Paddy Carter (ODI) | Africa: What have economists got wrong?

Paddy Carter (ODI)

09 September 2016

Morten Jerven made a splash with his exposé of the woeful state of economic data in the developing world, Poor Numbers, and his second act Africa: Why Economists Get it Wrong has won him more fans. In this book he argues that economists were misled by cross-country growth regressions into thinking that Africa is incapable of development, and that by seeking to explain a failure of growth economists missed the chance to study historical growth episodes and show how Africa can grow.

In a forthcoming edition of Development Policy Review we are lucky enough to host a debate between Morten Jerven and three eminent economists, Denis Cogneau, Jonathan Temple and Dietrich Vollrath.

There is a lot of agreement around about the limitations of cross-country quantitative research and the need for closer study of individual countries. There is also some disagreement, with Jon Temple and Dietz Vollrath in particular arguing that empirical economic research is more useful than Morten is willing to admit. Morten mounts a methodological critique and our respondents concentrate on those points; none really tackles what exactly it is that Morten thinks economists have got wrong about Africa. That’s what I am going to take on here.

Morten’s call for more useful research into African economic growth, for better data and for richer, historically grounded, studies, should galvanise the profession, but I am going to argue that the gulf between how economists think about Africa and how Morten thinks they ought to is not as large as it may appear. Morten thinks economists are unduly pessimistic about Africa because they have concentrated on explaining why the continent is so much poorer than others today (or why growth averaged from 1950 until now is lower), which has led them to form the view that there is something about Africa that is inimical to growth. That something is ‘bad initial conditions’. Morten writes that the bottom line is that there is no bottom billion (those who Paul Collier wrote are trapped in countries with conditions inimical to growth). As Morten has it, part of the reason economists have gone wrong is that they have ignored the fact that African countries have often experienced periods of growth, when external conditions have been favourable (p.86).

I contend that what most economists actually think is that most African countries are saddled with circumstances that hamper their chances of achieving sustained growth, which requires structural transformation and continuing productivity improvements. I do not think there is an economist on the planet who believes ‘bad institutions’ mean a country is incapable of riding a commodity boom.

Unfavourable initial conditions do not imply growth is impossible. The two economists most closely associated with the credo ‘institutions rule’, Daron Acemoglu and James Robinson, emphasise that what they term ‘extractive’ economic and political institutions are compatible with decades of economic growth (as evident in China) and also that their theories do not imply ‘historical determinism’: two similar societies can drift apart and then, when a critical juncture arrives, head off in quite different directions.

If certain initial conditions are found to have explanatory power in a long-run growth regression, this only tells us something about what happened on average. Morten writes that such regressions tell us ‘that growth needs private property rights and that the Gold Coast and Ghana could not grow [as they did] under such institutional frameworks’ (p.86). That is not what regressions tell us.

Morten describes how political and economic institutions should be seen as adaptations to individual circumstances, and quotes Anthony Hopkins to argue that asking whether initial conditions determine growth is ‘rather like trying to decide if life is more difficult for penguins in the Antarctic or camels in the Sahara’. But economies that achieve long-run growth may resemble each other more than penguins do camels. Sustained growth requires investment, productivity improvements and structural change. Striking metaphors cannot rule out the possibility that some institutional adaptations are more conducive to this than others. There are many paths to growth, but there may also be some commonalities across country experiences, and cross-country empirical research tests this possibility.

If initial conditions are understood to mean whatever it is that makes the transition from externally driven growth to structural transformation and sustained growth more (or less) likely, what economists think about Africa does not seem to me very different from what Morten thinks, when he writes, ‘change is required in the political economy of African nations to enable them to weather difficult external conditions more effectively. Finally, a shift towards self-reliance and self-sustained growth is required. This means building institutions that can invest and reinvest returns from more prosperous times that can then be used to keep economies afloat when conditions are less favourable’ (p.88).

Morten focuses on regressions that seek to explain growth averaged over a long period or, equivalently, to explain income levels today, conditional on historical variables, and argues that research should concentrate more on what explains episodes of growth within Africa. But economists have been doing this for decades: the bulk of econometric growth research employs ‘panel regressions’ that identify which variables are correlated with growth episodes within countries. Useful research has been done on the consequences of trade liberalisation, public investment in infrastructure and recent signs of structural changes in Africa, for example.

One gets the impression that Morten thinks economists are unaware that many African countries have seen periods of growth. This is not so. Morten’s bête noire Paul Collier wrote in a survey article from 1999 ‘African performance has been strongly episodic’, and there are countless papers looking at what drives episodes of growth and the role of external factors (Easterly et al., 1993; Blattman al., 2004; Raddatz, 2007). Morten cites a paper on growth episodes by Lant Pritchett that I think every growth economist knows well. Readers may not come away from the book with the impression that economists are well aware that African economies have experienced periods of growth.

Moreover, if economists are Afro-pessimists who place too much importance on institutions, both these views are possible without going anywhere near a growth regression. Acemoglu and Robinson are at pains to point out that their theories draw on other forms of evidence. As for whether there is a bottom billion, forecasts see uncomfortably close to 1bn still living in extreme poverty by 2030.

Characterising what economists think about Africa is difficult because there are so many of them and they disagree about much. Cross-country regressions represent only a fraction of the research economists have done on growth in Africa. Some names that spring to mind, who look at African growth from different perspectives, are Stefan Dercon, currently DFID Chief Economist, Christopher Udry, Christopher Woodruff, David McKenzie and the doyenne of development economics, Esther Duflo.

So I think the picture is brighter than Morten paints. No doubt some economists have oversold their ideas based on flimsy evidence, but academics everywhere are prone to that sin. Cross-section regressions have well-known limitations but can be informative if interpreted with care, and quantitative research into African growth moved beyond that approach many years ago. On the specifics of what economists are wrong about, Afro-pessimism and the role of institutions and other initial conditions, I’d say it’s all a question of degree, and economists are not alone in thinking Africa faces many challenges.

Morten writes that economists have been trying to explain something that did not happen: African chronic growth failure. But if the problem is seen as failure to achieve sustained growth through structural transformation, as opposed to episodic externally driven growth, and questions around initial conditions and institutions are taken as asking what determines the chances of sustaining growth, then, I think, much of the apparent disagreement between Morten and the economics profession can be resolved.


Photo credit: Antony Robbins, 2008