Forecasting Finance professionals need to learn efficient and effective data forecasting methods in order to make effective decisions Almost all managerial decisions are based on forecasts of future conditions Forecasts are needed throughout an organisation – and they should certainly not be produced by an isolated group of forecasters Forecasting is never “finished” Forecasts are needed continually, and as time moves on, the impact of the forecasts on actual performance is measured, original forecasts are updated, variance analysis assessed and decisions modified, etc.
Forecasting considerations Managers are required to make decisions under uncertainty about the future In order to make those decisions, it is necessary to forecast key variables The choice of forecast models can have a significant impact on the accuracy of forecasts It is necessary to understand forecasting methods (and their limitations) in order to make reliable and timely business decisions
Rolling forecast: overview Typically, a 12-month budget which is prepared and revised on a regular basis during the year Applications: 12-month rolling forecasts in material pricing Weekly projections for cash-strapped companies
Regression analysis: overview Establish the linear relationship between variables Predict the value of the dependent variable from one (or more) independent variables Example applications: Predict sales from advertising Predict consumption from income
Moving average: overview Average of data points from a specified number of consecutive periods Moving average is “updated” when new information becomes available Applications: Moving average cost (inventory costing method)
Weighted moving average: overview There are two types: Weighted Moving Average (WMA) and Exponentially Weighted Moving Average (EWMA) They are similar to the Simple Moving Average (SMA), but they assign more weight to recent observations than older observations WMA assigns more weight to recent events than SMA, and EWMA assigns more weight to recent events than WMA
Assessing accuracy of forecasts Forecast errors represent differences between actual values and the estimated values We need to analyse them to determine the accuracy of our forecasts Measures of Forecasting Errors Mean Squared Error (MSE) Mean Absolute Deviation (MAD) Cumulative Forecast Error (CFE) Mean Absolute Percentage Error (MAPE)
Difficulties in forecasting Often we do not know the underlying nature of our data (e.g. linear or non-linear) Forecasts made on the basis of historical data may be biased and not forward-looking It may be difficult to choose appropriate forecasting models It is imperative to validate the usefulness of the models we use, and to test their appropriateness to our business data