Energy efficiency, sustainability and economic growth

Energy efficiency, sustainability and economic growth

ARTICLE IN PRESS Energy 32 (2007) 634–648 Energy efficiency, sustainability and economic growth Robert U. Ayresa,b, Ha...

353KB Sizes 0 Downloads 73 Views


Energy 32 (2007) 634–648

Energy efficiency, sustainability and economic growth Robert U. Ayresa,b, Hal Turtonb,, Tom Castenc a

INSEAD, Fontainebleau, France IIASA, Schlossplatz 1, A 2361 Laxenburg, Austria c Primary Energy LLC, 2000 York Road, Suite 129, Oak Brook, IL 60523, USA b

Received 16 April 2006

Abstract This paper explores two linked theses related to the role energy in economic development, and potential sources of increased energy efficiency for continued growth with reduced greenhouse gas (GHG) emissions. The first thesis is that, while reduced GHG emissions are essential for long-term global sustainability, the usual policy recommendation of increasing energy costs by introducing a carbon tax may be relatively ineffective under current market structures and have an unnecessarily adverse impact on economic growth. Our second thesis is that there exists a practical near-term strategy for reducing GHG emissions while simultaneously encouraging continued technology-driven economic growth. Moreover, this strategy does not require radical new technologies, but rather improved regulation or—more precisely—better deregulation of the electric power sector. In respect to the first of our two theses, this paper addresses a deficiency in neoclassical economic growth theory, in which growth is assumed to be automatic, inevitable and cost-free. We challenge both the assumption that growth will continue in the future at essentially the same rate (‘‘the trend’’) as it has in the past, and the corollary that our children’s children will inevitably be richer and better able to afford the cost of repairing the environmental damages caused by current generations [[1]Simon et al., The state of humanity. Cambridge MA: Blackwell Publishers Ltd.; 1995]. r 2006 Elsevier Ltd. All rights reserved. Keywords: Energy efficiency; Economic growth; Sustainability; Exergy

1. Economic background Energy (actually exergy) is as essential to the functioning of the global economic system as gasoline is to a car or electricity to a light bulb. The evidence is visible and pervasive. It is the sun’s energy that drives the most fundamental process in nature, photosynthesis, whereby carbon dioxide and water are converted into carbohydrates, lipids and proteins (biomass). The biomass produced in the distant past was converted by natural geological processes into the fossil fuels we utilize today to power the world’s factories and transport systems, heat its homes and generate much of its electricity. It is also (mainly) the wastes from combustion that are now driving climate change, perhaps irreversibly.

Corresponding author. Tel.: +43 2236 807493; fax: +43 2236 71313.

E-mail address: [email protected] (H. Turton). 0360-5442/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/

Nevertheless, macro-economists traditionally underestimate the importance of energy. Standard textbooks include such curious illustrations as the theoretical bakery that employs capital (rented) and labor to produce bread, without either flour or fuel [2].1 To be sure, economists have not completely ignored energy or more precisely, energy that is available to do useful work (technically, exergy), especially since the events of the 1970s. However, this is not the place to review the history of attempts to incorporate energy explicitly as an independent variable in the standard production function approach. Suffice it to 1 This typical example can be found in the popular recent textbook entitled ‘‘Microeconomics’’ by Harvard Economics Professor N. Gregory Mankiw, who is currently the Chairman of the Council of Economic Advisors to the President [2]. The purpose of the example is to explain why the national accounts of an economy in equilibrium will pay capital and labor, respectively, in accordance with their respective productivities. Energy is not mentioned in the example, the theorem or the national accounts.

ARTICLE IN PRESS R.U. Ayres et al. / Energy 32 (2007) 634–648

say that the first attempts were not successful in endogenizing and explaining historical economic growth. In retrospect, it appears that this failure was mostly due to a doubtful assumption, namely that the marginal productivities of the production inputs must be proportional to the shares of payments to factors (e.g. capital, labor and energy), in the national accounts.2 Energy does not appear explicitly on the payments side of the national accounts. However, payments to ‘energy’ can be roughly equated to revenues of certain industries, such as coal mining, petroleum and gas drilling (and value added in refining) and—possibly—electricity generation and distribution. Based on this approximation, it turns out that the energy ‘share’ of payments in industrial countries is negligible, not more than a few percent of GDP—around 4% for the US, for instance. Most economists in the past have assumed, despite intuition to the contrary, that (due to the income allocation theorem mentioned previously) energy could not be very ‘productive’ as compared with capital or labor.3 The standard escape from this dilemma has been to assume that energy is not really a primary variable (i.e. a factor of production) but, rather an intermediate product of the economy. The argument is that capital and (human or animal) labor—combined as in the construction of a watermill, or digging a coal mine—produce the materials and energy intermediates that are subsequently converted into products and final services. This explanation seems plausible at first sight, and has rarely been challenged. However, labor and capital are also produced, and every process and activity in the economy requires exergy, some of which (like sunlight for photosynthesis) is by no means an intermediate product. Considering labor and capital alone ignores the fundamental role of solar exergy as a direct input to agriculture, or as the source of natural forces like wind and flowing or falling water. The food or feed consumed by the human or animal workers is essentially pure exergy. Moreover, exergy is embodied in finished materials, like steel, aluminum or plastics, or in silicon chips, and thus in all material goods. Finally, exergy is needed to transform raw materials into finished goods and to move them to consumers (and finally to dispose of wastes). In fact, neither labor nor capital could exist without exergy. Exergy is a complement—as well as a substitute—for capital, while capital and labor are often complementary as well. Thus the traditional assumption that the factors of production must be substitutes in the production function must be reconsidered. It has been well known since the 1950s that economic growth cannot be explained by the simple accumulation of invested capital, still less by a growing labor force [5,6]. The unexplained ‘Solow residual’ is now called total factor 2 It has been convenient to divide all payments into two categories, namely wages and salaries (returns to labor) and dividends, royalties, interest and rents (returns to capital). 3 For instance, see [3,4].


productivity (TFP). Much econometric effort has been devoted to uncovering the determinants of TFP [7,8] and with remarkably little success. The reasons for this failure cannot be analyzed here, except to note that econometricians have consistently neglected to include energy or useful work among the variables tested. However it has been standard practice for model builders and forecasters to assume that TFP is an exogenous ‘driver’ that will continue to increase automatically, and even more recent efforts to model TFP endogenously—while still ignoring the role of useful work—have been unable to account for a significant proportion of growth [7,8]. This, in turn, means that many of today’s models of long-term economic activity assume that changes in the supply of energy or the demand for energy services (useful work) have no significant impact on economic growth. We challenge this view.

2. Digression: on exergy, power and useful work Before introducing an alternative perspective on the relationship between energy and economy activity (and growth) it is important to focus on the essential distinction between ‘raw’ energy (e.g. from the sun) and the services actually performed by energy and utilized by the economy. These services are called useful work.4 The more familiar term, power, is simply the rate at which useful work is produced by a prime mover, such as a steam-electric generator or an internal combustion engine. A technical explanation of the term work would involve reference to textbooks on thermodynamics. For present purposes, however, useful work can be divided into three major categories, which will (hopefully) also help to clarify the meaning. The first category consists of muscle work, performed by humans or animals. The most common examples, in the West, would be the use of horses for transport or plowing. The second major category is mechanical work. Mechanical work is mostly done by heat engines (so-called ‘prime movers’) such as steam turbines, gasoline engines, diesel engines or gas turbines. Electric power is the output of a combination of a prime mover and an electric generator. Electric power can, in turn, be reconverted into mechanical work at some distant point by means of a motor, or it can be reconverted into light, heat (as in a toaster) or into chemical work (as in electrolysis). Finally the third major category consists of heat, whether 4 In this entire discussion, the term ‘energy’ is really a misnomer, despite its common usage. What we really mean in this context, is ‘available energy’ which has the technical name ‘exergy’. Available energy is that fraction of total energy that can be converted into ‘useful work’, as defined in physics and engineering texts. Examples of work, in the technical sense, would be lifting a weight against the force of gravity or overcoming rolling or sliding friction. Unfortunately, ‘work’ is another ambiguous and misleading term, often confused (especially by economists) with what human workers do for employers or, if self-employed, for purposes of providing food, shelter and otherwise improving their well-being.


R.U. Ayres et al. / Energy 32 (2007) 634–648

at high temperature or low temperature, delivered to a point of final use. We note that in common language usage, energy is ‘converted’ into work. But the units we use to measure useful work (e.g. horsepower-hours, or kilowatt-hours) are also ipso facto units of energy. This is a source of much confusion, and probably accounts for the fact that most economists have failed to distinguish clearly between exergy, which is an input, and useful work, which is an output. However, the ratio of the output (useful work) to the input (raw exergy, or fuel) is a pure number that lies between zero and one (unity). This ratio is known as the efficiency of conversion. As an example, the efficiency of conversion of heat energy from fuel to electric power delivered to consumers in industrial countries is roughly 33–35%.5 We have already noted that economic activity is driven by available energy, or exergy, in various forms. In fact, it is useful work that drives production, along with capital and labor as traditionally defined. (Here we must again emphasize the distinction between labor in the economic sense, and useful work in the physical-engineering sense.6) Before the industrial revolution, human or animal muscles provided most of the work, with a tiny contribution from windmills, sails and watermills. Economic growth since then has been a long process of substitution of machines, driven mostly by the combustion of fossil solar energy, for human and animal muscles. Fossil fuels also drive almost all of the metallurgical and chemical processes that create modern materials from steel to plastics. 3. The growth connection What is the link between useful work output and economic growth? It is conceptually very simple, though much less simple to demonstrate quantitatively. In brief, as technological progress makes the conversion process from ‘raw’ exergy (e.g. fuels) to useful work more efficient, the cost of ‘useful work’ tends to decline. As costs fall, in a competitive market, prices fall also. Declining prices of work generate increased demand for useful work, throughout the economy. (We have already noted that work and/or embodied exergy, are required at every stage of the economic system from extraction to finished goods and final services.) Increased demand, in turn, requires bigger production units with greater economies of scale, economies of 5 The astute reader may notice that if the electricity delivered is subsequently converted to a secondary service, like electric light (only 10% efficient) the overall efficiency with which the economy produces the service of illumination is only 3%. In fact, one of the authors has estimated that the US economy is currently operating at an efficiency (in the technical sense) of the order of 3%, or less [9]. Clearly there is still a lot of potential for energy conservation. 6 What most human workers do nowadays in industrialized countries involves sensory inputs, eye–hand coordination, judgment and decisions, with very little use of muscle strength as such.

adoption, and learning-by-doing (increasing experience). All of this requires more capital investment, of course, but not necessarily more labor. Economists have generally simplified this relationship by reversing the implied causality, i.e. by assuming that investment drives growth. This is another way of expressing Say’s law, the idea that ‘supply creates its own demand’. In reality, we argue that supply and demand drive each other, although their relative importance may fluctuate from one time period to another, as in the business cycle. In the short run consumer demand tends to drive employment. On the other hand, employment generates income which drives demand, which is why Keynes recommended deficit spending as an antidote for recession. In the longer run, increasing exergy conversion (to useful work) efficiency drives growth via declining prices, thus increasing demand for all products and services. This, in turn, spurs new investment, further economies of scale, learning-by-doing, R&D, and further declines in prices, leading to additional demand (according to the price elasticity) as illustrated in Fig. 1. This is a positive feedback cycle wherein growth, per se, generates further growth.7 The long-run importance of this positive feedback cycle—the primary ‘engine’ of economic growth—has been quantified in earlier publications [10–15]. We explicitly introduce ‘useful work’ (in the thermodynamic sense) as a third variable in a suitable production function that satisfies the usual constant returns-to-scale condition (Fig. 2), noting the useful work is the product of exergy input multiplied by the efficiency of conversion. The estimation procedure for calculating useful work and underlying data have been published elsewhere recently and are now available in the economics literature. The published details cannot be reproduced here for obvious reasons,8 but one important thing to note here is that the LINEX production function in Fig. 2 does not assume that output elasticities are constant, in contrast to the


We note, immediately, that economic growth before the industrial revolution was very slow, because the primary source of useful work—by human and animal muscles—could not be increased in efficiency. In other words the productivity of muscle work remained virtually constant. It follows that growth in GDP in every country is directly correlated with the mechanization of agriculture and the movement of rural agricultural workers—no longer needed to work the soil—to towns and cities, where they became available to work in factories using machines. One reviewer has helpfully noted that the cycle in Fig. 1 cannot necessarily be considered absolute, and for example diseconomies of scale are theoretically possible and quite possibly already occurring. In fact, in some respects the main focus of this paper is to explore means to overcome some of the diseconomies of scale in electricity production. 8 The calculation of historical time series on useful work for the US economy is in Ref. [12]. Econometric estimations of US economic output and comparison with actual data from 1900 through 1998 are in Ref. [15]. Historical estimates of the efficiency of electricity use (secondary work) in the US from 1900 through 2000 are in Ref. [16]. The methodology is currently being applied to other industrial countries for which long timeseries data are available, notably Japan, the UK and Sweden.

ARTICLE IN PRESS R.U. Ayres et al. / Energy 32 (2007) 634–648


Lower unit price P = mC

Declining unit costs (due to scale economies & learning-by-doing) C = (c + N) -b t


Increased consumer demand for products (due to price elasticity) ln P ln Y =-O t t

Increasing labor productivity





Investment to increase physical capacity (& scale of production); substitution of capital & natural resources for labor

C = unit cost c = initial unit cost N = cumulative number of units b = learning exponent P = price m = ratio of Price to Cost Y = total output (demand) t = time sigma = price elasticity

Fig. 1. Salter cycle growth engine.

{( (

Y t = U exp a 2 -




(( + ab ( U -1 ({

For the USA, a = 0.11, b = 5.18 Corresponds to Y = K 0.45 L0.07 U 0.48 • Y = total output • U = usefulwork • L = labor • K = capital • a, b = parameters, t = time

standardized value (1900=1)

30 25 20

GDP Capital (K) Labor (L) Work (U)

15 10 5

Fig. 2. The production function can be either CD, or LINEX. 9

Cobb–Douglas production function. The GDP history for the US from 1900 through 1998 and the factors of production (including work) are shown graphically in Fig. 3. It turns out that US economic growth can be ‘explained’ quite accurately throughout the 20th century by a three-factor production function (with only two independent parameters) (Fig. 4). Note that this model essentially eliminates the need for an exogenous ‘total factor productivity’ (TFP) multiplier, such as is required by conventional economic models (e.g. [7])—that is, this model fully explains economic growth from factor inputs alone. 4. Policy implications of the new perspective on growth The policy implications of this new perspective on growth are quite simple and stark. One major implication 9 In the LINEX function the output elasticities are functions of the factor inputs (K, L and U).

0 1900






Fig. 3. GDP and factors of production of US, 1900–2000.

is that current policy prescriptions by most environmental economists and ‘greens’ are wrong. The need to reduce consumption of fossil fuels and the output of so-called greenhouse gases (GHGs) cannot be denied. But the usual prescription is to ‘get the prices right’—e.g. to raise them— perhaps via a carbon tax. This would presumably cut demand (e.g. [17]). If the standard assumptions about marginal productivity and payment shares in the national accounts were correct, higher energy (exergy) prices would cut energy consumption as required, and yet have little or no effect on the economy as a whole (e.g. [3]). This is because commercial exergy in the form of fossil fuels, nuclear and hydroelectric power accounts for such a small share—around 4% (as noted above)—of the gross national expenditures.

ARTICLE IN PRESS R.U. Ayres et al. / Energy 32 (2007) 634–648



GDP (1900=1)


empirical data including electricity use efficiency excluding electricity use efficiency

15 10 5 0 1900






year LINEX parameters: a=0.11, b=5.18. Fig. 4. LINEX results of US, 1900–2000.

But, because the production of useful work is such an important driver of growth, from the long-term perspective, this prescription for cutting carbon emissions—ceteris paribus—would put the mechanism of the economic ‘growth engine’ into reverse gear. Higher prices would cut demand, not just for fuels, as such, but for downstream products and services at every stage. The result would be an economic recession. It follows that, to address the real problem of climate change, and to reduce the consumption of fossil fuels while encouraging continued economic growth, it is essential to focus simultaneously on three things: (1) increased economic efficiency of useful work, so that additional economic output is generated with the same (or a lower) amount of useful work; (2) increased conversion efficiency, so as to produce more useful work with less carbon-based energy (exergy) inputs and (3) to continue to do so at lower cost. Evidently petroleum and gas (or even coal) are unlikely to become cheaper in the future, especially because demand from rapidly industrializing China and India is now overtaking traditional supply. Nor is nuclear power the panacea, since it has become more, not less, costly in recent decades, due to the need for more elaborate safety precautions than were envisioned in the 1950s. The lack of investment in nuclear power in recent decades has more to do with rising cost than public objections. Moreover, notwithstanding talk about electric cars, electricity is not a direct substitute for liquid fuels, especially for mobile power. Transportation, mechanized farming and most construction are entirely dependent on the availability of liquid fuels. New and emerging technologies under discussion (e.g. [18]) are unlikely to achieve this double objective, either, at least in the near term. Hydrogen fuel cells, in cars, may be more efficient than gasoline engines, but there remain a number of significant barriers to this technology [19]. For the next few decades at least, the main source of hydrogen fuel will be from fossil hydrocarbons, yielding waste carbon dioxide and resulting in only a modest advantage,

at best, in terms of net carbon emissions.10 Carbon sequestration is technically feasible, at least for central generating facilities, but at significant added cost. Combined cycle generating plants are an attractive possibility where natural gas is readily available on long-term contract at low prices. But if the fuel must be obtained by coal gasification, the efficiency advantage is modest, at best, and the cost is significantly higher. Nuclear power remains the first choice of some experts, but even where it has become the dominant source (as in France) it is unloved. The public worries about accidents, environmentalists and others worry about long-term disposal of wastes, while political leaders worry about the diversion of nuclear materials to weapons that might fall into the hands of terrorists or rogue states. Wind power and solar PV are not yet cost competitive (except in special and somewhat rare circumstances). Moreover, both are intermittent, which rules them out for base load applications. Biomass, by contrast, has the potential for use as base-load but, large-scale application may be problematic because of competing demands for land. The natural rate of market penetration for any of these is likely to be slow at best. These technologies are unlikely to become costcompetitive with central electric power generators unless either (1) the scale of production can be increased by orders of magnitude, or (2) the (mostly hidden) subsidies presently received by electric utilities are eliminated. The first option would require extensive subsidies or regulatory interventions to favor the newer technologies that would, themselves, introduce costs into the economy. The second option we address below. There are two present alternatives with the potential for achieving increased economic conversion efficiency at lower cost. One is ‘energy conservation’ at a profit— Amory Lovins has introduced the evocative term ‘negawatts’—as applied primarily to end-uses.11 The problem with this option is that the opportunities for achieving negative cost are limited, although more widespread than skeptics will admit. Clearly these negawatt options should be exploited, but there are many obstacles to overcome. Perhaps the most fundamental economic barrier to conserving energy (exergy) in any specific downstream application (such as home heating) relates to the fact that exergy-related operating costs are typically a very small fraction of total ownership costs. The other alternative, applicable at the production end of the spectrum, is the tried and true technology known as decentralized combined heat and power, or DCHP. Unfortunately, the widespread application of this technology is inhibited by the existing structure of the economy in general and the electric power industry in particular. We focus now on the latter issue. 10 Granted it is possible, in principle, to obtain hydrogen by electrolysis or thermal decomposition of water, but these processes are still far too costly to be serious contenders. 11 See for instance [20–24].

ARTICLE IN PRESS R.U. Ayres et al. / Energy 32 (2007) 634–648

5. The case for and against centralized generation of electric power The majority of the world’s electricity is produced at large centralized plants utilizing steam turbines fuelled with fossil energy. A century ago the justification for such centralized generating plants was straightforward: early 20th century steam turbines were capable of converting heat energy into mechanical energy with efficiencies much higher than the small reciprocating steam engines then installed in most factories (and railroad locomotives) with variable loads. Centralized turbines, operating at steady loads, achieved up to 20% thermal conversion efficiency, compared to 8–10% for smaller reciprocating engines. Moreover, the conversion and distribution losses associated with centralized electricity generation were also much lower than the energy losses associated with delivering mechanical power from on-site reciprocating engines in factories to the machine tools by means of a flexible belt (the usual mode of transmission). Overall, it is estimated that the centralized system achieved efficiencies of the order of 12%—around three times the net efficiency of an on-site reciprocating steam power system—making it thermodynamically and economically superior.12 By today’s standards, however, the centralized power plants of the early 20th century were small and relatively inefficient. Importantly though, these small generating plants also sold heat (as steam) to local customers, allowing them to achieve much higher effective net efficiencies. For example, although Edison’s first plant (at Pearl Street, New York City) only achieved 6% thermal efficiency, Edison was also able to sell much of the by-product heat locally, for an overall thermal efficiency roughly of the order of 50%.13 However, increases in the size of power plants (which contributed to increased thermodynamic efficiency of electricity production) eventually precluded generating plants being sited in built-up city areas. The application of rejected heat for useful purposes declined in parallel [25]. During that past four decades large central generating plants in the US and other major countries have continued to reject large amounts of waste heat, and the average efficiency of delivered electricity in the US has not 12 Quite apart from the efficiency gains, the ability to balance loads provided another compelling reason to centralize in the early days. Streetcar companies needed electricity mainly during morning and evening rush hours; office buildings and factories needed electric power during the working day; apartments and homes needed electric power (for light) mainly in the evening hours; and a few municipal and industrial users could use power in the dark hours of the night. Centralization did not solve the load problem perfectly, but it helped quite a lot. Samuel Insull, at the Commonwealth Edison Company of Chicago, was one of the first to see the advantages of centralizing and load balancing. By systematic interconnection he managed to lower prices dramatically, resulting in rapid increases in demand. 13 It has been estimated that the power sector as a whole achieved maximum overall efficiency (of the order of 65%) in or near 1905, thanks to high levels of waste heat use, when the efficiency of the steam-turbogenerators was only about 15%.


increased significantly since 1960. Now, as then, around 32–34% of the heat content of the fuel burned is converted to electric power, and there is a further loss of about 6–9% of that in transmission and distribution (T&D) [26].14 At the same time, industry, commercial buildings and houses must burn extra fuel inefficiently to produce low grade heat, much of which could, in principle, be recovered from other processes. The failure to do so is not due to technical difficulties, but rather to a lack of incentives and competition. Specifically, in the US the option of cogenerating electricity and selling surplus heat or electricity locally to other users has been inhibited, or outright prohibited by law, for the benefit of the monopoly electric utilities. This prohibition was first established in a 1920s agreement between Commonwealth Edison Company of Chicago and the Illinois Legislature, which granted Edison monopoly rights in exchange for regulation of electricity rates by an independent commission. This became the pattern for electric utilities across the US (and the world). In addition to being technically viable, decentralized combined heat and power generation is now becoming increasingly commercially attractive, due to rising real energy prices. Natural gas and oil prices are now more than three times 1999 levels and coal prices are up significantly. The fuel price increases are largely explained by the rising demand for fossil fuel to support growing economies in the developing world, especially China and India, and the long-term decline in the rate of discovery of new oil and gas resources. In other words, these rising prices may represent the beginning of a long-term trend. On top of increasing fuel prices, for the past 20 years US peak electric loads have grown faster than transmission system additions, leaving little spare capacity in the distribution system. This is evidenced by the Northeast blackout in 2002. A number of similar, though smaller, blackouts have occurred in Europe. The massive financial investments necessary in order to supply new electric load growth with centralized generation will inevitably be reflected in higher retail electricity prices. The ‘bottom line’ is that ceteris paribus, the limited efficiency gains available in centralized generation, the additional electricity distribution costs and rising fuel prices will curtail the long trend of falling real costs of useful work. These developments threaten future income growth and could even cause per capita income reductions. To avoid these developments and return to steady reductions in the cost of useful work, the US must significantly increase the efficiency with which fuel exergy is converted to useful work, thereby dramatically decreasing the quantity of hydrocarbon fuels burned. Fortunately, decentralized electric generation technologies may provide 14 Improvements in generators are possible, primarily by introducing socalled ‘combined cycle’ systems, which means putting a gas turbine at the front end and using the hot exhaust to run a steam turbine. However, combined cycle plants require natural gas or gasified coal to drive the gas turbine. This is not yet economic in many locations.


R.U. Ayres et al. / Energy 32 (2007) 634–648

a proven means by which to achieve this objective, since they facilitate the utilization of presently wasted energy. We now show that the use of currently wasted heat is blocked by the mind-set of the central generation paradigm and its associated regulations and procedures. 6. Obsolete assumptions about central generation A fundamental problem blocking reform is that policymakers, regulators and the public assume that central generation is optimal. This was true early in the 20th century but is true no longer. Several unquestioned and incorrect assumptions underlie this fallacy. They include the following: 6.1. False assumption: economies of scale guarantee that central power generation is more cost effective It is still true that a new large plant can be built for fewer dollars per kW of generating capacity ($500–1500) than a smaller plant using the same fuels and technology. However, this statement applies only to capital cost for the plant itself, not the fuel and not transmission and distribution (T&D). Hence larger scale does not cut the cost of power delivered to the user, under present conditions. There are several reasons. First, existing transmission system capacity in the US (and almost everywhere else) is overloaded. Adding a new central plant is not sufficient: T&D capacity must also be increased to accommodate the added load. But each new kilowatt of central capacity requires an average of $1300 for new T&D wires. This increases the investment to $1800–2800/kW of new capacity. Moreover, average line losses from central power plants are 9% in the US. But the losses under conditions of peak demand are closer to 25%. This means that to supply a kilowatt of new peak load from a central plant to distant consumers one must construct 1.33 kW of new generation and transmission capacity. This drives the investment up to $2400–3500/kW. In addition, central generation requires at least 15% reserve margin on top of expected peak loads (as compared to 4–5% for decentralized CHP). The cost of this reserve capacity ranges from $360 to 525/kW. It follows that to supply 1 kW of new demand with central generation one must actually build 1.56 kW of new central capacity and T&D at a total cost between $2760 and $4025, before paying for any fuel. In short, the real capital investment of new central capacity can be over 5.5 times the supposed minimum of $500/kW, and nearly 3 times the supposed maximum of $1500/kW. Now consider a decentralized plant across the street from a customer. There is no need to add T&D capacity and there are virtually no line losses, because power is consumed either by, or virtually next door to, the producer. The new decentralized plant will require only the last leg of the distribution system. So it may require the addition of only $100–200/kW for wires, far less than the capital

needed for adding the same capacity to a central system. These savings are additional to the fuel saving that arises from the fact that the decentralized plant saves boiler fuel that would otherwise be needed to provide space heat or hot water to the customer. The World Alliance for Decentralized Energy (WADE) has built a model to take account of all of these factors to determine the optimal way to provide for expected electric load growth [25]. The results from a number of model runs indicate that meeting all anticipated US load growth with decentralized generation would save $350 billion over the next two decades, representing a reduction of 40% from the likely cost of meeting the demand from central generation alone [25]. Nonetheless this may represent an ambitious scenario, so to illustrate more realistically the potential impact of decentralized generation in the medium-term we present in Appendix B some simple illustrative market penetration scenarios for the residential and commercial sectors in the US. 6.2. False argument: DCHP plants could fail at the same time the grid is experiencing peak loads, so the grid must have available enough additional redundant capacity to meet peak demand This is the argument regularly made by monopoly utilities to justify high standby rates. It applies only in the hypothetical case that there is one local generation plant on the entire utility grid, and that plant has only one generator. Since there are already 800 large distributed generation plants operating in the US, or an average of 16 per state, this hypothetical case is academic. Most DCHP plants have 2–8 generators with random failure rates of between 2% and 4%. A grid with 100 DCHP plants connected thus has a probable failure rate of 2–4% of the gross system capacity at any one time, including the system peak load time. And, with a significant amount of decentralized capacity connected to a grid, system reliability can be achieved with less than the 15–20% redundancy needed by a grid served by only a few large central plants. A recent study by the Carnegie–Mellon Center for Electric Industry Analysis shows that a system based on many decentralized generation units located near users can achieve desired reliability with only 4–5% reserve margin, rather than the standard 15% margin [27]. Putting it another way, depending on the percentage of power generated locally, the need for dedicated reserve capacity from the grid could range from 0% to 4%. The study cited above concluded ‘‘Even without considering the benefits of robustness y a DG system offers substantial cost savings. Based on current IC engine cogeneration, and with utilization of only half of the cogeneration capabilities of IC engines, savings of up to 20% can be realized in the cost of electricity y These savings increase if more cogeneration is used’’ [27, p.129]. Other studies have led to similar conclusions [28,29].

ARTICLE IN PRESS R.U. Ayres et al. / Energy 32 (2007) 634–648

7. Why is so little ‘waste’ energy used in the US and many other OECD countries? Given the above-market financial returns available in CHP investment (for example, see [30]), not to mention the obvious benefits of reducing air pollution and decreasing grid vulnerability, a question arises: Why is 93% of world wide electric power generated in central plants that reject large quantities of thermal energy? The simple answer is that the costs and losses of moving thermal energy per se long distances (from central plants) are unaffordable. However the technology that permits economical production of electric power locally, on a small scale, is of recent origin. It was not available three quarters of a century ago when the current pattern was established. If it is possible to recycle industrial waste heat to produce 100 gigawatts (GWe) (probably an underestimate), why is only 2.5 GWe actually installed? The reasons differ in detail from case to case, but there is a clear pattern. We offer several explanations, some of which apply in one case, some in another. But all together, they provide clues as to policies that would encourage better utilization of waste energy and lower the cost of useful work, thus powering economic growth in the years to come and increasing per capita income. The generation and distribution of electricity has enjoyed nearly 100 years of monopoly protection, worldwide. In principle, the Federal Public Utility Policy Regulatory Act of 1978 (commonly known as PURPA) supposedly required the state regulators to allow third parties to generate electricity and required the monopoly grid companies to purchase that electricity at ‘avoided cost’. But the law left it to the local state regulatory commissions to determine how to calculate each utility’s avoided cost. Most regulators have essentially left it up to the utilities themselves. In effect, many ancient barriers to decentralized or local generation are still in force, largely in state regulations. Herewith a short list of interconnection barriers with notes about how the particular barrier stifles competition and thus efficiency [31]. (1) All states in the US ban private electric wires that cross any public street or public property. This prevents a local power plant from selling power to neighboring users at retail, so any power in excess of the adjacent user’s needs must be sold back to the utility at wholesale price, even though the current will automatically flow to the nearest users and thus avoid transmission and distribution losses. (2) Since only the monopoly distribution company can connect to a plant, the only way to assure the availability of backup power is to operate in parallel with the grid. Utilities have consistently delayed design approval of interconnection, demanded excessive equipment based on worst case assumptions (that backup power will be required at peak load), and then imposed backup interconnect charges so high that they






often kill the projects. Regulatory commissions are inundated with other rate cases and seldom pay much attention to the question of fixing standby rates. This effectively allows the regulated utilities to block competition. Fifteen (15) states in the US have outright bans on any third party generating power for sale to anyone but the monopoly utility. In these states, a CHP developer is not allowed to invest in a plant to provide retail power to an industrial or commercial user, even on the same site. Not one of the 50 US states has considered the savings to the grid and other users from the addition of new decentralized power. This is so, even though a number of studies have concluded that the benefits from avoided transmission and distribution wires, avoided line losses, and voltage stabilization far outweigh the costs of backup services from the grid to the user. (By contrast, Portugal has built these benefits into the prices paid for electricity produced from generators who utilize rejected heat.) The EPA has historically regulated pollution based on plant design, and current available technology, when built, giving grandfather rights to all existing plants. In effect, existing generators are allowed to pollute at their historic levels or the levels allowed by the permit when the plant was constructed (and permitted, if after 1976). An output-based pollution standard would credit each plant for the boiler fuel displaced by utilizing waste heat. Such a standard would make it easier and more economical to build CHP plants in many, if not most, cases than to build new central plants. The Federal government provides debt guarantees to publicly owned facilities, including the Tennessee Valley Authority, Bonneville Power Administration and a few other federal power agencies. This enables them to obtain capital for roughly 5% versus the 10–12% cost of capital for private sector developers of local plants. Similarly, 2800 municipal governments have established city power companies that borrow taxexempt municipal funds, costing roughly 3% for the construction of central generation plants. In addition, US DOE subsidizes the nuclear industry with roughly $13 billion per year for research and for waste disposal, giving central nuclear plants an advantage over local plants that recycle ‘waste’ energy. These subsidies lower the cost paid by industry for power by transferring costs to taxpayers. The savings achieved by a decentralized CHP plant for displacing centrally generated electricity are consequently often below the cost of producing and delivering that power without capital subsidies.

In spite of the barriers noted above, it may be surprising that there are still many opportunities to earn abovemarket rates of return on decentralized power plants. However, industry leaders are often reluctant to invest the


R.U. Ayres et al. / Energy 32 (2007) 634–648

intellectual and financial resources to make their own power, even when the expected paybacks are far above market returns. The reasons are obscure. Perhaps this is attributable to the fact that big companies rarely have an executive (below the CEO level) with authority and financial resources to initiate such a project, because it is outside the firm’s core business. A number of companies, including those of one of the authors (Casten) have offered to build and operate recycled energy plants with their own capital. These third parties appear to own most of the 68 GWe of existing combined heat and power plants. But these firms have had difficulty gaining size and credibility to compete effectively in the face of regulatory restrictions [31]. 8. Conclusions This review makes two major points, which are worth repeating and emphasizing. Whereas conventional theories of economic growth do not depend on energy supply, or the demand for energy services, a new, quantitative, endogenous theory of economic growth does incorporate these factors explicitly [11,13,14]. In particular, it has been shown that US and Japanese economic growth, since 1900, can be explained endogenously by a production function incorporating a third factor (along with traditional capital and labor). This third factor is a measure of the production and delivery of useful work to consumers, where useful work is the product of exergy inputs multiplied by the efficiency of conversion of exergy to work. This conversion efficiency tends to increase over time, whence it is also a measure of technology. While this result is interesting in itself, the key point here is that the endogenous theory has certain policy implications very different from the implications of the standard theory in which energy and energy services play no explicit role. The important implication of this new endogenous theory, for our present purposes, is that economic growth in the future depends fundamentally on continuation of the historical decline in costs of energy services. But this longterm trend, which has been driven both by discoveries (e.g. of oil and gas) and by technological progress (e.g. in electric power generation and in the production of energy services), cannot be expected to continue indefinitely at historical rates. In fact, the chances are very good that energy costs and the costs of energy carriers (e.g. electric power) have bottomed. If energy service costs begin to increase significantly as a fraction of GDP, economic growth is likely to decrease or even turn negative. It follows that continued economic growth depends on continued declines in the cost of energy services (useful work). Hence a policy of increasing the cost of energy (e.g. by imposing carbon taxes) in order to reduce consumption—for legitimate environmental reasons—could be economically counter-productive. On the other hand, this logic suggests that a policy of encouraging energy conservation at negative cost, (i.e. exploiting opportunities

for ‘double dividends’) would be productive. Luckily, there is at least one such opportunity worthy of serious consideration. This leads us to the second major point of this paper: that utilization of waste heat in decentralized co-generation (combined heat and power or CHP) systems, and industrial waste energy streams, can have a very significant economic benefit, not only in terms of energy conservation but simultaneously, in terms of GHG (and other) emissions reduction and also in terms of increased grid stability. We point out that central steam-electric plants are really much less cost effective than they seem to be, and that they are also being subsidized (through low-cost borrowing) whereas the decentralized alternatives are not subsidized and are more cost effective than they appear at first. But the crucial point is that there is already a substantial potential for the profitable introduction of decentralized combined heat and power (DCHP) plants, and that the result would be less fossil fuel use and less pollutant emissions, and greater stability of the grid, without any need to develop new technology. The key to opening this reservoir of opportunities is real deregulation of the electric power industry, especially by eliminating as many as possible of the barriers to competition listed in the text [32,33]. It should be emphasized that there is no need to eliminate conventional centralized capacity. The DCHP suppliers will have plenty to do in just meeting the demand increases already forecast for the next 20 years. Nor is there any reason to weep for the established utilities. In a deregulated environment there would be nothing but their own inertia to prevent them from getting into the DCHP business themselves. The end result will be a sharp increase in the overall efficiency of the US power system, and that of many other countries, with lower costs to consumers and less pollution of the environment. Appendix A. Using waste energy: combined heat and power (CHP) Extraction of useful work from waste energy streams is accomplished in two ways. The first is to conduct otherwise waste heat from conventional steam electric generating plants to large nearby heat users, to displace boiler fuel. The second is to convert industrial waste energy streams into useful heat and electric power that can be utilized onsite or nearby. The first strategy was widely adopted, mainly in central and eastern Europe, as ‘district heating’. In many cities steam from centralized coal-burning power plants is piped to nearby apartment blocks for domestic heating purposes. Unfortunately, even though the pipes are insulated, there are major losses along the way and it is difficult to regulate the temperature at the end of the pipe, resulting in further inefficiencies. Also, the diversion of steam at a temperature above 100 1C also reduces the efficiency of the electric power generating unit. District heating based on steam

ARTICLE IN PRESS R.U. Ayres et al. / Energy 32 (2007) 634–648


poly created to sell electricity, and not heat, under the (false) assumption that centralized power generation is more efficient than decentralized power generation. This was clearly true early in the 20th century, but is no longer necessarily true. Relatively small gas turbines and diesel engines are nearly as efficient as large steamgenerating systems, especially when transmission and distribution losses are taken into account. For instance, a 5 MW Solar gas turbine (32% efficient) with a highpressure heat recovery steam generator equipped with a back-pressure steam turbine can deliver electricity with 45% efficiency, plus useful heat. The Solar Mercury gas turbine is already 42% efficient and has a very clean exhaust [36]. The next generation of high-temperature fuel cells may be even more efficient. We comment further on these points below.

from central plants has not been widely adopted in the West for these reasons. However, the logic of using otherwise wasted lowtemperature heat from power generation to replace boiler fuel is inescapable. Fig. A1 illustrates the system as it exists today. Fig. A2 illustrates the benefits schematically: to produce 35 units of electric power and 50 units of heat via CHP (co-generation) requires of the order of 100 units of fuel exergy, as compared to 189 units if the electricity and heat are provided separately. The overall losses in this example are reduced from 104 units to only 15 units, an enormous gain. The theoretical gains illustrated in Fig. A2 are not quite as easily achieved in practice as one might hope. There are multiple reasons, but the basic one is that the electric utility industry in most countries is a regulated ‘natural’ mono-

Wasted energy

Coal 17075 Conversion losses from thermal production 24726 Oil 3214

Total primary energy input for electricity production 40180

Gas 8384 Renewable biomass geothermal 1024

t 963 plan wer o p use 339 Own es 1 loss n io t u istrib nd d ion a s is ansm

Gross electricity production 15454

Net electricity production 14491

Tr Electricity delivered to customers 13153

Industry 5683 Non industry 7470

Nuclear 7777 Hydro 2705

Less then 33% total efficiency of the world's electricity supply system [34]

Fig. A1. Electricity generation worldwide (in TWh) (source: IEA, 2002 [34]).

Conventional Generation Losses (95)

Combined Heat and Power

power station fuel (130)


(1 MW natural gas reciprocating engine) electricity



CHP fuel CHP


boiler fuel (59)

100 New boiler




Losses (9) Source: Ramani [35]

Losses (15)

Fig. A2. Typical fuel inputs using conventional separate heat and power (source: Ramani, 1997 [35]).


R.U. Ayres et al. / Energy 32 (2007) 634–648

Most of the existing DCHP capacity has used coal and steam turbines or gas fired combined cycle plants. Some of the capacity burns pulp and paper waste biomass. The average capacity of distributed generation plant in the US is 84 MW. The only change to the technology from central generation is the extraction of steam at low pressures to displace boiler fuel. By giving up electric production equal to perhaps 5% of the energy in the fuel, the extracted steam displaces boiler fuel equal to 40–50% of the initial fuel. In other words, the overall electricity plus heat demand can be met with just over half of the fuel that would be used by the same technology in a central plant. As regards the second strategy, it starts from the fact that many industrial processes produce under-utilized byproduct energy streams. These include (1) hot exhaust gases, (2) combustible flare gases and (3) high pressure gases. Hot exhaust gases are produced by coke ovens, pelletizing ovens, glass furnaces, petroleum refineries, ammonia plants or hot rolled steel ovens. Flare gases are mainly produced by blast furnaces or petroleum refineries. High-pressure gases are mainly from steam or natural gas pipelines, which must be reduced to ambient or near ambient pressure at the point of use. All of these can be used to generate electricity using commercially available equipment. Nevertheless, virtually all industrial facilities also utilize electric power, entirely purchased from the grid. The problem is to utilize the in-house waste energy streams to save on purchased electricity. Unfortunately, the electric utilities typically set discouragingly high prices on grid inter-connections and they pay low prices for purchased surplus power, mainly to discourage such competition. It is not necessary here to detail specific examples of costeffective cogeneration options, but it seems clear that there exists an enormous potential to realize energy cost savings from exploiting the heat energy that is wasted unnecessarily by the arbitrary separation of electricity generation and process heat—for example, see [25,30]. Moreover, these energy savings also lead to savings from avoided transmission line investment, avoided transmission losses due to the lightened remaining load on local and regional wires and reduced air pollution. Official reports of the Energy Information Agency (EIA) indicate that only 9% of the total electric power generated in the US in 2000 was produced by facilities that utilized waste heat. Those facilities accounted for only 7.5% of total capacity. However, the potential is much greater. A recent study has identified 44 gigawatts (GWe) of additional capacity that could be powered with waste heat from just three source categories: flare gas from petrochemical processes, exhaust heat from the gas turbines that drive transcontinental gas pipeline compressors, and the recovery of steam pressure drop in industrial steam systems [37]. That study did not explore other industrial heat sources. Another recent study done for EPA has estimated that 96 GWe of electric power could be provided in the US by recycling industrial waste heat in 19 industries. This would

amount to 11.5% of current generating capacity in the US. Yet, according to the 2003 Energy Information Agency data, only 2.5 GWe of co-generation capacity was actually installed. Exploiting the full potential could save the US economy $4–10 billion per year after full capital cost recovery and reduce fuel consumption by 1.2–3 EJ per year, not to mention the sulfur oxides, nitrogen oxides and carbon dioxide associated with the combustion of this fuel. As of 2002 the EIA reported 806 combined heat and power plants in the US with 68 GWe of ‘nameplate capacity’, this was 7.13% of total installed capacity in the US. These CHP plants averaged 84 MW each. (The data do not include plants with less than 1 MW of capacity, so many micro-turbines and small diesel plants are not included.15) By contrast, in the US there were 3855 utility-owned or municipal electricity-only power plants with nameplate capacity of 863 GWe, averaging 224 MW per plant. The electrical supply industry has forecast that the US will need 137 GWe of new capacity by 2010, costing $84 billion, plus $220 billion for additional transmission and distribution. One of us (Casten) has estimated that meeting this demand with decentralized CHP would cost only $168 billion, with no additional needs for transmission and distribution [25]. There is a great disparity between states with power produced by CHP, ranging from 0% in three states to 22% in California and 25% in Hawaii. Of course the discrepancies are partly attributable to the mix of power production facilities, since hydroelectric plants are incompatible with CHP, and nuclear plants—as in France—are usually sited too far from cities to be able to provide district heating economically. International data show that combined heat and power (CHP) facilities generate roughly 7.2% of the world’s electric power, similar to the US percentage. But CHP accounts for over 50% of the electric power generated in Denmark,16 39% in the Netherlands, 37% in Finland and 31% in Russia; Germany gets 19% and Poland, Japan, and China are at 18%. (Admittedly the high co-generation (CHP) percentages in Russia, Poland, former E. Germany and China reflect long-standing policies of encouraging district heating combined with coal-burning central power plants.) These data indicate that CHP can be installed, provided that the policies are friendly, the plants are significant in size, and it is possible to utilize effectively heat energy from a majority of electric generation facilities.

15 As a matter of interest, in the Netherlands (1998), DCHP plants of less than 1 MW capacity accounted for 6% of the national total, while cogeneration plants with capacity between 1 and 50 MW accounted for 35% [38]. 16 There is some doubt about this statistic which may also apply to the Dutch and Finnish cases. We have been told that the statistic actually refers to power generating facilities with the capability for selling heat as well as electric power, but that many power-generating plants do not actually do so.

ARTICLE IN PRESS R.U. Ayres et al. / Energy 32 (2007) 634–648

Appendix B. CHP scenarios for the US to 2025 This appendix describes the construction and results of a simple scenario analysis to explore the potential impact on energy consumption of large-scale deployment of CHP systems in the residential and commercial sectors in the US. Importantly, this analysis is meant to be illustrative only, and readers are referred to other studies (e.g. [39]) for more technical specifics of possible CHP applications. We start by outlining the key assumptions and data sources used. Currently, space and water heating in the commercial and residential sectors in the US consume a significant amount of energy, accounting for 13–14% of total delivered energy use [40]. Around one-quarter is for water heating, and the remainder space heating. Energy consumption for space and water heating is projected to grow steadily throughout the next 20 years to around 12 EJ by 2025 [40]. In this analysis, our interest is in the potential impact of replacing conventional space and water heating equipment with CHP systems in the residential and commercial sectors. We assume that conventional space and waterheating technologies and fuels exhibit the useful heat efficiencies presented in Table B1. These efficiencies were chosen to reflect an assumed combination of fuel and technology characteristics [40]. Currently, around twothirds of US heat demand is supplied with natural gas, about 12% from diesel and 11% from electricity, with LPG and other fuels accounting for the remaining share [40]. The alternative CHP systems are assumed to achieve an average electricity generation efficiency of 35%; in the range of a number of existing CHP technologies (see [42]). We also assume that CHP systems have a useful heat generation efficiency of 40%. Accordingly, the total useful energy efficiency was assumed to be 75%, consistent with estimates from the Department of Energy [42]. To explore the potential deployment and penetration rate of CHP in commercial and residential sector space and water heating, we constructed two simple scenarios for the medium term (to 2025), based on the energy demand projections from the US Department of Energy [40]. In these scenarios, we assume that deployment of CHP is

Table B1 Assumed heating efficiencies of conventional space and water heating technologies in the commercial and residential sectors Fuel and technology

Efficiency (%)

Electricity (electric resistance) Electricity (electric resistance), including losses in generation and transmission Natural gas (typical central heating) Diesel (typical central heating) LPG (estimate for central heating) Other (wood, kerosene, coal) (combustion)

97 30–32

Source: Based on EIA [40]; USDOE [41].

85 80 70 40


restricted to new and refurbished/renovated buildings, or major equipment replacement. In the rest of the market, we make the conservative assumption that CHP is either technically or commercially uncompetitive. Projections of the number of new residential dwellings were taken from EIA [40], and range between 1.8 and 2 million annually over the next 20 years. Annual changes in projected total commercial floor space (from [40]) were assumed to represent new additions to the commercial sector building stock, which grows approximately 40% between 2005 and 2025 under the EIA projection. In addition to new buildings, refurbishment and equipment replacement in old buildings represents a potential market for CHP. In the commercial building sector, we assume that an average of 5% of the total floor area is refurbished every year, so that over a 20-year period all buildings are refurbished, and major plant replaced. In the residential sector, we assume that in each year water- and space-heating equipment is replaced in 3.5% of the total housing stock. Importantly, however, the rate at which CHP technologies can be deployed in these markets is currently very limited by scale and expertise in small-scale CHP system installation, and market acceptance. Accordingly, in the first of the scenarios presented here (scenario (1)) we assume initially a low (5%) penetration of CHP technologies in new or refurbished buildings, but that this increases according to a simple logistic function, presented in Fig. B1. Fig. B1 also shows how this penetration rate labeled (1) combined with the assumptions described above translates into potential CHP deployment in the commercial and residential sectors. We also present in Fig. B1 a more pessimistic market penetration scenario labeled (2). Preliminary results By combining the market penetration rates presented in Fig. B1 with projections of space and water heating energy demand (from [40]), we can estimate the amount of fuel used in conventional heating technologies that could be saved by deployment of CHP under the two market penetration scenarios. In constructing these estimates we assume that conventional space and water-heating technologies are displaced by CHP in accordance to their market shares. That is, we assume on average that in new and refurbished buildings the mixture of technologies does not differ from the overall market average technology choice (from [40]). In addition to displacing energy used in heating, CHP technologies produce electricity, which avoids the need to generate electricity from conventional sources. Accordingly, we also need to account for this energy saving to calculate the full impact of CHP technologies. To illustrate the potential overall energy savings, we present in Table B2 a detailed accounting of the impact of CHP on total energy use in 2025 under scenario (1), the more optimistic market penetration scenario (see Fig. B2). Table B2 first shows the

ARTICLE IN PRESS R.U. Ayres et al. / Energy 32 (2007) 634–648



Penetration rate (1)

0.9 Penetration rate (2)

Share of market

0.8 0.7

Potential CHP share of commercial market


Potential CHP share of residential market

Commercial (1)

Residential (1)


Commercial (2)

0.4 Residential (2) 0.3 0.2 0.1 0 2006










Fig. B1. Assumed CHP technology penetration and impact on potential deployment scenarios (1) and (2).

Table B2 Impact of CHP on energy consumption in 2025, optimistic scenario Electricity (EJ) Fuel displaced from conventional heat production Electricity Natural gas Diesel LPG Other (wood, kerosene, coal) Total Fuel used in CHP production CHP electricity generation Generation plus savings (1) Avoided conventional electricity generation (2) Net energy savings

Other (EJ)

0.72 5.03 0.85 0.20 0.27 6.36


14.77 5.17 5.89 5.89

17.67 9.26

Note: Positive numbers indicate energy savings or additional production; negative numbers indicate additional energy consumption or reduced energy production. (1) Including electricity displaced from conventional heat production. (2) Based on average generation efficiency [40].



Net energy saving - scenario (1)



Net energy saving - scenario (2)



GHG abatement - scenario (1)



GHG abatement - scenario (2)

120 Scenario (1)




80 60

3 2

Scenario (2)

GHG abatement (Mt C)

Energy saving (EJ)


40 20



20 0 20 2 0 20 3 0 20 4 05 20 0 20 6 07 20 0 20 8 0 20 9 1 20 0 1 20 1 1 20 2 1 20 3 14 20 1 20 5 16 20 1 20 7 1 20 8 19 20 2 20 0 2 20 1 22 20 2 20 3 2 20 4 25


Fig. B2. Scenario of distributed generation in the residential and commercial sectors, US 2002–2025 (based on building turnover).

ARTICLE IN PRESS R.U. Ayres et al. / Energy 32 (2007) 634–648

amount of each conventional heating fuel displaced, based on projections from EIA [40], and the amount of fuel required by CHP to produce the same quantity of useful heat (calculated according to the efficiencies of the conventional technologies from Table B1, and our assumed CHP heat efficiency of 40%). Table B2 then reports the electricity cogeneration from CHP, and the total amount of avoided conventional electricity generation, accounting also for displaced electricity used in conventional space and water heating. From this amount we estimated the avoided fuel consumption in conventional electricity generation, based on the projected average fuel mix from EIA [40]. Finally, we report in Table B2 the total net energy savings from deployment of CHP under this scenario in 2025. The net energy savings from CHP deployment across the whole time period to 2025 under our two scenarios are presented graphically in Fig. B2. Fig B2 also shows net GHG abatement under these scenarios, assuming that CHP systems are fueled with natural gas (which is also the main conventional heating fuel). Under the more optimistic scenario net energy savings amount to over 9 EJ, or around 7% of total US energy consumption [40] by 2025, and carbon dioxide emission savings exceed 180 Mt C, or around 8% of projected total US energy emissions [40,43].17 Even under the pessimistic market penetration scenario (2), annual GHG abatement reaches almost 120 Mt C by 2025, with significant additional potential considering that market penetration rates are around 35–45% in 2025 under this scenario. Accordingly, under the assumptions described above, deployment of CHP in the residential and commercial sectors can result in substantial reductions in energy use and GHG emissions over the medium term. References


[12] [13]

[14] [15] [16] [17] [18] [19] [20] [21] [22] [23]

[24] [25] [26] [27]

[28] [29]

[1] Simon JE, Beisner C, Phelps J. The state of humanity. Cambridge, MA: Blackwell Publishers Ltd.; 1995. [2] Mankiw NG. Macroeconomics. Worth Publishing; 1997. [3] Denison EF. Explanation of declining productivity growth. Survey Curr Bus 1979;59(8-Part II):1–24. [4] Jorgenson DW. Productivity and economic growth in Japan and the United States. Am Econ Rev 1988;78(2):217–22. [5] Solow RM. A contribution to the theory of economic growth. Q J Econ 1956;70:65–94. [6] Solow RM. Technical change and the aggregate production function. Rev Econ Stat 1957;39:312–20. [7] Barro RJ, Sala-i-Martin X. Economic growth. 2nd ed. Cambridge, MA: MIT Press; 2003. [8] Sala-i-Martin X. I just ran two million regressions. Am Econ Rev 1997;87(2):178–83. [9] Ayres RU. Energy inefficiency in the US economy: a new case for conservatism. Laxenburg, Austria: International Institute for Applied Systems Analysis; 1989. [10] Ayres RU, Ayres LW, Warr B. Is the US economy dematerializing? Main indicators and drivers. In: Janssen A, van den Bergh JCJM, 17

Emissions savings would drop to 115 Mt C (or 5% of total emissions), if we assume diesel fuel is used to power CHP systems.



[32] [33]

[34] [35]

[36] [37]


editors. Economics of industrial ecology: materials, structural change and spatial scales. Cambridge, MA: MIT Press; 2004. Ayres RU, Ayres LW, Ra˚de I. The life cycle of copper; its coproducts and byproducts. In: Tukker A, editor. Eco-efficiency in industry and science. Dordrecht/Boston/London: Kluwer Academic Publishers; 2003. Ayres RU, Ayres LW, Warr B. Exergy, power and work in the US economy, 1900–1998. Energy 2003;28(3):219–73. Ayres RU, Warr B. Economic growth models and the role of physical resources. In: Bartelmus P, editor. Unveiling wealth: on money, quality of life, and sustainability. Dordrecht/Boston/London: Kluwer Academic Publishers; 2002. p. 171–88. Ayres RU, Warr B. Useful work and information as drivers of growth. Fontainebleau, France: INSEAD; 2003. Ayres RU, Warr B. Accounting for growth: the role of physical work. Struct Change Econ Dynam 2005;16(2):181–209. Ayres RU, Ayres LW, Pokrovsky V. On the efficiency of electricity usage since 1900. Energy 2005;30(7):1092–145. von Weizsaecker EU, Jesinghaus J. Ecological tax reform: a policy proposal for sustained development. London: Zed Books; 1992. IEA. Renewables for power generation: status & prospects. Paris: International Energy Agency/OECD; 2003. Keith WK, Farrell AE. Rethinking hydrogen cars. Science 2003;301:315–6. Lovins AB. Negawatts: twelve transitions, eight improvements and one distraction. Energy Policy 1996;24(4):331–43. Lovins AB. Energy efficiencies resulting from the adoption of more efficient appliances: another view. Energy J 1998;9(2):155–62. Lovins AB, Lovins HL. Energy: the avoidable oil crisis. Atlantic Month 1987;260(6):22–9. Lovins AB, Lovin HL, Krause F, Bach W. Least-cost energy: solving the CO2 problem. Andover, MA: Brickhouse Publication Co; 1981. Fickett AP, Gellins CW, Lovins AB. Efficient use of electricity. Sci Am 1990;263(3) 64–68, 71. Casten TR, Downes B. Economic growth and the central generation paradigm. Dialogue (US Assoc Energy Econ) 2004;12(2):7–13. EIA, USDOE. Annual energy review 2004. Washington, DC: United States Government Printing Office; 2005. Zerriffi H. Electric power systems under stress: an evaluation of centralized vs. distributed systems architectures. PhD, Engineering and Public Policy, Carnegie-Mellon University, Pittsburgh, PA, 2004. AD Little Inc. Reliability and distributed generation. Cambridge, MA; 2000. Lovins AB, Datta EK. Small is profitable: the hidden economic benefits of making electrical resources the right size. Snowmass, CO: Rocky Mountain Institute; 2002. EERE. Distributed energy case studies database. Energy efficiency and renewable energy. US Department of Energy: /http://eeredev.; 2006. Alderfer RB, Eldridge MM. Making connections: case studies of interconnection barriers and their impact on distributed power projects. Golden CO: Competitive Utility Strategies, LLC and Kelso Starrs and Associates LLC for the National Renewable Energy Laboratory; 2000. Morgan MG, Zerriffi H. The regulatory environment for small, independent micro-grid companies. Electricity J 2002;15:52–7. King D, Morgan MG. Guidance for drafting state legislation to facilitate the growth of independent electric micro-grids. Pittsburgh, PA: Carnegie-Mellon Electricity Industry Center; 2003. IEA. World energy outlook. Paris: OECD/IEA; 2002. Ramani RV. Cogeneration: are we missing out on an opportunity? In: Dadhich PK, editor. Cogeneration: policies, potential and technologies. New Delhi: Tata Energy Research Institute (TERI); 1997. Casten TR. Personal communication, 2005. Casten TR, Collins MJ. Optimizing future heat and power generation. Co-generat On-site Power Product 2002;6(3):71–7.


R.U. Ayres et al. / Energy 32 (2007) 634–648

[38] Strachan ND, Zerriffi H. System implications of distributed generation. In: Thissen WAH, Herder P, editors. Critical infrastructures: state of the art in research and application. Boston: Kluwer Academic Publishers; 2003. [39] ONSITE SYCOM Energy Corporation. The market and technical potential for combined heat and power in the commercial/institutional sector. Washington, DC; 2000. [40] EIA, USDOE. Annual energy outlook. Washington, DC: United States Government Printing Office; 2005.

[41] USDOE. Comparing heating fuels: information resources, office of energy efficiency and renewable energy. US Department of Energy; 2005. [42] USDOE. Review of combined heat and power technologies. ONSITE SYCOM Energy Corporation for the California Energy Commission with the US Department of Energy, Office of Energy Efficiency and Renewable Energy, 1999. [43] IPCC. Revised 1996 IPCC guidelines for national greenhouse gas inventories: workbook 2: Intergovernmental Panel on Climate Change. 1997.