Abstract The attempt to prescribe accuracy criteria for the publication of sales and net-income forecasts not only raises the issue of feasibility but would not supply investors with the essential data relevant to the investment decision. In the rationale of investor models, the fact that forecasts of sales or net income should not err by more than 10% does not greatly advance the decision process unless reinforced by probabilistic estimates of deviations more or less than 10%. Probabilistic budgets for multi-product firms aggregate product line budgets, and as such the distribution of revenue and expense depends upon the covariance relationships within and among the product line estimates. As illustrated, with this in- formation and the major assumptions underlying each product forecast, the investor may manipulate product mixes which conform to his prognostications about the economy and the industry and calculate the net-income effects. In this context, the historical accuracy of company forecasting has marginal value since the investor commands the appropriate information to act upon his own insights and assumptions. Historical accuracy-the dispersion of actual and forecasted sales in prior years- is significant to the analyst only as bench- mark data against which to argue his special projections regarding the future behavior of sales. Thus, if in recent years total or product line sales had never fallen short of the management forecast by more than 6%, the investor might well ponder his rationale for a higher downside variation. Such comprehensive disclosure would also tend to reduce the potential for management manipulation of forecast information, although in particular circumstances short term competitive disadvantages might accrue. Financial reporting has primarily emphasized the verification of past events. The transition to a system which will allow for the prediction of future outcomes merits an extensive research effort into the format and content of budget disclosures. This paper has attempted to highlight some areas of investigation with particular reference to the design of investor oriented budgets (forecasts of sales and net-income). At best, it represents a threshold discussion of an area with far reaching implications for financial reporting.
Clark et al. (Mon,) studied this question.