Few questions in comparative political economy have proved more durable, or more resistant to settlement, than whether democratic government promotes economic growth. The intuition that it should is old and appealing: governments accountable to voters, the argument runs, must respect property, curb the predation of officials, and supply the public goods - courts, roads, schools - on which prosperity depends. An autocrat, by contrast, faces no such discipline and may plunder at will. Yet the empirical record has stubbornly declined to cooperate. For every democracy that has grown rich under stable institutions, one can adduce an authoritarian state that has industrialized at breakneck speed, and a democracy that has stagnated amid factional paralysis. Confronted with this untidiness, an influential line of argument reverses the causal arrow. On this "modernization" account, it is not democracy that produces wealth but wealth that produces - or at least sustains - democracy. As societies grow richer, the reasoning goes, they acquire a literate middle class, dense networks of voluntary association, and a citizenry with the leisure and confidence to demand a voice; democracy is less the engine of development than its eventual passenger. The correlation between prosperity and representative government, so often cited as evidence that the latter causes the former, is on this view better read the other way around, or as the joint product of some deeper cause. A third position, now ascendant, holds that both formulations mistake the object of inquiry. What matters for growth, its proponents contend, is not the presence of elections but the character of a country's institutions - whether they are, in a common phrasing, "inclusive" rather than "extractive." Inclusive institutions secure broad property rights and open economic opportunity to the many; extractive ones concentrate both in the hands of a few. Democracies may or may not build inclusive institutions, and some autocracies have approximated them for a time. The variable that does the causal work, on this account, is institutional, and the regime type is at most a rough and unreliable proxy for it. This institutionalist turn is a genuine advance, but it is not without difficulties of its own. To say that inclusive institutions cause growth invites the question of what causes inclusive institutions - and the most common answers, which point to historical contingencies such as patterns of colonial settlement or the timing of enfranchisement, threaten to push the explanation so far back in time that it loses its grip on any policy a present-day government might adopt. There is, moreover, a suspicion of circularity: if institutions are judged inclusive partly by the prosperity they are meant to explain, the thesis risks defining its own confirmation. The most careful proponents have labored to break this circle by measuring institutional quality independently of outcomes, with uneven success. What can be said is that the debate has moved, however haltingly, from the blunt question of whether democracies grow faster to the more tractable one of which specific arrangements - and under what conditions - reward productive effort rather than punishing it.