A nurse marks a child after administering de-worming medicine. If you can't count it, it becomes nearly impossible to manage. (Photo/USA Africa/Kat McDowell/Flickr).

Africa’s Has A Data Problem, But Both Bad And No Numbers Can Still Be ‘Good’ Things

THE unreliability and paucity of African data continues to vex.

In the space of one week alone The Conversation published two articles on it; Morten Jerven’s “Studying Africa by numbers can be misleading. What can be done about it”, bemoaning the continent’s patchy and unreliable economic data, and Damtew Teferra’s “Ranking African universities: hypocrisy, impunity and complicity”, slating the research on which Africa university rankings are based.

However, I still find cause for celebration in the face of bad or absent data.

If a government or institution lies about data to show poverty has reduced where it has instead risen, it shows that it recognises that numbers – as opposed to speculation and opinion – are important enough to lie about. That is much better than denial, or a refusal to even do bad research.

Secondly, it is easier to know if data is lying or lousy, as Teferra’s article aptly demonstrates. But if you are arguing with the minister of the Economy, and he asserts that there is growth, and only people who don’t wish the country or lack faith deny it, you are in a long fruitless night.

In addition, bad or no data reveals state (in)competence. A government that cannot count how many classrooms there are in the country, is unlikely to develop a world class education system. There is no reason to believe it if it says it has met its 4% annual inflation target.


An army that thinks it has 100,000, while 60,000 are “ghost” soldiers, and goes to war on that basis, will probably lose. If you fool around with data and believe your own lies, you get punished.

You can protect yourself better against an incompetent who reveals himself, than one who masquerades as an expert.

However, perhaps the biggest problem with data in Africa is not the figures and even reliability of the research, but the philosophical and ideological context in which it happens.

In 2009 Kenya had a population census, and the results were released in 2010.


On the outskirts of the Ifo Refugee Camp in Dadaab, Kenya, Somali women work in the field where they are able to learn to grow their own crops. (Photo/Kate Holt/Africa Practice).

The results for the Somali population in the semi-arid northeastern districts near the border with Somalia shocked many.

Kenya’s Somali population, the census showed, had shot up to 2,385,572 from less than 800,000 within 10 years.

There was an uproar, with media commentators and Parliamentarians alleging that the sharp rise was accounted for by an influx of “illegals” and refugees from war-torn Somalia.


The Planning minister promptly cancelled the Somali numbers.

It was plausible, but not definite, that the number had been inflated by “illegals”.

According to the census, at that point Kenya had a population of 38,610,097.

The Somali had shot to the 6th largest national group, from close to bottom of the pile at the last census. The census found that the largest seven ethnic groups were the Kikuyu at 6.62 million, Luhya 5.33 million, Kalenjin at 4.96 million, Luo 4.04 million, Kamba (3.89 million), Kenyan Somali (2.38 million), Kisii (2.21 million).

I asked a highly regarded Somali academic in Nairobi what the truth might be. He said he thought the Somali figures were “roughly accurate”. He explained that the sources of the problem were the earlier census counts.

In the past, the Somalis in the northeast were very nomadic, and could only be counted if officials went the extra mile to do so. Because of the marginalisation of the region, they didn’t, so they were undercounted.

Secondly, because poverty rates were high in the region, birth rates too were.

By 2009, due in part to the collapse over 10 years earlier of Somalia “Proper”, northeastern Kenya had changed as it had become the fallback region for servicing what remained of the broken Somalia economy up north, and as formalisation grew in the area when humanitarian agencies moved to cater to the swelling army of Somali refugees.

The 2009 census was also more rigorous, reaching more people. With a more settled population, the numbers could only be “very high”.

If he was correct, then the problem with the Kenya Somali population data was not that it was wrong. Rather it was accurate, but inconvenient.

Kenya’s politics has historically been organised along a sometimes fine, and sometimes blurry, line that divvied up the political spoils between the largely Christian Nilotic-Luo and the Bantu elite.

The Somalis were crashing the party. They were messing up the formula.

However, their rejection vividly revealed what is Kenya’s number of the future.


    Leave a Reply