MODULE 5

TIME SERIES ANALYSIS

TIME SERIES.
The time series is a record of observations or measurements of a variable over a period of
time. Such a record discloses relationship between two variables, one of which being time.
Most of the time series are from the field of Economics and Business. Examples of time
series are yearly national income data, quarterly inventories in an industry, physical volume
of production, yearly bank rates etc.
ANALYSIS OF TIME SERIES
Analysis of time series is a statistical device which can be used to understand, interpret, and
evaluate changes in economic phenomena over time, with the hope of more correctly
anticipating the course of future events.
A time series is a dynamic distribution that reveals good deal of variations over time.
Through analysis of time series, we try to examine the effect of time on various variables.
Changes in time affect the economic, political and social conditions of the economy and these
conditions, in turn, affect statistical data.
Time affects income, habits, fashion, tastes, birth and death rates, production, population etc.
Analysis of time series is done to understand the dynamic conditions of these variables.
Therefore, Time series analysis consists of identifying or determining the various forces or
influences whose interaction produces variations in the time series. (Isolating, studying,
analysing and measuring them independently.
A study about the effect of these forces will help us to understand the past behaviour of the
data, general tendency of the data and thus to forecast future behaviour of the data.

IMPORTANCE OF TIME SERIES ANALYSIS

The analysis of time series is important not only to businessmen and economists, but also to
scientists, social scientists, biologists etc., because of the following reasons:

(1) Time series analysis discloses changes in time and changes in the value of the
variable.

(2) It helps in understanding past behaviour: By observing past data, one can understand
the past behaviour of the variable under study.

(3) It helps in Predicting, Forecasting and Planning: Quantitative data on the past and
present behaviour of an activity, in combination with knowledge about the various
 factors that influence it, make it possible to forecast the activity's future magnitude.
Thus, time series data on a given activity are useful for a comprehensive study of the
factors affecting it, learning of its past behaviours, interpreting its current behaviour
and forecasting its probable future magnitude. Time series analysis can play an
important role not only in making short range estimates for a year or two ahead but
also as a basis for long-range operational planning.

(4) It helps in evaluating current programmes: The actual performance can be compared
with the expected performance and the causes for variations can be analysed.

(5) It facilitates comparison: Different time series are often compared and important
conclusions are drawn from them.

COMPONENTS OF TIME SERIES

The statistical series are usually affected by a multiplicity of causes.
The change in the values of a variable related to time can be the result of a large variety of
factors like changes in the tastes and habits of people, changes in population, reduction in
cost of production, increase in income of people etc. The value of a variable changes due to
the interaction of such forces. These forces are interconnected and cannot be distinguished
easily. The effects of these forces on a time series are called the components of a time series.
These components are
(1) Secular trend
(2) Seasonal variations
(3) Cyclic variations
(4) Irregular variations.

1. Secular trend
Trend refers to long-period changes. It shows the definite and basic tendency of the statistical
data with the passage of time. It is smooth, regular and long-term movement. It refers to the
general tendency of statistical data to rise or to fall or to remain the same. For example, in a
series concerning population or national income, an upward tendency can be noticed while in
data on birth or death or illiteracy a downward tendency.
2. Seasonal variation
Seasonal variations are those variations which occur with some degree of regularity within a
specific period of one year or shorter. Climatic conditions, social customs, religious functions
etc. are the factors responsible for seasonal variations. The prices of rice will go up in the
showing season and will come down during the harvest season. This is the seasonal variation
in the case of price of rice.
 3. Cyclic variation
Cyclic variations are periodic movements. These variations occur at intervals (or periods) of
more than one year. Cyclic movements pass through different stages like prosperity,
recession, depression and recovery. During these different stages, the time series show
changes. These changes are called cyclic fluctuations. They are visible in the case of most of
the business and economic activities.
4. Irregular fluctuations
Irregular fluctuations are those caused by unusual, unexpected and accidental events. Effects
vi earthquake, strike, flood etc lead to irregular variations. These events cause sudden
transition of affairs from one state to another. Irregular fluctuations are random in their
nature. Their occurrence cannot be predicted unlike the other components of time series.

METHODS OF MEASURING TREND

Trend can be measured by the following methods.
(a) Free hand curve method
(b) Method of semi averages.
(c) Method of moving averages
(d) Method of least squares.

1. Free hand curve method (Graphic method)

Freehand curve method is the easiest of all the methods for measuring trend. According to
this method, original data are plotted on a graph paper. Usually, original data when plotted on
a graph gives a wave-like curve, but it depicts a general tendency of the data. The normal
tendency can be either upward or downward. Time is always taken along the horizontal axis.
We draw a smooth freehand line in such a way that it clearly indicates the tendency of the
original data. Such a line will describe the long period movement or the general tendency of
the original data.
Merits:
1. The freehand method is the simplest method of measuring trend.
2. It is purely non - mathematical and involves no calculations at all.
3. Since it is the easiest method, it can be easily understood and used by ordinary people.
4. A freehand curve fitting enables us for the understanding of characteristics of time
series and guides as to which type of mathematical trend will be appropriate.
5. It takes little time to measure the trend.

Limitations
1. This method involves no calculation or mathematical formula; hence the trend values
cannot correctly be obtained.
 2. This method is highly subjective because the trend line depends on the personal
judgment of the investigator. Therefore, different trend lines are obtained from the
same set of data.
3. This method does not give definite trend values and so cannot be relied upon. This
method is unsuitable for further analysis of time series.
4. This method is not suitable where a high degree of accuracy is desired. The method
gives us an approximate picture of the tendency in the long run.

2. Semi - average method

This method is also quite simple and easy to understand. It is employed when freehand curve
method is inadequate to explain the long term tendency. According to this method original
data are divided into two equal parts.
When the number of years is even, there is no problem. But if the number of years is odd, the
total number of years is divided into two groups, omitting the middle year. For example,
when the data given is for 1998 to 2007 (10 year) then the two parts are 1998 to 2002
(5 years) and 2003 to 2007 (5 years).
If the data is for 1998 to 2006 (9 years) the two parts are 1998 to 2001(4 years) and 2002 to
2006 (5 years), 2002 is eliminated to get equal number of years on both parts. The figures
from the two parts are averaged. The average of each part is plotted on the graph against the
middle year of each part. A straight line shall then be drawn to pass through these two points
plotted on the graph. The straight line shows the trend. Trend values can be measured from
this line.
Merits:
(a) This method is quite simple and easy to understand. It requires little time.
(b) There is no scope for deliberate bias or personal prejudice of the person
computing it.
(c) This method is better than freehand curve method because trend values are
definite.
Limitations:
(a) Joining of two points will give us a straight line only and not a curve because this
method assumes a straight line relationship between the two points plotted on the
graph.
(b) This method involves averaging of the data and is supposed to suffer from all
those defects from which an average suffers.

3. Moving Average Method.

Freehand curve method and semi-average method are not suitable where a high degree of
accuracy is desired. The moving average method is an improvement over these two methods.
It is quite simple and is used for smoothing the fluctuations in curves. The trend values
obtained by this method are more accurate.
Moving averages of a time series is a new series obtained by finding out successive items
chosen on the basis of periodicity of fluctuations, dropping off one item and adding the next
item at each stage.
Merits
1. It is quite simple and easy to calculate as compared with the method of least squares.
2. There is no scope for the bias or the personal prejudice.
3. It is most commonly used and in a very long series it is the only practical method.
4. This method is associated with a high degree of accuracy and it can be made the basis
for further analysis of time series.
5. When the period of moving average is equivalent to the period of the cyclic
fluctuations, such fluctuations are completely eliminated.
Demerits
(1) Trend values cannot be computed for all the years. The moving averages for the first
few years and last few years cannot be obtained. It is often these extreme years in
which we may be interested.
(2) The selection of proper period is a great difficulty. If a wrong period is selected, there
is every likelihood that conclusions may be misleading.
(3) Since the moving average is not represented by a mathematical function, this method
cannot be used for forecasting.
(4) It can be applied only to those series which show periodicity.

4. Method of least squares (Curve fitting by mathematical equations)

For many types of data it is better to obtain trend values by fitting a mathematical curve than
by finding out the moving average. If the increase or decrease in the values of a particular
series is of equal absolute amount year after year or if the increase or decrease is always a
constant percentage, it is better to establish mathematical equations and to find trend based on
the values obtained by such equations. In many studies relating to economic phenomena, the
data conform to definite laws of growth or decline, and in such cases, the mathematical curve
gives the best possible trend values.