HISTORY, GEOGRAPHY, & THE AFRICAN SLAVE TRADE: AN EMPIRICAL ANALYSIS

We empirically examine the effect of bad geography (measured as rugged terrain) on well-being outcomes through its historical interaction with the African slave trade. This paper builds on a broad steam of research that examines how geography interacts with historical events in shaping contemporary outcomes. While bad geography (rugged terrain) is in general negatively correlated with economic outcomes, previous literature has empirically argued that paradoxically in Africa, countries with a rugged landscape tend to (paradoxically) perform better than flatter ones, because rugged terrain in the case of Africa shielded those being raided during the slave trades and hence would have a historic indirect positive effect on income. In this paper, we further examine the robustness of this positive ruggedness effect for Africa by introducing time-series data to examine the magnitude of this effect over time via calculating marginal effects (as opposed to static approaches in the previous literature). We also exploit novel non-GDP measures of well-being as dependent variables to further test the robustness of ruggedness’ positive effect for Africa. Finally, we also make improvements on various techniques in the previous literature, such as incorporating multiple imputation methods to deal with missing data (vs. list-wise deletion). After incorporating new data, dynamic measurements, and improved empirical techniques, the positive ruggedness effect for Africa as suggested by the previous literature appears to be less unambiguously robust than implied. We find some evidence to suggest that insofar that this positive ruggedness effect for Africa exists, it is one that diminishes as the historical distance from the precipitating event (slave trade) increases, the direction of this effect does not always hold for other non-GDP measures of well-being, and its effect may often be overpowered by various time-period specific events.本研究通过考察恶劣地理条件（以崎岖地形（rugged terrain）为衡量指标）与非洲奴隶贸易的历史互动关系，实证检验其对福祉结果的影响。本文依托于一系列探讨地理与历史事件如何互动以塑造当代发展结果的研究脉络。尽管恶劣地理条件（崎岖地形）总体上与经济发展结果呈负相关，但既往学术研究通过实证提出了一项悖论：在非洲地区，地形崎岖的国家往往比平坦国家表现更优，究其原因在于非洲的崎岖地形为在奴隶贸易时期遭受劫掠的群体提供了庇护，进而对居民收入产生了历史性的间接正向影响。本文通过引入时间序列数据，计算边际效应（区别于既往研究中的静态分析方法）以考察该效应随时间变化的幅度，进而进一步检验非洲地区崎岖地形正向效应的稳健性。此外，本研究还采用创新性的非GDP福祉衡量指标作为因变量，进一步验证非洲地区崎岖地形正向效应的稳健性。最后，本研究还对既往研究中的多种实证方法进行了改进，例如引入多重插补法处理缺失数据（替代以往的完全病例删除法）。在纳入新数据、动态测量方法与改进后的实证技术后，既往研究提出的非洲崎岖地形正向效应，其稳健性似乎并不如此前研究所暗示的那般明确。本研究发现的部分证据表明：若非洲地区的崎岖地形正向效应确实存在，那么该效应会随着与触发事件（奴隶贸易）的历史距离增加而逐渐减弱；同时，该效应的正向性并不总能在其他非GDP福祉衡量指标中成立，且其影响往往会受到特定时期各类事件的抵消。