Forecasting with Indices

Accounting

Forecasting Summer Demand Using Indices

Assessing the inventory needs for a forthcoming may be undertaken using a number of methods. Some managers will simple guess, using a mixture of experience and judgment, which is likely to be a weak approach (Shmueli, 2012). A different method of forecasting is to use the results from the previous year, adapting it using judgment (Shmueli, 2012). However, this also has the potential to be erroneous, as the results may vary on a year to year basis, especially if the firm faces a volatile demand. Where there is volatile demand, an effective method of forecasting may be to make use of statistical assessment using the results from all the available previous years, looking at the different seasons or months (O'Connell and Koehler, 2004). No method is likely to be completely accurate, but basing the forecast on the previous few years increases the potential for an acceptable forecast.

The data supplied indicates that the demand falls and rises over time. To assess the overall volatility and identify any potential patterns the data can be placed on a graph showing demand for each month for each of the four years provided (O'Connell and Koehler, 2004).

Figure 1; Monthly demand over the four years

Figure 1 demonstrates that there is a seasonal demand, peaking during months 4-6 and showing the lowest demand between month 11 though to month 1 or 2 of the following year. The pattern indicates that it is not consistent; there are rises and decreases of different amounts in different years. The importance of seeing this pattern tells the forecaster that rather than looking at annual figures or even quarterly figures should look at the monthly figures in order to make an assessment.

The development of an index which can be used to forecast demand may be very useful. The index is created by taking the mean demand for each month over the four-year period to use as the base line (1), the actual demand over the period of time may them be calculated against that base line, showing the demand as a proportion of that mean. This is shown in table 1.

Table 1; Calculations to create the index

1

2

3

4

Month

Average

Year 1

Year 2

Year 3

Year 4

1

39,600

0.45

1.14

1.51

0.90

2

37,080

0.53

1.25

0.83

1.38

3

30,000

0.52

0.74

1.59

1.15

4

59,210

0.91

0.70

1.25

1.15

5

64,375

1.29

0.71

0.94

1.06

6

57,750

1.26

0.72

0.96

1.06

7

47,370

1.17

0.84

0.68

1.32

8

56,638

1.01

1.13

0.68

1.17

9

29,855

0.52

1.59

0.84

1.05

10

39,638

0.70

1.09

1.29

0.92

11

27,323

0.78

1.44

1.16

0.61

12

19,350

0.88

0.53

1.61

0.98

With the creation of the index for each moth, this may then be used to assess the most likely demand. The most appropriate method is the use of the least square regression. This uses the data from the previous years and places them on a graph, drawing a straight line through the points so it has the least distance from the different points. The future forecasts are assumed to be on this line. The equation for the line can be used to calculate forecasts. The graph for month 1 is shown below in figure 1.

Figure 2; Graph for Month 1 demand