1. Introduction

Corporations use their annual reports to disclose environmental information. These disclosures are mostly narrative, cover various environmental themes, and can be found throughout the corporate annual report. An environmental disclosure theme is characterized by placing an environmental disclosure into a given category. The theme represents a recording unit, which is the first step in designating units to be analyzed using content analysis. Content analysis uses a set of procedures to make valid inferences from messages (Weber 1990, p. 9). The most useful unit of content analysis is the theme, because it leads to much more detailed and sophisticated comparisons (Holsti 1969, p. 116; Weber 1990, p. 22).

Through understanding the quantity and quality of environmental disclosure themes, the direction of future research and possible regulation to improve environmental disclosures may be established. For instance, discretionary environmental disclosures usually made in the nonfinancial section of the annual report may provide important information aside from regulated disclosures usually found in the financial section of the annual report. These nonfinancial environmental disclosures made at the discretion of management may inform the accounting profession and authoritative bodies of generally accepted environmental themes and practices that could be improved upon. Through more adequate authoritative guidance, improved financial environmental disclosures can be more informative, and may make consistency within firms and comparability between firms possible. Previous research has not addressed adequately the issue of environmental disclosure themes from a financial versus nonfinancial perspective, or provided a reasonable basis to make comparisons of disclosure themes feasible.

From a financial perspective, the possible effect that environmental disclosures can have on the perception of a firm's cash flows and earnings potential is reason for concern. Blacconiere and Patten (1994, p. 374) suggest that investors may interpret more extensive environmental disclosures as a positive sign of the firm managing its exposure to future regulatory costs. "Environmental issues can dramatically impact a company's short-term financial position and its chances for long-term success" according to John P. Surma (1992, p.22), a partner and director of environmental services for a major accounting firm.

Environmental issues have been and continue to be a major social issue facing many corporations (e.g., Post 1991, p. 33). The Financial Accounting Standards Board (FASB), the Securities and Exchange Commission (SEC), and others have begun to realize and address the importance and necessity of providing appropriate environmental information in a manner helpful to assessing properly the impact on cash flows and earnings. For instance, Todd Johnson (1993, p.123), a research manager at the FASB, believes that "the time is right for research that would help the FASB and others address the financial reporting questions associated with environmental costs and obligations." There is a need to understand and improve environmental disclosures through research.

The objective of our study is to provide insight into environmental disclosure practices by examining the nature and extent of disclosure themes in terms of quantity and quality between the financial and nonfinancial sections of corporate annual reports. The year of study is 1989. In addition, our study seeks to explore the relationship between the quantity and quality of disclosure themes and environmental performance.

The next section presents relevant studies for this research. The design methodology and hypotheses appear in section 3. Section 4 presents the findings. Concluding discussion and implications of the study appear in Section 5. An Appendix details the content analysis and coding scheme used.

2. Relevant Studies

2.1 Introduction

By 1989, numerous environmental laws and regulations had been enacted and public policy regarding the environment had changed (Roussey 1992, p. 48). The disaster of the Alaskan Exxon Valdez oil spill on March 24, 1989 represents an event that has focused much attention on environmental issues and public policy debates (Benoit 1989, p. 82; Grover 1989, p. 34). The magnitude of this disaster resulted in the "Valdez Principles" (Gray 1990, p. 67; Sanyal and Neves 1991, p. 883). The level of disclosures in corporate annual reports has also changed. Findings by Walden and Schwartz (forthcoming) and Gamble et al. (1995, p. 47) suggest significant positive differences in the level of disclosures for 1989 and for several years thereafter. Therefore, there is a need to analyze and evaluate the nature and extent of environmental disclosure themes made by firms in their corporate annual reports beginning with 1989.

Post (1991, p. 33) suggests that the prevalence and seriousness of environmental problems are becoming increasingly evident to business. Environmental issues can no longer be ignored (Post 1991, p. 33). An objective of financial reporting is to provide information to potential investors, current owners, creditors, and others that will aid in assessing the future earnings and cash flows of an enterprise (FASB 1986). Management may communicate environmental information to those outside of an enterprise by means of various types of disclosure. The information is required to be disclosed by regulatory rule or because management considers it useful to those outside the enterprise and discloses it voluntarily (FASB 1986).

The most notable environmental legislation established "Superfund" in which potentially responsible parties (PRPs) are made liable for remedial clean-up of hazardous waste sites (Roussey 1992, p. 49). These potential liabilities along with other environmental costs and clean-up of hazardous wastes may affect earnings and cash flows, and may not be disclosed even though required by regulatory rule (Roussey 1992, p. 50). SEC Regulation S-K requires disclosure of a material potential environmental liability in the "Legal Proceedings" section of the SEC Form 10-K (annual filing). In 1989, SEC Financial Reporting Release No. 36 (SEC 1989, Section III.B.) required firms to disclose in the "Management Discussion and Analysis" section the effects of PRP status and quantify them whenever reasonably practicable. The SEC position of disclosing environmental problems has been strengthened by an agreement with the Environmental Protection Agency (EPA) beginning in 1989 that the EPA will provide information on PRPs to the SEC in order to target environmental disclosures for enforcement (Roussey 1992, p. 53). The SEC disclosure requirements are at a "reasonably likely to occur" level as opposed to the FASB Statement No. 5 threshold regarding contingent liabilities that is based on "probable of being asserted."

Epstein (1991, p. 11) and Epstein and Freedman (1994, p. 106) suggest that investors want environmental disclosures in annual reports. According to Epstein and Freedman (1994, p. 106), the vast majority of investors want environmental disclosures, and a large minority of those surveyed would like these disclosures audited. Ideally, disclosure should relate the environmental outcomes that corporate actions have on stakeholders in a manner that is meaningful, fair, and provides an adequate basis for decision making (AAA 1976, p. 48). Unfortunately, most environmental disclosures are narrative, cover a variety of topics, and are oftentimes difficult to analyze or interpret.

2.2 Environmental Disclosures and Environmental Performance

Many researchers investigating environmental disclosures and environmental performance have found little support for a relationship. A number of prior studies used the Council on Economic Priorities' (CEP) 1977 Pollution Audit of environmental performance indices, which are based on investigations of the pollution control records of the leading firms in five industries known to have environmental problems (Ullmann 1985, p. 544). According to Freedman and Wasley (1990, p. 185), the CEP indices are the most comprehensive objective measures of firms' environmental performance. The following are representative of content analysis studies with most involving the CEP, the quantity and/or quality of disclosures, and a potential relationship with environmental performance.

Ingram and Frazier (1980, p. 615) examined the relationship between measures of firms' environmental performance devised by the CEP in 1977 and the environmental disclosures found in annual reports for a sample of 40 firms. The insignificant results suggested that the content analysis scores of firms' disclosures do not relate strongly to CEP indices of firms' environmental performance. The only important relationship identified was that, except for disclosures related to litigation, the annual reports of the poorer performers contained more environmental disclosures than the better performers. The greater disclosure by the poor performers appeared as narrative disclosures in the discretionary section of the annual report.

Freedman and Jaggi (1982, p. 170) used content analysis to investigate the association between environmental disclosures in 10-Ks and the environmental performance by the CEP for 109 firms in four highly polluting industries. The results confirmed earlier findings that indicated there is no association between environmental disclosures and environmental performance (Freedman and Jaggi 1982, p. 174).

Wiseman (1982, p. 54) examined the annual report disclosures made by 26 firms in three industries with their environmental performances as ranked by the CEP in 1977. Her content analysis measured the extent of disclosures using 18 items and four categories to evaluate the quality and accuracy of environmental disclosures (Wiseman 1982, p. 55). These findings indicated that the voluntary environmental disclosures were incomplete, providing inadequate disclosure for most of the environmental performance items included in the content analysis (Wiseman 1982, p. 62). Her findings also demonstrated that no relationship existed between the contents of the firms' environmental disclosures and the firms' environmental performance (Wiseman 1982, p. 62).

Rockness (1985, p. 342) used 128 participants to conduct a field experiment to examine the reliability of voluntary environmental disclosures by testing whether annual report users were able to make accurate comparative judgments among U.S. firms' environmental performance based on their annual report disclosures. The CEP 1977 Pollution Audit provided an external evaluation of environmental performance for 26 firms in three industries included in her study (Rockness 1985, p. 343). Results showed that users formed consistent comparative evaluations of firms' environmental performance within an industry, but these evaluations were inaccurate interpretations of actual performance as measured by the CEP (Rockness 1985, p. 350). Her study concluded that environmental disclosures are incomplete or inaccurate reports of actual performance (Rockness 1985, p. 350).

Freedman and Wasley (1990, p. 184) examined the relationship between environmental performance and environmental disclosures made in both annual reports and in the 10-K reports filed with the SEC. They used the CEP indices from 1977 for 50 firms in four industries (Freedman and Wasley 1990, p. 185). The results of the content analysis, similar to Wiseman (1982), indicated that neither voluntary annual report environmental disclosures nor mandatory 10-K report disclosures are indicative of actual firm environmental performance (Freedman and Wasley 1990, p. 191). Their findings implied that if environmental disclosures are to be useful to financial statement users then regulation of these disclosures may be necessary (Freedman and Wasley 1990, p. 191).

Freedman and Stagliano (1995, p. 165) using content analysis examined the 1987 SEC 10-K environmental disclosures of firms named potentially responsible parties (PRP) under Superfund. The objective was to detail both the existence of disclosures and type of discussion provided by identified firms (Freedman and Stagliano 1995, p. 165). A computerized search by key words of all 1987 10-K reports and other documents resulted in a sample of 193 firms (Freedman and Stagliano 1995, p. 167). Freedman and Stagliano (1995, p. 170) suggested that firms potentially impacted in the same way report the environmental event in a variety of ways and often in a conflicting manner. These disclosures fail to help stakeholders to reach an informed judgment as to the potential impact Superfund will have on the firm (Freedman and Stagliano 1995, p. 170). Their study concludes that even when disclosure regulations exist, corporations still may disregard them to avoid disclosing potentially damaging information (Freedman and Stagliano 1995, p. 171). Stakeholders must be more diligent in the pursuit of "full disclosure" regarding the financial impacts of Superfund according to Freedman and Stagliano (1995, p. 171).

2.3 Summary

Studies using the CEP indices from 1977 have suggested that no relationship exists between environmental disclosures and environmental performance. Environmental disclosures have been incomplete and inaccurate in relation to performance. Public policy has changed over the last decade with respect to the environment. Freedman and Stagliano (1995, p. 171) suggested that, even with regulation, environmental disclosures can mislead stakeholders due to the variety of reporting options available and the fact that some firms provide conflicting information about their PRP status resulting from Superfund. Therefore, environmental disclosures may be misleading stakeholders because there seems to be no relationship between the disclosures and the environmental performance of the firms. This concurs with the various studies cited by Ullmann (1985, p. 545) that find no relationship between social disclosures and social performance.

3. The Study

3.1 Methodology

The CEP's Corporate Environmental Data Clearinghouse (CEDC) selected 1989 to monitor, gather, and analyze information on corporate environmental performance for firms in the Fortune 500 (e.g., CEP 1992). Firms in our study are chosen based on firms and industries previously analyzed by the CEP's CEDC. These firms represent leaders in their respective industries as defined by the CEDC. The CEDC provides information on these firms; the task of accumulating this information is being accomplished over a period of years. By the middle of 1992, environmental reports for 57 firms were available from the CEDC, and grouped into four industries by the CEDC. The first four available industries were "chemical," "consumer products," "forest products," and "oil" ("CEDC reports" 1991, p. 3; "New CEDC reports" 1992, p. 6). Using content analysis, 1989 corporate annual reports for the 57 firms selected initially by the CEDC are analyzed for environmental disclosures using a computerized search. From preliminary analysis, a keyword descriptor listing of environmental terms was developed to search for the disclosures.

The text of the environmental disclosure is read, and the content of the disclosure is coded in stages using content analysis as described in the Appendix. Descriptive analyses and statistics summarize and characterize the data from the research task. They provide insight into environmental disclosure in terms of quality, as well as quantity, between the financial and nonfinancial sections of corporate annual reports for 1989. One measure is used to capture the quantity of disclosure, and two measures are used to capture the perceived quality or information content of disclosure (see Appendix). The quantity measure is the number of theme occurrences (NTO), which measures the number of specific themes occurring within five broad categories. The quality measures are a four-element quality index (QI) with a maximum of six points for each specific theme occurrence, and a disclosure score (DS), which represents the summation of the QI for each specific theme category.

Separate analyses for the financial and nonfinancial sections of the annual report provide information about the usage of these sections for disclosure. It is believed that accounting and regulatory requirements should affect the disclosures made in the financial section of the report, since this section is reviewed by auditors. It is also believed that the disclosures located in the nonfinancial section should be affected more by social change (discretion) than regulatory effects, since management has wide discretion to include in the nonfinancial section what they consider to be important. Analyses are performed in total for all four industries and by each industry to investigate the nature and extent of the disclosures.

3.2 Hypotheses

Correlation analyses are used to explore the relationship between the quantity and quality of disclosures and environmental performance. The following two null hypotheses of no association are tested:

The availability of environmental performance factors is problematic. Reports from the EPA regarding such information as Superfund-PRP status and toxic releases often are very difficult to obtain and compile. Fortunately, the CEDC does provide information on Superfund-PRP sites and toxic releases per $1,000 of sales for 1989. These two factors are used as proxy variables for a firm's environmental performance. Neither the Superfund-PRP sites nor toxic releases proxy variables indicate the severity of the Superfund-PRP sites or toxic releases to the firm, which is an inherent weakness in these two proxies.

Beyond the hypotheses above, the study seeks to investigate the relationship between the environmental contingencies disclosures in the financial section and the two environmental performance proxies, Superfund-PRP sites and toxic releases per $1,000 of sales provided by the CEDC. Two additional null hypotheses of no association are tested:

Accounting and regulatory requirements should affect the environmental disclosures made in the financial section of the report, which are reviewed by auditors. The disclosures related to environmental contingencies are of particular interest. If these disclosures are useful and informative, we should expect an association between the environmental contingencies disclosure and environmental performance.

A potential relationship between the quantity and quality of disclosures and size is considered in addition to the environmental performance factors used. Recent studies by Patten (1991, p. 305; 1992, p. 475) suggest that the level of social disclosures including environmental disclosures is related to public policy pressure (see also Walden and Schwartz, forthcoming). Patten (1991, p. 303; 1992, p. 474) used the natural log of sales as a size variable, and it is used in our study as well. This proxy variable may represent many other things besides public policy pressure. The following two null hypotheses of no association are tested:

Each fundamental hypothesis dealing with the quality of disclosure has allied with it a specific hypothesis for each quality measure, the average quality index, or the disclosure score. These specific hypotheses are simply denoted as "A" for the average quality index, and "B" for the disclosure score. For example, the null hypothesis 2A can be stated as follows:

4. Findings

4.1 Sample and Search Procedure

The final sample for our study consists of 53 companies across four industries as reported in Table 1 along with their standard industrial classification codes (SIC). Four companies listed separately in Table 1 of the original sample of 57 did not have annual reports due to privatization and reorganizations. The final sample consists of 11 companies each from the chemical and consumer products industries, 16 companies from the forest products industry, and 15 companies from the oil industry.

(Insert Table 1 here)

The keyword descriptor listing discussed above was used to perform a computerized search by company using the electronic annual report database from the National Automated Accounting Research System (NAARS) Service. Once the searches for a company were complete, the disclosures were marked clearly for analysis. Coding was done for each disclosure using two systems of enumeration discussed in the Appendix. Due to the subjective nature of the coding process, two other coders besides the researchers independently coded the data for reliability purposes.

4.2 Descriptive Analyses

Table 2 presents a descriptive analysis for the four industries in total. The financial and nonfinancial sections of the annual reports are analyzed by five broad environmental theme categories. A total of 51 of the 53 companies analyzed had environmental disclosures in their annual reports for 1989. The quantity and quality of disclosure themes varies across the financial and nonfinancial sections. The number of companies reporting various themes also varies.

For environmental contingencies in both sections of the annual reports, this disclosure theme was reported by 24 companies (45%). Environmental contingencies information is noticeably absent from the nonfinancial section with only 5 of 53 companies reporting. For firms reporting, the average number of specific theme occurrences (ANTO) for environmental contingencies is 2.79 times per annual report, and the average quality index (AQI) is 3.24 out of a possible six points (the maximum quality index is six, see Appendix).

(Insert Table 2 here)

The environmental expenditures disclosure theme was reported by 39 companies (74%) in both sections. This information appears to be balanced between the financial and nonfinancial sections. The ANTO for environmental expenditures is 3.28 times per annual report, and the AQI is 3.55 for firms reporting.

For the pollution abatement disclosure theme, 46 companies (87%) reported on this theme in both sections of the annual report. The vast majority of this information is reported in the nonfinancial section. For firms reporting, the ANTO for pollution abatement is 4.26 times per annual report, and the AQI is 2.33.

Thirty-one companies (58%) reported on environmental preservation in both sections. Only one firm reported this theme in the financial section. The ANTO for pollution abatement is 2.32 times per annual report, and the AQI is 2.55 for firms reporting.

Other environmental information was reported by 45 companies (85%) in both sections of the annual report. The majority of this information is disclosed in the nonfinancial section. For firms reporting other environmental information, the ANTO is 4.31, and the AQI is 2.08.

Tables 2A through 2D present the descriptive analyses for each industry (chemical, consumer products, forest products, and oil respectively). The consumer products industry shows the smallest number of environmental disclosures. For all industries, the majority of disclosure themes reported are pollution abatement (46, 4.26) and other environmental information (45, 4.31), based on firms reporting and the ANTO, respectively. They are found mostly in the nonfinancial section. The highest perceived average quality of disclosure (based on AQI) appears to be associated with environmental expenditures (3.55) and environmental contingencies (3.24) respectively. Environmental contingencies information is reported mostly in the financial section of the annual report.

(Insert Tables 2A through 2D here)

4.3 Descriptive Statistics for Disclosure Score, Environmental Performance Factors, and Size

Table 3 presents the descriptive statistics for the DS and the various performance factors for our study. Tables 3A through 3D present this information by each industry. A large degree of variability exists for the DS and the performance factors across industries and sections of the annual report. The largest average DS is found in the oil industry (59.40), followed by the chemical (35.73), forest products (27.56), and consumer industries (6.00) respectively. The largest average Superfund-PRP sites disclosures are found in the chemical industry (27.82) followed very closely by the oil industry (26.13). Across industries, the chemical industry (10.93) has the largest average toxic releases followed by the forest products (3.51), consumer (1.23), and oil industries (1.21) respectively. Size on average appears to be comparable across industries.

(Insert Tables 3, 3A through 3D here)

4.4 Association Between Disclosure, Environmental Performance Factors, and Size

The Spearman rank correlations are presented in Table 4 for the four industries in total. The strongest significant positive associations are between the number of Superfund-PRP sites and the quantity of environmental disclosures as measured by the NTO (H1). There also appears to be significant positive associations between the number of Superfund-PRP sites and the quality of the disclosures measured by the AQI (H2A) and the DS (H2B) found in the financial section. There are no associations between toxic releases and the quantity or quality of disclosure. There are significant positive associations between size, the quantity of disclosure (H5), and the DS (H6B). The strongest associations are in the nonfinancial section. There is no association between size and the AQI.

(Insert Table 4 here)

Tables 4A through 4D present the Spearman rank correlations for each industry. The correlations for each industry show a lack of consistency in relationships across industries.

(Insert Tables 4A through 4D here)

4.5 Association Between Environmental Contingencies in the Financial Section and Environmental Performance Factors

Table 5 presents the Spearman rank correlations for environmental contingencies disclosures reported in the financial section of the annual report and the environmental performance factors, Superfund-PRP sites and toxic releases. This disclosure theme shows no association with toxic releases. For Superfund-PRP sites, there appears to be a significant positive association in total with the four industries combined for all measures (H3, H4A, and H4B). On an industry by industry analysis, significant positive association in terms of quantity and quality for these disclosures appears only for the chemical industry for all measures (H3, H4A, and H4B). The forest products industry indicates a significant positive association to the quality of disclosures (H4A and H4B), while the oil industry shows a significant positive association to the quantity of disclosures (H3) and the quality of disclosures as measured by DS (H4B). There are no associations for sites and these disclosures in the consumer industry.

(Insert Table 5 here)

5. Discussion and Implications

From the findings, it is evident that environmental disclosures occur throughout the annual report with a great degree of variability across the four industries studied. The vast majority of disclosures are found in the nonfinancial section. Based on five broad categories of disclosure themes in total, 87% of the firms in the sample disclosed pollution abatement information, followed by 85% of the firms disclosing other environmental information. Both of these categories occur most often in the Letter to Shareholders and other portions of the nonfinancial section of the annual report. In the financial section, the majority of disclosures involve environmental expenditures and environmental contingencies information, but only 58% and 43% of the firms respectively in the sample reported this information.

The first set of hypotheses related to the quantity and quality of disclosures and the environmental performance factors, H1 and H2 respectively. For the four industries in total and H1, a significant positive association was found between the quantity of disclosures and Superfund-PRP sites, but not toxic releases. A significant positive association is supported for H2B between the disclosure score (DS) and Superfund-PRP sites, but only for the financial section's average quality index (AQI) and Superfund-PRP sites, H2A. For H2, there appears to be no association for the quality of disclosures and toxic releases. When these two hypotheses are analyzed across industry, there are inconsistent results found. These findings suggest definite industry effects with regard to the quantity and quality of disclosure and the use of environmental performance proxies. It further suggests that the quantity and quality of environmental disclosures may make differentiation between high polluting and low polluting industries feasible, but not between firms within a given industry.

The second set of hypotheses was tested for the environmental contingencies disclosures in the financial section of the annual report with similar results. For the four industries in total, all three measures for quantity and quality showed significant positive associations with Superfund-PRP sites (H3, H4A, H4B), but not for toxic releases. However, across industries only the chemical industry confirmed this finding (H3, H4A, H4B). These findings also suggest definite industry effects and variability of disclosures made between industries and companies. In addition, the environmental performance factors used may represent noisy proxies for actual environmental performance.

A third set of hypotheses was tested to determine the relationship between the quantity and quality of disclosures and size, H5 and H6 respectively. For the four industries in total and H5, a significant positive association is supported between the quantity of disclosures and size. For H6, a significant positive association is supported only between the disclosure score and size (H6B), not the average quality index. For individual industries, the results are inconsistent and mixed. Of particular interest is the oil industry for 1989; significant positive associations are found for the quantity of disclosure (H5) in total, and the disclosure score and size (H6B) in the nonfinancial section and total annual report. This lends support to Patten (1992) who suggested that environmental disclosures in the oil industry were related to public policy pressure in association with the 1989 Alaskan Exxon Valdez oil spill. As with the previous findings, a definite industry effect is found.

Subject to limitations, the quantity of environmental disclosure themes varies across sections of the annual report, and the quality of themes is relatively low and varies across themes as well. There is no relationship between the quality (as measured by average quality index) of environmental disclosures and environmental performance. It appears that the quantity of disclosures may increase the perceived information content of environmental information (as measured by the number of theme occurrences and the disclosure score), but these disclosures may only make differentiation between high-polluting and low-polluting industries feasible at best. Researchers should concern themselves with possible industry effects in future studies.

It is interesting to note that the majority of environmental disclosures in the nonfinancial section are concerned with pollution abatement and other environmental information, not the environmental expenditure and environmental contingencies information found in the financial section of the annual report. Management appears to be attempting to inform their stakeholders of environmental information not prescribed currently by regulatory requirement. The information in the nonfinancial section tends to be more positive than negative based on the disclosure themes between sections.

In conclusion, many firms and industries do not appear to provide adequate and informative environmental disclosures. It is doubtful that present environmental disclosures provide environmental information in a manner appropriate for assessing properly the impact on cash flows and earnings for any firm. Perhaps it is time to consider the recommendation of the 1989 American Accounting Association (AAA) Committee on Accounting and Auditing Measurement:

In addition, the minority report of the Committee went further by suggesting a report that would serve a variety of stakeholders and would use, for the most part, data that already are available to regulatory bodies such as the EPA. This type of environmental disclosure report would use a multidimensional approach and use units of measure that are of interest to particular stakeholders, e.g., tons of particulate emissions, and gallons of acid discharge, in combination with dollar amounts.

Further regulation of environmental disclosures is necessary as suggested by the issuance of SEC Staff Accounting Bulletin (SAB) 92 (SEC 1993). The reporting of environmental issues must improve, since many firms minimize the disclosures not covered by specific accounting pronouncements from the FASB and/or the SEC according to Stanko and Zeller (1995, p. 21). For instance, a recent survey reported that 36% of U.S. companies did not plan to mention current and potential environmental liabilities in their annual reports as required by SAB 92 ("Many companies fail" 1994, p. 12). Only nine percent of the 200 companies surveyed indicated that environmental information would be given significant space in their annual reports presently.

Since the majority of nonfinancial environmental disclosures deal with pollution abatement and other environmental information, it appears appropriate to direct future research into attempting to develop more meaningful reporting practices in the direction of these two theme categories for possible inclusion into the financial section of annual reports. The pollution abatement and other environmental information disclosures also scored lowest in quality in our study, even though they were used frequently by management. It also is evident that more meaningful reporting practices are needed for environmental contingency disclosures in the financial section. These disclosures represent potentially large environmental liabilities for companies. Unfortunately, firms provide presently very limited information to assess adequately the potential impact of these liabilities on a company's future earnings and cash flows. Finally, the environmental information disclosed presently by companies makes it difficult to differentiate between firms with good versus bad environmental performance, particularly in the same industry. It is also apparent that a company's measure of environmental performance is multidimensional, and may require more complex measures than may be available currently.

Appendix

Content analysis uses a set of procedures to make valid inferences from messages (Weber 1990, p. 9). A method of creating and testing a coding scheme is developed similar to other studies (e.g. Wiseman 1982, p. 55; Freedman and Wasley 1990, p. 185; Freedman and Stagliano 1992, p. 114). Coding is the process by which raw data are transformed systematically and aggregated into units that permit precise description of relevant content characteristics (Holsti 1969, p. 94).

The development of a coding scheme begins with the selection and definition of categories (Holsti 1969, p. 95). Besides the selection and definition of categories into which content data are to be classified, the units to be coded must be designated according to Holsti (1969, p. 116). These units include the recording unit, the context unit, and the system of enumeration (Holsti 1969, p. 169). Figure 1 includes the details of the coding scheme used in our study.

(Insert Figure 1 here)

The choice of recording unit is the first step in designating units to be coded, and is the specific segment of content that is characterized by placing it in a given category (Holsti 1969, p. 116). The recording unit chosen for our study is the theme of the environmental disclosure. According to Holsti (1969, p. 116), the most useful unit of content analysis is the theme, but it is also the most time-consuming to develop and code. Weber (1990, p. 22) suggests that this form of coding is labor-intensive but leads to much more detailed and sophisticated comparisons. Wiseman (1982, p. 55) and Freedman and Wasley (1990, p. 185) used the following four broad theme categories: (1) "litigation," (2) "pollution control equipment and facilities," (3) "pollution abatement," and (4) "other environmentally related information." Our study used an updated version of these four categories and a new broad category for "environmental preservation," which includes recycling and conservation of natural resources. These five broad theme categories include specific theme categories as presented in Figure 2.

(Insert Figure 2 here)

Categorizing of a recording unit or theme relies on the context unit, which is the largest body of content that may be examined to categorize a recording unit (Holsti 1969, p. 117). Our study used the financial and nonfinancial sections of the annual report as the context unit. Patten (1992, p. 474) also used these sections of the annual report as the context unit. For the purposes of our study, the financial section of the annual report includes the Management Analysis and Discussion, financial statements, supplemental schedules, and footnotes. The nonfinancial section of the annual report includes the Letter to Shareholders and all other portions of the report not classified as financial.

Freedman and Stagliano (1992, p. 115; 1995, p. 168) suggest that it is more important for a method of content analysis to focus on what is included in the theme, rather than how much is said. The meaning of the message is what is important. Considering only the quantity of disclosures does not convey the importance of the disclosures. Most studies use only quantity as the "system of enumeration." For example, Patten (1991, p. 301; 1992, p. 473) used only a portion of the page (1/100th of a page intervals) containing the disclosure. Blacconiere and Patten (1994, p. 368) used simply the presence or absence of statements relating to five areas of environmental concern as a measurement scheme.

Two systems of enumeration are developed for our study to capture not only the quantity of disclosures but also to establish assessments of the quality. Freedman and Stagliano (1992, p. 115; 1995, p. 168) suggest that the latter are more important since this method of content analysis focuses on what is included in the theme, rather than how much is said.

The first system of enumeration in our study considers the quantity of disclosures made by counting the number of specific theme occurrences (NTO) related to the themes. Each time a specific theme was found it was counted. Blacconiere and Patten (1994, p. 368) used a similar method. This was done to indicate the strength of usage of various themes.

The second system of enumeration in our study considers the quality or perceived information content of the disclosures. Our study uses a four-element quality index (QI) that attempts to assess the quality of disclosures related to the themes similar to a study by Freedman and Stagliano (1992). The four elements used are: (1) effect--significant or not significant, (2) quantification--monetary or not monetary, (3) specificity--specific as to actions, persons, events, or places, or not specific, and (4) time frame--past, present, or future. For our study, significant effects were based on location within the annual report. Those disclosures found in the Letter to Shareholders and financial sections of the annual report were deemed significant. The remaining three elements of each disclosure were judged independently by the researchers and two coders. Each element of the index that is present in the disclosure receives one point. If the disclosure involves the current reporting period, it receives one point. No points are given if the disclosure involves the past or the element is not present. Therefore, each environmental disclosure assessed by the quality index (QI) by theme could receive a minimum of zero points and a maximum of six points based on the four-element index.

In addition, a second quality measure is computed, which is referred to as a disclosure score (DS). The DS is a summation of the quality index (QI) per environmental theme. For example, assume that company X has two observed environmental disclosures assessed for quality by theme (two QIs) for year one. If QI1 and QI2 are rated as five and three points respectively, the DS is eight for year one. This procedure is the same as used by Walden (1993, p. 76) and Walden and Schwartz (forthcoming, p. 9). The type of rating procedure is based on the presence or absence and the degree of specificity for each environmental disclosure theme. This type of differential rating is consistent with other studies (e.g., Wiseman 1982, p. 55; Freedman and Wasley 1990, p. 185; Freedman and Stagliano 1992, p. 114).

The nature and extent of environmental disclosures are examined to measure objectively the information contained in the disclosures and to provide a systematic numerical basis for comparing environmental disclosure themes. Once the ratings are tabulated for quantity and quality assessment, the environmental disclosure themes are compiled by various sections of the annual report for use in further statistical analysis.

References

• American Accounting Association (AAA), Committee on Accounting for Social Performance. 1976. Report of the committee on accounting for social performance American Accounting Association 1975. The Accounting Review Supplement: 38-69.
• American Accounting Association (AAA), Committee on Accounting and Auditing Measurement. September 1991. American Accounting Association committee on accounting and auditing Measurement, 1989-90. Accounting Horizons 5(3): 81-105.
• Benoit, E. June 27, 1989. The Valdez legacy. Financial World 158 (13): 82-83.
• Blacconiere, W. G. and Patten, D. M. 1994. Environmental disclosures, regulatory costs, and changes in firm value. Journal of Accounting and Economics 18 (3): 357-77.
• Boulding, K. E. 1978. The legitimacy of the business institution. In Rationality, Legitimacy, Responsibility: Search for New Directions in Business and Society (E. M. Epstein and D. Votaw, eds.) Santa Monica, CA: Goodyear Publishing, pp. 83-97.
• CEDC reports now available. Oct. 1991. Research Report, p. 3.


• Council on Economic Priorities (CEP). May 1992. Westvaco: A Report on the Company's Environmental Policies and Practices (research and writing conducted primarily by K. Scott) New York: Council on Economic Priorities.
• Early, K. May 1994. The SEC's toxic solution to environmental reporting. Corporate Cashflow 15 (5): 60.
• Epstein, M. J. April 28, 1991. What shareholders really want. The New York Times, Forum 140: 11.
• Epstein, M. J. and Freedman, M. 1994. Social disclosure and the individual investor. Accounting, Auditing and Accountability Journal 7 (4): 94-109.
• Financial Accounting Standards Board (FASB). 1975. Statement of Financial Accounting Standards No. 5: Accounting for Contingencies. Stamford, CT: FASB.
• Financial Accounting Standards Board (FASB). 1986. Accounting Standards: Statements of Financial Accounting Concepts 1-6. Stamford, CT: FASB.
• Freedman, M. and Jaggi, B. 1982. Pollution disclosures, pollution performance and economic performance. Omega 10 (2): 167-76.
• Freedman, M. and Stagliano, A. J. 1992. European unification, accounting harmonization, and social disclosures. The International Journal of Accounting and Education 27 (2): 112-22.


• Freedman, M. and Stagliano, A. J. 1995. Disclosure of environmental cleanup costs: The impact of the Superfund Act. In Advances in Public Interest Accounting (C. E. Lehman, ed.). Greenwich, CT: JAI Press, 6: 163-76.
• Freedman, M. and Wasley, C. 1990. The association between environmental performance and environmental disclosure in annual reports and 10-Ks. In Advances in Public Interest Accounting (M. Neimark, B. Merino, T. Tinker, eds.). Greenwich, CT: JAI Press, 3: 183-93.
• Gamble, G. O., Hsu, K., Kite, D., and Robin, R. R. Sept. 1995. Environmental disclosures in annual reports and 10Ks: An examination. Accounting Horizons 9 (3): 34-54.
• Gray, R. June 1990. The accountant's task as a friend to the earth. Accountancy 105: 65-68.
• Grover, R. May 22, 1989. Fighting back: The resurgence of social activism. Business Week: 34-35.
• Guthrie, J. and Parker, L. D. Autumn 1989. Corporate social reporting: A rebuttal of legitimacy theory. Accounting and Business Research 19: 343-52.
• Guthrie, J., and Parker, L. D. 1990. Corporate social disclosure practice: A comparative international analysis. In Advances in Public Interest Accounting (M. Neimark, B. Merino, T. Tinker, eds.). Greenwich, CT: JAI Press, 3: 159-75.
• Holsti, O. R. 1969. Content Analysis for the Social Sciences and Humanities. Reading, MA: Addison-Wesley Publishing Company.
• Huck, S. W., Cormier, W. H., and Bounds, Jr., W. G. 1974. Reading Statistics and Research. New York: Harper and Row Publishers.
• Ingram, R. W. and Frazier, K. B. 1974. Environmental performance and corporate disclosure. Journal of Accounting Research 18 (2): 614-22.
• Johnson, L. T. Sept. 1993. Research on environmental reporting. Accounting Horizons 7 (3): 118-23.
• Klonoski, R. J. July-Aug. 1991. Foundational considerations in the corporate social responsibility debate. Business Horizons 34 (4): 9-18.
• Many companies fail to disclose environmental liabilities. July 1994. Journal of Accountancy 178 (1): 12-13.
• Ness, K. E. and Mirza, A. M. Sept. 1991. Corporate social disclosure: A note on a test of agency theory. The British Accounting Review 23 (3): 211-17.
• New CEDC reports available. June 1992. Research Report, p. 6.


• Patten, D. M. Winter 1991. Exposure, legitimacy, and social disclosure. Journal of Accounting and Public Policy 10 (4): 297-308.
• Patten, D. M. July 1992. Intra-industry environmental disclosures in response to the Alaskan oil spill: A note on legitimacy theory. Accounting, Organizations and Society 17 (5): 471-75.
• Post, J. E. July-Aug. 1991. Managing as if the earth mattered. Business Horizons 34 (4): 32-38.
• Rockness, J. W. Autumn 1985. An assessment of the relationship between U.S. corporate environmental performance and disclosure. Journal of Business Finance and Accounting 12 (3): 339-54.
• Roussey, R. S. Spring 1992. Practice note: Auditing environmental liabilities. Auditing: A Journal of Practice and Theory 11 (1): 47-57.
• Sanyal, R. N. and Neves, J. S. Dec. 1991. The Valdez principles: Implications for corporate social responsibility. Journal of Business Ethics 10 (12): 883-90.
• Solomon, C. Sept. 19, 1994. Exxon is told to pay $5 billion for Valdez spill. The Wall Street Journal 224 (55): A3, A6.
• Stanko, B. B. and Zeller, T. L. Jan.-March 1995. Environmental liability in financial reporting. Business and Economics Review 41 (2): 19-23.
• Surma, J. P. Fall 1992. Tackling corporate America's environmental challenge. Review 36 (2): 10-23.
• Ullman, A. A. 1985. Data in search of a theory: A critical examination of the relationships among social performance, social disclosure, and economic performance of U. S. firms. Academy of Management Review 10 (3): 540-57.
• U. S. Securities and Exchange Commission (SEC). May 1989. Financial Reporting Release No. 36: Management's Discussion and Analysis of Financial Condition and Results of Operations; Certain Investment Company Disclosure. Washington, DC: Securities and Exchange Commission.
• U. S. Securities and Exchange Commission (SEC). June 1993. Staff Accounting Bulletin No. 92: Accounting and Disclosures Relating to Loss Contingencies. Washington, DC: Securities and Exchange Commission.
• Walden, W. D. 1993. An Empirical Investigation of Environmental Disclosures Analyzing Reactions to Public Policy and Regulatory Effects. Ph.D. dissertation, Virginia Commonwealth University.


• Walden, W. D. and Schwartz, B. N. forthcoming. Environmental disclosures and public policy pressure. Journal of Accounting and Public Policy.
• Weber, R. P. 1990. Basic Content Analysis. Newbury Park, CA: Sage Publications.
• Welch, I. Aug. 18, 1994. Green reports go into decline. Accountancy Age, pp. 8-9.
• Wiseman, J. 1982. An evaluation of environmental disclosures made in corporate annual reports. Accounting, Organizations and Society 7 (1): 53-63.
• Zeghal, D. and Ahmed, S. A. 1990. Comparison of social responsibility information disclosure media used by Canadian firms. Accounting, Auditing and Accountability Journal 3 (1): 38-53.

Table 1. Leading Companies Reported on by the CEDC in Four Industries with SIC Codes (n=57)

Table 2. Descriptive Analysis for Four Industries (n=53)

Table 2A. Descriptive Analysis for Chemical Industry (n=11)

Table 2B. Descriptive Analysis for Consumer Industry (n=11)

Table 2C. Descriptive Analysis for Forest Products Industry (n=16)

Table 2D. Descriptive Analysis for Oil Industry (n=15)

Table 3. Descriptive Statistics for Four Industries (n=53). Disclosure Score (DS) and Performance Measures

Table 3A. Descriptive Statistics for Chemical Industry (n=11)

Table 3B. Descriptive Statistics for Consumer Industry (n=11)

Table 3C. Descriptive Statistics for Forest Products Industry (n=16)

Table 3D. Descriptive Statistics for Oil Industry (n=15)

Table 4. 1989 Spearman Rank Correlations between Disclosure by Section and Environmental Performance for Four Industries (n=53)

Table 4A. 1989 Spearman Rank Correlations between Disclosure by Section and Environmental Performance for Chemical Industry (n=11)

Table 4B. 1989 Spearman Rank Correlations between Disclosure by Section and Environmental Performance for Consumer Products Industry (n=11)

Table 4C. 1989 Spearman Rank Correlations between Disclosure by Section and Environmental Performance for Forest Products Industry (n=16)

Table 4D. 1989 Spearman Rank Correlations between Disclosure by Section and Environmental Performance for Oil Industry (n=15)

Table 5. 1989 Spearman Rank Correlations between Environmental Contingencies and Environmental Performance in the Financial Section by Industry

Table 6. Summary of Significant Positive Differences by Industry

Figure 1. Coding Scheme

Figure 2. Broad and Specific Theme Categories *