Fragile States
Paul and Anke's work on civil wars was the starting point for a broader investigation into the political economy of different causes of persistent underdevelopment. This work, informed by Paul's former undergraduate student, Tim Besley (see, for example, Besley and Persson 2011), aims to account for the different underlying causes of State fragility.
First, there is a connection between the literature on civil wars and the literature on the “natural resource curse”.[218] Compared with other types of economic activity, mineral production tends to be highly geographically concentrated, and intensive in physical rather than human capital. Therefore, minerals are highly lootable, creating an incentive not only for corruption and criminality, but also for violent conflict: war is a continuation of rent-seeking by other means. Evidence for the association between the incidence of civil war and mineral dependence appears in studies by Paul and Anke and by James Fearon (see Collier and Hoeffler 2005; Fearon 2005). Such rent-seeking behaviour (along with Dutch Disease, as outlined in Controlled Open Economies') leads to disinvestment in other sectors of the economy, and an even greater reliance on minerals: another vicious circle. Secondly, Paul's work with Steve O'Connell has highlighted the particular challenges facing landlocked resource-poor countries. These countries' international trade connections depend not just on the ability of their maritime neighbours to maintain adequate port facilities with good regional road and rail connections, but also on the quality of governance in these neighbours, and the absence of roadblocks created by protectionist central governments or by corrupt local officials. The misery caused by a country's poor governance can easily spread beyond its borders (see Collier and O'Connell 2007; Collier 2007a).At this point, Paul's research might seem to paint a rather depressing picture: countries that are landlocked, or that have a history of war, or that rely heavily on natural resources, can become persistently impoverished and have persistently poor government. Under these conditions, on average, foreign aid has had very little effect. Paul's work with Lisa Chauvet does something to alleviate this depression by investigating the variation in foreign aid effectiveness around this mean. Their most striking result is that the effectiveness of World Bank aid is highly dependent on the amount of time devoted by Bank staff to preparation and supervision. When more time is given, the project is much more likely to be successful, and this effect is significantly larger in the aftermath of civil conflict. There is also some evidence that aid to build infrastructure is more likely to succeed than is aid for education. Investment in physical capital is a better bet than investment in human capital, although the long-run returns to successful human capital investment are likely to be especially high, if it creates the capacity for countries to provide efficient supervision of their own. A balanced donor investment portfolio might reasonably include some of both (see, for example, Chauvet et al. 2010).
In many of Paul's contributions over the years, the angel is in the detail. Vinerian customs union theory is a fine thing, but it will lead to mistaken policy conclusions if its simplistic aggregation of commodities is taken at face value. The Harris-Todaro model is equally fine. However, it too will lead policymakers astray if the aggregation implicit in its characterisation of labour markets is not relaxed. Even more importantly, the finding that aid is relatively ineffective in poor policy environments, on average, is not to be taken as a counsel of despair, but rather as a reason to disaggregate aid in poor policy environments, in order to discover the correlates of variation around the mean.