Forecasting Technique Part3

ACCA F2 - Management Accounting


Time series is sequence of numbers collected at regular interval over a period of time. Time series analysis records changes in value of economic variable over a period and often is used where; there are no relative independent variables. This is done by extrapolation. Extrapolation is the extension of values beyond the known (sequence of numbers) range.

Customized synopsis form the maltamanagement article (link provided below)

Four components of variation in time series are

Ø  (T) The trend,  =  E.g. linear trends, logistic trends, compound interest trends

Ø  (S) The seasonal components,  = E.g. revenue/cost trends for seasonal components

Ø  (C) The cyclical component, = E.g. variation caused by business life cycle (introduction, growth, maturity and decline)

Ø  (R) The residual (or irregular/random) component, = E.g. unpredictable trends like sudden boom and bust of market or a fall in product demand

Different model of forecasting:

1.       In additive model the variable Y is given by

Y = T + S + C + R                                 for C and R is equal to Zero          Y = T + S

2.       In multiplicative model the variable Y is given by

Y = T * S * C * R                                 for C and R is equal to Zero          Y = T * S

3.       Forecasting linear trend:

Use regression analysis

4.       Forecasting seasonal components

S = Y - T                                                                 S = Y /  T

5.       Residual component

R = Y - T - S - C                                    R = Y / (T * S * C)

6.       Moving average: It is used where the trend is not linear. The moving average approach require trend extrapolations using judgement before getting into future. The moving average is adjusted for seasonal and cyclical variances using the expertise of the users.

Interpolation is the stretching of values within the known (sequence of numbers) range.