Uncertainty - Definition and Classification for the Task of Economic Forecasting

The aim of this text is to establish a working definition and classification of uncertainty for the task of economic forecasting. This is necessary in order to arrive at a common understanding of the term, reduce semantic ambiguity and define a clear distinction when it comes to quantifying forecast uncertainty. Two fundamental sources on uncertainty by John Maynard Keynes and Frank H. Knight are reviewed from the perspective of the classification of uncertainty into aleatoric and epistemic, which is a separation of growing use in engineering and machine learning. The concepts of aleatoric and epistemic uncertainty are explored and the possible ambiguity and interaction between them are discussed. Finally, a working definition and classification of uncertainty is laid out and refined for practical use in the context of economic forecasting.