Barriers to the spread of prosperity

Enrico Spolaore, Romain Wacziarg

10 February 2017



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 diffusion of wealth is arguably the most consequential development for human welfare in recent history. Since the Industrial Revolution, modern prosperity has spread from its European birthplace to many corners of the world.1 Yet the technologies, institutions and behaviours associated with this process of economic modernisation have diffused unequally over space and time. Why?

A recent literature has documented the important role played by deeply-rooted factors as predictors of the current world distribution of income and other economic outcomes. These factors include geographic conditions and historical events that sent different societies on different economic trajectories — the effects of bio-geographic endowments (Olsson and Hibbs 2005, Ashraf and Galor 2011), the legacy of colonialism (Acemoglu et al. 2001), the persistent effect of pre-colonial traits and institutions (Michalopoulos and Papaioannou 2013), the durable cultural impact of traditional agricultural practices (Alesina et al. 2013), and the effects of long-term history and movements of populations across the globe (Spolaore and Wacziarg 2009, Putterman and Weil 2010, Ashraf and Galor 2013), to name but a few. Many of these historical determinants are summarised in the new VoxEU eBook in which this column features (Michalopoulos and Papaioannou 2017). However, the mechanisms by which deeply-rooted factors influence current prosperity remain elusive. Moreover, studies that emphasise the persistence of historical legacies and long-term determinants raise questions about the scope for change. As pointed out, for instance, in an excellent discussion by Banerjee and Duflo (2014), there is an inherent tension between historical determinism and the ability of policy to affect outcomes. If the past casts such a long shadow, can contemporary societies escape from factors and constraints that may have historically limited their economic development?

In this column, we argue that the divergent historical paths followed by distinct populations led to barriers between them. The more divergent the historical paths of different populations, the greater the barriers. And the greater the barriers, the more difficult it was for innovations, institutions and behaviours to spread from society to society. Hence, on average, countries that are richer today are those more closely related to the frontier society where modern technologies, institutions and behaviours first arose. In order to prosper, more distantly related societies need to overcome the barriers that separate them from societies that are closer to the frontier. However, while such barriers are deeply-rooted, their effect is not permanent and immutable. Historical factors do not constitute permanent limits to the growth potential of those with disadvantageous historical legacies. Instead, barriers resulting from distinct historical trajectories can be gradually overcome, suggesting a substantial role for action and positive change.

Measuring human barriers

In principle, barriers to the transmission of prosperity can arise from numerous sources. Geographic barriers are likely to be important for several outcomes, and they are perhaps easiest to measure and control for in empirical work on the diffusion of development. Measuring human barriers – those that prevent, at a given geographic distance, the spread of innovations, institutions and behaviours – is much more challenging. In our past work, beginning with Spolaore and Wacziarg (2009), we employed a variety of measures of historical separation among populations to capture human barriers. Chief among them was FST genetic distance, a measure that captures separation times between populations: when humans migrated out of Africa, groups splintered as they moved across continents, and the groups that separated earlier had more time to drift apart genetically than groups that separated more recently. Hence, genetic distance is correlated with how long populations have had a common history. Pairs of societies with smaller genetic distance are expected to have lower human barriers to the spread of development.2

The idea behind the use of genetic distance as a general proxy for human barriers is that human traits – not only biological but also cultural – are mostly transmitted, with variation, from generation to generation (i.e. vertically). Thus, the longer two societies have drifted apart, the greater the differences in traits between them, and the greater the barriers that separate them. Of course, genetic distance is by no means the only measure of intergenerational separation times. Linguistic distance is a closely related class of measures, again based on a trait that is mostly transmitted vertically (language). Another possibility is to look directly at differences in culture, as revealed by surveys: values, norms, and attitudes (including but not limited to religion). Cultural values  can be transmitted in a number of ways – vertically, from generation to generation; obliquely, across biologically unrelated members of the same society; or horizontally, i.e. across societies (Richerson and Boyd 2005). The vertical dimension of transmission is a common feature of genetic traits and language, as well as of norms and values. Thus, metrics of distance between societies that are based on these three classes of measures, while distinct from each other, should be positively correlated. This is indeed what  we find in Spolaore and Wacziarg (2016a), where we further discuss and document empirically the complex links between various measures of human relatedness. In a nutshell, the vertical transmission of genes, language and culture accounts for the positive correlations between human distance metrics based on each of these traits. Yet these measures are not perfectly correlated because: i) there are differential rates of drift in genes, language and values, ii) some of these traits are transmitted horizontally, and iii) different methodologies are used to compute distances across the three classes. In our ongoing research on the diffusion of development, we use all three classes of measures.

Three examples

What is the evidence that these measures of human relatedness matter when predicting differences in prosperity? In recent work we have found such evidence in a variety of contexts. Here we will discuss three: technology, institutional quality, and fertility behaviour.

The diffusion of development

In Spolaore and Wacziarg (2009, 2014a), we documented a strong correlation between genetic distance between countries relative to the technological frontier and their differences in levels of development: two societies are predicted to have similar levels of development if they happen to be at relatively similar distances from the global technological frontier (in our applications, either the US or northwestern Europe). We interpreted this correlation as indicative of barriers to the spread of the Industrial Revolution. We showed in particular that the effect of barriers was largest just after the Industrial Revolution, when some but not all countries had transitioned to economic modernity. The effect declined as more and more societies, at successively greater genetic distances from the innovation frontier, became rich. In the age of globalisation, when barriers became easier to overcome, the effect fell further (Figure 1).

Further, in Spolaore and Wacziarg (2012, 2014a) we found that this pattern held true not just for the overall level of prosperity, measured by per capita income, but also for specific technologies (mobile phones, computers, etc.). In sum, societies that are historically distant from the technological frontier have a harder time adopting better technologies, and consequently take longer to become prosperous.

Figure 1 Standardised effect of genetic distance relative to the UK on bilateral differences in per capita income over time, 1820–2005

Source: Spolaore and Wacziarg (2014a)

The diffusion of institutions

In Spolaore and Wacziarg (2016b), we conducted a similar exercise to understand the worldwide diffusion of democracy during the Third Wave of Democratisation that took off in the 1970s. The manner of this diffusion process was similar to the spread of the Industrial Revolution: genetic distance relative to the institutional frontier (the US) matters increasingly after the onset of the third wave, and declines gradually as more countries, at greater distances from the institutional frontier, become democratic (Figure 2). What deserves further research is the precise mechanism whereby institutional change spreads from one country to the next.

Figure 2 Standardised effect of genetic distance relative to the US on bilateral differences in Polity 2 Democracy scores, 1960–2005

Source: Spolaore and Wacziarg (2016b)

The diffusion of the fertility transition in Europe

In the two examples above, the effect of distance from the frontier fades away after some time, but does not disappear entirely. Yet a prediction of our diffusion model is that the effect of ancestral distance should disappear after the most distant societies have finally overcome the barriers and adopted modern technologies, institutions and behaviours. The case of the European fertility transition, starting in the early 19th century in France, affords an example where the entire diffusion process can be observed within our sample. In Spolaore and Wacziarg (2014b), we analysed this process in a panel of European regions from 1831 to 1970. We measured ancestral distance using linguistic distance, since this was more readily available for the regions of Europe than genetic distance. Initially, only regions that spoke a language close to French adopted the fertility behaviour first observed in France in the late 18th to early 19th centuries. Later, regions at successively greater distances from France adopted the new behaviour. By the end of our sample period, virtually every region in Europe had adopted modern behaviours regarding fertility (i.e. 2-3 children per household). The interpretation of this particular diffusion process is different than for our other examples for two reasons. First, the frontier society in this case was not England, but France. This fact highlights how different innovations may start at different frontiers – implying different barriers to their diffusion. Second, fertility behaviours likely diffused as the result of a process of social influence regarding appropriate norms of fertility, rather than the diffusion of specific technologies (although the diffusion of birth control methods – broadly defined – may have played a complementary role). Whatever the precise mechanism, the lesson is clear: ancestral barriers, measured by relative linguistic distance from French, predicted the diffusion of modern fertility behaviours across Europe.

Figure 3 Standardised effect of linguistic distance to French on marital fertility through time

Note: Overlapping samples of 30 years centred on the date displayed in the x-axis.The sample is a balanced sample of 519 European regions
Source: Spolaore and Wacziarg (2014b)

Conclusion: Barriers and the scope for policy

As we have argued in this column, populations that are historically and culturally more distant face higher barriers to adopting each other’s technologies, institutions, and behavioural innovations. Such barriers – measured by genetic, linguistic and cultural distance – stem from long-term historical divergence, and thus capture the effect of deeply-rooted historical factors that sent different populations on different historical trajectories. However, we have also seen that the effect of barriers is not permanent and immutable, but changes over time, as societies that are farther from the frontier also learn and adopt novel technologies and innovations.

Moreover, the frontier itself is not immutable, but changes over time, and may differ depending on the specific innovation – for example, the frontier was originally England for the Industrial Revolution, but France for the societal changes in norms and behaviour associated with Europe’s demographic transition.

If such historical barriers can be overcome – and they have indeed been overcome by many societies over time – there is room for optimism regarding the scope for change and progress, even when dealing with persistent historical factors.3 While distances themselves may be deeply-rooted in history, their impact on contemporary outcomes can, in principle, be affected by current actions and policies. For instance, policy can reduce obstacles to interactions and communication between people from different cultural and linguistic backgrounds. Our research suggests that the effect of barriers to the spread of prosperity has diminished in the age of globalisation. The ease with which ideas, people, goods and capital can flow across societal borders helps to reduce the ancestral barriers that kept populations from learning from each other. Facilitating these flows, therefore, offers the promise of lower barriers to the spread of prosperity.


Acemoglu, D, S Johnson and J Robinson (2001), “The Colonial Origins of Comparative Development”, American Economic Review 91(5): 1369-1401.

Alesina, A, P Giuliano and N Nunn (2013), “On the Origins of Gender Roles: Women and the Plough”, Quarterly Journal of Economics 128(2): 469-530.

Ashraf, Q and O Galor (2011), “Dynamics and Stagnation in the Malthusian Epoch”, American Economic Review 101(5): 2003–41.

Ashraf, Q and O Galor (2013), “The ‘Out of Africa’ Hypothesis, Human Genetic Diversity, and Comparative Economic Development”, American Economic Review 103(1): 1-46,

Banerjee, A and E Duflo (2014), “Under the Thumb of History? Political Institutions and the Scope for Action”, Annual Review of Economics 6: 951-971.

Galor, O (2011), Unified Growth Theory, Princeton: Princeton University Press.

Michalopoulos, S and E Papaioannou (2013), “Pre-colonial Ethnic Institutions and Contemporary African Development”, Econometrica 81(1): 113-152,

Michalopoulos, S and E Papaioannou (2017), The Long Economic and Political Shadow of History, Volume 1, CEPR Press.

Mokyr, J. (2005), “Long-Term Economic Growth and the History of Technology”, in P Aghion and S N Durlauf (eds), Handbook of Economic Growth, Volume 1B, Amsterdam: Elsevier, North-Holland.

Olsson, O and D A Hibbs, Jr. (2005), “Biogeography and Long-Run Economic Development”, European Economic Review 49(4): 909–38.

Putterman, L and D N Weil (2010), “Post-1500 Population Flows and the Long-Run Determinants of Economic Growth and Inequality”, Quarterly Journal of Economics 125(4): 1627–82.

Richerson, P. J. and R. Boyd (2005), Not By Genes Alone: How Culture Transformed Human Evolution, Chicago: University of Chicago Press.

Spolaore, E and R Wacziarg (2009), “The Diffusion of Development”, Quarterly Journal of Economics 124(2): 469-529.


1 For example, see Mokyr (2005) for an insightful historical discussion and Galor (2011) for a unified account of the growth take-off.

2 Of course, since geographic and genetic distances are correlated - because groups splintered gradually as they moved farther and farther away from East Africa, while conquering other territories - it is imperative to control for geographic distance in any work that uses genetic distance as a measure of human barriers.

3 That said, we should add that inter-population barriers do not always play a negative role in human history. They may also prevent the spread of deleterious innovations, such as hateful ideologies or disruptive behaviors, and may reduce international conflict over territories and resources (see Spolaore and Wacziarg 2016c).



Topics:  Economic history Poverty and income inequality

Tags:  Inequality, diffusion of wealth, prosperity, Industrial Revolution

Professor of Economics, Tufts University

Professor of Economics, Anderson School of Management, UCLA; and Research Fellow, CEPR