The waterbed effect is a feature of what economists refer to as two-sided markets. These types of markets - which include telecommunications platforms, electronic marketplaces, credit cards, newspapers, and clubs - generate the highest value of welfare to society when prices charged to the parties on each 'side' of the trading platform differ from the marginal-cost standard that's used for welfare-maximisation in one-sided markets. Oddly, welfare is maximised when consumers on one 'side' of the platform (the 'money side') are charged a price which is above the costs their usage imposes, whilst consumers on the other (the 'subsidy side') are charged less than the costs their use incurs.
Howell gives the following example to illustrate this point:
[...] newspapers charge high prices to advertisers and subsidise the price of the paper to readers. Circulation increases, which means greater value is offered to advertisers (more readers to view the advertisements). So advertisers can be charged even more – and prices to readers become even lower. Ultimately, newspapers can even be provided free to the reader. Forcing newspapers to charge advertisers only the actual costs incurred to print the copy would inevitably force up the price of newspapers to readers, reduce the number of papers sold, and impose a net loss to society.Applying this idea to the mobile phone market Howell writes,
[...] mobile telephony network operators often charge prices below cost for calls made between subscribers on their own network (on-net customers), who are the consumers on the platform’s ‘subsidy side’. More on-net customers means the high fixed costs of network construction and operation can be spread over a larger customer volume, reducing prices further and inducing even more customers to join. Welfare increases as more connections are sold and more calls are made – more than than would occur at cost-based call pricing. Shortfalls in calling revenues are recovered by charging a price above cost for services provided to customers who call across networks (cross-net customers) and who are the consumers on the platform’s ‘money side’. Examples of such pricing include calls from other networks terminating on the network in question (mobile termination fees for cross-net calls) or charges enabling other operators to use network resources (mobile roaming).Howell goes on to explain that evidence from the European Union - see C. Genakos & T. Valletti (2009) - suggests that in practice, the waterbed effect is alive and well in mobile markets. She writes,
If the market for mobile calling was ‘one-sided’, then forcing the termination price down would be likely to result in a reduction in retail charges. But the greater the extent to which cross-net calling subsidises on-net calling, the greater the likelihood that reducing termination charges will result in increased charges for on-net calls, with a net loss in welfare to those customers whose calling patterns have evolved to take advantage of the on-net discounts.
Figure 1 shows the average price paid relative to the world average, both before (periods T-6 to T-1) and after (T+1 to T+6) regulated reduction of mobile termination charges in twenty EU countries. In all cases, regulation was introduced because it was perceived that the unregulated rates were ‘too high’. The actual prices paid by consumers in the studied countries before regulation were lower than the rest of the world. When mobile termination rates were regulated downward, the prices paid actually increased relative to other countries. This is likely to be because of the reduction or elimination of on-net discounts by operators reacting to reduced income from cross-net calling.The point to take from this is that regulation may have unintended consequences in that for some customers prices could rise, not fall, as planned:
[...] because of the waterbed effect, regulated reduction of mobile termination rates will not automatically be positive for all consumers. Cross-net calling charges may fall; but if the bulk of calls made are discounted on-net calls, then the average prices paid for calls may actually rise – even as network operator profits reduce.Reference:
- C. Genakos & T. Valletti (2009) Testing the ‘waterbed effect’ in mobile telephony (www.sel.cam.ac.uk/Genakos/Genakos&Valletti-Waterbed%20effect%20v_2(core).pdf).