Chapter 6. Energy Supply: How Much Will We Have? How Much Will We Need?
There is simply no way to accurately forecast exactly how much total energy is likely to be available in our 100 percent renewable future. There are too many variables at play—some technical, others economic or political. A few of the factors impacting future energy supply are favorable—including falling prices, technical improvements, and a generally favorable public attitude toward solar and wind. However, other factors that we have just surveyed pose challenges, including source intermittency, the need for storage and grid redesign, and the difficulties of electrifying heavy transport and many industrial processes. On balance, we believe the preponderance of factors support the assumption that energy quantities will be lower, perhaps significantly lower, than business-as-usual global energy demand projections from official agencies such as the International Energy Agency (which expects demand to rise 1.5 percent a year through 2035, nearly doubling over 2009 levels by 2050).
This chapter explores in more detail why energy supplies are likely to be constrained in an all-renewable future, and then examines what this means. Some of the questions we’ll address along the way include the following:
- Will we need less energy due to more efficient usage and the reduced need for energy conversion?
- Can we decouple energy growth from economic growth?
- Can we use renewable energy to build more renewable energy production capacity?
Perhaps the overall challenge of replacing all fossil-derived energy while continuing to grow the economic benefits of energy supplied to society can best be appreciated in historic terms. Humanity’s past energy transitions (from wood to coal, coal to oil and natural gas) were driven by economic opportunity, not policy, and new types of energy were usually additions to, rather than replacements for, existing energy sources. Total supplies expanded quickly as the mix of energy sources evolved. However, what the world needs to do now is largely unprecedented—to force a rapid, policy-driven replacement of existing energy resources without sacrificing the benefits of the incumbent energy system, knowing that the characteristics of new energy resources may partially compromise the outcome.
The theoretical potential for solar and wind is vast. As noted earlier, the total amount of energy absorbed annually by Earth’s atmosphere, oceans, and land masses from sunlight is approximately 3,850,000 exajoules—whereas humanity currently uses just over 500 exajoules of energy per year from all sources combined, an insignificant fraction of the previous figure. If only 0.014 percent of the energy flow of sunlight could be captured, it would be enough to satisfy current world electricity demand. The potential is vast; however, as we have noted, limits are likely to be encountered in scaling up the technology required to harvest these enormous ambient energy flows, and in adapting current energy consumption patterns to using variable sources of electricity. And constructing enormous numbers of solar panels and wind turbines requires materials and energy, as well as financial capital.
The goal of this chapter is not to forecast future energy supplies quantitatively (as already noted, there are just too many variables), but rather to explore some additional factors that will impact levels of supply—and also to explore the relationship between energy supply levels and economic growth.
Energy Returned on Energy Invested of Renewables
To know how much useful energy we will have in an all-renewable world, we have to adjust assumed gross energy production figures by subtracting the energy invested in energy-producing activities. This tells us the net energy available to do useful work (as discussed in chap. 1). If the energy returned on energy invested (EROEI) ratio for the future renewable energy system (including not just panels and turbines but storage technologies and grid enhancements as well) is significantly lower than that of our current energy system, then even if total energy production stays the same, the amount of useful energy will decline.
When considering EROEI study results, it is helpful to keep two threshold numbers in mind. Charles Hall and others argue that returns above 3:1 are needed for an energy resource to be viable, while society needs a much higher overall EROEI (above 7:1) to support energy-consuming activities like education, health care, research, and the arts.
Unfortunately, although EROEI studies are key to the economic evaluation of energy sources, the status of the net energy literature is far from satisfactory. Differences in methodology tend to yield widely ranging EROEI estimates for the same energy source. For example, EROEI studies of wind power have yielded results varying from 1.27:1 in a 1983 German study to 76.92:1 in a Danish study in 2000. (In fairness, the technology and economics of wind power changed significantly between 1983 and 2000.)
EROEI is a system-level evaluation of a particular energy production pathway embedded in a specific industrial/economic network. Still, when the inherent complexities of the discussion and methodological differences are accounted for, it seems clear that some renewable energy production pathways have a much lower EROEI than those for most commercial fossil fuels. Indeed, the EROEI of some renewables is too low for them to serve as viable, self-sustaining energy sources; this is almost unquestionably the case for corn-based ethanol production in the United States, for example.
There is some controversy as to whether solar photovoltaic (PV) systems also have too low and EROEI to power industrial societies. A study by Marco Raugei concludes that the EROEI of PV technologies ranges from 19:1 to 38:1; if these numbers are verified, then PV systems should easily be able to provide energy to operate industrial societies while in energy terms also “paying for” their own production and maintenance. However, a comprehensive operational study conducted by a pioneed of EROEI research (Charles Hall) and the manager of several of Spain’s largest industrial PV power-generation facilities (Pedro Prieto) came to starkly different conclusions. Prieto and Hall calculate an EROEI for Spanish PV of 2.4:1 to 7:1, depending on boundries chosen. Graham Palmer arrives at similar results in his EROEI analysis of PV in Australia. If verified, the Prieto-Hall and Palmer estimates would be very discouraging for the energy transition. However, it should be noted that the Prieto-Hall study has been criticized for its methodological inconsistency with other studies. The authors start with a project-level analysis (of a single panel or PV farm, which would produce a full life cycle energy profitability metric) but then switch to an analysis of the entire PV industry in Spain for a given year without discussing the implications of the switch–that is, that energy-flow analysis is dynamic and factors like industry growth rate will impact the result. Critics of the Prieto-Hall study argue that it makes little sense to compare a flow-based EROEI with a full life cycle EROEI without correcting for the growth rate of the industry, which was not done. A more recent meta-analysis by Bhandari et al. suggests a range of EROEI for PV of 8.7 to 34.2, depending on the technology and its siting. These figures are generally supportive of Raugei’s results and, while lower than the energy return numbers for conventional fossil fuels during their heyday, are still high enough to support an industrial economy.
The EROEI of wind has been the subject of less controversy, with a meta-analysis of fifty studies suggesting a likely value of 19:1 for systems in place.
Examination of the EROEI of energy sources per se may not give us an accurate view of the energy costs associated with different energy systems. If electricity storage and redundant capacity are required to buffer the intermittency of solar and wind, then these systemic energy costs need to be taken into account as well. A study by Weissbach et al. showed that the EROEI of a solar or wind energy system is reduced roughly by half when energy storage is added to the analysis. This confirms a conclusion many energy analysts have already arrived at on the basis of economic calculations alone: that in designing renewable energy systems it is preferable to minimize the need for storage and redundant capacity through demand management wherever possible.
We are still at too early a stage in renewable energy deployment to know how much storage and capacity redundancy will be needed, and we are at too early a stage in EROEI studies to be able to judge whether the more optimistic or more pessimistic results for PV are more accurate. However, if it turns out that high levels of storage are required and that the middle-of-the-road EROEI figures for solar PV of 10:1 and for wind of 19:1 (without storage) are justified, then as society transitions away from high-EROEI fossil fuels its overall economic efficiency may decline, as a somewhat higher proportion of produced energy will have to be reinvested into further energy production. This may have implications for the possibility of further economic growth, as we will consider later in this chapter.
Building Solar and Wind with Solar and Wind
The rapid build-out of renewables constitutes an enormous infrastructure project that will itself consume significant amounts of fossil-fuel energy (fig. 6.1). While it is possible to imagine a solar panel or wind turbine factory operating solely on electricity supplied by renewable electricity, it is much harder to envision entire supply chains—from the mining of ores to the final delivery and installation of panels and turbines—functioning without fossil energy, at least in the early stages of the transition.
As we saw in chapter 4, fossil fuels are currently used for mining raw materials, constructing roads and factory buildings, and transporting raw materials and finished products. Theoretically solar and wind technologies could supply the energy for these processes, using electric mining, manufacturing, and hauling equipment (perhaps, for example, electricity could be produced on-site and transmitted via cables to mining equipment). Fossil fuels are also used to supply high levels of heat for extruding aluminum, making copper wire and plate, and producing iron and cement. Solar and wind electricity can in principle produce high heat for these purposes. However, as discussed in chapter 5, it would be much more expensive to generate the temperatures needed with electricity from solar panels or wind turbines than from burning fossil fuels. This would add to the cost of renewable energy technologies. To the authors’ knowledge, no real-world pilot projects exist in which all the industrial processes involved in making renewable energy technologies are powered by renewable energy.
A bootstrap transition scenario (in which renewables provide the energy needed to build more renewables, while still supplying much of the rest of the energy that society needs) seems daunting in principle. Where will the energy for the transition come from, then? Realistically, most of it will have to come from fossil fuels—at least in the early-to-middle stages of the process. And we will be using fossil fuels whose economic efficiency is declining due to the ongoing depletion of existing stocks of high-quality oil, gas, and coal. Again, this implies higher overall costs. But using only renewable energy to build renewables would be slower and even more expensive.
The faster we push the energy transition, the more energy will have to be diverted to that gargantuan project, and the less will be available to all the activities we’re already engaged in (running the food, transportation, manufacturing, communications, and health care sectors, among others). Moreover, a faster transition will delay the point at which large amounts of useful net energy are available from newly installed renewable energy generators.
If fossil fuels will be required for constructing solar panels, wind turbines, and the infrastructure that enables us to use them, then high build-out rates for renewable energy technologies may have implications for carbon emissions. The faster we push the transition, the higher the emissions—unless we rapidly curtail current uses of fossil fuels in the meantime (reducing fossil energy consumption faster than it can be replaced by renewable energy), implying a reduction in energy consumption and therefore in gross domestic product (GDP).
A realistic assessment of future energy availability would also have to take into account the requirement for financial investment capital. While solar and wind have enjoyed rapidly increasing rates of installation during most of the past decade, transition plans envision an even more rapid shift, involving much higher levels of investment in generation capacity, storage, grid upgrades, and transport alternatives. Will sufficient money be available?
The affordability problem is finessed in some published energy transition studies. For example, in a recent plan for a conversion of the U.S. economy to 100 percent renewable energy by 2050, Mark Jacobson et al. count savings from avoided costs of climate change and health damage in their estimate of the affordability of such a comprehensive and rapid conversion. However, as discussed earlier, avoiding externalized costs associated with fossil fuel consumption might make the renewable energy transition more affordable on a society-wide basis, but that does not actually mean the transition will be affordable on its own terms.
Estimating how much a total energy transition would cost is difficult. The problem is simplified greatly by including only the direct cost of solar panels and wind turbines, but doing so is unrealistic. Actual costs would include required investments in new technology for the transportation, agriculture, and manufacturing sectors; in new equipment for building operations, and for energy efficiency retrofits; in grid redesign; in energy storage; and in redundant generation capacity. For the average American household, costs for installing insulation, an air-source heat pump, an electric stove (assuming they currently have a gas stove), and a solar water heater with on-demand electric water heater backup would run into many thousands of dollars; this does not include the cost of an electric car (we assume the average family will be trading out its current car at some point anyway) or solar panels and batteries (our hypothetical family may choose to buy grid-supplied renewable electricity). Just multiplying these outlays by the number of American households yields figures in the hundreds of billions of dollars, but this does not include the far greater costs to utilities, or the research and development and retooling costs in the energy-consuming industries just mentioned. In a Scientific American article in 2009, Mark Jacobson estimated the total cost of the transition at about $100 trillion, spread over 20 years. However, this includes primarily energy supply requirements and excludes the necessary investment in revamping all economic sectors on the consumption side. This latter could easily match the necessary investment in energy supply.
Actual rates of investment in renewable energy globally have leveled off in the past four years (fig. 6.2), with investment rates in North America and Europe stalling or shrinking while China continues to surge ahead.
In 2014 the world’s nations invested $270 billion in renewable energy (92 percent of that was for solar and wind), which represented roughly one-sixth of all energy spending. Overall, investments in conventional fossil fuel production continue to dominate.
Jacobson’s estimate of the energy supply cost of the transition ($100 trillion over 20 years) amounts to $5 trillion per year in required investment. With the current investment rate stuck at around $270 billion per year, it is clear that rates of investment will have to increase by a factor of more than 10 if we are to come close to supplying sufficient energy from renewables to replace all current energy supplied by fossil fuels. The world currently spends $1.8 trillion annually on military activities, so the required investment rate should not be ruled out as unrealistic in principle. However, the scale of what is needed is breathtaking.
Funding for enormous new infrastructure spending projects is difficult to organize unless economic and financial systems are stable and expanding. One of the authors of this book has argued elsewhere that three converging factors (too much debt, rising energy costs, and increasing environmental stress) are leading to the end of economic growth as it was known during the latter half of the twentieth century. Real economic growth has indeed slowed in the world’s wealthy industrial nations in the past couple of decades. Since the 2008 crisis, central banks have deployed low interest rates and quantitative easing, and governments have bailed out banks and major industrial firms while engaging in deficit-funded stimulus spending. In theory these actions should have produced a robust recovery, but the result has instead been more commodities, stock market, and real estate bubbles—with almost all the benefits going to society’s wealthiest. Further, the marginal productivity of debt—the amount of additional GDP produced by one dollar more of debt—has plummeted from around $3.00 in the 1950s to near zero today, indicating that debt is no longer providing the economic boost it did in the past. Meanwhile, real living standards in the United States and much of Europe drift lower. A fairly robust literature is developing to attempt to account for this “secular stagnation,” which many economists now think could continue for decades.
When economic growth ceases, as it does in times of recession, investment capital tends to become scarce. Thus scarce investment capital could pose a barrier to a robust renewable energy transition. The Keynesian solution for recession is for governments to become the borrowers and spenders of last resort in order to prime the growth pump. Could governments and central banks, following the Keynesian formula, simply print the money needed to fund the energy transition? This is just one of many currently unanswerable questions we are likely to encounter along the path toward a renewable future.
The Efficiency Opportunity: We May Not Need as Much Energy
In the production of electricity from coal and natural gas, about 60 percent of the primary energy contained in the fuel is lost in the conversion process. Solar and wind electricity sources do not require a conversion process and therefore do not incur these high losses. This amounts to a substantial amount of potential energy savings: out of 197 billion gigajoules of primary energy currently flowing to the entire global electricity sector, 117 billion gigajoules wind up as conversion losses; this is energy that will no longer be needed in an all-renewable future.
In addition, electric motors are significantly more efficient than internal combustion engines. While the latter are only 20 to 30 percent efficient (with most of the energy contained in gasoline lost as waste heat), electric motors can be 92 percent efficient at translating energy into motive force. Thus the more we electrify transportation and other uses of combustion engines, the less energy we will need in order to produce the same economic and social benefits. This has practical implications for the energy transition. In the United States, passenger vehicles currently use about as much energy in the form of gasoline as is consumed in the entire electricity sector. But transitioning to electric cars would not require a doubling of electricity generation; we could do it with about 29 percent additional electricity.
As discussed in chapter 5, a great deal of energy could also be saved in space conditioning if all homes and buildings had passive house levels of efficiency. Assuming a generous 90 percent cut in energy use for this purpose, consider another 350 Mtoe or so energy saved.
Of course, to obtain a realistic estimate of overall energy savings we should also consider some inefficiencies that an all-renewable energy system might bring with it. One of these is tied to storage: storing a ton of coal or a gallon of gasoline implies little direct loss (though there are costs for the tanks and other storage infrastructure), while electricity storage always involves losses. The percentage of electricity that would be lost in storage annually in an all-renewable future would depend on a range of factors, including the types of storage used and the degree to which storage is used to buffer intermittency (as opposed to using capacity redundancy or demand management for this purpose). Also, if grids were expanded to enable load balancing over longer distances, this would entail higher electricity transmission losses. Still, on balance, there are very large opportunities for energy savings, though many of these would take time and substantial investment to realize. An electrified, optimally efficient society might need only half to two-thirds of current primary energy consumption to yield similar economic benefits.
All published renewable energy transition scenarios highlight this opportunity for obtaining equal economic benefits from reduced primary energy consumption. Most go further and assume that even greater reductions in energy use can be achieved while still supporting economic growth. But this assumption is controversial, as we are about to see.
Historically, there has been a close correlation between energy use and economic activity (see fig. 2.2). Increased energy consumption is associated with economic growth; during times of economic recession, energy consumption often declines. This correlation makes sense, as everything we do requires expenditure of energy. Policy makers do not want to sacrifice prospects for economic growth in order to curtail fossil energy sources in favor of solar and wind. Yet there are good reasons to conclude that the energy transition will leave us with less useful energy than historic trends would lead us to expect. Is it possible to stretch the link between energy consumption and GDP growth so as to have more of the latter with less of the former?
Energy intensity (measured as the ratio of the consumption of final energy, meaning usable forms of energy, such as heat or electricity, to GDP) varies from nation to nation. There is evidence that the energy intensities of both the United States economy and the global economy have indeed been falling (fig. 6.3), though a recent study by Wiedmann et al. suggests that historic “decoupling” of economic growth from increased energy usage has been significantly overstated. The reasons for energy intensity improvements are summarized as follows in a paper by Jesse Jenkins and Armond Cohen:
- Sectoral shifts in the composition of the global economy, such as the increasing importance of services as a share of global GDP, which tend to expend much less energy per unit of economic activity than heavy industry or agriculture;
- Substitution of other economic inputs for energy, such as an increased reliance on capital or labor in productive processes in lieu of energy inputs;
- Improvements in primary to final energy conversion efficiency, or the efficiency at which primary energy supplies, such as coal, oil, or renewable energy inputs, are converted to usable, final forms of energy such as heat or electricity;
- Improvements in end-use energy efficiency, or the amount of final energy inputs needed to deliver a given energy service, such as heating, cooling, transportation, or industrial process energy inputs.
Is there reason to think energy intensity can be reduced significantly as we transition to renewable sources? The energy savings from slashing energy conversion losses (no. 3 in the preceding list) and from the replacement of combustion engines with electric motors (no. 4 in the preceding list) discussed in the previous section would almost certainly drive considerable further improvement in energy intensity. But these strategies have limits.
In a review of seventeen published decarbonization scenarios, Loftus et al. found that all of the scenarios rely upon improvements in energy intensity that are unprecedented in history. Three scenarios that exclude nuclear and carbon capture and storage technologies (i.e., the ones that depend almost entirely on growth in wind and solar power) require the fastest energy intensity improvements. The authors also noted that all of the studies they surveyed offer little detail on how to decarbonize the industrial and transportation sectors, and on needed energy system transformations.” Loftus et al. conclude with the following comment, with considerable relevance for this book: “To be reliable guides for policymaking, scenarios such as these need to be supplemented by more detailed analyses realistically addressing the key constraints on energy system transformation.”
A 2014 report by PriceWaterhouse Coopers notes that the decoupling of emission growth from economic growth has averaged only 0.9 percent per year since 2000. This raises questions about the prospects for meaningful reductions in energy intensity beyond what can be achieved by reducing conversion losses and replacement of combustion engines with electric motors. Industry already has a cost-cutting incentive to improve efficiency; policy makers may have limited ability to increase the rate of efficiency improvements above this “exogenous” background rate.
The Role of Curtailment and the Problem of Economic Growth
If we won’t have as much energy, and we can’t improve efficiency at a continuous and dramatic rate—and therefore energy intensity cannot be reduced at unprecedented rates—then the economy will likely shrink. Rather than merely streamlining economic activities, we will have to curtail them, at least to a certain degree. Perhaps aviation offers the most pertinent example: as we have seen (in chap. 4), there are no easy or inexpensive substitutes for kerosene-based jet fuels, and so it is difficult to imagine the continued growth of this industry as carbon-based fuels are fairly quickly eliminated. Altogether, it is difficult to avoid the conclusion that an all-renewable future will offer less economic growth, no growth, or negative growth. But then again, the world is already seeing a reduction in economic growth rates. Since fossil fuels are finite, they cannot fuel perpetual growth in any conceivable instance. Thus it would be specious to argue that we face a choice between renewable energy and reducing greenhouse gas emissions on one hand, and economic growth from continued reliance on fossil fuels on the other.
A few climate scientists have already suggested that dealing with global warming could have serious implications for the economy. Kevin Anderson and Alice Bows of University of East Anglia’s Tyndall Centre for Climate Change Research have calculated that a “carbon budget” consistent with a threshold of 2°C would entail an 8 to 10 percent annual reduction of emissions in industrialized nations, which would be, in Anderson’s words, “incompatible with economic growth.” The 1.5°C goal referenced in the international agreement reached in Paris in December 2015 makes the challenge of achieving growth while massively reducing greenhouse gas emissions even more daunting.
The Intergovernmental Panel on Climate Change’s (IPCC’s) Fifth Assessment Report (2014) admits the difficulty of the renewable energy transition in this regard. “No single mitigation option in the energy supply sector will be sufficient,” the report warns. To stabilize the climate at an average global surface temperature no higher than 2°C above the preindustrial level, scenarios relying almost entirely on solar and wind energy would, in the opinion of the IPCC report’s authors, require global energy supply to be radically curtailed below currently projected demand.
Again: we cannot estimate how much energy will be available in an all-renewable future, other than to suggest that it will probably be significantly less than business-as-usual demand projections. Thus the energy transition constitutes an important challenge not just for scientists and engineers but for economists and policy makers as well. How shall we maintain social and material benefits to the world’s people as population continues to grow, but energy availability declines and economies stall and contract?
The tapering of economic growth really should come as no surprise: a long-standing school of thought says that physical expansion cannot continue forever on a finite planet. However, tapering presents serious challenges not just for political and economic systems but for the renewable transition itself: how are societies to obtain sufficient funding for the rapid and dramatic expansion of renewable energy infrastructure if their economies are stagnant rather than growing? Perhaps the worst outcome of all would come from a failure to plan for economic tapering: in that case, societies would deploy futile strategies to restart growth, while frittering away opportunities to prepare for a renewable, postgrowth future.
Of course, the fossil fuel lobby uses fears of economic hardship as an excuse to say that the energy transition should be delayed as long as possible. However, the reverse is true: the longer the transition is delayed, the more expensive and perilous it will become. The world’s remaining high-quality and inexpensive-to-produce fossil fuels are depleting rapidly, so if a transition to alternative energy sources is not organized rapidly, economic contraction will still result. But in that case, we eventually end up with catastrophic climate change and no viable energy system.
As many economists have pointed out, GDP growth is a poor indicator of societal progress or well-being. For example, if power plant emissions are reduced due to the expansion of renewable energy, this could result in a decline in hospital stays and drug prescriptions related to asthma attacks, and this would in turn lead to lower GDP, even though it reflects an improvement in well-being. Extreme storms damage buildings, which then need to be repaired, increasing GDP—but well-being has of course declined in the process. GDP is the sum of all consumption in the economy (household, business, and government, along with net exports); thus we measure our well-being by how much we consume, and we have trapped ourselves into believing that this quantity must increase year after year.
Replacing GDP with a more robust and realistic measure of economic success is just one of the tactics proposed by postgrowth economists, such as Peter Victor, who recognize that the rapid expansion of population and consumption that characterized the twentieth century will inevitably subside in the decades ahead. Victor and others propose ways to promote full employment and higher quality of life as consumption of energy and materials declines.