Adequacy or inadequacy of accounting information in annual financial reports: methodological analysis.
The authors' study focuses on information needs of (groups of) individuals and organizations both in terms of what is and what should be--such needs must be discovered and analyzed if the services rendered by information reporting are to be improved. The authors' research suggest that information contained in Annual Financial Reports (AFRs) in its present form explain small proportion of variation in market prices of publicity-quoted companies, indicating that the information contained in these reports may be inadequate for the users' needs. Policy-wise, the findings suggest that mere additional disclosure may not always enhance the adequancy of information contained in AFRs--rather, they suggest that the evolution of an efficient accounting information disclosure process requires an evolution of the likely impacts of existing information terms on the one being considered for disclosure.
Keywords: Information needs & reporting; Inadequate accounting information, Efficient accounting information disclosure
JEL Classification: C52; E62; G3I; M41; 055; NI7
The Accounting profession essentially involves the supplying of information to decision-makers. Hence, it is no surprise that accountants have always been involved in many information-related problems. The need for this involvement is underscored by the avowed aim of accounting which, according to the American Accounting
Association's Committee to prepare A Statement of Basic Accounting Theory (ASOBAT), is the provision of relevant economic information to permit informed judgements and decisions by users of the information (1). The attainment of this objective necessarily requires both the identification as well as the disclosure of an 'adequate' amount of information items considered relevant by the users. Consistent with this goal, ASOBAT has predicted that, in future, accounting may well merge with other disciplines with the result that a new information profession will evolve. For this transition to occur, and to enable the accountant to maintain his/her present relative position, San Miguel (2) notes the need for research into the following areas:
* information needs of (groups of) individuals and organizations both in terms of what is and what should be (emphasis added). Such needs, he advised, must be discovered and analyzed if the services rendered by information reporting are to be improved;
* Impact of (accounting) measurements on human factors; and
* improvement of measurement and communication techniques
Although research to date has produced evidence bearing on the last two areas (3&4) no research evidence has, to these writers' knowledge, been reported in respect of the first suggested area.
The importance of research in this area become obvious if one realizes that an ability to provide information relevant to decision-making requires an identification of the users' felt needs. This will, of course require an assessment of how well the existing information structure is satisfying the users' needs, as well as the identification of the information items considered important for decision tasks. It is on these that attention would then be focused regarding efforts at evolving the most appropriate measurement and communication techniques. Hence, research evidence relating to this area will enhance an efficient management of the information supplying function.
The need for evidence bearing on this issue in the Nigerian environment is underscored by the concern expressed by some observers to the effect that in its present form, the published annual financial reports (AFRs) of public companies in Nigeria contain less than adequate information required for the users' needs. Consequently, they have advocated the disclosure of additional accounting information considered relevant to the users' needs. However, there is as yet no established procedure either for measuring information adequacy, or for the determination of the incremental value of any additional information item being considered for disclosure to enhance this adequacy. This article reports the preliminary findings of a suggested framework aimed at filling this gap. Specifically, the paper employs the Burnswick's Lens Model (5) to provide evidence bearing on
i) the extent of adequacy of information, and
ii) the nature and extent of redundancies, contained in published AFRs of public companies in Nigeria in their present form.
For example, in the Nigerian context, Inanga (6) indicated that the information content of published AFRs of public companies are inadequate for the users' needs. He has, therefore, suggested the disclosure of other information such as protected cash flow and future dividend levels.
The discussion above suggests that mere additional disclosure may not always be the only or even the best, solution towards efforts aimed at providing an adequate information set for users needs. Instead, it would be desirable to evolve a well managed information disclosure process based upon the users' felt needs. This will be enhanced through an evaluation of :
* the extent to which the existing information structure is meeting the users' needs,
* the nature and extent of redundancies in the existing structure, and
* the information items considered most relevant by the users.
Modelling Users' Investment Decision Processes
Share investment activities involve making decisions under uncertain conditions. Hence, investors (or prospective investors) decide whether, and how much, to invest in the shares of say, a company after an evaluation of its (foreseeable) future on the basis of a number of cues (information items)' which are probabilistically related to the future (true) state of nature of the given company. A measure of their adequacy for the user's needs is the proportion of variation in users' investment decisions explained by these cues. However, to enable one to determine the relative importance of each cue to the users' decision process, a model is required to represent the underlying decision (6) process. Research in modelling of human information processing has suggested the use of the multiple regression as a model for human use of information (7). It was first recommended as a tool for modelling human judgement process by Brunswick (5) (1940) based on his theoretical interpretation of information processing in the face of uncertainty. The assumption underlying the model is that man must attempt to function effectively in an environment about which he/she receives only equivocal information. The task is to learn to utilize this information in order to infer the current state of the environment and thereby to be able to behave in the most advantageous manner. Multiple regression has been suggested as the model for optimal use of the equivocal information. In particular researchers (8) have indicated that the linear correlation-regression model in probably the most commonly used statistical tool for modelling human judgement processes. It is also the simplest additive model that explicitly provides for correlations among independent variables (cues, or sources of information).
Firstly, interpretation of the results of a regression analysis within the context of this model will elucidate its appropriateness for our purposes here. Let us look at a regression equation of the form
Ye = [B.sub.e1][X.sub.1] + [b.sub.e2][X.sub.2] + ... + [b.sub.en][X.sub.n]
Ye represents the predicted environmental state (in this case the future of a given company), given [X.sub.1], through [X.sub.n] and which represents the cue values from each of the cue sources K = 1 through n, and [b.sub.e]k is an optimal weight determined by the validity of each cue source. Within the context of the Brunswick's model [b.sub.e]k indicates the extent to which the associated cue influenced the users' decisions.
It therefore, provides an insight into the relative importance of each cue. The regression result also will indicate the proportion of variation in users decisions explained by these factors, as a measure of the information set to the users' needs.
Second, the correlation matrix derivable from the regression analysis represents the extent of intercorrelations among the environmental cues. According to Beach, such intercorrelations measure the extent of the cues' redundancy, and could mean a number of things in that they may disclose subjects' (investors) assumptions about supposed redundancy among environmental cues, or they may indicate that investors' use of a particular cue is conditional upon, or modified by, one or more of the latter cues. It should be noted that the latter interpretation of inter-cue correlation is consistent with findings reported by Libby (9).
Finally, a step-wise regression procedure may provide evidence regarding the manner in which the presence of a cue may influence the relevance of another cue in the set to the decision task. Consequently, the multiple regression statistical model is considered appropriate, and hence adopted, for this study.
Methodology Used Operationalization of Users' Decisions and Data Collection Procedure (9)
After an evaluation of the information contained in published AFRs of a set of companies, a prospective investor may decide to invest in the shares of some or none of these companies. Similarly a current investor may decide to sell some of his/her current shareholdings in some companies, or buy additional shares in those companies where he/ she presently has some holdings. The net effect of this network of share investment decisions is reflected in the quoted marker prices of the shares concerned. For example, the investors may decide to buy the shares of the company for which they had projected a rosy future relative to others, on the basis of the information contained in the annual reports. Hence, for the purpose of this study, the net effect of the users' investment decisions is assumed to be reflected in the current market price of the company's shares. Consequently, the average market price of the share of each company in the study's sample constitutes the dependent variable.
To eliminate the distortion due to differences in the nominal values of the shares, only companies with identical par value were used for the study. It also was ensured that any company in the sample had been in continuous existence for not less than ten years, a time period which was considered long enough to enable users to form expectations regarding each company's future. About two hundred AFRs for 1981-82 and 1982-83 were initially collected for the study. Of these, only seventy-nine satisfied the above-stated requirements. It is noteworthy, however, that the selected reports covered all the major industrial groups.
Choice of Cues : The cues required for this type of study should either on theoretical or empirical bases be of direct relevance to the dependent variable. The following cues were deemed relevant to this study for the reasons indicated.
* Earnings per share (EPS): Some studies (e.g. Ball and Brown) (9) have indicated that a few accounting variables, notably earnings, explain the bulk of the story told by all accounting variables. The EPS derived were adjusted for stock dividends and stock splits.
* Dividend per share (DPS): indicates that results of relevant studies are consistent with the 'capital market viewing dividend announcements and changes as providing information pertinent to the evaluation of share prices. It was concluded that at a minimum, corporate forecasts and dividend change announcements also provide information pertinent to the evaluation of share prices.
* Gearing ratio (GR): otherwise called financial leverage: this is defined as the ratio of long-term debt to equity as a measure of the firm's financial risk. Borrowing by a firm while maintaining a fixed amount of equity, increases the risk of the stock to the investor. Hamada further notes that (financial) leverage explained as much as twenty-one to twenty-four per cent of the value of the mean Beta (i.e. the firm's systematic risk).
* Earning power (EP): is conventionally defined as the ratio of net profit after taxes to the capital employed. For this study, an average of net profit after taxes and capital employed over a five-year period for each company was used, since the figure for a single year may not adequately reflect a firm's earning power.
* Cash flow per share (CFPS) : Some writers have advocated the disclosure of cash flow data in published AFRs. This call was predicated on the argument that accounting profitability (earnings) indicate operating performance, whereas cash flow measures signify solvency and financial flexibility For this study, cash flow was defined as the sum of net income less depreciation.
Analysis of Data
In what follows, we present the results of an analysis aimed at providing evidence on this study's subject-matter. In particular, the analysis is supposed to provide evidence bearing on the following questions:
* What proportion of variation in share prices is explained by the above-noted cues extractable from published annual reports?
* What is the extent of intercorrelations among these cues?
* In what manner does the presence of a cue affect the importance of another cue in the determination of share prices?
To provide evidence bearing on the first question, a multiple regression analysis using cross-sectional data was performed, the results of which are reported in Table 1. The associated F-values are shown in patentheses. This equation per se is, however, not of significance, but the adjusted coefficient of determination ([R.sup.2]) is, in view of the following. Correlations among the cues interfere with estimation of regression weights. However, as Gibbins" indicates, taking [R.sup.2] as the evaluative criterion avoids this problem because only the joint performance of all the cues is evaluated. Furthermore, although the estimate of [R.sup.2] is biased upward due to sampling factors, the adjusted [R.sup.2] ([R.sup.2]) is free from this problem. Hence, [R.sup.2] can be regarded as an unbiased estimate of the extent of the adequacy of the cues in influencing share prices.
The reported R of 0.2112 indicates that only about 21 per cent of the variation in determinants of share prices was explained by the chosen cues. By implication, this finding seems to suggest that the information content of the annual reports in its present form is inadequate for the users' needs, as .some writers (e.g. Inanga (6)) alleged. With regard to the second research question, the result of the correlation matrix among the variables used for the regression analysis are shown in Table 2. The results indicate that there exists a high degree of intercorrelation among the following variables: DPS and EPS (0.84), CFPS and EPS (0.82), EP and EPS (0.76), EP and DPS (0.66), EP and CFPS (0.62).
The results suggest a great deal of redundancies in the simultaneous disclosure in annual reports of data relating to earnings, dividends, and cash flow. This idea is buttressed by the fact that EPS, CFPS and EP had an almost identical correlation index of 0.46, and 0.44 and 0.46 respectively with the market price of shares, followed closely by DPS with an index of 0.38. The results further indicate that the gearing ratio data is not significantly correlated with any of the other cues, suggesting that information conveyed by GR is different from those of other cues.
The results contained in Tables 1 and 2 provide a lesson and a guide regarding information disclosure decisions. Although in its present form, the AFRs of public companies in Nigeria do not seem to provide adequate information for users' needs, the disclosure of additional data bearing on future dividends and projected cash flows may not convey any new information. It is no wonder that other researchers (e.g., Ariyo (10)) have indicated that, based on the results of an experimental study, the disclosure of projected cash flow data in AFRs has no significant impact on users' investment decisions.
Table 3 presents the results of the (forward) stepwise regression analysis aimed at providing evidence bearing on the third research question. The results indicate that EPS was considered first by the analysis which, by itself, accounted for about 20 per cent of variation in share prices. The introduction of the EP variable causes a drastic reduction in the F-value of EPS, without a significant change in [R.sup.2]. In Step III, the introduction of CFPS further dissipates the relative importance of EPS, as shown by the F-value, while in step IV, the results indicate that the introduction of GR may even have a negative effect on [R.sub.2]. It also should be noted that the regression model considered DPS as irrelevant to the analysis.
The following are the possible implications of these results. First, as indicated earlier, there is a high degree of redundancies among the cues, especially those relating to earnings and dividends. The results in Table 3, however, indicate that data relating to earnings dominate those relating to dividends. Second, the results suggest that, in trying to eliminate these redundancies, it might be desirable to retain data relating to earning power (EP) which seems to be less affected by the disclosure of other earning-related data. It also will be desirable to evaluate the relative importance of earnings per share (EPS) and cash flow per share (CFPS) data to determine which one dominates the other. Third, although the gearing ratio conveys a message different from those related to earnings or cash flow, it appears that the market places less importance on data relating to the financial risk of the companies concerned.
Fourth, data related to dividends per share (DPS) seem to have no relevance to the determination of share price, a finding which seems puzzling in view of the research evidence bearing on the information content of dividends reported in the literature. However, since the studies reported on market reaction to dividend announcements were conducted in other countries, the results of this study suggest the possible impact of the environment on the relative importance of an accounting information system. It also suggests that results of studies reported on other environments should not be uncritically regarded as being applicable to a different environment. The observed differences may also be due to the fact that studies of market reaction to accounting information do not normally consider the effect of the presence of other competing information items on the" dependent variable.
The concern with the quality and quantity of information contained in AFRs is a reflection of its importance at enhancing the quality of economic decisions of its numerous users. However, like any other conventional product, accounting information should be guided by the basic principles of economics. This requires that not only should its production (supply) be tailored towards the needs of users (demand) as may be determined from time to time, but also should be provided in the most efficient manner. It is then that providers of accounting information can ensure that users do suffer from neither a lack of relevant information nor over-abundance of irrelevant information.
To enhance the adequacy and usefulness of information contained in AFRs, most accounting regulatory bodies and writers have always advocated for data expansion. This orientation, however, seems to overlook the problems of information overload and other limitations of human beings at processing information noted in the literature. It also overlooks the possible existence of redundancies among the cues. Hence, consistent with the suggestion by San Miguel (2), this paper has described a procedure for identifying and satisfying the users' desire for additional information in an efficient manner.
The results suggest that information contained in AFRs in its present form explain small proportion (about 21 per cent) of the variation in market prices of publicly-quoted companies, indicating that the information contained in these reports may be inadequate for the users' needs, as earlier writers opined. The results also indicate that this inadequacy may be attributable to the existence of intercorrelations among and, hence, redundancies in, the information set. A major import of the finding is that, although users may be suffering from a lack of adequate relevant information it is desirable to identify the type of additional information required to avoid an over-supply of redundant information. The study seems to have provided a framework for achieving this goal.
Policywise, the findings suggest that mere additional disclosure may not always enhance the adequacy of information contained in AFRs. Rather, they suggest that the evolution of an efficient accounting information disclosure process requires an evaluation of the likely impacts of existing information items on the one being considered for disclosure. The higher its degree of intercorrelation with those being currently disclosed, the less useful would the new information item be at enhancing the adequacy of information for users' needs. Hence, the cost of such additional disclosure' may be greater than its incremental benefit. The reader is, however, advised to be cognizant of the following major limitations in appraising the study's findings.
First, although their choice was guided by the literature, the chosen cues might not have constituted an exhaustive set of cues available in published AFRs that could have influenced users' decisions. The identification of other eligible cues is therefore left for future research.
Second, the study implicitly assumes that only quantifiable data could influence users' decisions. However, the Libby study (9) has indicated that there are intangible factors which may influence judgements under uncertainty which are not amenable to modelling. The relative importance of these intangibles is positively related to the degree of complexity of the environment. To provide evidence bearing on this issue in' the Nigerian context, future research should examine the functional value of the residual portion of this class of judgement, as Einhorn (13) has suggested.
(1.) American Accounting Association, Committee to Prepare A Statement of Basic Accounting Theory A Statement of Basic Accounting Theory (1966)
(2.) Miguel, S.J.G., Accounting Research in Information Processing and Communication, The Accounting Forum (May 17-24, 1977)
(3.) Ijiri. Y., Jaedickek. R. and Knight. K.E., The Effects of Accounting Alternatives on Management Decisions', in Research on Accounting Measurement: American Accounting Association, (1966)
(4.) Jensen, R. E., An Experimental Design for the Study of Effects of Accounting Variations in Decision Making'. Journal of Accounting Research (4 : 1966)
(5.) Brunswick. E. Thing Constancy as Measured by Correlation Coefficients, Psychological Review (47 : 1940)
(6.) Inanga. E.L., The Information Content of Published Accounts of Nigerian Public Limited Companies, Nigerian Journal of Economic and Social Studies (1976)
(7.) Beach, L. R., Multiple Regression as a Model for Human Information Utilization. Organizational Behaviour and Human Performance (2 : 1967)
(8.) i) Ball. R. and Brown, An Empirical Evaluation of Accounting Income Numbers, Journal of Accounting Research (6 : 1968)
ii) Hamada R. S., The Effect of the Firm's Capital Structure on the Systematic Risk of Common Stocks, Journal of Finance (May 1972)
iii) Gibbins. M., Regression and Other Statistical Implications for Research on Judgment Using Intercorrelated Data Sources, Journal of Accounting Research (20 : 1982)
iv) Einhorn, H.J., Cue Definition and Residual Judgment, Organizational Behavior and Human Performance (12 : 1974)
(9.) Libby. R. Accounting Rations and the Prediction of Failure: Some Behavioral Evidence. Journal of Accounting Research (13 : 1975)
(10.) Ariyo.A., An Experimental Study of the Effect of Projected Cash Flow Data on (Share) Investment Decisions. Nigerian Journal of Economic and Social Studies (1983).
Professor A. SOYODE, Ph.D.
Professor A. ARIYO Ph.D.
Department of Economics
University of Ibadan
(*) Revised text of paper presented at the Fifth Annual Conference under the theme, Managing Nigeria's Economic Systems, organised by the Nigerian Association of Schools of Management Education and Training jointly with the Centre for Management Development and held at the University of Ibada.
The authors own full responsibility for the contents of the paper.
TABLE 1 ADEQUACY OF ANNUAL REPORTS INFORMATION CONTENT TO SHARE PRICE DETERMINATION [R.sup.2] MKpt = 0.467 + 0.697EP + 0.39 CFPS .2112 (2.64) (1.07) + 0.24EPS + 0.066DPS (0.09) (0.01) -0.075GR (0.79) MKpt = Market share price in time EP = Earning power CFPS = Cash flow per share EPS = Earnings per share DPS = Dividend per share GR + Gearing ratio TABLE 2 MATRIX OF CORRELATION COEFFICIENTS AMONG THE CHOSEN CUES MKP EPS DPS CFPS GR EP MKP 1.00 EPS 0.46 1.00 DPS 0.38 0.84 1.00 CFPS 0.44 0.82 0.58 1.00 GR -0.17 -0.14 -0.16 -0.13 1.00 EP 0.46 -0.76 0.66 0.62 -0.14 1.00 TABLE 3 RESULTS OF STEP-WISE REGRESSION ANALYSIS Step Variable Constant B-coefficient F I EPS 0.494 1.26 20.78 II EPS 0.478 0.724 3.01 EP 0.719 2.88 III EPS 0.285 0.246 EP 0.724 0.929 CFPS 0.442 0.393 1.234 IV EPS 0.279 0.236 EP 0.700 2.715 CFPS 0.384 1.173 GR 0.469 0.070 0.820 Coefficient of Determination Step Variable Unadjusted Adjusted [R.sup.2] [R.sup.2] I EPS 021 020 II EPS EP 024 020 III EPS EP CFPS 0.254 0.223 IV EPS EP CFPS GR 0.262 0222
|Printer friendly Cite/link Email Feedback|
|Author:||Soyode, A.; Ariyo, A.|
|Publication:||Journal of Financial Management & Analysis|
|Date:||Jan 1, 2014|
|Previous Article:||Financial management analysis of dividend policy pursued by selected Indian manufacturing companies.|
|Next Article:||Domestic debt and strategic financial management imperatives in Nigeria: causal diagnosis.|