This book is an original, important, and in many ways impressive study that will make a contribution to both electoral and Africanist scholarship. The author's central argument is that repetitive elections (three or more) increase the democratic qualities of regimes and broaden and deepen civil liberties in societies. This argument is carefully placed within democratic theory and rigorously tested by a dense and generally sophisticated empirical analysis, although I have criticisms regarding the tightness of the theory and two variables used in the empirical analysis.

Theoretically, democracy is a regime, a set of rules that is both constitutive and regulatory: “Within regimes, actors have room to maneuver, and their subjective interests, goals and calculations matter for outcomes, but within limits not of their own choosing” (p. 7). Elections are partial regimes, which, Steffan Lindberg argues, are likely to produce complete democratic regimes. Democracy is defined procedurally as self-government comprised of participation, competition, and legitimacy in the eyes of the people. The author wisely chooses to define democracy as a matter of degree, rather than to dichotomize between democracies and nondemocracies. But he surreptitiously introduces dichotomous analysis in the form of a cutting point between two categories on the continuous variable of free and fair elections that contain 91% of the cases (37, 48). These categories are distinguished from each other only by whether irregularities affected the outcome, which is a separate variable from the degree of irregularities and is strongly affected by the closeness of the outcome. Participation is measured in terms of turnout, opposition participation, and the questionable criterion of whether former authoritarian rulers or their close associates are among the presidential candidates. This criterion makes the questionable assumptions that democratic values are essential for democratic behavior, that former rulers' values do not change, and that current rulers not involved in previous regimes do not have authoritarian tendencies. Competition is measured in terms of the winner's percentage of the vote, the largest two parties' percentages of the seats, and turnover of power. Finally, legitimacy is measured in terms of losers' acceptance of results, peacefulness of elections, and electoral system survival.

The problems with the author's use of free and fair as a dichotomous variable and his excessive emphasis on the presence of former authoritarian politicians can be effectively illustrated by his classification of the 1991, 1996, and 2001 elections in Zambia. There were almost as many irregularities in 1991 as in 1996 and there were fewer in 2001 than in either of the previous elections. The election in 1991 is classified as free and fair because the one-sided results meant that irregularities by the losing ruling party did not prevent its defeat. The election in 1996 is classified as not free and fair because manipulations by the ruling party led to a boycott by the former ruling party, even though that party's inclusion would have led to the same presidential result and only slightly different parliamentary results. The 2001 election is classified as not free and fair, though all parties participated, because the very close results may have allowed the ruling party to steal the election through relatively minor manipulations. Certainly, 1996 was the least democratic of these elections. However, this was because the apparently democratic victor in the 1991 presidential election had turned authoritarian in the face of increased opposition, which originated mainly from within his own party, rather than because members of the former one-party regime participated.

Additional theory presented later in the book involves hypotheses relating repetitive elections to increased democratic qualities of society. These hypotheses (pp. 111–16), guided too loosely by a rational choice institutionalist framework, suggest that repetitive elections may lead to 1) broader societal participation by citizen-voters; 2) greater acceptance of democratic values; 3) realization by initially nondemocratic actors that democratic expectations predominate among those with whom they interact, and that it is therefore to their advantage to conform to those expectations; 4) the strengthening of pro-democracy civil society groups; 5) stimulation of postelection pro-democracy activities by individuals within then; 6) greater incentives for judges, police, and the military to reinforce democratic norms; and 7) stimulation of more pro-democracy content in the media. The author suggests that these potential effects probably interact with one another in positive ways.

Empirically, Lindberg first shows that the annual frequency of elections between 1990 and 2003 fluctuated but increased slightly on average over the period, while the percentage that were free and fair and had democratic characteristics fluctuated but remained constant on average. Second, he shows, through both the analysis of all elections and a time series panel-group comparison that groups together countries that have had one, two, three, or four or more elections, that—although there is a variety of specific findings—the democratic qualities of elections improve on average with each election. Finally, the effect of the number of elections on the democratic characteristics of society (operationalized by the Freedom House civil liberties scale) is analyzed in five different tests. The dichotomous free and fair variable is wisely omitted from this analysis. All tests support the hypothesis that a greater number of elections—whether free and fair or not—increases civil liberties. The final chapter offers interesting speculations about the broader implications of the book's findings and points out additional research that needs to be done to answer some of the questions raised in it. One interesting speculation with which this reviewer agrees is that the effects of culture on the transfer of electoral institutions has been less extensive than culturalists and Afro-pessimists have claimed.

Further research is clearly needed to demonstrate how strongly the positive effects of even more repetitive elections under transferred institutions will lead to further democratization. The author does not deal with this issue specifically, but his central argument probably implies that these positive effects will continue. My criticisms are not meant to detract unduly from the real contribution that this book makes. It will be cited positively in studies of African elections for many years.