The Penrith problem – structural incentives and mobile service networks
Eh, Charlie Stross’s blog is a machine for destroying time. Anyway. This post is going to be so wonkish it’s to not come back from.
An occasional theme on this blog has been the intersection between the Bush wars and the mobile phone industry. In fact, looking back, that’s not been so much an occasional theme as more of an obsession, and I’d have written more if I hadn’t been subject to non-compete clauses.
Everyone who reads this blog probably knows that Afghanistan got GSM coverage very quickly after 2001, with Roshan and the Afghan Wireless Communications Company or AWCC in the lead. Things went so fast that for a while there were four operators with licenses and a good half-dozen pirate networks. The explanation of this is pretty simple – in the early 2000s the mobile industry had developed a whole package of technology, business models, methods, and personnel that made it possible to unfurl a GSM network pretty much anywhere and make an absolute killing.
Thaksin Shinawatra’s career is a case in point – who knows how a Royal Thai Police colonel raised the money to come up as the holder of a GSM licence, but he did, and there were consulting engineers and contractors who would build the network and equipment vendors who would supply the parts with 100% vendor finance. Once it was up, it rained money and he was off to the races.
Of course Thailand is nothing like Afghanistan – a solid middle-income, industrialising economy with the kind of institutions that function by corruption rather than failing because of it. By 2001 there weren’t so many plums like that one to pluck and the buccaneers who were first in were beginning to think about cashing out.
On the other hand, the gear kept getting cheaper and the success-stories made it easier and easier to borrow from the World Bank or other friendly local multilateral financial institution, as at this point it looked like about the only development success in 40 years or so. Thanks to people like Mo Ibrahim and the rest at Mobile Systems International, the level of average revenue per user that made it viable to build a GSM network was driven down until now we’re operating below $5/month and there is no country that doesn’t have at least a little bubble of coverage around the capital city.
So that’s why it happened. There was a reliably deployable package of technology and economics and legalities, with a global workforce of Sven-units with frequent flyer points on every-damn-thing, and a set of reliable sources of capital. As well as the Aircom or Ericsson Professional Services guys who would design the network, and the contractors who would recruit the people who dug the foundations on the knolls and warps in the landscape that the radio planners made obscurely significant, there were others who would write the formal licence proposal to fit through the newly established bureaucracy of “regulators” and public procurement systems redesigned to please the IMF and other princes of the Washington consensus. No doubt there were people who specialised in operating the other, informal procurement systems. If you know what I mean. There was a product that sold and that, once sold, became one of the markers of modernity and status. The wheel of capital intensification kept turning, recapitulating the development of the Grand Banks fishery in the 1500s. Or something like what Erik Lund would say.
Of course, there were some problems with the package. Most of all, it structurally favours creating a new operator over extending an existing one’s network, which is why Uganda has six mobile phone networks (and two WiMAX DSL-substitute not-officially-mobile networks) when a lot of people who ought to know think the UK only needs three. The turn-key vendor contract is meant to give you all the bits you need to call yourself an operator; the MFI funding is released when the licence application is accepted; the money starts flowing when the 15% or so of the cells that carry 50% of the traffic are on line. Increasing population coverage is mostly cost, which is why a coverage requirement is typically laid down in the licence.
And that’s why supposedly (and that should be a big “supposedly”) Kabul has better mobile service than Rory Stewart’s constituency. Rory may need to consider what kind of mobile service places that stand in the same relation to Kabul as Penrith does to London get, and we’re going to discuss this (and some other stuff) in the next post.