Development’s False Divide
Giving Internet access to the world's poorest will cost a lot and accomplish little.
Politicians, business people, donors, and the press have all proclaimed the digital divide between the developing-world poor and the wealthy West as one of the century’s most significant development challenges. In fact, 99.6 percent of the populations of Africa and South Asia did not use the Internet in 2000. "The digital divide threatens to further marginalize the economies and peoples of many developing countries," concluded the U.N. General Assembly in June 2002. The Group of Eight declared at Okinawa in 2000 that "everyone should be able to enjoy access to information and communications networks."
The best of motives may drive a concern to equalize global Internet access, but not the strongest of logic. True, tools of communication are important to the world’s poorest, and one can also find many examples of effective Internet use in developing countries. For instance, the Internet has been used to inform farmers of crop prices in Argentina, to register deeds in India, to educate children in rural Uganda, and to sell woodcarvings and sandals in Kenya. But it is a large leap to conclude that global Internet access is a sensible goal. Uplifting anecdotes are not enough to justify the high costs of universal Internet access, costs that would be at their highest in the least developed countries.
One reason for the high cost of providing widespread Internet access to low-income countries is that about 69 percent of their population is rural. Providing networked services like electricity and telephony to rural areas is expensive — and because rural people are largely poor, it is hard to justify that cost in terms of potential revenues. Solar power and satellite connections are a potential alternative, but such technology further increases the cost of Internet access. In Costa Rica, for example, one off-grid telecenter carried an annualized cost per Internet-enabled computer of about $10,000. By way of comparison, the average person living on $1 a day (and there are 1.5 billion such people worldwide) spends about $10 per year on communications when he has access to it.
Subsidized public access is one answer. The subsidies would have to be large, however. Ensuring one hour a week of access at a telecenter such as the one in Costa Rica might cost as much as $50 per year per capita — or about 10 times public spending per capita on health in low-income countries and 10 times discretionary spending per primary student. On this basis, the worldwide subsidy for everyone living on $1 a day to get one hour of access a week might reach $75 billion — considerably more than the global total of aid flows each year. Given that providing widespread Internet access will be complex and expensive, attempting to provide ubiquitous service will be a costly mistake.
Some argue that access could be provided more cheaply — perhaps at levels that only equal average health expenditures in low-income countries. Nonetheless, costs of access would probably still outweigh benefits because the digital divide encompasses far more than a physical lack of access; it also relates to deficits in skills and the broader economic environment.
Lack of education is a major barrier to productive Internet use, for example. In Ethiopia, 98 percent of Internet users in 1998 had a university degree, yet 64.5 percent of the overall population is illiterate. Worldwide, most people living on $1 a day are illiterate. Further, they usually speak a minority language in their own country — few speak a major global language. For example, about 17 million people in Nigeria speak Igbo. My search for Web pages in Igbo turned up only five sites: a translation of the Universal Declaration of Human Rights, a translation of a document called "The Four Spiritual Laws" (theological provenance undetermined), a translation of the food pyramid, a two-page Igbo phrase book, and a prayer manual. There isn’t an Igbo translation service on the Web, so an Igbo speaker would be limited to these five. None involved sound or video, so the illiterate Igbo speaker would gain nothing. Bridging the gaps in language and technical skills as well as basic literacy will be difficult, considering the small per-student spending available in the poorest countries’ primary schools, where the discretionary budget per student is as little as $5 a year.
Even if poor people are lucky enough to be literate and conversant in a major world language, their use of the Web for activities such as e-commerce is likely to be limited by their lack of credit cards, not to mention the challenge of persuading FedEx and ups to start delivery services in their neighborhoods. Limitations in relevant content and ability to use that content perhaps best explain why only 2.2 percent of India’s Internet users have ever engaged in buying or selling over the Web. Similarly, a survey of Tanzanian firms found that among the 30 percent of companies with access to the Internet, less than half use it frequently and only 9 percent rated it as a very effective tool for promoting products.
Communications matter to the poor. A system of well-regulated, competitive communications services will reduce costs and extend access. In many cases, it may well be worth extending access to telephony with limited, targeted, carefully designed subsidy programs. But pursuing universal access to the Internet would be a misallocation of considerable resources. To draw an analogy, another technology boasts a 70-fold difference in access rates between the United States and India, and economists link that technology to increased productivity as well. But no one is setting up a U.N. task force to overcome the Air Conditioner Divide.
Poor countries face many serious divides, including those in education, healthcare, and transportation. The relevant question for the poorest is, does the lack of access to a particular good provide a significant barrier to becoming more wealthy? The answer is yes for the tools of communication in general but no for the Internet in particular.