• If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • Stop wasting time looking for files and revisions. Connect your Gmail, DriveDropbox, and Slack accounts and in less than 2 minutes, Dokkio will automatically organize all your file attachments. Learn more and claim your free account.



Page history last edited by Helmut Mueller 9 years, 8 months ago

Back to Project Overview

1. Introduction

This document outlines the basic methodology for the international media concentration project in order to enable us to compare countries and trends.


We acknowledge that not all of the media industries can be researched equally in all participating countries and for all years, but we would like to set a common framework upfront rather than end up with too many disparate apples and oranges.



2. Industries

TV Broadcasting

Daily Newspapers

Wireline Telecommunications

Wireless Telecommunications

Internet Service Providers (ISP)

Radio Stations and Networks


Magazines and Periodicals 

Book publishing

Search Engines

Multichannel(Cable TV Operators/ Satellite Operators)

Video Network, Cable Network


Final Deadline for all tables  March 11th, 2011

(Exceptions for new participants; otherwise only in agreement with Executive Committee)


3. Market Definitions

There will always be gray zones between media industries. For example, are weekly newspapers to be included in “daily newspapers” or in “magazines and periodicals”? (In this case, probably under the latter.) We suggest that you do not overly agonize about such questions but go ahead, state your assumptions clearly in footnotes or the text, and in some cases notify the entire group of the question if it applies to all. We can then make general determinations so as to keep the results comparable.



4. Geographic Definitions

The main unit of analysis is national markets. This has several implications;

  1. We ignore, for now, the fact that several media are local in nature, for example local newspapers.
  2. We focus on national, not global, revenues. For example, Bertelsmann’s book revenues and market shares would be calculated separately for Germany, USA, and France in the respective country chapters. (Later, we would aggregate global market shares.)
  3. We focus on all participants in national markets, not on those domestically owned. Thus, Sony BMG or EMI are part of the American music market, even though they are not American-owned.



5. Time Frame

To observe trends over time, we will cover a period of about 20 years, with one observation for every 4 years. We want to cover, going back from today, the following years: 2008 (or as close as possible), 2004, 2000, 1996, 1992, 1988, and 1984.



6. Measures of concentration

To assess the level of concentration in an industry for a given year, we use two indices that can be calculated using the same underlying numbers: the C4 ratio and the “Herfindahl-Hirschman Index” (HHI). A third index, referred to as a Noam index, divides the HHI by the square root of the number of voices, for a Media Ownership and Concentration Index.


C4 ratio

This index aggregates the market share percentage of the largest four companies in an industry. If the market shares of the top 4 firms were, for example, 40%, 30%, 10%, 10%, with 5 other firms holding 2% each, the C4 index would be 90. The formula for the C4 ratio is:

Where              Si = firm's i market share of a given industry j and where firms are ordered by size of market share.


This index is equal to the sum of the squares of the market shares of all market participants. Please note that this index only includes known market shares, so the category others (see below) is excluded from the calculation. The HHI can range between zero—where the share of each firm is infinitely small—and 10,000, when a single firm accounts for 100% of the market. In the example above, the HHI would be HHI = 402 + 302 + 2 × (102) + 5 × (22) = 2,720. The formula for this index is:

Where              f = number of firms participating in an industry,

                      Si = each firm’s market share

                       i = firm in a given industry


For your information, the US Department of Justice’s antitrust enforcement guidelines classify market concentration levels as follows:


HHI < 1,000                                                  Unconcentrated Market

1,000 < HHI < 1,800                                       Moderately Concentrated Market

1,800 < HHI                                                  Highly Concentrated Market


Media Ownership and Diversity Concentration Index (the Noam-index)

This index takes into account the number of voices available. Market power alone does not reflect media diversity. The availability of alternatives, even if they are small, permits different viewpoints. A media market consisting of four major and four minor firms is more diverse than one of merely four major firms, even if the market power indices would be virtually identical. This can be defined by the equation:

Where              n = number of firms participating in an industry;

                      Si = each firm’s market share

                       i = firm in a given industry


Using our numeric example, this index would result in:

The important thing to remember is that market definitions and indices are never perfect, and most useful if applied consistently over time and across countries. If the thermometer drops it is getting colder, whether the scale is Celsius, Fahrenheit, or Reaumur.


The following example illustrates the use of the C4, HHI and Noam Index. The market shares in this table were calculated based on circulation. The Newspaper Association of America reports the total revenues (www.naa.org).



Source: Noam, E. (forthcoming) Media Ownership and Concentration in America, Oxford University Press.


Total Revenue

In order to average the indices across industries and possibly countries, we must also obtain the total revenue number for each industry.



Weighted Averages

The concentration trends of individual industries’ concentration trends are important by themselves. Each industry experiences specific developments and transactions that create particular trends. Thus, the concentration in newspapers might rise, while that of magazines declines and that of books remains stable. To get the big picture then, it is important to look at trends for larger media categories, in this case that of “Print Publishing,” which is a weighted aggregate of the several industries that comprise the print sector. Accordingly, we aggregate the individual industries along the dimensions of industry segments and broader sectoral categories. Such measures aggregate the various industries, with weights based on size in terms of revenues.


Here is an example:


Source: Noam, E. (forthcoming) Media Ownership and Concentration in America, Oxford University Press. 


By weighing the indices in this way, we can assess the larger concentration trends within the print industry, here defined as five separate industries.



Notation Guidelines

To maintain transparency throughout the project, we use a few guidelines on how to organize the tables.


  1. All tables should be organized by company size, large to small, in the most recent (right-hand) column.
  2. Tree Structure (indentation): Media companies tend to merge – that’s one of the underlying reasons for this study. In the table, the ultimate present parent is placed at the top, with each acquired or merged subsidiary indented under it, based on chronology of purchase. In Table 1, McClatchy acquired Knight Ridder’s newspaper operations in 2006. Knight Ridder, in turn, might have ( though not in this case) aquired another company earlier, which would add a second level of indentation
  3. Please notate revenues in local currencies (e.g. using ‘€’ for European countries).
  4. The category ‘others’ should be the last line item and add market shares to 100%.




It is not realistic to expect many “official” numbers, except perhaps for telecom, where  government regulation includes the collection of industry numbers. Therefore, one must find numbers in a variety of creative ways. Search tools like Google or Nexis make life easier.  It should also be said that no two industries can be approached in quite the same way, each has its own sources and institutional peculiarities.


Here are some source categories we found useful:


-        Industry-wide reports by investment banks and financial institutions

o      Pro: detailed, in-depth

o      Con: not always easily to get, sparse historic data

-        Company reports: annual reports, financial profiles

o      Pro: detailed, historic data

o      Con: takes a long time to survey, and old ones may not be online; also: the numbers often aggregate across media lines and across national boundaries, thus requiring dis-aggregation.

-        Industry association reports

o      Pro: industry-wide overview, often good for past years, too.  

o      Con: bias

-        News articles

o      Pro: easy to search (e.g. LexisNexis), great for identifying relevant industry events (i.e. mergers & acquisitions, summarized annual reports), and often provides market share tables.

o     Con: spotty data

-        Consulting reports

o      Pro: in-depth information

o      Con: can be difficult to obtain

-        Official reports

o      Pro: relatively reliable, authoritative, with historic time series.

o      Con: will often not identify individual companies



Sometimes exact numbers are not available in a way that makes sense for our research. An international company may, for instance, not report its newspaper operations narrowed down to country-level, but rather make it available by continent. In these cases there often is other data that may help to establish a solid estimate. If a company earns in the year 2000 20% of its total revenue in publishing, and generates 40% of it in Germany, one may assume that this company yields 20% of 40% of its total revenue in German publishing. If there is no evidence that the proportions have shifted much, the same proportions might be used for 1996 if no further information is available. Similarly, in those cases when data is available for 1995 and 1998 but not for 1996 (one of the years for our methodology), then one may want to calculate the missing data point by looking at the trend between the other two years.


Reports or tables frequently provide comparative numbers, for example Orange’s market shares in several countries. In those cases, it would be much appreciated if such data would be shared with among authors, to reduce workload and increase consistency.


Back to Project Overview

Comments (0)

You don't have permission to comment on this page.