Toward an "Events" Theory of Accounting.

Abstract The article examines several problems in the events approach to accounting theory using distinction among terms. Terms used in any science theory may be grouped with reference to the "mode of verification." Verification of a value-inference term would require testing the consistency, necessity, and sufficiency of the inferential logic. The term "event" refers in a general way to an action, occurrence, or happening that could be described by one or more of an infinite number of properties, attributes, or characteristics. A primary need for any accounting theory is a taxonomy that satisfies the criteria of exclusiveness, exhaustiveness, efficiency, and effectiveness. A forecast may relate one kind of event to an expectation of that same kind of event. It can be temporal, as when the antecedent market values of a firm's common stock are used to forecast its subsequent market value. Like extrapolation, prediction may make formal use of statistical techniques. Any accounting observation or inference that is useless for forecasting, is useless for guiding action. If no accounting data were accurate in forecasting, then accounting would be useless for decision. If accuracy in forecasting cannot be established for accounting data, then the claim that accounting is useful for decision is an empty boast.