The last 500 years have witnessed a massive rearrangement of the world’s population. Somewhat over half of the globe’s inhabited surface, comprising almost all of Asia, Africa, and Europe, consists of countries populated almost entirely by descendants of the same territory’s pre-1500 inhabitants. The remainder consists in similar parts of:
- Countries like the US, Canada, Argentina, Brazil, Australia and New Zealand in which less than 10% of the population is of indigenous descent, and
- Countries like Mexico, Ecuador, Bolivia and Peru having indigenous fractions in between the two extremes.
As the attention of students of economic growth to history, geography and institutions has grown in recent years, the geographic and cultural origins of populations have been dealt with in a variety of ways, including a “neo-Europe” dummy variable (Hibbs and Olsson, 2004), controlling for the proportion speaking European language (Hall and Jones, 1999), and using dummy variables for “predominant European settlers” and “lesser European settler predominance” (Comin, Easterly and Gong, 2008).
The 5 century migration matrix
Suspecting that a more systematic way of accounting for the population movements since the year 1500 was called for, we constructed and studied a matrix of world migration in which the columns correspond to countries of origin, the rows to countries of current residence, and the cell entries indicate the proportion of the ancestors of each country today who lived on the territory now comprising that and each of 164 other countries in 1500.
To construct the matrix, we consulted a large number of secondary sources including three conventional encyclopaedias, specialized encyclopaedias and databases on ethnicity such as the World Christian Encyclopaedia, and sources such as the CIA World Factbook, Nationsencyclopedia.com, and Everyculture.com.
These sources break country populations down into ethnic, linguistic, or racial categories, with respect to which we made required inferences about origins. For example, we assumed that African-Americans are mainly descended from populations in one of twenty-two sub-Saharan African countries and we attributed their ancestors to those countries in the proportions indicated in Eltis et al.’s “Trans-Atlantic Slave Trade” (1999). No doubt, a better resourced effort could generate somewhat more precise estimates, but for now our matrix appears to offer the only general compilation that lets one map present-day populations into ancestries prior to the rearranging of world demographics in the era of European expansion.
What the matrix says
The matrix data tell us that today’s countries are extremely diverse in terms of the importance of immigration. The fraction of today’s population that has ancestors who lived in their present country in 1500 are:
- 3% for US, Canadian, Australian and New Zealand;
- 94% to 98% for Southeast Asia, South Asia, and East Asia;
- 94% for Europe on average;
- 89% for North Africa and West and Central Asia;
- 65% for Mexico and Central America;
- 20% for South America;
- 0%, for the Caribbean.
Armed with population estimates for 1500 and 2000, we can also generate estimates, for each country, of the number of present-day descendants per person of 1500 and the proportion of these who live today in the same versus other world regions.
Using the McEvedy and Jones (1976) estimates of populations in 1500, the Penn World Tables 6.2 population estimates for 2000, and the matrix’s estimates of origins, we find that there were 31.6 million descendants of the Irish population of 1500 alive in 2000, with less than 12% of these living in Ireland, 76% in the US, 5% in Australia, and so on. In contrast, the vast majority, although not all, of the descendants of China’s year 1500 inhabitants live today in China.
We can also trace the origin of a particular nation’s population today. For example:
- About 4.4% of the current US population’s ancestors lived in France in 1500,
- About 8.5% lived in Ireland,
- About 1.1% in Senegal, and
- About 0.9% in China.
It also tells us that:
- about 97.9% of the ancestors of today’s Indians lived in India in 1500,
- about 77% of the ancestors of today’s Singaporeans lived in China,
- about 45% of the ancestors of today’s Fijians lived in India.
Migration and GDP
The map summarise the rows of the migration matrix, i.e. the origins of today’s population going back to 1500. This map graphically illustrates the terms ‘old world’ and ‘new world’; today’s populations on the Eurasian land mass consists almost entirely of indigenous populations, while the Americas and Australia are predominantly populated by immigrants.
Origin matters for GDP
The power of regional origins is illustrated by the fact that in a 125-country regression, 44% of the variance in the natural logarithm of year 2000 per capita GDP is accounted for by entering only the share of the population’s ancestors that lived in Europe in 1500. If the share of the population speaking English, French, German, Portuguese or Spanish is substituted, the explained variance falls to only 12%. (The matrix but not the Hall-Jones European language share variable permits estimates on a larger sample of 152 countries; here 41% of the variance of current income is explained by the share of European ancestors variable.)
Although the 1500 cut-off was chosen with reference to the Columbian encounter, there are important examples of post-1500 population movement involving actors other than Europeans. These include not only the populating of Haiti and other Caribbean islands by people of African descent (through involuntary movements organized by Europeans), but also the populating of Taiwan (which had a mainly non-Chinese indigenous population in 1500) by ethnic Chinese and the establishment in Fiji and Guyana of large ethnic Indian populations. To attempt to explain the 20th Century success stories of Taiwan and Singapore or the struggles of Haiti and the Dominican Republic without reference to their non-European population movements seems comparable to trying to explain the economic emergence of an Australia or a Canada with no mention of European colonization.
Impact on income differences within and between nations
We perform various exercises to investigate the effects of post-1500 population movements on income differences both among countries and within countries. The results indicate a degree of persistence of capacity to generate or to capture income that is almost breathtaking. The power of theories positing the persistence of developmental advantages acquired hundreds or even thousands of years ago is dramatically augmented by incorporating information on ancestry into one’s analysis, and inequalities within countries are much better predicted by heterogeneity and linguistic distance between ancestors than by heterogeneity and distance between contemporary ethno-linguistic groupings.
Some of our analysis uses measures of early advantage such as duration of agricultural experience and history of state-level organization (we report this in a Vox column that will appear tomorrow). We also carry out exercises using a more “blank slate” approach that lets the data speak for itself with regard to the differences among countries and world regions. In one exercise, we divide the world into twelve regions based on geographic proximity and similarity of population origins. We then regress the year 2000 log of GDP per capita of 152 countries on either dummy variables for country location by region, a familiar approach, or on the fraction of each country population’s ancestors who lived in each of eleven regions in 1500 (a remaining region serving as omitted category). Either approach “explains” the better part of income variation: 58% in the case of the region dummies. But the source region shares explain somewhat more – 63%.
Why it matters
Why should we care about the apparently powerful influence that population origins exert on country and sub-national incomes levels?
First, if this influence is indeed as significant as our findings suggest it to be, then efforts to sort out the roles that geographic, institutional, and other factors play in explaining income levels and growth rates may produce misleading results unless we properly control for it.
Second, the influence of population origins suggests that there is something that human families and communities transmit from generation to generation -- perhaps a form of economic culture, a set of attitudes or beliefs, or informally transmitted capabilities -- that is of at least similar importance to economic success as are more widely recognized factors like quantities of physical capital and even human capital in the narrower sense of formal schooling. If we understand which culturally transmitted factors are important and what contributes to their emergence and propagation, we might be able to design policy interventions that could help less successful groups and countries to close their developmental gaps.
Comin, Diego, William Easterly and Erick Gong, 2008, “Was the Wealth of Nations Determined in 1000 B.C.?” Harvard Business School Working Paper 09-052.
Eltis, David, Stephen Behrendt, David Richardson, and Herbert Klein, 1999, The trans-Atlantic slave trade: a database on CD-ROM. Cambridge: Cambridge University Press.
Hall, R., Jones, C. (1999). “Why do some countries produce so much more output than others?” Quarterly Journal of Economics 114 (1), 83–116.
Hibbs, Douglas A., and Ola Olsson, 2004, “Geography, biogeography, and why some countries are rich and others are poor,” Proceedings of the National Academy of Sciences 101: 3715-3720.
McEvedy, C. and R. Jones (1978). Atlas of World Population History, Viking Press.
Putterman, Louis and David N. Weil, 2008, “Post-1500 Population Flows and the Long Run Determinants of Economic Growth and Inequality,” NBER Working Paper 14448.