environmental performance, the development of an accounting for these externalities is a priority for research and practical experimentation.

The potential for quantification of targets and achievements through physical measures—tons of hazardous wastes, proportions of recyclable to nonrecyclable materials, concentrations of particulate emissions, etc.—is clear. Such measures are already illustrated in the publicly available reports on pollution control including, increasingly, companies' annual environmental reports (Collier et al., 1993). Internally, such measures may also be used as part of an array of targets and performance indicators within a balanced scorecard (Epstein, 1994). The increased use of nonfinancial measures, at least at lower organizational levels, is also a feature of modern management control systems with their focus on quality and continuous improvement and is increasingly important where organizations promote bottom-up empowerment rather than top-down control and are downsizing and flattening their structures (Epstein, 1994; Tyson, 1994). However, the power of the financial bottom line has always made it accountancy's strongest weapon, both in its apparent capability to summarize organizational performance across a diverse range of divisions, activities, and products and in its behavioral linkages to incentives and rewards (Ezzamel et al., 1990). Despite the major reorientations of management accounting systems in recent years, top managements are likely to continue to manage by the financial numbers (Tyson, 1994). The need to capture internal environmental considerations in terms of financial consequences (as in TCA) and to attempt to measure impacts on the environment financially external to the organization is a major challenge for the further development of environmental accounting (Cope and James, 1990).

Increasing quantification (whether physical or financial), however, carries its own dangers. It gives a spurious objectivity to numbers that often reflect highly subjective and judgmental assumptions and estimates. It marginalizes qualitative factors whose subjectivity is thereby further emphasized but that may be more important. The interpretation of accounting numbers remains as important as, if not more important than, the actual numbers. The numbers should provide the means to sharpen analysis and questioning but do not in themselves provide the answers and certainly not the complete answers.