Competition and concentration - The distribution of market power in the global cellular industry

In this report, we analyse the levels of market concentration or fragmentation over the past decade across all cellular markets worldwide. The Herfindahl-Hirschman Index (HHI) is a commonly accepted measure of market concentration and is often used by telecom regulators to monitor the distribution of market power and anticipate the effect of merger and acquisition activity on competition. The values of the normalised index run between 0 and 1 and a market’s HHI is calculated based on an assessment of competitors’ market shares, enabling us to look at how the competitive landscape has evolved – and is likely to change – over time.

In the case of the mobile sector, our analysis is based on the market share of connections. We consider a market to be concentrated when a limited number of mobile operators own the largest share of total connections. In contrast, a market is fragmented when there is no dominant player and a reasonable disparity in market shares. Common terminology defines a market as unconcentrated for values of the index lower than 0.1, moderately concentrated between 0.1 and 0.18 and highly concentrated above 0.18.

Our findings show that cellular markets in Western Europe and the Middle East have higher levels of competition in comparison to other global regions. The majority of markets in these two regions have a normalised HHI below 0.1 which positions them as fragmented markets. In addition, over the past five years most of these markets have also shown a decrease in HHI, which indicates a reasonable disparity in market power among mobile operators and therefore intensifying competition.

In contrast, North America has moved from being a fragmented market to a high level of concentration over the past decade, indicating that the increased dominance of a few mobile operators in both the USA (with a normalised HHI of 0.18) and Canada (0.26) has been limiting competition. In the USA, this is mainly due to the wave of consolidation which inflated the market share of AT&T Mobility, Verizon Wireless and Sprint Nextel. These operators currently control a joint 80% share of the country’s total connections. The trend is similar in Latin America where the average concentration index (0.19) is very close to that recorded in North America.

Ranking alongside the USA as some of the most concentrated cellular markets in the world are Norway, Ukraine, China, Mexico, Kenya and Iran. These markets are typically dominated by a single player whose market power has not been significantly eroded by competition. Only Ukraine, Mexico and Iran have shown signs of decreasing HHI over the past five years, indicating a decrease in market concentration and therefore an increase in competition.

Among the most fragmented cellular markets in the world are Italy, Hungary, Australia, Venezuela, Syria and Cote d’Ivoire. With the exception of the latter, all of these markets have recorded intensifying competition levels over the past five years.

Furthermore, our findings show that the average effective price per minute in highly concentrated markets is slightly lower than in fragmented markets, suggesting that price pressures for competing operators are more intense in markets where a single operator is dominant. In addition, despite strong price pressure in the prepaid segment, the blended ARPU of dominant mobile operators in highly-concentrated markets tends to be around 30% higher than that of competing local operators. These trends indicate a certain relationship between market concentration levels and price erosion in cellular markets across the globe.

USA/Canada

Over the past decade, the USA clearly moved from being a fragmented cellular market to a concentrated market dominated by a few mobile operators. The increase in HHI indicates a decrease in competition and an increase in market concentration.

Until 2003, the USA was considered to be a fragmented market with a normalised HHI below 0.1. Between 2004 and 2005, the index increased to 0.15 positioning the country as moderately concentrated. This was due to the acquisition of AT&T Wireless by Cingular Wireless (formerly AT&T Mobility) in Q4 2004, followed by the purchase of Nextel by Sprint in Q3 2005. Following these acquisitions both operators inflated their market dominance, with their market shares increasing by 12 and 9 percentage points respectively. Furthermore, the acquisition of Alltel Wireless by Verizon Wireless in Q1 2009 increased the country’s HHI to 0.183, up from 0.155 in the previous quarter. Over this period, Verizon increased its market share from 27% to 32%. The knock-on effect of this change was a reclassification of the USA as a highly concentrated cellular market with three main dominant players - AT&T Mobility, Sprint Nextel and Verizon Wireless. At present, 80% of the total US connections are jointly controlled by these three operators, compared to 48% in 2000.

The trend is similar in Canada, which jumped from being moderately to highly concentrated in 2004 following Rogers Wireless’ acquisition of Microcell Telecommunications (formerly Fido). Between Q3 and Q4 2004, the operator increased its share of connections from 27% to 35%. Two years later, Bell’s acquisition of Aliant Mobility also inflated the country’s HHI as the operator grew its market share by 1 percentage point between Q2 and Q3 2006. Although Bell’s acquisition increased the index by just 0.018 compared to 0.041 following Rogers’ acquisition, both jumps reflected an increase in market concentration from dominant players and a decrease in competition. In 2010, Bell and Rogers jointly controlled over two thirds (65%) of Canada’s total connections, compared to 58% in 2000.

Western Europe

Over the past decade, Western Europe recorded a fairly flat HHI below the 0.1 threshold, positioning the countries in the region as mostly fragmented. The HHI regional average indicates that most countries in Western Europe tend to have well-distributed market power among all mobile operators with large groups having no significant dominance over local operators. Furthermore, most markets recorded a decline in the index, which places Western Europe as the most competitive cellular region in the world.

Ten markets were clearly fragmented in 2010: Belgium, Italy, Greece, the UK, France, the Netherlands, Austria, Germany, Ireland and Spain. These markets all had an HHI of less than or equal to 0.1, which indicates a high level of disparity in market power and thus intense competition. Over the past five years, the change in market dynamics in Spain, Germany and the UK have led to an increase in HHI but since the index did not cross the 0.1 mark, no competition concerns have been raised. In the UK, this trend is linked to the entry of 3 (Hutchison) in the market in 2003 and the recent joint venture between Orange UK and T-Mobile UK to form Everything Everywhere.

The situation is different in markets such as Sweden, Norway, Denmark, Switzerland and Portugal, which all have an HHI of greater than 0.1. These markets are considered to be concentrated since they have all had a dominant operator controlling the vast majority of total connections at some point over the past decade (TeliaSonera, Telenor, TDC, Swisscom and TMN respectively). However, over the past five years these markets have all shown a decrease in HHI which indicates that market power is being redistributed and competition is increasing. This trend has been mainly driven by the wide deployment of mobile virtual network operators (MVNOs) in those countries with a high impact on operators’ market shares.

In Denmark, the entry of 3 (Hutchison) in late 2003 and Telia’s acquisition of Orange in late 2004 has increased the HHI over the past decade. Despite recording changes above 0.025 – which according to EU rules could have raised competition concerns – both changes were aimed at hampering the dominance of TDC which controlled almost half of the market at the time. In Sweden, the entry of 3 (Hutchison) in late 2003 also had a similar desired effect to compensate the dominance of Telia (TeliaSonera). Switzerland is a typical example of a highly concentrated market with an HHI over the 0.18 threshold and a dominant player, Swisscom, controlling two thirds of the country’s total connections at present. Nonetheless, Norway appears to be the most concentrated cellular market in Western Europe with an HHI of 0.27 which is the result of Telenor’s dominance. However, over the past decade Telenor’s connections market share in the country has dropped by 8 percentage points to 58%. The increase in HHI between 2005 and 2010 is a result of the launch of new entrant Network Norway in 2008.

Eastern Europe

Since 2004 the average HHI in Eastern Europe has generally stood at just under 0.18, positioning most countries in the region as moderately concentrated, including Russia.

Over the past five years, Bulgaria, Croatia and Romania moved from being concentrated to fragmented markets. Their indices were in the range of 0.2-0.3 in 2005 but by 2010 had all reduced to 0.1 or less. This was due to strong competition from local players eroding the market shares of incumbent operators. In Bulgaria, M-Tel (Telekom Austria) controlled more than two thirds of the market up to 2004, but since then Globul (OTE) and VIVACOM have increased their joint market share to half of the country’s total connections, reducing M-Tel’s market dominance. In Croatia, Tele2 launched in 2005 and reached a market share of 12% by 2010, eroding T-Mobile’s leading share by almost 10 percentage points. In Romania, Vodafone and Orange previously controlled 97% of total connections in the market prior to 2005, but over the past five years, Cosmote (OTE) and DigiMobil (RCS&RDS) increased their combined market shares to 32%, thereby reducing the dominance of the incumbent operators.

Poland, Slovakia and Serbia are unique cases with regards to their respective evolution of market concentration. Poland recorded an HHI of 0.00 in 2006 as all three operators had a market share of 33%, and the situation would be the same today were it not for the launch of four new operators in the interim. Era (Polska Telefonia Cyfrowa), Orange (Telekomunikacja Polska) and Plus (Polkomtel) managed to keep a similar equilibrium in 2010 with market shares close to 30%, but the launch of Play (P4), Sferia, Aero2 and Centernet Mobile increased the number of smaller players in the country, making the top three operators appear more dominant. Nevertheless, with an HHI of 0.16 in 2010, the market is only moderately concentrated and does not present any competition concerns.

Prior to the launch of O2 (Telefonica) in 2007, market power in Slovakia was well-distributed as Orange controlled 55% of the market and T-Mobile 45%, reflecting an intense level of competition and no concentration. However, by 2010 O2 had reached a market share of around 13%, eroding further the share of the two incumbent operators.

Serbia was fragmented until 2007 as m:ts (Telekom Srbija) and Telenor controlled approximately 60% and 40% of the market respectively. However, the market now appears to be moderately concentrated even though Vip (Telekom Austria) launched in 2007 and reached an estimated market share of 13% in 2010. The launch of the new operator did not erode the market share of the leader, m:ts - which maintained a share close to 60% - but instead eroded that of second placed Telenor which dropped by almost 10 percentage points.

Russia appears to be moderately concentrated and has maintained an HHI of 0.17 over the past five years. Competition therefore has remained quite intense among all 12 key operators even though the market is clearly dominated by the top three – Beeline (VimpelCom), MTS (Sistema) and MegaFon – who jointly contribute 80% of total connections in the country. Consolidation in the country has yet to happen and the HHI is likely to increase if the top three operators inflate their market power by acquiring smaller local players and reducing competitive pressures.

Finally, four markets in the region qualify as highly concentrated with an HHI above 0.18, namely Slovenia, Kazakhstan, Ukraine and Uzbekistan. Each of these markets are dominated by a couple of operators that jointly control at least 80% of total connections. In both Slovenia and Kazakhstan, the launch of new entrants has not eroded the market share of incumbent operators to sufficient levels as to significantly reduce the weight of their market power. In Ukraine, VimpelCom’s acquisition of Kyivstar in Q2 2010 increased the country’s HHI by 0.03 – which according to EU rules would have raised concerns and qualified for further scrutiny.

Asia Pacific

In Asia Pacific, the majority of markets are considered to be concentrated with an HHI regional average of 0.18 in 2010. The most important change related to the regional index is linked to the large-scale restructuring of the Chinese cellular market in 2008. Between 2007 and 2008, the normalised index in China jumped from 0.15 to 0.39 clearly positioning the country as highly concentrated. Initially, the restructuring did not have the desired effect, as market leader China Mobile actually became more dominant, but over the past three years we have observed that the index has decreased by almost 0.1 - indicating an increase in competition and a reduction in China Mobile’s market power. As we have predicted in previous reports, China Mobile is set to dominate the market for a few more years and it will take time for China Unicom and China Telecom to weaken the incumbent’s market power. Nevertheless, the effect of the restructuring is moving in the right direction in terms of increasing competition.

The main reason behind China Mobile’s market share increase in 2008 was due to the transfer of CDMA users between China Unicom and China Telecom, which resulted in a drop of 39 million connections in China Unicom’s installed base. Consequently, China Unicom’s market share decreased by almost 10 percentage points to 21% in 2008, yet China Telecom only garnered an additional 4% of the market while China Mobile’s share increased to 74%. At the end of 2010, the leader’s market share stood at an estimated 69%.

Thailand, Indonesia and the Philippines all appear highly concentrated, each having an HHI somewhere between 0.2 and 0.3 in 2010. This is due to the dominance of one operator controlling almost half of each market (AIS, Telkomsel and Smart respectively) with the remaining market share spread across an average of five additional operators. In contrast, six markets appear to be fragmented, namely Australia, Hong Kong, Taiwan, Pakistan, South Korea and Sri Lanka. Those markets all had an HHI equal to or below the 0.1 threshold last year, indicating intense competition and evenly-distributed market power among key players.

India is a unique case considering the sheer size of the territory and the large number of players in every state. The country is geographically divided into 22 telecom service areas on the basis of their revenue generating potential: category A, B and C Circles (which have five to seven circles each) and three Metros (the three largest cities in India - Delhi, Kolkata and Mumbai). There are six mobile operators competing in Metro service areas, 9 in Circle A service areas, 11 in Circle B service areas and 8 in Circle C service areas. It is interesting to note that a high level of fragmentation applies to most circles and that no one operator has a considerable dominance over its competitors. For example, Tamil Nadu is home to ten operators and registered an HHI of 0.08 in 2010, compared to 0.04 for Mumbai which hosts nine operators. This high level of market fragmentation across each circle clearly reflects the intense competition occuring in India. One could therefore argue that if a wave of consolidation happens in the coming years, market power would still remain fairly well-distributed.

Finally, markets such as Japan, Vietnam and Bangladesh had an HHI averaging 0.15 in 2010, which positions them as moderately concentrated. Each of those markets were home to one dominant operator which controlled more than half of total connections at some point over the past decade - NTT Docomo, VNPT (Vinaphone and Mobifone) and Grameenphone (Telenor) respectively. Nevertheless, the decline in HHI observed in all three markets since 2005 clearly indicates that incumbent operators are suffering from competition from local players and thus market power is being redistributed.

Americas

In the Americas, half of all markets (excluding the Caribbean) can be classified as highly concentrated with normalised HHI figures ranging from 0.18 to 0.4 in 2010, the average being very close to that recorded in North America (0.19). Furthermore, compared to other regions, the Americas has registered a higher number of HHI increases (eight markets) over the past five years, indicating a decrease in competition and an increase in the market power of dominant operators. Those eight markets represent 55% of total connections in the region.

In Brazil, the wave of consolidation that happened in 2005 changed the competitive landscape from fragmented to concentrated, with the HHI increasing by 0.08 between 2005 and 2006 to 0.14. TIM and Vivo (Telefonica) had respectively four and five local operations that they consolidated under their respective brands in Q1 2006, hence reducing the absolute number of operators in the market from 17 to 10. At present, there are eight operators in Brazil, although 97% of the country’s total connections are jointly controlled by four key operators - Vivo, Claro (America Movil), TIM and Oi - which explains the high level of concentration. Nevertheless, market power is well-distributed among those four operators and there is no dominant player, with Vivo owning 29% of connections in 2010, followed by 25% for Claro, 24% for TIM and 19% for Oi. In addition, the fact that the HHI has remained fairly flat at 0.14 since 2006 reflects the intense level of competition among these key players.

Guatemala recorded a high level of concentration with an HHI of 0.18 in 2010, mainly linked to the dominance of Tigo (Millicom) who controlled almost half of the country’s total connections. In Guatemala, Paraguay and Honduras, Tigo has also maintained its market leader status over the past five years, therefore exhibiting an increase in HHI and a reduction in the level of competition in those markets.

In contrast, the low and decreasing HHI figures in Argentina, El Salvador, Bolivia, Uruguay, Venezuela and Nicaragua indicate a high level of market fragmentation and intense competition among mobile operators. All six markets have a fair equilibrium in market power. For example, the three operators in both Uruguay and Bolivia make up similar market share patterns of 40%, 40%, 20%, while in Venezuela, the shares of the three operators are split 45%, 35% and 20%. Similarly, Argentina is home to three GSM players with markets shares at 35%, 31%, 31% and an iDEN player, Nextel, with a 2% market share.

Mexico is the most concentrated market in the Americas with an HHI of 0.4 in 2010. This is due to the dominance of Telcel (America Movil) which has 70% of total connections in the country. Over the past five years, competition from Iusacell (CDMA) and Movistar (Telefonica) (GSM) has only dented Telcel’s market share by around 7 percentage points. This trend in the index reflects how Telcel’s market power is hampering competition levels in Mexico, indicating that the entry of new GSM operators is an option that the regulator could consider in order to boost competition in the country.

Middle East

In the Middle East, the vast majority of markets have been reporting a decline in HHI over the past five years, a trend which reflects competition levels in the region being the most intense globally. Most countries initially had a dominant player, however since 2005 competition has intensified across the region leading to a redistribution of market power. Only three markets – Kuwait, Iraq and Lebanon – have shown an increase in HHI over this period.

Half of the countries in the Middle East can be considered to have fragmented cellular markets, including Afghanistan, Israel, the UAE, Yemen, Kuwait, Oman, Lebanon and Syria. The most radical changes have happened in the UAE, Yemen and Oman, markets that have moved rapidly from concentration to fragmentation. These three markets were previously dominated by a single mobile operator whose market share has since reduced rapidly due to the entry of new operators. In the UAE, incumbent Etisalat witnessed the end of its monopoly with the launch of du in 2007, the new operator acquiring over one third of the country’s total connections in less than four years. Similarly, in Oman, Nawras (Qtel) launched in 2005 and has subsequently captured 45% of the country’s total connections, considerably reducing the market share of incumbent Oman Mobile (Omantel). In Yemen, new entrants Yemen Mobile and Y launched in 2004 and 2008 respectively and increased their joint market share to almost 40% by 2010, providing strong competition for Sabafon and MTN. If we apply a similar benchmark to Qatar - where Vodafone launched in 2009 to provide competition for the first time to incumbent Qtel - we expect that the newcomer’s market share will increase to 40% by 2012, compared to an estimated 24% in 2010.

Nevertheless, the large number of fragmented cellular markets only contributes 24% of the region’s total connections. Three key concentrated markets make up two third of connections, namely Iran, Saudi Arabia and Turkey.

Iran is the most concentrated market in the Middle East with a normalised HHI of 0.32 in 2010. Overall, it has taken a decade and the entry of four new mobile operators to erode the incumbent MCI’s (TCI) market share to an estimated 55%. Similarly, the launch of Mobily and Zain in 2005 and 2008 respectively reduced the market share of Saudi Arabia’s incumbent STC (Saudi Telecom) to 45%. Finally, Turkey appears to be a concentrated cellular market since Turkcell still dominates over half of the country’s total connections despite competition from more recent entrants Avea (18%) and Vodafone (27%).

Africa

In Africa, highly concentrated cellular markets contributed 40% of the region’s total connections in 2010, compared to 35% for moderately concentrated markets and 25% for fragmented markets. Overall, more than two thirds of African cellular markets are considered to be concentrated.

Highly concentrated markets all have one dominant mobile operator that controls between 50% and 65% of the country’s total connections, and in most cases this is a large operator group such as MTN, Airtel (Bharti) or Orange. However, the vast majority of these markets have recorded an increase in HHI since 2005 which denotes a decrease in competition and a sustained disparity in market power – despite the fact that all have seen the launch of at least one new operator since 2008. Kenya is the most concentrated cellular market in Africa – with a normalised HHI of 0.58 - due to the dominance of Safaricom. This operator has maintained its connections market share at around 80% since 2008 despite the launch of Orange (Telkom Kenya) and yu (Essar Telecom). Zambia, Ghana, Niger and Sudan are the only highly concentrated markets in Africa that have recorded a decline in HHI over the past five years.

Most of the fragmented markets in the continent are located in Western Africa, namely Cote d’Ivoire, Burkina Faso, Benin, Togo and Liberia. The rest are widespread across other sub-regions: South Africa in Southern Africa; Algeria in Northern Africa; Angola, Chad and Central African Republic in Central Africa; and Mozambique and Somalia in Eastern Africa. All these markets (with the exception of Cote d’Ivoire, Benin and Liberia) have recorded declines in HHI over the past five years.

In the near term, we anticipate that the competitive landscape is likely to change in a number of markets where there is still room for connections growth and where increasing levels of market concentration have been limiting competition. Such markets include Cameroon, Malawi, Sudan and Zimbabwe. Although the levels of development and the challenges differ, these four cellular markets are all likely to be transformed by regulatory initiatives in the coming years.

 Price competition trends

To fully understand the level of competition and changes in market power across global cellular markets, we have to correlate the Herfindahl-Hirschman Index results with pricing trends.

Effective price per minute (EPPM) is defined as the average revenue per user (ARPU) divided by minutes of use per user (MoU) per month, taken as a rolling average over four quarters. We have analysed changes in this operational indicator in both fragmented and concentrated markets. The results show that the average EPPM in highly concentrated cellular markets is lower than in fragmented markets, suggesting that price pressures are more intense in markets where a single mobile operator dominates. In both cases, price erosion has occurred at a similar rate over the past decade - with EPPM decreasing from US$0.25 in Q4 2000 to US$ 0.11 in Q3 2010 for concentrated markets and from US$0.33 to US$ 0.13 over the same period in fragmented markets. However, the delta between EPPM in concentrated and fragmented markets has declined from US$ 0.08 in 2000 to just US$ 0.02 in 2010.

In addition, we have analysed the delta between the dominant Tier 1 operator’s blended ARPU against a local Tier 2 competitor (where Tier 1 is defined as an operator with a market share greater than 20% and Tier 2 those under 20%) across 17 highly concentrated markets. The results show that in these markets, despite strong competition Tier 1 operators’ ARPU tends to be on average around 30% greater than that of Tier 2 operators. The delta has remained flat between Q3 2009 and Q3 2010 which denotes a certain relationship between market power and price erosion.

Overall, price competition is more intense in concentrated markets where new entrants or local operators often rely on discounted prepaid voice offers to reduce the dominance of incumbents and acquire substantial market shares. This approach is common in highly concentrated emerging markets. Operators in more mature markets however tend to simplify their offers and pricing (e.g. SoftBank Mobile in Japan) or rely on innovative bundle deals (triple-play) and/or MVNO strategies to address new or existing demand with more affordable voice services. In addition, economies of scale and customer loyalty programs play a critical role in price competitiveness, especially in markets where growth is driven by handset replacement. These factors clearly contribute to decreases in HHI in both concentrated and fragmented markets.

In a concentrated cellular market like China, both China Unicom and China Telecom are slowly eroding China Mobile’s connections market share but the incumbent’s dominance is still expected to remain strong in the near term. All three operators have kept blended ARPU stable over the past year with China Mobile at around US$ 10, China Telecom at US$ 8 and China Unicom at US$ 6. Furthermore, all three mobile operators have registered an EPPM around US$ 0.02 in Q3 2010 which indicates the level of price elasticity that the incumbent is maintaining over its competition.

However, in a fragmented market such as Italy with no dominant player, ARPU is subject to more seasonal activity and operators’ EPPM follows the same steady trend. Despite the quarterly seasonal fluctuations, Vodafone has maintained its ARPU at around US$ 27 since 2008, compared to US$ 25 for TIM and US$ 22 for WIND. Over the past year, Vodafone’s EPPM stood at around US$ 0.23 compared to US$ 0.18 for TIM and US$ 0.13 for WIND, showing that changes in price erosion happen at a slow pace in fragmented markets as the mobile operators’ life cycle is driven by the balance between marginal cost and profitability. In addition, these differences in EPPM indicate that consumers in Italy have access to a wide choice of voice tariffs, suggesting that the level of market fragmentation and maturity has a further direct impact on price elasticity.

Methodology

The Herfindahl-Hirschman Index (HHI) is a commonly accepted measure of market concentration and is often used by telecom regulators to monitor the distribution of market power and anticipate the effects of merger and acquisition activity on competition. The values of the normalised index run between 0 and 1 and a market’s index is calculated based on an assessment of competitors’ market shares, enabling us to look at how the competitive landscape has evolved – and is likely to change – over time.

In the case of the mobile sector, our analysis is based on the market share of connections. We consider a market to be concentrated when few mobile operators own the largest share of total connections. In contrast, a market is fragmented when there is no dominant player and a reasonable disparity in market shares.

The index itself is calculated by taking the sum of the squares of the market shares of the top 50 firms competing in the market, or if there are fewer than 50 firms in a given sector, all companies are included in the calculation. For example, if a market has three mobile operators with market shares of 40%, 30% and 30%, HHI would be calculated as 0.402 + 0.302 + 0.302, resulting in an index of 0.34. In this form, the index ranges from 1/n to 1 where n is the number of firms in the market, but in this analysis we have normalised the index to within the range 0 to 1.

The index of each geographical region in our analysis is then based on the weighted average of the HHI within the regions’ respective countries.

Common terminology defines a market as unconcentrated for values of the index lower than 0.1, moderately concentrated between 0.1 and 0.18 and highly concentrated above 0.18. In addition, regulators are also interested in the relationship between price and competition as prices should typically drop when competition increases, thus boosting consumer choice. Typically, an increase in HHI indicates a decrease in competition and an increase in the market power of a dominant mobile operator. Conversely, a decrease in HHI indicates an increase in competition and a decline in the power of that operator.

Different measures are used by different regulators when determining if a merger meets competition regulation guidelines. In the USA, a merger raises concern if it increases the HHI by more than 0.01 (Horizontal Merger Guidelines), while in the EU the guidelines state that a 0.025 increase in markets where the index is already above 0.1 must be flagged for further potential investigation (Guidelines on the Assessment of Horizontal Mergers). It is unlikely that a merger would adversely impact competition in a market where its HHI is below 0.1 post-merger.

Criticisms

When applying the Herfindahl-Hirschman Index to mobile operators’ market share data, a number of limitations should be considered:

• The HHI is dependent on a proper definition of a market which hinges on the notion of substitutability. The index does not balance the adjacent niche segments that operators address outside of their main cellular activities – which can limit substitutability to some extent. For example, a convergent operator will probably have more market power as it addresses demand for fixed telephony as well as broadband and TV services, rather than a traditional operator which mainly focuses on mobile voice services.
• The market definition also depends on the geographical scope. Mobile operators could cover different areas across the territory, and are thus not directly competing against each other. If a mobile operator covers exclusively one city or community, it is not necessarily substitutable with an operator in another city. This is particularly true for markets like the USA, Canada, Brazil or India where many operators (including local players) tend to cover a few states spread across a wide territory. For those markets, the HHI could be calculated at a state level.
• In the European Regulatory Framework, the European Commission defines seven distinct markets in retail telecommunications. It is clear that a deeper analysis using HHI would need to go further than computing a single HHI figure for a country based on connections market shares.
• In some markets, certain operators only report active subscribers whilst competitors report all registered SIM cards, thus distorting the distribution of market shares and as a consequence the level of market concentration according to the HHI method.
• The HHI does not take into account various levels of network deployments, legacy systems and priorities in terms of investments (capex) and cost savings (opex). In addition, factors such as range and quality of services, brand awareness, cultural preferences and GDP per capita are not considered.
• Finally, the Herfindahl-Hirschman Index is one commonly used diagnostic tool that measures market concentration, but to fully understand the level of competition or market power, other factors should be taken into account such as the Lerner Index - and its derived version, L2 - which correlates changes in market share with changes in price elasticity and supply elasticity.