The third-wave transitions to democracy in Latin America were met with optimism that governments in the region would become more accountable to their citizens, as well as trepidation that these transitions might collapse in the face of military coups. In most of these countries, democracy's long endurance has allayed fears of the latter. Questions about the quality of these now not-so-young democracies, however, remain extremely relevant. Scholars, citizens, and other observers increasingly recognize that it is no straightforward task to make democracy work well, and that there may be significant variation in government performance even within a single country. In spite of this growing concern with the quality of subnational democracy, our understanding of why, within a democracy, some governments perform better than others remains incomplete.

In this book, I use the case of clientelism both to illustrate the breadth of variation in government quality and to test a theory of what explains improvements in local government performance. Clientelism undermines ties of accountability between citizens and politicians. It narrows the issue space about which clients make political decisions to the question of access to particularistic goods; in doing so, it insulates politicians from the need to be responsive to the full potential range of citizen preferences. Although the costs of clientelism to accountability are widely recognized, those costs may be unevenly distributed across a country's territory.

Chapter 2 illustrated how clientelism can work as a vote-getting tool on a mass scale. Local incumbents who so choose can claim credit for the distribution of valued benefits, and citizens, worried about future access to goods crucial to their well-being, will be willing to exchange their political support for these benefits. However, though there are many strategies Argentine mayors can use to make clientelism work, not all local incumbents rely on clientelism in social policy. As the anecdotes that began this book suggest, clientelism in Argentina is widespread, but it is not ubiquitous. Officials in a minority of towns and cities explicitly deny personal responsibility for government programs and opt not to make political ties or behavior a precondition for receiving government benefits. In this chapter, I seek to understand these exceptions: within a single national political environment, why do some local incumbents opt out of clientelism?

My answer to this question relies on what I call the “audience costs” of clientelism. Although clientelism is typically targeted at the poor, there are other, nonpoor citizens in the political environment. The reactions of these nonclients to the practice of clientelism will affect a politician's incentives to rely on the practice. I argue in this chapter that the non-poor are likely to punish clientelist politicians in the voting booth.