Matthew Cleary and Susan Stokes begin their book by pointing out the recent popularity of neo-Tocquevillian theories that link the quality of democracy to widespread civic association and social capital. As the authors acknowledge, this popularity is not surprising in light of the intuitive appeal of the idea that an engaged populace full of citizens who trust their neighbors and participate in civic affairs leads to better democratic performance. In this book, the authors take issue with the literature that links a densely connected, trusting citizenry to a well-functioning democracy. Instead of widespread trust, they argue that an “engaged skepticism” is in fact the key to a healthy democracy. According to Cleary and Stokes, a society marked by engaged skepticism is one in which citizens are skeptical of the motives of individual politicians, but in which the former nonetheless remain interested and active in political matters and trust institutions to hold politicians accountable.

Empirically, the authors test their major hypothesis using subnational data from two Latin American countries—Mexico and Argentina. Although they begin their empirical work by examining the quality of democracy in all of the states/provinces in each country, they ultimately concentrate on eight subnational units: Mar del Plata, Buenos Aires, Córdoba, and Misiones in Argentina, and Baja California, Chihuahua, Michoacán, and Puebla in Mexico. First, they order the four units in each country according to their respective quality of democracy (in Chapters 2 and 3). They then use detailed survey data to examine whether citizen attitudes and behavior vary in the expected manner across the four units in each country (in Chapters 4 and 5), before concluding in Chapter 6.

The book's greatest conceptual contribution is to clarify the idea of trust and its likely relationship to the quality of government. In the first chapter, the authors separate personal trust—which they consider a strong form of trust that relies on some “inherent quality” of the person being trusted—from a weaker form of trust that relies instead on the risk of “third-party punishment,” such as that meted out by elections (p. 13). They offer the intuitive—and compelling—argument that weak trust is in fact crucial for a well-functioning democracy, while strong trust—especially when placed in individual politicians—is a likely sign of low quality democracy.

Drawing on data from a few subnational units in two middle-income countries has the advantage of allowing the authors to rely on detailed quantitative and qualitative data for their analysis. Nonetheless, their first task—ordering the regions by their quality of democracy—suggests some of the difficulties with this research design. In Mexico, they focus on each state's degree of democratization, ordering the four states largely on the basis of data on alternation in power and the historical ruling party's margins of victory or defeat in elections in recent decades. In Argentina, where democratization took place simultaneously throughout the country in 1983, the authors instead focus on the quality of democracy, assessing this using information on municipal fiscal behavior, the volatility of support for the Peronist party, and other variables. Though the measures used in each country are reasonable on their own, they create some strange juxtapositions. Most notably, the authors consider Mar del Plata to be the most democratic of the Argentine regions they examine, in spite of the fact that a single party held power over the municipal government since the 1983 return to democracy (p. 71). In Mexico, by contrast, they cite the lack of party alternation in Puebla as evidence of that state's low-quality democracy (p. 49). Although the parties that held power in each region differ, the lack of clear criteria for ranking the four subnational units across the two countries is likely to make more difficult the replication of these rankings in other countries.

In Chapters 4 and 5, Cleary and Stokes rely on original survey data to examine their major hypotheses. The surveys themselves are extremely rich, and some of the questions they ask allow them to poke holes in the microfoundations of the neo-Tocquevillian argument. They find evidence, for example, that individuals in clientelistic relationships are more likely to trust individual politicians (p. 104), undercutting the argument that high levels of trust are a hallmark of high-quality democracy. However, they focus much of their efforts on explicating not individual patterns of behavior but, rather, the differences in aggregate responses among the four units in each country. Although this approach allows them to have large samples in each of the units, the inclusion of only four units in each country reduces their ability to make claims about the relationship between a region's level of democracy and patterns of trust in politicians and government.

So, for example, the authors find that, consistent with their expectations, Mexicans living in Baja California, the most democratic of their regions of interest, are more likely than their counterparts in each of the other three regions to believe that when governments perform efficiently, it is because they are monitored by institutions (p. 92). However, as Cleary and Stokes acknowledge (p. 91; see also p. 99 for a similar point), the proportion of residents who hold this belief does not consistently increase as the level of democracy rises across these four units, as their theory would predict. Even when the pattern of responses conforms to their expectations and multivariate regressions confirm that these cross-regional differences are significant, the inclusion of only four units in each country means that it is not possible to measure alternate explanatory variables—like poverty or economic development—at the regional level and include them in regression analyses. Although the authors are able to include myriad controls at the individual and municipal level, the reader would be more confident in their conclusions if they had data available on a larger number of units in each country.

In spite of the limitations to the data, Cleary and Stokes succeed in making a compelling argument that strips away the romance of narratives that argue that all good things—trust and democratic quality—go together. Although more data are needed, they have charted out a model of how we might go about determining the relationship among participation, different types of trust, and the quality of democracy. Democracy and the Culture of Skepticism will help set the agenda for continued research in this field.