Towards a classification framework for accounting information

Abstract

This thesis is concerned with the classification of accounting information from the recording phase through to the reporting phase. Various criticisms on the way information is presented in financial statements, especially in the balance sheet and income statement ware found in the literature. Classification models for accounting information have been proposed in the literature but the main disadvantages of these models are that they are inherently static in nature and do not embed the notion of time. To further test the various criticisms of present classification structures, 1) a comprehensive literature survey of the criticisms of present accounting classification structures as presented by researchers in the literature was conducted, 2) a questionnaire was prepared and sent out to companies in industry, academics as well as analysts in order to test these criticisms, and 3) an analysis of the financial statements (balance sheet and income statement) of 93 companies listed on The JSE Securities Exchange South Africa (JSE) was undertaken. The outcome of the above three (3) actions led to the following hypothesis: The current classification of accounting information, from the recording phase to reclassification in the reporting phase, does not supply users of such information with the necessary information for decision-making purposes. In this regard a comprehensive classification framework for accounting information is proposed, with the following properties: • A well-defined set of attributes will be used, ultimately to classify a transaction into a static sub-framework to aid decision-making. • Time will be used to classify a transaction at the time of recording and later to reclassify it at the time of reporting. • The proposed framework will guide the classifier as to how an item finds its way into a static structure. Classification has as its step the identification of all the attributes known at the time of recording and reporting of the items implicitly involved in a transaction. These attributes in essence define the items that are to be classified into a structure as well as allowing for relationships to be drawn among the said items. One of the advantages of classification is the creation of (new) knowledge or information. The utility of such information depends on the quality of the classification performed. Normally the responsibility for classification at both the recording and reporting phases rests with the accountant. The resultant classification should provide enough information to users of financial statements, especially financial managers, to allow them to reclassify the given information to suit their own needs. An important influence on a classification framework for accounting information is the various requirements put forward by a variety of users of such information. Some of these requirements may conflict with one another. It is proposed in this thesis that a distributed union of all requirements of users be taken and all conflicting requirements be removed from the union (i.e. simply put all requirements into one group and remove all those that conflict with one another). Classification is then performed for the result. Additional information may be supplied in the statements to cater for the requirements outside the result. The classification framework for accounting information proposed in this thesis is made up of three (3) sub-frameworks as follows: • A normative sub-framework that defines a number of attributes for a transaction and which is based on the recording (past) and reporting (present) phases of accounting information. This framework also takes transactions with future aspects into account • A decision sub-framework which follows directly on the normative sub-framework and which directs the classifier as to where a transaction should be classified in a final, static sub-framework. • A static sub-framework which shows where items implicitly involved in a transaction are finally classified. Classification frameworks are proposed for both the balance sheet and the income statement. The balance sheet framework embodies a temporal component, a decision component and a static structure. A different classification layout for the income statement, including a temporal component and a static structure, is also proposed in this thesis.