UNIT 6 INDEX NUMBERS 

Structure
6.0 Objectives
6.1 Introduction
6.2 Steps in Construction of Index Numbers
6.2.1 Selection of Base Period
6.2.2 Choice of a Suitable Average
6.2.3 Selection of Items and their Numbers
6.2.4 Collection of Data
6.3 Method of Construction of Index Number
6.3.1 Relative Methods
6.3.2 Aggregative Methods
6.3.3 Quantity or Volume Index Numbers

6.4 Merits of the Various Aggregative Measures

6.5 Tests for Index Numbers
6.5.1 Time Reversal Test
6.5.2 Factor Reversal Test
6.5.3 Chain Index Number and Circular Test

6.6 Cost of Living Index Number (CLI) or Consumer Price Index Number (CPI)
6.7 Worked-Out Examples
6.8 Let Us Sum Up
6.9 Answers or Hints to Check Your Progress Exercises

6.0 OBJECTIVES
After going through this Unit, you will be able to:
 define index numbers; and
 Construct and calculate them.

6.1 INTRODUCTION
An “index” in the common sense of the word is an “indicator” and not anything
more than that. “Index numbers” or “indices” are forms of the plural, but they all
mean the same thing.
An index number represents the general level of magnitude of the changes
between two (or more) periods of time or places, in a number of variables taken
as a whole. In this definition, the world “variable” refers to numerical variables
which can be measured in quantity, such as the prices of commodities.


Adapted from IGNOU study material of EEC 13: Elementary Statistical Methods and Survey
Techniques, Unit 10 written by J Roy with modifications by Kaustuva Barik
Summarisation of For example, we may like to compare the price level of an article between 2010
Bivariate and Multi-
variate Data and 2020 or between Mumbai and Kolkata. Let us consider the yield of rice in
2015 and in 2020 as 50,000 and 60,000 tons respectively. The year 2015 is taken
as base for comparison of yields that is 2015 = 100. The corresponding figure for
,
2020 will be , × 100 = 120. This is a single-commodity index number in its
simplest form, being just a relative number. In practice, however, we deal usually
with a number of commodities for the construction of an index.
Index numbers are ratios that are usually expressed as percentage in order to
avoid awkward decimals. Thus if one commodity costs Rs. 45 in 2019 and Rs.
150 in 2020 the ratio would be or 3.33. If instead of this, we express the ratio
into a percentage × 3.33, we say that the index is 333, based on 2019, which
is 100.

6.2 STEPS IN CONSTRUCTION OF INDEX NUMBERS

Many government and private agencies are engaged in computation of index
numbers or indices as they are often required for the purpose of forecasting
business and economic conditions, providing general information, etc.
It is not always the case that the comparison should be over time, but most
common types of index numbers measure changes over time. Similarly, index
numbers may be constructed for studying changes in any variable such as
intelligence, aptitude, efficiency, production, etc. However, the time series of
prices is frequently used. Therefore, in the discussion below we will concentrate
mostly on prices of commodities. The principles of construction are, however,
quite general in nature, and may thus be applied to other areas of interest.
There are various uses of price index numbers. The wholesale price index
number indicates the price changes taking place n wholesale markets. In the other
hand, the consumer price index number or the cost of living index number tells us
about the changes in the prices faced by an individual consumer. Its major
application is in the calculation of dearness allowance so that real wage does not
decrease; or in comparing the cost of living in, say, different regions. It is also
used to measure change in purchasing power of money. The reciprocal of a
general price index is known as purchasing power of money with reference to the
base period. For example, if the price index number goes up to 150, it means that
the same amount of money will be able to purchase 100/150 = 0.67 times or 67%
of the volume of goods being purchased in the base period.
6.2.1 Selection of Base Period
Since index numbers measure relative changes, they are expressed with one
selected situation (e.g., period, place, etc.) as 100. This is called the base or the
starting point of the series of index numbers. For example, a date is first chosen
and all changes are measured from it. The base may be one day such as with
index or retail prices, the average of a year, or the average of a period. While
selecting a base period the following aspects should be taken into consideration.
130
 1) The base period must be “normal” in the sense that the data chosen are Index Numbers

not affected by any irregular or abnormal situations such as natural

calamities, war, etc. It is desirable to restrict comparisons to stable
periods for achieving accuracy.
2) It should not be too back-dated as the patterns of trade, imports or
consumer preferences may change considerably if the time-span is too
long. A five to ten year interval is likely to be suitable for one base date,
and after that the index becomes more and more outdated. Greater
accuracy is attained for moderate short-run indices than for those
covering greater span of time.
3) For indices dealing with economic data, the base period should have some
economic significance.
6.2.2 Choice of a Suitable Average
An index number is basically the result of averaging a series of data (e.g., price-
relatives of several commodities). There are, however, several ways of a
averaging a series: mean (i.e., arithmetic mean), mode median, geometric mean
and harmonic mean.
The question naturally arises as to which average to chose. The mode has the
merit of simplicity, but may be indefinite. The median suffers from the same
limitations. Moreover, neither of them takes into account the size of the items at
each end of a distribution. The harmonic mean has very little practical
applications to index numbers. As a result, mode, median and harmonic mean are
not generally used in the calculation of index numbers. Thus, the arithmetic mean
is most commonly used. However, the geometric mean is sometimes used despite
its slight difficulty in calculation.
6.2.3 Selection of Items and their Numbers
The number and kinds of commodities to be included in the construction of an
index numbers depend on the particular problem to be dealt with, economy and
ease of calculations. Various practical considerations determine the number and
kinds of items to be taken into account. For a wholesale price index, the number
of commodities should be as large as possible. On the other hand, for an indexed
meant to serve as a predictor of price movement rather than an indicator
movement rather than an indicator of changes over time, a much smaller number
of items may be adequate. Care should, however, be taken to ensure that items
chosen are not too few which make the index unrepresentative of the general
level. A fixed set of commodities need not also be used for a very long period as
some items lose their importance with the passage of time and some new items
gain in significance. In general, the commodities should be sensitive and
representative of the various elements in the price system.

131
Summarisation of 6.2.3 Collection of Data
Bivariate and Multi-
variate Data As prices often vary from market to market, they should be collected at regular
intervals from various representative markets. It is desirable to select shops
which are visited by a cross section of customers. The reliability of the index
depends greatly on the accuracy of the quotations given for each constituent item.

6.3 METHOD OF CONSTRUCTION OF INDEX

NUMBER
Various methods of construction of index numbers are as follows:

1) Relative methods

a) Simple average of relatives

b) Weighted average of relatives

2) Aggregative methods

a) Simple aggregative formula

b) Weighted aggregative formula

i) Lasperyres’ index

ii) Paasche’s index

iii) Edgeworth-Marshall’s index

iv) Fisher’s Ideal index.

6.3.1 Relative Methods
If we record prices of a variety of commodities at a given date and at a later date
record the prices of similar items, the change in price can be simply expressed as
a percentage of the new compared with the old for each commodity.
This provides us with price relatives and if weights are available the next step
will be to multiply the relatives by the weights. Finally, an index number can be
produced if we add together the weighted relatives and calculate an average. It is
unrealistic to assume that the consumption of each commodity has been equal. So
most indices take account of the proportions of each item actually used. This
method to weighting shows the relative importance of each in the series.
Given k commodities with base year prices of
Po1, Po2,. . . . Pok,
and current prices of
Po1, Po2,. . . . Pok,
the price relative for the ith commodity will be where i = 1, 2, ……, k and the
subscript 0 refers to the base year and subscript n refers to the current year.

132
a) Simple Average of Relatives Index Numbers

The arithmetic mean of the price relative is given by

Index = 100 ∑ …(6.1)

For simplicity we can omit the subscript ‘i' and write

index = 100 ∑

b) Weighted Average of Relatives

The most suitable weights to use are the value of each item, which is denoted by
wi for the i-th commodity. We may use the value of the base year quantities sold
at the base year prices (wi = poi qoi) or current year quantities sold at current
prices (w1i = p1i q1i)
If we omit the subscript ‘i’ for simplicity, a weighted arithmetic mean of price
relatives using base year values as weights is given by
∑ ×
index = ∑
× 100 …(6.2)

You should note that the base year weighting preserves continuity, but it loses
“up-to-dateness” in the course of time.
Example 6.1: The table below presents the average fares per railway journey.
Using 2010 average = 100, calculations are made according to base year weights.

Class of No. of passenger Fare Weights Price 𝑃. 𝑤

Ticket journeys in 2010 relatives
𝑤 =𝑝 𝑞
(in millions) 2010 2020 𝑃
P= . 100
𝑃

Full Fare 23 12 60 276 500 138000

Excursions 25 6 30 150 500 75000
Festival 20 4 15 80 375 30000
Season 32 5 14 160 280 44800
tickets
Total 666 287800

Applying formula (6.2), we get

Index for 2020 = = 432.13.

Using Current year values (𝒘𝒏 = 𝑷𝒏 𝒒𝒏 ) as weights, the index is given by

𝑃
∑
𝑃 ×𝑤
index = × 100 … (6.3)
∑𝑤

133
Summarisation of Example 6.2: The table below shows the average fares per railway journey.
Bivariate and Multi-
variate Data Using 2010 average = 100, calculation are made according to current year
weights.

Class of No. of Fare Weights Price 𝑃. 𝑤

Ticket passenger relatives
journeys in 2010 2020 𝑤 = 𝑝 𝑞 𝑃
P= . 100
2010 (in 𝑃
millions)

Full Fare 25 12 60 1500 500 750000

Excursions 26 6 30 780 500 390000
Festivals 9 4 15 135 375 50630
Season-
27 5 14 378 280 105800
tickets
Total 2793 1296430
Applying formula (6.3), we get
1296430
index = = 464.17
2793
6.3.2 Aggregative Methods
In this method, the aggregate (sum-total) of the prices of all commodities in the
current or given year is expressed as a percentage of the same in the base year.
Thus, in the case of simple aggregative index, we have:
aggregate prices in current year
index number = × 100
aggregate prices in the base year

𝑃 + 𝑃 + … … … .. 𝑃
= × 100
𝑃 + 𝑃 + … … … .. 𝑃
∑𝑃 ∑𝑃
× 100 = × 100 … (6.4)
∑𝑃 ∑𝑃
where the summation symbol ∑ extends over all the selected commodities
numbering k. On the other hand, in the case of weighted aggregative index we
have,

𝑝 + 𝑝 𝑞 + … … … .. 𝑝 𝑞
General index = × 100
𝑝 𝑞 + 𝑝 𝑞 + … … … .. 𝑝 𝑞

∑𝑝 𝑞
× 100
∑𝑝 𝑞
∑𝑝 𝑞
or simply = × 100 … (6.5)
∑𝑝 𝑞
The weights used should be actual quantities bought or sold, and these are kept
unchanged until such time as the index requires to be revised.
134
There are many formulae for weighted aggregative index, but depending on the Index Numbers
type of weights used, we discuss four indices which are commonly used.
a) Laspeyres’ index
If we use base period quantities (q0) as the weights in the general weighed
aggregative index formula (6.5), we get what is known as Laspeyre’s formula
(L).
∑
𝐿 = ∑
× 100 … (6.6)

It can be seen that this index has fixed base year quantity as weights (𝑞 ) and is
equivalent to arithmetic mean of price relatives given at formula (6.2). Thus, we
can also write (6.6) as
∑ ×
𝐿 = ∑
× 100

b) Paasche’s index
If we use current year quantities (𝑞 ) as weights in the general aggregative index
formula(6.5), we get what get what is known as Paasche’s formula(P).
∑
𝑃 = ∑
× 100 … (6.7)

Where 𝑞 (actually 𝑞𝑛1 , 𝑞𝑛2 , … . . 𝑞𝑛𝑘 )are the quantities bought or sold in the
current period.
c) Fisher’s Ideal Index
An index number obtained as geometric mean (i.e., square root of the product) of
indices obtained by Laspeyres’ and Paasche’s formulae, satisfies certain
important properties (to be discussed later), is known as the Fisher’s ideal
formula.
∑ 𝑝𝑛 𝑞0 ∑ 𝑝𝑛 𝑞𝑛
𝐹 = √𝐿 × 𝑃 = ∑ 𝑝0 𝑞0
×∑ × 100 … (6.8)
𝑝0 𝑞𝑛

d) Edgeworth-Marshall Index
If the mean of the base period and the current period quantities is used as weight,
i.e.,
𝑤 = (𝑞0 + 𝑞𝑛 ), we get a compromise formula of ‘Edgeworth-Marshall index’.
∑ 𝑝 (𝑞 + 𝑞 )/2
𝐼 = × 100
∑ 𝑝 (𝑞 + 𝑞 )/2

∑ ( )
= ∑
× 100 … (6.9)
( )

We take some hypothetical data and calculate the above indices from it (see
Table 6.1).

135
Summarisation of Table 6.1: Illustrative Calculation of Laspeyres’
Bivariate and Multi-
variate Data Edgeworth-Marshall]s and Fisher’s Indices

Item Base Year Current Year

(𝑝 𝑞 ) (𝑝 𝑞 ) (𝑝 𝑞 ) (𝑝 𝑞 )
Price Quantity Price Quantity
(𝑝 ) (𝑞 ) (𝑝 ) (𝑞 )

A 20 7 25 9 140 175 180 225

B 42 6 40 8 252 240 336 320
C 30 17 25 4 510 425 120 100
D 8 15 14 10 120 210 80 140
E 10 8 13 5 80 104 50 65
Total 1102 1154 766 850
From the above table we calculate various indices as follows:
∑ 𝑃𝑛 𝑞0
1) Lasperyres’ price index = ∑ × 100 = × 100 =
𝑃𝑛 𝑞0

104.72 = 105

∑ 𝑃𝑛 𝑞𝑛
2) Paasche’s price index = × 100 = × 100 =
𝑃0 𝑞𝑛

110.97 = 111

∑ 𝑃𝑛 𝑞0 +∑ 𝑃𝑛 𝑞𝑛
3) Edgewroth-Marshall’s price index = ∑ ×
𝑃0 𝑞0 +∑ 𝑃0 𝑞𝑛

100 = × 100 =

× 100 = 107.28 = 107

∑ 𝑃𝑛 𝑞0 ∑ 𝑃𝑛 𝑞𝑛
4) Fisher’s ideal index = ∑ 𝑃0 𝑞0 ∑ 𝑃0 𝑞𝑛
× 100 =

[(𝐿) × (𝑃)] =
(104.72 × 110.97) = 107.8 =
108
Note that for the same price change, different formulae provide different values.
Moreover, when prices are increasing, Laspeyres’ index gives the lowest value
while Paasche’s index gives the highest value. Therefore, it is often said that
Laspeyres’ index is an under-estimate while Paasche’s index is an over-estimate
of true price change.
6.3.3 Quantity or Volume Index Numbers
We can get a quantity or volume index number, which measures and permits
136
comparison of quantities of goods, from corresponding price index number Index Numbers
formulae simply by replacing p. by q and q and q by p.
1) Quantity relative = × 100

2) Arithmetic Mean (A.M.) of quantity of relatives = 100 ∑ /𝑘

3) Weighted A.M. of quantity relative index:

∑ 𝑞𝑛 /𝑞0
a) Base year weights: ∑
× 100 where 𝑤0 = 𝑝0 𝑞0
∑ 𝑞𝑛 /𝑞0 ×𝑤𝑛
b) Current year weights: ∑ 𝑤𝑛
× 100 where 𝑤𝑛 = 𝑝𝑛 𝑞𝑛
∑ 𝑞𝑛
4) Simple aggregative quantity index = ∑ × 100
𝑞0
∑ 𝑞𝑛 𝑃 0
5) Laspeyres’ quantity index = ∑ × 100
𝑞0 𝑃 0
∑ 𝑞𝑛 𝑃𝑛
6) Paasche’s quantity index = ∑ × 100
𝑞0 𝑃𝑛

∑ 𝑞𝑛 𝑃0 ∑ 𝑞𝑛 𝑃𝑛
7) Fisher’s ideal index = ∑ 𝑞0 𝑃0 ∑ 𝑞0 𝑃𝑛
× 100
∑ 𝑞𝑛 (𝑃0 +𝑃𝑛 )
8) Edgeworth-Marshall’s index = ∑ 𝑞0 (𝑃0 +𝑃𝑛 )
× 100

Check Your Progress 1

1) What do index numbers seek to measure?
………………………………………………………………………………....
……………………………………………………………………………....…
…………………………………………………………………………………
…………………………………………………………………………………
……………………………………………………………………………....…
..........................................................................................................................

2) Discuss the various problems involved in construction of index numbers with

particular reference to price indices.
………………………………………………………………………………....
……………………………………………………………………………....…
…………………………………………………………………………………
…………………………………………………………………………………
……………………………………………………………………………....…

3) The following are the prices of six different commodities for 2020 and 2021.
Compute the price index by (a) aggregative method, and (b) average of
price relatives method by using arithmetic mean.
137
Summarisation of
Commodities Price in 2020 (Rs.) Price in 2021 (Rs.)
Bivariate and Multi-
variate Data
A 40 50
B 50 60
C 20 30
D 50 70
E 80 80
F 100 110
………………………………………………………………………………....…
…………………………………………………………………………....………
……………………………………………………………………………………
……………………………………………………………………………………
……………………………………………………………………………………

4) Calculate Fisher’s Ideal Index Number from the following group of items.
Base Year Current Year
Item No. Price Quantity Price Quantity
(in Rs.) (in Kg.) (in Rs.) (in Kg.)
1 4 1.0 3 4
2 8 1.5 7 5
……………………………………………………………………………………
……………………………………………………………………………………
……………………………………………………………………………………
……………………………………………………………………………………
……………………………………………………………………………………

5) Calculate Laspeyres’ and Paasche’s Index Number from the following data.

Base Year Current Year

Item Quantity Price Quantity Price

Bread 6.0 40 7.0 30

Meat 4.0 45 5.0 50

Tea 0.5 90 1.5 40

……………………………………………………………………………………
……………………………………………………………………………………
……………………………………………………………………………………
……………………………………………………………………………………
138
 Index Numbers
6.4 MERITS OF THE VARIOUS AGGREGATIVE
MEASURES
The different index numbers serve different purposes and, therefore, the
appropriateness of a particular index number depends on the purpose at hadn.
The Laspeyres’ index calculation is simpler, since this uses the base period
quantities as weighs which are not difficult to get and the denominator needs
calculating only once. But in this index a rise in pries tends to overstated, since
it does not take into account corresponding falls in demand or changes in output.
Indices such as Paasche’s on the other hand, use current period quantities as
weights which are difficult o get and the weights need to be constructed afresh
for every year. Moreover, Paasche’s index tends to understate the rise in prices
because it uses current weights. The Laspeyres’ index is probably more
commonly used, since it is convenient to employ fixed weights. But with the
passage of time the weights are rendered out of date. For example, in 1995 the
number of mobile phones in Odisha was nil. In 2020 there are more mobile
phones than the number of land line connections. The Paasche’s index uses the
preferable current weights, but since-up-to-date information on quantity of goods
produced or consumed or marketed or distributed are not readily obtained, the
Laspeyres’ index has a great advantage.

6.5 TESTS FOR INDEX NUMBERS

A perfect index number, which measures the change in the level of a
phenomenon from one period to another, should satisfy certain tests. There are
three major tests of index numbers: (1) Time reversal test, (2) Factor reversal test,
and (3) Circular test.
6.5.1 The Time Reversal Test
According to this test, if we reverse the time subscripts (such as 0 and n) of a
price (or quantity) index the result should be the reciprocal of the original index.

Symbolically,

I0n × I0n = 1

where I0n = index number for period n with the base period 0.

In0 = index number for period 0 with the base period n.

If from 2010 to 2020 the price changes from Rs. 4 to Rs. 16, the price in 2020 are
400 percent of the price in 2010, and the price in 2010 is 25 percent of the price
in 2020. The product of the two price relatives is 4×0.25 = 1. The test is based on
the analogy that the principle, which hold good for a single commodity, should,
also be true for the index number as a whole.

139
Summarisation of There are five methods which do satisfy the time reversal test. These are:
Bivariate and Multi-
variate Data 1) Simple geometric mean of price relative
2) Aggregative indices with fixed weights
3) Edgeworth-Marshall formula
4) Weighted geometric mean of price relatives if fixed weights are used
5) Fisher’s ideal index
∑ 𝑃𝑛 𝑞0 ∑ 𝑃𝑛 𝑞𝑛
Fisher’s ideal index F = ∑ 𝑃0 𝑞0
× ∑ 𝑃0 𝑞𝑛

If time subscripts are reversed,

∑ 𝑃0 𝑞𝑛 ∑ 𝑃0 𝑞0
F’ = ∑ 𝑃𝑛 𝑞𝑛
× ∑ 𝑃𝑛 𝑞0

Since F × F’ = 1 the test is satisfied

6.5.2 The Factor Reversal Test

With the usual notations, a “value index” formula is given by

∑ 𝑃𝑛 𝑞𝑛
Iv = ∑
𝑃0 𝑞𝑛

Now, for example, Laspeyres’ index for prices and quantities are given
respectively by
∑ 𝑃𝑛 𝑞0 ∑ 𝑞𝑛 𝑃0
Ip = ∑ and Iq = ∑
𝑃0 𝑞𝑛 𝑞0 𝑃0

The factor reversal test desires that Ip. Iq = Iv

But for Laspeyres’ index

∑(𝑃𝑛 𝑞0 ) ∑ 𝑃𝑛 𝑞0
Ip.Iq = ∑(𝑃0 𝑞0 )2
= Iv

On the other hand, Fisher’s ideal index satisfies this test, as shown below.

∑ 𝑃𝑛 𝑞0 ∑ 𝑃𝑛 𝑞𝑛
Ip = ∑ 𝑃𝑛 𝑞0
×∑
𝑃0 𝑞𝑛

∑ 𝑞𝑛 𝑃0 ∑ 𝑞𝑛 𝑃𝑛
Iq = ∑ 𝑞𝑛 𝑃0
×∑
𝑞0 𝑃𝑛

∑ 𝑃𝑛 𝑞0 ∑ 𝑃𝑛 𝑞𝑛 ∑ 𝑞𝑛 𝑃0 ∑ 𝑞𝑛 𝑃𝑛
Ip.Iq = ∑ 𝑃0 𝑞0
×∑ ×∑ ×∑
𝑃0 𝑞𝑛 𝑞0 𝑃0 𝑞0 𝑃𝑛

∑ 𝑃𝑛 𝑞𝑛 ∑ 𝑃𝑛 𝑞𝑛 ∑ 𝑃𝑛 𝑞𝑛
= ∑ 𝑃0 𝑞0
×∑ =∑ = Iv
𝑞0 𝑃0 𝑃0 𝑞0

140
To understand this principle further, we take the following example. Index Numbers

If the price and quantity per unit of an item changed in 2020, as compared to
2010, from Rs. 16 to Rs. 32 and from 100 units to 200 units respectively, then the
price and quantity in 2020 would both be 200% or 2.00 times the price and
quantity in 2010. The values (product of price and quantity) would be Rs. 1600 in
2010 and Rs. 6400 in 2020, so that the value ratio is 6400/1600 = 4.00 Thus, we
verify that 2.00 × 2.00 = 4.00, that is, the product of price ratio and quantity ratio
is equal to the value ratio.

Only the Fisher’s ideal index satisfies this test.

Example 6.3 we show with the following data that the Fisher’s ideal index
satisfies the factor reversal test:

Item Base Year Current Year

(𝑝 𝑞 ) (𝑝 𝑞 ) (𝑝 𝑞 ) (𝑝 𝑞 )
Price Quantity Price Quantity
(𝑝 ) (𝑞 ) (𝑝 ) (𝑞 )

I 6 50 10 56 300 500 336 560

II 2 100 2 120 200 200 240 240

III 4 60 6 60 240 360 240 360

IV 10 30 12 24 300 360 240 288

V 8 40 12 36 320 480 288 432

Total 1360 1900 1344 1880

Let us calculate the following from the data given in the above table.

∑ 𝑃𝑛 𝑞0 ∑ 𝑃𝑛 𝑞𝑛
Price Ratio: Ip = ∑ 𝑃𝑛 𝑞0
×∑ = ×
𝑃0 𝑞𝑛

∑ 𝑞𝑛 𝑃0 ∑ 𝑞𝑛 𝑃𝑛
Quantity Ratio: Ip = ∑ 𝑞𝑛 𝑃0
×∑ = ×
𝑞0 𝑃𝑛

∑ 𝑃𝑛 𝑞𝑛
Value Ratio: Iv = ∑ =
𝑃0 𝑞0

Ip.Iq = × × × = × =

= Iv which shows that the test is satisfied.

141
Summarisation of 6.5.3 Chain Index Number and Circular Test
Bivariate and Multi-
variate Data Two types of base periods are used for the construction of index numbers,
namely, (a) fixed base, (b) chain base. Most commonly used indices use fixed
base method. This method cannot take into account any changes in price or
quantity in any other year. It fails to include new commodities gaining
importance at a later date or exclude commodities losing significance in course of
time. These problems can be overcome by chain index numbers.
Using a suitable index number formula (say, Laspeyres’ index), link indices,
defined as follows, are first calculated: Link index = Index number with previous
period as base. The chain index is obtained by multiplying link indices
progressively. Thus, the chain index number Ion for period n with base period 0 is
given by

I01 = I01

I02 = I01 × I12

I03 = I01 × I12 × I23 = I02× I23

……………………………..

…………………………….

I0n = I01 × I12 × I23 ……..× I(n-1)n = I(n-1) × I(n-1)n

Example 6.4 The calculation of chain index numbers is illustrated with reference
to the following data:
Year Link index Chain index (Base 2010 =100)

2010 100 100

2011 I01 = 80 100× = 80

2012 I12 = 120 80× = 96

2013 I23 = 75 96× = 72

Thus, the chain index numbers for the years 2011 to 2013 with 2010 as the base
are 80, 96 and 72 respectively.

Circular Test: The circular test is an extension of time reversal test over a
number of years. It states that the chain index for the year 2013, calculated above,
starting from the base year 2010 will be same as the index number directly
calculated with fixed base period of 2010. In symbols,

142
I01= I12 × ……..× I(n-1)n ×= In-0) = 1. (Notice that I0n = Index Numbers

Considering an aggregate index with fixed weights

∑ 𝑃1 𝑞
∑ 𝑃0 𝑞

We can illustrate the test as follows:

With base period 0, we can trace the above formula from 1 to 3 years:

∑ 𝑃1 𝑞 ∑ 𝑃2 𝑞 ∑ 𝑃3 𝑞 ∑ 𝑃0 𝑞
× × × =1
∑ 𝑃0 𝑞 ∑ 𝑃1 𝑞 ∑ 𝑃2 𝑞 ∑ 𝑃3 𝑞

The formulae satisfying the requirements of circular test are:

1) Simple aggregative index

2) Simple geometric mean of relatives

3) Weighted aggregative index (such as Laspeyres’ index with constant

weights)

4) Weighted geometric mean of relatives with constant weights.

Fisher’s ideal index does not satisfy this test. It has been proved that no index
satisfies both the factor reversal and the circular test.

Check your progress 2

1) Compute the chain index number with 2010 prices as the base from the
following table giving the average wholesale prices of commodities A, B
and C for years 2010-2014.

Commodity Average whole sale Price (in Rs.)

2010 2011 2012 2013 2014

A 20 16 28 35 21

B 25 30 24 36 45

C 20 25 30 24 30
…………………………………………………………………………...…
…………………………………………………………………………...…
…………………………………………………………………………...…
…………………………………………………………………………...…
……………………………………………………………………………..
…………………………………………………………..…....................…

143
Summarisation of 2) Construct Fisher’s ideal Index number from the following data and
Bivariate and Multi-
variate Data show that it satisfies Factor and Time Reversal Tests.
Base Year Current Year
Commodities Price Expenditure Price Expenditure
Per unit (Rs.) per unit (Rs.)
A 2 40 5 75
B 4 16 8 40
C 1 10 2 24
D 5 25 10 60
……………………………………………………………………………
……………………………………………………………………………
……………………………………………………………………………
……………………………………………………………………………

6.6 COST OF LIVING INDEX (CLI) OR CONSUMER

PRICE INDEX (CPI)
This is an index of changes in the prices of goods and services commonly
consumed by a homogeneous group of people, such as families of industrial
workers. The major items of consumption that are considered for the construction
of CLI are:

1) Food

2) Fuel and Light

3) Clothing

4) House rent

5) Miscellaneous
The common method for obtaining the consumption basket is to conduct a family
living survey among the population group for which the index is to be
constructed. Prices of selected items are also collected from various retail
markets used by consumers in question. It may be noted that each of the above
broad groups contains several sub groups. Thus, ‘food’ includes cereals, pulses,
oils, meat, fish, egg, spices, vegetables, fruits, non-alcoholic beverages, etc.
‘Miscellaneous’ includes such items as medical care, education, transport,
recreation, gifts and many others. When more than one price quotation is
collected for a single commodity, a simple average is taken. Index number is
constructed for each of the five groups using weighted average of the price
group; the weighs used are proportional to the expenditure on the consumed item
buy an average family. Next, the overall index (CLI) is computed as an weighted

144
average of group indices, the weights being again the proportional expenditure on Index Numbers

different groups (e.g., 50 per cent on food).

Using Laspeyres’ formula
∑ ×
Cost of living Index: I = ∑
where w =∑

The CLI or consumer price index (CPI) numbers have significant practical
implications and extensive public use. Its use as a wage regulator is the most
important. The dearness allowance (DA) of employees is primarily determined
by this index. When wages or incomes are divided by corresponding CLI, the
effect of changes in prices (inflation) is eliminated. This is known as the process
of deflation, which is used to find ‘real wages’ or ‘real income’ As mentioned
earlier the reciprocal of CLI gives us the purchasing power of money.
Example 6.5: Construction of an index for food
Item Prices Weights
Pn P0 P = (Pn × P0) w P×w
Rice 50 40 125.0 30 3750.0
Wheat 45 30 150.0 20 3000.0
Pulses 60 40 150.0 10 1500.0
Sugar 40 20 200.0 5 1000.0
Oil 75 60 125.0 15 1875.0
Potato 60 50 120.0 15 1800.0
Fish 200 150 133.3 5 666.5
Total 100 13591.5

∑ ×( ) ∑
Index (food) = ∑
× 100 = ∑
× 100

.
= = 135.915 = 135.92

Example 6.6: Construction of a final Cost of Living index number.

Item Weight Index Weight × Index
(Percentage
Expenditure)
Food 45 130 5850
Clothing 15 140 2100
Housing 20 170 3400
Fuel 5 110 550
Misc. 15 125 1875
Total 100 13591.5
,
Cost of Living Index = = 137.75 = 138

145
Summarisation of Check your progress 3
Bivariate and Multi-
variate Data 1) Calculate a number which will indicate the percentage change in volume
of traffic from October 2019 to October 2020, when account is taken of
the relative values of the different types of traffic.
Type of traffic Tons(‘000) Receipts(Rs.’000)
Oct. 2019 Oct. 2020 Oct. 2019
Merchandise 1246 1206 776
Fuel 4794 4229 562
……………………………………………………………………………
……………………………………………………………………………
……………………………………………………………………………
……………………………………………………………………………
………………………………………………….................………..……

2) Compute Paasche’s price index number for 2020 with 2015 as the base
from the following data:
Commodity Unit Price(Rs.) per unit Quantities sold
2015 2020 2015 2020
A kg. 4 5 95 120
B kg. 60 70 118 13
C kg. 35 40 50 70

……………………………………………………………………………
……………………………………………………………………………
……………………………………………………………………………
……………………………………………………………………………
…………………………………………………………………………...

3) From the following data, compute the Laspeyres’ price index number for
2021 with 2019 as base:
Item Price(Rs.) Total Value(Rs.)
2019 2021 2019
A 12.50 14.00 112.50
B 10.50 12.00 126.00
C 15.00 14.00 105.00
D 9.40 11.20 47.00
146
 …………………………………………………………………………… Index Numbers

……………………………………………………………………………
……………………………………………………………………………
……………………………………………………………………………
……………………………………………………………………………

4) Calculate Marshall-Edgeworth index number from the following data:

2010 2015
Commodity Price Quantity Price Quantity
Rice 9.3 100 4.5 90
Wheat 6.4 11 3.7 10
Jowar 5.1 5 2.7 3
…………………………………………………………………………...…
…………………………………………………………………………...…
…………………………………………………………………………...…
…………………………………………………………………………...…
.......................................................................................................................
.......................................................................................................................

6.7 WORKED OUT EXMAPLES

In this section, we shall provide worked out examples so as to further familiarize
you with the topic.
Example1.7: Construction of Price Index
Item Unit Price per unit(Rs.)
2010 (P0) 2020 (Pn) (Pn ÷P0)×100
Rice quintal 100 220.00 220
Wheat kg. 1.50 2.40 160
Fish kg. 15.00 28.00 187
Bread lb. 0.60 1.35 225
Milk litre 2.50 4.00 160
Total 119.60 255.75 952
a) Aggregative Method
Index number for 2020 (base 2010 =100)
Average price per unit in 2020
× 100
Average price unit in 2010

147
Summarisation of ∑ / .
Bivariate and Multi- =∑ × 100 = × 100 = 213.84
/ .
variate Data
b) Method of Price Relative
Index number for 2020 (base 2010 =100)
∑ ×
=

= 952 5 = 190.4

Example 6.8: Calculate price index numbers from the following information,
using (a) weighted aggregative formula, and (b) weighted arithmetic mean of
price relatives:
Item Unit Price per unit(Rs.)
Base year Current year Weight
A quintal 85 115 19
B kg. 15 15 25
C dozen 45 61 40
D litre 55 100 20
E Lb 17 23 21

Calculation for Index numbers

Item P0 Pn w P0w PnwI = ( Pn÷P0)×100 Iw

A 85 115 19 1615 2185 135.3 2570.7

B 15 20 25 375 500 133.3 3332.5
C 45 61 40 1800 2440 135.6 5424.0
D 55 100 20 1100 2000 181.8 3636.0
E 17 23 21 357 483 135.3 2841.3

Total 125 5247 7608 17804.5

∑
a) Weighted aggregative index = ∑ × 100 = × 100 = 144.99 =
145
b) Weighted arithmetic mean of price relatives
∑ .
= ∑
= = 142.44

Example 6.9: Given below are the data on prices of some consumer goods and
the weights attached to the various commodities. Calculate price index numbers

148
for the year 2021 (base 2020 = 100), using (a) simple average, and (b) weighted Index Numbers
average of price relatives.

Price (Rs.)
Commodity Unit 2020 2021 Weights
Wheat Kg. 0.05 0.75 2
Milk Litre 0.60 0.75 5
Egg Dozen 2.00 2.40 4
Sugar Kg. 1.80 2.10 8
Shoes Pair 8.00 10.00 1
Calculations for price relative index.

Commodity Unit P0 Pn I=( Pn÷P0)×100 w I

Wheat kg. 0.50 0.75 150 2 300
Milk Litre 0.60 0.75 125 5 625
Egg Dozen 2.00 2.40 120 4 480
Sugar kg. 1.80 2.10 117 8 936
Shoes Pair 8.00 10.00 125 1 125

Total 637 20 2446

∑ ×
a) Simple average of price relative index = = = 127.4
∑
b) Weighted average of price relative index = ∑
= = 123.3

Example 6.10: On the basis of the following data, calculate the wholesale price
index number for the five groups combined (Base: 2015-16 = 100).
Group Weight Index no. for
the week ending 31.01.2021

Food 50 241

Liquor and tobacco 2 221

Fuel, power, light and lubricants 3 204

Industrial raw materials 16 256

Manufactured commodities 29 179

∑
We compute: General index =
∑

where I = Group index, and w = Group weight

149
Summarisation of
Bivariate and Multi- Group weight (w) Group Index(I) I×w
variate Data

Food 50 241 12050

Liquor and tobacco 2 221 442

Fuel, power, light and lubricants 3 204 612

Industrial raw materials 16 256 4096

Manufactured commodities 29 179 5191

Total 100 22391

Index number of wholesale price = = 223.91

Example 6.11: Annual production (in million tons) of four commodities are
given below:
Commodity Production Weight
2015 2019 2020

A 160 200 216 20

B 24 42 45 30
C 50 72 68 13
D 120 168 156 17

Calculate quantity index numbers for the years 2019 and 2020 with 2015 as base
year, using (a) simple arithmetic mean, and (b) weighted arithmetic mean of the
relatives.
Quantity relatives for 2019 with base year 2015 (=100)
𝑞 𝑞
𝐼= 𝑞 × 100 = 𝑞 × 100
Commodity A: × 100 = 125

Commodity B: × 100 = 175

Commodity C: × 100 = 144

Commodity D: × 100 = 140

Quantity relatives for 2020 with 2015 = 100
𝑞
𝐼= 𝑞 × 100
Commodity A: × 100 = 135

Commodity B: × 100 = 187.5

Commodity C: × 100 = 136

150
Commodity D: × 100 = 130 Index Numbers

Commodity Quantity relatives(I) Weight I×w

2019 2020 w 2019 2010
A 125 135.0 20 2500 2700
B 175 187.5 30 5250 5625
C 144 136.0 13 1872 1768
D 140 130.0 17 2380 2210
Total 584 588.5 80 12002 12303
∑
a) Simple arithmetic mean of quantity relatives = × 100

(where k = number of commodities)

Index number for 2019 = = 146
.
Index number for 2020 = = 147
∑
b) Weighted arithmetic mean of quantity relatives
∑

Index number for 2019 = = 150

Index number for 2020 = = 154

Example 6.12: From the following price(p) and quantity (q) data, compute
Fisher’s ideal index number.
Commodity 2015 (Base Year) 2020 (Current Year)
Price Quantity Price Quantity
A 12 10 17 10
B 14 9 16 11
C 11 12 13 10

Calculations for Fisher’s ideal index:

Commodity P0 q0 Pn qn P0 q0 Pnqn P0qn Pnqn

A 12 10 17 10 120 170 120 170

B 14 9 16 11 126 144 154 176
C 11 12 13 10 132 156 110 130

Total 378 470 384 476

∑ 𝑃𝑛 𝑞0
Laspeyres’ price index = ∑ × 100 = × 100 = 124.34 = 124
𝑃𝑛 𝑞0

∑ 𝑃𝑛 𝑞𝑛
Paasche’s price index = ∑ × 100 = × 100 = 123.96 = 124
𝑃0 𝑞𝑛

Fisher’s ideal index = √𝐿 × 𝑃 = √124 × 124 = 124.

151
Summarisation of
Bivariate and Multi- 6.8 LET US SUM UP
variate Data
In this unit you have been introduced to the concepts and methods involved in the
construction of Index Numbers. You have been shown how to use the
Laspeyres’, Paasche’s and Fisher’s formulae for calculating price as well as
quantity indices. You also know now how to measure changes in consumer price
or cost of living.

6.9 ANSWERS AND HINTS TO CHECK YOUR

PROGRESS EXERCISES
Check Your Progress 1
1) and (2): Do it yourself.
3) Simple Aggregative Index Number = 117.14
Average of Price Relative Method = 122.9
4) 84.2
5) Laspeyres’ Index Number = 86.02
Paasche’s Index Number = 81.25
Check Your Progress 2
1) 108.33. 135.41, 160.23, 165.56
2) Do it yourself
Check Your Progress 3
1) We find quantity for Oct. 2020 with Oct. 2019 as base. The required index
may be obtained as the weighted arithmetic mean of quantity relatives, using the
receipts in 2019 as weights.
Type of traffic q0 qn Weight Quantity (4)×(5)
(w) (Pn÷ q0)×100
(1) (2) (3) (4) (5) (6)
Merchandise 1246 1206 776 97 75272
Minerals 1125 981 252 87 21924
Fuel 4794 4229 562 88 49456

Total 1590 146652

∑ × ×
Quantity index = ∑
= = 92

2) Calculation for Paasch’s price index

152
Commodity P0 Pn q0 qn P0 qn Pn qn Index Numbers

A 4 5 95 120 480 600

B 60 70 118 130 7800 -----
C 35 40 50 70 2450 2800
Total 1590 146652
∑ 𝑃𝑛 𝑞𝑛
Paasche’s Price index = ∑ × 100 = × 100 = 116
𝑃0 𝑞𝑛

3) We are given the base price(P0), current price (Pn) and value in the base year
(P0q0). To find base year quantity (q0), we can use the relation
𝑃𝑞
𝑞 =
𝑃
Using P0, Pn and q0, we can find Laspeyres’ index as
∑ 𝑃𝑛 𝑞0
𝐿= × 100
∑ 𝑃0 𝑞0
Calculation for Laspeyres’ price index
Item P0 Pn P0 q0 Pn Pn q0

A 12.50 14.00 112.50 9 126.00

B 10.50 12.00 126.00 12 144.00
C 15.00 14.00 105.00 70 7 98.00
D 9.40 11.20 47.00 5 56.00
Total 390.50 424.00
∑ 𝑃𝑛 𝑞0 .
Laspeyres’ price index = × 100 = × 100 = 109.
𝑃0 𝑞0 .

∑ 𝑃𝑛 𝑞0 +𝑞𝑛
4) Marshall-Edgeworth index = ∑ × 100
𝑃0 𝑞0 +𝑞𝑛

∑ 𝑃𝑛 𝑞0 +∑ 𝑃𝑛 𝑞𝑛
=∑ × 100
𝑃0 𝑞0 +∑ 𝑃0 𝑞𝑛

Let us take 2010 as base and 2015 as current year.

Commodity P0 q0 Pn qn P0 q0 P0 qn Pn q0 Pn qn
A 9.3 100 4.5 90 930.0 837.0 450.0 405.0
B 6.4 11 3.7 10 70.4 64.0 40.7 37.0
C 5.1 5 2.7 3 25.5 15.3 13.5 8.1
Total 1025.9 916.3 504.2 450.1
504.2+450.1
Required Index = × 100 = 49.1.
1025.9+916.3

153