Is Africa Deindustrializing Prematurely?
John Stuart, tralac Associate, discusses Africa’s recent industrialization path by examining trends in industrial employment and manufacturing value added in relation to other world regions
In order for a country to advance from a poor, agrarian state into a modern post-industrial state, experience has taught that it must follow a specific pattern of development, in which industrialization is a key phase. This pattern was observed in the economies of the developed West, Japan and the ‘Asian Tigers’ of South East Asia. Until recently, it was assumed that the balance of the developing world would follow the same pattern in moving from low to middle to high income nations. However, recent research appears to contradict this assumption and to suggest instead that many developing nations are instead ‘deindustrializing’ and headed down an unchartered economic path towards potential low-income/low development traps.
Development theory has evolved over the last 50 years from a position of assuming industrialization was a result of policy-created internal diversification and trade barriers. Since the United States was a relatively closed economy until the 1970s, and since it had industrialised with high diversification and in the absence of major export growth, it was assumed that this pattern could be repeated for pre-industrial economies.
The subsequent experiences of several economies in East Asia – first Japan and then Taiwan and South Korea – illustrated, however, the importance of exports as a driver of growth. However, these latter two economies had used trade barriers to protect ‘infant industries’ and so there was no demonstration effect that trade barriers could not still be an important component of trade and industrial policy, despite the emphasis of policy shifting from an import-substituting (ISI) focus to an export-oriented focus (EOI).
The developing world therefore tried to imitate the success of the developed world by attempting to follow the same path of industrialization, based on export-led growth. Eventually the inherent logic of the law of comparative advantage permeated trade policy and with the founding of the World Trade Organization (WTO) the world began a multilateral process of reducing trade barriers and ‘globalization’.
The results of globalization have, however, been mixed. Whilst some nations – such as China, India and the ‘Asian Tigers’ have undoubtedly benefitted and industrialized rapidly, much of the balance of the developing world – including Africa – has not managed to follow the same industrialization path.
Of more concern however, is the recent assessment that much of the developing world – especially Latin America and Africa – are in fact ‘deindustrializing’, or experiencing reductions over time in relative industrial employment and/or manufacturing value added. Dani Rodrik (see his profile) is a political economist who has closely examined this phenomenon and concluded that much of the developing world is deindustrializing ‘prematurely, i.e. at much lower levels of per-capita income than was the case for the developed world. Among the implications of this are:
Inability to create urban and formal jobs to facilitate urbanization from the agrarian, subsistence state
Economies unable to leverage the dynamism and scalability of industry to accelerate growth
Developing economies see an increasing importance of economic activity in the services sector, but many of these are ‘petty’ services that are low-tech and can’t be scaled or traded. The services sector is therefore lacking in dynamism
Potential social and political instability.
In order to assess the situation Africa finds itself in with respect to its industrialization path, it is useful to examine the data. Whilst the data sets for Africa are not complete, a surprising amount of data exists and it is possible to draw conclusions and make meaningful comparisons with other regions. Two sets of data are accessed in order to analyse Africa’s path of industrialization:
Total employment in industry as reported by the International Labour Organization (ILO). Rather than present the raw data, industrial employment is presented as a proportion of total population. This gives an idea of the changing importance of industrial employment within a region’s population.
Manufacturing value added (MVA) as a proportion of GDP. This gives an idea of relative contribution to output of manufacturing, as opposed to the primary and tertiary sectors. Whilst it would have been preferable to have reported this statistic in real terms, only the nominal version of the data set was available for most African countries.
Figure 1: Industrial employment (% population) for four regions
The evolution in industrial employment shares since 1991 is depicted in Figure 1 for four regions. It is clear that East Asia has grown most rapidly of the regions, and this region alone has ‘pulled’ the world aggregate upward. When one nets off the effect of East Asia, it is clear that globally, industrial employment is falling as a proportion of the population. Although there is marginal growth in the indicator for Sub-Saharan Africa (SSA) and the Middle East and North Africa (MENA), the increase is so slight in both cases as to be effectively a ‘sideways’ movement. What is more concerning is that there is no convergence between the patterns of East Asia and those of SSA/MENA. For these regions, there is no discernible movement in the variable that would indicate a pattern of progressive industrialization.
When considering industrial employment in the current context, the effects of automation need to be taken into account. The incidence of automation tends to increase with the level of value-add of a particular manufacturing technology, but this curve is ‘shifting down’ over time as technology advances. This means that it would be possible and in fact likely, to observe that employment in industry would necessarily grow more slowly than industrial output. This pattern is a particular concern for Africa, with its high levels of unemployment and large youth population cohort.
These data series in Figure 1 are influenced by cyclical as well as trend factors. The East Asia and World indicators show a turning point in 2002, when the fallout from the 1998 Asian financial crisis had finally given way to a new cycle of industrial-led growth. Clearly though, the trend in East Asia has been upward, whereas that for SSA and MENA has been flat.
Figure 2: Manufacturing value added for four regions
Figure 2 depicts industrialization from a different perspective – as manufacturing value added (MVA) a share of GDP. East Asia showed a marginal increase in MVA share up to 2006, where after there was a small decline. The other regions show a definite decreasing trend however. This pattern is not necessarily a bad thing when considered overall, because advanced economies typically display an increasing share of services in overall value-added. This is an expected and even desirable trait of economic development and carries with it benefits to the consumption rates of non-renewable resources and to pollution levels.
However, what is true in net should not be true for developing countries. East Asia’s MVA share is already high – around 30%, indicating the importance of industrial activity to the economy. This is a healthy structural condition, the far lower and declining shares for MENA and SSA are of concern – these regions are not, as is the developed world, ‘post industrial’, but are in fact pre-industrial. Their path, over time, should be a rising MVA share from the levels it is at now. If instead, these economies de-industrialize at this stage of their development, they may become trapped in a state of low income and low dynamism.
Figure 3: SSA relative to East Asia
In order to get an idea of the extent to which specifically SSA is lagging East Asia, Figure 3 presents the data series for SSA net of East Asia’s figures. East Asia has a very large population – much of it still rurally-based - but with rapid urbanization. By comparison, SSA lags East Asia in industrial employment population share by almost 10% at the end of the sample. Of equal concern is that the net indicator has a downward trend over the sample.
The net indicator for MVA share shows an even larger gap – an average for the sample of around 18%. The indicator has begun to converge in the last six years but remains way short of the East Asia level. The recent convergence in the indicator could reflect the increasing importance of China’s tertiary sector (China dominates the East Asia regional group).
In conclusion it is clear from the data presented that Africa is not industrializing as expected by the standard economic growth model. By comparison with the East Asia region, which is following the traditional economic growth and industrialization path, it could be said that Africa is in fact de-industrializing. There are various risks – both economic and political (the latter specifically dealt with by Rodrik) – associated with this pattern and it should be taken very seriously by policy makers in the continent and further afield. A forthcoming tralac Working Paper by this author will explore the issue in more detail and attempt to identify causes and recommend solutions.