اجبار نیروی کار و انباشت سرمایه های انسانی

This paper examines the effect of labor coercion on human capital accumulation. We use micro data from Puerto Rico, where unskilled laborers were forced to work for landowners during 1849–1874. Using variation in municipality-level suitability for coffee cultivation and international coffee prices, we estimate the response of schooling to exogenous increases in relative demand for unskilled labor in regimes with and without forced labor. During the coercive regime, increased coffee prices had no effect on individuals' literacy rates in coffee growing regions. Following the abolition of forced labor in 1874, similar changes in coffee prices reduced literacy rates by 12%, consistent with a diminished skill premium in the free labor market regime relative to the coercive period.

Throughout history, many types of labor arrangement have involved the use of coercion — the threat or use of force to compel workers to enter into an employment relationship. Slavery was a common way of organizing labor in plantation economies throughout the Americas and has long been the focus of economic historians.2 Other forms of forced labor, such as debt peonage and the state-sanctioned forced recruitment of laborers, also played an important role in the hacienda system found in Latin America during the colonial and post-colonial periods.3 Episodes of coercion are not confined to the historical record: they have persisted in many forms across both industrialized and developing countries throughout the twentieth century (Andrees and Belser, 2009). While the conditions determining the degree and precise form of coercion have been studied extensively, in particular the role of international trade, less is known about the consequences of coercion for economic development.4 Scholars have argued that coercive labor institutions can lead to socially inefficient outcomes as they involve a costly transfer of resources from workers (e.g. Engerman and Sokoloff, 1997, Coatsworth, 1999, Conning, 2004 and Acemoglu and Wolitzky, 2010). Consistent with these arguments, empirical studies have found a negative relationship between the historical prevalence of coercion and measures of current economic development such as income and levels of education.5 In contrast, less empirical analysis has been devoted to identifying the mechanisms that may lead to these differences in human capital accumulation. It is difficult to overstate the importance of this channel given that it is arguably one of the most prominent determinants of development in the modern period (e.g. Galor, 2005). This paper provides empirical evidence to answer the question of what are the consequences of labor coercion for individuals' education decisions in the context of increasing commodity export trade. We study legislation by the Spanish government in Puerto Rico that forced free unskilled workers – jornaleros – to seek employment on legally titled farms during 1849–1874, followed by abolition of this legislation after 1874. Using micro data on cohorts of individuals across municipalities, we exploit variation in the suitability of coffee cultivation across municipalities and in coffee prices over time to identify exogenous changes in unskilled labor demand following the insight of Domar (1970) that the incentive to coerce is increasing in the demand for unskilled laborers. We then examine whether and to what extent the availability of legally sanctioned coercion by landowners against unskilled workers affected how literacy patterns responded to these changes. Consequently, identification of the effect of coercion on human capital accumulation comes from a ‘triple difference’ strategy that explicitly takes into account the possibility that regions possessing high demand for unskilled labor might have different levels of human capital accumulation regardless of whether labor market coercion exists. We start by providing suggestive evidence that, during the coercive period, coffee-region local governments allocated more resources toward the enforcement of coercive labor measures. We then show that literacy rates were significantly higher among cohorts of schooling-age in coffee growing regions in response to an increase in coffee prices during the coercive regime relative to comparable cohorts in the non-coercive regime. Specifically, during the coercive period, higher coffee prices had no effect on individuals' literacy rates in coffee growing regions, whereas, after the abolition of coercion in 1874, similar changes in coffee prices reduced literacy rates by 12%, consistent with a diminished skill premium. These results suggest that the abolition of forced labor had important consequences, eliminating landowners' ability to extract unskilled workers' income but also reducing laborers' incentive to accumulate human capital, as the abolition of coercion of unskilled workers raised the relative wages of unskilled workers by lowering the skill premium. While our finding of increased human capital accumulation coincident with coercion might seem perverse and surprising, we believe it is intuitive and can be easily rationalized by standard economic theory. Our results are consistent with models where, in the context of a free labor market, positive coffee price changes lead to a decrease in the skill premium and therefore a reduction in human capital investments. However, in a coercive regime, unskilled wages are artificially suppressed and therefore this effect is mitigated as workers avoid coercion and invest in human capital. After these coercive institutions are removed, similar changes in coffee prices reduce literacy rates, consistent with a diminished skill premium. This does not imply that workers are better off as a result of coercion. In fact, unskilled and skilled workers' welfare, as measured by lifetime income, is lower under the coercive labor regime than in a free labor market. Our results are fully consistent with the negative relationship between the prevalence of historical coercive institutions and long-run levels of human capital accumulation but suggest an alternate interpretation of this correlation.6 As argued by Domar (1970) and Acemoglu and Wolitzky (2010), labor coercion is more likely to exist in regions that experience high unskilled labor demand. To the extent that certain regions possess fixed or persistent characteristics (e.g. geography) that encourage specialization in the production of unskilled labor-intensive goods, these regions are more prone to historical episodes of coercion, and in the cross-section should also experience a higher demand for unskilled labor. Consequently, these regions will also possess lower demand for education, despite the possibility that, under a forced labor regime, this negative relationship may be partially mitigated due to a higher skill premium. The research design and the richness of the data allow us to distinguish our explanation for educational outcomes from numerous competing explanations that we assess in the robustness section. We do so by exploiting variation in the suitability for coffee cultivation across otherwise similar municipalities. First, we show that the observed patterns of literacy in response to changes in unskilled labor demand are not due to a change in the availability of schooling, consistent with our demand-side hypothesis. Second, from a purely technological perspective, there are disincentives to establish coffee plantation economies, as there are constant (or even decreasing) returns to scale in coffee cultivation.7 Third, using unique land ownership distribution data, we show that coffee region municipalities did not experience a greater degree of concentration of land ownership relative to a comparison group of municipalities, as would be the case with the traditional factor endowments–economic inequality hypothesis.8 We also rule out many other possible explanations for our results including differences in natives' and immigrants' sorting patterns across regions, in transportation cost changes and other technological improvements, and alternate means of the establishment of a coercive political system. In summary, our findings emphasize changes in relative wages and the incentive of elites to create and maintain a coercive labor regime as important mechanisms affecting workers' incentives to accumulate human capital. The paper is structured as follows: Section 2 describes the historical background and context. Section 3 discusses the theoretical background of our study. Section 4 describes the data used in the analysis and discusses the empirical strategy. Section 5 presents the main empirical results of the paper and evidence supporting our identifying assumptions. Section 6 assesses alternative explanations. Section 7 concludes.

This paper studies whether changes in the incentive for elites to enforce coercive labor institutions affected individuals' human capital accumulation decisions during coffee booms throughout the second half of the nineteenth century in Puerto Rico. We find that, under a coercive labor regime, local governments in coffee growing regions allocated more resources to the enforcement of coercive measures arguably with the goal of depressing the realized wage of unskilled labor. While this reduced the public provision of education, it also, surprisingly, induced workers to demand significantly more schooling leading to no change in the equilibrium amount of educational attainment in response to rising coffee prices. Following the abolition of these coercive measures in 1874, rising coffee prices correlated with declining literacy rates, consistent with a ‘resource curse’ in which individuals' perceived opportunity cost of schooling increases during commodity booms. Our results are consistent with models where, in the context of a free labor market, positive coffee price changes lead to a decrease in the skill premium and therefore a reduction in human capital investment. However, in a coercive regime, unskilled wages are artificially suppressed and therefore this effect is mitigated as workers avoid coercion and invest in human capital. After these coercive institutions are removed, similar changes in coffee prices reduce literacy rates, consistent with a diminished skill premium. This does not imply that workers are better off as a result of coercion. In fact, unskilled and skilled workers' welfare, as measured by lifetime income, is lower under the coercive labor regime than in a free labor market. In addition, our results raise a fundamental identification issue in assessing the relationship between episodic coercion and the accumulation of human capital. Following the insight of Domar (1970), coercion is likely to be present when demand for unskilled labor is high. However, demand for human capital is also likely to be low when demand for unskilled labor is high. Consequently, when looking at a negative (contemporary) relationship between some measure of human capital and whether coercion has occurred in the past, one must be careful to rule out if that relationship is simply coming from the demand side. Finally, the paper also provides evidence broadly consistent with North–south models of divergence with endogenous skill acquisition as trade can exacerbate initial skill differences across countries by raising the return to the relatively abundant factor. While we do not consider such mechanisms here, Stokey (1991) considers positive externalities to education such that trade can lead to lower human capital accumulation and economic growth in unskilled labor abundant countries. Although coercion of unskilled workers is not an acceptable institution because of the resulting limitations on the freedom of an individual (Sen, 1999), such mechanisms would suggest that policies that increase the skill premium and induce human capital accumulation may be socially beneficial. Empirically considering the long run relationship between historical labor market institutions, human capital accumulation, and growth remains an important next step in this line of research.