Regulatory and mixed policy options for reducing fuel use and carbon emissions

Joel N. Swisher

UNEP Collaborating Centre on Energy and Environment

Risø National Laboratory, Denmark

1 Introduction

As the Framework Convention on Climate Change (FCCC) goes into effect, some 20 industrialised countries have made commitments to stabilise or reduce future carbon emissions. To accomplish the stated goal of the FCCC, which is to stabilise greenhouse gas (GHG) concentrations in the atmosphere, global carbon emissions would have to be reduced by 60 percent or more from present levels (IPCC 1990). This would require much more drastic reductions by the industrialised countries, and emissions from developing countries would eventually have to be limited as well.

To achieve the existing reduction targets, not to mention those necessary to stabilise the atmosphere, technological changes will be necessary to reduce the fossil fuel-intensity of most countries' energy system (supply-side measures) and to improve the efficiency with which fuels and electricity are used (demand-side measures). The possible policy instruments with which to stimulate these changes are many. At the international level, most discussion has centred on various forms of carbon emission taxes and to some extent on tradable emission offsets or permits. At the national level, where most real energy policy changes would have to be implemented, other non-fiscal policy instruments are common.

These other policy mechanisms, which use regulation or a mix of regulation and financial incentives, are the topic of this paper. Some of these mechanisms are already in widespread use, while some are more innovative and have only recently been applied to energy technologies. They include energy performance standards, technology procurement programs, utility demand-side management (DSM), as well as familiar development and demonstration activities. To a large degree, these types of programs are designed to capture the potential demand-side energy-efficiency improvements. Supply-side measures seem to be addressed relatively effectively through technology development programs and fiscal measures such as emission taxes and subsidies for less polluting energy sources. Most energy models agree on results of such measures but offer little help with estimating parameter values regarding the inherent size of the resources or potential rates of technical progress.

One of the principal reasons for pursuing energy-efficiency improvements is that energy consumption leads to pervasive externalities, ranging from local pollution and global greenhouse gases to energy and nuclear security risks, that are not reflected in energy supply costs and planning efforts. By mitigating these problems with technical improvements that are cost-effective relative to new energy supplies, innovative energy-efficiency programs appear to offer a "win-win" solution. Moreover, it appears that such cost-effective technical opportunities are also abundant in developing countries (US OTA, 1992).

The investment behaviour of most energy users does not account for the long-term costs of energy supplies, even without the externalities. The implicit discount rates applied by energy-users to energy-efficiency investments range from 20% to 200%, compared to utility discount rates of 6-10% (Ruderman et al., 1987). Thus, one cannot expect that carbon taxes and energy price increases, although a necessary measure in many cases where energy is subsidised, will lead to sufficient investment in energy-efficient technologies. The need for other measures to help implement energy efficiency is the justification for direct government policy measures and utility programs such as DSM.

2 Background: are there alternatives to high carbon taxes?

There continues to be a great deal of disagreement between studies that focus on the costs of reducing carbon dioxide emissions from fossil fuel production and use (Wilson and Swisher, 1993). Technical-economic (bottom-up) models identify substantial cost-effective emission reduction potential in most countries, under the assumption that existing barriers to energy efficiency can be reduced. The total emission reduction potential in most industrialised countries over the next decades is estimated at 10 to 30 percent, at no or low cost to society, and larger if increasing costs are accepted (Johansson and Swisher, 1993). Similar potential has been identified in several developing countries (UNEP, 1994).

Studies based on macroeconomic (top-down) models, on the other hand, generally conclude that significant macroeconomic losses would result from the imposition of carbon emission limits (Gaskins and Weyant, 1993). The policy measures that the macroeconomic models evaluate with respect to their impact on the energy system and carbon emissions are energy-price changes through, for example, carbon taxes. As modelled in top-down analyses, such measures result in a transfer of inputs to other sectors, revenue increases to governments, and an economic efficiency loss to society. Other policy interventions (e.g., regulations and other measures aimed at overcoming barriers to energy-efficiency improvements) are assumed to be expensive and sub optimal, because they are not part of the assumed economically-efficient baseline (Manne, 1992).

While the theory of economic externalities dates back to Pigou, the economic theory of environmental protection has developed most over the past thirty years (Pigou, 1920). Coase demonstrated that, in the absence of transaction costs, bargaining between the source and the victim of an externality will produce the economically efficient outcome if either party holds the relevant property right (Coase, 1960). This result, that self-interest expressed through the market will efficiently capture an externality, can be applied to environmental protection, for example via tradable emission permits. Similarly, for a given level of emission reduction, an emission charge rate that achieves that level can satisfy the theoretical conditions for minimum-cost emission reduction (Baumol and Oates, 1975). The difficulty is finding the optimum level of emission reductions and the right level of emission charges to achieve the desired reduction. Ideally, this level would be equivalent to the marginal value of the environmental damage caused or avoided, but such values are difficult to estimate.

Top-down models suggest that a direct (Pigouvian) tax on carbon emissions, channelled through general government spending and large enough to constrain emissions, would be an expensive strategy. Bottom-up analysts would probably agree, because a tax would do little to remove the market barriers to energy efficiency improvements. Both groups would likely agree that a tax, perhaps revenue neutral or channelled to investment, to slowly increase the price of energy is justified to capture the many environmental and other externalities from energy use. The bottom-up models, however, identify additional emission reduction potential under the assumption that the barriers to energy efficiency can be reduced.

3 Barriers to energy efficiency

The economic situation regarding demand for energy services, which drives the demand for energy supplies, is fundamentally different from product markets. Energy efficiency is one of many characteristics of the products using energy. It therefore does not exist as a tradable good in any market in the traditional sense. Energy consumers demand energy services, but they buy and sell energy commodities as fuel and electricity. The resulting energy end-use "market failure" includes the following categories of barriers (Fisher and Rothkopf, 1989; Jochem and Gruber, 1990; Reddy, 1991; Krause et al., 1993).

1) lack of information and experience among both energy users and equipment suppliers;

2) difference in economic and risk criteria between energy users and suppliers;

3) separation between those who pay the costs of efficiency improvements and those who receive the benefits;

4) fuel and electricity prices that are subsidised and do not reflect environmental and national security externalities.

Little of the energy-efficiency potential identified by bottom-up studies will be realised in the absence of government policies to reduce these barriers. This "gap" suggests policy measures focused on stimulating innovation and investments in energy-efficiency improvement. Such policies include accelerating technology development and demonstration, stimulating product demand via procurement policies, applying efficiency standards to information-poor sectors, encouraging utility DSM programs, and in general finding ways to create markets for energy savings, thus stimulating further innovation. Efforts to implement many such measures are already underway in some industrialised countries, and this policy area is evolving rapidly.

An advantage of a market approach to emissions reduction is that it shifts much of the information burden from regulators to polluters. In centralised industries, such as motor vehicle manufacturers and electric utilities, the relative costs of changing processes or reducing emissions are known, but not necessarily available to the public or to regulators. On the other hand, when the polluters are individual households and firms, who generate emissions via their energy use, information about the technologies and costs for saving energy and emissions is generally lacking.

The need for non-price policies is evidenced by the lack of response to energy-prices observed in several groups of energy users, indicating small price elasticities of demand (Levin Kruse, 1991; Nielsen et al., 1992). In some energy-demand sectors, income effects dominate over price effects. For example, European cars did not increase significantly in fuel efficiency despite the fuel price increases of the 1970s. In the absence of regulatory measures that drove fuel efficiency improvements in the U.S., the gradual increases in income were enough to cancel the effect of large price increases. This suggests that other policies, which do not rely on consumer price-response, may work better if they can stimulate technical innovation. These policies, and their potential relationship with innovation, cannot be addressed in macroeconomic models, which deal with the entire issue through one or two aggregate parameters such as the price elasticity.

Many energy end-uses such as lighting are essentially saturated and are not very sensitive to marginal budget variations, which helps explain the low reported price elasticities in non-industrial sectors. Thus, higher prices will not discourage existing uses, and there is little room for lower prices or higher efficiency to stimulate new uses. The latter "takeback" effect is limited by the same information barriers that discourage energy-efficiency investments, which also make it difficult for customers' to know what behavioural changes can reduce their energy costs. This is essentially the same issue of elasticity of demand, and its effect is also limited by the fact that reduced energy costs are mitigated somewhat by higher equipment replacement costs. If consumers are astute enough to react to the energy cost savings, there is no reason to think that they will ignore total costs or overlook possibilities such as using compact fluorescent lamps (CFLs) preferentially to reduce the hours of use of their less efficient lamps.

4 Energy performance standards: Overcoming the information barriers

The government policy role may involve increasing consumer information and instituting energy-efficiency standards. Examples of information programs are ongoing or planned appliance performance labelling programs in both North America and Europe. Evaluations have shown that information programs alone are generally insufficient to stimulate significant changes in technology, although they can complement and amplify other program approaches (Nadel et al., 1990). An additional government role is the sponsorship of research and development in energy-efficient technologies. Such expenditures have been increasing in recent years but still represent less than 10% of the energy research budgets in most OECD countries (IEA, 1987).

There are two basic types of energy-efficiency standards: prescriptive standards that mandate that a certain technology must be used, and performance standards that require a certain level of overall performance. Generally, prescriptive standards are simpler and are used to govern the efficiency of various types of components and stand-alone equipment. Performance standards are more flexible and are used to govern the overall system efficiency of, for example, a building or functional areas within a building. These standards are more complex and generally impose greater requirements for compliance and verification. Prescriptive component standards are also advantageous for improving the energy performance of equipment replacements in existing buildings and facilities (Atkinson et al, 1993). In new buildings, however, system performance standards are appropriate to allow designers flexibility to exploit system interactions that provide better overall energy efficiency with lower cost and greater comfort than simply by following component standards.

U.S. appliance energy-performance standards began with state-level standards in California and were first adopted at the national level in 1987, for application in 1990 to residential refrigerators, freezers, water heaters, furnaces and air-conditioners. The national standards were strengthened in 1993, and further tightening is being discussed. The standards have had a major impact, for example reducing the energy use of new refrigerators and freezers by as much as 60 percent at low cost to consumers, less than $0.03/kWh saved including administrative overhead (McMahon et al., 1990). There is no evidence that these efficiency improvements would have resulted in the absence of the standards.

In addition, energy performance standards have been extended to some classes of lamps, motors, plumbing fixtures and air-conditioning equipment. Meanwhile, many U.S. states and municipalities have adopted building energy standards, and California has a particularly sophisticated code that combines prescriptive measures with performance-based compliance options. Building codes that extend to lighting and air-handling efficiency are also used in several states including California and are being developed at the national level.

Elsewhere, Sweden has some of the strictest building thermal standards in the world, and as a result Swedish housing is among the most comfortable and energy-efficient despite the severe climate. Although Danish building energy-efficiency at one time lagged far behind that in Sweden, the strengthening of thermal standards and investments in building retrofits have made Danish buildings nearly comparable in energy efficiency to those in Sweden (Schipper and Johnson, 1993).

A recent international survey identified approximately 30 countries that have mandatory building energy efficiency standards in place, and another 15 where such standards are presently proposed or voluntary (Janda and Busch, 1994). Most of the mandatory standards are found in industrialised countries, although several developing countries, particularly in Southeast Asia, have at least proposed or voluntary standards. Many of the standards are derived from those developed in other countries, such as the U.S. or Germany, and adapted to the local climate and building practice. The development of standards is generally hindered, especially in developing countries, by a lack of data on energy consumption and building practices, lack of compliance verification and enforcement procedures, and lack of performance testing and development facilities.

For household appliances, energy efficiency standards are an effective way to remove energy-wasting products from the market and capture a large share of the least-expensive energy-saving potential. Engineering-cost analysis of refrigerators and freezers in Scandinavia indicates that an energy-consumption level somewhat less than today's best-available models can be achieved with consumer life-cycle costs no higher than the average new model on the market today (Pedersen, 1992). If mandated by standards, such efficiency levels could be the norm as soon as 2002 (see Figure 1), many years before this level would be reached without policy intervention (Swisher, 1994). These levels of efficiency are technically feasible, and their economic penalty is not likely to be significant.

Figure 1. Historical trends and energy-efficiency standards analysed for fridge-freezers in Sweden. Energy use is normalised per adjusted litre of refrigerator space, as reported by Nutek (1992). Past trends show best-available, average new-model, and average stock efficiency. Average stock efficiency is projected into the future based on two sets of standards: with and without technology procurement (Swisher, 1994).


It is difficult to evaluate the economic impact of mandating further efficiency levels, although they are clearly technically feasible, but it is likely that both purchase price and life-cycle cost would increase, based on present engineering data. However, it is also quite possible that after more than ten years of development, manufacturers would find less expensive ways to produce energy-efficient products than those presently available. Preliminary results from the US suggest that the actual costs to manufacturers of complying with the 1993 appliance efficiency standards were less than estimated in engineering analyses supporting the development of the standards (Turiel, 1993).

Although these performance levels represent a dramatic improvement over the appliances sold today, the average stock energy consumption values would not be reduced to such efficient levels until after 2010, because many less efficient models would still be in service in the meantime (see Figure 1). The rate of energy-efficiency improvement is limited by the rate of turnover of existing equipment. It is also subject to constraints on market penetration rates, as there is little incentive for appliance producers to develop and introduce new models with efficiencies greater than required by existing standards.

Other examples of energy-efficiency standards include the U.S. Corporate Average Fuel Economy (CAFE) standards, which doubled automobile fuel efficiency in less than ten years. Most improvements were achieved through technical efficiency and design, with little reduction in interior space or safety (Ross, 1989). These improvements are sometimes attributed to price increases during the 1970s, but it seems curious that the average fuel economy values exactly met the targets in most years. Statistical analysis of the manufacturers' fuel-economy progress shows that the CAFE standards were the primary cause for the improvements (Greene, 1990).

Can energy-efficiency standards be economically efficient, even though they impose additional regulations on consumer behaviour? The high engineering benefit/cost ratios and low administrative costs of existing programs suggest that they can indeed. The key advantage is that standards overcome information barriers in diffuse markets such as home appliances, and they reduce the risk to vendors of stocking and selling energy-efficient products. To the consumer, standards overcome the uncertainty and invisibility of energy efficiency improvements. To the vendors, standards convey the essential information that yes, customers will buy efficient products. While such regulations suggest that consumer choice is reduced, in fact, there is just as much chance that producers will introduce additional models that surpass the standards, and this has been the case in the U.S.

Energy-efficiency standards are appropriate for product lines with relatively uniform characteristics and functions, in products that already exist in relatively mature markets. Although standards can be technology-forcing, this is generally not politically acceptable with producers, and standards are therefore not a mechanism to introduce more efficient products than those already on the market.

5 Technology procurement: reducing the risk of introducing efficient products

Other governmental policy measures that further enhance the rate at which energy efficient equipment is introduced to the market include stimulating manufacturers to produce efficient equipment via procurement and demonstration, and stimulating the adoption of existing energy-efficient technologies through voluntary agreements and by encouraging utility demand side management (DSM) programs. An innovative energy-efficiency policy intervention is public technology procurement, which has been developed in Sweden by NUTEK, the technology and industrial development ministry.

This process combines government incentives with guaranteed orders from organised buying groups (such as apartment managers) in a competitive solicitation for improved energy-efficient products (Westling, 1991). Manufacturers are invited to enter prototype models with certain features, including a specified minimum energy efficiency, and the entries are judged according to their efficiency and how well they satisfy the other requirements (Nilsson, 1992). The winner(s) receive incentive payments and a guaranteed initial order sufficient to begin production of the new model, thus removing a large part of the risk of introducing new energy-efficient models in their product line.

This process was successfully completed in 1991 for fridge-freezers, with the winning model's energy use 30 percent below the previous best available and 50 percent below the average in the market (Fritz, 1990). Although the winning model entered the market with nearly a 50 percent price premium, within one year the price premium was reduced to about 10 percent and a competing firm offered a new model with energy use comparable to the winner and a price close to other models on the market (NUTEK, 1992). The administrative costs of the procurement process are rather low, about $300,000 in the case of the fridge-freezer procurement (Lewald and Bowie, 1993).

The procurement process has also been applied in Sweden to high-performance windows, high-frequency lighting ballasts, computer displays that turn off automatically, and most recently to washing machines tailored to use in small households. The winning window products have about three times the thermal resistance of conventional glazing products, and these improved products are now entering the market both in Europe and North America.

NUTEK has recently conducted a successful public technology procurement for automatic shut-off computer monitors (Lewald and Bowie, 1993). There are large potential energy-efficiency improvements in computers and other office equipment that can be achieved at very low incremental cost. The automatic shut-off monitors and other energy-efficient office-equipment products are expected to gain a large market share in the next generations of office-equipment technology (Dandridge et al., 1993), with energy savings of more than 50 percent compared to current equipment models. These savings can take place quickly, because of the fast turnover of electronic equipment. The improvements are driven by the rapid technical advances in this area, leaving little need for additional programs to further accelerate the market penetration of efficient products, once they have been introduced.

A similar U.S. program, Energy Star by the Environmental Protection Agency (EPA), is expected to move the computer market from zero to nearly 100% power-managed desktop computers in 4 years, at essentially no cost. This voluntary program, which certifies power-managed computers and efficient peripheral equipment with the "Energy Star" label, nevertheless achieves efficiency improvements that would probably not have happened as soon without EPA involvement with the computer manufacturers.

The Energy Star program grew from the EPA's "Green Lights" program, a voluntary corporate lighting energy-efficiency program. Already hundreds of large commercial firms, representing several percent of the national commercial floorspace stock, have joined the program and committed to lighting energy-efficiency retrofits covering 90% of their floor area. The demand generated by this program should have a significant effect in terms of moving the lighting equipment industry in the direction of greater energy efficiency.

The Super Efficient Refrigerator, or "Golden Carrot," Program in the U.S. is a variation on the Swedish technology procurement program. In this case, several large utilities created a pooled incentive, which was offered to manufacturers as the prize in a competition to develop a highly-efficient CFC-free fridge-freezer. The incentive will be paid as each unit of the winning model is delivered in the utilities' service areas. The program will also be expanded to clothes-washers and air-conditioners. The technical progress stimulated by this program might make it possible to further tighten the fridge-freezer energy-efficiency standards by 1998, a move that would probably not have been possible without the program.

The effect of technology procurement and related "technology push" approaches is to accelerate energy-efficiency gains by raising the performance of the high-efficiency end of the market, which serves to accelerate energy-saving potential earlier in time and is particularly effective in combination with energy performance standards. Standards eliminate the least efficient models from the market, but their energy-saving impact is limited by presently available technology because they cannot improve the high-efficiency end of the market. It is therefore possible that, without a "technology push" mechanism such as the technology procurement process, development of new products with further energy-efficiency improvements would not occur. The introduction of new models at the high-efficiency end of the market pulls the average efficiency upward, even without the imposition of energy efficiency standards, but their overall market impact can be amplified by presence of progressive standards to remove less efficient products (Swisher, 1994).

Thus, energy performance standards and technology procurement are complementary to each other. The procurement process accelerates the introduction of more efficient equipment on the market, while standards help remove less efficient models and increase the market penetration of their more efficient replacements (see Figure 1). Together, these two policy instruments provide clear signals to appliance producers that the future market will reward the development of more efficient new models with growing demand and sales, thus removing much of the risk associated with the introduction of high-efficiency models.

Technology procurement and related programs are appropriate for available technologies that have yet to be introduced in commercial products. In such cases, the risks are high for producers to introduce a new product without knowing if customers will buy it. Overcoming this type of risk is the most important effect of programs such as technology procurement. For products with clear side-benefits, such as power-managed computers, that can be expected "sell themselves" once introduced, technology procurement is sufficient to ensure their adoption in the market. Other products may need an additional mechanism to increase further their market penetration. This function can be achieved by standards, as described above, or by other types of incentives such as those provided by utility DSM programs.

6 Demand-side management: utility investment in energy efficiency

In the U.S., the last ten years have seen a drastic shift away from the construction of large central power stations by electric utilities. This has been the result of two trends, which are somewhat opposed to one another. The first is the advent of the Public Utilities Regulatory Policy Act (PURPA) legislation, which was meant to stimulate the development of renewable energy sources for utility power. In addition to introducing more renewable energy, PURPA has stimulated a general deregulation of the electric supply industry, allowing greater access to smaller, independent power producers (IPPs) and the advent of competition for supplying utility power. The second trend is the widespread adoption in the U.S. and Canada of demand-side management (DSM), to reduce electricity demand while still meeting customer energy-service needs. The arrival of DSM, however, has been driven by the close regulatory oversight of the state-level public utility commissions (PUCs) in several states and provinces, especially those with ambitious environmental agendas.

According to the Electric Power Research Institute (EPRI), "DSM is the planning, implementation, and monitoring of those utility activities designed to influence customer use of electricity in ways that will produce desired changes in the utility's load shape, i.e. changes in the time pattern and magnitude of a utility's load. Utility programs falling under the umbrella of DSM include: load management, new uses, strategic conservation, electrification, customer generation, and adjustments in market share" (Gellings, 1985, p. 1468). The reason that utilities began to pursue DSM programs was that they wanted to influence their customers' load profile to better match their production capacity, mostly by reducing peak demands, while some utilities have practised DSM as a way of selling additional power during times of low demand.

However, it is strategic conservation measures that can reduce the amount of electricity required to meet a given end-use. This allows the utility to meet its customers' energy-service needs with less fuel use and generally less pollution. As regulatory bodies in the U.S. and Canada have allowed and even encouraged utilities to recover the costs of DSM programs, more utilities have implemented programs to capture cost-effective efficiency opportunities that reduce the total cost of electric service. The cost of saving a kWh is often less than the cost of generating and transmitting one. In general, DSM programs can increase customer satisfaction and reduce the environmental impact of power plant siting and use.

These programs range from relatively passive informational programs, to financial incentives via rebates or loans for customer investment in end-use efficiency measures, to active investment by the utility and bidding programs where utilities request proposals from outside parties to contract for demand reductions. Generally, the less direct the program, the more it relies on the customers' knowledge and financial criteria, and the less reliable are the results.

Financial incentives range from low-interest and deferred-payment loans to subsidies and rebates for the purchase of energy efficient equipment. Loan programs have generally not been very successful, as relatively low numbers of customers are typically willing to take on debt in order to save energy (Nadel et al., 1990; Larsen et al., 1993). Rebate programs have been more successful and are best applied to fully saturated equipment markets such as refrigerators or lighting, where the rebates would not be likely to increase the total number of units installed, regardless of efficiency. This criterion would not apply, for example, to air-conditioners. Although more reliable than loan programs, rebate programs still do not provide the utility with any direct control over the level of energy or peak demand savings, and these programs sometimes suffer from high administrative costs, especially in the residential sector.

Direct installation programs are more expensive, but they have the potential to be simpler and therefore more cost-effective than incentive programs. This approach is often implemented by contracting with an energy service company (ESCO), which provides the energy end-use equipment in exchange for a share of the energy savings. Direct installation programs avoid the problem of consumers' lack of information, and they are particularly attractive for difficult sectors such as rental housing and offices.

Such programs have greater consumer participation rates than incentive programs. For example, more than 90% participation was achieved in residential retrofit programs in Hood River, Oregon in the mid-1980s and more recently in Espanola, Ontario (Goeltz and Hirst, 1986; Sharpe et al., 1993). In the commercial sector, the New England Electric System's (NEES) lighting retrofit program provides both financing and comprehensive training to electrical contractors for advanced lighting systems, thus increasing the ability of the contractors to apply efficient technology in their work outside the DSM program as well (Miller et al., 1992).

The successful implementation of end-use energy efficiency measures requires at a minimum that their adoption does not depend solely on the customers' present economic criteria. As discussed above, the implicit discount rates applied by energy-users to energy-efficiency investments are much higher than utility discount rates, and they are evidence of the many barriers to the adoption of cost-effective efficiency improvements by customers. High consumer discount rates suggest that the consumer treats energy investments as rather risky. This perception is due to imperfect information, doubt that the investment will be recovered before the consumer moves, and other reasons. From the societal perspective, however, such investments appear to be a relatively low risk opportunity to provide energy services, especially given the increasing supply-side uncertainty in the electric utility industry.

In a perfect financial market, such low-risk investments should normally be evaluated at a relatively low discount rate, rather than the high rate typically applied by consumers. As a result, it appears that other criteria, closer to the supply-side criteria, must be applied for even a fraction of the potential for energy efficiency to be realised. In other words, some energy-supply entity (i.e., generators, distributors or energy-service companies) must be in a position to choose, on a relatively equal basis, between financing energy efficiency or paying the potentially high marginal costs of new electricity.

A key element of utility DSM programs is thus to bring the economic evaluation of energy efficiency onto an equal basis with supply expansion through a process of integrated (supply and demand-side) resource planning (IRP). In North America, active province and state-level regulation of vertically integrated utilities has made this kind of evaluation a routine aspect of utility business. The resulting DSM investments may reach US$3 billion in the U.S. in 1994 and C$1 billion in Canada by 1996 (Bauer and Eto, 1992; Canadian Electrical Association, 1992). The effect of DSM programs can be seen in equipment markets. For example, the majority of electronic lighting ballasts and about half of the integral compact fluorescent lamps (CFLs) have been bought with utility incentives (Geller and Nadel, 1994).

Recently, North American DSM programs have begun to evolve away from general information and rebate programs, which are common but have been criticised for high costs and low participation rates, to more targeted programs that better respond to customers' needs and produce more reliable energy savings (Nadel, 1992). In addition, more attention is being paid to capturing long-term efficiency improvements by influencing equipment manufacturers and building designers to offer more energy-efficient products.

In Europe, however, DSM has not developed so quickly, partly because vertical integration is not as common. A more common utility model is municipal distribution utilities that buy their power from larger (often nationalised) generation utilities. In Scandinavia, for example, the few fully-vertically integrated utilities, such as Oslo Energi and Stockholm Energi, may be broken up as a result of the deregulatory policies now taking effect. Moreover, the regulatory intervention that made DSM possible in North America is not apparent in Europe, where utilities operate more independent from government oversight. Nevertheless, energy suppliers in some countries, such as Denmark and the Netherlands, have become more active in encouraging energy efficiency.

In Sweden, Vattenfall AB, the largest electric generator and wholesaler and formerly the national power board, recently completed "Uppdrag 2000," (project 2000) a field study of energy efficiency options. The study produced the first detailed statistical picture of energy in the Swedish service sector and evaluated the potential cost and performance of a wide range of energy-saving measures (Hedenström et al., 1992). The Uppdrag 2000 results suggest savings potential in the buildings sectors of about 10-15 percent based on retrofit measures that could be implemented immediately (Hedenström, 1991). While the achievable energy efficiency potential is not as great as some studies of technical potential suggest, Vattenfall's results are an underestimate because they ignore the longer-term efficiency potential in new buildings and replacement equipment. The efficiency measures identified under Uppdrag 2000 are not presently being carried out, partly due to the continuing surplus of electric supply in Sweden. Presumably, these measures could be implemented later when supply constraints appear, or to defend the utility's market under the threat of new supply competition (Swisher et al., 1994).

In both Sweden and New Zealand, the former state electric power boards have been corporatised, with the eventual aim of privatisation. For the time being, this situation is the worst of both worlds for DSM: the new corporate monopoly faces neither competitive market pressure nor public oversight via binding regulation. These organisations' business strategies thus become a matter of positioning themselves to defend and perhaps increase their market position in case of real competition being introduced in the future. One possibility is that the utility uses DSM as a tool to defend its market share against new competition by making their service more attractive and economical to the customer. Rather than reducing the customers' rates to respond to competition, the utility can thus reduce their bills, including perhaps maintenance and other fuel costs, making continued service from the utility relatively attractive compared to the competition. Although the utility could recover only part of its costs and lost revenues from the DSM recipients and variable cost savings, the lost revenues must be compared to the alternative case where customers are lost, taking all of their revenues with them (Swisher and Hedenström, 1993).

In a country such as Norway, with lower power supply costs than other countries, there would appear to be relatively little incentive for DSM. However, if such a country is able to sell power to the countries with more expensive supplies, the value of saved energy should increase to the level of costs in the importing countries, not just the low domestic costs. This raises the possibility of using DSM to "rescue" low-price electricity for export to higher-paying customers. A less attractive alternative, from the national perspective, would be to allow power exports to force up the domestic electric prices.

Another strategy is for electricity distributors to use DSM for reducing area-specific costs, which depend most on distribution and local transmission costs. Recent analytic advances in accurately determining utilities' area-specific costs allows the precise targeting of programs in areas where the avoided costs are relatively high (Orans et al., 1992). A utility DSM strategy, built on intensive campaigns with limited geographical coverage, rather than broad coverage but modest impact, can avoid capital investment costs in transmission and distribution that are driving the current investment needs of utilities with relatively slow total load growth. This may become an increasingly important strategy as national and other integrated utilities are broken up into smaller, competitive units.

7 Free riders and transaction costs

Every three years, the Northwest (U.S.) Power Planning Council (NWPPC) produces 20-year demand forecasts and electricity-resource plans. The NWPPC has made two important changes in the region's energy planning process. First, energy forecasts now explicitly consider the uncertainty of demand. Second, energy efficiency improvements are now treated as part of the electricity-supply resource. The plans account for new and existing electricity-supply resources, energy-efficiency opportunities already captured by consumers and by the Council's and other programs, and future energy-efficiency potentials. Estimates of future energy-efficiency potentials take into account evaluations of their achievable market penetration over time and the administrative costs and uncertainties associated with their implementation (NWPPC, 1991).

Because the efficiency measures must be implemented in a fixed time interval to defer construction of new generating plants, this is a more realistic and conservative approach to energy end-use analysis than simply estimating technical and economic potential. The most recent plan identifies as its lowest cost resource the immediate acquisition of energy savings sufficient to meet all the new demand in forecasts based on low and medium rates of economic growth (high growth requires additional generating resources) (NWPPC, 1991).

The fact that significant savings would still be achieved without DSM programs means that there are customers who will take advantage of utility-sponsored incentives, even though they would have invested in efficiency measures without the incentives, i.e. there will be "free riders." Although free riders do not impose additional costs on society (except administration costs), the utility must still pay both the incentives and the program administration cost for the free riders' participation. Evaluations of North American DSM programs report free rider fractions from less than 10 percent to more than 50 percent, depending on the type of program, and experience has shown that programs can be designed to avoid excessive free-ridership (Nadel et al., 1990).

When a utility tries to capture the lowest-cost efficiency measures, which are generally found in new and replacement equipment rather than retrofits of existing buildings and equipment, there tends to be a high fraction of free riders, which increase utility costs. However, such measures are one-time opportunities that are lost or become much more expensive if they are not exploited when the equipment is first installed. Capturing this inexpensive "lost opportunity" energy-efficiency resource should therefore be a high priority, despite the risk of free-riders. These are measures that are appropriate for energy performance standards, for end-uses such as household appliances and service-sector lighting. Even relatively modest standards helps to capture many low-cost "lost opportunity" resources. This allows utility DSM planners to set thresholds for incentives that exceed the standard by some degree, requiring more elaborate measures where fewer free riders are likely (Swisher et al., 1994).

In addition to free-riders, significant transaction costs exist for many of the energy-efficiency improvement options. It is sometimes argued that these are high and, in fact, the reason why energy efficient technologies have not already been adopted more by the markets (Sutherland, 1991). In some studies, transaction costs have implicitly been assumed to be zero (Lovins and Lovins, 1991). In fact such costs and limitations exist and can be measured and explicitly included in bottom-up analysis, as shown by the NWPPC example above. However, penetration limits and implementation costs vary widely with the technology and type of program applied.

Transaction costs can be estimated both in terms of consumers' "search cost," the time and trouble it takes to acquire more efficient products, and actual administrative costs of energy-efficiency programs. At the low end of the program costs, the procurement of energy-efficient and environmentally sound refrigerator/freezers promoted by the Swedish government reduced energy use in new models by 30 percent with an estimated transaction cost of about $300,000, or less than $0.001/kWh (Nilsson, 1992). The $50 million spent by the U.S. government on appliance standards amount to even lower costs per saved kWh (Levine et al., 1994). For US utility demand side management (DSM) programs, administrative costs add on average 10-30 percent to the technology costs (Berry, 1989; Nadel et al., 1993). These costs tend to decrease with increasing participation rates, which reduce the importance of fixed program costs, up to very high participation rates where decreasing returns drive costs up.

Consumer transaction costs for acquiring energy efficiency are more difficult to estimate, but they appear to be in the same range of 15-30% of technology costs (Bjorkqvist and Wene, 1993). This is significant but certainly does not explain the consumer payback "gap" shown by the very high implicit discount rates. This gap must be explained by the other barriers described earlier. Assuming that most energy-efficiency measures include transaction costs on either the consumer or the program side, or perhaps a mix of both, it seems that a conservative estimate of the additional cost is about 30% of the technology cost, although many programs can be implemented with lower administrative cost and with little effort by the consumer. While transaction costs are a real cost that should be included in DSM planning, as illustrated by the NWPPC, they cannot explain the barriers to energy-efficiency investment, nor do they indicate that such barriers represent irreducible costs that are thus economically efficient.

The costs of DSM programs very widely and are somewhat larger than the simple technology costs, as discussed above. Most programs report costs of saved energy of $0.02/kWh or less (Nadel et al., 1990). Generally, the programs with high rates of free-riders involve measures that are highly cost-effective and therefore have very low technology costs. While some critics point out the uncertainty in these costs and argue that DSM programs are much more expensive than claimed by utilities (Joskow and Marron, 1993). However, they ignore uncertainties that could reduce such costs, such as free-drivers, described below (Hirst and Reed, 1991).

Another source of DSM cost data is the competitive bids that have been tendered for DSM projects at around 4-7 cents/kWh, decreasing over time (Goldman and Kahn, 1993). The firms involved are small and undercapitalised, with relatively few projects to diversify risk, so there is a high risk premium (perhaps 40%) included in their costs. Thus, these costs should be seen as the high extreme for what saved energy, or "negawatts," cost, although they are less than the costs suggested by critics. In any case, there is presently a great deal of effort underway to evaluate the effects and costs of DSM programs, in order to resolve some of these uncertainties and to improve program design (Hirst and Reed, 1991).

8 Market transformation: one-time change or continuous process?

Utility DSM programs can transform markets by encouraging equipment distributors to stock, advertise and sell efficient products, even to customers who do not participate in the programs. Additional adopters of efficient products, outside of existing programs, are referred to as "free drivers," and they act to increase the effect of utility programs by investing in energy efficiency without participating in the DSM program.

One example of this transformation process is British Columbia Hydro's efficient industrial motors program, which used both customer and distributor rebates to increase the market share of efficient motors from 4% to 64% in 3 years, at a cost of $0.012/kWh saved (Nelson and Ternes, 1993). BC Hydro has reduced the rebate payments twice after transforming the market, and they are now applied only to even higher performance levels. Because vendors tend to stock one line of motors, this transformation has made the efficient motors the norm, making further utility expenditures for future energy savings unnecessary. Free riders were estimated at about 10% initially, but now there are many times more free drivers. Similar results were achieved in a similar program at Ontario Hydro.

Market transformation, in which a permanent shift is made to more energy-efficient products, is becoming an explicit goal of DSM programs. Utilities want to use their investment in these programs to achieve a long-term improvement in energy efficiency and not have to continue paying for energy savings indefinitely. Market transformation can also be achieved by other policy instruments, as illustrated above in the cases of technology procurement for power-managed computers and refrigerator energy-efficiency standards. The effect of the latter can be amplified further via technology procurement, as shown in the Swedish and U.S. examples.

The Super Efficient Refrigerator Program in the U.S. is aimed explicitly at market transformation. Similarly, a consortium of U.S. gas utilities are providing support for the commercialisation of a gas heat pump, which will improve the efficiency of gas-fired home heating to more than 100 percent. The utilities will pay incentives directly to the selected manufacturer, in order to buy down the retail price, and will be repaid from this model's sales after a certain level of sales have been reached (Geller and Nadel, 1994).

The Model Conservation Standards were developed by the Bonneville Power Administration (BPA) and the NWPPC to encourage builders to improve the energy efficiency of new houses and to make such improvements sufficiently acceptable to builders to allow the adoption of stronger mandatory building codes in the region, thus moving from incentives to consensus standards. BPA and NWPPC also collaborated with building manufacturers, paying rebates for the first four years, to get a commitment of 90% compliance with the voluntary standard. The cost of saved energy for the program, from 1983-2003, is estimated at $0.03/kWh, plus about $0.005 for implementation costs (Geller and Nadel, 1994).

The U.S. EPA is assembling a consortium of utilities and NGOs to pay manufacturers incentives to buy down the cost of integral CFLs, together with public education programs, which over time can be expected to increase consumer awareness of CFLs as a household energy- and money-saving measure. Although information programs are rarely sufficient to change consumer purchasing habits, they can complement and amplify incentive programs. A study of utility CFL promotional campaigns in Sweden found that participants were willing to accept almost twice as long a payback time for CFLs as non-participants (Bülow-Hübe and Ankarljung, 1991).

The process of market transformation is usually described in terms of an S-shaped curve, such as one commonly uses to describe technology diffusion, with an early introduction phase, a phase of growing adoption, and a slowing phase as the ultimate market penetration is reached. Another way to look at this process is to consider the dynamics of market penetration as a function of technology performance. As shown in Figure 2, this approach allows us to describe the effects of the different program approaches discussed above.

Figure 2.Effects of energy conservation programs on distribution of product efficiency (solid line is without program, dashed line is with program).

The appropriate policy instrument to encourage energy-efficiency improvement depends strongly on the technology and the stage of commercialisation at which it faces market barriers. For technologies that are not fully developed, of course, research and development is necessary to make the technology available to the market. For young technologies that have not been commercialised, a "technology push" approach (A) such as illustrated by the Swedish technology procurement program is appropriate and has the effect of increasing the availability of products at the high-efficiency end of the energy-efficiency spectrum.

Commercialised products with a small market share are appropriate for incentives (B) such as DSM programs that encourage the selection of more efficient available technologies over less efficient ones, without changing the availability of different models on the market. This effect is important for energy end-uses such as lighting and pumping systems, where there are many different components to be selected and potentially improved.

The market share of mature technologies can be increased via efficiency standards (C) that eliminate less efficient products from the market. It is also possible to design "fleet" efficiency standards (D) such as the U.S. CAFE standards that govern the average efficiency of the entire product line. This effect is similar to that which should be achieved by general incentives such as energy price or tax measures, but as discussed above their impact can be limited by the existing market barriers.

The goal of market transformation is to achieve a permanent technological shift that requires little or no further intervention. If this process is successful, will it eventually exhaust the potential for energy-efficiency improvements? Or, will new technical opportunities provide additional potential that can be captured through further interventions? The experience of the last 10-15 years suggests that energy-efficiency in most areas is far from reaching any thermodynamic limit, and that new technological options are being developed as fast as existing ones are being implemented. This continued technical progress is partly the result of rapid progress in electronics and control technologies.

The appearance of new technologies is also an indirect result of the energy-efficiency policy measures already in place. In addition to achieving market transformation in a few selected technologies, policies such as standards, technology procurement and DSM have for the first time created a market for energy efficiency improvements, which did not exist before in any real sense. The existence of a market for energy efficiency can stimulate new progress and innovation, providing the conditions in which a continuous process of technical improvement is possible. The policy instruments that stimulate market transformation are needed only temporarily, but technical innovation can provide an on-going stream of new markets to transform, eventually making previous advances obsolete.

This postulated process is shown in Figure 3, which represents past energy demand growth and forecasts (A). These projections of very high energy demand growth, which led to plans for large expansion of nuclear and coal-fired electricity generation, did not develop accordingly, stimulating the development of bottom-up analysis to understand why the projections did not reflect reality. One hypothesis was that the reduction in energy growth was a temporary correction caused by the oil price increases of the 1970s, after which growth would resume (B).

However, the general observations that many of the energy-intensive goods and services were reaching saturation in industrialised countries, and that many technical improvements in energy end-use efficiency were becoming available, suggested that economic growth and material welfare could be maintained with significantly less energy supply and environmental impact than the existing projections indicated. During the following decade, until the mid-1980's, nearly every OECD country revised its energy-demand projections drastically downward (Baumgartner and Midttun, 1987). Still, conventional wisdom suggests that, even if much of the technical efficiency potential is captured, demand would eventually have to increase (C).

Figure 3.Possible future trajectories of energy demand.

The question raised here is whether the creation of a market for energy efficiency, through the variety of policy instruments described above, can stimulate a sufficient rate of innovation to make the process of market transformation a continuous path of energy-efficiency improvement (D). Rather than simply supporting research and development and responding to energy-technology innovations, energy policy would stimulate continuous innovation by the private sector, creating new innovations and opportunities.

9 The roles of policy, technology and institutional innovation

Weak environmental standards have the effect of making natural resources (energy, assimilative capacity, etc.) cheaper. This effect is often cited as an economic argument against environmental standards, especially by macroeconomic modellers with a static view of technological and institutional innovation (US NRC 1992). Because the world economy is becoming increasingly knowledge-based rather than resource-based, innovation, not cheap resources, make a country competitive. As an example of the unimportance of cheap natural resources, consider Germany's or Japan's success in international markets compared to the floundering former USSR. The centrally planned economies provide an example of the narcotic effect of subsidising energy and natural resources through low prices and weak environmental standards.

Resource efficiency can be a stimulus of progress and competitiveness, not a threat to economic welfare. Although the situation is more complex in developing countries, much of this strategy still applies. It is a popular economic myth that developing countries must come to use resources as profligately as the industrialised countries have in order to develop. In fact, if they do, they will be priced out of the market for resources such as oil. Thus, a resource-efficient development path may be the only feasible course. Moreover, non-fiscal policies such as safety and environmental standards can create the competitive conditions for technical innovation, leaving market competition to complete the technology selection and implementation process.

What is the role of energy prices and information in stimulating innovation? Clearly, energy price changes, through policy instruments such as emission taxes, create a similar type of pressure on producers of energy-using equipment. Because this pressure is exerted via the customers' demand, the effect can be diluted by customers' response to energy prices. Energy prices can directly influence technology development in the case of jet aircraft engines, but are much less likely to affect household appliances without the added stimulus of non-fiscal measures.

Econometric-based studies that enlarge the possible responses to energy-price changes to include income, substitution and innovation effects suggest that innovation is more important than substitution, and that policy measures other than price measures are more likely to stimulate innovation and technological change (Siniscalco, 1993). Other studies, however, find that such energy-productivity increases are more than overcome by reduced productivity in other areas due to higher energy prices (Hogan and Jorgenson, 1991). These results may be questionable because they extrapolate relations derived from the oil-shock years of the 1970s to the possible future case of a smooth, predictable energy or carbon tax increase.

It seems clear that certain policy instruments can help accelerate the technical innovation needed to reduce energy consumption and emissions. In fact, technical innovation may demand further innovations in policy and institutions, which can be considered as a type of "enabling technology" for more tangible technological developments. Institutional enabling technologies could range from specific regulatory standards to such intangibles as a democratic society that promotes sufficiently open exchange of information for innovation to occur. To speak of standards and other non-fiscal policies in general as enabling technology may seem vague and unfamiliar, but such institutional changes are not unheard of historically. Indeed institutional changes accompanied, facilitated, and were essential to many of the important technological changes in U.S. industrial history. These changes have occurred in the areas of law, finance, manufacturing, marketing and management (Gross, 1990).

In law, several innovations have been important to trigger technical innovation. The invention of the telephone led to the concept of a public utility and the national and international standards and network interconnections that allow the system to function and grow (Paine, 1921). The invention of the steam engine and the accompanying danger of explosions led to the first U.S. regulatory laws. Like environmental regulations today, these early regulations were supported by independent background research, some of it commissioned by the U.S. Congress. Although the regulations were opposed by the industry, it was only after realistic and enforceable standards took effect that sufficient public confidence in the industry permitted its rapid growth. A key legal innovation necessary today is to make energy and environmental regulation more efficient and less adversarial.

In finance, the railroads triggered the growth of Wall Street's stock markets to finance the new capital-intensive industry, the development of which in turn depended on the growth of the market. More recently, the U.S. venture capital industry co-evolved with the entrepreneurial electronics industry based in California's "Silicon Valley" (Gross, 1990). Financial innovation is needed to develop market mechanisms that handle energy savings (Fisher and Rothkopf, 1989; Krause et al., 1993). The difficulty is in trading a service that cannot be directly measured. This problem of intangibility is not unique to energy efficiency, however. The insurance industry sells "security" that often can be measured only according to what did not happen. Indeed, this relatively intangible service has been essential to the development of tangible and valuable goods and services that involve high degrees of risk.

In manufacturing, the concept of interchangeable parts for the purpose of efficient mass-production was first applied on a large scale to the manufacture of rifles for the U.S. Army (Rosenberg, 1969). This led to the development of the machine tool industry and the mass-production of bicycles, sewing machines, etc. Also, statistical methods of quality control were developed to assist the rapid production of naval fleets that became necessary during the second world war, such that a "Liberty Ship" could be launched five days after laying its keel. These methods became an important tool in increasing the efficiency of post-war industrial growth (Deming, 1982). Manufacturing innovation is required especially in the area of renewable energy technologies, the application of which is still limited until capital costs can be reduced through economies of production (mass-producing a large number of units) rather than economies of scale (increasing the size of each unit).

In marketing, the early electric utility industry used mass marketing techniques to identify products such as refrigerators with a higher standard of living. This increased their short-term sales, created the demand needed to develop base-load electricity, and provided the rapid sales growth that made possible the industry's strong economies of scale and improvement in reliability. Marketing innovations are needed to proliferate utility DSM and the innovative technology procurement practices described earlier. This includes identifying opportunities to develop products and implement services that improve energy efficiency, convincing customers to participate in markets for such products and services, showing producers that efficient products need not be risky, and changing utilities' culture to accept and learn to benefit from such activities.

In management, the invention of electric lighting and telecommunications both stimulated and required many innovations, including the development of the modern industrial research laboratory, as well as many systems of information management. The large private companies and public utilities that grew from these industries thrived on such innovations, which facilitated further technological progress. As mentioned above, the changing economics of the electric utility industry has led to the introduction of DSM throughout North America and its spread to other regions. Management innovations are required to advance the development of "small science," especially in areas related to energy efficiency and biological fields such as agriculture and silviculture.

10 Why does economic theory ignore these opportunities?

As noted earlier, existing macroeconomic (top-down) studies consider price effects, via measures such as a carbon tax, as the principal policy instrument to reduce energy use and emissions. Thus, they tend to ignore the possibility of exploiting the policy instruments described in this article. Is this neglect the result of a modelling framework that cannot treat these policy measures in detail, or simply the modellers' bias against the feasibility of such policies? In the top-down models, the most important parameters are gross domestic product (GDP), costs and prices, and parameters such as price and income elasticities that, aggregated into a macroeconomic framework, describe the overall energy-economy interactions (Wilson and Swisher, 1993). The present energy economy is presumed to be in an optimal state.

Bottom-up analysis treats the structure of the economy and performance of technologies in detail. No assumption is made a priori about which set of future options will be optimal, nor is the present energy-economy itself presumed optimal (Johansson and Swisher, 1993). However, essentially all top-down energy analyses are based on exactly these points as assumptions. By implicitly assuming that the present energy economy (like that of the past) is economically efficient, and that energy-economy interactions will be the same in the future as they were in the past, top-down models exclude some types of policies and institutional innovations, such as those aimed at improving energy efficiency, which would change the existing energy economy.

Thus, policy interventions such as those described above aimed are assumed to be economically inefficient. The possibility of efficient intervention is often equated to "belief in free lunch," in ironic reference to the ecologist's law, and this view has become dogma among some macroeconomic modellers (Manne, 1992). However, this view that "the current system is operating optimally" is not proven or even being tested empirically. While there appear to be ways to model the inefficiencies in the energy economy in a top-down framework, to date no studies have taken up this challenge. The apparent non-existence of potentially efficient policy interventions in support of energy-efficient technologies remains an untested assumption.

Part of the economic modellers' resistance to the possibility of inefficiencies or disequilibria, and therefore to efficient policy intervention, is that such conditions would introduce multiple solutions to their models, making it more difficult if not impossible to determine the "cost" of one scheme relative to another. However, single-point optimality in economic systems is a mathematical convenience, not an empirically proven fact. Chaos theory tells us that, in non-linear systems like the economy, small variations in initial conditions can lead to huge divergence in results. This means that a long-term single-point forecast is fantasy, because it would require precise measurement of variables that are completely invisible at the scale of macroeconomic analysis. Meanwhile, complexity theory indicates that non-linear systems tend to have many possible and nearly equivalent solutions, i.e., that a single optimum is impossible to find or even recognise (Davies, 1989). It is not difficult to imagine various historical paths along which different combinations of goods and services could be produced, using somewhat different sets of inputs, that would satisfy consumers at least as well as those which they consume today.

Although rather sketchy, there is some empirical evidence of the potential economic efficiency of non-fiscal mechanisms. There are numerous examples of energy-efficiency improvements, for which no significant "hidden costs" apply, that were not widely adopted until the imposition of standards or other non-fiscal policy measures (Levine et al., 1994). In some cases feasible and cost-effective technologies are not even introduced, because manufacturers consider them too risky or simply not worth the trouble. In particular, energy performance standards can be efficient because they eliminate much of the consumers' transaction costs of purchasing energy efficiency.

One explanation for the existence of such options, in terms of a macroeconomic framework, is that many energy-demand subsectors are relatively price-inelastic. From an empirical viewpoint, aggregate elasticity values are very uncertain and can vary considerably by sector and end-use (Bohi and Zimmerman, 1984). However, this elasticity parameter alone can have a dramatic effect on the magnitude of the national (or global) economic losses predicted by macroeconomic models (Hogan and Manne, 1979). Unfortunately, price elasticities are rarely measured or even defined at a disaggregated level that would apply to individual consumer classes or end-uses.

From a bottom-up perspective, one can model customers' investment response to energy price changes, i.e., a price elasticity of demand. Because it is based on technology decisions, this elasticity effect is relatively long-term and permanent. It is more complex than a constant elasticity, however, because its value depends on the availability of energy-efficiency measures at different cost levels, i.e., the "supply" elasticity of energy savings. Because of the many barriers to energy-efficiency investments, described earlier, this elasticity is generally low.

The other effect captured by a price elasticity, changing energy-using behaviour, is a short-term and reversible effect. For many energy end-uses such as those in buildings, this effect also appears to be small, because 1) the observed behavioural price response is small (Levin-Kruse, 1991; Nielsen et al., 1992); 2) lighting and other building services are essential services that are practically saturated in industrialised countries, and therefore not sensitive to marginal budget variations; and 3) the same information barriers that discourage energy-efficiency investments also make it difficult for customers to know the degree to which behavioural changes can reduce their energy costs.

11 Recommendations for future abatement costing studies

An exceptional study from the Netherlands applies detailed econometric analysis to the problem of firms' adoption of energy-efficiency measures (Velthuijsen, 1993). The results show that non-residential energy-users display variable but generally unresponsive behaviour with respect to energy-efficiency opportunities and energy price changes. The variations in observed adoption rates and price elasticities can be explained according to the presence or importance of several of the common energy-efficiency barriers mentioned earlier, such as information gaps, the small size of the energy bill in the total budget, etc.

The Dutch study is exactly the type of research that is needed to resolve some of the empirical issues raised here. There is generally a need for more empirical bottom-up work, backed by such detailed econometrics, in order to treat separately the subsectors or end-uses with relatively high and low price elasticity values. Those categories with relatively inelastic energy-price response indicate opportunities where non-fiscal policy instruments can be effective in reducing emissions and energy-intensity. Such measures would reduce the economic impact of energy or carbon taxes, and the reductions would be little influenced by the "rebound" or "takeback" effect. Meanwhile, relatively elastic categories would respond to energy or carbon taxes by reducing energy use and emissions with little economic impact.

An energy system model that allows for energy-efficiency programs to be applied preferentially to inelastic subsectors or end-uses would tend to show lower energy and emission reduction costs than a traditional undifferentiated macroeconomic model. Thus, it would allow for the prospect of efficient policy intervention to improve technical energy efficiency. Another approach to modelling such interventions would be to assume that consumer behaviour forces energy "market-clearing" to occur at a price lower than the supply price, causing excess consumption. One could then allow partial reconciliation of the market-clearing price, based on interventions to reduce the barriers to energy-efficiency investments. Again, energy and emission reduction costs would be relatively low.

While both of these approaches have potential to improve the realism and utility of energy models, both introduce problems in terms of data requirements and consistency with the theoretical explanations of the processes involved. Either approach would require a great deal of detailed empirical analysis of the types described above. Also, there is no consistent theory explaining just how the inefficiencies and disequilibria in the energy-services market are quantitatively related to familiar concepts such as prices, quantities and production functions. This uncertainty would make the formulation of the modified model somewhat arbitrary and subject to severe limitations to its empirical calibration. Nevertheless, it would still be an advancement compared to the perfectly-efficient equilibrium assumption and would allow for some interesting exploratory work.

One interesting question to explore with such a model is whether the process of creating a market for energy efficiency, captured in the model by convergence of the market-clearing price or by efficiency improvements in inelastic subsectors, would cause a one-time reduction in energy intensity or a continuous long-term transformation, as shown in Figure 3. The answer would likely depend on the assumptions made about technological progress. If technology advances in response to endogenous changes, for example increases in the demand for or the value of energy-efficiency improvements, then the model might show that long-term transformation of energy demand is possible. In any case, it would be possible to explore some of the areas that now separate top-down and bottom-up modelling work.

This prospect raises the question of whether the differences between top-down and bottom-up analyses result from different assumptions about costs, rates of growth and technology performance, or from the fact that the analysts are asking different questions, based on the different perceptions of reality. These perceptions lead to assumptions regarding which of the values used to describe the energy economy 1) cannot change (fixed constants), 2) can change (parametric variables), and 3) can be changed (decision variables) (Wilson and Swisher, 1993).

A model with the characteristics described above could be developed from either a top-down or bottom-up perspective. It would explicitly address the differences between these three types of parameters, and it might make it possible to reconcile bottom-up and top-down models by adjusting the assumptions regarding such parameters. To the extent that differences between models can be reduced to commonly-used input assumptions, the debate among different groups of modellers would become more constructive. At present, differences in results regarding energy-system performance and costs on the supply-side are mostly attributable to assumptions, rather than model structure (Edmonds et al., 1994). However, demand-side differences are model-based and appear to be irreconcilable unless the type of modifications proposed above can be introduced.

As noted above, allowing non-equilibrium solutions in macroeconomic models can create problems of multiple solutions, where choices of the best or most likely solutions are path-dependent, i.e., partly determined by historical precedent, whether efficient or not. Although inconvenient to the modeller, multiple path-dependent solutions are more realistic than single optima. Moreover, path-dependent solutions can lead to positive feedback, in which early choices are magnified in a way that strongly influences the later outcome, regardless of the merits of other potential outcomes (Arthur, 1990). Thus, policy decisions that change the course of technical development might have minor effects or they might lead to a bifurcation, i.e., a permanent divergence between multiple equal but exclusive paths. Such a bifurcation might appear if it is possible to use both fiscal and non-fiscal policy instruments to establish a long-term transformation of energy demand, for example by catalysing future technological advances.

12 Prospects in developing countries

Policy mechanisms to increase energy-efficiency investments, including DSM, have begun to be implemented in developing countries and the formerly planned economies of Eastern Europe and the former Soviet Union. From a technological perspective, these countries are generally less energy-efficient than Western industrialised countries and therefore present great opportunities for efficiency improvement. Although developing countries use much less energy per capita than industrialised countries, they can still benefit from the economic savings and environmental advantages of energy-efficiency by following a course of development that adopts efficient technology in the early stages. This approach would avoid repeating the industrialised countries' energy-intensive and polluting development path, and it would capture energy-efficiency opportunities at the least expensive point, when new equipment and facilities are first put into service.

Unfortunately, all of the barriers to energy efficiency discussed earlier apply to developing countries, and some additional barriers appear. Energy prices are often highly subsidised, industries are often (as in former planned economies) not subject to competitive pressure to reduce costs or introduce new products, the availability and flow of information is restricted, and norms and regulations are neither transparent nor widely enforced. As a result, price reform and information programs are insufficient to stimulate investments in energy efficiency, and regulatory measures such as performance standards are often ineffective.

Despite these difficulties, DSM and other innovative approaches to energy efficiency have appeared in some developing countries and formerly planned economies. Four energy efficiency centres in Poland, the Czech and Slovak republics and Russia were opened in 1991, and recently a similar centre opened in China. These centres have encouraged the adoption of least-cost energy planning to include energy-efficiency measures, developed information programs and energy-efficiency legislation, and catalysed private-sector investments and joint ventures in energy-efficiency projects. Other successful energy efficiency centres have been set up in Pakistan and Korea (World Bank, 1993).

The Electricity Generating Authority of Thailand (EGAT), which is having difficulty keeping up with its customers' explosive demand growth, has committed to a five-year DSM program, partly financed by the Global Environment Facility (GEF). The program is expected to reduce peak demand by 160 MW and serve as a pilot for more ambitious DSM implementation in the future (IIEC, 1991). In Brazil, the government and national electric supply utility Eletrobras began a national electricity efficiency program, Procel, in 1985. During its initial years, Procel was involved with research projects, demonstrations, information programs and some direct installation of energy-saving measures such as efficient street lamps (Geller, 1990). More recently, some Brazilian utilities have begun to finance energy savings in DSM programs, such as a major lighting energy efficiency program in São Paulo.

Despite the barriers noted above with regard to energy efficiency in developing countries, one should not disregard the potentially positive role of the government in providing information and incentives for energy efficiency. In particular, it is important to develop programs to work with manufacturers and importers at the upstream end of the equipment-supply chain and to encourage the introduction and promotion of energy-efficient products (World Bank, 1992). In addition, there is still a need for public programs to stimulate the development of ESCOs and investment in DSM by electric suppliers. Where public sector involvement remains, simple rules are needed to allow utility DSM cost recovery. In more deregulated environments, incentives are needed at the national level to encourage private-sector investment in DSM, not just in new combustion turbines. As marginal supply costs increase and the fragmented power industry begins to re-integrate, DSM options, if they are credible and easily available, may become more attractive than some supply-side options.

13 Recommendations for abatement program implementation and financing

It seems clear that regulatory and mixed policy options such as DSM can be effective in reducing energy consumption and carbon emissions. Moreover, it is possible that such policy measures could stimulate accelerated technical development to make further reductions larger and less expensive. However, most of this discussion focuses on experiences and prospects in industrialised countries. The limited efforts to date in developing countries suggest that, just as technical efficiency tends to be lower, it is also more difficult to apply these policy instruments. Many existing barriers to energy-efficiency investments are likely to hinder the implementation of DSM and other regulatory or incentive measures. In many cases, the needed capacity to implement such measures simply does not exist in government or utility institutions. This capacity, however, can be developed through training and investment.

Some investments can contribute indirectly to energy savings and emission abatement through overall economic efficiency gains, and must therefore be made in advance of investments directly aimed at emission abatement. In addition, some aspects of the implementation process, even for cost-effective measures such as energy-efficiency improvements, will require investments in information programs, demonstration projects, testing facilities, training and outreach, and general capacity building. The non-fiscal policies discussed here rely on exactly these types of capabilities. Although the net cost of a given level of resulting energy savings and emission abatement may be small or negative, the necessary implementation costs appear as positive cost barriers, because it is difficult to allocate these costs to individual measures and to the stake-holders in those measures.

Thus, overcoming the implementation barriers to energy-efficiency improvements and other potential low-cost options requires institutional changes and involves transaction costs. Such implementation issues and costs have generally not be treated explicitly in national costing studies (UNEP, 1994). Nor is it yet possible to model synergistic benefits from introducing information, technological learning, removing risk through comprehensive regulations and incentives, and reduced emissions of non-GHG local pollution.

If developing countries are to carry out the types of energy-efficiency and other abatement measures already identified as cost-effective, however, such implementation costs and external benefits will become important considerations in their strategy. Investments in institutional capacity-building and other implementation steps will facilitate additional investments in abatement measures that would be considered cost-effective. Thus, the initial investment elements should be considered incremental and necessary for protecting the global environment and therefore eligible for international funding from, for example, the GEF. This kind of investment would also provide a role for bilateral development assistance in the implementation of carbon emission reduction measures.


Arthur, W.B., 1990. "Positive Feedbacks in the Economy," Scientific American, pp. 80-85, February.

Atkinson, B.A., et al, 1992, "Analysis of Federal Policy Options for Improving U.S. Lighting Energy Efficiency: Commercial and Residential Buildings," Energy and Environmental Division Lawrence Berkeley Laboratory, December.

Bauer D., and J. Eto, 1992. "Future Directions: Integrated Resource Planning," ACEEE Summer Study on Energy Efficiency in Buildings, Monterrey, CA, September.

Baumgartner, T., and Midttun, A., 1987. The Politics of Energy Forecasting, New York, Oxford Press.

Baumol, W.J., and W.E. Oates, 1988. The Theory of Environmental Policy, Second Edition, New York, Cambridge University Press (first edition, 1975).

Berry, L., 1989. "The Administrative Costs of Energy Conservation Programs," ORNL/CON-294, Oak Ridge National Laboratory.

Bjorkqvist, O.and Wene, C.O., 1993. "Transaction Costs for Energy Efficiency Investments in the Residential Sector," Proceedings of the Second International Energy Efficiency and DSM Conference, Stockholm, September, pp. 315-321.

Bohi, D., and Zimmerman, M., 1984. "An Update on Econometric Studies of Energy Demand Behavior," Annual Reviews of Energy, vol. 9.

Bülow-Hübe, H., Ankarljung, T., 1991, "A Programme Targeted at More Permanent Conversation to Compact Fluorescent Lamps in the Residential Sector," Proceedings of the 1st European Conference on Energy-Efficient Lighting, Stockholm, May.

Canadian Electrical Association (CEA), 1992. Demand-Side Management in Canada, Montreal.

Coase, R.H., 1960. "The Problem of Social Cost," Journal of Law and Economics, vol. 3, pp. 1-44.

Dandridge, C., et al, 1993. "Trends in Energy Efficiency in Office Technology," Proceedings of the 1993 ECEEE Summer Study: The Energy Efficiency Challenge for Europe, Denmark, June.

Davies, P., 1989. The New Physics, Cambridge Univ. Press.

Deming W.E., 1982. Quality, Productivity and Competitive Position, Cambridge, MIT Press, Cambridge.

Edmonds, J. M. Wise, C. MacCracken, 1994. "Advanced Energy Technologies and Climate Change," Battelle Pacific Northwest Laboratory, Washington.

Farstad, H. and Ward, J., "Wind Energy in Denmark," in T. Baumgartner and T.R. Burns (eds), Transitions to Alternative Energy Systems, Westview Press, Boulder, CO, 1984.

Fisher A.C. and Rothkopf M.H., 1989. "Market Failure and Energy Policy," Energy Policy, pp. 397-406, August.

Fritz, P., 1990, Effektivare Elanvändning, Statens Energiverk, Stockholm, June.

Gaskins D., and J. Weyant, 1993. "Tentative Conclusions Energy Modeling Forum Study Number 12 on Controlling Greenhouse Gas Emissions," in Kaya et al, eds., Costs, Impacts and Benefits of CO2 Mitigation, International Institute for Applied Systems Analysis, Laxenburg, Austria.

Geller, H.S., 1990. Efficient Electricity Use, A Development Strategy for Brazil, American Council for Energy-Efficient Economy, Washington.

Geller, H. and Nadel, S., 1994. "Market Transformation Strategies to Promote End-Use Efficiency," American Council for Energy-Efficient Economy, Washington.

Gellings, C., 1985. "The Concept of Demand-Side Management for Electric Utilities," Proceedings of the IEEE, October.

Goeltz, R., and Hirst, E., 1986. "Residential Retrofit Measures in the Hood River Conservation Project," Oak Ridge National Laboratory, ORNL/CON-208.

Goldman, C. and E. Kahn, 1993. "Integrated Resource Planning: the Role of Competitive Forces," Proceedings of the ECEEE Summer Study: The Energy Efficiency Challenge for Europe, pp. 33-44, Copenhagen.

Greene, D.L., 1990. "Cafe or Price?:An Analysis of the Effects of Federal Fuel Economy Regulations and Gasoline Price on New Car MPG, 1978-89", The Energy Journal, vol. 11, pp. 37-57.

Gross W.A., 1990. "Six Dimensions of Successful Technologies," Proceedings: American Solar Energy Societ Conference, Albuquerque, NM.

Hedenström, C., 1991, Från Krog till Kontor: Slutrapport från Uppdrag 2000-lokaler, Vattenfall, Vällingby, Sweden.

Hedenström, C., Göransson, A., Lindahl, U., Forsman, G., 1992. Lokalerna och Hushållningen, STIL-studien, Vattenfall, Vällingby, Sweden.

Hirst, E., and J. Reed, 1991. Handbook of Evaluation of Utility DSM Programs, ORNL/CON-336, Oak Ridge, Tennessee, December.

Hogan, W., and D. Jorgenson, 1991. "Productivity Trends and the Cost of Reducing CO2 Emissions," Energy Journal, vol. 12, pp. 67-86.

Hogan, W., and A. Manne, 1979. "Energy-Economy Interactions," Advances in the Economics of Energy and Resources, vol. 1.

Intergovernmental Panel on Climate Change (IPCC) Working Group 1, 1990. Scientific Assessment of Climate Change, Geneva, U.N. Environment Programme, June.

International Energy Agency (IEA), 1987. Energy Conservation in IEA Countries, OECD, Paris.

International Institute for Energy Conservation (IIEC), 1991. "Demand-Side Management: Master Plan for Thailand's Electric Power System," IIEC, Bangkok.

Janda, K. and J. Busch, 1994. "Worldwide Status of Energy Standards for Buildings," Energy: the International Journal, vol. 19, pp. 27-44.

Jochem, E. and E. Gruber, 1990. "Obstacles to rational electricity use and measures to alleviate them," Energy Policy, pp. 340-350, May.

Johansson, T.B. and Swisher, J.N., 1993. "Perspectives on Bottom-Up Analysis of Costs of Carbon Dioxide Emission Reductions," OECD International Conference on the Economics of Climate Change, Paris, June.

Joskow, P. and Marron, D., 1993. "What Does a Negawatt Really Cost? Evidence from Utility Conservation Programs," The Energy Journal, Vol. 13, No. 4.

Krause, F., E. Haites, R. Howarth, J. Koomey, 1993. Cutting Carbon Emissions: Burden or Benefit, International Project for Sustainable Energy Paths, El Cerrito, Calif.

Larsen, A., Nielsen, L., Jørgensen K., Jordal-Jørgensen, J., 1993. Virkemidler og Elbesparelser, AKF, Copenhagen, March.

Levin Kruse, K., 1991. "Från Mätare till Människa," Vattenfall, Vällingby Sweden, November.

Levine, M., E. Hirst, J. Koomey, J. McMahon, A. Sanstad, 1994. "Energy Efficiency, Market Failures, and Government Policy," Lawrence Berkeley Laboratory, March.

Lewald, A., Bowie, R., 1993. "What Is Happening with the Swedish Technology Procurement Program," Proceedings of the 1993 ECEEE Summer Study: The Energy Efficiency Challenge for Europe, Denmark, June.

Lovins, A.B., and H.L. Lovins, H.L., 1991. "Least-Cost Climatic Stabilization," Annual Review of Energy and the Environment, vol. 16,.

Manne, A., 1992. "Global 2100: Alternative Scenarios for Reducing Carbon Emissions," OECD Economics Dept. Working Paper 111.

McMahon, J., D. Berman, P. Chan, J. Koomey, M. Levine, S. Stoft, 1990. "Impacts of U.S. Appliance Efficiency Standards on Consumers, Manufacturers, Electric Utilities and the Environment," ACEEE Summer Study on Energy Efficiency in Buildings, Monterrey, CA, September.

Meier, A., et al, 1983. Supplying Energy Through Greater Efficiency, University of California Press, Berkeley.

Miller, M., Blake, W., Dagher, S. and Schutte, J., 1992. "Development of Measured Savings for a Small Commercial/Industrial Lighting Program,"ACEEE Summer Study on Energy Efficiency in Buildings, Monterrey, CA.

Nadel, S., 1992. "Utility Demand-Side Management Experience and Potential - A Critical Review," Annual Review of Energy and the Environment, Vol. 17,pp. 507-535.

Nadel, S., Wolcott, D., Smith, P., Flaim. T., 1990. " A Review of Utility Experience with Conservation and Load Management Programs for Commercial and Industrial Customers," New York State Energy Research and Development Authority, Albany, NY, April.

Nadel, S., Atkinson, B., McMahon, J., 1993. "A Review of US and Canadian Lighting Programs," Energy: the International Journal, vol. 18, no. 2, pp. 145-158.

Nelson, D. and M. Ternes, 1993. "Flipping the Industrial Market: The Move to High Efficiency Motors," Proceedings: 6th National Demand-Side Management Conference, EPRI TR-102021, Electric Power Research Institute.

Nielsen, L., Jørgensen,K.and Jordal-Jørgensen, J., 1992. "Elbesparelser i Boliger," AKF, Copenhagen, August.

Nilsson, H., 1992. "Market Transformation by Technology Procurement and Demonstration," ACEEE Summer Study on Energy Efficiency in Buildings, Monterrey, CA, September.

Northwest Power Planning Council (NWPPC), 1991. "Northwest Conservation and Electric Power Plan," NWPPC, Portland, Oregon.

NUTEK, Kansliet för effektivare energianvändning, 1992. "Eleffektiv Disk-Tvätt-Kyl-Frys," Stockholm, June.

Orans, R., Woo, C.K., Swisher, J.N., Wiersma, B., Horji, B., 1992. "Targeting DSM for Transmission and Distribution Benefits," Electric Power Research Institute, EPRI TR-100487, May.

Paine A.,1921. In One Man's Life, Harper, New York.

Pedersen, P.B., 1992."Engineering Analysis Concerning Energy Efficiency Standards for Refrigerators and Freezers in Denmark," dk-TEKNIK, Copenhagen, February.

Pigou, A.C., 1920. "The Economics of Welfare," London, Macmillan.

Reddy, A.K.N., 1991. "Barriers to improvements in energy efficiency," Energy Policy, December , pp. 953-961.

Rosenberg N., 1969. The American System of Manufacturing, Edinburgh Univ. Press, Edinburgh.

Ross, M., 1989. "Energy and Transportation in the United States," Ann. Rev. of Energy, 14: 131-172.

Ruderman, H., M. Levine, J. McMahon, 1987. "The Behavior of the Market for Energy Efficiency in Residential Appliances Including Heating and Cooling Equipment," The Energy Journal, vol. 8, no. 1, pp. 101-124.

Schipper, L., and F.X Johnson, 1993. Energy Use in Sweden: An International Perspective, LBL-33819, Lawrence Berkeley Laboratory.

Sharpe, V., D'Angelo, P., Ruhnke, W., Hydro, O., 1993. "Working With Customers: The Espanola Experience," Proceedings of the Second International Energy Efficiency and DSM Conference, Stockholm, September, pp.339-348.

Siniscalco, D., 1993. "Environmental Policy and Technological Innovation," Fondazione Eni Enrico Mattei, Milan.

Sutherland, R., 1991. "Market Barriers to Energy-Efficiency Investments," The Energy Journal, vol. 12, no. 3, pp. 15-34.

Swisher, J., 1994."Dynamics of Appliance Energy Efficiency in Sweden," Energy: the International Journal, vol. 19, no. 11, pp. 1131-1141.

Swisher, J., Hedenström. C., 1993, "The Role of Energy Efficiency for Utility Planning in a Deregulated Market," Proceedings of the Second International Energy Efficiency and DSM Conference, Stockholm, September

Swisher, J., L. Christiansson and C. Hedenström, 1994. "Dynamics of Energy-Efficient Lighting", Energy Policy, vol. 22, no. 7, pp.581-594, July.

Tureil, I., 1993. "Personal communication," Lawrence Berkeley Laboratory, June.

UNEP 1994. UNEP Greenhouse Gas Abatement Costing Studies. Phase Two Report. UNEP Collaborating Centre on Energy and Environment, Risø, Denmark.

U.S. National Research Council, 1992. Global Environmental Change: Understanding the Human Dimensions, National Academy Press, Washington.

U.S. Office of Technology Assessment (US OTA), 1992. Fueling Development: Energy Technologies for Developing Countries, OTA-E-516, US Govt. Printing Office, Washington.

Velthuijsen, J.W., 1993. "Incentives for Investments in Energy Efficiency: An Econometric Evaluation and Policy Implications," Environmental and Resource Economics, vol. 3, pp. 163-169.

Westling, H., 1991. Technology Procurement for Innovation in Swedish Construction, Byggforskningsrådet, Stockholm.

Wilson, D., and Swisher, J., 1993. Exploring the Gap: Top-down versus bottom-up analyses of the cost of mitigating global warming, Energy Policy, vol 21, no. 3, pp. 249-263, March.

World Bank, 1992. Financing Energy Services for Small-Scale Energy Users, Asian and Pacific Development Centre, Kuala Lumpur, Malaysia.

World Bank, 1993. "Energy Efficiency and Conservation in the Developing World," World Bank, Washington.

Back to Contents