By Daron Acemoglu, Professor of Applied Economics, MIT and James Robinson, Professor, University of Chicago Harris School of Public Policy. Originally published at VoxEU
Editor’s note: This column first appeared as a chapter in the Vox eBook, The Long Economic and Political Shadow of History, Volume 1, available to download here.
The immense economic inequality we observe in the world today didn’t happen overnight, or even in the past century. It is the path-dependent outcome of a multitude of historical processes, one of the most important of which has been European colonialism. Retracing our steps 500 years, or back to the verge of this colonial project, we see little inequality and small differences between poor and rich countries (perhaps a factor of four). Now the differences are a factor of more than 40, if we compare the richest to the poorest countries in the world. What role did colonialism play in this?
In our research with Simon Johnson we have shown that colonialism has shaped modern inequality in several fundamental, but heterogeneous, ways. In Europe the discovery of the Americas and the emergence of a mass colonial project, first in the Americas, and then, subsequently, in Asia and Africa, potentially helped to spur institutional and economic development, thus setting in motion some of the prerequisites for what was to become the industrial revolution (Acemoglu et al. 2005). But the way this worked was conditional on institutional differences within Europe. In places like Britain, where an early struggle against the monarchy had given parliament and society the upper hand, the discovery of the Americas led to the further empowerment of mercantile and industrial groups, who were able to benefit from the new economic opportunities that the Americas, and soon Asia, presented and to push for improved political and economic institutions. The consequence was economic growth. In other places, such as Spain, where the initial political institutions and balance of power were different, the outcome was different. The monarchy dominated society, trade and economic opportunities, and in consequence, political institutions became weaker and the economy declined. As Marx and Engels put it in the Communist Manifesto,
“The discovery of America, the rounding of the Cape, opened up fresh ground for the rising bourgeoisie.”
It did, but only in some circumstances. In others it led to a retardation of the bourgeoisie. In consequence colonialism drove economic development in some parts of Europe and retarded it in others.
Colonialism did not, however, merely impact the development of those societies that did the colonising. Most obviously, it also affected the societies that were colonised. In our research (Acemoglu et al. 2001, 2002) we showed that this, again, had heterogeneous effects. This is because colonialism ended up creating very distinct sorts of societies in different places. In particular, colonialism left very different institutional legacies in different parts of the world, with profoundly divergent consequences for economic development.
The reason for this is not that the various European powers transplanted different sorts of institutions – so that North America succeeded due to an inheritance of British institutions, while Latin America failed because of its Spanish institutions.
In fact, the evidence suggests that the intentions and strategies of distinct colonial powers were very similar (Acemoglu and Robinson 2012). The outcomes were very different because of variation in initial conditions in the colonies. For example, in Latin America, where there were dense populations of indigenous people, a colonial society could be created based on the exploitation of these people. In North America where no such populations existed, such a society was infeasible, even though the first British settlers tried to set it up. In response, early North American society went in a completely different direction: early colonising ventures, such as the Virginia Company, needed to attract Europeans and stop them running off into the open frontier and they needed to incentivise them to work and invest. The institutions that did this, such as political rights and access to land, were radically different even from the institutions in the colonising country. When British colonisers found Latin-American-like circumstances, for example in South Africa, Kenya or Zimbabwe, they were perfectly capable of and interested in setting up what we have called ‘extractive institutions’, based on the control of and the extraction of rents from indigenous peoples. In Acemoglu and Robinson (2012) we argue that extractive institutions, which strip the vast mass of the population of incentives or opportunities, are associated with poverty. It is also not a coincidence that such African societies are today as unequal as Latin American countries.
It wasn’t just the density of indigenous peoples that mattered for the type of society that formed. As we showed in Acemoglu et al. (2001), the disease environment facing potential European settlers was also important. Something that encouraged the colonisation of North America was the relatively benign disease environment that facilitated the strategy of creating institutions to guarantee European migration. Something that encouraged the creation of extractive institutions in West Africa was the fact that it was the ‘white man’s graveyard’, discouraging the creation of the type of ‘inclusive economic institutions’ which encouraged the settlement and development of North America. These inclusive institutions, in contrast to extractive institutions, did create incentives and opportunities for the vast mass of people.
Our focus on the disease environment as a source of variation in colonial societies was not because we considered this to be the only or even the main source of variation in the nature of such societies. It was for a particular scientific reason: we argued that the historical factors that influenced the disease environment for Europeans and therefore their propensity to migrate to a particular colony are not themselves a significant source of variation in economic development today. More technically, this meant that historical measures of European settler mortality could be used as an instrumental variable to estimate the causal effect of economic institutions on economic development (as measured by income per-capita). The main challenge to this approach is that factors which influenced European mortality historically may be persistent and can influence income today, perhaps via effects on health or contemporary life expectancy. There are several reasons why this is not likely to be true however. First, our measures of European mortality in the colonies are from 200 or so years ago, before the founding of modern medicine or the understanding of tropical diseases. Second, they are measures of mortality faced by Europeans with no immunity to tropical diseases, which is something very different from the mortality faced by indigenous people today, which is presumably what is relevant for current economic development in these countries. Just to check, we also showed that our results are robust to the controlling econometrically of various modern measures of health, such as malaria risk and life expectancy.
Thus, just as colonialism had heterogeneous effects on development within Europe, promoting it in places like Britain, but retarding it in Spain, so it also had very heterogeneous effects in the colonies. In some places, like North America, it created societies with far more inclusive institutions than in the colonising country itself and planted the seeds for the immense current prosperity of the region. In others, such as Latin America, Africa or South Asia, it created extractive institutions that led to very poor long-run development outcomes.
The fact that colonialism had positive effects on development in some contexts does not mean that it did not have devastating negative effects on indigenous populations and society. It did.
That colonialism in the early modern and modern periods had heterogeneous effects is made plausible by many other pieces of evidence. For example, Putnam (1994) proposed that it was the Norman conquest of the South of Italy that created the lack of ‘social capital’ in the region, the dearth of associational life that led to a society that lacked trust or the ability to cooperate. Yet the Normans also colonised England and that led to a society which gave birth to the industrial revolution. Thus Norman colonisation had heterogeneous effects too.
Colonialism mattered for development because it shaped the institutions of different societies. But many other things influenced these too, and, at least in the early modern and modern period, there were quite a few places that managed to avoid colonialism. These include China, Iran, Japan, Nepal and Thailand, amongst others, and there is a great deal of variation in development outcomes within these countries, not to mention the great variation within Europe itself. This raises the question of how important, quantitatively, European colonialism was, compared to other factors. Acemoglu et al. (2001) calculate that, according to their estimates, differences in economic institutions account for about two-thirds of the differences in income per-capita in the world. At the same time, Acemoglu et al. (2002) show that, on their own, historical settler mortality and indigenous population density in 1500 explain around 30% of the variation in economic institutions in the world today. If historical urbanisation in 1500, which can also explain variation in the nature of colonial societies, is added, this increases to over 50% of the variation. If this is right, then a third of income inequality in the world today can be explained by the varying impact of European colonialism on different societies. A big deal.
That colonialism shaped the historical institutions of colonies might be obviously plausible. For example, we know that, in Peru of the 1570s, the Spanish Viceroy Francisco de Toledo set up a huge system of forced labour to mine the silver of Potosí. But this system, the Potosí mita, was abolished in the 1820s, when Peru and Bolivia became independent. To claim that such an institution, or, more broadly, the institutions created by colonial powers all over the world, influence development today, is to make a claim about how colonialism influenced the political economy of these societies in a way which led these institutions to either directly persist, or to leave a path dependent legacy. The coerced labour of indigenous peoples lasted directly up until at least the 1952 Bolivian Revolution, when the system known as pongueaje was abolished. More generally, Acemoglu and Robinson (2012, Chapters 11 and 12) and Dell (2010) discuss many mechanisms via which this could have taken place.
Finally, it is worth observing that our empirical findings have important implications for alterative theories of comparative development. Some argue that geographical differences are dominant in explaining long-run patterns of development. In contradistinction, we showed that once the role of institutions is accounted for, geographical factors are not correlated with development outcomes. The fact that, for instance, there is a correlation between latitude and geography, is not indicative of a causal relationship. It is simply driven by the fact that European colonialism created a pattern of institutions that is correlated with latitude. Once this is controlled for, geographical variables play no causal role. Others argue that cultural differences are paramount in driving development. We found no role at all for cultural differences measured in several ways. First, the religious composition of different populations. Second, as we have emphasised, the identity of the colonial power. Third, the fraction of the population of a country of European descent. It is true, of course, that the United States and Canada filled up with Europeans, but in our argument this was an outcome of the fact that they had good institutions. It is not the numerical dominance of people of European descent today that drives development.
See original post for references