I’ve asked this question of people who know lots more than me about telecommunications economics. And they say ‘double marginalisation’. Anyway, David Levine is a clever fellow and he’s had a crack at answering this. It’s an outrage of course. And is so egregious the Gupment should do something about it. A rule that you couldn’t charge more than – say – three times cost would undo most of the damage. And how about a bit of corporate social responsibility? If rich business people are paying these fees and don’t care to buy themselves a local SIM card that’s one thing. But Levine suggests a lot of the explanation lies in the capacity to trick people into paying $1,000 or so when they leave their data roaming on and then have to cough up when they get back. So large companies with reputations to maintain are tricking their customers out of a thousand dollars or so every chance they get. Not a good look.
David’s blog post is here on his own blog and also reproduced below the fold (It’s OK, as you’ll notice from the title of his blog he’s against IP).
Why are international data roaming rates so high? There does not seem to be a lot of careful analysis. I have found one EU study – in which I have been unable to find one interesting or relevant fact or analysis. So here are my own back of the envelope calculations.How high are international data roaming rates? I have direct evidence from two providers: an Italian provider TIM charges about $10 per megabyte; a U.S. provider T-mobile charges $15 per megabyte. The typical business user uses receives about 15 megabytes per day of email. My smartphone uses about four times this. By way of contrast, you can buy a SIM from Vodafone UK with 30 megabytes of data for about $30. Wifi at the airport or a hotel runs about $10-$60 per day. Over-the-air prices charged to local customers is much lower: TIM charges $25 per month for 5 gigabytes of data, of which probably about 2 gigs is actually used, so the effective rate is about $0.0125 per megabyte. T-mobile in the US charges a similar amount for similar service.,
We can summarize the data this way: for a five day trip the typical smartphone user will pay about $1000 for international roaming or about $100 for wifi. The local user pays about $4 for the same service. In other words, the markup over the competitive cost is about 250, and the markup over close substitutes is about 10.
No doubt some small number of customers are very wealthy and happy to pay several hundred dollars a day for the convenience of keeping the same phone number, not buying a new SIM, not having to locate wifi and so forth – despite the tenfold cost reduction by doing so. Surprisingly only 40% of international roamers turn off their data roaming. In the end it seems reasonable to conclude most people use data roaming not on purpose, but by mistake. Notice that TIM advertises their price as 0.8 cents per kilobyte – which doesn’t sound like much, and has meaning only if you have some idea how much data you use. When I was lost in Paris, knowing the high cost of roaming, I turned on the data service to locate myself on the map. I expected it would be expensive. What I did not expect is that a single viewing of the map would cost 100 euros – cleaning out my account and making my phone unusable in the process.
No doubt telephone companies are good at setting prices to maximize their profits. Apparently such a large number of people accidentally turn on their data services that it is profitable to charge them 250 times marginal cost rather than charge a much lower fee to the much larger group of people who would be willing to pay for the service. Such a business strategy is possible only with substantial monopoly power and illustrates well of the problem of monopoly. Each international roamer pays a thousand dollars once for which they receive practically no value. After that they know to keep the data service turned off.
How can such a business strategy be profitable? The answer lies in the phenomenal growth rate of smart phones. Smart phone sales are growing at about 50% per year. If the stock is growing at a constant rate, it must also be growing at 50% per year. Suppose the typical international roamer spends one month per year roaming, and pays $1000 for one week before shutting off the service forever. If they were willing to pay $50 per week, then charging that would bring in $200 for each member of the entire stock of roamers. Since the stock of roamers is two and a half times the number of new roamers charging $50 per week would bring in $500 rather than $1000.
The international roaming market, lacking competition, is dysfunctional: rather than providing a useful service it taxes consumers one-time lack of awareness of prices. Eventually this will sort itself out: the market for smartphones is not likely to grow at 50% a year forever. Once smartphones start to satiate the market and growth drops to more modest levels – by the calculation above, about 25% per year – it will be profitable to start charging for services received rather than for consumer errors. A similar transition took place in the within U.S. market for voice roaming years ago; originally roaming outside the local area code was extremely expensive; now all providers offer national coverage.
There is an irony in all this. The market for selling smartphones is relatively competitive, and likely the international roaming “profit” is largely given back to consumers in the form of lower prices for contract smart-phones. Relatively rich international roamers subsidize smartphones for their less wealthy and well-travelled brethren.
There is a case for government price regulation here. Of course as smartphone sales diminish, government price regulation will enhance rather than reduce collusion. Ironically the lack of competition in the mobile phone market is largely due to government policy. Governments – wishing to control the flow of information – tightly control bandwidth. Some of this is done through monopolies given or sold to television stations, usually with a quid pro quo in the form of free or subsidized political advertising as well as restrictions on content. As another example, free wifi is not an option in Germany as the government insists on knowing exactly who is logged on to the internet. Copyright law also plays a role here, as the copyright oligarchs are eager also that nobody access the internet anonymously.
A sharper analysis should probe the source of the monopoly power in roaming. Why cannot different mobile providers in the country you are visiting compete to offer you lower data rates? Unregulated countries like Panama greet visitors with giant signs promising lower long-distant rates if you roam with them. The reason this kind of competition is not more broadly possible is because SIM cards are tied to your home provider. Other mobile providers can bill you only through that same home provider. In other words: what your home provider has is a monopoly over your good credit record and means of payment. On the face of it this sounds absurd, and it is a wonderful illustration of the fact that not all property rights are created equal. Some property rights (copyright anyone?) serve to foster monopoly, while other property rights foster competition. This has been recognized in most countries to a limited extent: you now own the property rights in your phone number. But your phone company still owns the rights to your reputation.
There many organizations who compete for the rights to your reputation: Amazon, Google Checkout, PayPal, credit card companies, mobile phone companies, and banks all compete to provide you with direct online payment. The problem is that the SIM card technology prevents mobile providers from billing any organization other than your phone company. Here is a modest proposal. The SIM card standard was imposed (as part of the GSM standard) by European governments. So let them modify the standard to require that the user be allowed to provide alternative means of billing to over-the-air telecommunications providers: the small eagerly wishing to sell providers should be able to recognize my Google Checkout to bill me for cheap data service while I keep my usual phone number. The problem of roaming costs should quickly vanish in the dust of competition.
[I am grateful to a conversation with Philippe Jehiel of the UCL although he is in no way responsible for this. After going through numerous alternatives, we agreed that the most likely explanation for high data roaming rates is exploitation of consumer error in the face of a rapid influx of new consumers. The data on growth in the smartphone market seems to bear this out.]