Double diversification

Thorvaldur Gylfason, Per Wijkman

06 February 2017



Gains from trade arise from specialising domestic production in those goods in which a country has a comparative advantage. There can, however, be too much of a good thing.

Excessive specialisation creates risk. Should an over-dominant specialised sector be hurt, its problems may damage the rest of the national economy. Hence the need for economic diversification. Also, the specialised sector needs to be one without adverse spillover effects on other sectors. Exploitation of natural resources often creates negative externalities in the form of rent seeking and Dutch disease (Spence et al. 2008).

Natural resource exploitation may attract rent-seekers to politics who transfer rather than create wealth. They often impart a false sense of security, a feeling that natural resource wealth makes it unnecessary to build up human and social capital to create sustainable, inclusive economic growth.

Most large open economies export a diverse range of merchandise to other countries, mostly manufactures. Small countries, especially those specialising heavily in the production of primary commodities for export, often fail to develop parallel lines of production. This suggests that diversification is especially important for small countries. 

Figure 1 uses the Herfindahl-Hirschman index of market concentration to illustrate the difference between countries such as the US, the UK, France, Germany, and Canada, and countries such as Chile, Iran, Saudi Arabia, Nigeria, and Iraq. The countries on the left in the figure have an export structure similar to the world average, while those on the right have an export structure that is much more concentrated and less diversified than the world average. Notice that New Zealand’s specialisation in agricultural exports means that its export structure less diversified than that of Norway, even though Norway is one of the world’s largest exporters of oil. Notice also that Iceland’s exports of fish and aluminium mean that its export structure is as concentrated as those of Saudi Arabia and Nigeria.

Figure 1 Herfindahl-Hirschman index of market concentration 2012

Note: The Herfindahl-Hirschman index shows the extent to which the structure of exports by product of a given country differs from the world average. The index ranges from 0 to 1, with values closer to 1 indicating a bigger difference from the world average. Several other indices are closely correlated with the Herfindahl-Hirschman index: the Finger-Kreinin export diversification index (UNCTAD), the IMF’s export diversification index based on Theil’s index, the IMF´s product quality index, and the economic complexity index developed by Hidalgo and Hausmann (2009). For a comparison of these indices and how they relate to democracy and growth, see Gylfason (2017).
Source: UNCTAD.

Figure 2 illustrates the cross-country relationship between export diversification as measured by the average share of manufactures in total merchandise exports in the period 1962 to 2012, and the log of the purchasing power of per capita gross national income in 2012 in 139 countries. The use of only the end-of-period value of per capita GNI for each country rules out reverse causation from incomes to export diversification.

Figure 2 Export share of manufactures and income per person, 1960-2012

Note: Each country is represented by a bubble whose size reflects the country´s population.
Source: Authors‘ computations based on World Bank, World Development Indicators.

The relationship is significant in a statistical sense (rank correlation = 0.62) as well as in an economic sense. The slope of the regression line (0.029) suggests that a 20-point increase in the manufacturing share (for example, from 40% to 60%) goes with an increase in real per capita GNI by 58%. The rank correlation between the share of manufactures in exports in Figure 2, and one minus the Herfindahl-Hirschman index in Figure 1 – an index that rises as exports become more diversified – is 0.71.

Economic diversification is rising. The world average share of manufactures in total exports rose from 59% to 69% from 1962 to 2012, according to the World Bank’s World Development Indicators. This is good for global growth.

The importance of political diversification

Just as economic diversification spurs growth by transferring labour from low-paying jobs in low-skill-intensive farming and mining to more lucrative jobs in skill-intensive occupations, political diversification spurs growth by redistributing political power from ruling elites to the people. Replacing an extended monopoly of power with a democracy reflecting political pluralism has economic effects. Political diversification exemplified by the promotion of electoral competitiveness, openness, and popular participation contributes to growth. Put differently, broader-based and more political organisation can be viewed as an investment in social capital, including strong civil society, good governance, and societal institutions that people can trust (Paldam 2000).

Figure 3 illustrates the cross-country relationship between political diversification through democracy and per capita GNI in the same 137 countries as in Figure 2, from 1960 to 2012. Two countries were not represented, due to lack of data. Democracy is measured by the average of the Polity2 variable in each country between 1962 and 2012, a variable that ranges from -10 in autocracies to +10 in full democracies. As in Figure 2, the vertical axis shows the log of purchasing power of per capita GNI in 2012 to rule out reverse causation from income to democracy.

Figure 3 Democracy and income per person, 1960-2012

Note: Other measures of political democracy include the Political Rights Index and Civil Liberties Index compiled by Freedom House. These are closely correlated with Polity IV Project’s Polity 2 variable. For a comparison of these indices, see Gylfason (2017). 
Source: Authors‘ computations based on data from Polity IV Project and World Bank, World Development Indicators.

The relationship is significant in a statistical sense (rank correlation = 0.62) as well as in an economic sense. The slope of the regression line (0.135) suggests that a four-point increase in Polity2 (for example, from 2 to 6) goes with an increase in real per capita GNI by 54%.

Figure 4 Economic and political diversification, 1960-2012

Note: The rank correlation between the share of manufactures in exports and democracy across the 141 countries shown is 0.49, and rises to 0.52 if China is omitted from the sample.
Source: Authors‘ computations based on data from Polity IV Project and World Bank, World Development Indicators .

The resemblance between Figures 2 and 3 suggests that economic and political diversification go hand-in-hand with strong economic performance. Also, the two types of diversification go hand-in-hand across countries. The more diverse the economy, the more diverse the political structure (Figure 4). If the absence of democracy enables rent-seekers to hold back economic as well as political diversification, the advance of democracy – that is, political diversification – seems likely to create conditions for economic diversification, and vice versa.

Notice that China is an outlier in the upper left of the figure, while India is mainstream.

Double ascent

Modern economies need to diversify to offer ordinary people a steadily improving standard of life. Agriculture tends to perpetuate poverty, and excessive dependence on natural resources delays the development of more productive industry and services.

Common property resources require public management. Norway used its oil wealth to build up a Pension Fund of nearly $900 billion, invested overseas, which is equivalent today to about $170,000 per person (Gylfason 2008). Iceland, with abundant fish and energy resources, carries a crushing debt burden due to poor public management. Thus, modern economies also need democratic pluralism with broad political participation. Without unfettered democracy, bad governments tend to last too long and do much damage.

The need for political diversification is particularly urgent in natural resource-rich countries that often face a double jeopardy – that is, natural wealth that is concentrated in the hands of relatively small elites that block both economic and political diversification (Ross 2001). Rent-seekers typically resist reforms – economic as well as political diversification – that would redistribute the natural resource rents to their rightful owners (Wenar 2008, Gylfason and Wijkman 2016).

Among the ten Spanish or Portuguese speaking countries of Latin America, the number of democracies (countries with Polity2 scores from 6 to 10) grew from three in 1961 to eight in 2012, while the share of manufactures in total merchandise exports rose from 8% to 46%. This pattern is repeated around the world, except in Europe and Central Asia where manufactures were about 70% of total exports 50 years ago, and the same today.

The double ascent of economic diversification and democracy generally supports growth. Between 1962 and 2012, democracy was rising around the world (Diamond 2015). But indicators of democracy (for example, freedom of speech, confidence in public institutions, lack of corruption) have decreased in many countries in the past ten years. Freedom House has recently demoted the US from its top rank for the first time. Indices reflecting economic diversification also threaten to weaken in many countries. Unless corrective action is taken, these trends may impair global economic growth.


Diamond, L (2015), “Facing Up to the Democratic Recession,” Journal of Democracy 26 (7), 141-155.

Gylfason, T (2008), “Norway´s Wealth: Not Just Oil”, VoxEU, 6 June.

Gylfason, T (2017), “From Double Diversification to Efficiency and Growth,” Comparative Economic Studies (forthcoming).

Gylfason, T and P M Wijkman (2016), “Double Diversification with an Application to Iceland,” Chapter 11 in S Mahroum and Y Al-Saleh (eds) (2016), Economic Diversification Policies in Natural Resource Rich Economies, London and New York: Routledge.

Hidalgo, C A and R Hausmann (2009), “The Building Blocks of Economic Complexity,” in P S Dasgupta (ed.), Proceedings of the National Academy of Sciences of the United States of America 106 (26), June 30, 10570-10575.

Mahroum, S and Y Al-Saleh (eds) (2016), Economic Diversification Policies in Natural Resource Rich Economies, Routledge, London and New York.

Paldam, M (2000), “Social Capital: One or Many? Definition and Measurement,“ Journal of Economic Surveys 14 (5), 629–653.

Ross, M (2001), “Does Oil Hinder Democracy?” World Politics 53, April, 325-361.

Spence, M et al. (2008), The Growth Report: Strategies for Sustained Growth and Inclusive Development, Commission on Growth and Development, World Bank, Washington, D.C.

Wenar, L (2008), “Property Rights and the Resource Curse,” Philosophy and Public Affairs 36 (1), 1-32.



Topics:  Development International trade Politics and economics

Tags:  specialisation, diversification, growth, Dutch disease, rent-seeking, concentration, democracy

Professor of Economics, University of Iceland; Research Fellow, CESifo and CEPR Research Fellow

Former Director of Economic Affairs at EFTA Secretariat, Geneva, Switzerland and Adjunct Professor of International Economic Policy at University of Göteborg, Gothenburg, Sweden.