Books I have co-authored

Exorbitant data roaming as an n-person Prisoner’s Dilemma

One of the major hurdles for further growth of mobile data services, and LBS in particular, is the exorbitant data traffic fees customers incur when using their phones abroad.

Corporate users can pretend that the service is free but their employers certainly notice if the mobile Net bill is larger than the hotel bill. In the private market segment, price sensitivity is much higher. Prices have to come down for this market to reach its potential.

As long as the price elasticity of demand is larger than one, lower traffic fees will lead to such high growth in traffic volume that the total revenue increases, despite the lower prices. The entire industry would benefit from an agreement to remove most of the roaming charges, but that is not even on the horizon.

A single operator could see the benefits of offering affordable data traffic abroad for its own customers. But as long as other operators don’t reciprocate, the industry will be stuck in a sub-optimal stalemate.

The problem is that the same operator who wants to offer an affordable rate to its own customers has no incentive to lower prices for other operator’s customers who travel into the operator’s coverage area.

It would not help much if two operators were to reach a bilateral agreement about lower roaming fees. In most countries there are three or more operators and as long as the handset connects to the strongest signal a bilateral agreement would just create a random patchwork of affordable and exorbitant fees. It would still be confusing, add uncertainty and deter usage.

Even if most major operators cooperated and formed a club with mutually lower prices, it would undermine the arrangement if only a few defectors refused to participate. The minority of defecting operators would be able to both reap exorbitant data rates from other operators’ customers as well as benefit from a larger total market and customers’ expectations of low prices.

For the operator market as a whole, the optimal end-state would be if all operators chose the strategy of cooperation, but the optimal individual rationality for one operator is to defect from any agreement. The result is that all operators are worse off in a sub-optimal state of mutual defection.

This can be modeled in game theory as an n-person Prisoner’s Dilemma. Here is a simplified numerical example with one operator (player A) playing against all the other operators (player B). Both the players can chose between two strategies: Cooperate (charge affordable data roaming fees) or Defect (charge exorbitant data roaming fees). Player A is the interesting active player in this game and Player B should more be viewed as a passive dummy.

If both players cooperate they will each receive a payoff of 10, denoted as (10, 10) for Player A and B respectively (where 10 for Player B is the payoff for each operator in the operator pool). If Player A chooses to defect while B cooperates the payoff will be (20, 9). The defecting player gains significantly and receives a payoff of 20 while Player B loses and only receive 9. If both Player A and B defect, the payoff will be (5, 5) which means that the total payoff to the players is significantly lower. One the other hand, if Player A is the only player that cooperates in an environment where Player B defects, the payoff will be (2, 6), and A only receives 2 (instead of 5). The figure below shows the game matrix described in normal form.

Player B
Cooperate Defect
Player A Cooperate 10, 10 2, 6
Defect 20, 9 5, 5

The payoff matrix shows the strategy of one operator (A) versus all other operators in the end states where either all chose to cooperate or all chose to defect. It is also possible to model the payoffs when the number of cooperating telcos moves from 0 –> 100 percent (described here). Regardless of the share of operators that chose to cooperate vs. defect the conclusion for Player A is the same. The dominant strategy for Player A is to defect, which makes it a classic Prisoner’s Dilemma for the active player. For Player B (all the other players) this is not a Prisoner’s Dilemma as the dominant strategy is to cooperate even if one Player (A) decides to defect. The problem is that for each individual operator it is a dominant strategy to defect even though the common good would be maximized if everybody cooperated.

I don’t have a clear answer for how to escape from this sub-optimal state. A few years ago, European voice roaming showed a similar pattern until the EU Commission mandated a price cap and forced down the prices. If the industry can’t solve this themselves the politicians may intervene again, at least in Europe.

The industry might be able to handle this on its own if the major operators form a club and then exert strong peer pressure on the remaining operators. One way is to punish defecting operators with very unfavorable roaming deals or refuse roaming. However, that will leave the club’s customers with inferior network coverage. It could possibly also be considered anticompetitive if a cartel of market dominants bullied smaller operators.

To be effective and protect the customers from accidentally connecting to a network with exorbitant prices, an operator club probably needs to have a technical software solution installed on their customers’ handsets. This software would ensure that the handset only connects automatically to networks that are members of the club, even if a renegade network has a higher signal strength. This should be manageable for operator branded handsets but if customers just have SIM cards it will be complicated to mandate that they install an app that can control low-level functionality on almost any device.

Comments are closed.