Conclusion

The development of Paul's research over the last 50 years reflects some trends in economics as a whole (albeit there are enclaves that resist these trends). Neoclassical economic theory is no longer ubiquitous, but remains as a tool to be used in specific circumstances, and there is an increasing awareness of the importance of engaging with (and sometimes provoking) academics from other social sciences.

This is not a point of view that could easily have been predicted from Paul’s first work on customs unions. It reflects the fact that much of the broadening of academic economics over the last 30 years can be traced through the work of individual economists: it is an age effect at least as much as a cohort effect. The new eclecticism is seen most clearly in Paul’s work with Pedro Vincente on intimidation and corruption during African election campaigns. Their theoretical work to explain the conditions under which intimidation and cor­ruption appear has a strongly neoclassical flavour, with rational electoral can­didates making strategic decisions about the employment of their resources on the basis of the expected costs and benefits of different types of electoral manipulation (see Collier and Vincente 2012). The applied work which fol­lows, and which is motivated by a search for practical policies to reduce the incidence of electoral manipulation, has a rather different flavour, drawing on the experimental methods of behavioural economics to test the effectiveness of different policy interventions in Nigeria and Mozambique (see Collier and Vincente 2014; Aker et al. 2017).^[220]

Notwithstanding this eclecticism, Paul’s current work is driven by the same enthusiasm that motivated his undergraduate essay on Malawi, even if the undergraduate’s unbridled optimism about the ability of economics to solve the problems of the post-colonial world has disappeared. Last, but not least, several generations of graduate students are indebted to him for his thought­ful support and encouragement. The family tree of his PhD children, grand­children, great-grandchildren and great-great-grandchildren has not been documented, but it is surely extensive.