With a Single Market Act[1], the European Commission are taking initiatives to strengthen a single market for different services in the European Union including energy, transport and electronic communications. This is a positive signal towards the harmonisation of regulation in the concerned markets, since it will push for more regulatory transparency and give businesses more confidence in driving their activities.

However, caution should be taken regarding the interpretation of motives underlying these new initiatives. Regulation harmonisation is different from market homogenisation. The first principle means the promotion of transparency by having the same regulatory frameworks across Member States. The latter could result in large companies taking over smaller companies thanks to structural advantages like economies of scale to expand their pan-European activities and profit from their market power, to the detriment of consumers and other industry stakeholders. The European Union is made up of 27 different Member States, each one with their own rich history, culture and different market characteristics such as consumer habits. This is a strength that should be promoted to maximise consumer welfare, innovation and each Member’s comparative advantages in various fields, whether in energy, transport, electronic communications, food processing or entrepreneurship. 

The same is true for the mobile communications markets, which has seen calls for measures towards a European single market by the Digital Agenda for Europe[2]. The mobile market consists of a variety of operators in the Union’s national markets. Each market typically consists of three or four operators with the leading two defined as first entrants and the others, later entrants. This structure allows for fringe competition from the later entrant operators who hold on average 25% of subscription market share[3]. On one hand, regulation harmonisation of these markets would mean having the same regulatory framework, for example the same asymmetric measures for later entrants in order to preserve the competitive dynamics provided by these players. On the other hand, market homogenisation would very likely transform the variety of national markets into a handful of pan-European operators who consolidate the market by M&A activities, especially by acquiring smaller, later entrant operators.

This article highlights the importance of making a distinction between the two concepts, with a numerical illustration from the European mobile markets. We present our findings on the potential impacts of consolidation in the case of market homogenisation. First, it was found that there would be a significant increase, about 66%, in the market power of European first entrants[4]. Second, this entails an equally important loss of consumer welfare. The scale of the loss of consumer welfare if later entrant operators do not exist, in the period of 2011-2020 would amount to the order of hundreds of billions of Euros. This is by no means negligible and implies important policy considerations for the Digital Agenda. 

The paper is structured as follows. In section 2 we provide a review of economic literature on why asymmetric measures are needed to address the imperfect competition imposed by the unequal sizes between first and later entrants. We then proceed to test the importance of challenger operators in the market by an economic model that calculates the increase in market power of European first entrants and then the loss of consumer surplus in the absence of the challengers. The methodology of the model will be explained in details in section 3, including a review of other quantitative or qualitative studies that have similar goals of quantifying the consumer surplus in mobile telecommunications market. Section 4 presents and discusses the results of our model. A short conclusion is provided at the final section.

Literature review on the benefits of small local operators

§Competition and the promotion of economic efficiency

Generally people use the term competition in reference to markets in which firms must compete strongly for sales. Each firm attempts to gain customers at the expense of other firms, and through their competition, market price and quality are affected to the benefit of those customers. The extreme case for competition is called perfect competition, which is the situation in which no individual supplier in the market can individually influence the market price (i.e., each firm is a price taker) and each supplier can sell however much it wants at the prevailing market price. While the telecommunications industry almost never fits this perfectly competitive market paradigm, it is possible for the telecommunications industry to exhibit efficient, or effective, competition, which is often defined as the situation where:
 
- Buyers have access to alternative sellers for the products they desire (or for reasonable substitutes) at prices they are willing to pay,

- Sellers have access to buyers for their products without undue restraint from other firms, interest groups, government agencies, or existing laws or regulations,

- The market price of a product is determined by the interaction of consumers and firms. No single consumer or firm (or group of consumers or firms) can determine, or unduly influence, the level of the price, and

- Differences in prices charged by different firms (and paid by different consumers) reflect only differences in cost or product quality/attributes.

 In such a market, suppliers are able to freely enter and exit, there are a large number of firms, the market price reflects marginal costs, and the service quality and services provided are determined by market forces. The establishment of effective competition therefore should result in the presence of these characteristics.

The goal of effective competition is to promote economic efficiency. Economic efficiency incorporates three concepts of efficiency: Allocative efficiency, technical efficiency, and dynamic efficiency. Allocative efficiency is the situation where the economy’s resources are put to their most valuable uses. It is important because it helps maximize the value that customers receive from the services provided. Economists refer to this as maximizing net consumer surplus, which is the difference between the prices paid and the value customers expect when purchasing the service.

When a firm exercises market power, there is an efficiency loss to the economy that results from resources being misallocated. The misallocation comes from the firm exercising market power restricting its output to increase profits. When output is restricted in this way, the resources that should be used for this market are used in other parts of the economy where economic value is lower. For example, suppose that a country allowed its mobile services to be provided by a monopoly. The monopolist, if it wanted to maximize its profits, would limit the size of its customer base so that it would sell service only to those customers willing to pay high prices. As a result of this supply restriction, some marketers, engineers, managers, etc. who would be very good at providing mobile services are not put to work providing the services. Instead they work in other jobs where they provide less value to the economy. This is a loss in allocative efficiency because the economy would be better off if these marketers, engineers, managers, etc. worked in the telecom sector.

In addition to this misallocation of resources, some of the consumer surplus that would be obtained under a competitive environment disappears. Many economists believe that the loss of allocative efficiency is the primary detriment of market power. To measure misallocation, data on the costs of production, specifically how fast average cost increases or decreases with output; the degree of mark-up in price over marginal cost; and elasticity of demand for the product.
 
Economic efficiency also includes internal efficiency, also called productive efficiency or x-efficiency, whereby costs are kept at a minimum for all levels of output and service quality. X-efficiency addresses the tendency for managers of firms with market power to become less vigilant about keeping costs as low as possible. In large dominant companies, positive profit hides inefficiencies better than possible in a truly competitive firm. X-inefficiency occurs when employees do not work at maximum levels, and when inputs are wasted (for example when the firm buys more inputs than it would have had the managers taken greater care to contain costs). The result is actual costs that exceed the minimum possible cost. This difference is the amount of x-inefficiency within the market (Leibenstein, 1966).

Finally, economic efficiency includes dynamic efficiency, which occurs when product and production innovations occur at their most appropriate rates. Product innovation is occurring at the proper rate when the extra costs of developing and implementing innovation are equal to the extra value created by the new products. Production innovations occur at the appropriate rate when the extra costs of the innovation are equal to the production cost savings that they create. In short, effective competition results in a competitive balance in which no firm remains dominant and the industry exhibits efficiency of all types.
 
Effective competition is desired to promote customer benefits through improved efficiency; however when competition is ineffective, regulatory involvement can improve efficiency. This is not to say that regulatory oversight always improves the situation – regulation can be imperfect just as markets can be imperfect – so a difficult decision for policy makers and regulators is to know when regulatory intervention is appropriate and what form of regulatory intervention can actually benefit customers.

§The market structure of the mobile telecommunication industry: an oligopoly with fringe

In practice, the mobile telecom industry in the US as well as in European markets can be described as an oligopoly with a fringe (see e.g. Benzoni et al. 2011). An oligopoly corresponds to a market structure of imperfect competition, characterized by the fact that the market is dominated by a small number of producers so that one company alone has any power to influence the price as well. The fringe means that there exist size asymmetries between operators. Some big leaders firms coexist with smaller challengers. One can observe that this asymmetry is mainly explained by a first mover advantage. From the preceding discussion on economic efficiency, it is quite intuitive to consider that the existence of the small operators is clearly beneficial. Dewenter (2007) states that first mover cost advantages typically result from structural advantages, such as economies of scale and learning curves, higher degrees of advertising appeal or better access to input markets. Therefore, as Peitz (2005) stated the late entrants initial position with respect to coverage, installed consumer base, quality of service and reputation is different to the incumbent’s position. This creates asymmetric market environments. The solution to eliminate the potential negative effects of asymmetric market on competition can be to apply asymmetric price regulation in order to try and address the imbalances in the mobile market.

§A large body of economic literature found structural advantages for first entrants in the mobile telecommunications market

There is a long tradition in economics to study the question whether incumbent firms have a first-mover advantage vis-à-vis later entrants. Von Stackelberg (1934), for example, showed already in the context of quantity-setting firms that a leader (first mover) is able to get a larger market share and higher profits than a follower (second mover). There are several reasons why first-mover advantages may arise. In the economic theory (see, e.g., Mueller, 1997) and strategic management (see, e.g., Lieberman and Montgomery, 1988) literature the following factors are mentioned. On the demand side, there is the pre-dominant importance of switching costs (Klemperer, 1987). Switching costs can arise (i) from the fact that consumers have to make some initial investment in adapting to a seller’s product or services, (ii) from firm specific learning on how to use the product (habit formation), or (iii) because of contractual costs imposed by the firm. With switching costs, firms that already exist in the market when consumers have to make their first adoption choice benefit as later firms have to convince consumers to switch. Other sources for first-mover advantage on the demand side include network externalities and uncertainty about the quality new firms offer (Schmalensee, 1981). On the supply side, there are similar factors working to the advantage of early entrants in the market, most notably sunk costs and economies of scale (Schmalensee, 1982) and cost efficiencies through learning by doing. There are also possible strategic effects implying that firms find it more difficult to gain market share in markets where already many firms are active. There is also a large empirical literature on first-mover advantages, mostly in management and marketing journals (see, e.g., Urban et al., 1986, Kalyanaram and Urban, 1992, and Robinson et al., 1994, for an overview).

 

An important part of this literature studies how market shares at a particular moment in time depend on the firm’s position in entering the market: what is the typical (long-run) market share of the pioneer firm, the second entrant, etc.? These studies are based on cross-sectional data from many different markets. Kalyanaram and Urban (1992) are the first to study a combination of cross-section data and time series. They also consider the impact of factors such as price differences, relative advertising expenditures, relative quality perception, etc. on the development of market shares. Their model allows a comparison in the rate of convergence to an asymptotic market share level by later entrants. The vast majority of empirical studies find that market pioneers have substantially higher market shares than later entrants.

§To effectively promote competition and economic efficiency, remedies are needed to address the potential anti-competitive practices resulting from the market structure

The analysis above defines the foundation for most specialists to introduce symmetric price regulation measures. This implies to set different MTRs between operators, generally in favour of late market entrants. According to ERG (2008), asymmetry refers to charge levels, rather than to differentiation of regulatory remedies e.g., price control versus fair and reasonable. In recommendation study, the European Commission (2009) recognized that in certain exceptional cases asymmetry might be justified by objective cost differences outside the control of the operators concerned. For example, encouraging the growth of a new entrant on the market which suffers from a lack of scale due to late market entry, necessitate asymmetric MTRs.

In fact, asymmetric charges in favour of smaller operators, allow higher expected profits in the short term for new entrant and in the long term results a more intense competition to the benefit of consumers. In other words, it is appropriate for a regulator to allow asymmetric termination rates. In such circumstances, asymmetric MTRs contribute to dynamic efficiency and let new entrants to invest more in networks hence, favour infrastructure based competition. The spectrum allocation licences that permit operators to build mobile networks around the country are awarded sequentially. Thus, it automatically generates first and late entrant players in the market. It is generally, accepted that the mobile communications market where licences have frequently been granted sequentially is characterised by first mover advantages (Dewenter, 2007).

First mover advantages stem from early adoption by users, allowing a firm to capture a large market share early on. Thus, by the time competitors can enter the market, the first-mover will have already established advantages in brand-loyalty or awareness as well as cost advantages in distribution and/or infrastructure systems (Benzoni and Geoffron, 2007). Studies also support this view. For example, Bijwaard et al. (2005) concludes that depending on specific entry conditions, it seems fair to conclude that the first entrant may still gain a large market share and that subsequent entrants encounter more difficulties in gaining market share.
In another study, for the Swiss regulatory authority which compares the development of the Swiss telecommunications market with the rest of Europe, the WIK Consult also stated that a sequential award of mobile licences negatively impacts competition dynamics because of major disadvantages of later entrants (especially, linked to network coverage and to the high switching costs for business customers of early entrants) compared to early entrants (Benzoni and Geoffron, 2007). Several factors have been identified which put late market entrants at an economic disadvantage in supplying mobile telecommunication services including:

- The cost advantages of early entry due to economies of scale, favourable selection of base station sites and favourable access to spectrum.

- Switching costs faced by consumers moving from an established operator.

- Price-induced network effects whereby the price differential between on-net calls and off-net calls makes it attractive to consumers to use the network with the larger subscriber base (Peitz, 2003).

 

• appear to be really important for the efficiency of mobile industry is to maintain an effective competition. Efficient firms will see the need to innovate and become even more efficient. Small operators clearly contribute to innovation in cost-reducing activities and prices which are expected to reflect costs in the industry. If this dynamic disappears, at the end the ultimate losers will be customers as a whole.

Methodology

§Concentration and market power

The Hirschman-Herfindahl Index (HHI) has long been a traditional tool to measure market concentration, in research as well as in assessing competition cases by regulatory authorities (see for example Albarran and Dimmick (1996), or the European Commission’s 2004 “Guidelines on the assessment of horizontal mergers” and the and Federal Trade Commission’s 2010 “Horizontal Merger Guidelines”). It is calculated by summing the squares of each firm’s market share. The index ranges from close zero, in the case of a perfectly competitive market with many suppliers and well-informed consumers, to 1 (or 10,000 if market share is expressed in percentage) as in the case of a pure monopoly.

Hendricks et al. (2007) and Gans (2007) proposed an approach to evaluating the competitive effects of horizontal and vertical mergers using the HHI. By applying the Cournot model in the relevant market, the HHI can be used to calculate the industry average price-cost margin, or the Lerner Index, which reflects the industry’s pricing power in relation to price elasticity of demand. In this case, the level of welfare distortion in a market can be said to be based on the level of market power resulting from a merger. The Lerner index is calculated by dividing the sums of squared market shares (the HHI) by the market demand elasticity.

§Consumer surplus

There have been other academic as well as consulting studies that have calculated consumer surplus, all of which are based on data from the past. The goal of study is to quantify the consumer surplus that would occur in a future scenario. Our database will therefore be under certain data availability constraints and a number of important hypotheses. However, a look at other methodologies would give a good guide to the base of our own model. 

One study that gives a general background to different methodologies of calculating consumer surplus is Kim and Kim (2010). The authors analysed the competitive implications of horizontal mergers between Korean MNOs. Most notably, the merger between Korea Telecom Freetel and Hansol in 2000 and the acquisition of STI by SKT in 2001 led the HHI to increase from 2577 to 4113. The article explained and measured the change of consumer surplus consequent to these mergers, with three types of calculations: Marshallian, exact, and ACM Consumer Surplus.

The Marshallian consumer surplus is based on the ordinary demand curve, using the log-linear functional form that integrates price, income, a vector of variables that affect the demand, and the price and income elasticities of demand.

The exact consumer surplus is derived directly from an estimated ordinary demand curve. This type of calculation captures the implicit change in real income due to price changes through including the income elasticity of demand in the consumer surplus equation.

The ACM (Australian Communications Authority) methodology is different to the previous two in that it includes a parameter to account for the diffusion of technology, therefore accounting for a shift in the demand curve either by a parallel shift or a pivotal shift in the slope of the curve.

The three methodologies yields different results, with the first two result twice as much change in consumer surplus than the ACM method.

It should be noted that the study above is based on past mergers that took place within one market. The authors used data that come directly from the market before and after the mergers, and therefore could build a demand curve integrating reliable statistics including price and income elasticities of demand within that market. The scope of our own study encompasses 15 different markets with future forecasts of demand and consumer surplus, the model must therefore be simplified and based on certain assumptions to make up for the unavailability of real market data.

A paper by Ovum in 2010, “The benefit of the wireless telecommunications industry to the Canadian economy”, details a simple method to construct the demand curve. Ovum used past data for the post-paid market in Canada between 2000 and 2009 to estimate the lower bound demand curve, with the revenue per minute as the price variable and the billable minutes as the quantity demanded in the mobile market as a whole. Similar use of variables have also been employed in other papers, such as ACMA (2009), with average revenue per call minute as the price variable and call minutes per subscriber as quantity variable. 

One obvious problematic issue of the Ovum’s method is that the current (i.e. of 2010 market) demand curve extrapolated from a set of annual data between 2000 and 2009. Given advanced technological changes and other non-price factors’ changes (such as consumer income, consumer habits, the decline of fixed telephone usage as a substitute goods, etc.) that took place during this period, the demand curve is likely to have shifted to the right, as also rightly pointed out by Ovum.

Since the calculation of consumer surplus invariably requires the construction of the demand curve, we have used average revenue per call minute as the price variable and minutes of use as the quantity variable. These data are forecasted based on data of previous quarters right up to the fourth quarter of 2010. We aim to improve the methodology by mapping a demand curve for mobile services for each year, thus accounting for a demand change in mobile services from year to year in the period of 2012 - 2020.

§Data and variables

The data we use in this paper is compiled from different sources: the European Commission 15th Progress Report on the Single European Electronic Communications Market, European operators’ annual reports, and a recent study by TERA Consultants “A comparative study of the national mobile telecommunications markets in the European Union”. The different elasticities of demand are sourced from various academic papers as discussed further below.

To calculate the change in consumer surplus, we construct a demand curve for the mobile telecommunications market before and after the acquisition of later entrants by first entrant operators. The demand curve of a typical consumer is built for each of the EU15 market each year between 2012 and 2020.

The price level is proxied by the Average revenue per minute (ARPM). It is calculated by dividing Average revenue per user (ARPU) by the monthly minutes of use (MOU), both sourced from TERA Consultants 2011. A weighted average national ARPU is calculated based on the market share of each operator in the market. For our study, we have chosen the quarterly ARPU data between Q1 2008 and Q4 2010 as a base for our ARPU logarithmic forecast of 2012-2020. 

The quantity demanded is proxied by the monthly minutes of use (MOU). As for MOU forecast of 2012-2020, we have based the MOU growth forecast on a UMTS Forum study and MOU evolution between 2008 and 2010. The UMTS Forum document, “Mobile traffic forecast 2010-2020”, outlined the main trends and drivers of mobile broadband for years 2010-2020. It indicated the general trend that “voice traffic growth should remain limited compared to traffic growth from 2010 to 2020”.

Furthermore, based on MOU quarterly evolution between Q1 2008 and Q4 2010 across the EU15 countries, it is observed that MOU growth is indeed quite limited, even flat for several countries including Austria, Belgium, France, Spain.

We have therefore taken the stability scenario as the base case for MOU growth forecasts. For other scenarios, we have calculated the change in consumer surplus based on an annual MOU growth between 0% and 3%.

As for the price elasticities of demand, we have taken two types of estimates. The first is the market-specific elasticities of the EU15 countries calculated by Grzybowski (2004) (Table 1: Price elasticities of demand of EU15 countries). The research in this article is based on extensive data from the yearly reports on the implementation of the Telecommunications Regulatory Package available on the website of the European Commission and from the cross-country data on the telecommunications industry by the International Telecommunications Union, among others.

Table 1: Price elasticities of demand of EU15 countries

Country	Price elasticity
Austria	-0.352
Belgium	-0.337
Denmark	-0.285
Finland	-0.196
France	-0.324
Germany	-0.308
Greece	-0.375
Ireland	-0.262
Italy	-0.288
Luxembourg	-0.177
Netherlands	-0.355
Portugal	-0.413
Spain	-0.282
Sweden	-0.292
United Kingdom	-0.256

Source: Grzybowski (2004)

Some particular studies have estimates of price elasticity of demand in a higher range for high income or European countries, between -0.47 and -1.1 (Growitsch et al 2010, Dewenter and Haucap 2008, and Madden et al 2004, see Table 2: Price elasticities of demand from other studies). We therefore have taken -0.55 as the symbolic average of these estimates for the second figure of price elasticity of demand for mobile telecommunications services in the EU15 countries.

Table 2: Price elasticities of demand from other studies

	Country		Study		Price elasticity
	UK		DotEcon (2002) – as reported in Competition Commission (2003)		-0.37   -0.40
	UK		Frontier Economics (2002) – as reported in Competition Commission (2003)		-0.54  -0.30
	UK		Holden Pearmain (2002) – as reported in Competition Commission (2003)		-0.08
	UK		Rohlfs – as reported in Competition Commission (2003)		-0.30
	UK		CRA – as reported in Competition Commission (2003)		-0.45
	UK		Competition Commission (2003)		-0.30
	USA		Rodini, Ward and Woroch (2002)		-0.43
	64 countries with relatively high per capita GDPs		Ahn and Lee (1999)		-0.25 (connection fee – not statistically significant)   -6.1 (monthly charge)
	Japan		Okada and Hatta (1999)		-3.36
	New Zealand		Danaher, P. (2002)		-0.06 -0.10 -0.20 -0.35 (Increasing access price and usage price)
	EU15 + Czech Republic		Growitsch et al (2010)		-0.52 -0.61 (long term)   -0.097 (short term)
	56 countries		Madden et al (2004)		-0.55 (global)   -0.53 (high-income countries)
	Austria		Dewenter and Haucap (2008)		-0.47  -1.1
	US		Ingraham and Sidak (2004)		-1.12 -1.29
Source: Vodafone New Zealand (2003), Growitsch et al (2010 (this study only gives the elasticity of the whole sample and not country-specific elasticities), Dewenter and Haucap (2008), Ingraham and Sidak (2004). In grey are the studies used for second type of elasticity in model

We have also checked these elasticity estimates used in our model against other studies and have found them to be within the range of the estimates of these studies (Table 2).

Again, the studies of Table 2 are based on past data, sometimes well past data. It is not possible to have elasticity estimates of our sample countries of more recent periods, such as the ones of 2009 or 2010.

The task of estimating the price elasticity of demand in the mobile communications market is a tricky issue. Rapid technological development take place very often and are likely to change the nature of the market, in particular the consumer behaviour. Many other economic factors that affect demand such as pricing, handset subsidies could also make it difficult to have a correct estimate of the elasticity. Therefore, the results of our model should be treated as indicative rather than an exact quantification of the change in consumer surplus.

Consumer surplus is calculated based on the consumer surplus of a typical consumer in a national market, in each of the EU15 country, multiplying by the total number of consumers in that market. With the demand curve, the ARPM as the price level and the MOU as the quantity demanded, the consumer surplus is the triangle formed by these three factors as described in Figure 1 below.

Figure 1: Consumer surplus in a generic mobile market

Note: i = each of the Member State in the EU15, t = each year in the period 2012-2020

 

Figure 2: The change in consumer surplus due to an increase in price

Source: TERA Consultants, for illustration purposes only

The change in consumer surplus is the difference in the consumer surplus before and after the acquisition of the later entrant operators.

§Hypotheses

To make up for the unavailability of future data, we have had to make important assumptions that are detailed and explained below.

• Hypothesis 1: Later entrants’ acquisition by first entrants

This is a trend that has been observed in the US mobile telecommunications market in particular in recent years (GAO 2010 or FCC 2011). The largest mobile operators in the American market have been acquiring smaller operators in order to expand their nationwide market share, as shown in Figure 3, which highlight some major mergers and acquisition in the market between 2000 and 2010. 

The M&A strategy lead by the top operators have consolidated the US market such that the four leading nationwide operators hold 90% of subscriber market share in 2009 (GAO 2010) and the concentration index HHI has went up 30% within 6 years from 2151 to 2811 in 2009 (FCC 2011). In addition, an analysis of how market share has evolved of the top two operators vs. small operators clearly indicates a trend that top operators’ market share expand due largely to M&A strategy whereas small operators’ market share has been diminishing (Figure 4). This decline can be explained by either exits or acquisition by larger operators. 

 

Figure 3: Major M&As in the US mobile market in the last decade

Source: Reproduced from GAO 2010

 

Figure 4: Market share of top operators vs. small operators[5]

Source: Operators’ annual reports

A similar scenario is therefore can be envisaged in the European single mobile. We have taken the assumption that with the regional markets in the EU opened up to form the single market, first entrants in their quest to expand market share and propped up by their financial strengths, would engage in M&A activities and acquire the smaller operators, which is not far from what has happened in the US market.

• Hypothesis 2: After the acquisition of the later entrants, the price level increases to the first entrants’ price level

We have taken the assumption that prices would increase after the acquisition when the market share and thus market power of first entrants would be significantly more important. This is supported by a wide body of theoretical and empirical research. Andini et al 2006, for example, found that market share has a positive and significant impact on prices and margins, based on an international sample of 177 mobile operators in 45 countries over the period of 1994-2004. More concretely, across the sample, one percentage point increase in the market share results in a 0.23% increase in the price, or ARMP and 0.7% increase in margins. In the Eurozone countries sub-sample, the results are even more important: one percentage point increase in average market share results in a 0.46% increase in ARPM and 1.1% increase in average margins. 

This hypothesis also corresponds to market-power seeking behaviour of large MNOs. Benzoni et al 2011 observed that pan-European operators, such as Vodafone, Telefonica, Deustche Telecom and France Telecom, have aimed to achieve first-entrant status and thus first-entrant pricing power in national European markets. When they entered new markets and failed to have first-entrant position, they either merged with another operator to increase market power or left the market entirely. This strategy ensures that they will be able to make high profit margins as price makers and not price takers.

This is also the reason why operators always claim impressive amount of cost savings to justify M&As strategies. However, when the M&As are complete, consumers have not witnessed the cost savings translated to lower retail prices. Not only it is because these operators wish to retain high profit margins but also there exists very little economies of scale for mobile network operators. A large proportion of their cost consists of retail distribution network and mobile local loop, for which economies of scale hardly applies. With respect the retail distribution network and the mobile local loop, costs are mostly variable with the number of subscribers. In the end, the ultimate motive underlying M&A strategies is the maintenance and expansion of pricing power and profit margins.

In our model, we have defined the price increase after the acquisition as the price level of first entrant operators, which is in general higher than the price level of the market where both first and later entrants exist and compete to win customers. The higher market price level will inevitably lead to a lower level of demand for mobile services, which we account for by using the demand curve and the price elasticity of demand. 

• Hypothesis 3: The demand curve is linear

This hypothesis is also employed in other studies, for example ACMA 2009.

We make use of a Marshallian demand curve in which income and the prices of other goods are assumed to remain constant. Thus any change in the price level of the mobile telecommunication services, or ARPM, is reflected directly in a proportional change in demand, or MOU, based on the price elasticity of demand. 

• Hypothesis 4: Price elasticity of demand is constant over the period 2012-2020

This is a strong assumption and it poses a major issue for our model. However, as discussed earlier, a price elasticity of demand in applied economics is never a straightforward question. It is different along the demand curve, across national markets and over time. However, for the task of estimating the change in consumer surplus due to a price change, we have tried to choose the best option by integrating market-specific elasticities to our model and verifying our numbers against other studies.

Furthermore, our study is the only one so far to our knowledge that aims to quantify consumer surplus in a future period. The results are therefore supposed to give an indication, an order of economic importance of the change, rather than an exact amount of consumer welfare.

§Change in HHI and Lerner index

The HHI calculation is based on the analysis of the number of subscriptions of all MNOs within the EU15 countries in 2010. It is the sums of squares of market shares of each operator.

The objective is to calculate the weighted averaged Lerner Index in Europe following two scenarios: a Lerner Index with all operators currently in the market, the other is a Lerner Index based on a fictive European market distributed to first movers only.

In the first case, all market shares are squared then summed to calculate a national HHI. The Lerner Index is obtained by multiplying this HHI by the national elasticity of demand.

In the second case, the market is supposed to be composed only of the current first movers. The remaining subscribers initially attributed to later movers are spread among first movers according to their market shares. The calculation of the Lerner Index follows thereafter the same process.

§Change in consumer surplus

The change in consumer surplus is calculated by taking the difference of consumer surplus before and after the acquisition. This is done for each of the sampled countries over the period 2012-2020.

The difference in consumer surplus before and after the acquisition is calculated using the demand curve and the corresponding price level

with p₁ = ARPM₁, the price level before the acquisition, p₂ = ARPM₂, the price level after the acquisition, or the price level of first entrant operators.

Result analysis and discussion

§First entrants’ market power would increase by two thirds and loss of consumer surplus could amount to 300 billions euros

In the context of the 15 European mobile markets, we measure the change in market power and the scale of the consumer welfare loss that would result from the fringe competition, which is also what would have been the reduction of consumer welfare if later entrant operators do not exist, in the period of 2012-2020.

First, it was found that the HHI and the Lerner Index would increase by around 65%. This indicates an important degree to which the disappearance of fringe competition would reinforce the market power of first movers, which is also an indication of the degree to which consumer welfare could be lost.

Table 3: The change in HHI and Lerner Index with and without later movers

	With all operators	Without later movers	Variation
Weighted HHI	­­­3107	5095	+64%
Weighted Lerner Index	1.036	1.717	+66%

Source: TERA Consultants

 

Secondly, we calculated that the loss of surplus would be between €170 billion and €300 billion depending on different elasticity values. Clearly, this is by no means negligible and implies important policy considerations for the Digital Agenda.

Table 4: The loss in consumer surplus without later movers

Change in consumer surplus (nominal billions of Euros)	Different elasticities*	­­­-298,6
	Elasticity = -0.55**	-173,5

Source: TERA Consultants

 

* Country-specific price elasticities of demand (elasticities from Grzybowski 2004)

** Price elasticity of demand for all sampled countries as -0.55 (representative elasticity from Growitsch et al 2010, Dewenter and Haucap 2008 and Madden et al 2004)

 

Both changes in market power and consumer surplus are critical and demonstrate the importance of having the right approach to regulation in order to encourage competition for the benefits to the consumers. This confirms, as analysed above, the necessity of differentiating between regulation harmonisation and market homogenisation. While the first facilitates the replication of best practices among regulatory authorities, the latter would bring about serious consequences to competition and consumer welfare. 

§The results reinforces the justification for a harmonised asymmetric regulatory measures for smaller industry actors across Europe 

Given the benefits they bring to the market, it appears reasonable to allow smaller operators to charge higher MTRs, at least for a period until they gain economies of scale. Another issue that justifies the application of asymmetric access price regulations in mobile telecommunication sector is the use of differentiated conditions of spectrum allocation (Haucap, 2007).

In spite of the supportive arguments regarding, the use of asymmetric price regulations on mobile network operators, also there are some drawbacks such as possible increase of off-net tariffs of the more efficient mobile operators, lower incentives to invest and innovate, risk of inefficient entry. In asymmetric price regulations, smaller operators have higher termination charges than the efficient first entrants. Therefore, off-net tariffs of first entrants tend to be higher because they pay higher wholesale charges to terminating operators. However, practical observations indicate that there is a confusing relationship between termination rates and retail prices (see e.g. Genakos and Valletti, 2008). On the other hand concerning the efficiency losses, as we have explained above there are two efficiency losses which may emerge in the mobile industry. The first one is allocative efficiency losses which is associated with the exercise of market power through increasing prices to consumers. The second one is productive inefficiency which a dominant firm has an incentive to produce but which is also below the social optimum production amount (Waverman and Schankeman, 1997; Valletti, 2006).

Moreover, some empirical economic studies may reach to the different results. For example, Peitz (2005), Kim and Park (2007) and Lee et al. (2010) show that asymmetric access price regulation with a cost-based access price for the incumbent and an access mark-up for the entrant is more successful than cost-based symmetric access price regulation applied to incumbent and entrant, also De Bijl and Peitz (2002, 2004) reached to conclusion that asymmetric regulation brings higher consumer surplus due to lower price and is the source of higher entrant market share and profit. Social welfare is lower but not very notably, hence asymmetric regulation can be seen as the best device to promote competition.

However, if the benefits of the existence of small operators are not disputable, the question of asymmetric regulation is still open. Dewenter and Haucap (2005) has argued based on an analysis of MTRs for 48 European mobile operators between 2001 and 2003 that asymmetric regulation of the larger operators will induce the smaller operators to increase their termination rates even higher which could result in welfare loss. They argue that smaller operators tend to charge higher termination rates than larger operators, even if there are no differences in costs. The argument goes that consumers are generally ignorant about the termination rates of individual operators and are only influenced by the average MTR across all operators. This average figure reflects the relative traffic volumes terminating on each operator’s network and so the MTRs of the smaller operators carry less weight. Since, a small operator’s impact on the weighted average price is quite minor they have greater discretion to increase their prices significantly without a major reduction in demand (Curien, 2007). Dewenter and Haucap (2005) conclude that if MTRs are to be regulated, a symmetrical solution should be applied to both small and large operators although, it does not necessarily imply that their termination rates should be at the same level insofar as late entrants face cost disadvantages due to exogenous factors. The debate is mainly about the possibility to maintain an asymmetric regulation in the long run.