We frequently hear calls for the leaders of struggling societies to “tackle” corruption. Aid donors back project after project aiming to improve institutions and foster better “governance” on two- or three-year time lines. Academic researchers reinforce that short-term mindset by issuing reams of cross-sectional research regressing a recent year’s corruption index scores on some set of equally recent statistical indicators of national characteristics. But what if corruption—however we may define or measure it—has far deeper roots, and is a function of several intertwined developments over much longer spans of time?

In his new book, Eric Uslaner offers a strong argument that such is precisely the case. For a substantial group of countries, he shows that one of the strongest predictors of a country’s level of corruption in 2010, as measured by the widely used Transparency International Corruption Perception Index (TI-CPI), is its level of mass educational attainment in 1870—nearly a century and a half ago. That statistical relationship is stronger than similar calculations using later years’ educational data, and is robust with respect to a range of competing hypotheses, such as productive factor endowments, experiences with colonialism, and so forth. The effects of education upon corruption, Uslaner shows us, are “sticky”: Countries with high levels of educational attainment, and more effective at checking corruption a century ago, tend to be found at the higher end of such rankings in the current era, while those faring less well on these measures in, say, 1900 still struggle today.

After a Preface, The Historical Roots of Corruption begins with a chapter laying out the theoretical framework, which is followed by a chapter presenting the quantitative evidence. There follow three chapters that develop important regional variations on the overall thesis: “Education in Developed Europe” (Chap. 3, pp. 56–81); “Education Beyond Developed Europe” (Chap. 4, pp. 82–100), which takes up Latin America, Africa, and several individual non-Western countries; and “The United States and Other ‘New’ Anglo-American Countries” (Chap. 5, pp. 101–30). A concluding chapter, “Is Path Dependence Forever?” (Chap. 6, pp. 131–63), answers its own question with an overall assessment of “no,” and then moves on to analyze several different processes of change, wisely reminding us to look beyond the data in understanding and responding to corruption in diverse settings. There are also a brief data Appendix, References, and an Index.

What are the causal linkages behind Uslaner’s provocative findings and arguments? They are numerous and reflect the “embeddedness” of corruption that most analyses mention but usually do not examine in any depth. Mass education is associated with wider economic and political opportunities, giving citizens means and incentives to resist corrupt ways of doing things. It affects and reflects patterns of inequality, and is an aspect of greater state capacity: A higher-capacity state will, all else being equal, be more able to provide wider and more persistent educational opportunities. Conversely, enhancing mass educational attainment is one way to build greater state capacity over time. Extensively corrupt societies, by contrast, tend to invest significantly less in higher education. The overall significance of education for corruption control is nested, for Uslaner and colleagues such as Bo Rothstein, within broader findings about the long-term value of universal social welfare programs for societies, their inequalities, and their quality of government.

A skeptic might challenge the notion of assigning so much weight to a single causal factor over such a long time span but, as suggested here, Uslaner does point out a range of specific causal connections underlying the apparent anticorruption effects of mass education. More important criticisms might revolve around treating corruption as a single generic problem, rather than a phenomenon occurring in significantly different forms with distinctive underlying causes and extended effects.

At a methodological level, the TI-CPI faces a great many criticisms, including that it records perceptions of corruption, rather than directly measuring the phenomenon itself according to a specified set of criteria (to be fair, few if any of its competitor indices go much further toward direct measurement). That measurement question is important, however, particularly for embedded causal relationships over long time spans; arguably, TI-CPI scores are influenced, directly or indirectly, by a wide range of perceptions (affluence, state capacity and efficacy, regime durability and stability, and the like) that are themselves shaped by, and contribute to, long-term education levels. If that is the case, then the long-term statistical relationships might be said to be between historical levels of education and their own downstream effects, with corruption being embedded in, and difficult to distinguish from, a great many such consequences. To minimize that risk, however, Uslaner makes use of the superb V-DEM time-series scores on corruption and numerous other aspects of politics and government, and also incorporates data on news reports of corruption in the American states, to minimize the risk (in the U.S. case at least) of treating corruption as a single monolithic national characteristic.

Methodological issues are important but should not obscure the valuable core argument of the book: Corruption is not just a law-enforcement or administrative-process problem at any one given time, nor is it solely a function of the other sorts of political and economic influences one might approximate using national-level indicators. It is, rather, a deeply embedded and long-term aspect of relationships between state and society; politics and the economy; opportunities and constraints upon the pursuit of self-interest; and the administrative, economic, and political capacity of the state to establish and maintain an effective and legitimate system of public order.

Mass education is arguably implicated in all of those relationships and tensions, both as an influence and as an outcome. Looked at that way, it is not surprising that long-term trends in education might have major and persistent effects upon corruption—and upon much else. Therein lies the most important anti-corruption lesson of The Historical Roots of Corruption: not (of course) to somehow go back to 1870 and build schools, but rather to look at education, corruption control, and many other policy and political processes as ways to build long-term demand, and support, for open, fair, and honest government in society at large. For drawing attention to these large-scale and long-term phenomena, Uslaner is to be congratulated.