Recent post-electoral violence in Kenya and massive irregularities in the 2007 elections in Nigeria seem to confirm the widespread view that African elections are principally flawed or have little meaning in the African context: democratisation, it is often concluded, has not taken root south of the Sahara. This view has to be put into perspective, and Staffan Lindberg's book is well positioned to correct this popular Afro-pessimistic cliché.

The book is well structured and, notwithstanding a thorough theoretical discussion in the first chapters, very accessible to the reader; it remains close to individual cases by providing rich anecdotal evidence. However, its main strength is its solid and systematic empirical foundation. More than 200 elections in Africa between 1989 and 2003 are systematically assessed vis-à-vis a large number of pertinent variables (such as turn-out, acceptance by losers, electoral violence, subsequent turnovers, opposition boycott, competitiveness etc.), which are then grouped into three dimensions describing the democratic quality of elections: participation, competition and legitimacy. This sound database is then used to demonstrate that since 1989 the quality of African elections has, on average, improved. Particularly when a regime manages to hold more than three elections, it is fairly likely that these will be free and fair, and that no breakdown will occur afterwards. Apparently, elections in Africa have a ‘self-reinforcing power’. Lindberg then moves to what is probably his most provocative hypothesis, ‘that improvement in democratic qualities tends to occur not only in conjunction with but as a causal effect of electoral activities’ (p. 117). An empirical analysis demonstrates that increases in democratic quality, measured independently from electoral activities, tend to be stronger directly after elections than in non-electoral periods. A multivariate analysis shows that, even when other important determinants of democracy are controlled for, the number of previous elections remains statistically significant, and explains about a quarter of the variance in democratic quality. Unlike many other books on African politics, this one has a happy ending: from the empirical evidence presented, Lindberg draws the conclusion that donors are right to focus on elections to promote democracy in Africa.

The ‘democratisation by elections’ thesis and the related policy recommendation, however, leave a somewhat uneasy feeling. In empirical terms, this is mainly because some explanatory variables for democratisation – particularly the role of the military and the quality of inter-communal relations (measured only by the relative number of ethnic groups), not to mention political culture – are left out of the multivariate analysis. Thus, the causal relationship between repeated elections and democratic quality may be spurious. Also, the results are probabilistic, and exceptions such as Zimbabwe or, as already mentioned, Kenya and Nigeria, exist. Even if we knew the exact weight of repeated elections in causing improved democracy on average, the findings tell us little about the individual country case, and how a promising strategy for democracy promotion should be designed.

These uncertainties and possible shortcomings, however, should not be overrated. It has to be reiterated that this book is based – unlike many other books on democracy and elections in Africa – on a sound empirical foundation (have a look at the Appendix!). Despite the widespread view that formal institutions have little meaning in Africa, it provides strong empirical support for the assumption that (formal) institutions do matter. Without doubt, thus study provides a strong stimulus for future research, and this is precisely what excellent scientific work is meant to be.