نابرابری درآمد و استعمار

This paper proposes that colonialism is a major explanation behind today's differences in income inequality across countries. We argue that income inequality has been higher in the colonies where the percentage of European settlers to total population was higher, as long as Europeans remained a minority. The countries where Europeans became the majority of the population did not suffer from high inequality. These initial differences continue to hold today. The empirical evidence we provide strongly supports our thesis.

In the last few years a considerable number of papers have increased our understanding of the way the colonial experience of today's developing countries has influenced their subsequent economic evolution. While the analysis of many short term issues might be safely done without considering the historical context of each country, such approach would be questionable for long term problems like differences in income levels between rich and poor nations. It is precisely the careful consideration of history in general, and of colonial history in particular, that sets papers like Acemoglu et al., 2001 and Acemoglu et al., 2002 or Engerman and Sokoloff (2002) aside among those addressing the question of why southern countries have not attained the economic well-being of the world's industrial countries. Acemoglu et al. (2001) argue that the pattern of European settlement in the colonies determined the type of institutions that these countries developed and that these institutions are a major factor behind their economic backwardness. In those regions where few Europeans settled, Europeans created “extractive states” and the resulting institutions “...did not introduce much protection for private property, nor did they provide checks and balances against government expropriation.” On the other hand, the authors also propose that the countries that received a large number of settlers “tried to replicate European institutions” and therefore created the right set of rules encouraging future economic growth. European migration to a given region was largely determined by the mortality rates that future settlers would face. This provided the authors with an instrument to test the effect of institutions on the level of GDP per capita, namely the mortality rates of European settlers in colonial times. Engerman and Sokoloff (2002) provide us with a convincing account of how many of the differences in economic outcomes between the northern and the southern part of the American Continent are rooted in colonial times. The authors identify as the initial cause of this divergence the differences in factor endowments that Europeans found in the New World. The southern part of the continent, with its aptitude for sugar production and mining, started with a much more pyramidal society than the northern one, with its family-size agricultural units. Engerman and Sokoloff “document—through comparative studies of suffrage, public land and schooling policies—systematic patterns by which societies in the Americas that began with more extreme inequality or heterogeneity in the population were more likely to develop institutional structures that greatly advantaged members of the elite classes (...) by providing them with more political influence and access to economic opportunities.” The present paper places itself in the same line as the two aforementioned ones. Like them, it studies a long run distinction between developed and developing countries in the light of the colonial past. The object of our study is the large differences in income inequality that exist among countries. Using the most popular measure of inequality, namely the Gini coefficient, we can observe that the countries of Latin America and sub-Saharan Africa have much higher levels of inequality than those prevailing in Europe or Asia. The average value of the Gini coefficient is 51.80 for Latin America and 46.85 for sub-Saharan Africa while for Western Europe and South and East Asia the values are 31.83 and 37.71, respectively.1 The only explanation for these important differences that has ever gained wide approval among the economics profession is the “Kuznets’ hypothesis”. Proposed by Kuznets (1955), it asserts that inequality should evolve following an “inverted U” pattern as countries develop. Although initial tests of Kuznets’ hypothesis gave contrasting results,2 the issue could finally be settled once comprehensive time series data on inequality became available for a large number of countries. This took place in the mid-nineties thanks to the work of Deininger and Squire (1996). Tests using this data set showed Kuznets’ hypothesis to be wrong.3 The main proposition of this paper is that different experiences during the colonial period are a major explanation behind today's differences in inequality across countries. We use an objective measure in order to differentiate the diverse colonial experiences the World has known in the last five centuries. This measure is the percentage of European settlers living in the colony at its apogee. We hypothesize that the higher this percentage the greater the inequality in the country, as long as Europeans remained a minority. The second section of the paper develops the argument in more detail and presents a brief historical account. The third section presents the econometric evidence in favor of our thesis. The central point we want to make is that the effect of colonialism on inequality is: (i) statistically significant, (ii) quantitatively large (and in particular, larger than the effect of most other variables) and (iii) robust to several specifications. Our paper can be related to a hefty literature that treats inequality and growth as endogenous outcomes that depend on initial conditions and some market imperfections. Several papers in this literature suggest that an economy can have multiple equilibria that greatly differ in terms of both inequality and growth. Thus, the initial conditions of the economy, for instance the initial distribution of wealth, might lead the country into an undesirable equilibrium of low growth and high inequality. The models of Banerjee and Newman (1993), Galor and Zeira (1993), Aghion and Bolton (1997) and Bénabou, 1996 are good examples of this literature. These models use an imperfect credit market as the source of their results. In Banerjee and Newman (1993), an economy with an initial distribution of income that is strongly unequal will see a large part of its population excluded from the credit market. This exclusion will not allow poorer people to invest in physical and human capital, reducing the growth rate of the economy and perpetuating the high inequality situation. Bénabou, 1996 also sees inequality as potentially self-reinforcing, but he stresses other channels. He argues that in high inequality countries, policies of redistribution would not find enough support from rich individuals since the transfers they would be required to make would be too large with respect to the potential gains that stem from a faster growing economy. Other authors who have also studied the political dimension of the problem are Alesina and Rodrik, 1994 and Persson and Tabellini (1994). Their view is somewhat opposed to that of Bénabou, 1996 in the sense that they see redistribution as a nuisance for the economy, arguing that it implies a higher level of taxation and distorts economic decisions. This literature offers several possible explanations for the long lasting effects of colonialism on inequality that we advocate. Credit market imperfections and redistributive issues are certainly part of the economic environment in third world countries. These and other features could have “trapped” countries in the high inequality state in which they found themselves at the end of the colonization era. We can also relate our paper to a broader literature studying the consequences of colonialism. Many accounts of colonialism have stressed its large costs on colonized societies: Direct human costs, slave trade and exploitation of natural resources. In addition to this, some authors have put the blame for the Third World's underdevelopment on the colonial rule. Bairoch (1993), who argues that the imposed free-trade with the colonies caused a large deindustrialization in India, states: “There is no doubt that a large number of negative structural features of the process of economic underdevelopment have historical roots going back to European colonization”. A similar argument is used in a recent work by Galor and Mountford (2004). On the other hand, some recent work on the subject have tilted the balance towards a more positive view of colonialism. Ferguson (2003), for instance, stresses that more investment flowed into poor countries in the colonial period of one century ago than today. The reasons for these large investments point to other benefits of colonialism: The introduction of legal protection of property rights, effective public services and the rule of law. One can find a certain degree of contradiction between this last view and the hypothesis made by Acemoglu et al. (2001). According to these authors, colonialism brought good institutions but only in the countries of large European immigration. The majority of the colonies are not in this group. Our paper distinguishes itself from the aforementioned in important ways. Acemoglu et al. (2001) is concerned mainly with institutions and growth; inequality is not really analyzed. Engerman and Sokoloff (2002) do discuss the issue of inequality but they limit themselves to the American Continent and do not offer a quantitative analysis like we do. We can also mention the empirical works of Bertocchi and Canova (2002) and Grier (1999), who are concerned with the effect of colonialism on economic growth. One aspect that the present paper shares with the studies by Acemoglu et al. (2001) or Engerman and Sokoloff (2002) is that it stresses the persistence of the colonial order after colonization was over, at least in some places. This is at the center of the thesis of Acemoglu et al. (2001), where one of the three premises their paper is based on is “The colonial state and institutions persisted even after independence”. Engerman and Sokoloff (2002), discussing the case of Mexico, state: “The evidence obviously conforms well with the idea that in societies that began with extreme inequality, such as Mexico, institutions evolved so as to greatly advantage the elite in access to economic opportunities, and they thus contributed to the persistence of that extreme inequality.” Our paper also relies on such an assumption, but on a relatively milder version. We assume that the main characteristics of the colonial order persisted after independence in the countries of important European settlement. We do not require this to be true in the many colonies where Europeans settlers were a small fraction of the population. This and other views on colonialism will be presented in the following section.

Our paper's main point is that colonial history is a major explanatory factor behind today's large differences in inequality among the world's countries. We have reviewed the different colonial experiences of the last five centuries and have classified them in three broad categories. Of these three, we argued that one clearly produced and sustained highly unequal societies. This high inequality group is the one where colonialism brought into the country an amount of European settlers whose number was considerable but still inferior to that of the local population. This minority was able to concentrate most of the countries’ income in their hands, mainly by excluding the rest of the population from owning land or mining resources. Moreover, and with the exception of Algeria, it was this minority who took all political power once these countries became independent. This allowed high inequality to remain a characteristic of these countries up to our times. We believe that the empirical evidence strongly advocates our view. The number of European settlers as a percentage of total population has clearly a statistically significant and an economically substantial effect on inequality. The effect is robust to the inclusion of a large number of control variables, including controls for climate and some natural endowments, dummies for Latin America and sub-Saharan Africa and for the identity of the colonial power. While some aspects of the econometrical work can always be changed, the results seem to be so consistent that we can be confident in them. One might be tempted to infer, given the above conclusions, that European settlement was a negative phenomenon for the countries where it took place (excluding the four cases of Australia, Canada, New Zealand and the U.S.) since the society that came out of these countries was highly unequal. In addition, one could add that such a message is in contradiction with the results of Acemoglu et al. (2001), where more European settlements have rather positive consequences since they are associated with better institutions and thus with a higher level of GDP per capita. Such a conclusion would not be correct since there is no contradiction whatsoever between the fact that Settler colonies became highly unequal and the one that these same colonies achieved a higher level of production per capita than Peasant ones. This relative economic success was precisely the result of the European settlers’ growth record. By their cultural background they were able to put at least partially in place the technology and institutions that made the economic superiority of Europe and the New Europes. It should also be kept in mind that the higher inequality of the former Settler colonies does not imply that the poorer majority of these countries is less well-off in absolute terms than the people from the less unequal Peasant colonies. Inequality is a relative concept and it does not inform us about the absolute level of GDP per capita. There exists a trickle-down effect in the former Settler colonies since the rich minority demands and consumes many services provided by the poor and in this way makes them profit at least in part of their well being. If this effect is strong enough then it might be better to be poor in a former Settler colony than in a former Peasant colony, at least in what concerns absolute levels of income. A final comment on the “Kuznets’ hypothesis” is also in order. Our paper tells us that the colonial experience of developing countries is of the highest relevance when analyzing inequality issues while the level of development, measured by an indicator like GDP per capita, is not. The empirical record refutes the “inverted-U” relationship between inequality and income while it supports the one between inequality and European settlement. It might then come as a surprise that some intellectual support for the ideas we present here can be found in the same paper of 1955 where Simon Kuznets formulates his famous hypothesis. Kuznets had certainly noted that his analysis was not appropriate for every country in the world and that there was something particular about some colonized regions. In Kuznets (1955, p. 21) we read that he is excluding of his analysis countries that are characterized by “... large native populations with small enclaves of nonnative, privileged minorities, e.g. Kenya and Rhodesia, where income inequality, because of the excessively high income shares of the privileged minority, is appreciably wider than even in the underdeveloped countries cited above.” The problem is that this case of a small enclave of non-native, privileged minority is far from being an exception and concerns an important number of developing countries in the world. Kuznets, after all, had seen it right.