Democratization in Africa has forced a rethinking of general theories of
democratic transition. Geddes suggests that one of the few things we have learned from
twenty years of theorizing about democratization is that it is associated with middle
levels of income. This claim was first made by Lipset in 1959, and has been refined by
scholars and consistently supported by empirical data into the 1990s (Dahl 1971, Bollen
and Jackman 1985, Huntington 1991, Callaghy 1994, Ottaway 1997, Geddes 1999,
Lawson 1999). Yet, in Joseph's assessment, the 'third wave' demonstrated little respect
for income distinctions (1997: 9). Indeed, now that the dust has settled, the most
successful new democracies in Africa appear to be among the poorest (Benin, Mali,
Niger). African democratization also challenges the necessity of other commonly
accepted "prerequisites" of democracy, such as a strong civil society, and an urbanized,
well-educated population. Similarly, transitions from African patrimonial regimes in the
1990s followed a different course from that identified in general theories of transition
(Bratton and van de Walle 1997, Solt 2001, O'Donnell and Schmitter 1986, Przeworski
1991). Transitions appear to be shaped as much by where a country starts as by where it
is going. Van de Walle (2003) takes the analysis one step further, noting that emerging
party systems in Africa continue to reflect the region's patrimonial inheritance, retaining
strong elements of presidentialism and clientelism (supported by the appropriation of
office).
The broad challenge that African experience presents to commonly held
understandings of democratization suggests not that existing theories are misspecified,
but rather that they rest upon an unidentified antecedent condition. Though states
undergoing democratization have varied on any number of characteristics, all (or at least
the vast majority) outside of Africa have had a bureaucratic core. Sub-Saharan African
2
Convention