Down to Earth Again: The Third Stage of African Growth Perceptions

Income Levels

While sub-Saharan Africa has not reached the Millennium Development Goal of halving the 1990 poverty rate by 2015, the region as a whole was lauded for having made inroads by bringing the figure down by a quarter, to 40.9 per cent of the population of SSA, which will cross the one-billion-inhabitants mark one of these days. So, there has been progress on poverty eradication, but the somewhat neglected reality is that the absolute number of poor people earning under USD 1.25 per capita daily is constantly rising: from 290 million in 1990 to 403 million in 2015 (representing a 40 per cent increase). As Africa south of the Sahara is the only world region with an absolute rise in poverty, the highest share of destitute people globally is now found here rather than in South Asia; SSA contains almost half of the developing world's poor (48.3 per cent). ⁴ This is factually new. Another upward revision of the international poverty line by the World Bank will present this trend in an even grimmer light.

Trade and Hyperglobalisation

The main engine of prosperity over the period under review was what is now called “hyperglobalisation,” as international trade growth considerably exceeded that of domestic production. The recently terminated globalisation round, which historically paralleled the long spell from 1870 to 1914, was driven by an extended prosperity cycle in the United States and by the emergence of one mega-trader, China (Subramanian and Kessler 2013). Africa benefitted all along. The relative growth rates of emerging countries are now slipping back to those of old industrial nations. The headwinds returned at the sudden end of Chinese hyper-growth, which was halved from 14 per cent in 2007 to approximately 7 per cent a year later and has remained at the latter level since. The World Trade Organization (WTO) has observed a slowdown that may well represent, as is currently being debated, “peak trade” for the world economy (WTO 2014). Merchandise exports grew almost three times more quickly than GDP in the 1990s, and up until 2009 still almost two times more quickly. In the current decade, such trade integration has levelled out; in 2012 and 2013 the growth of the volume of goods traded and GDP growth were on par with one another. The important corollary of this levelling-off is that it occurred well before the end of another feature associated with hyperglobalisation – the commodities super-cycle, which came to a halt in 2014.

It is not at all clear whether this evening-out will last and what the underlying reasons for it are. The simplest explanation refers to the basic distinction between internationally tradeable and non-tradeable goods and services. When there are no trade barriers, competition in globally tradeable goods is fiercer than it is for non-tradeables, over time driving their relative prices down – hence, the current value of exports and imports relative to GDP. Other scholars have emphasised, instead, a growing number of non-tariff barriers that have been erected, in particular since the crisis years 2008/2009, which has slowed down trade – an observation manifestly at odds with the dominant view that developed and developing countries alike have withstood the protectionist temptation. The Economist quotes related analysis, but also Paul Krugman, who perceives the trend as less damaging, and James Manyika (of McKinsey Global), who argues that the growing kind of immaterial trade components is not properly counted in trade statistics (The Economist, A Troubling Trajectory, 13 December 2014). ⁵ When we include software, design, and logistics, real trade in goods is arguably far bigger than the accounted-for merchandise trade. While there is a consensus on this, which has triggered efforts to remedy the situation statistically, it does not tip the balance back in favour of developing countries. The classic case is the iPhone or iPad, manufactured by Foxconn in China. Of an average sale of USD 200 in the United States, a meagre USD 10 ends up in the Chinese economy, because the lion's share of design, software, and high-end components comes from four industrialised countries (Karabell 2014), and what applies to China applies a fortiori to Africa.

Regarding the global trade peak and its negative impact on developing-country growth, one genuinely Africa-specific growth accelerator springs to mind: intra-regional trade, which could help. Trade integration within Africa lags considerably behind that within other world regions, despite numerous regional economic communities at work. Trade facilitation measures can allow intra-African formal trade to catch up considerably and thus counteract the dampening global trade effect, simply by putting African countries on the same footing of exchange with their neighbours, as is the case within other world regions, above all the EU. However, low intra-regional trade is endogenous to production structures; more precisely, it is endogenous to the persistent lack of an agricultural and industrial division of labour, limiting complementarity in cross- and intra-sectoral trade. Mobilising catch-up potential from regional trade is therefore conditional on good agro-industrial strategies in the respective regions. At the present time, though, no such strategies are being employed.

Structural Change

The long growth spell that Africa has experienced since the mid-1990s has led to some structural change. But is that change in the right direction, and is it decisive enough? Secular development of economies has long been portrayed in terms of change among economic sectors. Since Clark, Fourastié, and Kuznets, the advancement from agricultural to industrial and possibly service-driven economies has been the stylised development path. Lewis-style analysis of the dualism between a traditional (mainly agrarian) and a modern (mainly industrial and service) sector completed the pattern for less developed countries. However, closer analysis of such patterns of growth, along with its companion input–output analysis, fell largely into disrespect and became almost an orphan discipline sometime after the end of Hollis Chenery's stint as the World Bank's vice president for development policy in 1982 and the copublication of his textbook (Chenery and Srinivasan 1988). Using tools of structural change analysis to favour some (sub)sectors over others – particularly industry – seemed to no longer give traction to development – at a time when the markets were supposed to discover opportunities for themselves in perfect sector neutrality. ⁶ Patterns of structural change were in fact not a good predictor of successful development. While the emerging East Asian economies presented an accelerated replication of the Western light and heavy industry model, manufacturing in Latin America stalled after the 1950s; and who could have predicted that Africa's latest growth episode would rely on commodities, telecommunication, trade, and finance?

Regarding manufacturing's share in African GDPs, there is a rather sobering rule of thumb: on average, manufacturing has accounted for approximately 15 per cent of GDP since the 1960s, regardless of the prevailing policy (Lawrence 2005), but in Africa's Least Developed Countries, or when subtracting South Africa from the total, the share is at 10 per cent and falling. In Africa's now-biggest economy, Nigeria, manufacturing value added lingered at approximately 2 to 3 per cent at the beginning of the new millennium. Characteristically, the global commodities boom of the 2000s helped Africa to sustain growth but contributed to keeping the share of manufacturing close to the depressed long-term average. Manufacturing does not even mean “industry” proper, as artisanal activity in much of the informal sector is counted as well. Bakeries remain the most widespread “industrial” enterprise in a number of African countries. So, the share occupied by manufacturing industry hovers somewhere in the single digits. For most African countries, modern industry is still negligible, in stark contrast with the South and Southeast Asian experience. Interestingly, the continent's industrial giant, South Africa, is also declining towards the average without plausibly having gone through any structural transformation that would allow for the shutting down of old manufactures in favour of some bright post-industrial future.

Employment and Labour Productivity

New research now deepens our knowledge on structural change by providing insights into shifts in employment and labour productivity, and this represents an important turn. Data on employment and unemployment is notoriously scarce for African countries, and there has not been much attention paid to labour statistics, apart from the International Labour Organization's efforts to compile reliable figures for some countries. Groundbreaking work has been conducted at the University of Groningen, whose Growth and Development Centre has produced a consolidated global database of employment data from household and labour force surveys that was first used in analyses of Asia and Latin America, but later for Africa as well (Timmer and de Vries 2007). The database has since been used by several authors in combination with other new or updated sources. The key issue is that labour productivity gains ⁷ can be generated either by modernisation in one industry (say, by the introduction of a new technology) or at given productivity levels by factor movement from less productive to more productive sectors. As Peter Timmer established, agriculture secularly lags in modernisation, and in employment more than in output, so the gain is mainly generated by moving both from agriculture to industry and among industries (Timmer 1988). Services are a difficult subject in this respect. McMillan and Rodrik (2014) recall that such inter-sectoral moves do not necessarily mean progress; they can entail “productivity reducing structural change” when labour moves in the wrong direction, back into less productive sectors or into forms of unemployment with zero or negative marginal productivity, outpacing output growth in the modernised sectors. Based on the new work with African countries added to the initial Groningen database, they contend that exactly this happened in sub-Saharan Africa, and partly in Latin America: structural change reduced productivity and, in the end, growth.

Here, economists rediscover in technical terms a bitter reality that marked East Germany after reunification and Eastern Europe, as well as Latin America and Africa, the latter two under structural adjustment programmes, leaving industrially deserted landscapes behind. Industry shrank due to forced modernisation, and labour moved in the wrong direction – out of industry and most notably into rural and urban informality. Whilst both the Groningen team (De Vries et al. 2013) and McMillan and Rodrik initially saw this happening in Africa over the entire time span from 1990 into the early 2000s, in the 2014 paper the latter team subdivided the period and found a net move of labour out of traditional agriculture and into modern services or even industry for the years 2000 to 2010. Structural change then contributed positively approximately 1 per cent to labour productivity growth in sub-Saharan Africa. Moreover, overall labour productivity growth in Africa was second only to Asia's, as was absolute GDP growth. McMillan and Rodrik call this “Africa's turnaround” – an “encouraging” message consistent with the rest of the “African growth miracle.” ⁸ To what extent is this really good news, and why did the most prominent author defect from the optimist bench right after publishing it?

As background work for IMF/WB regional reports, some authors used a different methodology to estimate employment growth in Africa. Main results are published under the deliberately ambiguous title “Africa's got work to do” (Fox et al. 2013). This team, too, found that agricultural employment had shrunk to 60 per cent of total employment between 2005 and 2010, and conclude that “at least some type of employment transformation is underway.” However, 60 per cent of the workforce produces only approximately 20 to 30 per cent of the value added. Wage labour in modern industry (informal activities counted separately as part of the household economy) remained stagnant at 3 per cent of employment, except for the upper-middle-income countries (UMIC: Botswana, Cape Verde, Equatorial Guinea, Gabon, Mauritius, Namibia, South Africa), where the share is approximately 15 per cent, oil and mines included. Industry is definitely not the mainstay of employment in Africa, because both modern non-artisanal mining and manufacturing absorb relatively few workers. The third sector, “services,” next to acclaimed modern finance or trade, also comprises general and local government. To get at least an idea of how job creation evolved along the lines of private and public sector job creation, the WB/IMF analysts estimated the share of public service in wage employment across African country groups for 2005: while the sub-Saharan weighted average stands at a bit above one third, after the adjustments in civil service imposed in the 1980s and 1990s, a whopping 69 per cent of all wage employment in the resource-rich countries was offered by the public sector – a striking confirmation of the resource curse, as oil riches do not even generate a private service boom, let alone a manufacturing one. Resource rent income has instead created government jobs (ibid.).

The authors cited take the present GDP growth rate of 6 per cent as an optimistic baseline for a projection until 2020 and nonetheless spell out a sobering prognosis: the African working-age population will predominantly remain in the farm and non-farm household economy (70 to 80 per cent); just 15 per cent will work in wage-paying (= “modern”) services (work for the government again included); and a meagre 3 per cent will be in wage-paying industry. Neither an industrial nor a post-industrial service economy is in sight. Based on employment/growth elasticities bigger than unity, the World Bank and IMF present an alternative scenario aptly labelled “game change.” Unfortunately, the uppercase scenario barely arrives at doubled salaried industrial employment in 2020. McKinsey-style fanfare is not warranted.

Industrialisation

Game change indeed is what is required. Current employment trends point to the need for strong industrial policies to push modern labour force absorption in both resource-rich and resource-poor African countries. Some African countries have started implementing targeted industrial promotion – for instance, Botswana, Ethiopia, Namibia, and Rwanda. However, another problem with such a strategy has resurfaced. That opportunities for latecomers to replicate the industry-reliant growth path of today's fully industrialised countries may be limited (in the extreme: to the handful of East Asian newly industrialised countries as well as China) has long been a subject of debate, and, in a way, the low numbers of industrial jobs created in Africa are a manifestation of that. New insights have now reinforced the doubts. ⁹ Industry as a share of GDP and employment follows an inverted U-curve – kind of a hump – which peaks at certain national income levels to give way to services. The new pieces of comparative research indicate two stylised facts:

In other words, the inverted U-curve shifts left and downwards. Countries simply tend, over time, to be less industrial, and in huge countries like India this can be demonstrated even among federal states (Amirapu and Subramanian 2014). Countries thus face premature de-industrialisation compared to historical growth experience (Rodrik 2014, 2015). The IMF authors give an indication for the height of the manufacturing industry: “Manufacturing value added peaks at a level of GDP per capita (in logs) of 8.5 (around USD 5,000)” (Dabla-Norris et al. 2013: 7). The peak, although occurring later in Amirapu and Subramanian's and Rodrik's output estimates (while at the same stage for the employment share: USD 5,500 in 1990), comes right after the turn from a lower-middle-income country to an upper-middle-income country (currently USD 4,085 GNI per capita) ¹⁰ – a stage that Angola, Botswana, Gabon, Mauritius, Namibia, Seychelles, and South Africa have reached, some of them aspiring to ambitious industrial goals. ¹¹ All other things being equal, they could just exploit the remaining difference between the stylised historical peak and their current levels. The small class of industrial entrepreneurs in those countries will, ceteris paribus, not expand anymore. Entrepreneurial spirit would have to be redirected to the provision of services – touted anyway as the future in tech hubs all around Africa. In the less upbeat analysis, premature de-industrialisation is a key feature of the so-called “middle-income trap” – a loosely defined pattern of stalled development, initially discussed for countries belonging to this income group in Latin America and Asia.

As we are not quite at the entrance of Dante's inferno, however, we should not abandon all hope. Some of the same sources attributed part of the premature de-industrialisation to the outcome of premature liberalisation – that is, to bad policy: the dubious side of structural adjustment. The whole cloud of country cases could have shifted towards the upper right, as industrialisation is not a zero-sum game. Things can be reversed – we just do not know to what extent. If we take a second look at the facts reported, first at the shifting industrial path, we can observe a relatively broad spread around the stylised, inverted U-curve, with many countries way above the hump even in 2010 (the statistical scattering being barely reduced). Also from this angle, it is obvious that the positioning on the curve can be positively influenced by good trade and investment policy. Altogether, the bulk of sobering new evidence on African growth and employment amounts to a massive plea for the necessity of consequential, sector-specific economic policy. ¹²

Social Prospects

Finally, a significant change of mind in the social arena is being recorded as well. In the international debate, we observe the same triple jump as with the economics of Africa. The late 1980s and 1990s saw an emphasis on poverty reduction in a sea of misery. Albeit initially contested by African elites, the focus on poverty became well established with new concepts, data, policies, and aid. Then, at the dawn of the millennium, many discovered the blossoming of a “new African middle class.” Definitions of that class diverge enormously – ranging from per capita, per day earnings of USD 2 all the way to a minimum of USD 10–15 – and so do estimates on the actual size of the new middle class, indicating that prospects might be less bright than many previously thought. ¹³ Now we are at the third stage. There is growing recognition among analysts and institutions that, definitely, “the growth was not inclusive” ¹⁴ in Africa, with a robust causal link at least to the disheartening youth employment prospects. Along these lines, scholars and development practitioners today are paying considerably more attention to problems of social inequality, and not just in Africa. The not-quite-official new World Bank position with the twin goals of complete eradication of absolute poverty by 2030 and some “shared growth” for the bottom 40 per cent of developing-country populations encapsulates the shift well. In operational terms, it is most evident for social transfers.

Direct social transfer programmes have been around for quite some time, but for those who followed the cautious debate among donor agencies from early cases in Zambia or Malawi on cash transfers and the conditions under which they might work may be surprised to note the new attitude towards not only conditional cash transfers (CCT), but even unconditional cash transfers (UCT), youth wage subsidies (as foreseen in South Africa), and other social transfer schemes (World Bank 2013). This is manifestly influenced by successful empirical cases outside Africa – namely, Mexico's Progresa/Oportunidades and Brazil's Bolsa Família, positive evaluations of which contributed to the spread of such programmes. In particular, it was found that transfer programmes themselves contribute to private investment and growth, reducing their own cost over time. It is not so easy to prove that the opinion shift in social policy and the more realistic economic analysis not only occur in tandem but are, in fact, causally linked. The link is rarely made as clear as in a recent policy proposal for resource-rich countries in Africa. Most of them are not translating their mineral wealth into mass employment and shared growth, as seen above. Authors, again from the World Bank, convincingly argue that direct payments from resource dividends can reduce poverty through increased private consumption and provision of public goods. ¹⁵

Such a programme would take care of one half of a well-known transfer problem in the many resource-rich countries of Africa – transfer of mineral revenue to a broad layer of private households. The other half is transfer from the same extractive industry to other industries – agriculture and manufacturing, typically marginalised in oil-producing countries. Though badly needed, not only for resource-rich countries, productive subsidy schemes to these two core sectors remain suspicious for the same institutions now excited about social subsidies. Neither targeted agricultural subsidies (e.g. for seeds and fertilisers) nor East Asian–style selective industrial promotion generate remotely the same enthusiasm. A World Bank Africa chief economist believes in unconditional cash transfer to needy people rather than in targeted industrial policy. In sum, a strange sort of new Washington Consensus seems to be emerging for Africa, comprising three elements:

The first element clearly distinguishes this third-stage perception from the Afro-pessimism of the first stage: Africa's growth record as such is not questioned. All three elements recognise that dramatic economic progress is underway in Africa, and that there is an emerging core of modern sectors, growing domestic entrepreneurship, and economic self-assertion on the continent as well. Yet, this alone might not suffice. The new bundle of perceptions leaves a crucially important twin gap, as new strategies are needed for the next generation of African policymakers (a) to accelerate growth and the emergence of domestic entrepreneurship in the fledgling sectors of their economies, paying special attention to the manufacturing and services-related components of industry, and (b) to radically alter the path of population growth. These issues have the potential to fuel another big debate on African soil.