10
A Hypothetical Model for Improving Aggregation and Presentation of Environmental Performance Metrics

Two characteristics of financial metrics that have made them very widely useful is that they can be applied nearly universally and they treat all financial issues thought to be of importance. The credibility of financial metrics has been hard won and is based on a foundation of generally accepted principles. Industrial environmental performance metrics, on the other hand, are nonstandardized and very sector dependent.

This chapter explores ways in which environmental performance metrics might be made more broadly useful through the selection of common metrics topics, normalization onto a common scale, and more effective presentation and aggregations. The generation and standardization of a set of metrics in this way are presented as a hypothetical model for aggregating proxies for environmental performance. Given the number of uses and users of metrics, it is unclear whether it would really be possible to collapse a wealth of metrics information into a handful of composite numbers. In addition, it would be difficult using such an approach to account for differences in geographic location or local circumstances or to characterize ecosystem-level impacts in any universally acceptable way. Nonetheless, this chapter illustrates the potential advantages of such a scheme over the present system of metrics used to gauge industrial environmental performance metrics.

Guidelines For Generic Metrics

Unlike measures of financial performance, environmental metrics do not automatically lend themselves to a common unit. They tend to be recorded in



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 169
--> 10 A Hypothetical Model for Improving Aggregation and Presentation of Environmental Performance Metrics Two characteristics of financial metrics that have made them very widely useful is that they can be applied nearly universally and they treat all financial issues thought to be of importance. The credibility of financial metrics has been hard won and is based on a foundation of generally accepted principles. Industrial environmental performance metrics, on the other hand, are nonstandardized and very sector dependent. This chapter explores ways in which environmental performance metrics might be made more broadly useful through the selection of common metrics topics, normalization onto a common scale, and more effective presentation and aggregations. The generation and standardization of a set of metrics in this way are presented as a hypothetical model for aggregating proxies for environmental performance. Given the number of uses and users of metrics, it is unclear whether it would really be possible to collapse a wealth of metrics information into a handful of composite numbers. In addition, it would be difficult using such an approach to account for differences in geographic location or local circumstances or to characterize ecosystem-level impacts in any universally acceptable way. Nonetheless, this chapter illustrates the potential advantages of such a scheme over the present system of metrics used to gauge industrial environmental performance metrics. Guidelines For Generic Metrics Unlike measures of financial performance, environmental metrics do not automatically lend themselves to a common unit. They tend to be recorded in

OCR for page 169
--> such disparate units as pounds of waste generated, liters of water used, or hectares of forest harvested. Techniques exist for establishing common ground, however, if one relates performance on a percentage basis or some other uniform system. A generic metric is sufficiently general to be used without regard to the industrial sector of the user, and a generic metrics set is a group of metrics that forms a reasonably complete picture of a corporation's environmental performance. A single such set, while it might provide useful guidance, is unlikely to serve all the metrics needs of all companies, or even all the metrics needs of a single company. However, a generic metrics set does contribute to the ability to assess competitive performance and is of considerable value in that regard. Metrics come in broad classes, related to resources, environmental burdens, and human health and safety. While perhaps not strictly environmental, health and safety are often monitored by the same organization within a corporation that tracks environmental performance and are presented in corporate environmental reports, so it seems useful to consider them concurrently. Other metrics deal with products, suppliers, and broader environmental issues, such as sustainability. In Chapter 8 the committee suggested a generic set of metrics divided into seven categories: resource metrics for manufacturing, products, and product packaging; environmental burden metrics; human health and safety metrics; supplier performance metrics; and sustainability metrics. The seven categories are described more fully below. Resource Metrics in Manufacturing Resources are a natural focus for environmental metrics because of the potential, established to a greater or lesser degree for different resources, of eventual unavailability; the potential for some resource acquisition to entail unacceptable environmental impacts; and the feeling that a "single use and discard" approach to resources, especially nonrenewable resources, is inherently unsound. For purposes of this example, three resource metrics, related to the use of materials, energy consumption, and water use, are considered. The units used to express these metrics are important. One cannot merely report mass of materials used, for example, because a yearly increase or decrease in the reported figure might simply reflect different production volumes. Rather, material use should be normalized to take production into account (e.g., pounds of material per pound of product rather than pounds of material per automobile manufacturing facility). One may also wish to consider the relative abundance of the materials. The minimum value for a materials-use metric is obviously unity; that is, every molecule entering the facility leaves as part of a product. The maximum value is more difficult to specify, as it is almost certainly dependent on the industry sector. The smelter that extracts gold from ore is much less likely to avoid substantial residues than the goldsmith who crafts fine jewelry, for example. Performance is also heavily dependent upon suppliers. If a manufacturer receives

OCR for page 169
--> metal produced by "near net shape" casting, for example, there will be less residue than if the same manufacturer begins with ingots. Energy metrics relate both to the industrial processes employed and to the materials provided by suppliers. If minor assembly of components manufactured by others is the function of a facility, as is often the case in the electronics industry for example, the minimum energy may be close to zero. Water metrics are directly related to the type of industrial processes in question. The minimum is obviously zero (no water use per unit weight of product); the maximum depends on the particular industry. In some industries, soft-drink manufacture for example, water is part of the product. In others it is used for cooling or as a constituent in chemical solutions. Water is often returned to its source cleaner than when it was removed, so an appropriate metric might be something like "gallons of degraded water discarded." In addition to complexities resulting from the uniqueness of each industry sector, data availability may also limit the degree to which resource consumption-related metrics can be employed. Incoming materials and outgoing products may not be customarily available, for example. One might then consider related metrics, such as the cost of incoming materials divided by the cost of outgoing product, or the cost of incoming materials per unit of outgoing product. Such "second-level" metrics are much less desirable than "first-level" metrics because they are influenced by nonenvironmental factors such as financial negotiations with suppliers and customers, global financial oscillations, and so forth. The situation is equally unsatisfactory with regard to water and energy, for which geographically influenced supply conditions as well as financially related factors could blur the environmental evaluation. Ideally, then, resource consumption-related metrics would be established by having industry consortia or other appropriate groups provide first-level maximum and minimum normalized values (for example, on a scale of 0 to 10). These scales will need to be periodically reviewed and updated to keep pace with improvements in knowledge and process technology. The performance of individual facilities or corporations could then be compared with those standards. Suppose, for example, that an industry were to establish standards as follows:   Minimum Maximum Materials (lb./$ of sales) 3 10 Water (gal./$ of sales) 0 50 Energy (BTU/$ of sales) 1,000 2,500

OCR for page 169
--> Graphically, this would give Thus, a corporation using 6.5 lb. of material, 35 gal. of water, and 1,375 BTU per dollar of sales would have metrics of M1, = 5, M2 = 3, and M3 = 7.5. Materials use might also be expressed as yield, for example pounds of product per pound of input materials, or as pounds of input materials per unit of product. The choice should be made to maximize the potential for comparability among factories and companies, to allow long-term tracking, and to permit the easy acquisition of the necessary information. The conservation of resources is strongly linked to several aspects of product design and customer interaction. Accordingly, metrics that measure these attributes are helpful indicators of environmental performance. The World Business Council for Sustainable Development (1996) has suggested several metrics of this type including the use of recycled materials, the use of renewable resources, and the provision of services rather than goods. Metrics based on these measures may be strongly sector dependent, and industry consortia or other appropriate groups will need to establish more-descriptive definitions. One metric that seems particularly likely to be useful, however, is the percent of recycled materials in products (weight basis), or M4. Both preconsumer and postconsumer recycling could be included. This metric can be transformed to the same 0–10 rating scale used for resource consumption metrics by dividing the percentages of recycled material by 10. Resource-Related Metrics for Products A distinct group of resource-related metrics are those related to the use of products rather than their manufacture. The most common concern in this connection is with the use of energy, as with television sets, washing machines, or

OCR for page 169
--> elevators. The unit of measure is generally the energy consumption of a single product over a given period of time, such as BTU/yr. Water and other ancillary inputs such as detergent or oil are other inputs with potential environmental significance that might be tracked. As with other resource-related metrics, an industry group or corporation needs to establish energy consumption standards. For a television set, for example, the minimum consumption might be set as 0.5 kWh and the maximum at 1.0 kWh. The relative scoring scale is then In-service energy use (M5) (kWh) Thus, a television set drawing 0.75 kWh would have a metric of M5 = 5.0. Other product-focused metrics could be related to water use (a washing machine), resource consumption (use of lubricants), or other topics of interest. Resource Metrics for Product Packaging Packaging requirements differ greatly across industry sectors: It is obviously necessary to package a liquid chemical, but it is not so obvious that an automobile requires very much in the way of packaging. Within all sectors, however, the use of packaging is a useful measure of environmental performance, and a reasonable (though not technically comprehensive) metric is weight of packaging per dollar of sales. Thus, if an industry consortium set a packaging standard of 0.1 lb. of packaging per dollar of sales, the relative scoring scale becomes Packaging quantity (M6) (lb./$ of sales) A corporation whose packaging use is 0.02 lb./$ would have a rating of M6 = 8.

OCR for page 169
--> Environmental Burden Metrics Compliance with regulatory requirements is a central feature of corporate environmental performance metrics, and such metrics are of obvious utility. However, simple reporting of tons of emissions gives little perspective on environmental performance. Normalization is needed. As with the water resource metric, the minimum level for the emissions metric is zero. The maximum level can be set by an industry consortium or some other appropriate group. These levels are then transformed to the 0–10 scale. For example, suppose that an industry-sector standard for maximum emissions were 0.5 pounds of emissions per dollar of sales. The rating scale would look like this: Emissions (M7) (lb./$ of sales) Emissions of 0.35 pounds per dollar of sales would have a metric of M7 = 3.0. Human Health and Safety Metrics One or more metrics related to corporate health and safety performance is generally included in a company's environmental, health, and safety report. A common metric is the "recordable incidence rate," often expressed as the number of job-related injuries or illnesses per 100 employees. Zero is obviously the minimum for this metric. The maximum will be industry specific and should be established by an industry consortium or some other appropriate group. For example, suppose that an industry set the maximum rate at three incidents per 100 employees. Graphically, this gives Health/safety (M8) (incidences per 100 employees)

OCR for page 169
--> If a particular corporation had an incidence rate of 1.8, the metric value would be M8 = 4. The use of several metrics is common in the human health and safety arena. Examples include annual worker fatalities and the number of lost workdays per employee. Supplier Performance Metrics Corporations evaluate their suppliers on the basis of many characteristics, including financial stability, product quality, and manufacturing capacity. Some corporations are now also beginning to assess their suppliers' environmental performance. There are many variations on the practice, with some companies considering only whether their suppliers are in compliance with environmental regulations and others examining some of the aspects of the metrics described in Chapter 9. Using the 0–10 scale, a middle-of-the-road supplier might be rated M9 = 5. Sustainability Metrics Several industry sectors have suggested that one or more metrics dealing with sustainability might be useful. A potential metric of this type could convey information about the uses made of corporate lands. Such a metric would be very industry specific, since forest products companies or agricultural organizations obviously treat land much differently than do mining companies or electronics manufacturers. One way to define a land-use metric would be to divide corporate lands into three categories: α—land maintained in essentially a natural state; β—land actively utilized for corporate purposes but with consideration given to habitat disruption, erosion, and related topics; and χ—land actively utilized for corporate purposes without significant consideration of long-term ecosystem impacts. If α, β, and χ are the fractions of corporate land determined (perhaps by an outside auditor) to fit the classifications as stated, we might arbitrarily assign a value of 1 to α lands, 0.5 to β lands, and 0 to χ lands. This would result in the following scale:

OCR for page 169
--> Sustainability A weighted calculation is thus: M = (1 × α) + (0.5 × β) + (0 × χ) For example, if a corporation has 10 percent α lands, 20 percent β lands, and 70 percent χ lands, the calculation gives M = (1 × 0.1) + (0.5 × 0.2) + (0 × 0.7) = 0.2 This metric is transformed to the uniform scale by multiplying by 10: M10 = 10[(1 × α) + (0.5 × β) + (0 × χ)] = 2 Other potential metrics for sustainability can relate to specific impacts. A widely used metric is the emissions of particular chemicals, such as the chlorofluorocarbons and hydrochlorofluorocarbons that contribute to ozone depletion. Still other metrics might refer to habitat disruption, ambient noise generation, or some other impact of interest. Use of Weighting System Some metrics may be more useful if weighting factors are applied. In the case of materials, for example, M1, as initially defined, values a pound of platinum (a rare metal often in short supply) the same as a pound of silicon (an extremely common material). Without complicating the metric too much, one might take account of scarcity. One way to do this is to put all materials in three categories: those that are abundant, those whose supplies are moderately constrained, and those that are scarce. Arbitrarily, one can assign weighting factors

OCR for page 169
--> of 1.0, 0.5, and 0, respectively, for these categories. The weighted version of metric M1 then becomes M1 (weighted) = M1(1 × ε) + (0.5 × λ) + (0 × μ) where ε, λ, and μ are the fractions (by weight) of abundant, constrained, and scarce materials, respectively, in the product under study. A second metric for which weighting seems appropriate is that of emissions. Here, weighting factors can help distinguish materials that are nonhazardous, moderately hazardous, and highly hazardous. The weighted version of metric M7 then becomes M7 (weighted) = M7 (1 × π) + 0.5 × ρ) + (0 × σ) where π, ρ, and σ are the fractions (by weight) of nonhazardous, moderately hazardous, or highly hazardous emissions, respectively. If weighting factors are to be used, it will be important to get community agreement on which material resources should be assigned to the scarcity categories and which emissions should be classed as hazardous. Such designations will have a significant impact on the usefulness of the metrics, and agreement on their use will be vital to their widespread acceptance. Expanded Metrics In many cases a corporation may wish to develop detailed, as opposed to generic, metrics information. For example, a company may wish to express emissions to air, water, soil, and deep-well injection rather than to present a composite emissions figure. As before, minimum and maximum values would need to be established by an individual corporation or industry consortium, such as   Minimum Maximum Air emissions (lb./$ of sales) 0.1 1.0 Water emissions (gal./$ of sales) 3.0 7.0 Soil emissions (lb./$ of sales) 1.0 2.0 Deep-well injection (gal./$ of sales) 20.0 50.0  

OCR for page 169
--> Graphically, this gives Thus, for a corporation emitting 0.7 lb. of air emissions, 4 gallons of water emissions, 1.8 lb. of emissions to soil, and 29 gallons of deep-well injections, the expanded metrics would be M7a = 4, M7b = 7.5, M7c = 2, and M7d = 7. Metrics Displays and Metrics Aggregation The way in which metrics data are displayed and aggregated can increase their meaningfulness and utility. The danger of such approaches is that they can be simplistic and misleading; however, because of the potential payoff, they are worth exploring. The simplest level of reporting conveys uniform values and trend information about individual metrics. The major benefit of having a common metric is that the uniform rating scale permits the environmental performance of corporations in the same industrial sector to be directly compared and the environmental performance of a single corporation to be tracked over time. It is not a significant problem that the scales will differ for different sectors. Such financial measures as debt equity and dividend payout ratios are routinely expected to be sector dependent, for instance. To get an overview of a corporation's environmental performance, metrics can be displayed in a grouped chart. Suppose that 10 metrics have been selected and that all have been normalized on a scale of 0–0. We can now create a triangular metrics scoring display divided into 10 areas (one for each of the 10 metrics) (Figure 10-1). Within each area we express the rating score by a color, as follows: Red Orange Yellow Cyan Green 0–2 2–4 4–6 6–8 8–10 Rating score

OCR for page 169
--> FIGURE 10-1 Triangular scoring display for metrics data. A colored triangular display could show the metrics in a form that directs the user's attention to areas of highest and lowest performance, and it might be an efficient way to communicate the ensemble of industrial environmental metrics information applicable to an individual facility or a corporation as a whole. Such a display has the potential to become a uniform and readily understood presentation of corporate environmental performance, much as the U.S. Department of Agriculture's chart of daily minimum nutritional requirements efficiently presents the characteristics of foods or food products. A final though potentially controversial step might be to aggregate the results from a metrics set (Table 10-1) into a single indicator of corporate environmental performance. The data in Table 10-1 can be shown in a series of triangle plots, as seen in Figure 10-2. The figure exploits the human ability to group patterns more quickly than numbers and colors more readily than shadings. Improvement in metrics 3, 5, 6, and 7 and a decline in metric 9 are quickly appreciated. It is unclear whether a single environmental score for a corporation represents appropriate or inappropriate aggregation. Nevertheless, one could derive an overall environmental rating (Msum) for a facility or corporation by adding the 10 metrics. In this model a perfect score would total 100. The result could be expressed as a trend over time (Table 10-2) or plotted as shown in Figure 10-3.

OCR for page 169
--> It is important to note that three distinct steps were involved in generating the composite metric shown in Figures 10-2 and 10-3. The first step was to select the classes of metrics. Three classes—related to resources, environmental burden, and health and safety—were chosen. For aggregation to work, agreement needs to be reached up front on which classes of metrics should be analyzed. The classes chosen will almost certainly be dependent on the industry sector. The second step is to select specific metrics within the chosen classes. Ideally, this step would also be independent of sector, though in some cases that may not be possible. The ideal number of metrics is not obvious, but more than 10 would likely be too confusing and expensive, while fewer than five seems unlikely to cover the necessary ground. This step will be harder to accomplish than the first. The third step is to assign minimum and maximum values for the chosen metrics; this is clearly sector dependent and is likely to be the most difficult to accomplish. Summary A set of corporate environmental metrics based on the model described in this chapter would have several uses. The aggregate rating would be universal, and the environmental performance of corporations in completely different industrial sectors could be compared at this level. At the level of individual metrics, corporations within the same industrial sector could be directly compared. At the intermediate level, that of metric classes, comparisons would be dependent or independent of industry sector, as a function of how the metric classes were established. For a number of metrics groupings, arriving at a generic set of metrics appears to be a reasonable expectation. These groupings, however, measure corporate activity related to the environment, not corporate impacts upon the environment. For the foreseeable future, this is probably as much as can be expected, and if such metrics are widely used by corporations, the environment will benefit. As noted in the beginning of this chapter, the generic set of metrics and its possible use as a tool for comparison are but a model of what should be a relatively simple and straight forward process. Yet, as the four sector studies indicate, different users cast metrics in different forms, sometimes normalized to production, sometimes not, sometimes focused at the product or process level, other times at an entire facility, business unit, corporation, or nation. This model is presented to show what may be possible and how such a scheme may be used. The steps outlined illustrate the need for consensus building on a set of simple, finite metrics; on the upper and lower limits used to rank performance; and on the relative positions of the various metrics within the triangle. This chapter is presented as food for thought, not as a final solution. It recognizes that there is a need for better metrics and better presentation of the information they convey. Arriving at a set of metrics that measures corporate

OCR for page 169
--> TABLE 10-1 Hypothetical Metrics Data, 1995–1998   1995 1996 1997 1998 M1 2.1 2.3 2.3 2.4 M2 4.6 4.5 5.7 5.8 M3 5.9 6.1 6.2 6.3 M4 5.1 5.2 5.4 6.1 M5 3.6 4.5 5.8 6.2 M6 3.5 3.7 4.1 4.3 M7 1.6 1.8 2.1 2.2 M8 7.2 7.6 7.9 8.1 M9 4.4 4.4 4.3 3.9 M10 8.4 8.6 8.7 8.8 Figure 10-2 Triangular scoring display of hypothetical data, 1995–1998.

OCR for page 169
--> TABLE 10-2 Hypothetical Aggregate Metrics Scores, 1995–1998 1995 1996 1997 1998 46.4 48.7 52.5 54.1 Figure 10-3 Hypothetical aggregate environmental metrics rating (Msum), 1996–1999. activity related to the environment is a difficult task. Measuring corporate impacts on the environment, a much more difficult yet important topic, is the subject of the next chapter. Reference World Business Council for Sustainable Development (WBCSD). 1996. Eco-efficient Leadership for Improved Economic and Environmental Performance. Conches-Geneva, Switzerland: WBCSD.