Telecommunication networks interconnect so customers of one network can call customers of a different network, but there are considerable interconnection payments between networks that customers do not directly observe but significantly affect the prices they pay. The most important of these interconnection services is “call termination”. This allows a subscriber on one network to call a subscriber on another network.
In this column, we focus on fixed-to-mobile termination rates (MTRs) – the charges mobile operators levy on fixed network operators for call termination. Such termination rates have become a hotly debated issue among regulators and academics worldwide. The level of these charges is perceived to be high both in absolute terms and in relation to similar prices charged by fixed network operators. Bomsel et al. (2003) estimate that there has been a transfer of €19 billion from the fixed to the mobile sector in excess termination charges over costs over the period 1998-2002 in France, Germany, and the UK alone.
Why regulate a potentially competitive industry?
A large theoretical literature has demonstrated that, independent of the intensity of competition for mobile customers, mobile operators have an incentive to set MTRs that will extract the largest possible surplus from fixed users.1 This is due to the fact that the calling party (the fixed customer) usually has no influence on the receiving party’s choice of mobile network. Since the party paying the termination charge (indirectly through retail prices) has little to no influence on the choice of mobile network, there is little downward pressure on fixed-to-mobile termination rates. A mobile operator holds a monopoly over delivering calls to its subscribers and therefore charges high rates. As a result, regulators are increasingly finding mobile operators to be dominant in setting MTRs. This has provided a rationale for imposing a remedy and cutting these rates. However, reducing such charges could increase prices for mobile subscribers – what is known as the “waterbed effect”.
The waterbed effect
The idea behind the waterbed effect is intuitive. Each mobile customer makes and receives calls. A mobile firm can thus be seen as a “platform” that supplies various services to customers and can set different prices to the various sides that use the platform.2 As argued above, a mobile network is a bottleneck for received calls, and money can be made over their termination. When considering its overall pricing policy, a mobile operator will take these termination revenues into account. The higher these revenues, the lower the total price an operator would charge its customers. It makes sense for any mobile operator to pass-through some of its revenue to consumers, because by lowering prices it increases the number of mobile subscribers, which in turn increases the termination revenues earned. Of course, the reverse is also true. If regulation reduces termination charges and hence revenues, operators will have to raise their prices to subscribers. This is the “waterbed effect” – pressing down prices in one part of firms’ operations causes another set of prices to rise.
Regulating mobile termination rates in the real world
Fixed-to-mobile termination rates have been at the centre of attention ever since 1997 when the regulatory debate started in the UK. Price controls on the two largest operators were put into effect from 1998 to 2002, requiring termination charges to be reduced by 9% per year in real terms. After a lengthy consultation and investigation, the UK telecommunications regulator, Oftel (now called Ofcom), concluded at the end of 2001 that mobile termination charges were still substantially in excess of cost. It proposed additional price controls on the four major mobile companies (Vodafone, O2, Orange, and T-Mobile).
Other countries have followed the UK’s lead. The European Commission introduced a New Regulatory Framework for electronic communications in 2002. Every member state must conduct a market analysis of the mobile termination market and, to the extent that market failures are found, introduce remedies. Indeed, all the countries that completed the analysis did find problems and imposed (differential) cuts to termination rates. In 2005, the New Zealand Commerce Commission introduced similar regulation, and while it was convinced that the “waterbed” effect is a theoretically general phenomenon, it doubted its empirical importance. The most recent termination rate proposals by UK’s Ofcom acknowledged the importance of the “waterbed” effect but questioned whether the effect is “complete”, arguing that this can only be the case if the retail market is sufficiently competitive.
Watch out for the ‘waterbed’ effect!
Despite the importance of the waterbed effect for welfare calculations, no systematic evidence is available on its existence or magnitude. Detecting this effect with casual empiricism is difficult. Both termination rates and mobile prices have come down over time, but that does not rule out a waterbed effect. It may have been hidden by a number of countervailing factors, such as tougher competition due to new entrants, technological progress, or economies of scale from traffic growth.
The main purpose of our research is to examine the existence and magnitude of the waterbed effect in the mobile telephone industry. More specifically, we analyse the impact of MTR regulation on mobile prices and profit margins using a newly constructed dataset of mobile operators across more than twenty countries during the last decade. Our identification relies on the assumption that regulation influences retail prices only indirectly via reducing the termination rates and on the knowledge that regulators did not intervene in any other direct manner on consumer prices. Our quarterly data allows us to quantify the waterbed effect for the first time.
Figure 1 plots the average (time and usage-country-operator demeaned) prices in countries that experienced a change in regulation at time T. Notice first that, compared to prices in the rest of the world, average mobile retail prices in countries that experienced a change in regulation at time T were lower before the introduction of regulation.3 Most importantly, in line with our “waterbed” prediction, the introduction of regulation at time T has a clear positive impact on retail prices, and the impact becomes stronger as regulation becomes progressively more binding.
Figure 1. Average price around the introduction of regulation
Notes: Data from Teligen. Figure 1 plots the average (time and country-operator-usage demeaned) logarithm of the PPP-adjusted price paid per usage profile in countries that have experienced a change in regulation, six quarters before and after the introduction of regulation of fixed-to-mobile termination charges. The two continuous grey lines indicate the linear trend before and after the introduction of regulation. Regulation takes the form of “glide paths”, in which termination charges are allowed to fall gradually towards a target over the period.
The full empirical specification allows us to control for time-invariant country-operator characteristics that may influence both regulation and prices or profits. Importantly, it allows us to control for cost differences across mobile operators due to differential access to spectrum frequencies (e.g., some operators have access to 900 MHz spectrum, other only to 1800 MHz) or differences in the cost of network deployment. The specification also accounts for common global trends.
Our results suggest that although regulation reduced MTRs by about 10%, this also led to a 5% increase in mobile retail prices. This roughly translates to a €25 increase on the yearly bill per subscriber, or some €750 million extra in total in our sample.
The waterbed effect is significant but not “complete” – accounting profits are positively related to MTR. Mobile firms tend to keep part of termination rents instead of passing them through to customers and thus suffer from cuts in termination rates. Figure 2 plots the average operators’ profits4 in countries that introduced regulation. Regulation significantly lowered profits.
Figure 2. Average profits around the introduction of regulation
Notes: Data from Merrill Lynch. Figure 2 plots the average (time and country-operator demeaned) logarithm of the EBITDA in countries that experienced a change in regulation, as in Figure 1.
The existence and magnitude of the waterbed effect is key in understanding the social costs and benefits of the regulation of fixed-to-mobile termination rates. Regulators have generally accepted that the waterbed effect is likely under full competition, but they had doubts regarding its validity and empirical significance under conditions of imperfect competition. Our research shows that the waterbed effect is significant and strong in practice.
Our results have important implications for the mobile industry (and other markets). First, it implies that any welfare analysis of regulation of termination rates cannot ignore the presence of the waterbed effect. Regulation clearly benefits fixed users calling mobile subscribers, but it increases the (unregulated) retail prices for mobile subscribers.
Second, our analysis is also important in the current debate of regulation of international roaming charges. The European Commission has imposed caps to roaming charges when making and receiving phone calls within the EU. A reduction in roaming charges may cause a similar “waterbed” phenomenon, whereby prices of domestic calls may increase as operators seek to compensate for their lost revenue. Mobile subscribers bear the burden, as calling abroad becomes cheaper but calling at home could become more expensive. A similar waterbed effect could arise in the banking industry, where there is currently a debate whether to regulate specific prices, such as overdraft charges that many consumers do not “see” when opening a bank account. Our results show that such regulation could unintentionally increase other fees.
Third, we have also shown that the waterbed effect is not complete, as mobile companies do not fully pass through termination rents to their customers. This is an important empirical question that may vary among industries. The more incomplete the waterbed effect, the easier it is to justify regulatory interventions.
Armstrong, Mark (2002), “The Theory of Access Pricing and Interconnection.” In M. Cave, S. Majumdar and I. Vogelsang (eds.) Handbook of Telecommunications Economics, North-Holland, Amsterdam.
Armstrong, Mark and Julian Wright (2009), “Mobile Call Termination.” Economic Journal, 119: 270-307.
Bomsel, Olivier, Martin Cave, Gilles Le Blanc and Karl-Heinz Neumann (2003), How mobile termination charges shape the dynamics of the telecom sector. Final report, WIK Consult, University of Warwick.
Genakos, Christos and Tommaso Valletti (2007), “Testing the “Waterbed” Effect in Mobile Telephony.” CEPR Discussion Paper 7611 (forthcoming in the Journal of the European Economic Association).
Rochet, Jean-Charles and Jean Tirole (2006), “Two-Sided Markets: A Progress Report.” RAND Journal of Economics, 37: 645-67.
Valletti, Tommaso and George Houpis (2005), “Mobile termination: What is the “right” charge?” Journal of Regulatory Economics, 28: 235-258.
1 See, for example, Armstrong (2002), and Valletti and Houpis (2005). Armstrong and Wright (2009) provide an excellent overview of mobile termination theoretical literature and policy in the UK.
2 In this sense, a mobile firm can be seen as a “two-sided” platform, showing similarities with other industries such as credit cards or advertising-funded media platforms (see Rochet and Tirole, 2006).
3 This is important because it refutes the argument that regulation was introduced as a result of high retail mobile prices (reverse causality).
4 Our proxy for profits is earnings before interest, taxes, depreciation and amortisation (EBITDA).
These are the charges made to customers when using their phones outside their home country, i.e., an Italian customer making/receiving a phone call in Greece.