Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CHAPTER 5 HOME ENERGY RATINGS Building technology is too complex for the average home purchaser to be able to make an informed decision about the energy costs of operating different homes. Therefore, home energy ratings have been advocated to help consumers, as well as builders and lenders, take energy costs more fully into account in their decisions. This com- mittee agrees. After careful study of the factors affecting energy-related decisions by individuals and organizations, we recommended that "federal or private agencies . . . develop simple, understandable indices of energy efficiency, comparable to miles-per- gallon, for . . . building shells" and that "labeling, rating, and certification programs should be supported to ensure that indices of energy efficiency come into common use" {Stern and Aronson, 1984: 197,1981. This chapter addresses some of the problems involved in developing such rating programs and getting them used. Research on energy ratings has so far focused mainly on developing ratings that meet three major technical criteria: reliability, validity, and accuracy. A reliable rating system is one that gives the same building the same rating regardless of who does the rating. A valid rating system is one in which the higher a building's rating for energy efficiency, the less energy it in fact uses to provide given levels of heat and cooling (holding constant climate, occupant behavior, and other external factors). An accurate rating system is one in which a building's rating is borne out by measured energy use: if a building's energy use is rated to be double that of another building or double what it would be after retrofit, that predicted use will be the same as the measured energy use (when external factors are held constant). However, reliability, validity, and accuracy do not necessarily add up to success for an energy rating system. In fact, the chief barriers to success at present are not technical. According to a recent study of six existing rating systems by the Consumer Energy Council of America (McCarty and Willner, 1985), the most significant barriers they face are funding, inadequate involvement among some sponsor groups, and low levels of consumer awareness. It is essential to keep in mind that the goal of a rating is to affect the decisions of home purchasers, builders, real estate agents, primary and secondary mortgage lenders, and other intended audiences so that energy efficiency is reflected in 64
65 a home's market value or so that new and retrofitted homes reach higher standards of energy efficiency. Technical adequacy is important, but it is equally important that a rating system be understandable to its intended users, available when needed for decision making, and credible. This chapter addresses three of the five major behavioral questions that arise in informational programs (see Chapter 1), focusing primarily on issues of designing programs to diffuse information and get it used and also on the issue of program evaluation. It is premature to address the remaining questions--about forecasting the effects of programs and about attributing these effects to one cause or another--until careful evaluations have been done. There is little direct knowledge about how design and implementation aspects of home energy rating systems affect whether a system is used. Over 40 home energy rating or certification systems of various types were in operation in the United States by 1982 (Hendrickson, Garrett- Price, and Williams, 1982~; however, few if any of these were designed on the basis of systematic analysis, and there has been little or no careful evaluation of their effects. Since 1982 the situation has improved only a little, but enough is known about the context of energy ratings to identify the critical questions in designing and implementing energy ratings, make some educated guesses about possible answers, and offer ways to check the accuracy of the guesses. This chapter begins by discussing the characteristics of a home energy rating system that would make it ideal for its various users. It then discusses some major issues in designing and implementing home energy ratings and offers suggestions for research and programs that can move home energy ratings toward wider use. Design questions concern characteristics of the ratings themselves that can make them more or less understandable, interesting, meaningful, or relevant to potential users. Implementation questions concern ways of delivering ratings to their potential users. Although it-may seem to make sense to think about implementing a program only after its design is clear, it makes as much sense to reverse the order: a rating system designed so that it is irrelevant to the organizations that will implement it is destined for failure. CHARACTERISTICS OF AN IDEAL HOME ENERGY RATING SYSTEM Home energy rating systems can have many users and many uses. Thus, an ideal rating system must be intelligible to home builders, primary and secondary mortgage lenders, appraisers, real estate agents, retrofit contractors, and the buyers, sellers, and occupants of homes. It must be relevant to decisions about building design, financing, retrofitting, advertising, and purchase. And it must be applicable to both new and existing homes. Although it is probably impossible to design and implement a system with all these characteristics, we offer as a guide to what may be sacrificed in terms of a program's acceptance and usefulness as a result of technical decisions. Based on our earlier work on the way individuals and organizations use energy information, we offer the . .
66 following list of character istics of a home energy rating system the t would be ideal from the users ' perspectives: (1) it is simple; (2) it is expressed in familiar units; {3 ~ it it easy for the user to verify; (4) it is readily translatable into cost; (5) z~ can provide meaningful regular (e.g., monthly) feedback to indicate changes in the energy intensity of the home; and (6) it allows the user to examine either the efficiency of the building's construction, the effects of occupant behavior, or both, as desired. The last two characteristics are relevant only if the system will be used by building occupants; they are not necessary in a rating designed only to influence purchases. The obverse of this point is that a rating system that is not designed to give feedback or take behavior into' account is much less useful as a stimulus to retrof it existing buildings. This list of ideal characteristics sets the context for our discussion. QUESTIONS ABOUT DESIGNING RATINGS In What Units Should a Rating be Presented? Home energy ratings have been expressed in energy units, in dollars, and on var ious arbitrary scales . Energy Units From a technical s tandpoint, energy units have a clear advantage: energy use is what is to be rated. They also have a disadvantage: a difficult decision must be made because different fuels are measured in different units and because there might be debate about what constant should be used to compare electricity with fossil fuels. For example, if all homes are rated for energy use in kilowatt-hour equivalents, the ratings will confuse many purchasers of gas- or oil-heated homes. Behaviorally, there is no advantage to using energy units: people are not likely to understand them. People understand the meaning of miles-per-gallon because both miles and gallons are familiar units and because a mile of travel is an understandable index of what gasoline produces for consumers. But the units of measure for household energy are not so intuitively meaningful. As a result, people generally think of household energy use in dollars per month, weeks between oil deliveries, or other budget-based units (Kempton and Montgomery, 1982)e There is no evidence that householders readily understand information presented in~Btus per degree-day, therms per winter, or other physical units, and little reason to believe that bankers or builders are much different. Home energy ratings presented in energy units are more likely to confuse people than enlighten them.
67 Dollars Rating homes in dollars of energy cost, or dollars per square foot, has the advantages of familiarity, relevance to purchase decisions, and comparability across homes. A representative of the Federal National Mortgage Association (Fannie Mae) told Pennsylvania researchers that for the needs of lenders, it was important to have estimates of costs of energy conservation investments and savings in terms of a home's operating costs (Gallagher and Desmond, 1984~. Such information is relevant to decisions about whether to raise a borrower's debt limit to allow a mortgage on an energy-efficient home. Unfortunately, however, there are technical problems with ratings in dollars. A determination must be made of whether a rating should use current energy prices or expected future prices, and if the latter, what estimates should be used. For example, if a mortgage lender takes future energy costs into account in setting debt limits, the viability of the mortgages will be affected by how well the energy rating system anticipates the price of fuel. Also, a meaningful dollar rating must assume constant climate and occupant behavior. Such ratings may mislead householders who expect to pay the rated energy cost in the first winter of occupancy. Users would have to be educated about what they can and cannot expect from a dollar rating. Arbitrary Scales Ratings can be given on a binary scale {e.g., pass-fail) or in other arbitrary scales of varying sophistication. An advantage of arbitrary ratings is that, unlike dollar ratings, they are unaffected by changes in fuel prices, weather, and occupant behavior. But their arbitrariness can be a disadvantage unless an effort is made to give them meaning, possibly by "anchoring" them to certain meaningful values. We know of only one empirical effort to examine user response to alternative units for use in home energy ratings. In a pilot project in Massachusetts, a home rating system was developed for use in the RCS energy audit program (Ackerman et al., 1983~. The researchers discussed the concept of ratings with selected lenders, appraisers, and real estate agents, both before and after developing eight possible formats for a rating system. The formats were also the subject of two focused group discussions among local homeowners. The system chosen used a 0-10 scale in which zero represented a home with no energy-saving features and 10 represented a home with no energy bills (Ackerman et al., 1983~. All the groups, as well as the organization doing the rating, found the "anchored" 0-10 scale, combined with estimated annual energy costs, understandable and acceptable. The Massachusetts 0-10 system was constructed with users in mind and was accepted by users in early field testing. For this reason, it is worthy of further test and adaptation. Other rating scales may prove equally acceptable. One of these is a five-star rating system developed by Western Resources Institute (Luboff, 1983), which also arose from an effort to seek consensus among segments of the building
68 industry. Both these approaches, which have the advantage of sim- plicity, have had good user response in f ield tr ials, and both deserve systematic evaluatic~n under f ield conditions . Groups interested in developing new rating systems can learn from the methods used to develop them even if they do not choose to use either of them. How Much Precision Should a Home Energy Rating Offer? Is it better to use a binary (certification) scale or one with more levels presented? If the latter, how f ine should the detail be? One principle to keep in mind is that a rating should not offer precision that exceeds its accuracy. If a rating is only accurate within 10 f inures can undermine its _ percent, presenting it to two significant __, credibility; if it is accurate to within 20 percent, it should not be presented as anything more precise than a f ive-point scale ~ A binary scale (a pass/fail or certification approach), such as utilities sometimes use when they certify Energy-efficient homes, n has the advantage of s implicity. Certif icates are meaningful to builders, realtors, and others who can use the rating as an advertising point. They may be more useful in the new home industry than in influencing the home resale market. The ability of users to take advantage of a more detailed rating system depends on whether they can relate different meanings to is, , ~ ~ tenant ratings. In a category system (e.g.' poor, good, very goon, excellent ~ the number of categor ies should be kept small, probably not more than five. This sort of rating probably has meaning for comparing homes, but it is likely to be less useful for deciding whether to invest in a retrofit because the categories are so coarse. Numerical ratings add further detail, but can be confusing if the meaning is unclear. Ratings in dollars are easy for people to inter- pret, although they can lead to unrealistic expectations of predictive accuracy because homeowners will compare ratings with actual energy bills, which are affected by weather and occupant behavior. Arbitrary scales promise less, but tend to be less meaningful. The notion of anchoring them to some understandable values, as done in the Massachusetts project, seems wise. Decisions about the precision of a rating system should depend on its goals and on what is meaningful to the intended users. Binary scales may be appropriate for certifying new buildings because the dividing line can be set slightly above ordinary building practice to provide an incentive for improving energy efficiency. More detail is appropriate, however, when a program aims to influence retrofits of existing houses before resale. There are usually many effective retrofits that can be made in an old home that would still leave it below an energy-efficiency criterion set for new buildings. Arbitrary numerical scales should probably be anchored to points that have some meaning for all the intended users. But these conclusions are all tentative . Prototype scales should be presented to samples of the use r populations to get their reactions before a rating system in put into use. l
69 Should a Rating Explicitly Estimate the Effects of Retrofits? The Massachusetts home energy rating system experiment gave each home two ratings: as audited and as they would be after taking the energy efficiency measures recommended by RCS. It also offered estimates of average annual energy costs with and without the retrofits. Because this approach makes it easy for the seller or buyer of an existing home to judge the likely cost and savings from recommended retrofits, it makes good sense when a rating system is intended to encourage retrofits of existing housing before resale and to increase the market value of energy-efficient homes. To remain credible, though, such estimates must be accurate, which requires a full energy audit of the home. Although such an audit normally costs $100-150, an estimate of the effect of retrofit can be done for an additional $15 per home when a rating is attached to an RCS energy audit (McCarty and Willner, 1985), and for $25-35 when included in a mortgage lender's appraisal process {Hoskin, 1983~. Accurate estimates from ratings can be useful, but the additional information may confuse people who have no interest in undertaking retrofits. This possibility should be addressed empirically in pilot projects with rating systems. What Energy Uses Should a Rating Reflect? Ratings may attempt to estimate only energy use for heating, or energy for heating and cooling, or a larger proportion of home energy use. It is reasonable to expect that the interest of potential users in one or another package of information depends on climate, but there are no relevant data on this point. As more energy uses and more discretionary appliances are subsumed in an energy rating, it becomes more difficult to produce an accurate rating. The potential for confusion among users probably also increases, especially when ratings are used to compare homes with different appliances included. For example, a home with a central heat pump air conditioner may use more energy than a home that is only partly cooled by room air conditioners, even though it uses the energy more efficiently. It will probably be difficult to express these differences simply and clearly in a rating. If an energy rating is to encompass more than space and water heating, careful thought, technical research, and assessment of user reactions should precede introduction of the rating. QUESTIONS ABOUT IMPLEMENTING RATING PROGRAMS Technically adequate ratings have no effect unless they are delivered to users in a way that encourages their use in decision making. Careful design of ratings is necessary, but some institutional questions must also be addressed. 1'
70 Who Should Rate Homes? Experience with other energy programs suggest that a rating system will be implemented best when the organization responsible can do the job at low cost; has the appropriate knowledge, sufficient resources, and the motivation to do the job well; and is credible to the user communities. These considerations do not point unequivocally to a single most appropriate institution, but they do suggest some approaches to try in pilot projects. Cost considerations suggest that ratings be implemented as part of other activities, such as those of home energy auditors, real estate appraisers, or ''house doctors." Since the groups that carry out such activities vary in knowledge and resources, some could provide good energy raters only if a reliable, valid, accurate, and sufficiently simple rating system were developed by an expert group for use in the routine of energy auditing or appraisal. For example, in the Western Resources Institute's program (Luboff, 1983), trained real estate appraisers use a heat loss methodology already developed by the Bonneville Power Administration. To ensure credibility, two strategies seem promising: to have the rating conducted by a credible institution or to get the rating system accepted by a credible institution. It is possible to tell which organizations would be credible sponsors by asking for the reactions of potential users to a list of possible rating organizations. But a more promising strategy involves building credibility into the system. The Western Resources Institute's rating system seems to have accomplished this by convincing local mortgage lenders to accept Bonneville Power's heat loss methodology. The banks Day their appraisers to calculate _ _ _ _ _ energy ratings using the methodology, and they otter larger mortgage loans for purchasers of homes that receive high enough ratings. The banks' acceptance creates instant credibility for the system among home purchasers, and people considering selling their homes are likely to feel an incentive to retrofit because an energy-efficient home will be financially attractive to a larger number of buyers. A sponsor's motivation should also be considered. House doctors or other purveyors of retrofit services have a vested interest in making a house look more energy efficient after they perform their service than before, so their ratings may be suspect. Real estate appraisers are motivated to accurately assess market prices, so may be good candidates if a rating system can be made convenient for them. However, unless energy ratings are used by banks in their lending decisions or appraisers conclude that energy-efficient homes are worth more, appraisers may be reluctant to conduct energy ratings. Home energy auditors can rate homes conveniently. Their motivation to rate accurately depends on the context of the audit. Auditors for a utility company that is motivated to conserve energy, whether because of a prospect of supply shortage or because of regulatory pressure, would have a motive to produce accurate ratings. Utilities that are not motivated to conserve energy would not be motivated to document the need for it. An electric utility with a large overcapacity would be unlikely to offer ratings, and if it did, its ratings might not be
71 credible. Local nonprofit groups have sometimes produced more thorough energy audits than utilities at lower cost (Polich, 1984), probably reflecting in part a stronger motivation to cut energy costs for consumers. Such groups might be effective raters if a rating system were designed to be attached to energy audits. The problem of who should rate homes becomes most difficult when the organizations that have the motivation and credibility lack resources or expertise. In some areas, utilities have low motivation or low credibility, but more credible and motivated groups, such as community organizations or local government, lack resources and expertise. The RCS experience shows that collaborations that link the resources of utilities or state government with the motivation and credibility of local organizations can be effective (Cowell and Rebitzer, 1984; Polich, 1984; Stern and Aronson, 1984~. What are the Key Institutions for Getting a Rating System Accepted? Home energy rating systems can become effective in various ways: they can be used directly by home purchasers in making decisions; homeowners, developers, and real estate agents can use them to help advertise energy-efficient homes; they can add value to energy- efficient homes if used by appraisers; they can make homes salable at higher prices if bankers use them in calculating total cost of owner- ship and in setting mortgage limits; and operators of home retrofit programs might use them to market their services. In short, rating systems operate in a complex institutional environment, in which it is not immediately clear whether any particular institutionts acceptance is the key. Institutional acceptance, however is a central question and a major barrier to the acceptance of rating systems (McCarty and Willner, 1985~. We are aware of only two studies that have directly asked for reactions to energy ratings from major institutions in the building industry. A study in Massachusetts involved discussions with primary and secondary mortgage lenders, appraisers, real estate agents, utility representa- tives, and homeowners before and after implementation of a pilot home energy rating system (Ackerman et al., 1983~. A more recent study in Pennsylvania surveyed homeowners and had discussions with the major industry groups before implementing a "home energy scorecard" (Gallagher and Desmond, 1984~. This limited evidence and the experience of the Western Resources Institute rating system suggest that secondary mortgage lenders are a keystone of the complex institutional arrangement. Several banks were willing to use ratings in mortgage decisions only if such action would not jeopardize the sale of the mortgage to a secondary lender. Thus, some would raise debt limits to qualify buyers for mortgages on energy- efficient homes only if secondary lenders would accept that judgment in repurchasing the mortgages. It proved important in Massachusetts that secondary lenders were willing to change a previous policy and allow retrofit loans to be included in first mortgages before a retrofit is complete. In the Western Resources Institute program, acceptance of a
72 particular rating system by secondary lenders was essential to primary lenders ' acceptance of the ratings. Once the primary lenders accepted the rating system it was a fairly easy step to include a rating in the appraisal process. The involvement of secondary mortgage lenders can create incentives for other actors in the housing market and can make home energy ratings work without any effort by home buyers. The Massachusetts project was approached by a nonprofit building corporation to rate some new homes being constructed for low-income buyers; the builder had learned that the local bank would allow higher debt-to-income ratios for borrowers if the homes had high ratings. Thus, an energy-efficient home would become affordable for people who otherwise would have remained in less- efficient, older housing. In this way the energy efficiency of the housing stock improves without the home purchaser needing to consider energy costs explicitly. Real estate agents have been reluctant to take the lead on home energy ratings. In Pennsylvania, where a state survey found that 85 percent of residents believed that energy efficiency would be an impor- tant consideration in selecting their next home and 90 percent felt a home energy labeling program would encourage people to select energy efficient homes, real estate brokers believed that the customers did not want ratings (Gallagher and Desmond, 1984~. In Massachusetts, realtors raised concerns about their possible liability for information in the ratings. The Massachusetts realtors wanted to follow the lenders. Even after the pilot program began, they wanted assurances that the banks were committed to it. Real estate appraisers often follow the market as they see it. In Massachusetts, some appraisers did not believe that energy efficiency affected home values (Ackerman et al., 1983~. Even those who supported ratings did not want to do the retina themselves because they believed ~ ~ ~ an., ~ ha ~ ~ ~ ~ _ _ _ _ ~ Allay ~~U1Q llU ~ ~ MU ~11= ~~ ~ . In Pennsylvania two years later/ appraisers were very positive about rating systems: many reported that they were being asked to appraise the value of energy conservation features of homes and were alad to be offered a ready ~ ~~ ~ tool for doing the job (Gallagher and Desmond, 1984~. The experience of Western Resources Institute shows that appraisers follow the desires of their customers, the banks, as well as the housing market. Less information is available on the reactions of utility repre- sentatives, trade associations, and builders. In both Massachusetts and Pennsylvania, the utilities were concerned mostly with technical points in the rating calculations. The Pennsylvania report concluded that it was important to involve these groups early because it improved the rating system and gained the acceptance of groups whose opposition might be influential. The importance of possible utility company opposition depends on the implementation of the rating system: opposition is critical if a program is to be implemented by utility-run RCS programs but not if it is part of the home appraisal process.
73 What Other Institutional Arrangements Might Strengthen a Rating System? We have noted that home energy ratings exist in a complex institu- tional environment: this suggests that yet other parts of that environ- ment may make the difference between success and failure for rating systems. For example, a rating system can make local building codes or standards for energy-efficient construction more feasible. If a rating system is incorporated in building codes, it could be implemented and funded by local governments as part of the process of granting building permits. Or, if the accuracy of a rating system is backed by insurance, the organizations responsible for conducting or using the system could stand very firmly behind their ratings. Energy service companies now take out insurance to back their guarantees of low levels of energy use; builders can do the same. The cost of the insurance would depend on the record of the energy service company or the builder in achieving promised savings. A rating system with proven accuracy in the aggregate might lower the cost of the insurance and thereby improve the position of energy service companies in the home retrofit market and of energy- efficient builders in the new home market. DEVELOPING EFFECTIVE HOME ENERGY RATING SYSTEMS It is not possible to offer precise prescriptions for a home energy rating system. Just like a new technology, an energy rating system requires research, development, and demonstration before an effective model is widely accepted. Rating systems can be based on relevant technical and behavioral knowledge and refined by testing successive approaches in the physical and social reality in which they will operate. This point was discussed generically in Chapter 2; this section offers suggestions on which approaches are now ready for testing and on how the testing might be conducted. What to Test Our suggestions are confined to issues of rating design and implementation. In choosing this focus, however, we do not wish to suggest that further technical research and development on ratings is unnecessary nor to overlook the fact that design and implementation issues are intertwined with the technical quality of ratings. Our comments assume that rating procedures exist or will soon be developed that can produce ratings accurate enough to justify a numerical rating of considerably more detail than a five-point category system. The five-star system developed by Western Resources Institute deserves careful and quantitative evaluation as it is now operating and as it might be adopted in additional communities. This system has been among the most successful of existing rating systems and has already had a measurable effect on the approval rates of mortgage loans by
74 banks {Hoskin, 1983~. An evaluation should assess its effects on home sales and home retrofits--even though the program is aimed only at the former--in order to make comparisons possible with other rating systems. Evaluation studies should also be conducted on programs modeled on the Department of Energy's Massachusetts Home Energy Rating System Project (Ackerman et al., 19837. Evaluation of the original program was not carried far enough to assess the effects on retrofits or home sales, and a thorough evaluation of such a project is needed. We believe that the screening of design alternatives that was done in Massachusetts, which selected a 0-10 scale anchored by "no energy improvements" and "no energy bills," has enough validity to serve as a basis for future pilot studies. It is also worth expressing ratings in annual heating costs and these ratings should be tested both with and without a 0-10 rating scale. It is worth testing both the dollar-based scale and the arbitrary scale in home energy audits, with and without estimates of the costs and savings from recommended retrofits. It is advisable to begin with pilot studies in which householders are randomly offered different packages of information from the above alternatives. When the field of possible rating systems is narrowed, field tests should use only one rating design per housing market because using several designs at once will confuse real estate agents, lenders, and others who will see ratings of many different homes. In ratings delivered with energy audits, the sponsor of the audit will make a difference. The choice of sponsor should be made to suit the location of the study and should take into account the credibility, motivation, resources, and expertise of candidate groups. Marketing of rating systems is also critical to implementation. Thus, it is desir- able to fit a rating process into the routine of a local organization that already has a market for its services. Real estate appraisers are one example; a successful energy auditing program is another. How to Conduct a Test One clear lesson from the experience with rating systems and other energy programs to date is that all the relevant actors should be involved from the start. For a home rating system, this may include local officials, bankers, builders, appraisers, real estate agents, utility companies, and home buyers, sellers, and occupants. Even if it seems clear that a rating should have a particular design, it is important to get the reactions of the target groups, which may bring to light important local conditions that escaped the analysts who drafted the project plan. For example, information in a rating system that is valued by people in a cold climate may be seen as incomplete or mislead- ing in a climate where cooling or water heating is the major household energy use. Of course, the most important reason for involving target groups is that the process makes them aware of and committed to the rating system that is chosen. Such involvement is the first step in marketing, which has so far been a serious problem for rating systems (McCarty and Willner, 1985~.
75 Tests of rating systems should be considered part of a learning process. Thus, a rating project should be monitored from the start to assess the reactions of users to the rating design, the conduct of the rating, its marketing, and other details of program implementation. This information can help rating system operators improve their programs and inform the designers of future rating systems. Such process evalua- tions focus on implementation issues and assess both quantitative and qualitative factors. User surveys can determine whether users believe the ratings are done professionally and reliably, whether raters are available on short notice, and so forth. Information from the organiza- tions that operate rating programs can quantify the training given to raters, the amount of time they spend conducting ratings, and the operator's level of marketing effort. Qualitative judgments must also be made about the sponsor's commitment to the program, the rating system's emerging reputation for reliability in the community, the kind and degree of communication about the rating system between key institutions, and so forth. How to Assess Outcomes A successful home energy rating system will take time to show effects on energy use, and the effects will not all be of one kind. Evaluation, therefore, should carefully look for effects that can be expected at the time of assessment. Below we note three kinds of effects to examine, and some examples of each. Early Indicators Initial outcome evaluations might look for such indicators as: (1) Adoption of ratings by mortgage lenders as criteria to qualify buyers for mortgages; (2) Requests for ratings, including information as to whether the requests are coming from developers of new buildings, homeowners planning to sell or to retrofit, real estate agents, appraisers, or others; (3) Mention of ratings in advertising by real estate agents or mortgage lenders; (4) Appearance of ratings on real estate agents' summaries of house characteristics; (5) Reports from real estate agents that home buyers are asking about ratings of homes on the market; (6) Percentage of houses sold that have ratings. Effects on the Sale of Energy-Efficient Homes Once rated homes begin to sell, it becomes possible to assess effects of the rating more directly. One possible effect is that
76 highly rated homes sell sooner, reducing the costs to the seller. Another is the debt/income ratios that banks use to qualify purchasers for mortgage loans. A more important effect, though it is harder to measure, is on sale prices of homes. One way to estimate this effect is to match recently sold rated homes with nearby unrated homes that were also sold recently. This method should include comparisons of house size and other factors affecting value and control these statis- tically through regression or other techniques, matching should be used to control directly for the value of location. Another method for estimating the price effect is econometric modeling of sale prices in a housing market, with the presence of a rating treated as a dummy variable in regression analysis. These methods would not Yield definitive results because they do not assess the possibility that ratings might encourage retrofits or energy-efficient building practices in homes that are not rated. We have discussed the evaluation of this contagion effect in Chapter 4. Time-series data on a given housing market and a comparison market are needed to evaluate that possibility. The most definitive research design is experimental: ratings would be included in an RCS audit or an appraisal process, but would be made available to only a randomly selected half of the relevant homeowners and 1enderse This approach to analysis was tried (but not completed) in the Massachusetts pilot project. Effects on the Energy Efficiency of Homes The new home market may be more easily influenced by ratings because it is easier to achieve impressive savings for new homes than for older homes that are retrofitted and because very simple rating systems can be effective in setting goals for home construction. To assess effects in either the new or existing housing market requires a full energy audit of a sample of rated and unrated homes. To assess contagion to builders and homeowners who do not use ratings, it would be useful to evaluate improvements in energy efficiency in comparison with another area. The creative experimental design tried in the Massachusetts project is especially appropriate for assessing effects on home retrofits. Retrofit activity after the RCS audits in the two sets of homes would give a good index of the effect of the rating on retrofits. It would be a stringent test, since both groups received detailed retrofit recommendations and the only difference was the addition of a rating. A comparison of rated and unrated homes in the absence of the information from a RCS audit would probably show a greater effect. It is also possible to assess the effects of ratings on home energy use by collecting energy bills rather than measuring retrofit activity. The two variables, energy use and energy efficiency, have quite differ- ent meanings for policy. If ratings cause builders to produce more enerav-efficient homes and lead buvers to pay a higher price for them, this may increase or decrease energy use within a housing market. If people purchase energy-efficient homes that are larger or more appliance-intensive than they could otherwise afford, a rating system _ ~ ~ _ ~ . . , ~
77 that is a success in terms of efficiency may not produce any net energy savings. If some people purchase detached houses, which most people prefer, rather than less energy-intensive and less costly attached houses, there might be the same effect. If some people move from older rental units to new energy-efficient homes, there may be either an increase or a decrease in their energy use. Although some of these outcomes do not save energy, all of them represent an increased standard of living for the purchasers of new energy-efficient homes. Thus, the choice of a measure of effect depends on whether the policy goal is to reduce energy consumption, to improve energy efficiency, or some combination of the two. CONCLUSIONS AND RECOMMENDATIONS Research on home energy rating systems has focused mainly on improving their accuracy. Relatively less attention has been given to the design and implementation of ratings, and it is in these areas that rating systems have hit their most stubborn barriers (McCarty and Willner, 1985~. Experience and behavioral research show that implemen- tation is the first consideration. The best strategy for implementation is to gain the cooperation of the range of actors that will be affected with a rating system: mortgage lenders, builders, real estate agents, appraisers, homeowners and buyers, retrofit contractors, and the sponsors of home energy audit programs. Which of these groups is most critical depends on who will be conducting ratings and on whether the primary target of the system is the new home market, the home resale market, or home retrofits. In ratings intended to affect home sales, the participation of mortgage lenders has been essential. We offer four recommendations for implementing home energy rating systems and for research to identify effective rating systems and improve their delivery. 1. Advice and cooperation should be solicited from all interested groups in a housing market well before a rating system is put into place. 2. Several rating designs should be evaluated in experimental field trials. Three ways of presenting ratings have so far shown promise in pilot programs: a five-star rating derived from a heat loss methodology, a 0-10 rating with anchor points at "no energy-saving features" and "no energy bills," and an annual cost estimate derived from an RCS energy audit. The first two of these should be given experimental field trials alone and in combination with cost estimates. Before using other possible formats under field conditions, potential users should be asked for their reactions. 3. Experimental tests of ratings should be conducted only in the context of ongoing activities within which ratings might become routine, _ such as energy audits, appraisals, or the approval of building permits. Only in such settings are energy ratings likely to be funded and institutionalized.
78 4. Resources should be made available for controlled, quantitative evaluation of the effect of energy ratings on rated homes and housing ... .. marketse To date, no home energy rating system has been carefully evaluated to determine whether it produces the results it is designed for: increasing the salability and market price of energy-efficient homes and increasing the energy efficiency of new and existing housing stock. The ideal research design for assessing effects on retrofits and sale prices is an experiment in which a number of homes are rated but the ratings are made available for only a random sample of those homes. Whole housing markets should also be studied to see if a rating system affects the energy efficiency of the population of homes offered for sale; econometric approaches are useful for this purpose . Studies shout d assess both effects on energy eff i ciency and on measured energy use. i
