Why Africa’s economies continue to grapple with ‘statistical tragedy’

Economic data in Africa has been under sharp focus for its inaccuracies. Without a statistical revolution, economists argue, Africa’s renaissance is being built on shaky ground. Business Beat had an exclusive interview with Canada-based Prof Morten Jerven, an economic historian whose development statistics have been widely published. He has worked in Ghana, Nigeria, Uganda, Kenya, Tanzania, Malawi, Zambia and Botswana. Excerpts:

How is economic data in the African nations you have worked inaccurate?

Gross domestic product (GDP), and other development metrics, rely on timely and accurate reporting on a range of variables, such as food production, food consumption, construction of rural roads, beer brewing, cement consumption, retail mark-ups for clothing, sweet potatoes and spare automobile parts, as well as imports, exports, wages, population, how the government taxes and how it spends, and so forth.

While this data is most of the time readily available through personal and corporate tax records, as well as administrative data in most developed countries, many of these areas are not well accounted for in poorer economies. That means that data is hard and expensive to come by. That data may not be frequently updated, and in many instances, producers are making qualified or weak guesses instead of delivering facts.

How can one tell if economic numbers are being manipulated?

There are three main safeguards for accurate and reliable facts being delivered from national statistical offices. The first is the legal independence of the statistical office. Only a minority of statistical offices in the African region are formally independent from the central government.

The second is funding. Does the statistical office have a sufficient flow of funds that is not dependent on the data it collects and disseminates? Most statistical offices in Africa rely heavily on donor funding, which in turn means that they are often required to collect very specific data that is needed for ad hoc projects, rather than responding directly to data needs from citizens.

Third, and this is preconditioned on the two above, access to needed data sources to independently deliver facts. Most statistical offices are legally dependent on the government, financially dependent on donors and have limited access to data that would allow them to collect, process and disseminate the facts. This leaves statistical offices vulnerable to manipulation and political pressure.

What are some of the ways of manipulating data?

It is not straightforward to detect data manipulation, and not all data is manipulated. But if the data matters a lot, and the incentives are strong, there is ample opportunity to exert pressure or otherwise affect data when there are large unknowns and so much remains uncounted.

When countries have an interest in receiving aid and concessional loans, they can refrain from updating macrostatistics to seem poorer, whereas if there is a need to appear wealthier — to improve leverage ratios, for instance — they can update macroeconomic statistics frequently and also exert pressure and choose ways in which data is aggregated to benefit the users of the data.

Administrative data — such as numbers of students enrolled in school, or food production, which are reported from government officials in districts that are rewarded for reporting ‘good results’ — can often be an overstatement of real effects. We have evidence of such effects by comparing survey data with administrative data. Note that in this case, the data is manipulated from below rather than from the top.

Kenya has just rebased its GDP. Should citizens be concerned?

It was a nice opportunity for Kenya to rebase now, in line with issuing bonds for international financial markets. But I also think that there was good reason to believe that GDP was understated. Whether the current GDP now accurately depicts the size of the Kenyan GDP is another matter.

The new GDP number raises a lot of new questions. What about the poverty data? If GDP is higher, how come poverty headcounts are still higher in Kenya than in Tanzania and Uganda. It shows that we need to know more about the income distribution and the political economy of growth in Kenya. Who benefits from the rise in Kenyan GDP? It is also important to not only focus on the validity of the new numbers but think outside the box. What is counted and what is not? What information do we not have access to?

What can Kenya and other African countries do to make their economic data as valid as possible?

Citizens should demand more and better data, and media should play a key role. Journalists should engage African academics, think tanks and NGOs to facilitate a rich and diverse debate on statistics and data for development. This is a long process of stimulating a healthy and accountable debate.

An independent statistical office may be used by journalists and citizens to hold governments accountable, and it comes down to citizens and journalists making intelligent use of the resources, or demanding more independent information to spur this process further.

Beyond that, I think that there is need for legal reform in statistical offices and ways to re-structure the financing of statistical offices so that these institutions can provide the needed day-to-day data for governments and citizens.