This article shows that a shift towards a more centralized school system can benefit countries that are characterized by poor levels of human capital and large regional disparities in education. In 1911, Italy moved from a fully decentralized primary school system towards centralization through the Daneo-Credaro Reform. The design of the Reform allows us to compare treated municipalities with those that retained school autonomy. Our quasi-experiment, based on propensity score matching (PSM), shows that centralization substantially increased the pace of human capital accumulation. Treated municipalities were characterized by a 0.43 percentage-point premium on the average annual growth of literacy between 1911 and 1921. We discuss some of the channels through which the new legislation affected primary schooling and literacy, with important implications for long-term economic growth.H uman capital formation is widely acknowledged as one of the main factors of economic growth, and this is supported by several well-established theoretical models. 1 Empirical evidence has confirmed that human skills and abilities are central to economic performance in the present day, as well as in the past. 2 As a result of this, one of the crucial themes investigated by the literature is the way that school systems are organized; that is, the extent to which schooling is managed and funded by local or central levels of government. Several pros and cons of * Authors' Affiliations: Gabriele Cappelli, University of Siena and University of Barcelona; Michelangelo Vasta, University of Siena. † We wish to thank one of the editors of this journal, Giovanni Federico, and three anonymous referees, for their constructive and very useful comments. The article also benefited from helpful feedback from

Italy's regions experienced different rates of human capital accumulation in the latenineteenth and early-twentieth centuries. Although Southern regions were very disadvantaged when the unification of the country took place in 1861, they caught up at a very slow pace -and a remarkable regional divide in education persisted until the Interwar period. While previous hypotheses have focused on the role played by fiscal capacity, this paper sheds new light on the effect that enfranchisement had on the growth of schooling. The presence of large regional disparities in local electoral franchise is confirmed by the data; however, the relationship between voting rights, the intensity of local direct taxation and municipal fiscal capacity is weak at best. Furthermore, if the impact of these factors is analysed separately through a number of econometric models, fiscal capacity stands out as the most significant determinant of education across Italy's provinces. Against recent hypotheses, these results show that the different distribution of political voice within Italy's municipalities did not determine the persistence of regional inequalities in schooling in the long-run: it was Italy's national education system, together with remarkable and pre-existent regional disparities, that slowed down the development of human capital in rural and Southern regions, with immense costs in terms of future prospects for economic growth and human development.

The present paper explores the role of public policy in the development of Italy's human capital in the late 19 th century and the Interwar period. It aims at understanding whether a system of decentralized primary education slowed down regional convergence in schooling. This work puts forward the hypothesis that, under such a system, the country was subject to a human capital trap-since poor and backward areas could not afford to invest a suitable amount of resources in schooling. Additionally, it investigates whether a more centralized system, introduced in 1911, loosened up the trap, fostered the accumulation of human capital and reduced the country's regional disparities. Original qualitative evidence and new data on schooling confirm the existence of such a trap, and underline the positive role of centralization in the Interwar period. The econometric model implemented strengthens these findings: poor regions could not improve the quality of education, which in turn would give rise to a vicious circle. Centralized primary education certainly fostered the development of Italy's schooling in the Interwar period. However, human capital regional disparities across the country persisted, a result that calls for further research on the topic.

Despite recent research about the link between social capital and economic growth, the degree to which social norms influenced the first phase of Italy’s regional economic divergence remains largely unexplored. A methodology based on a multifaceted definition of social capital, employing data about charity, mutual aid, and crime permits estimates of the differences in the strength of trust and cooperative norms across Italy’s provinces at ten-year intervals between 1871 and 1911. Further analysis of trust and cooperative norms via regression models of conditional convergence in industrial value added per capita shows that, although regional disparities in social capital were large during the late nineteenth century, they are not strongly correlated with industrial growth. Instead, the evidence indicates that human capital, innovation, and formal institutions were far more instrumental in determining the economic fortunes of Italy’s provinces before the World War I.