LECTURE 1 EC100

INDEX NUMBERS

Definitions
(1) Are convenient devices for measuring relative changes or differences from time to
time or from place to place. In the latter case this is done by simultaneously collecting
data in different locations and then comparing the data. The comparative cost of living
index for example may show that the cost of goods and services it is cheaper to live in
Gweru than in Harare.
(2) Are devices to simplify comparison over a period of time, and to show the relative
movements of data concerned and to replace what may be complicated figures by
simple ones calculated on a percentage basis.

Base period
This is a specified period used for comparison purposes. Calculations of changes are made
relative to this period. Index numbers are generally easy to follow since they use the value of
100 for the base year and changes from the base year can easily be calculated as percentage
changes.

Rules for the selection of the base period

It is necessary to choose the base year carefully as it is used as a point of reference in
comparing various index numbers and describing their individual behaviour. There are
essentially two rules:
(1) It should be a period when prices and other factors are stable i.e. a “normal” period. In
Zimbabwe 1980 was used as a base year (This is written as (1980= 100)). It was a
period of relative stability. The base period has since been changed because it has
become too outdated.
(2) Base period must be relatively recent. The more distant the base year is in the past, the
more difficult it will be to find comparable data for later periods. It is better to use a
more recent period because we are more concerned with changes relative to the
present than with fluctuations related to the distant past.

Problems Associated with Index Construction

(1) Selection of the base period. The base period should be as close to the trend as
possible, since an abnormal base year distorts the index in subsequent years. Always
consider how and why the base period was selected before a claim based on the result
of comparing index numbers.
(2) It is essential to define the purpose of the index carefully and clearly so that the index
can be correctly designed and problems of collection, collation and data interpretation
anticipated.
(3) Data selection. What data to include will depend on the purpose for which the index is
required. The data must be relevant. Sometimes there is difficulty in finding suitable
data to compute an index.
(4) Selection of the commodities. The items must be carefully selected as representative
of all relevant items.
(5) Selection of weights. The weights associated with each item must be carefully
considered so that the most appropriate weighting (e.g. volume or value weighting) is

1
 used. Some variables are considered more important than others in order to come up
with appropriate weight.
(6) Representative Sampling. In constructing a simple index or in comparing the price of
a single commodity in two different periods, the index numbers are constructed from
samples but not necessarily random ones, but selected according to the compiler’s
judgment. In a composite index the process may involve hundreds of variables and the
selection of a sample may be an arduous task. The bottom line is that the samples
must be as representative as possible.
(7) Comparability. The data must be comparable. Incomparability of indices occurs when
attempts are made to compare one index with another after there has been a basic
change in what is being measured. Newer cars appear more expensive than old ones.
This does not take into account technological advances in the quality of cars.
(8) Monitoring of the data. Accurate data is required and skilled staff are needed to
ensure the index is authoritative.
(9) Calculation of the index. The method depends on the formula used.

Simple Index Numbers

These are computed for a single variable. They include the following:
(1) Index Relatives. An index relative, sometimes simply called a relative is the name
given to an index number which measures the change in a single distinct commodity.
Pn
×100
Price Relative= P o where Po is price in the base period and Pn is price in any
other period.
Qn
×100
Quantity Relative= Qo where Qo is quantity in the base period and Qn is
quantity in any other period.

(2) Simple aggregative indexes. If this is a price index that is to be calculated, the sum of
various prices for each of the periods under consideration should be divided by the
sum of the base period.
∑ P n ×100
Price Index = ∑ P 0
Table 57 Prices of commodities
Commodity Price in 1995* Price in 1996 Price in 1997
X ( per ton) $12.00 $15.00 $15.60
Y (per dozen) $ 3.00 $ 3.60 $ 3.30
Z (per kg) $ 5.00 $ 6.00 $ 5.70
Aggregate $20.00 $24.60 $24.60
Index 100.00 123.00** 123.00
* (1995 = 100).
24 . 60
×100=123
** is calculated as follows: 20 .00
It is usually sufficient to quote index numbers to one decimal place [Robertson and
McCloskey (2002: 74)]

(3) Simple Average of Relatives (Unweighted). To compute such an index, actual prices
for each variable are converted to percentages of base period. These percentages are
called relatives because they are relative to the value for the base period.

2
 ∑ ( Pn
Po
×100
)
Price Index = n where n represents the number of elements or products
in the composite. The price index is simply the arithmetic mean of relatives.
Table 58 Price Relatives
Commodity 1995 1996 1997
Po P1 P2
×100 ×100 ×100
Po Po Po
X 100 125 130
Y 100 120 110
Z 100 120 114
Aggregate 100 365 354
Index 100.00 121.7** 118.0
** calculated as follows: 365/3 = 121.7
The major disadvantage of an unweighted index is it does not attach greater
importance or weight, to the change of a high use item (e.g. eggs) that it does to a low
volume item (e.g. a calculator). A substantial price change for slow moving items can
completely distort an index.

Composite Index Numbers

These are also called weighted index numbers and are constructed from two or more
variables. These are more common than the simple index numbers. Simple index numbers are
often criticized for not showing importance indicators (to show the relative importance of an
item) and for adding up prices for different units. To avoid these pitfalls weighting is used in
compiling the composite index numbers. These include:
(1) Fixed weight Index . This employs weights that are based on a period or periods that
are considered representative. The weights and base prices may not come from the
same period.
∑ P1 w ×100
Price Index= ∑ P o w where w is the weight. The formula can be alternatively be
put as
∑ P1 Qf ×100
Price index = ∑ P o Q f where Qf is the quantity produced or sold in the base
period or period selected as representative.
Advantages-
(a) Flexibility in selecting the base price and the fixed weight (quantity).
(b)The index allows a company to change the price base without changing the fixed
weight. This is useful because quantity measures are often expensive or impossible to
obtain for certain periods.
(2) Simple Weighted (Aggregative) Index. This uses the base year for both price and
quantity in the numerator. This is also called the expenditure index or value index.
∑ P1 Q1 ×100
Value Index or Expenditure Index= ∑ P o Qo
Although it is simple to use, quantities sold or produced in the base period may not be
considered representative of overall performance.

3
(3) Paasche Index. This is a special case of a weighted aggregate index which uses
current time period weights.
(a) Paasche Price Index (Pp). This is a current year quantity weighted index.

P p=
∑ P1 Q1 ×100
∑ P0 Q1
Advantages: With this index the relative importance of the items making up the
basket of goods is continuously changing because the quantities are different each
year. It is more accurate than the Laspeyre’s Index because it shows what the industry
is actually using in the current year, and therefore takes account of the price changes
and the quantity changes. The Paasche price index indicates how much current year
costs are related to the cost of buying current year quantities at base year prices.
Disadvantages: There is need to tabulate quantity measures for each period examined.
It is difficult to find reliable data for say 100 food products in different parts of the
country each for several years. Each value for a Paasche price index is the result of
both price and quantity changes from the base period. Since the quantity measures
used for one period are usually different from the quantity measures for another index
period, it is impossible to attribute the difference between the two indices to price
changes only. Thus it is difficult to compare indices from different periods as
calculated by the Paasche method.
(b) Paasche Quantity Index (Pq). This is a current year price weighted index.

Pq =
∑ P1 Q1 ×100
∑ P1 Q 0
Although the index reflects changed patterns of production or consumption, a separate
index must be computed for each period of interest, substituting the quantity (for
quantity index) or the prices (for a price index) values of that period. New inputs of
quantity (or price) data would be required for each period to which we direct our
attention and this could be costly and time consuming.
(4) Laspeyre’s Index. This is another special case of a weighted aggregate index which
always uses base time periods weights.
( a) Laspeyre’s Price Index ( Lp): This is a base year quantity weighted index. It
indicates how much the cost of buying base year quantities at current price is,
compared with base year costs.

L p=
∑ P1 Q0 ×100
∑ P 0 Q0
Advantages: It compares the price of the quantity of goods purchased in the base year,
year by year, as prices rise. It assists comparability from year to year because the
quantity is the same every year. Many commonly used quantity measures are not
tabulated every year.
Disadvantages: It does not take into consideration changes in consumption patterns.
Items purchased in large quantities just a few years ago may be relatively unimportant
today.
(b) Laspeyre’s Quantity Index (Lq): This measures percentage changes in
quantities. The Laspeyre’s quantity index can also be looked at as a base year price
weighted index.

Lq =
∑ P0 Q1 ×100
∑ P0 Q 0
4
 The Laspeyre’s method is more often used than the Paasche’s method. It uses quantity
weights only for base period and therefore does require fresh inputs of quantity data
when calculating the price index of a given period. Index numbers for different
periods are directly comparable.
(5) Fisher’s Ideal Index: This is also called the True Index. The Laspeyre’s index tends
to exceed the Paasche index. In this context, Laspeyre’s index will tend to overstate
changes in prices or quantities whilst the Paasche will tend to understate the changes.
However, the price and quantity changes have to be quite dramatic for there to be an
extremely large difference. These two indices can be thought of as being at two
extremes of under – and overestimating the effects of price changes. The differing
effects of price changes on the quantity consumed have led to the use of a composite
measure, the Fisher’s Ideal index lies between the two, ,it is a compromise of the two.
The index is considered ideal in the sense that it makes a more accurate prediction of
the expenditure index than the other individual indices. In formula terms the Fisher’s
index is the Geometric Mean of the Lp and Pp or the Geometric Mean of the Lq and Pq.
F p= √ L p ×P p and
F =√ P × L
q q q
[NB. Make sure that when calculating the Fisher’s index you do not mix quantity
index and price index. If you make that mistake you would find a nonsensical result.].
(6) Retail Price Index. This is a weighted average of price relatives. This index is
calculated where there are a large number of items. Firstly a sample survey is carried
out to find what people buy. Each item is then priced as it is bought each year and a
simple price relative for each item is worked out. To make the index we must know
whether we want it to be base year weighted or current year weighted.
For a base weighted price relative index:

[( P
)
∑ Pn ×100 ( P o Qo )
o ]
RPI = ∑ Po Q o
For a current weighted price relative index:

[(P
)
∑ Pn ×100 ( P n Qn )
o ]
RPI = ∑ Pn Q n
Robertson and McCloskey (2002:82) say that by comparing the equations for the values,
price and quantity indices it can be seen that they are related to each other in that
Value Index = Laspeyre’s Price Index ¿ Paasche quantity index
= Paasche price Index ¿ Laspeyre’s quantity index

Some Concepts To Remember

Fixed Base Index
Each index here is calculated based on the same fixed time point. The approach can only be
used when the basic nature of the commodity is unchanged over the whole period. For
example if data for the periods 1990, 1991, 1992, 1993, 1994 and 1995 is given index
numbers for all these periods will be calculated based on one common time point such as
1990.

Chain Base Index

5
Each index is calculated with respect to the immediately preceding time point. In other words
the base period changes from one period to the next. For example if data for the periods 1990,
1991, 1992, 1993, 1994 and 1995 are given the index numbers for 1991 will use 1990=100
while that for 1992 will use 1991=100 ; 1993 will use 1992=100 etc. The chain base index is
often appropriate when the basic nature of the commodity changes over the whole time
period.

Changing the base

To change an index with base period X into an index with base period Y, all you need to do is
to divide each value of the index with base period X by the values of the index with base
period X at period Y.
Index x (t )
Index Y (t )= ×100
Index x (Y )
When two or more sets of data need to be compared and they have two different base years a
reasonable comparison can only be made when the data is changed to a common base period
preferably a later one. For example if television sets sold used 1984 = 100 and television
licences used 1995 = 100 then to make a reasonable comment on the requisite rates of
growth or decline we would need to use 1995=100 for both sets of data. The procedure is as
follows:
(a) Look up the index number relating to the new base year (row 1995 = 100 and
the column 1984 = 100 for televisions).
(b) Divide this number into each index number in the series and multiply the
result by 100. This results in a new series of index numbers with the new year
1995 as its base for televisions as well!

Table 59. A comparison of licences and radio purchases

TV licences TV sets
Year TV sets (1984 =100) (1995=100) (1995=100)
1990 126 87 69*
1991 135 89 74
1992 147 92 81
1993 159 95 87
1994 165 97 91
1995 182 100 100
1996 187 105 103
1997 195 109 107
1998 208 112 114**
* calculated as 126/182 x 100
** calculated as 208/182 x 100

Uses of Index Numbers

- The retail price index is used to show changes in the cost of living.
- The index of Industrial production reflects changes in the industrial output.
- Index numbers can be used by trade unions for wage negotiations.
- Index numbers can be used in the insurance industry to index link house policies.
- Index numbers can be used for deflating a time series to see whether when inflation is
removed wages have risen in real terms. Deflation is the division of a dollar-value time
series by an appropriate price index number series in order that one may see the amount
of the change in the series that was physical in character, ie. The result of changes in
quantities e.g. a retail-sales series could be divided by a retail-price index to determine

6
 how much of an increase in the former series resulted from an increase in the volume of
goods sold. A similar computation is the division of a monthly wage series by the
Consumer Price Index to determine “real wages”.
- Index numbers may be studied as indicators of general business conditions. Such a study
along with study of other indicators is a preliminary stage in business forecasting and
planning. The wholesale price index and its components are studied by a business to
follow basic price movements and plan the company’s price policies in light of this
analysis. The company can compare its own price movements with those of the index.
The index of industrial products and its components are studied to follow changes in the
physical volumes of production, to make comparisons and to plan accordingly.
- Escalators- Index numbers are used in the operation of “escalator clauses” in various
contracts. The CPI is used as an escalator in many wage contracts, wages being adjusted
up or down automatically with movements up or down in the index. The wholesale price
index, or some component of it may be used as an escalator for adjusting prices at the
time of delivery in long business contracts.
- Purchasing Power- the reciprocal of a price index may be used to show the purchasing
power of the dollar relative to the type of goods whose prices were measured by the
index. This purchasing power itself will be relative to the base period on which the index
is constructed, e.g. If in May 1997 the all item consumer price index was 180.6 on the
1987 base, the purchasing power of the 1987 consumer dollar in May 1997, was
100.0/180.6 or $0.553.

Limitations of Index Numbers

1. Weighting factors can become obsolete.
2. If sampling had been done to obtain data, information might be incomplete or false.
3. They give only general indications of changes- minority groups or professions may
not be catered for.
4. Can be misinterpreted by the uninformed layman if he does not know the difference
between the chain base and fixed base index.