Exploring the Waterbed Effects in Mobile Telephony Using Systems Dynamics more

MN 415 – Assignment 1 Submitted by: 201116896 01.12.2011 1. Introduction This paper analyzes a highly debated phenomenon called waterbed effect, which is used to describe situations under which regulation of price of one product of a multiproduct, or two-sided, firm causes a change in the price(s) of other product(s) of the firm. (Schiff) The waterbed analogy is used to stress out the unbalancing action we observe as when pressing down on one part of the bed causes another part to rise, just like the prices of a multi-product firm. Waterbed effect emerges when firms sell in multiple products whose marginal revenue and/or marginal costs are interdependent. (Oehling)A typical example of multiproduct companies is mobile operators which provide communication services to their own customers over the same platform and also sells connectivity, or call termination, to their customer base to other fixed and mobile networks. (Genakos and Valletti) Another example is banks which sell a bundle of account services to their customers, together with penalty fees for unauthorized overdrafts, with separate prices for each. (Schiff) This paper will be based on the investigation of waterbed effects on subscription rates due to downward regulation of mobile termination rates (MTR)1 by telecom authorities. The basic mechanism here is that a reduction in the termination rare causes the marginal revenue per mobile subscriber to fall at all subscription levels, which then causes the profit-maximizing subscription fee to rise. (Schiff)The organization of the rest of the paper is as follows. Section 2 investigates and simulates the waterbed effect from scratch. Sources of the charts and empirical data used to illustrate the phenomenon are mostly gathered from the discussion paper “Testing the ‘Waterbed’ Effect in Mobile Telephony” written by Christos Genakos and Tommaso Valletti and provided in the Appendix 2. Moreover, the theoretical proof of the waterbed effect is also provided in Appendix 3. Section 3 then discusses alternative regulation policies that can be used to evade the negative consequences of the original policy. Section 4 offers implications for regulators and concluding remarks. 2. The “Waterbed” Effect in the Mobile Telephony Market The possibility of an unintended price increase in other product(s) when trying to press down the price of another product has always been acknowledged by the telecom authorities, its actual presence, however, has always been subject to extensive debates. Although it has long been assumed that this type of effect could only be present when markets are perfectly competitive, it has been proved both theoretically and empirically that waterbed effect is present regardless of the market characteristics. (Schiff) The following behaviour-over-time (BOT) graphs are derived by using price data from the telecom operators who have been prone to an extent of MTR regulation. The graphs are selected to most clearly illustrate the waterbed effect from different perspectives. Figure 1 shows the average prices2 six quarters before and after the introduction of regulation, in countries where a downward regulation of MTRs have put into place. The upward shape of the line indicates that compared to prices in the rest of the world, average prices in countries that experienced a change in regulation were actually lower before the introduction of regulation; but introduction of regulation has a clear positive impact on prices. Figure 2 depicts this effect by showing how much the prices will increase, i.e. how much is the waterbed effect, when we lower MTR by 1 unit. It is seen in the graph that the waterbed effect becomes stronger as regulation becomes progressively more binding over time. (Genakos and Valletti) 1 Mobile termination rates (MTR) are the charges mobile operators levy on either fixed network or other mobile operators for terminating calls on their networks. 2 The prices are time and usage-country-operator adjusted. 2 Figure 1: Average Price Around the Introduction of Regulation 0.17 0.12 0.07 0.02 -0.03 -0.08 Average price paid Figure 1 plots the evolution of time and country-operator-usage demeaned average logarithm of the PPP adjusted price paid per usage profile six quarters before and after the introduction of regulation of fixed-to-mobile termination charges based on the Teligen data corresponding to the best deals available at every period. Source: (Genakos and Valletti) Regulation put in place T-6 T-5 T-4 T-3 T-2 T-1 T T+1 T+2 T+3 T+4 T+5 T+6 Figure 2: Evolution of the Waterbed Effect 0.4 0.3 0.2 0.1 0 T-6 T-5 T-4 T-3 T-2 T-1 T T+1 T+2 T+3 T+4 T+5 T+6 Waterbed effect Figure 2 plots the regression coefficients on binary variables six quarters before and after the introduction of regulation. The dependent variable is the logarithm of the PPP adjusted total bill paid by consumers with different usage based on the data corresponding to the best deals available at every period. Source: (Genakos and Valletti) Regulation put in place Figure 3: Evolution of Implied Total Welfare I Regulation put in place II T-6 T-5 T-4 T-3 T-2 T-1 Total welfare T T+1 T+2 T+3 T+4 T+5 T+6 T+7 T+8 T+9 T+10 T+11 T+12 Welfare of Buyer B Welfare of Buyer A Figure 3 depicts a variable that has been created arbitrarily to convey the welfare effect of the MTR regulation. Although it is very hard to quantify, this affect has been proven to be the sole focus of any regulation. Basicly, total welfare is the welfare of the buyers of the regulated product, in our case mobile-to-mobile or fixed-to-mobile calls, plus the welfare of the buyers of the unregulated 3 products. (Schiff) In the boundaries of our analysis, not the exact values of the total welfare in the market, but the shape of the curve, that is rising for some time (Period I), and decreasing afterwards (Period II), is what is important to assess the waterbed effect. Although it is not a generic definition, it is helpful for the purposes of this research to assume welfare as the consumer surplus, i.e. the amount of money customer is willing to pay minus the actual amount he is paying, which is inversely proporationate to the price levels. 2.1. What is Happening Deep Under? In this section, we aim to illustrate and simulate the both periods of Figure 3. First, for the aim of simplification, imagine that each network operator derives revenues from two possible sources. The first group of customers, called Buyer A, are the ones who use the services which operators provide to their own customers, i.e. subscription services, calls to customers in the same network, etc. All these services are bundled together and cost an average price to the buyer. The second group of customers, denoted as Buyer B, are the ones who make calls to other mobile operators. They also face an average price. Finally, the total welfare in this case is that: Level of Total Welfare= Level of Welfare of Buyer A + Level of Welfare of Buyer B. CLD 1: Shows the basic intension of regulators in intervening to the MTR levels CLD 1 shows the negative feedback loop, through which the regulators observe the welfare level of Buyer B and intervene in the market by cutting the MTR levels when they observe a gap between desired and actual levels of welfare of Buyer B. The delay between the reduction of regulated MTR level and the decrease in the actual MTR level is due to the typical 6-months period given to the operators for adjusting their prices. CLD 2 depicts the process, through which the reduction in MTR causes the Prices Paid by Buyer A to rise, i.e. the waterbed effect. There are three critical points to mention here. First, there are no delays in this negative feedback loop. The basic reason is that the firm needs to respond quickly to its deteriorating profit margin, and after assessing numerous market and demand characteristics, the firm acts quickly by adjusting the price its charges to Buyer A. Second, exogenous market variables penetration and saturation, and most importantly their interaction, affect firm’s tolerance against the MTR cut and the overall impact of the waterbed effect on prices. The waterbed effect is stronger the more intense competition is in the market and the higher the market penetration. (Genakos and Valletti) Third, the urge and the elasticity affects the firms ability to reflect the loss caused by MTR cut on its prices. This multiplier is the waterbed effect itself. 4 CLD 2: Shows the effects of the decline of MTR on Buyer A’s welfare. Notice that the exogenous market dynamics play an important role here. CLD 3 shows the two balancing loops together and gives the complete picture of MTR regulation and the waterbed effect. The two negative feedback loops start operating almost together at first. However, due to the 6-month adjustment period, MTR actually goes down gradually, which causes CLD 2 to be less strong at the beginning. CLD 1 was dominant in the Period I (Between T-6 and T+6) of Figure 3, and this positive welfare effect of the MTR regulation caused the total welfare to go up immediately after the MTR cut. After completion of the 6-monts adjustment period, however, MTR goes down to its new level and the price charged to Buyer A rises due to decreased profitability. CLD 2 dominates right after the adjustment period, Period II, and the negative welfare effect of this loop causes the total welfare to decline. The question of whether the total welfare could go below its initial level is very hard to answer empirically. However, it has been showed above that regulation of prices of multiproduct firms such as mobile network operators can trigger a waterbed effect, which causes unregulated prices to rise. Depending on the extent of the negative welfare effect, this may cause either a welfare transfer (in the case when the positive welfare effect of CLD 1 outruns the negative welfare effect of CLD 2) or a welfare loss (in the case when the negative welfare effect is bigger). (Schiff) CLD 3: Illustrates the whole mechanism of MTR regulation and the waterbed effect. 5 3. Optimizing the Regulation Regulators’ intension in cutting MTR is clearly increasing the welfare of Buyer B. The actions they take increases the welfare of Buyer B in the short run, but at the cost of welfare of Buyer A. Moreover, MTR regulation’s final effect on total welfare is ambiguous and it in the best scenario it creates an unfair welfare transfer. To address and solve all the problems above, regulators should set a new target that both includes the welfare of the both parties. Choosing Level of Total Welfare as the target variable is an appropriate step towards this end. Figure 4: Evolution of Total Welfare under Total Welfare Targeting I Regulation put in place II T-6 T-5 T-4 T-3 T-2 T-1 Total welfare T T+1 T+2 T+3 T+4 T+5 T+6 T+7 T+8 T+9 T+10 T+11 T+12 Welfare of Buyer B Welfare of Buyer A Figure 4 is the BOT graph showing how the total welfare will behave under the new regime. Targeting Total Welfare will not eliminate welfare transfers entirely. It is inevitable for multiproduct firms that the unregulated prices will rise due to the downward regulation of price. However, with total welfare targeting, it is possible to prevent continuous and uncontrolled transfer of welfare which could lead to net welfare loss at the end. CLD 4 below shows the new total welfare targeting regime, under which the urge to cut MTR is linked to the observed difference between the actual and desired total welfare. By controlling the total welfare variable it is now possible to observe the magnitude of the positive (1 st negative feedback loop) and negative (2nd negative feedback loop) welfare effects of MTR regulation. Another important aspect of the new regime is the big positive feedback loop that covers both the negative feedback loops. This 3rd loop is nothing but the mere assurance that the system will repeat itself until the total welfare gap disappears, i.e. actual total welfare equals desired total welfare. It is also important to remember that this new regulation mechanism will get to balance much faster than the old one, because the target variable total welfare is affected by both of the loops and in opposite directions. 6 CLD 4: Using total welfare as the target variable enables us to consider outcomes of both loops. Please notice the new positive feedback loop, encompassing the two negative feedback loops. 4. Conclusion and Takeaways This paper has analyzed the waterbed effect phenomenon in the mobile telecommunications industry. Briefly, the waterbed effect emerges when regulation of any one price set by a multiproduct firm cause changes in the other price(s) set by the firm, as a result of its profitmaximizing behaviour. (Schiff) The existence of such effects has important implications when assessing the effectiveness of the regulation. That is, as well as calculating the welfare effects in market for the regulated good or service, the effects on the price(s) charged for any other good(s) produced by the regulated multiproduct firm should also be estimated and included in the welfare calculations. A useful way of evading the waterbed effect and minimizing the chance of disruptive welfare transfer or welfare loss is to use total welfare as the target variable, which will enable us calculating for both the negative and positive effects of the price cut. 7 References Genakos, Christos and Tommasso Valletti. Testing the "Waterbed" Effect in Mobile Telephony. CEP Discussion Paper No. 827. London: LSE Centre for Economics Performance, 2007. Oehling, Daniel. "Is there a waterbed effect in conjunction with the regulation of mobile call termination?" 2011. "Regulating the Mobile Phone Industry: Beware the "Waterbed Effect"." CentrePiece (2007): 25-28. Schiff, Aaron. "The "Waterbed" Effect and Price Regulation." Review of Network Economics (2008): 392-412. 8 Appendix 1: CLD 4 9 Appendix 2: Original BOT Graphs (Genakos and Valletti) 10 Appendix 2: Original BOT Graphs (Genakos and Valletti) Continued 11 Appendix 3: Proof of CLD 2 (Genakos and Valletti) 12