Table of Contents
Introduction page 2
Review of Argumentspage 2 – 5
Critical Analysis of arguments and
Empirical evidence page 5 – 7
Conclusion page 8
References/Bibliographypage 9 – 10
A review of Daron Acemoglu, Simon Johnson and James Robinson (2001) The Colonial Origins of Comparative Development: An Empirical Investigation.
American Economic Association: The American Economic Review.
In this paper by Daron Acemoglu, Simon Johnson and James Robinson (hereafter, AJR) they propose a hypothesis linking European settler mortality rates to modern day GDP per capita of former colonies. Namely the influence that settler mortality rates had on the type institutions established in the colonies and how the type of institutions established has directly affected modern day GDP per capita of former colonies. European’s used very different colonization policies, in different colonies with different institutions. A determining factor in the type of institution that was established was the feasibility of whether or not Europeans could settle in the colony. Settler mortality rates hugely influenced whether ‘Extractive institutions’ were set up or if the colony modeled itself after European institutions. Colonies that had unfavourable disease environments were harder to settle in due to European susceptibility to diseases such as malaria or yellow fever, both of which held very high mortality rates among European settlers due to their lack of built of immunization.
Review of Arguments
AJR argue that initial settler mortality rates faced by Europeans in the colonies was the determining factor for the types of institutions that were established. In areas with a noxious disease environment colonists couldn’t possibly settle due to high mortality rates of diseases such as malaria and yellow fever which were both rampant at the time. In a colony with an unfavourable disease environment an ‘Extractive state’ was most often established; the objective of an extractive state was to transfer the largest amount of wealth from the colony to the colonizer for as little investment as required. Extractive states typically had poor property rights, no checks and balances on government power and incredibly high tax rates. An example would be that tax rates were four times higher in Tunisia than in France during colonial times (Young, 1994, p125); it’s estimated that fifty percent of Dahomeys GDP between 1905-1914 was extracted by France (Manning, 1982) On the other hand to extractive states is ‘Neo-Europes’ such as the USA, Australia, New Zealand or Canada. Neo-Europes were colonies that replicated European institutions with strong emphasis on private property laws along with checks and balances on government power. Using first stage estimates with no controls AJR estimate that approximately 25 percent of cross country income difference can be attributed to the differences in institutions established in former colonies (AJR, p1371) AJR argue that even when controls are used that there’s remarkably little change in the data even once you control for various exogenous factors. These results imply that Africa’s not worse off because of geographical factors as mentioned by some scholars but rather because of worse institutions that were set up as extractive institutions but have persisted to the modern day. AJR estimates imply that if you improved Nigeria’s institutions to the level of institutions in Chile, that in the long run lead to as much as a seven fold increase in Nigeria’s income (AJR, p1371). While this might seem incredibly high in reality Chile is over as eleven times as rich as Nigeria.
The strength of the core instrumental variable (hereafter, IV) as used by AJR, conditional on controls that the mortality rate faced by settlers over a hundred years ago have no effect on GDP per capita today aside from their influence through institutional development. Scholars such as Sachs (1998) contests this view arguing that malaria directly affects GDP and underperformance in modern day. AJR refutes this argument within the study highlighting that malaria primarily affected European settler because the majority of the indigenous population develops an immunity to it due to repeat exposure to the disease. These diseases were fatal to Europeans because they did not develop the same immunization as the indigenous people meaning it’s unlikely that these diseases are the main reason that many countries in Africa are poor today. AJR supports their claim with Curtin’s (1968) mortality rates of locals. Some scholars have presented primarily geographical arguments as to why Africa has developed comparatively much worse than other regions of the world. Distance from the equator is a commonly used metric but once you control for the effect of institutions on economic performance the colonies distance from the equator produces null results. AJR tests their hypothesis against several variables controls such as controlling for climate, religion, geography,identity of main colonizer, legal origin, soil quality, natural resources and measures of ethnolinguistic fragmentation. AJR’s results are robust to control variables but they admit it’s impossible for them to control for all possible variables that may be correlated with economic outcomes and settler mortality.
Rafael La Porta (1998/1999) emphasizes the importance of colonial and legal origin on current institutions. He argues that common law countries based on British common law have better property rights and more developed and sophisticated financial markets. Similarly Landes and North (1998) argue that British colonies prospered relative to former Spanish, French and Portuguese colonies because of the superior economic and political institutions along with the inherited British culture. AJR maintain that it’s not the colonial or legal origin that matter, but whether colonialists could safely settle in a particular location. There’s been several attempts to discover the link between institutions and development such as Mauro (1995) focusing on ethnolinguistic fragmentation, Hall and Jones (1999) using distance from equator as an instrument for western influence which leads to good institutions. Both studies suffer the problem that their instruments can plausibly directly affect performance today; if the instruments do have a direct effect than they become void. AJR’s approach has the advantage that their instrument does not have a direct effect on performance today because the settler mortality from over a hundred years ago should have no effect on output today, other than through its effect on institutions. Previously scholars misattributed other factors to the comparative development of former colonies but in this paper AJR show that these factors aren’t the determining factor.
AJR argue that the mortality rates were key in the type of institutions that were established and once these extractive states were established they persisted even after the colonies gained their independence and in certain cases it even worsened post independence.
Critical analysis of arguments and empirical evidence
The data used by AJR stems primarily from Curtain’s and Gutierrez (1986) mortality rates of soldiers, sailors and bishops between the seventeenth and nineteenth century. Some critics such as David Albouy (2012) contest the legitimacy of AJR’s results. Albouy highlights that in AJR’s study that includes 64 countries only 28 countries use mortality rates from inside their own borders, while 36 countries were assigned mortality rates based by AJR based on their inference of which countries share similar disease environment. Due to lack of reliable data available AJR had to infer the mortality rates for certain countries where they were not able to find any reliable sources. However, as Albouy points out this brings the entire legitimacy of the study in question because it makes the study prone to bias and that the data used by AJR artificially favours their own hypothesis. AJR admit in the study that there’s a lack of data available for mortality rates of Latin America due to poor record keeping by the Spanish and Portuguese militaries, and that they base the mortality rates for Latin America based on the works of Gutierrez. Using Gutierrez’s data on bishop mortality in Latin America as a proxy for the mortality rate of Latin America is questionable. Gutierrez’s data categorize the deaths by city into three climate categories of high, medium and low temperature climates assuming that cities with similar temperatures share similar disease environments, this is not a precise categorization and makes the assumption that areas with similar temperatures share similar disease environments but Gutierrez fails to establish that cities with similar temperatures actually share similar disease environments. The Bishop death rates aren’t particularly significant either comparable to the rest of the world at the same period. Gutierrez uses 3 different sets of bishop mortality rates based on low, medium, high temperature cities with 16.7, 17.5 and 30.2 mortality rates respectively (Gutierrez 1986). While the death rate of similarly aged Swedish men of the time was 18.32 and barracked soldiers in England and France had a rate of 15.3 and 20.17 respectively (Curtin, 1989). Albouy’s evidence suggest that settler mortality in Latin America was not much higher than that in Europe at the time. In Latin America at the time Yet AJR benchmark it to mortality rates faced by French soldiers campaigning in Mexico between 1862 and 1863, then 71 out of a thousand died from disease which is 4.25 the low temperature estimate provided by Gutierrez. Implying that Chile, Argentina and Mexico are four times more deadly than the United States because the United States is given a mortality rate of 15.
22 out of 36 countries with conjectured rates have their mortality rates inferred based off of the mortality rates of Mali and Mexico. This raises a lot of questions about how correct AJR’s inferences are in applying mortality rates to neighbouring countries. AJR applies a mortality rate of 14.9 to Hong Kong taken from data of a British force campaigning 1,200 miles north in China on a campaign of 107 days and when that rate is annualized is brings it up to 50.6 (Army Medical Department 1862). British soldiers during peacetime in Hong Kong actually died at a rate of 285 from 1842 to 1845 (Tulloch 1847). That makes AJR’s inferred mortality rate 19 times smaller than the actual rates at the time. The method of inference used by AJR seems incredibly unreliable with their estimates potentially being drastically different to the actual rates at the time. Even in colonies with data available from military records on mortality rates there’s inconsistencies. In some countries AJR uses at peace barracks mortality rates and in others he uses campaign mortality rates; “One of the fundamental facts of military medical experience is that troops in barracks are much healthier than troops on campaign, even disregarding losses from combat,” (Curtin 1989, p. 4). Typically on campaign soldiers were less likely to have quality food, shelter or sewage as well as taking less health precautions. AJR uses campaign rates more often in countries with lower GDP and higher expropriation risk possibly making the data aritificially favoring their own hypothesis. Curtin (1998) documents how during campaigns, mortality from malaria usually increases by more than 100 percent, gastrointestinal infections by more than 200 percent, and typhoid by over 600 percent, resulting in mortality rates 66 to 2000 percent higher on campaign than when soldiers are in barracks. Albouy argues that once you control for the source of mortality rates it substantially weakens the relationship between mortality rates and expropriation risk.
Once you look at the flaws in the data AJR it puts the entire legitimacy of their findings in question as certain data used has little basis in reality and may either be far too small an estimate or far to large. Due to the extensive inferring on the part of AJR it leaves room for bias and them artificially favoring their hypothesis via their methodology and data used to prove their hypothesis.
This research paper is incredibly influential and prior to it there had been little research on the effect that settler mortality and expropriation risk have had on the comparative development of former colonies. The instrumental variable of settler mortality more than 100 years, AJR uses is strong as it has no direct effect on GDP except through its influence on settlement patterns and institutions that were established by colonizers. The results of the research show that there is a correlation between the type of institutions established in the colonies and the GDP of said former colonies today. However, the extent of this difference is hard to quantify and as pointed out by critics there is flaws in the data used by AJR in the paper. The flaws in the data isn’t entirely fault of AJR either as their using the best information available to them. Despite inconsistencies and inferences on the part of the authors I feel that their findings do show a strong relationship between the type of institutions established in colonies and the GDP of the colonies today. This study really highlights the importance of our economic and political institutions within our society and economy.
Acemoglu, Daron, Simon Johnson and James Robinson. 2001. The Colonial Origins of Comparative Development: An Empirical Investigation. American Economic Review 91(5): 1369-1401.
Bloom, David E. and Sachs, Jeffrey D. ;Geography, Demography, and Economic Growth in Africa.; Brookings Papers on Economic Activity, 1998, (2), pp. 207-73.
Curtin, Philip D. The image of Africa. Madison, WI: University of Wisconsin Press, 1964. . ;Epidemiology and the Slave Trade.; Political Science Quarterly, June 1968, 83(2), pp. 181-216.
Curtin, Philip D. Death by migration: Europe’s encounter with the tropical world in the 19th Century. New York: Cambridge University Press, 1989. .
Curtin, Philip D. Disease and empire: The health of European troops in the conquest of Africa. New York: Cambridge University Press, 1998.
Landes, David S. The wealth and poverty of nations: Why some are so rich and some so poor. New York: W.W. Norton & Co., 1998
La Porta, Rafael; Lopez-de-Silanes, Florencio; Shleifer, Andrei and Vishny, Robert W. "Law and Finance." Journal of Political Economy, December 1998, 106(6), pp. 1113-55. . "The Quality of Government." Journal of Law, Economics, and Organization, April 1999, 15(1), pp. 222-79.
Manning, Patrick. Slavery, colonialism, and economic growth in Dahomey, 1640-1980. New York: Cambridge University Press, 1982.
Gutierrez, Hector. "La Mortalite des Eveques Latino-Americains aux XVIIe et XVIII Siecles." Annales de Demographie Historique, 1986, pp. 29-39.
Hall, Robert E. and Jones, Charles I. "Why Do Some Countries Produce So Much More Output Per Worker Than Others?" Quarterly Journal of Economics, February 1999, 114(1), pp. 83-116.
Mauro, Paulo. "Corruption and Growth." Quarterly Journal of Economics, August 1995, 110(3), pp. 681-712.
North, Douglass C.; Summerhill, William and Weingast, Barry. "Order, Disorder and Economic Change: Latin America vs. North America." Unpublished manuscript, Hoover Institution, Stanford University, 1998.
Young, Crawford. The African colonial state in comparative perspective. New Haven, CT: Yale University Press, 1994.
Army Medical Department. 1862. Statistical, Sanitary and Medical Reports of the Army Medical Department for the Year 1860. London: Harrison and Sons.
Tulloch, A. M. 1847. “On the Mortality of Her Majesty’s Troops Serving in the Colonies during the Years 1844 and 1845.” Journal of the Statistical Society of London 10 (3): 252–59.