One of the consequences (intended or otherwise) of the TED Global conferencein Arusha, Tanzania last month is that many of the bloggers I read regularly are spending a good deal of time thinking about the classic questions of development economics: What makes some countries rich and others poor? What’s the critical missing ingredient in development: more aid? better governance? infrastructure? entrepreneurship?

This second question is getting a workout as a wide range of commentators respond to the Bono-edited Africa edition of Vanity Fair. Several writers have pointed out that the issue falls squarely in the “more aid first” camp, featuring a largely uncritical portrait of Dr. Jeffrey Sachs’s Millenium Villages project, a project designed to demonstrate what could be accomplished with massive infusions of aid into rural communities. The governance-first camp gets widely discussed in World Bank and USAID circles, in my experience. And TED Global gave some good introductions to the infrastructure-first argument, with former Nigerian Finance Minister Ngozi Okonjo-Iweala pointing out that China’s path to development has relied heavily on infrastructure investment.

I’ve been thinking a great deal about the “entrepreneurship-first” path - possibly best exemplified by financier Idris Mohammed’s statement, “If you make Africans rich, they’ll be less poor. That’s my poverty reduction strategy.” Almost every discussion of business opportunity in Africa focused on the amazing growth of the mobile phone industry. That growth has been astounding, but it’s hard to know whether that growth will be replicable in other sectors. There’s a couple of circumstances that I think are critical to understand in the rise of mobile networks on the continent:

- You can build a mobile phone network one piece at a time. With a GSM license and a single tower, a company can begin earning revenue and start using this revenue to finance future expansion. An investment in the single-digit millions can turn into a multi-billion dollar business through reinvestment of revenues. That just isn’t true for creating container ports, major roads or large power generating facilities… or, at least, I’m not smart enough to figure out a model that allows me to build container ports a few million dollars at a time.

- Users financed a great deal of the infrastructure behind the mobile phone boom - specifically, they purchased the handsets, which represent the lion’s share of investment. (Thanks to Reuben Abraham for this important insight.)

- Sheer government incompetence helped the mobile industry by ensuring that most phone buyers weren’t replacing land lines with mobiles, but purchasing their first phones. It’s easier to sell someone a new, useful service rather than an improvement on an existing service, as mobile companies did in higher development nations.

I’m trying to figure out whether these criteria lead to an infrastructure investment strategy for Africa based on incremental infrastructure development. For-profit companies, many founded by expatriate Africans with a few million dollars, would provide the sorts of resources we’ve traditionally expected governments and parastatals to provide. Ideally, governments would work with these providers to bring services to areas of their countries not able to pay for them; given the mixed record of African governments in creating infrastructure, perhaps we’re better off hoping that most governments stay out of the way of innovative infrastructure providers.

Russell Southwood, the dean of African telecoms analysis, publisher of the indispensible Balancing Act newsletter, has evidently been doing some thinking along similar lines. In this week’s letter, he points out that African mobile phone companies are being forced to become power companies. In urban areas, phone companies have to equip every tower with diesel generators because of frequent power cuts. In more rural areas, where companies can’t rely on grid power, providers need to put in two generators - one to power the station, the second as backup. The cost of delivering diesel fuel to these locations is substantial - Southwood calculates that a grid and road-connected base station costs $2,500 a month to maintain, while a very rural station might cost $20,000.

It’s worth considering those figures for a moment - mobile providers have been expanding aggresively into lower-density parts of African nations. Sometimes they’re making these investments because their licenses require them to provide a certain level of coverage throughout the country; more often, they’re expanding because there’s money to make in these markets. That suggests that mobile telephony is so important to rural Africa that operators are able to make five-digit sums per month in fairly rural areas and recover the costs associated with providing this connectivity… talk about “bottom of the pyramid revenue”…

Southwood suggests that universal service funds - a tax on telephony revenue designed to subsidize deployment of telephone service in rural areas - could be used to build electric power networks, not just phone networks. He points out that these funds have raised $6.5 billion (on the continent, I assume), but that only $1.7 billion has been spent, leaving a large amount “in the pot” which could be redeployed.

If mobile phone companies - or a similarly entrepreneurial entity - could begin building larger, more efficient power generating facilities, they could service local communities with power as well as with telephony. If there were sufficient success for this model, it might start to resemble the “electranet” that some have suggested might alleviate African power problems.

Southwood isn’t proposing something quite so emergent - he’s basically suggesting that the Universal Service revenues could subsidize creations of private power operators. I think Russell’s really onto something here. His closing paragraph is a compact statement of a potentially transformative idea:

In a nutshell, the mobile operators appoint a private company to build and operate a power transmission network. This company would have as its anchor customers at least two mobile companies. Whatever surplus power it generated would either be sold to the national grid operator or be sold on to retail customers. There are private investors in Africa wanting to put money into private power generation and perhaps they might also come in as investors. The mobile operators have shown that it is possible to get things done on the continent and to make money doing it. Perhaps they should now pick up the gauntlet that will allow them to address their high operating costs and get lower taxes at the same time.

In the meantime, mobile phone networks are turning to other creative solutions to power their towers in the absence of reliable grid power. Afrigadget reports that Winafrique Technologies in Nairobi is designing windmills that power remote mobile towers as a complement to diesel power, cutting fuel costs by 70-95% a year. These are relatively small windmills - 7.5 kWatts - but may serve as proof positive of the utility of wind for power in rural Africa. Helius Energy, a UK-based biomass energy company, is looking at the same market, building small power generation facilities that could power a mobile phone tower with excess capacity for local energy users.

Wouldn’t it be remarkable if innovative wireless phone companies ended up as the key force to wire Africa for electric power?