There is a small but expanding literature on the link between fragility and development in Africa. Bertocchi and Guerzoni (2011) employ the OECD definition of fragility within a yearly panel dataset covering sub-Saharan Africa in the 1999-2004 period. Following the benchmark specification of Barro (1991) and Bertocchi and Canova (2002), they include in their growth regressions an initial condition for per capita income and a wide range of economic, demographic, geographic, and institutional factors. Their results indicate that the conventional measure of fragility employed by the OECD exerts no effect on economic development, once standard regressors are accounted for. However, when they apply a more severe definition of fragility, which only includes the countries in the bottom quintile, they find a clear, negative impact of this condition, even after controlling for endogeneity through instrumental variables estimation. Using a comparable sample and the OECD conventional definition, Baliamoune-Lutz (2009) highlights how fragility exerts a non-linear impact on per capita income and that it tends to interact with several other factors: in fragile countries, beyond a threshold level trade openness may actually be harmful to income, while small improvements in political institutions can have adverse effects. Fosu (2009) explores the growth impact of policy syndromes, which include among other components state breakdown, a concept which is in turn close to fragility since it refers to a condition involving civil wars and acute political instability (see Fosu and O’Connell, 2006, for a definition of policy syndromes). His findings are that the absence of policy syndromes encourages growth in Africa.
The potential endogeneity of fragility is a serious concern, which has been addressed by Bertocchi and Guerzoni (2010) by gauging the links between fragility and other standard growth determinants. They find that indeed, within Africa, fragility tends to be shaped by institutional development, a conclusion which questions its exogeneity. In particular, the probability of a country having a fragile state appears to decrease with the level of civil liberties and to increase with the number of revolutions, while economic factors do not matter. These findings differ sharply from those presented by Carment et al. (2008) for a world sample, over which per capita income appears to be the main driver of fragility. This radically different conclusion can be explained, once again, by the specificity of the African region, but also by the fact that the former study employs the OECD definition of fragility, while the latter employs the index of Failed & Fragile States.
Beside these empirical investigations, Besley and Persson (2011) propose a theoretical framework to understand how fragility can hamper development and growth. Their theory highlights how a state may become fragile in situations of external or internal conflict, high political instability, and heavy economic distortions, and how fragility may in turn lead to poverty traps.
While the above contributions focus on the direct link between fragility and development, others have looked at its indirect influence through aid allocation. Since the condition of fragility is a crucial determinant of the amount of aid a country receives from international organizations, growth can be affected by fragility also through this channel. The interaction between aid and fragility is addressed in a number of studies, none of which is specifically focused on Africa. However, given the preponderant role played by African countries among fragile ones, their results are still useful to the present perspective. Burnside and Dollar (2000) provide evidence that aid is most effective in developing countries with sound institutions and policies. However, this conclusion is questioned on several grounds by Hansen and Tarp (2001), Dalgaard et al. (2004), and Rajan and Subramanian (2008). McGillivray and Feeny (2008) study the growth impact of aid in a world sample of fragile countries and find that it depends on the relative degree of fragility. Chauvet e Collier (2008) analyze the preconditions for sustained policy turnarounds in failing states and show that aid matters, but its effect depends on its kind (e.g., financial aid vs. technical assistance).
As emphasized by the theory proposed by Besley and Persson (2011), fragility is closely associated with conflict. Therefore, the literature that has evaluated the growth impact of conflict is also relevant. Examples within this stream include Collier and Hoeffler (1998, 2002), who search for the economic causes of conflict and then establish that Africa is indeed more vulnerable to it, because of its poverty; Blanton et al. (2001), who focus on the relationship between colonial domination and post-colonial ethnic conflict in Africa; and Bleaney and Dimico (2011), who distinguish between the correlates of the probability of onset of civil war and the probability of its continuation.
To conclude, the introduction of the broad concept of fragility, which reflects a complex combination of the dysfunctions that are typical of several African countries, has stimulated renewed interest for research on the deep roots of development in the region. At the same time, however, a clear impact of fragility on economic outcomes has proved hard to assess, partly because of the different definitions employed and probably also because of its potential endogeneity.
6. Slavery
Another recent research line that has extended the perspective in a promising direction has focused on the long term impact of the African slave trades not only for Africa, but also for the recipient countries. This line of research is closely related to the previous ones since it emphasizes the role of historical factors and evaluates how their influence on economic outcomes may run through institutions. Nunn (2008a) first looks at the long term effects of Africa’s slave trades on Africa itself. On the basis of shipping records and historical documents, he constructs measures of the number of slaves exported from each country in Africa in each century between 1400 and 1900 and finds a robust negative relationship between the number of slaves exported and current economic performance. Thus, the African countries that are the poorest today are the ones from which the most slaves were taken, even after accounting for the possibility of selection into the slave trades. Moreover, the paper indicates that the procurement of slaves results in subsequent state fragility and ethnic fractionalization.