International election monitoring has become a near-universal practice in young democracies and hybrid regimes, but we still know little about the behavior of monitors and whether they matter for democratic development. Summarizing and extending her past work, Judith G. Kelley presents a wide-ranging and theoretically rich account of election monitoring. Whereas Susan Hyde's The Pseudo-Democrat's Dilemma (2011) concentrates on the strategic interaction of monitors, autocrats, and domestic audiences, Kelley's Monitoring Democracy focuses more closely on the characteristics of the monitors and how their assessments are derived. Although the author ultimately expresses optimism, the book presents a complex picture of monitors' biases and compromises.

Kelley addresses three main questions. First, why has election monitoring become so widespread? The answer is tackled from two perspectives: why international organizations have chosen to monitor elections and why governments grant them access (Chap. 2). She argues that the supply of monitors was driven by the international community's growing emphasis on elections as a component of human rights and the desire of Western actors to shepherd states transitioning to democracy at the end of the Cold War. Though rare before 1990, international monitoring is now present in roughly three in four elections in nonconsolidated democracies, effectively becoming an international norm expected of legitimate governments.

Why do countries, including many autocracies, invite election monitors? Besides the international rewards, Kelley argues that allowing monitors provides a valuable signal to domestic audiences. For weak democracies, an international stamp of approval encourages political losers to accept the result, as was criticial in Mexico in 2006. For autocracies, monitoring has become sufficiently common that “the stigma associated with not inviting monitors motivated even cheating governments to invite monitors to avoid an automatic stamp of illegitimacy” (p. 31). However, a number of regimes—such as Egypt, Singapore, and Russia—have spurned international monitors without any discernible penalties. In general, Kelley could have provided more evidence regarding the punishments that are associated with avoiding monitors or cheating in front of them.

Second, what influences monitors' assessments of elections? There are several components to this, including differences between monitoring organizations (Chap. 3) and the relationship between specific types of violations and overall quality ratings (Chap. 4). As Kelley explains, there is surprising variation in the leniency of different monitoring organizations, with some comprising a “shadow market” of observers willing to provide autocrats the illusion of international credibility. Later, she makes the important point that only the tougher monitors appear to have a positive effect on election quality

Monitors' assessments are, unfortunately, reported in the media as a simple thumbs up or thumbs down, but their election reports delve into much greater specificity. To analyze these reports, Kelley employs a thorough original coding of the monitors' data records, a previously untapped source of information. She finds that election-day cheating and legal problems are especially likely to lead to a negative overall assessment, whereas preelection irregularities are more often overlooked. Oddly, Kelley finds that preelectoral violence is positively related to overall quality ratings, which, she speculates, is due to monitors' fear of inspiring more violence in protest if they condemn the election.

An intriguing question for future research concerns geopolitical interests and how they factor into monitoring behavior. Kelley offers examples to suggest that American interests and alliances with target states blunted the pressure to allow monitors (as in Egypt) or softened the tone of election assessments (as in Bosnia in 1996). She also finds that foreign aid predicts more favorable reports despite being unrelated to actual election quality, suggesting that monitors are more lenient toward internationally favored states (p. 70). A more systematic analysis could add to the study of both autocratic durability and the interaction of norms and realist interests among international actors.

Third, does international monitoring improve the quality of elections? In particular, does the presence of monitors in a given election improve its conduct (Chap. 7)? Two major problems arise in this analysis. The first is the nonrandom determination of the elections that are monitored, a product of strategic selection by both observers and observed. Kelley uses matching in this analysis, but correctly acknowledges (at least in the appendix) that matching has no advantage in dealing with endogeneity relative to regression. Other researchers have reached firmer causal conclusions using within-election natural experiments. For instance, articles by Susan Hyde and Erik Herron exploit the random assignment of international and domestic monitoring to specific precincts. Generally, they find that the incumbent underperformed in monitored polling places, but acknowledge that this may have simply shifted fraud elsewhere. A potential way to address the endogeneity across elections would be to exploit some randomness in monitors' choices of elections, perhaps from funding availability or individual idiosyncracies that can be divined through fieldwork.

The second problem in testing the effect of monitoring is that one needs a measure of election quality whether or not monitors are present. Kelley uses the U.S. State Department's rating of each election, but does not resolve the considerable difficulty that it may partly base its ratings on the observers' assessments (and the fact that observers were allowed). In fact, it would be strange if it did not. Nevertheless, it is an important finding that monitored elections in multiparty autocracies are about 8% more likely to be deemed acceptable by the State Department. More convincing is Kelley's analysis of incumbent turnover, which is about 25% more likely when quality monitors are present, although it might be that weaker leaders are more vulnerable to pressure to invite monitors. A number of authors have argued that monitoring leads autocrats to simply substitute preelectoral manipulation for election-day fraud, but Kelley does not find evidence for this argument (Chap. 5).

Finally, the author analyzes whether monitors' recommendations improve elections over the long run (Chap. 8), on the basis of case studies of 15 countries. Ultimately, she is surprisingly sanguine about the effectiveness of international monitors, noting that about half display no improvement and only one-quarter show “significant improvements” (p. 152). This limited influence is perhaps not surprising once we recall that the median monitoring mission contains about 22 people, stays for only two weeks, and has no direct power whatsoever. Although most countries adopt at least some recommendations, these are often the least threatening to regime interests. Further, the three states she identifies as having made the most progress in reform (Panama, Mexico, and South Africa) faced intense international political pressures, calling into question the explanatory power of the monitors. However, she theorizes that monitoring will be most effective when combined with strong state capacity, domestic pressure, and a desire for integration with the West.

Kelley has produced a fine piece of scholarship that should be required reading for scholars interested in democracy promotion, as well as practitioners. The analysis is careful, broad, and admirably conversant in the details of specific countries and elections, although it struggles at times to reach firm causal conclusions. One of her greatest contributions is the associated data set, which is publicly available and codes for both the characteristics of the monitoring missions and their detailed evaluations. Hence, interested researchers are amply supplied with the theoretical and empirical tools to build on Kelley's work.