Economics of Energy Efficiency ADAM B. JAFFE Brandeis University and National Bureau of Economic Research Waltham, Massachusetts, United States RICHARD G. NEWELL Resources for the Future Washington, D.C., United States ROBERT N. STAVINS Harvard University Cambridge, Massachusetts, United States This article reviews economic concepts relevant to 1. Overview decision making about energy efficiency investments 2. The Process of Technological Change and related policy choices. We describe economic 3. Understanding the Energy Efficiency Gap perspectives on the process of energy-saving techno- 4. Market Failures logical change, the distinction between market fail- 5. Non-Market-Failure Explanations of the Energy ures and market barriers in energy-using product Efficiency Gap markets, and the important role that discounting 6. Discounting and Implicit Discount Rates plays in this area. 7. Summary 1. OVERVIEW Glossary Energy efficiency is defined here to mean energy externality An economically significant effect of an activ- services provided per unit of energy input (for ity, the consequences of which are borne (at least in example, gallons of water heated to a specified part) by parties other than the party who engages in temperature per British thermal units of natural gas the activity, and which are not accounted for through input). Within this framework, energy efficiency is trade. diffusion The gradual adoption of new process or product conceived primarily at the disaggregated, product innovations by firms and individuals. level, rather than at a more aggregated sectoral level. innovation The initial market introduction or commercia- As with virtually all economic problems, the lization of new process or product inventions. economics of energy efficiency is at its heart a invention The development and creation of a prototype question of balancing of costs and benefits. For the new idea, process, or piece of equipment. individual energy user, this involves weighing the market barriers Disincentives to the diffusion and/or use of higher initial cost of purchasing energy-efficient a good, such as high costs or prices, which may or may products against the expected benefits of future cost not represent market failures. savings when operating the products, among other market failures The failure of private markets to provide considerations. Due to this difference in the timing of certain goods at all or at the most desirable level, energy efficiency costs and benefits, issues related to typically arising from a situation in which multiple parties benefit from a good without decreasing one discounting feature prominently in analyses of another’s benefits, and in which those who have paid for energy-efficient technology adoption. For suppliers the good cannot prevent others from benefiting from it. of energy-using products, decisions regarding energy- technological change The process of invention, innova- efficient innovations likewise depend on the expected tion, and diffusion whereby greater and/or higher profits from such technology development. Profits quality outputs can be produced using fewer inputs. from innovation depend in turn on the expected Encyclopedia of Energy, Volume 2. r 2004 Elsevier Inc. All rights reserved. 79 80 Economics of Energy Efficiency demand for energy-efficient technologies and the significance to demand for particular types of energy. degree to which firms can appropriate the value The reason it is important to distinguish carefully created by their innovations. among these different conceptual steps—invention, On both the demand and supply side of this innovation, diffusion, and use—is that public policies market for energy-efficient technology, potential can be designed to affect various stages and will have market imperfections can lead to underinvestment specific and differential effects. in energy efficiency. This raises the possibility that corrective government policies could provide eco- nomic gains if they are practicable and provide net 2.1 Technology Adoption and Diffusion benefits after inclusion of all public and private implementation costs. The degree to which such Beginning at the end of the technological change opportunities exist in practice is the subject of process, research has consistently shown that diffu- significant debate. Finally, environmental pollution sion of new, economically superior technologies is associated with energy production—particularly car- never instantaneous. The S-shaped diffusion path bon dioxide emissions from fossil fuels—represents shown in Fig. 1 has typically been used to describe an additional reason why private markets might the progress of new technologies making their way underprovide energy efficiency if energy users do not into the marketplace. The figure portrays how a new face the cost of any resultant environmental harm. technology is adopted at first gradually and then with increasing rapidity, until at some point its saturation in the economy is reached. The explanation for this 2. THE PROCESS OF typical path of diffusion that has most relevance for TECHNOLOGICAL CHANGE energy conservation investments is related to differ- ences in the characteristics of adopters and potential Many readers may be unfamiliar with the way adopters. This includes differences in the type and economists typically view the process of technologi- vintage of their existing equipment, other elements of cal change, thus it is useful to first establish this the cost structure (such as access to and cost of labor, common understanding. Furthermore, to understand material, and energy), and their access to technical the potential for public policy to affect energy information. Such heterogeneity leads to differences efficiency, it is also necessary to understand the in the expected returns to adoption and, as a result, process through which technology evolves: inven- only potential adopters for whom it is especially tion, innovation, diffusion, and product use. Inven- profitable will adopt at first. Over time, however, tion involves the development of a new idea, process, more and more will find it profitable as the cost of or piece of equipment. This activity typically takes the technology decreases, quality improves, informa- place inside the laboratory. The second step is the tion about the technology becomes more widely process of technology innovation, in which new available, and existing equipment stocks depreciate. processes or products are brought to market. The longevity of much energy-using equipment Another way of describing this stage is commercia- reinforces the importance of taking a longer term lization. The third step is diffusion, the gradual view toward energy efficiency improvements—on the adoption of new processes or products by firms and order of decades (see Table I). individuals, which then also decide how intensively to use new products or processes. Tying this all together, it would be possible, for Adoption share example, to think of a fundamentally new kind of automobile engine being invented. This might be an alternative to the internal combustion engine, such as a system dependent on energy-efficient fuel cells. The innovation step would be the work carried out by automobile manufacturers or others to commercia- lize this new engine––that is, bring it to market, then offer it for sale. The diffusion process, then, would reflect the purchase by firms and individuals of automobiles with this new engine. Finally, the degree Time of use of these new automobiles will be of great FIGURE 1 The gradual S-shaped path of technology diffusion. Economics of Energy Efficiency 81 Several studies have explored the effect of energy focus equally on both, but uncertainty about future prices and technology adoption costs on energy energy prices or whether they will face these costs efficiency investments. From a policy perspective, (because they could move, for example) makes them the effect of higher energy prices can be interpreted give less weight to energy prices than they do to as suggesting what the likely effects of taxes on capital cost, which is known. A final interpretation energy use (or carbon dioxide) would be and the might be that consumers have reasonably accurate effects of changes in adoption costs can be inter- expectations about future energy prices, and their preted as indicating what the effects of technology decisions reflect those expectations, but the proxies adoption subsidies would be. As suggested by the for these expectations that are used by researchers economic cost–benefit paradigm, these studies have are flawed, causing their measured effect to be found that higher energy prices increase and adop- smaller than their true effect. For example, studies tion costs decrease the extent of adoption of energy- often use current realized energy prices as a proxy for efficient technology (for example, building insula- expected future energy prices. Current prices fluc- tion, more efficient home appliances, or more tuate more than expected future prices, however, efficient industrial motors). An additional interesting leading to a downward bias in the coefficient on the finding in this line of research is that the adoption of energy price proxy relative to the true relationship these technologies is more sensitive to the cost of the with expected prices. equipment than it is to the expected cost of energy. Although empirical evidence indicates that subsidies This implies that a policy of subsidizing the purchase may be more effective than comparable taxes in of new efficient equipment may be