This article describes and explains a previously overlooked empirical pattern in state revenue collection. As late as 1913, central governments in the West collected similar levels of per capita revenue as the rest of the world, despite ruling richer societies and experiencing a long history of fiscal innovation. Western revenue levels permanently diverged only in the following half-century. We identify the twentieth-century great revenue divergence by constructing a new panel data set of central government revenue with broad spatial and temporal coverage. To explain the pattern, we argue that sustainably high levels of revenue extraction require societal demand for an activist state, and a supply of effective bureaucratic institutions. Neither factor in isolation is sufficient. We formalize this insight in a game-theoretic model. The government can choose among low-effort, legibility-intensive, and crony-favoring strategies for raising revenues. Empirically, our theory accounts for low revenue intake in periods of low demand (the nineteenth-century West) or low bureaucratic capacity (twentieth-century former colonies), and for eventual revenue spikes in the West.

本文描述并解释了以前被忽视的州税收征管实证模式。迟至 1913 年，西方中央政府的人均收入水平与世界其他地区相似，尽管它们统治着更富裕的社会并且经历了长期的财政创新。西方收入水平仅在接下来的半个世纪内出现永久性差异。我们通过构建一个具有广泛时空覆盖的中央政府收入面板数据集来识别 20 世纪的巨大收入差异。为了解释这种模式，我们认为，可持续的高收入提取水平需要社会对激进国家的需求，以及有效官僚机构的供给。孤立的任何一个因素都是不够的。我们在博弈论模型中将这种见解形式化。政府可以选择省力的、易读性强的和偏爱裙带关系的策略来增加收入。从经验上看，我们的理论解释了低需求时期（19 世纪的西方）或低官僚能力（20 世纪的前殖民地）时期的低收入，以及西方最终的收入激增。