Method of Moving Averages
Suppose that there are $$n$$ time periods denoted by $${t_1},{t_2},{t_3}, \ldots ,{t_n}$$ and the corresponding values of the $$Y$$ variable are $${Y_1},{Y_2},{Y_3}, \ldots ,{Y_n}$$. First of all we have to decide the period of the moving averages. For a short time series we use a period of 3 or 4 values, and for a long time series the period may be 7, 10 or more. For a quarterly time series we always calculate averages taking 4-quarters at a time, and in a monthly time series, 12-monthly moving averages are calculated. Suppose the given time series is in years and we have decided to calculate 3-year moving averages. The moving averages denoted by $${a_1},{a_2},{a_3}, \ldots ,{a_{n – 2}}$$ are calculated as below:
Years (t)
|
Variable (Y)
|
3-year Moving Totals
|
3-year Moving Averages
|
$${t_1}$$
|
$${Y_1}$$
|
——
|
——
|
$${t_2}$$
|
$${Y_2}$$
|
$${Y_1} + {Y_2} + {Y_3}$$
|
$$\frac{{{Y_1} + {Y_2} + {Y_3}}}{3} = {a_1}$$
|
$${t_3}$$
|
$${Y_3}$$
|
$${Y_2} + {Y_3} + {Y_4}$$
|
$$\frac{{{Y_2} + {Y_3} + {Y_4}}}{3} = {a_2}$$
|
$${t_4}$$
|
$${Y_4}$$
|
$$ \vdots $$
|
$$ \vdots $$
|
$$ \vdots $$
|
$$ \vdots $$
|
$$ \vdots $$
|
$$ \vdots $$
|
$${t_{n – 2}}$$
|
$${Y_{n – 2}}$$
|
$$ \vdots $$
|
$$ \vdots $$
|
$${t_{n – 1}}$$
|
$${Y_{n – 1}}$$
|
$${Y_{n – 2}} + {Y_{n – 1}} + {Y_n}$$
|
$$\frac{{{Y_{n – 2}} + {Y_{n – 1}} + {Y_n}}}{3} = {a_{n – 2}}$$
|
$${t_n}$$
|
$${Y_n}$$
|
——-
|
——
|
The average of the first 3 values is $$\frac{{{Y_1} + {Y_2} + {Y_3}}}{3}$$ and is denoted by$${a_1}$$. It is written against the middle year $${t_2}$$. We leave the first value $${Y_1}$$ and calculates the average for the next three values. The average is $$\frac{{{Y_2} + {Y_3} + {Y_4}}}{3} = {a_2}$$ and is written against the middle years$${t_3}$$. The process is carried out to calculate the remaining moving averages. 4-year moving averages are calculated as:
Years (t)
|
Variable (Y)
|
4-year Moving Averages
|
4-year Moving Averages Centered
|
$${t_1}$$
|
$${Y_1}$$
|
——
|
——
|
$${t_2}$$
|
$${Y_2}$$
|
\[\frac{{{Y_1} + {Y_2} + {Y_3} + {Y_4}}}{4} = {a_1}\]
|
——
|
$${t_3}$$
|
$${Y_3}$$
|
\[\frac{{{Y_2} + {Y_3} + {Y_4} + {Y_5}}}{4} = {a_2}\]
|
$$\frac{{{a_1} + {a_2}}}{2} = {A_1}$$
|
$${t_4}$$
|
$${Y_4}$$
|
\[\frac{{{Y_3} + {Y_4} + {Y_5} + {Y_6}}}{4} = {a_3}\]
|
$$\frac{{{a_2} + {a_3}}}{2} = {A_2}$$
|
$${t_5}$$
|
$${Y_5}$$
|
$$ \vdots $$
|
$$ \vdots $$
|
$$ \vdots $$
|
$$ \vdots $$
|
$$ \vdots $$
|
$$ \vdots $$
|
The first average is $${a_1}$$ which is calculated as \[\frac{{{Y_1} + {Y_2} + {Y_3} + {Y_4}}}{4} = {a_1}\]. It is written against the middle of $${t_3}$$ and $${t_4}$$. The two averages $${a_1}$$ and $${a_2}$$ are further averaged to get an average of $$\frac{{{a_1} + {a_2}}}{2} = {A_1}$$, which refers to the center of $${t_3}$$ and is written against $${t_3}$$. This is called centering the 4-year moving averages. The process continues until the end of the series to get 4-years moving averages centered. The moving averages of some proper period smooth out the short term fluctuations and the trend is measured by the moving averages.
Example: Compute 5-year, 7-year and 9-year moving averages for the following data.
Years |
1990
|
1991
|
1992
|
1993
|
1994
|
1995
|
1996
|
19997
|
1998
|
1999
|
2000
|
Values |
2
|
4
|
6
|
8
|
10
|
12
|
14
|
16
|
18
|
20
|
22
|
Solution:
The necessary calculations are given below:
|
|
5-Year Moving
|
7-Year Moving
|
9-Year Moving
|
|||
Years
|
Values
|
Total
|
Average
|
Total
|
Average
|
Total
|
Average
|
1990
|
2
|
—
|
—
|
—
|
—
|
—
|
—
|
1991
|
4
|
—
|
—
|
—
|
—
|
—
|
—
|
1992
|
6
|
30
|
6
|
—
|
—
|
—
|
—
|
1993
|
8
|
40
|
8
|
56
|
8
|
—
|
—
|
1994
|
10
|
50
|
10
|
70
|
10
|
90
|
10
|
1995
|
12
|
60
|
12
|
84
|
12
|
108
|
12
|
1996
|
14
|
70
|
14
|
98
|
14
|
126
|
14
|
1997
|
16
|
80
|
16
|
112
|
16
|
—
|
|
1998
|
18
|
90
|
18
|
—
|
—
|
—
|
—
|
1999
|
20
|
—
|
—
|
—
|
—
|
—
|
—
|
2000
|
22
|
—
|
—
|
—
|
—
|
—
|
—
|
Example: Compute 4-year moving averages centered for the following time series:
Years |
1995
|
1996
|
1997
|
1998
|
1999
|
2000
|
2001
|
2002
|
Production |
80
|
90
|
92
|
83
|
87
|
96
|
100
|
110
|
Solution:
The necessary calculations are given below:
Year
|
Production
|
4-Year Moving Total
|
4-Year Moving Average
|
2-values Moving Total
|
4-year Moving
Average Centered |
1995
|
80
|
—
|
—
|
—
|
—
|
1996
|
90
|
345
|
86.25
|
—
|
—
|
1997
|
92
|
352
|
88.00
|
174.25
|
87.125
|
1998
|
83
|
358
|
89.50
|
177.50
|
88.750
|
1999
|
87
|
366
|
91.50
|
181.00
|
90.500
|
2000
|
96
|
393
|
98.25
|
189.75
|
94.875
|
2001
|
100
|
—
|
—
|
—
|
—
|
2002
|
110
|
—
|
—
|
—
|
—
|
Cosvan
June 23 @ 3:11 pm
I appreciate the work, but how I can get notes when I need it through my email?
zunaira
October 30 @ 7:36 pm
How to find 2 year moving average
Roger
November 13 @ 8:39 pm
This has been very helpful. But what happens if I don’t have data for certain years?