African middle class: myth or fact?
The continent’s statistics are said to be unreliable. Morten Jerven looks at Africa’s middle class figures.
On September 30, Kenya announced that it had revised its GDP upwards by 25 percent. Earlier this year Nigeria’s National Bureau of Statistics made an even bigger splash when a similar exercise showed the Nigerian economy to be 89 percent bigger than previously thought, displacing South Africa as the continent’s largest economy.
In 2010, Ghana revised its GDP by 63 percent.
In 2011 the African Development Bank declared that not only was Africa rising, but that statistics on income distribution revealed a sizeable middle class comprising 34 percent of Africa’s population – or nearly 327 million people.
Earlier that year, The Economist announced that International Monetary Fund forecasts predicted that seven of the 10 fastest-growing economies in the world over the next five years would be in Africa.
These statistical earthquakes, while good news, have shattered trust in Africa’s numbers.
Shanta Devarajan, chief economist of the World Bank’s Africa region, called it “Africa’s statistical tragedy” when he reflected on the quality and availability of quantitative evidence in the aftermath of Ghana’s GDP revision.
In retrospect, it may seem puzzling that bewilderment has greeted what has essentially been good news – Africa’s economies and its middle class are bigger than we thought.
But for too long we have neglected the accuracy of African economic statistics. We are only now waking up to the size of the knowledge gap because suddenly the numbers on African economies matter.
Investors and social scientists rely on accurate measurements. If 327 million middle-class Africans really existed, investors would consider the continent a potentially lucrative market for making deals in real estate, retail, wholesale and communications.
These huge numbers would force social science scholars to redefine and jettison hackneyed development phrases such as “subsistence”, “informal economies”, “food security” and “poverty eradication”.
However, the African Development Bank’s 2011 report conceded that about 60 percent of Africa’s middle class, about 199 million people, were barely out of poverty. This startling admission was based on its expansive definition of the middle class: individuals who spent between $2 (about R23) and $20 a day.
For political scientists, the middle class is the backbone of a democratic society. In Marxist theory the rise of the bourgeoisie permits progressive modernisation and industrialisation.
For investment banks, multinational corporations, real estate developers and traders, the middle class is defined by purchasing power and signifies a potentially untapped market.
For this reason, a more accurate definition of the middle class requires a higher purchasing-power bracket that shows that households are living beyond subsistence and that its members are also high school and university graduates.
Researchers affiliated with international organisations and investment banks have also tried counting Africa’s middle class.
Some surveys, such as accounting firm EY’s 2012 “Africa by numbers” report, dance around the actual size, and prefer instead to refer to “a growing middle class”.
Similarly, “The Rise of the African Consumer”, a 2012 report from McKinsey, a consulting and research company, stays out of the numbers game and makes no mention of the middle class. Standard Bank released a report in June assessing 11 sub-Saharan economies, or half this region’s total GDP, to measure the size of the continent’s middle class.
Based on these reports, the size of Africa’s middle class ranges from as few as 15.7 million households as estimated by McKinsey to the 327 million people the African Development Bank assessed in 2010.
Completely different monetary definitions of the middle class drive these differences.
The African Development Bank’s bottom threshold of $2 a day is far lower than McKinsey’s $55, Standard Bank’s $23 or the $10 a day used by the Organisation for Economic Co-operation and Development. Also, this organisation and the African Development Bank report their statistics in total number of people, while McKinsey and Standard Bank report on households without specifying their size.
It may appear puzzling that Standard Bank defines the middle class as households that spend between $8 500 and $42 000 a year, while McKinsey’s 2010 “Lions on the move” report defines this group as households that spend more than $20 000.
This can be reconciled: McKinsey includes all households above $20 000 in disposable income. This means they also count very rich households, which explains why their estimate is higher.
In its other report, “The Rise of the African Consumer”, McKinsey contends that 40 percent of spending power growth will come from households that earn more than $20 000 a year. It says “this group accounts for just 1 to 2 percent of total households”, but that this income cluster is “growing faster than the overall average, in numbers and in average income”.
So what are we left with? We went from a middle class that comprises 34 percent of Africa’s population to one of 1 to 2 percent.
But this tiny group is not middle class: they are rich households that have the fastest-growing incomes. Ultimately, what we are seeing is not a pyramid bulging in the middle as in the picture drawn by the African Development Bank.
The numbers from McKinsey and Standard Bank describe a society where the top spenders are getting richer. This may be good news for some banks and investors, but it does not carry the same connotations for social scientists.
None of the above, however, explains how these numbers were calculated or whether they are trustworthy. It is highly likely that many of the GDP growth numbers exaggerate increases in productivity and improvements in living standards.
Ghana and Nigeria’s GDP ballooned following the introduction of new benchmark years for estimating GDP in 2010 and this year.
How confident can one be about a 7 percent growth rate in a country like Nigeria when almost half of the economy was missing in the official baseline?
Some commentators proclaim that Africa is growing faster than its outdated measurements suggest. True, some countries’ economies are larger than those shown by these old numbers.
But that does not mean that recent growth has been faster too. The opposite is likely.
An outdated baseline means that “new” growth is more than likely “previously unrecorded” growth. When the base is too small, the proportion of economic growth will be overstated.
Moreover, when statisticians and politicians know that their numbers are minimising total GDP, it is tempting to add a bit each year to pre-empt a large upwards revision when the GDP numbers are ultimately corrected.
GDP growth estimates are also misleading because only parts of the economy are recorded.
Changes in exports and foreign direct investment are quantifiable and easily measured, while other important sectors that may be moving less quickly, such as food production, often remain unobserved.
In developed countries, like Norway, individuals and companies’ income, production and expenditure are reasonably well recorded and available through administrative records.
The government routinely collects this information as part of its day-to-day operations.
In poorer countries, few companies and even fewer individuals, households and farms record or report income, production and expenditure.
To get a measure of how income is distributed in a country and how many people earn less than $2 a day requires drawing a graph with income on the X-axis and population on the Y-axis.
On such a graph the share of households that earn below $2, $3 or $4 a day can be seen, as well as the income ratio of the top 1 percent and bottom 10 percent.
Drawing this graph presumes this information is reliable. In practice, however, these numbers are mostly non-existent because data collection is expensive and time consuming.
The most common audit, the Living Standards Measurement Study, is used by the World Bank to obtain poverty statistics.
It requires each household to spend a day filling out a long questionnaire.
A typical survey with a sample of about 2 000 households costs a few million dollars. From data collection to dissemination takes another two years.
According to a May 2013 report by the Brookings Institution, a Washington think-tank, six of sub-Saharan Africa’s 49 countries have not conducted a household survey and only 28 have done one in the past seven years.
Reports on the size of Africa’s middle class highlight presumptions and (mis)calculations.
The Standard Bank report, which provides a conservative estimate of the size of the middle class, is based on a sample of 11 sub-Saharan African countries.
The problem is that data availability is not random – it is biased because we know more about the richer economies, such as Nigeria and Ghana, than we know about poorer, more problematic countries such as the Democratic Republic of Congo, Somalia or the Ivory Coast.
It is undeniable that more goods are leaving and entering Africa than 15 years ago.
But does the increase in the volume of transactions result in a sustained lift in living standards? Some might argue that a positive African narrative and the power of self-fulfilling prophecies can make the vision of a huge middle class in Africa come true.
A fact-based outlook, however, is the best path. Does Africa’s population really have more spending power? Are fewer Africans hungry?
The evidence on income distribution does not provide accurate answers.
Everyone wants to know if the continent is better off, but proclaiming that it is without solid proof may backfire – particularly if poverty reduction and income distribution are slower and more unequal than publicised.
Impartial and inaccurate numbers, too, often lead to poor policy decisions. – Africa in Fact
** The views expressed here are not necessarily those of Independent Media.