1. Arrange the data in a table by quarters and years.
2. Calculate the average of quarters for years 1 and 5. Substract 5 from and divide the result by 16. Call it net upward movement over 16 quarters, not 20. And get the average.
3. calculate grand mean for every quarter
4. Elliminate num by adding 1.5 num to Q1, 0.5 to Q2, -0.5 to Q3, -1.5 to Q5 - get the adjusted average.
5. Calculate the mean of quartely grand means.
6. Divide the quartely average by the whole average and get the seasonal pattern
7. Divide original figgures by the seasonal pattern to get the seasonally adjusted data.
year \ Quarte |
Q1 |
Q2 |
Q3 |
Q4 |
Average |
1 |
100 |
70 |
60 |
90 |
80 |
2 |
120 |
85 |
70 |
110 |
|
3 |
150 |
105 |
85 |
130 |
|
4 |
170 |
125 |
100 |
155 |
|
5 |
190 |
150 |
120 |
180 |
160 |
average |
146 |
107 |
87 |
133 |
Grand-118.3 |
Net-ward 06ve0ent = 80, Average num=5
Adjust num |
1.5x5 |
0.5x5 |
-0.5 |
-1.5 |
|
Ratio to gran |
1.29 |
0.92 |
0.71 |
1.06 |
1.0(seasonal |
Estimation |
Q1 |
Q2 |
Q3 |
Q4 |
|
YEAR 1 |
77 |
75.6 |
83.9 |
84.8 |
|
2 |
92.4 |
91.8 |
97.9 |
103.7 |
It is essential to have the same number of quarters in every year. (ie not only first)
Awkward factors like easter
Changes in practice - holidays arrangement with Bank holidays
Do not use for more than 5 years