Morten Jerven Morten Jerven

IMF and Ebola: why we don’t have any good answers

The debate on whether IMF has any blame for Ebola has been frustrating, even infuriating. It is going to continue to be frustrating largely because: “Who is to blame for Ebola?” is really not a good question. In turn, we have not got good answers.

  1. IMF is to blame.

That’s a bad answer. And that’s the how many, including Chris Blattman summarizes the view taken by the authors of the Lancet piece.  But Kentikelenis, King, McKee and Stuckler carefully stated “it could be that the IMF had contributed to the circumstances that enabled the crisis to arise in the first place”. It should be noted that it was provoked by the IMFs announcement of US$ 430 million of funding to fight Ebola. The IMF was selling itself as the solution to Ebola. But you could plausibly argue that IMF was part of the problem as well.

  1. History and Institutions is to blame.

That’s the view taken by Blattman in his first response. It is about state capacity, and it is wrong to assume that Africa is a blank slate. He is implicitly evoking the economic growth and political science literature that argues that the reasons why ‘good’ or ‘bad’ external policy advice makes little sense in countries like Sierra Leone, Liberia and Guinea. I am not that impressed with the research this argument relies on. Sure, according to the regression results – high ethnic fragmentation, colonial institutions, geography, slave trade, just to mention a few – is correlated with ‘bad’ institutions, ‘bad’ policy and low income per capita today.

I got two main problems with this. One is that it carries the ‘why aren’t you more like Denmark’ policy implication. Second, it is often forgotten that if you read the regression results there is only a fraction of the development outcomes today that is explained by ‘historical’ factors. I do sympathise with the view that is wrong to ignore politics and history – but I think that the literature that warns against the blank slate approach is equally wrong about ignoring the fact that recent history and current actions do actually have decisive impacts. The causality is messy, and because it is cleaner econometrically, messy current day politics is treated as an outcome. Yes, the answer is in politics and history, but IMF plays a role in that history.

  1. IMF is not to blame.

I could not make up a weaker response for the IMF than the one they came up with. They ran some econometric tests (with the fiscal data they themselves help country authorities to aggregate by the way) – et voilà: “we find an increase in health spending as a percent of GDP. In Guinea, spending increased by 0.7 percentage points, in Liberia by 1.6 points and in Sierra Leone by 0.24 points (from 2010 to 2013).”

0.24 percentage points? Really? So in Sierra Leone, according to World Bank statistics, total health spending as a share of GDP (note that this is not public, but total) was 15.24 percent in 2010 and it was 15.08 percent in 2012, the World Bank does report 2013 yet. In my experience with error margins in official statistics like this (remember, the denominator is GDP which can be a bit wobbly), that is not enough to assert direction of change. Look here, what happens to health spending as a share of GDP according to World Development Indicators.




The pattern is stagnation. We start at 15 percent in Sierra Leone and end up at 15 percent. We start at 6 percent in Guinea and we end up with 6 percent. That’s the end of the civil war. So to find a percentage fraction in this history and say: ‘look that’s me’, is not a good answer.

Towards a better question:

Obviously I agree with Blattman in that one needs to take into the opportunity costs when making statements about how much should be spent. I also agree that we need better research to get better answers. The ones I reviewed here are all pretty sloppy, but the original question that actually should be posed is whether the Ebola crisis have shown that over the past decades IMF focus a bit too much on fiscal discipline and too little on the importance of capacity in education and health. Most agree that too much was cut in the 1980s and 1990s, and there bits and pieces of evidence that Kentikelenis, King, McKee and Stuckler can point to show that the priority given to increasing capacity to deliver health services Sierra Leone, Liberia and Guinea is not high enough.

Final note: In all this I was reminded by the excellent book by Griffiths on Sierra Leone in 1986. Worth a read.


6 thoughts on IMF and Ebola: why we don’t have any good answers

  1. Good points all, and no disagreement from me. My only question/comment is on the poor quality research on history and institutions to blame. Are you talking about the statistical literature? Or the historical and comparative literature? The statistical literature is thin and weak, while the historical and political one is rich.

    For me, it’s hard to read Africa-wide studies of pre-colonial history (like Herbst), accounts of the impacts of colonialism (like Mamdani or Young), descriptions of the dysfunctional impacts on post-colonial politics (Sawyer, Jackson & Rosberg, van de Walle, etc.) and not be struck by the overwhelming challenges of state formation and political stability.

    The specific history of place like Liberia (elite capture by Americo-Liberians for most of history, 1980 coup by a junior military officer, increasing disorder through the 1980s without any real IMF or World Bank engagement, complete state collapse in 1990 when invaded by rebels and neighbors, etc.). It’s hard to see an answer for underdevelopment in anything but history, politics and institutions.

    Add to this the simple variable “time” (non-coastal Africa attained access to global trade and technologies hundreds of years later than everywhere else, and it takes decades or centuries to build a modern state and polity) and even without colonialism or slavery it’s unlikely that West Africa could have developed coherent polities, stable social contracts, and states with capacity by now.

    In this perspective, it’s hard for me to see the marginal effect of IMF policies in one or two decades as very decisive. In some countries you could make a good argument. Not Guniea, Sierra Leone, or especially Liberia.

  2. Why is the focus only on the IMF? Why have country donors contributed so little health funding to countries with such manifest health needs and very low health staffing? According to the Development Assistance Committee Creditor Reporting System total annual health assistance over the period 2010 to 2012 averaged just US$4.60 per capita in Guinea, $24.60 in Liberia and $14.40 in Sierra Leone.

  3. First, 15% of GDP on health is ENORMOUS. Passed only by the US. If this figure is accurate, such a large share is clearly a result of low GDP but signals the high priority given to health. Second, public spending as a share of total health spending went up in real terms quite dramatically. You can hardly argue that Sierra Leone did not prioritize health spending, and the country did the right thing by shifting away from private, out-of-pocket payments to public spending.

    Second, my frustration with this literature on IMF and health spending is its lack of rigor. While it seems absurd to mention again, correlation is not causation. This has been an ongoing problem and I blogged it here:

    The IMF study shows that the presence of an IMF program has a slightly positive but -as you say- marginal effect on public spending on health. That means that the IMF does not matter for health spending, right?? The issues lie elsewhere….

  4. Ms Glassman wrote:

    “The IMF study shows that the presence of an IMF program has a slightly positive but -as you say- marginal effect on public spending on health. That means that the IMF does not matter for health spending, right?? The issues lie elsewhere….”

    The issue here is the counterfactual. To the extent that public spending on health could have been considerably higher than it is then the IMF matters much.

Leave a Reply

Your email address will not be published. Required fields are marked *

Privacy Preference Center


Cookies that are necessary for the site to function properly.



These are used to track user interaction and detect potential problems. These help us improve our services by providing analytical data on how users use this site.

_ga, _gat, _gat_, __cfduid