The Washington Post predicts a year full of coups in Africa
Serginho Roosblad | January 31st, 2014


The Washington Post’s foreign affairs blogger Max Fisher (about whose infatuation with coloured maps we blogged before here and here) posted an entry earlier this week entitled: ‘A worrying map of countries most likely to have a coup in 2014′. It is based on the work of political scientist Jay Ulfelder. The post includes a coloured map of the globe with countries coloured from light yellow to dark brown. And as you might guess, the darker the country, the more likely it will see a violent overthrow of the government some time this year.

In Ulfelder’s study he takes a number of variables, such as the political stability and infant mortality rate. But he also took into account variables like how long a country’s been independent or who the last colonizer was. It’s not very clear how and why these could have an effect on the result, but according to Ulfelder’s blog these variables don’t necessarily have to have a significant effect on the risk of a coup.

This being Max Fisher, it’s not the first time he has had fun with maps and with Ulfelder’s research. Last year they did exactly the same thing, only then the headline read: ‘The countries most at risk for a coup in 2013.” At the time, Fisher called Ulfelder “the Nate Silver of coups.” Egypt didn’t make the list then. I wonder what changed in the last three hundred somewhat days that we went from ‘risk’ to ‘worrying’ and ‘most likely’?

Anyway, Africans: brace yourself, because the continent is up for an orgy of coups. From Guinea-Conakry to Madagascar, it doesn’t look very pretty, especially for the Sahel region.

But does the continent really stands on brink of political and civil chaos? Some readers at least seems to have a hard time believing this gospel. One, in my opinion rightfully, comments:

A coup d’ Etat is highly unlikely in the following countries: Mali, Central African Republic, and Guinea.

The two first are already under French and international community supervision; the third could face not a coup , but a lower level of civil unrest (maybe civil war) because of the Fulani ethnic group (financial power of the country) and the people of the forest region marginalization.

Another one writes:

A coup in Somalia seems more like a rhetorical exercise than an actual undertaking. There’s not a heck of a lot of government to overthrow, is there?

And on and on it goes.

And indeed, because when is an overthrow of the government considered a coup? And maybe equally important, when is a coup considered to be successful? As we said, Egypt for example did not even make it to last year’s list. However, some still struggle with what to make of the army’s interference in the country’s politics and deposing of the president. When writing an article with a screaming head as in this case, it would have been nice to at least get some context.

Apart from a failure to explain what exactly is meant by a coup, it is also quite confusing when the article is supposed to be about “countries most likely to have a coup”, but where the research the article is based on looks at the risk of attempts and not actual successful coups. And finally, it’s a bit of a downer if after all the predictions, you read midway through:

[E]ven the most extreme cases are well below a 50 percent likelihood of a coup, meaning that a coup probably won’t occur.

That’s where I stopped reading.

Weekend Music Break 67
The Work of Kenyan Photographer Mimi Cherono Ng’ok
The following two tabs change content below.

Serginho Roosblad

Journalist and producer for RNW Africa.

2 thoughts on “The Washington Post predicts a year full of coups in Africa

  1. Wow, what an incredibly irresponsible piece. If Roosblad could only be bothered to read to the fourth paragraph, 2nd sentence of a nine-paragraph summary of Ulfelder’s work, why should we take his analysis seriously? Roosblad’s framing suggests that Fisher didn’t make that low risk clear early in the piece, which is clearly not the case; he covered that point early in the piece, before even delving into the issues with African states.

    Beyond unfair misrepresentations of Fisher’s reporting, thought, he deeper problem with this piece is that the author neither seems to understand how data analysis works nor care about the reasons it is valuable. The only evidence he presents in response to a well-constructed, entirely transparent, available to anyone who wants to download it dataset is online comments from WaPo readers? Seriously? Social scientists use data to identify trends. Like it or not, we know that there are many characteristics of fragile states that increase the likelihood that they will experience a coup over time. When those characteristics are present in a state, it would be irresponsible not to point this out – we can act to prevent instability, mass atrocities, and civil war, among other bad things, when we use data to identify those trends.

    Is data science perfect? Of course not. A year ago, Egypt did not have the signs of a state prone to a coup. This isn’t a sign of racism; it’s a fact of data. I suspect that Ukraine might appear on the analysis if Ulfelder were redoing it in light of the events of recent weeks, but that the data didn’t indicate that six weeks ago is not a failure of the researcher or the journalist who reports on it. That one dislikes the data – or what it says about institutions like the rule of law in African states – does not invalidate it. To do that, you would need to analyze the data using equal rigor and transparency to prove otherwise.

    • I just want to leave a comment saying I agree wholeheartedly. Terrible piece. Critiquing a piece of data analysis without a) engaging with the analysis or b) reading the thing is idiotic and, worse, pretty shoddy journalism.

Leave a Reply