This article "exposes" some interesting phenomena in my mind. Quoting the article:
When the big four cellular companies decided to hike the price of sending a text message, they all managed to settle on precisely the same increase. Sprint Nextelraised its price from 10 cents to 15 cents per message in 2006. AT&T quickly followed suit, as did Verizon Wireless and finally, in June 2007, T-Mobile. And now Sprint has raised its price again, to 20 cents.
Today those copycat price hikes are producing banner results for the carriers. In the most recent quarter, the Big Four's customers coughed up anywhere from 29% to 64% more for data services (that is, everything but regular phone calls) than they had the previous year. The four carriers collectively produced $17 billion in operating income on $104 billion in revenue in the first nine months of last year. The carriers say that consumers can buy a monthly package that lowers the cost of a text message. But without the huge surge in payments for data, revenue per user would have fallen at every company.
Their ability to hike prices on text messages certainly can't be explained by the companies' costs. On modern cellular networks the few hundred bits of information that make up a text message take up such a minuscule amount of capacity that they can be carried for a fraction of a cent.
The first paragraph basically points to implicit collusion in price setting. As any economist with tell you, this is one of the principal problems in oligopoly. In this case, it was evident even with four industry participants, which suggests either that the price for text messaging is quite inelastic or that this "ala carte" price doesn't reflect consumer experience. Since avid SMS-ers purchase packages, the actual price is far less, so, I believe the latter explanation dominates.
The end of the second paragraph makes little sense ... data services and SMS are quite different and appeal to different markets. It is clear, though, that non-voice services are the future for profits in this sector.