Screening competition in mobile telephony†

Abstract

This article presents a simple method for screening competition in differentiated products oligopoly with a small number of competitors. In many situations, estimation of price elasticities of demand may be impossible due to difficulties in defining demand or missing data on sales. However, even without information on price elasticities, in certain situations it is possible to test for the static noncooperative Nash–Bertrand equilibrium, which in the case of rejection, may be important screening information for antitrust authorities. The static noncooperative Nash–Bertrand equilibrium may be rejected when demand is linear and in the estimation of best-response functions, the coefficients on the competitors’ prices are statistically greater than 0.5. The application of this method is illustrated by the example of German mobile telephony using monthly data between January 1998 and December 2002. According to the estimation results, the observed prices in the segment of low-users cannot be the outcome of a static noncooperative Nash–Bertrand equilibrium.

Notes

¹ Other factors which influence incentives to collude are: market transparency, product differentiation, demand growth and fluctuations, buying power of consumers, multimarket contacts, network effects, etc.

² In France, for instance, the Competition Authority found written evidence on information sharing and on market shares fixing agreement. The Competition Authorities in Spain and Ireland found collective dominance in their mobile telecommunications industries.

³ Since competition and collusion are not contrasted directly this method may be classified as screening, see discussion in Harrington (2005).

⁴ This method is applicable to markets with a small number of competitors, in which demand is well approximated by a linear function. It should be possible to estimate best-response functions in these markets and the diversion ratios for some products must take particularly large values, i.e. products of colluding firms must be relatively close substitutes.

⁵ The German Competition Authority has not raised any concerns regarding competition in mobile telecommunications industry.

⁶ If the price elasticities may be estimated in a trustable way, the main problem remains lack of information on marginal costs. However, when the determinants of marginal costs are observed, markups may be estimated and a collusive market conduct can be tested against the competitive one (for a literature review and discussion, see Bresnahan (1989) and Reiss and Wolak (2005)).

⁷ As in Shapiro (1996), diversion ratio from product A to product B is the fraction of sales lost by product A which is captured by product B due to a price increase of product A. It may be also interpreted as the proportion of people buying product A who consider product B as their second choice.

⁸ This technology allows data to be transferred at much higher rates in order to satisfy the demands of multimedia applications.

⁹ Source: www.RegTP.de.

¹⁰ A similar approach was also used by the Irish Commission for Communications Regulation in its analysis of wholesale mobile access and call origination. This analysis was a basis for the assessment of collective dominance in mobile telecommunications industry in Ireland. See Commission for Communications Regulation (2004).

¹¹ As can be seen in Fig. 1, prices of T-Mobile and Vodafone did not react to these entries, while E-Plus reacted by price decreases.

¹² First, price indices for the network operators in France are computed in an analogous way for Germany. Then, industry-level price is constructed by weighting the individual prices by market shares.

¹³ List prices in these periods may be the same for all four network operators but, for instance, in 2003, the share of E-Plus was 14.3% and of O2 only 6.3% in an analogous way as for Germany.

¹⁴ It is assumed that mobile network operators set their prices independently to the prices of fixed-line services. In the case of a price increase by one network operator, the second choice would be almost certainly subscription to another network. There are very few consumers who completely give up the usage of mobile telecommunication services. Also the European Commission in a number of decisions has established that mobile telecommunication services cannot be regarded as a substitute to fixed line telephony services.

¹⁵ Given that p _{2 t} = p _{1 t} , the first two equations are the same. Nevertheless, both equations are estimated.