Moving Averages and Extrapolation of Trends

(Adapted from An Integrated Approach to Business Studies by Bruce Jewell

4th Edition)

Why Do Businesses Plan?

Forecasting is simply the process of predicting the future behaviour of a variable based upon projections of
future market trends. For example, if the manufacturer of a luxury car forecasted interest rates to decline
over the future months, they might increase production to meet anticipated higher levels of demand.

Forecasting enables a business to plan nearly all of their activities and is typically used in such areas as:

• Production scheduling
• Manpower planning
• Investment appraisal
• Cost projections
• Distribution planning
• Stock control

Forecasting can take many forms but typically this can be divided into two main types:

1. Qualitative Techniques

• Market research
• Personal insight or judgement
• Historical analysis

2. Quantitative Techniques

• Time series analysis

Time Series Analysis Explained

A time series is a set of data recorded over a uniform time period such as a year. It shows how the variable
has behaved over this period of time. The analysis of a time series involves establishing whether a pattern
exists within the data and if it does then there is a basis upon which future trends can be predicted in to the
future.

A time series is a sequence of observations which are ordered in time (or space). If observations are made
on some phenomenon throughout time, it is most sensible to display the data in the order in which they
arose, particularly since successive observations will probably be dependent. Time series are best displayed
in a scatter plot. The series value X is plotted on the vertical axis and time t on the horizontal axis.
 Simple Moving Average

An average (more precisely, an arithmetic mean) is calculated exactly as you would expect; add all the
values together and divide by the number of values. A moving average is a series of averages plotted
against time.

This is not complicated! Let's look at an example to see how it works.

In this example, we're going to construct a 3-day Simple Moving Average (MA). While three days is
too short to be useful in the real world, it illustrates how a simple MA works, and it makes the
calculations obvious.

In this example we will be using the closing price of shares on the stock exchange when calculating a
simple MA. For our example, the stock's closing prices are as follows:

Day # 1 2 3

Price 24 25 26

At the moment, we have only enough information to calculate the average (arithmetic mean) for a single 3-
day period. Add the prices together and divide by 3, the number of price points:

24 + 25 + 26
= 25
3

We don't have a moving average yet, just one single average, a single data point. We need more price
data to start a moving average. Let's augment our original example with the following closing prices:

Day # 1 2 3 4 5 6

Price 24 25 26 27 28 26

Then, calculate the average (arithmetic mean) for each of the 3-day periods:

• SMA for Day (1+2+3) = 25

• SMA for Day (2+3+4) = 26

• SMA for Day (3+4+5) = 27

• SMA for Day (4+5+6) = 27

Day # 1 2 3 4 5 6

Price 24 25 26 27 28 26

SMA 25 26 27 27

A Further Example
 Using closing prices of shares, day 10 is the first day
possible to calculate a 10-day moving average. As the
calculation continues, the newest day is added and the
oldest day is subtracted. The 10-day moving average for
day 11 is calculated by adding the prices of day 2 through
day 11 and dividing by 10. The averaging process then
moves on to the next day where the 10-day moving
average for day 12 is calculated by adding the prices of
day 3 through day 12 and dividing by 10.

The Trend
Definition : The underlying pattern of growth or decline within a pattern of data. Trend is a
long term movement in a time series. It is the underlying direction (an upward or
downward tendency) and rate of change in a time series, when allowance has
been made for the other components.

Moving averages smooth out a data series and make it easier to identify the direction of the trend. Moving
averages will not predict a change in trend, but rather follow behind the current trend. Therefore, they are
best suited for trend identification and trend following purposes, not for prediction.

The trend will experience fluctuations of the following type:

(i) Seasonal Regularly repeated fluctuations associated with seasons of the year,
days of the week or even hours of the day.

(ii) Cyclical Another repetitive cycle but over a medium term period i.e. 5 years. An
example would be the business cycle.

(iii) Random They occur as a result of unforeseen events e.g. war or tax cuts. They are what
is left after the effects of seasonality and cyclical variation are removed from the
trend.

ACTIVITY

On some graph paper draw a graph to represent each of the following patterns of
data:

(i) A constant series

(ii) Seasonality combined with a positive trend
(iii) Sudden impulse purchasing of goods.
(iv) Normal seasonality
 Centred Moving Averages

The moving average itself ‘smooths out’ the data and helps to isolate the trend from the fluctuations that
can influence it.

You are now going to learn how to calculate centred moving averages. These are important and quite
easy to understand. Take the example below:

Quarter Sales Per Quarter Moving Total Moving Average

1 100
2 100
3 120
4 190
1 105
2 140
3 160
4 200

We know how to calculate the moving total from our earlier examples but exactly where do we insert the
information in the table as the even number makes it difficult to find a central value. The answer is that we
place the information between the centre of Quarters 2 and 3. The moving average is also centred and this
is ALWAYS the rule whenever we are dealing with even number periods.

Example

Quarter Sales Per Moving Total Moving Average

Quarter
1 100
2 100
=Sum (100+100+120+190) = Moving Total / 4
3 120
=Sum (100+120+190+105) = Moving Total / 4
4 190
=Sum (120+190+105+140) = Moving Total / 4
1 105
=Sum (190+105+140+160) = Moving Total / 4
2 140
=Sum (105+140+160+200) = Moving Total / 4
3 160
4 200

ACTIVITY

Calculate both the Moving Total and Moving Average for the following data:

Year 1 Year 2

Quarter 1 50 60
Quarter 2 80 90
Quarter 3 90 110
Quarter 4 60 70
If we now wish to compare the original data for the quarter with the identified trend, we have to calculate a
centred moving average. This simply involves averaging the two moving averages shown at the boundary
of each quarter. For example:

Quarter Sales Per Moving Total Moving Average Centred Moving

Quarter Average
1 100
2 100
510 127.5
3 120 128.13
515 128.75
4 190 133.75
555 138.75
1 105 143.75
595 148.75
2 140 150
605 151.25
3 160
4 200

ACTIVITY

You should now be able to calculate the Centred Moving Average for the given data in
your table. Plot the information on to a piece of graph paper and then add
the trend line onto your graph. What do you notice?

Extrapolation

Extrapolation is when the value of a variable is estimated at times which have not yet been observed. This
estimate may be reasonably reliable for short times into the future, but for longer times, the estimate is
liable to become less accurate.

Example
Suppose Angela was 1.20m tall on January 1st 1975, and 1.40m tall on January 1st 1976. By extrapolation,
it could be estimated that by January 1st 1977 she would have grown another 0.20m to be 1.60m tall. This
however assumes that she continued to grow at the same rate. This must eventually become a false
assumption, otherwise by January 1st 1980, she would be a giantess.

ACTIVITY

Using your graph, extrapolate the trend to project the sales for Quarter 2 in Year 3.
Quarter Sales Per Moving Average Centred Moving Seasonal
Quarter Average Variation
1 100
2 100
127.5
3 120 128.13 120 - 128.13
128.75
4 190 133.75 190 – 133.75
138.75
1 105 143.75 105 – 143.75
148.75
2 140 150 140 – 150
151.25
3 160
4 200
What are the problems with this method of forecasting?

Seasonality and Random Fluctuations

It is ok extrapolating the trend into the future based upon the current data, but we have not taken into
account the factors that influence the trend and which could make our forecasts into very expensive
mistakes if we base decisions upon them.

To calculate seasonal variation we simply subtract the trend (CMA) from the original raw data. For
example:

With two or more seasonal variations for a particular time period, we can produce a figure for average
seasonal variation that can be used as the basis for calculating by how much the data will deviate from the
trend.

Example

Add up all of the seasonal variations for Q1 periods and then divide by the total number of Q1s.

Q = Σ Q
Q
RANDOM VARIATION

The final calculation is straightforward. Subtract the average seasonal variation from the seasonal variation
and you are left with the random variation for that period.

ACTIVITY

Calculate the seasonal and average seasonal variation for the data in your table and then
finish by calculating the value of the random variation.

Disadvantages of Moving Average Forecasting Methods

• Forecasting is always an imprecise science and businesses are advised NEVER

to base important decisions upon forecasts alone. You should always question the
extent to which you can forecast the future from the events of the past.

• Moving averages are best suited to short term forecasts in stable economic
environments. The further into the future they are projected or the more unstable the
economic environment then the more imprecise the data.

• Moving averages give equal weight to all values. It could be argued that the most
recent data is more relevant and should be given a higher weighting.

• No account is taken of data that falls outside of the moving average period.

• Large amounts of data are required costing time and money to enable moving
average forecasting to work.