Market concentration describes the extent to which the top firms in an industry, say in the car industry where the top five firms in the UK would account for nearly 90% of the market, take up a large portion of the market share.  There are various methods used to measure this, which will be discussed in turn.

‘The concentration ratio is the percentage of all sales contributed by the leading three or five, say, firms in a market.’ (Maunder, P. et al (1991) p561) So the concentration ratio can be calculated by using the cumulative share of the first three or five firms according to their sales revenue share, summarised in the following equation:

CR_k= SS_i , i=1…k

where S_i =sales revenue of ith firm/sales revenue of subsector

Looking at the following table we can see that between the largest five firms in each of the following markets there has been a significant increase in their market concentration from 1963 to 1977:

Product	1963/ %	1977/ %
Beer	50.5	62.2
Biscuits	65.5	79.7
Cars	91.2	98.4
Flour	51	85.7
Pharmaceuticals	53.9	63.2
Refrigerators	71.9	98.8
Washing Machines	85.2	96.2

 

(Griffiths, A. & Wall, S. (1991) p 109)

So as can be seen from the above figures in 1977 especially the car, refrigerators and washing machines industries had high market concentrations.  However high market concentrations are not present in all industries, and much variance can occur.  For example in the tobacco industry the five largest firms accounted for 99% output and 98% of employment in 1991, however at the same time in the leather goods industry the five largest firms accounted for only 10% of net output and employment in.

However is there a way of classifying certain industries as being oligopolistic when looking at the three or five firm concentration ratio?  Firstly a clear definition of an oligopolistic industry or market must be set.  ‘Concentration ratios rise as we narrow the definition of an industry and fall as we broaden it.’ (Maunder, P. et al (1991) p 379) So one must therefore be careful when concluding that a market is oligopolistic.  Concentration graphs can be drawn as a result of data on concentration ratios.  The graph below (showing data for the year 2000) displays that the five largest investment managers in insurance groups accounted for $125.1 billion (84% of funds under management by insurance groups), the five largest investment managers in banking groups accounted for $139.3 billion (52% of funds under management by bank groups) and the five largest investment managers in other groups accounted for $74.5 billion (55% of funds under management by other groups).

The conventional means of displaying income distribution, and thus income inequality, is through the Lorenz curve (developed by Max O Lorenz).  The horizontal axis (in the below graph) shows the percentages of the population, and the vertical axis shows the percentage of the total income that they receive:

However we can adapt this concept to show information regarding market concentration.  Thus on the horizontal axis would be the cumulative percentage of total number of firms in the market or industry, and on the vertical axis would be the cumulative percentage of sales revenue they receive from the total of the market or industry, as shown on the below graph:

 The above diagram shows that for the market represented by the yellow dots the top 10% of firms compose 20% of the market share, whereas the market represented by the red dots the top 10% of firms compose only 10% of the market share.  Therefore the value for the market concentration for the market represented by the yellow dots is much higher than that for the market represented by the red dots.

To assess the concentration over the whole range of the distribution, the Gini coefficient is calculated.  ‘It is the ratio of the enclosed area between the Lorenz curve and the diagonal, to total area underneath the diagonal.’ (Griffiths, A. & Wall, S. (1991) p 447) The ‘diagonal’ that the definition refers to describes a curve where there is consistency amongst all the firms’ portion of market share in a particular market (e.g. the market represented by the red dots above).  So if the value for five firm market concentration ratio was 0.2, the Lorenz curve would coincide with this diagonal, thus the Gini coefficient value would be 0.  The other extreme would be if the Gini coefficient value were equal to 1, which occurs when the entire market share is composed of the last firm along the horizontal axis (i.e. the Lorenz curve coincides with the horizontal axis until the last firm).  In this situation the market concentration ratio value would equal 1 (i.e. a pure monopoly).  So the Gini coefficient ranges from 0 to 1, and a rise in the coefficient value suggests a higher market concentration.

Another measure of market concentration is the Herfindahl index.  The value of the index can be calculated by the following formula, which shows the sum of the squares of the market shares of all firms in a market:

H=S₁ⁿ(share_i)²

This index was originally used by the US Justice Department as a ‘screening’ index to warn the Department whether merger proposals required closer inspection or immediate prohibition.  One can apply a grid to a firm and write in each grid what percent of market share that firm occupies, placing greater emphasis on the large firms in the market This has been done below for two different markets:

Market A:                               

	1%	1%	1%	1%	1%
	1%	1%	1%	1%	1%
70%	1%	1%	1%	1%	1%
	1%	1%	1%	1%	1%
	1%	1%	1%	1%	1%
	1%	1%	1%	1%	1%

 

Market B:

20%	1%	1%	1%	1%	1%
20%	1%	1%	1%	1%	1%
20%	1%	1%	1%	1%	1%
20%	1%	1%	1%	1%	1%

 

So with market A the top 4 companies have 73% of total market share, and in market B the top 4 companies have 80% of total market share.

Market A: Index= 1(70) + 30(1) = 100

Market B: Index= 4(20) + 20(1) = 100

In this case the markets may be equally as competitive, however looking at the Herfindahl index a case could be made that market B is more competitive:

Market A: Index= 1(70)²+ 30(1)² = 4,930

Market B: Index= 4(20)²+20(1)² = 1,620

As market B has a lower index value in this case it can be regarded as being more competitive than market A.  This is because the higher the value the higher the market concentration (thus less competition as there are fewer larger firms in the market, thus less motivation for industries to be more efficient).  So in one extreme where there is a pure monopoly situation, i.e. a market concentration ratio value of 1, the index would have a value of 100² because only one firm has all the market share.  In the other extreme there may be hundreds of firms in a market, each bearing a small portion of the total market share, thus the index value would be close to zero and the market concentration would be low.

So market concentration is the key element in market structure and an important determinant of conduct and performance and hence of the type of competition, and there are many different methods to measure the level of concentrations in different markets. The two extremes are: perfect competition, in which products are homogeneous and firms small in relation to the size of the total market so that they cannot individually sway price (concentration is low), and monopoly, in which there is a single seller (concentration is absolute).

It has been already been mentioned that a market with a concentration value closer to 1 has a higher level of concentration.  However one must bear in mind that this description of the level of concentration is relative.  A measurement may show one market to have a higher level of concentration than another market, but ambiguities in the chosen method of measurement may leave it unclear as to which market is more concentrated.  An economist will not just consider the number of firms present in a market, but their sizes as well.  For example one may argue that a few but equally large firms in a market is more competitive (i.e. lower concentration) than a market with more firms but where one firm is much more larger and thus dominant over the other smaller firms.  So it becomes clear that the number of firms and firm size inequalities act separately on market concentration.  Hannah and Kay (1977) suggest a set of criteria for judging the suitability of different concentration measures.  These are as follows:

1. A higher concentration measure value represents a higher level of concentration.
2. If a large firm wins a customer(s) from a small firm the level of concentration should rise.
3. The entry of a new firm, below a significant size, into a market should reduce the level of concentration.
4. If a merger takes place the level of the concentration should increase.
5. The contribution of any firm to the measure of concentration tends to zero with its market share.

The concentration ratio only considers the top three or five, say firms and should a merger take place, the total sales of the market may remain unchanged and thus the value of the concentration ratio may get unaffected.  So this method of market concentration may does not satisfy criteria number 4.  Also the entry of a new (small enough) firm into a market may leave the Lorenz curve, and thus the Gini coefficient unchanged, hence the concentration level may remain the same, when in fact it should drop.  So this method of market concentration may not satisfy criteria number 3. 

So it may be concluded that the Herfindahl Index is the only measure of concentration that successfully fits all of the above criteria, and thus may be the most suitable measure of market concentration. 

 

References:

Maunder, P., Myers, D., Wall, N. & Miller, R. (1991) Economics Explained. 2^nd ed. London, HarperCollins

Griffiths, A. & Wall, S. (1991) Applied Economics. 4^th ed. London, Longman

Sykes, J. (2000) Australian Bureau of Statistics article no. 5655.0. 1^st ed. Australia, Managed Funds

World Bank Institute’s Education Program (2000) Beyond Economic Growth. 1^st ed.

Spring, B. (1998) Comm 497F Media Institutions. 1^st ed. USA, Compaine

 Mottershead, P. (2001) Concentration, Cost Structures and Entry [Internet] November 21^st Accessible from: www.uel.ac.uk Accessed on 22/11/01