Citigroup Questions Whether US Spectrum Shortage Exists

alphadogg writes "For more than two years, the U.S. mobile industry has warned of an upcoming spectrum shortage, but two analysts at Citigroup don't buy it. AT&T, trade group CTIA and even officials with the U.S. Federal Communications Commission have talked frequently about a coming spectrum crunch, as mobile customers move to data-sucking smartphones and tablets. Smartphones use 24 times the spectrum compared to standard mobile phones, and tablets use 120 times the spectrum, FCC Chairman Julius Genachowski said in a speech on Tuesday. But Citigroup analysts Jason Bazinet and Michael Rollins questioned what has become the conventional wisdom in the mobile industry. The U.S. has plenty of spectrum for mobile broadband, but much of it is in the wrong hands, they said."

Comments

A Patchwork Of Spectrum Is Not Usable Spectrum

[rsmith-mac]

Citi's report is not wrong, but how they go about counting things is naive at best. The crux of the matter is that there's a lot of crap spectrum that carriers basically got for free or close to it. But before we get too far ahead, let's answer an easier question: what is good spectrum.

1. 1) The ideal spectrum is below 1GHz, as these frequencies have the best building and tree penetration. 1GHz-2GHz is usable, but it's not ideal because you start taking notable losses indoors and customers who've given up on landlines can't reliably use their phones indoors everywhere. Anything over 2GHz is effectively useless for mobile wireless because it's so poor at penetrating obstacles. It's best used for fixed point wireless where obstacles can be planned around and/or removed.
2. 2) The ideal spectrum is nationwide. A patchwork of spectrum is not usable spectrum because it means you can only use narrow (lower bandwidth) channels, and requires a great deal more effort to plan, operate, and maintain a wireless network.
3. 2b) Local spectrum is only useful when it abuts nationwide spectrum so that carriers can use it by simply activating more channels in high population areas.

Case in point, 194MHz of the spectrum Citi says is available is above 2GHz: "Citigroup's description of 194 MHz available in the Broadband Radio Service (BRS) and Educational Broadband Service (EBS) bands between 2.4 and 2.7 GHz". This also goes hand-in-hand with Citi's weird method of counting spectrum in use: they're multiplying it by the percent of the population that the spectrum covers. "The two used averages to come up with spectrum use estimates; if a carrier has a 10 MHz nationwide block, but is only delivering service to half the U.S. population, the report considers that 5 MHz of used spectrum, Rollins said."

Ultimately the carriers are being wasteful at times, but not nearly to the degree that Citi says they are. The carriers need more national allocations if they're to run a 3rd network simultaneously, and those allocations need to be at least 40MHz wide so that they can operate two sets of wideband (10MHz) LTE channels. Smaller allocations mean that they're going to have to use smaller channels, and that's going to greatly limit network performance.

Re:Spectrum sale by Market

[Miamicanes]

> They should have sold the frequencies by market area (city, zip-codes, etc.) and not nation-wide.

Great. So then we could have a situation like we did prior to the arrival of Sprint around 1999, when every city had different cellular carriers, and sometimes you couldn't go 50 miles away from home without paying extra to roam. In case anybody has forgotten, roaming charges in the US were still common AND punishingly expensive less than 10 (hell, 5 or 6!) years ago unless you were a Sprint customer. Sprint's network might have sucked in most places, but if you lived in a real city and 99% of your travel was to other real cities and the major highways between them, it was rare to end up someplace that literally had no service unless it was totally out in BFE. You might have had to go outside, or even climb up on a roof to get a usable signal, but at least you weren't getting charged $5 plus a dollar per minute the way people with Verizon or AT&T did. There's a reason Sprint achieved early popularity in Florida and Texas -- both states were horribly fractured between hostile, rent-seeking regional carriers, and Sprint was literally the only way to travel around the state without getting raped by roaming charges.