Why Africa is much richer than you think

We place so much reliance on statistics to work out if countries are growing or receding but just how accurate are the GDP figures that are released by various bodies?

Ghana revised its figures and discovered that it was much richer than it thought it was; Nigeria is in the process of doing the same. Richard Walker investigates the accuracy of African statistics.

Nigeria, already the most populous country in sub-Saharan Africa and the second-biggest economy in the region, is about to get a lot bigger. Not literally – but when the Nigerian authorities release their long-promised revised estimate of the size of the economy (it was due out in December 2013), the total of Gross Domestic Product (or GDP) is certain to rise, quite probably by a large amount. Nigeria is not the only African country busy re-estimating the size of its economy. So what is going on here?

When it comes to the economy, GDP is the biggest number of all. It attempts to capture everything that is going on in an economy in a single measure. It includes everything that a nation produces, all supposedly valued down to the last cent. The trouble is, GDP figures are also highly unreliable.

That does not apply only to African countries. GDP everywhere is based on a series of estimates and guesses – at least for the initial GDP data releases – and those estimates usually go on being revised and adjusted for anything up to 18 months.

But Africa is a special case. That is something that became apparent in 2010 when Ghana released its new estimates of GDP – when overnight, the Ghanaian economy grew by over 60%. Or so it seemed: what in fact had happened is that Ghanaian statisticians had revisited all their assumptions about what kind of economy had developed over the previous two decades, and found that they were seriously out of date.

Everywhere in the world the rate of change in GDP is calculated with reference to a ‘base year’. The base year tells economists what the size and the structure of the economy was at a given moment, and the new data for the latest year are compared to the previous year, and on back to the base year to give figures for GDP growth (or GDP shrinkage, if you are unlucky enough to be in recession).

In some economies where statistical offices are well-funded and the data are usually up to date, the base year is continually updated and may be no more than a year or two before the present. At a minimum, economists believe that the GDP statistics should be rebased every five years (according to the IMF, this is international best practice).

But in Africa the base year is often decades out of date. Based on survey evidence, the IMF believes that no more than seven countries (Burundi, Ghana, Malawi, Mauritius, Niger, Rwanda and Seychelles) rebase their GDP that often, and that most countries believe GDP is underestimated.

Just to make things more complicated, there may be more than one base year involved – another statistical base is the ‘reference year’ which gives a baseline for the prices used in estimating GDP. But it is the base year (or ‘benchmark year’) for GDP volumes that remains the most important.

 Out-of-date data
According to the African Development Bank (AfDB), there are 19 countries in Africa where the GDP base year is more than a decade old, and at least 10 countries where it is much older still – the Democratic Republic of Congo and Equatorial Guinea have base years that are 30 years old. None of these meets international standards for statistical accuracy, says the AfDB.

Does this matter? It certainly matters if you need to know what is going on in your economy. If a country’s base year is set at a point decades ago, the economic data will all be based on the assumption that the kind of economic activities that go on today are exactly the same as those of 20 or 30 years ago. That means no computers, no mobile phones, no internet, no private airlines, little international trade, little manufacturing, and in many cases, no energy industry. Figures for growth in all these areas will not be included in the GDP results, because they didn’t exist 30 years ago so there is nothing to compare them with in the base year.

It is easy to see how an African economy might end up seriously underestimating the size of its economy – and that is without taking into account the informal or black economy, which is also largely invisible in the statistics (when Mozambique undertook its first comprehensive survey of the informal economy in 2004, the result was a doubling of the annual private expenditure estimate). What is less easy is solving the problem.

Africa’s national statistical offices are often underfunded and understaffed – in fact, in many countries you will be hard pressed to find out even whether there is a statistical office, let alone where it is and what it does.

In fact, these departments should be undertaking enormous data-gathering operations. Consider, for example, the GDP data process in most developed economies, where GDP is estimated using three different measures: the output measure (how much is produced), the expenditure measure (how much is spent) and the income measure (how much is earned).

In theory all three measures should come up with the same answer, but the three sets of data come in at different times, so all three measures are used to provide early quarterly statistics as well as later and more accurate annual statistics.

In the UK, for example, the output measure which provides the earliest data involves at least 46,000 quarterly surveys of companies in manufacturing, services, retail and construction, together with additional figures drawn from government monitoring of agriculture, energy, health and education. That is a lot of work, and many African governments simply do not have the budgets to undertake it.

What happens when such detailed survey data are not available is that statisticians make guesses, based on ‘proxies’ for growth – that is, using measures that they do have to hand, measures which they think will be moving roughly in line with economic growth. One such proxy is taxation, which has the advantage that most governments are willing to spend money on tax gathering so at least there is an annual statistical base to work with.

But taxation is a flawed measure, because it does not include the informal economy, and does not take account of fraud by taxpayers or tax gatherers. Another proxy is population growth, but that is also a primitive proxy: population growth itself may be no more than a guess, depending on how frequent and how accurate the population census is, and it does nothing to fix the problem of the out-of-date base year.

So, the biggest problem is that GDP base years are out of date, and African statisticians know they are out of date, leading to wildly varying estimates of the real value of the economy.

For example, in the case of Ghana, politicians and economists had been arguing about the figures for many years. Back in 2004 the World Bank estimated that Ghana’s per capita GDP implied a per capita income of $380. Then President John Kufuor chipped in and said the real figure was $600. Not to be outdone, the finance minister said it was more like $1,000.

What happened next was a huge effort to update the GDP data and methodology, one that is now being reproduced in several other African countries, with the help of the IMF and the AfDB (which says it has spent $100m over the last decade building up better GDP data gathering across the continent).

Ghana remains the model for this effort. In Ghana it had been clear for a long time that something was wrong, simply because the figures looked wrong. Using a base year of 1993 (which itself had recently been updated from an earlier base year of 1968), the ratios of taxation to GDP, or expenditure to GDP, all looked suspiciously high compared to other countries, implying that the GDP estimate was much too small. This was confirmed by the 2003 Industrial Census, which showed that VAT receipts were far higher than GDP would suggest. The problem was: how to fix it?

Discovering the real story
Ghana had to embark on a very wide-ranging process designed to discover what the real structure of the economy was. Data were collated from the Industrial Survey, the 2005/6 Living Standards Survey, and a 2007 infrastructure survey. That was added to economic data like current VAT receipts, the accounts of the banking sector, and call volumes from the entire Telecom sector.

The result of all this was that a completely updated picture of the economy emerged in the new 2006-based GDP base year, showing how a modern African economy was evolving.

In the 2010 national accounts, agriculture accounted for 30% of the economy, down from over 35% when interpreted according to the old base year. The share of industry also fell, from around 28% under the old base year to just over 18%. The biggest shift of all was in services, which now accounted for over half the economy, whereas under the old base year it would have been around a third. In other words, Ghana was shown not only to have a much bigger economy, but also a more modern one, with services leading the pack.

And the impact of all this? Apart from giving a much clearer picture of what was happening in Ghana, important for policy makers and investors alike, there were also other effects. Overnight Ghana went from being a low-income country to a middle-income country (and incidentally that meant that Ghana was no longer eligible for concessionary World Bank lending).

Also, the debt-to-GDP ratio suddenly dropped, revealing Ghana’s finances to be much healthier than previously thought. At a stroke the risk ratings of fixed-income investment in government debt were cut, making Ghana look like a better investment bet, and more than cancelling out any loss from the move to middle-income status. In a word, Ghana was no longer poor.

Many economists believe that other African countries are similarly richer than the statistics suggest. But as the AfDB has pointed out, reproducing what has been achieved in Ghana is not going to be easy.

Both Liberia and Burundi have tried, but failed to get essential endorsement from the back-room statisticians in the World Bank and the IMF. Tanzania has revised its GDP base, but mistakes were made in collating historical and current statistical tables, leading economists to be sceptical of the result. Nigeria’s rebased GDP was due to be released last year, but is still awaited. Statistics, it turns out, aren’t easy: there is a lot of heavy lifting to be done.

And in the meantime, these exercises have, if anything, increased uncertainty over the real size of African economies. It is widely agreed that the Ghanaian revision was a good thing – the downside is that it is also widely agreed that comparative data on the size of economies and the wealth of Africans cannot be trusted until a standardised statistical practice is adopted across the continent.

africanbusinessmagazine