Step Function Model For Forecasting Project Cash Flow

Cash flow forecasting for projects has assumed a special interest among researchers for the past 30 years or more. Due to their peculiar execution cycle, projects follow a sigmoid or ‘S’ curve pattern in their progress from start to completion. Many researchers have developed different empirical and mathematical models and also computer programs that forecast the cash flows by using a variety of characteristic ‘S’ curves for different project activities and incorporating related time lags between expenditure and cash flow. However, it must be noted that the cash flows do not occur on a continuous basis but are discrete and occur at specific instances. Therefore mathematically, cash flows are not continuous functions but are step functions and any projection made on the assumption of them being continuous functions would be inaccurate and therefore unreliable in practice. While modeling cash flows as continuous functions may be a mathematically reasonable approximation, it tends to average out the values occurring at different times by a smooth curve and may hide the likely cash deficits that may occur between the two periods. This difference cannot be ignored since cash flow shortfall, even for an extremely short duration of a few days, can have acute real consequences on a project. The paper critically examines various mathematical approaches by different researchers, and suggests a novel approach for developing a step function model for discrete cash flow forecasting for projects. This lays a foundation for further work that needs to be done to develop the actual equations based on project data and validate the approach for practical use.