The conventional wisdom in the ICT4D community is that people in developing countries spend a higher percentage of their income on communications than the global average. For South Africa, this notion became is widely accepted (I think) thanks to the Vodafone study on mobile usage, in which Diane Coyle of Enlighenment Economics says:
“Expenditure on mobile phones as a proportion of total expenditure can give some broad information on their importance and impact on household budgets. In South Africa, 134 mobile phone owners were happy to provide information on their income and their mobile phone expenditure. These respondents spent on average between 10 and 15 percent of their income (or 89 to 108 Rand) on mobile phones (estimation was made using mid-points of income and expenditure brackets). However, as only one respondent identified mobile phones in their top three expenditure items, so these figures should be treated with caution. National data suggest the biggest items of expenditure for the poorest black South Africans (urban and rural) are food (about 50 per cent of the household budget), fuel and energy, and housing (each 7 to 8 per cent). Transport and communications follow after these categories, however, in the national statistics.”
From this one may infer just how important telecommunications is to people, its utility and the pent-up demand for access. However, I see in the recent survey of “Income and expenditure of households 2005/2006” released last week by StatsSA, they say this about expenditure on communication:
Nationally, 3,5% of household consumption expenditure was attributed to communication, the largest component of which was telephone and telefax services. There was little disparity between provinces regarding communication’s share of consumption expenditure. Limpopo had the lowest at 3,1% and KwaZulu-Natal at 4,0%. There was even less disparity between population groups. Black African households spent 3,3% of their consumption expenditure on communication (the lowest expenditure out of the four population groups) compared with 3,9% for Indian/Asian households (the highest).”
These statistics are more in line with statistics reported by Botswana Central Statistics Office. I guess sample size is everything 🙂 and to Diane’s credit she does offer a disclaimer yet it didn’t stop that figure finding its way widely into usage.
I am looking forward to the impending publication of the ResearchICTAfrica household survey results later this year. Hopefully they will shed more light on this.