2016, a decade later. The former represents primarily engineering advances, the latter presumably an inability to cheaply convert biomatter into biogas, that is, chemical innovation.

Reasons for Cost Changes

Humans have long had a hard time discerning the difference between a transient event and a long-term trend, and this is certainly visible in the many commodity bubbles that have occurred during modern history, each one of which has been described as a new paradigm instead of a cycle. “People got to eat,” is one common refrain, as is, “Buy land, they aren’t making any more of it.”  (I have described in detail how this was true of the political disruptions of oil supply in the past decade, which were misinterpreted as being due to “Peak oil”.)

But this can be a significant problem for renewable energy sources as well.  The case of the Solyndra bankruptcy is often (correctly) cited as an example of the government’s inability to choose successful technologies, but overlooked is the role that a short-term spike in the price of silicon panels (due to rising demand as a result of exorbitant government subsidies) made Solyndra’s previously avoided technology appear attractive.[8]

Similarly, the supposed “floor price” of oil, said to be $100 per barrel, was assumed due to the marginal cost of oil, ignoring the cyclical effect of the drilling boom.  To the extent that recent reductions in the price of solar panels reflects a move of manufacturing to subsidized Chinese facilities, then, would be very important in understanding how much of recent trends are sustainable or a one-time transient.

Learning Curve

Explaining falling costs for renewable energy as a result of the learning curve effect has been common among advocates of subsidies for renewable energy:  costs will eventually drop to the point of economic competitiveness, but early purchases must have incentives to kick-start the trend.

And typically, the correlation between costs for solar panels and the number installed has been demonstrated graphically in many places. This trend has been described by O’Connor (op. cit.) as “the biggest story in solar power”. Similar reviews such as Rubin et. al. have noted a wide range of learning curve estimates, including 10% to 53% for solar power.

But, as others have noted,[9] this approach is assuming correlation means causality:  Installations rise over time and costs fall, therefore attribute the latter to the former. This is a classic case of logical fallacy, and argues for caution in using the numbers estimated.

This is especially true when considering the impact of compounding upon long-term cost trend assumptions. When discussing likely costs a decade in the future, whether the decline is 6% per year or 10% a year makes an enormous difference. In the former case, costs drop 40%, in the latter 60%. The discrepancy grows with time.

Shape of the Curve

Which highlights the extent to which assuming a rate of decline in costs can hide the true nature of the cost curve. Although writers like Romm and Curtis will refer to an annual trend as if it were continuous, and others relate the trend to installed capacity, in reality costs tend to decline quickly with the initial development of a technology, with the trend flattening out over time.

Most predictions of lithium-ion battery costs assume that the first five years, but from the current point, costs are much slower to decline. Straubel (2015) shows a drop in battery costs of 60% in three years (2012-2015) but only 50% in the next decade (2015-2025).[10]

This type of curve is actually logical, but also reflects the uncertainty about costs in the initial phases of development.  The number of battery electric vehicles sold in 2012 in the United States was less than 15,000, so economies of scale had obviously not kicked in.  (Similarly, the cost of early photovoltaics were elevated because they were designed for use in outer space, exaggerating the improvements in the last few decades.)

And as Nyquist et. al. showed, estimates of the cost of lithium-ion batteries in 2012 were very broad, from $300 to $1100 per kwh.[11]

Conclusion

Predicting the costs of a given fuel or technology is fraught with uncertainty, which is exaggerated by the frequently casual approach taken by writers. Although costs have been dropping for the Cleantech products such as solar, wind power, and lithium-ion batteries, costs for a number of other long-advocated energies such as cellulosic ethanol have not.

Part of the problem is comparing early, experimental products with latter, mass-produced items. Another problem is assuming a roughly continuous rate of improvement, instead of rapid cost declines followed by a more gradual trend.

The ultimate lesson is an old one: skepticism should always be applied, especially to the more optimistic predictions.

———————–

[1]   Arbib, James, and Tony Seba, “Rethinking Transportation 2020-2030,” RethinkX, May 2017.

[2]   Hodges, Jeremy, “Electric Cars May Be Cheaper Than Gas Guzzlers in Seven Years,” Bloomberg March 22, 2018.  https://ift.tt/2DLIUh6

[3]   “The New Petroleum,” January/February 1999.

[4]   “In Energy and Transportation, Stick it to the Orthodoxy,” https://ift.tt/2xezcpa

[5]   M.I.T. Energy Initiative, 2015, p. 118.

[6]   Wlliams, Eric and Eric Hittinger, “If We Keep Subsidizing Wind, Will the Cost of Wind Energy Go Down?” August 3, 2017.  https://ift.tt/2wtULx3

[7]   Lynch, Michael C., “Determinants of Peak Oil Demand,” September 2018, p. 55.

[8]   For the cost trend, see O’Connor, Peter, “What is the Learning Curve—and What Does it Mean for Solar Power and for Electric Vehicles?” September 29, 2016.  https://ift.tt/2IdgkrT

[9]   A good review is in Hogan, William W., “Clean Energy Technologies:  Learning by Doing or Learning by Waiting?”  Energy Policy Research Seminar, September 29, 2014.

[10]   Straubel, J. B., “Energy Storage, EV’s and the Grid,” 2015 EIA Conference, 6/15/15.  Slide 22.

[11]   Nyquist et. al. Nature, cited in http://www.iflscience.com/technology/battery-costs-drop-even-faster-electric-car-sales-continue-rise/

The post On Energy Cost Trends (applying caution to the big talk of energy transformation) appeared first on Master Resource.

via Master Resource

https://ift.tt/2pM6Qx0

March 29, 2018 at 01:27AM

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s