Why agriculture and services are key to development

The literature on global growth convergence and divergence is vast and deep. And it is still evolving. Some have argued that global growth is actually diverging across countries. Pritchett (1977) called this “divergence, big time”, whereby the living standards of a few countries pulled away from the rest in the aftermath of the industrial revolution.

Others have found evidence in favour of growth convergence. Subramanian (2011) has argued that the number of developing countries experiencing catch-up has more than trebled (from 21 to 75 countries) and the rate of average catch-up has doubled from 1.5% per year to over 3%. However, the global growth convergence trend should not be taken for granted. Convergence may not persist in the future (Rodrik 2011), and the pace of global convergence may be adversely impacted by the global financial crisis (Dervis 2012).

While these insights and concerns are commendable and insightful, what might have been overlooked in this debate is the role played by different sectors–agriculture, manufacturing and services–in global growth convergence and divergence. Which of these sectors have acted as growth escalators? What can the latecomers to development in Africa learn about what drives global growth? In a recent paper we examined these questions for some 100 countries (Ghani and O’Connell 2014).

Our finding shows that global growth convergence has continued unabated. Just like the East Asian Tigers, the Lions of Africa are now growing much faster than the developed economies. However, the growth escalators in Africa are different than in East Asia. The East Asian Tigers benefitted from a rapidly expanding manufacturing sector. The African Lions are benefitting from others sectors.

Figure 1 plots the growth in service labour productivity for 100 countries–both developed and developing—over the last two decades on the vertical axis, and initial service labour productivity on the horizontal axis. The negative slope is unmistakable, suggesting that service sector growth exhibits a strong tendency for global convergence.

Figure 1. Global growth convergence in service is much faster

Source: World Development Indicators. Ethiopia data from Martins (2014). Labour productivity calculated by the ratio of total sector value-added employment in sector. Underlying accounts are in 2005 constant international USD. Values taken from earliest year available from 1990-1993 and available from 2005-9. Overall conclusions do not change if line is quadratic or cubic in ln(initial VA per worker).

The fitted line is a downward sloping line, implying that low-income countries and latecomers to development in Africa that started with a lower level of labour productivity in services, and were further away from the global labour productivity frontier, have experienced a much faster growth in service labour productivity. Take Ethiopia for example, a latecomer to development. It is above the trend line. So are China and India. They remain above the trend line even when estimates are changed from linear to non-linear regressions. This is good news for the latecomers to development. They have more room to catch up, and they are doing it faster, quicker, better and smarter.

Figure 2 plots growth in manufacturing labour productivity on the vertical axis and initial labour productivity on the horizontal axis. The fitted line is also downward sloping, implying that countries that started with a lower level of labour productivity in manufacturing have experienced a faster growth in labour productivity. This is consistent with other findings that manufacturing sector displays a clear tendency towards global growth convergence (Rodrik 2011).

Figure 2. Global growth convergence in manufacturing is a bit slower

Source: World Development Indicators. Ethiopia data from Martins (2014). Labour productivity calculated by the ratio of total sector value-added in sector. Underlying accounts are in 2005 constant international USD. Values taken from earliest year available from 1990-1993 and latest year. Overall conclusions do not change if line is quadratic or cubic in ln(initial VA per worker).

Unfortunately, figure 2 also shows that unlike the East Asian Tigers like China, Malaysia and Thailand which are above the trend line, the Lions of Africa like Ethiopia and Tanzania are well below the trend line, implying a much slower progress in their manufacturing sector compared to the East Asian Tigers.

What is striking when one compares figures 1 and 2 is that the strength of the global growth convergence in service is much stronger than in the manufacturing sector, i.e., the convergence line is steeper for service than for the manufacturing sector. This would suggest that the growth escalator in services sector can be potentially as powerful, if not more powerful, than in manufacturing sector for the latecomers to development in Africa.

What explains faster pace of global growth convergence in the service sector? Might it be the case that the latecomers to development are benefitting from better access to improved technology—mobile phone, big data, internet of things—that has revolutionised the services sector (Ghani 2010)? There are nearly 3 billion Internet users worldwide, and about a third of them are in the developing countries (Cerf 2014). Cross-border flows of data and communications have exploded, expanding by more than 50% per year since 2005, and people are now more interconnected than ever before (Tyson and Lund 2014).

Thanks to technology, services can now be unbundled and splintered in a value chain just like manufactured goods and they can be electronically transported. The number of services that can be transported digitally is constantly expanding – labour matching on platforms that can connect employers and employees across national boundaries (e.g. ODESK), call centres, e-learning, and much more. Indeed, services are now contributing more to growth than the goods sector, in both developed and developing countries.

However, not all is well with global growth convergence. There is no global growth convergence in agriculture. The fitted line linking labour productivity growth in agriculture with initial productivity level in Figure 3 is flat and could even be upward sloping. This means that labour productivity growth in agriculture has remained low in low-income countries. It is not catching up with more developed economies, unlike in the services and manufacturing sectors. This is a major concern, as this could compromise the goals of reducing poverty and achieving shared prosperity, given that agriculture is the main income source for the bottom 40% in many low income countries (Beegle 2104).

Indeed, yields per hectare have not improved in much of Africa (Stiglitz 2013). Yet, Africa’s potential in agriculture is bright. It has about 60% of the world’s uncultivated cropland. The sector has the potential to become a major part of the global agricultural value chain through continued expansion of commercial farming onto uncultivated land, shifting some production from grain crops to higher-value crops such as horticulture and biofuels, and by improving the productivity and yields of especially smallholder finance.

Figure 3. No global growth convergence in agriculture

Source: World Development Indicators. Ethiopia data from Martins (2014). Labour productivity calculated by the ratio of total sector value-added in sector. Underlying accounts are in 2005 constant international USD. Values taken from earliest year available from 1990-1993 and latest year. Overall conclusions do not change if line is quadratic or cubic in ln(initial VA per worker).

So what should latecomers to development do?

As a starting point, more emphasis should be placed on agriculture, rural and spatial development. It is also the main source of income for the bottom 40% of the population in the low income countries. So ending global poverty and boosting shared prosperity will require a greater focus not just on overall levels of growth, but more attention to the nature and patterns of growth that matter more for the poor, informal sectors and women.

In particular, growth that increases the returns to assets held by the poor–land and their labour—and boosts agricultural productivity are more likely to translate to effective poverty reduction. Farmers should have more access to the use of fertiliser and mechanised equipment. Food distribution and marketing channels should also be reformed, so that farmers can keep more of the proceeds from the sale of their crops. Policy makers should consider land title reforms aimed at opening more farmland without deforestation.

The Lions of Africa need not be a one-trick pony. They need to broaden their growth agenda to more pro-actively include both manufacturing and services. Indeed they are already doing it. The optimistic view is that Africa is already amongst the fastest growing regions in the world. Improved macroeconomic policy, governance, business environment, integrated trade and regional markets, demographic dividend, and rise of the middle class have all positioned the continent well for a brighter future.

But there is also a pessimistic view. Africa is simply not catching up fast enough with the rest of the world. Growth could be derailed by the slow pace of structural transformation, poor access to roads, energy, education, small manufacturing base, large informal sector, deep pockets of poverty and conflict and much more. And that a different path to structural transformation in Africa may fail.

Published in collaboration with Vox

Author: Ejaz Ghani, Lead Economist in Economic Policy and Debt, PREM Network, World Bank

Image: An employee checks newly-assembled electric cars to be exported to South America at an electric vehicle factory in Zouping county, Shandong province September 24, 2013. REUTERS/China Daily