Forecast Correction Using Organizational Debiasing in Corporate Cash Flow Revisioning

Abstract

Corporations that employ information systems, such as decision support systems for judgmental forecasts, have business objectives that require accurate forecasts. But the accuracy of these forecasts is most likely biased by organizational and individual structures within the corporation. These biases, such as revenue targets or personal objectives, may alter the forecasters’ prediction due to financial incentives in a predefined way. This paper argues that model-driven correction of forecasts – which typically utilizes only statistical methods – should incorporate organizational debiasing methods. In a case of an international corporation, local experts forecast cash flows for corporate risk management. The forecasts are later aggregated on a corporate level with subsequent debiasing techniques for decision support. Empirical results show that considering organizational objectives for debiasing techniques can strongly improve forecast accuracy. The total correctable expert error is reduced by up to 60 % for all forecasts of a month, providing better decision support for managers.