Investment, Wage Goods, and Industrial Policy
Investment, Wage Goods, and Industrial Policy
Abstract and Keywords
Raising and sustaining long-run growth rates is made more difficult by the complexity of economic growth and by the complexity of growth theory debates. Nonetheless, the investment rate is central to long-run growth and development. Growth sustained by high investment rates will also involve structural change: a shift of resources into high-productivity economic activities. This chapter combines discussion of investment—why it matters, what economic policies help to raise investment rates and keep them high—with discussion of ‘industrial policy’. But the terrain of industrial policy has expanded to take account of new high-productivity activities, of servicification, and of agribusiness; policy officials thus need to refine the criteria used to make resource allocation and incentive decisions accordingly. A particularly important political economy constraint on investment rates is the non-inflationary supply of wage goods.
What can African policy officials do to promote sustained improvements in output, productivity, wages, employment, and welfare (and how can they ‘catch up’ with the advanced capitalist economies)? If investment is crucial, then how can the investment rate be raised? And is investment in some economic activities more important to the broader economy than others? Put another way, what are the greatest constraints on growth, and are certain types of investment less effective than others in promoting wider development?
Raising the investment rate is a necessary but insufficient objective if a society is to accelerate the development of its productive forces: the tools, machinery, sources of energy, and premises used in production, as well as the productive capabilities of the labour force (skills and know-how). The ‘efficiency’ of investment—its productivity—matters too, though it is not the overriding factor some economists have argued it to be and their measures of productivity do not make sense. Also vitally important is that policy officials identify which activities investment incentives should encourage—specifically, those most clearly associated with increasing returns, spill-overs and linkages, direct and indirect generation of employment, and generating foreign exchange.
We further argue that while supply-side issues—and institutional issues that may facilitate or hold back the development of productive forces—do matter, they have been given excessive emphasis in most of the economics literature. Instead, we emphasize the overriding importance of demand.1 This in turn involves a discussion of what forms of demand are important and how they influence investment. One of these forms is external to a national economy: export demand. The dynamics of structural change in low- and middle-income countries typically generate a ‘balance of payments constraint’ that holds back whatever growth rate it is possible to sustain. An important determinant of developing countries’ ability to sustain high growth rates is how effectively their governments address this constraint: that is, whether they can sustain a rapid rate of increase in imports. (p.76) Chapter 5 takes up this issue. However, there is another type of demand that is important for welfare as well as being a powerful influence on investment levels: domestic demand for wage goods (especially food, but also other basic consumer goods).
We not only reject simple distinctions between macro and micro levels of the economy; we also reject the orthodox neoclassical economic argument that there are ‘micro-foundations’ of the macro-economy (i.e. that the macro-economy is just the aggregate of choices made by a large number of individual ‘agents’).2 ‘Society does not consist of individuals,’ Marx argued, ‘it expresses the sum of connections and relationships in which these individuals stand.’3 Similarly, we argue that an economy is the expression of a huge array of relationships (both domestic and global). Investment drives savings more than the other way round, but investment itself is shaped by a range of factors, including the distribution of income nationally and internationally, the state of productive forces (including levels of education and skill formation), and the level of demand. While investment decisions themselves may be ‘micro’, they are determined by macroeconomic balances and trends, as well as by institutions and policies that affect relative profitability across different activities, how restrictive monetary and fiscal policy are, the ease of capital flight, and so on.
An oft-repeated view guiding many policy recommendations throughout Africa is that savings are a prerequisite for investment. That is, low- and middle-income countries will not be able to achieve more investment unless they build up a pool of domestic savings to add to whatever foreign financial flows can be attracted. The implications of this apparently common-sense statement are wide-reaching and contentious. We contrast the ‘savings lead to investment’ approach, which is rooted in neoclassical economics, with another and—we argue—clearer way of thinking about growth and development, based on the thinking of economists such as Kalecki, Kaldor, and Keynes. Naturally, a different way of understanding the dynamics of growth and accumulation leads to a different way of thinking about how countries can improve investment, productivity, wages, employment, and welfare. This in turn leads to very different policy implications.
One question is how important it is to generate a high savings rate before and in order to secure rising levels of investment. Another is whether it matters where that investment is directed. The answer that emerged from a great deal of neoclassical growth theory was that, given diminishing returns, it is irrelevant which sector investment flows into. Another common argument is that rather than focusing on manufacturing or industrialization, better gains can now be (p.77) made by supporting the service sector, particularly activities such as information and communications, as well as tourism. According to this view, low-income developing countries may be able to skip the long process of industrialization—which has historically been socially disruptive and highly exploitative, as well as environmentally damaging—altogether.
In reality, the evidence from many countries’ experiences of rapid economic growth in the modern era suggests that the sectoral concentration of investment really does matter. Historically, there has been something ‘special’ about manufacturing, with a high rate of investment in the sector particularly important to overall growth and development. In short, manufacturing has been and is an ‘engine of growth’. We explain why.
In recent years, many neoclassical economists have ‘rediscovered’ industrial policy, with the result that it now means different things to different people. We do not agree, as others might, that there is a ‘convergence’ of views among economists of different traditions regarding appropriate industrial policy. We clarify why there are differences: for some economists industrial policy is only sensible if implemented in ways that stick close to the principles of neoclassical theory, in order to fix or correct market failures. For others, any sufficiently ambitious industrial policy involves significant departures from the orthodoxy.
Just as industrial policy became fashionable again, however, a new fear arose that a twenty-first-century trend—premature deindustrialization—will render manufacturing irrelevant as an engine of growth, employment, and structural change. Premature deindustrialization happens when the share of manufacturing value added (MVA) in gross domestic product (GDP) begins to fall at levels of income per capita significantly below those at which this shrinkage started historically in the now industrialized economies. We discuss this issue in Section 4.3 and argue that such evidence is not a reason for governments to give up on industrial policy.
4.2 Investment, Savings, and the Wage Goods Constraint
Investment Matters: ‘Even a Skilful Cook Cannot Make a Meal Out of Nothing’
It is not possible to generate sustained growth, structural change, wage and employment growth, and welfare improvements without raising the level of investment relative to national income (the investment ratio). Of course, there can be wasteful, ineffective investments: capital sunk in wayward schemes or frustrated by demoralized management lacking adequate technical know-how. Also, economic development and structural change involves a progressive shift of resources into a diverse range of higher-productivity activities. Even so, it remains (p.78) the case that long-run growth and structural change require substantial and sustained levels of ‘capital deepening’.
This point can get lost amid the complex debates economists have about growth. An influential but implausible idea was that the historically unprecedented East Asian economic growth surge in the 1960s and 1970s could have been even faster if fewer inputs (including capital) had been used, less wastefully. Some argued that growth was a function of the accumulation of capital and not in any significant way a reflection of productivity gains. Assuming diminishing returns, growth in Singapore, South Korea, Taiwan, and elsewhere would, supposedly, fizzle out, a victim of inefficient over-accumulation, rather like what had been observed in the Soviet Union. Others argued the opposite: that East Asian growth was the result of assimilating improved technologies from overseas and closing the ‘ideas gap’. However, these critics used essentially the same kind of model, just using different estimates of variables. Both sides kept faith with the idea that free markets played a leading role in the countries’ ‘miraculous’ growth.4
Behind this debate lay the history of growth theory. Thinking about what determines rates of growth and why they differ over time and between countries has long been a source of controversy. For many decades, perhaps the most influential approach among orthodox economists was the neoclassical production function. In an aggregate production function, output is a product of the factors of production (essentially capital and labour) and an estimate of their overall productivity, or ‘total factor productivity’ (TFP).
TFP estimates play a significant role in many analyses of growth in Africa. For example, a simple production function, with a ‘productivity growth residual’, lay beneath a major growth accounting exercise applied to 18 African countries in the mid- 2000s.5 If capital and labour inputs are accounted for, then any growth over time not accounted for by changes in these factor inputs is a ‘residual’, taken to represent TFP. Contributors to an African Economic Research Consortium (AERC) project on growth argued that TFP growth had been very slow in Africa, and that this could be addressed if more African countries adopted ‘syndrome-free’ policy regimes. This meant not only avoiding the syndrome of state failure, but also other syndromes, such as splurging natural resource windfalls on infrastructure or introducing ‘state controls’, which involved intervening in markets and ‘distorting’ prices.6
The many difficulties in collecting, say, reliable demographic data or working out the level and rate of growth of output, pale beside the problems associated with TFP data. Growth accounting exercises that focus on estimating TFP rest on assumptions that are at the core of neoclassical economics in its purer forms: (p.79) perfect competition, full employment, constant returns to scale and diminishing returns to individual inputs, and perfect, easy substitution between capital and labour. Aggregate production functions imagine a whole economy as a single sector. But even if trying to work out TFP in terms of a single output—say, wheat production in the Punjab, coal in South Africa, or steel in South Korea—this is an exercise that will produce meaningless results, given the extremely varied outputs and qualities of each product.7
Typically, explanations for perceived patterns of TFP have nothing to do with the underlying model, and, in fact, often violate the very assumptions on which the estimates are made. In some cases, wildly unrealistic explanations are given. The debate about East Asian economies was kicked off in the 1990s by a paper suggesting that Singapore was a free market economy (conforming to the model’s assumption of perfect competition). We know this is far from true. A TFP account of South Korean steel is ‘riddled with insoluble theoretical and empirical problems that conceal rather than reveal the nature of the South Korean steel industry’.8 Estimates of TFP trends in the South African coal sector are similarly misleading, constructed out of an imaginary model that, rather than being a technical simplification, is a complete misunderstanding of the sector. The market for coal in South Africa historically has not been perfectly competitive, but instead highly concentrated. One reason this matters is that the structure of the coal sector and its political links have shaped trends in investment over time. A tight coordination between a very limited number of mining (and energy) firms and the state, and the apartheid state’s institutional exploitation of labour, underpinned the expansion of the coal sector before political uncertainties and foreign disinvestment broke apart the ‘political settlement’ in the 1980s.9 At the aggregate level, it is fairly easily shown that the history of the modern Zambian economy cannot be understood in terms of the perfect competition, constant returns, and full employment assumptions deployed in growth accounting models. In cases where these conditions do not hold (and they have never held in Zambia), then it becomes unclear what it is that estimates of TFP are actually supposed to be measuring (and this not even mentioning problems with the underlying data).10
Although many well-known economists promoted efforts to quantify the sources of East Asian growth using measures of TFP, they paid no attention to the profound conceptual and empirical problems with aggregate production functions.11 A review of their work concludes that aggregate production functions ‘are a dead end because they ask the wrong question, which they try to answer with the wrong data and because the model that underlies these approaches can never be refuted empirically’.12 Unfortunately, there is a huge amount of research on (p.80) African economies that follows these East Asian examples, ignoring inconvenient methodological difficulties when attempting to explain growth.
Orthodox growth thinking later followed two different, if related, tracks. One of these built up the growth accounting approach used in earlier models, but sought to include an ever-increasing number of variables to ‘explain’ the remaining residual (i.e. the portion of growth outcomes not easily explained by the shorter list of variables). In Africa, this took the form of a flurry of growth regressions seeking to explain the difference between average growth rates over time in Africa compared to the rest of the world, as captured in a dependent variable (the variable to be explained) known as the ‘Africa dummy’. This voluminous literature also reached a dead end,13 with a key figure in neoclassical growth theory remarking that this growth regression literature was ‘statistically unprepossessing’.14
The other track was the development of ‘new’ or endogenous growth theory. This theory recognized that increasing returns do exist and are significant. In reality, economists had really known this but found it difficult to deal with technically. One of the key figures in neoclassical economics, Alfred Marshall, explained that there was both a Law of Diminishing Returns and a Law of Increasing Returns: ‘while the part which nature plays in production shows a tendency to diminishing return, the part which man plays shows a tendency to increasing return’. But the latter tendency faded into relative insignificance through successive editions of his Principles of Economics.15 Much earlier, Adam Smith also recognized the importance of increasing returns, but he too had turned away from them, to the point that Nicholas Kaldor would later quip that economics went wrong after Book 1, Chapter 4 of the Wealth of Nations, when Smith dropped the assumption of increasing returns.16
Thus, having for many years evaporated from economics, increasing returns made a comeback in ‘new’ growth theory models, which attempted to account for the origins and persistence of imperfect competition and make analytical sense of sustained divergence between economies. If increasing returns matter, and are a source of imperfect competition (because, for example, ‘first-mover’ firms build a competitive edge through reaching a scale that competitors may take a long time to achieve, if at all), then investment in economic activities with scope for increasing returns matters more than it does in models assuming constant or diminishing returns. Furthermore, states may play an important role in nurturing such investment in larger-scale innovative production.
Unsurprisingly, there is huge debate about the usefulness of new growth theory, with the most obvious critique being that it is not really new at all but is ‘old wine in new bottles’. After all, economists outside the neoclassical orthodoxy had long (p.81) argued for the importance of increasing returns in economic growth. Furthermore, critics argued that these non-orthodox economists—such as Kaldor—were more realistic in their ideas about growth as a process of cumulative causation, proceeding through tensions and disequilibria, rather than a matter of markets working themselves out through constant adjustments to a state of equilibrium (see Chapter 6 for a discussion of equilibrium and balance in development economics). Heterodox economists charge that ‘new’ growth theorists grafted a partial recognition of reality (increasing returns, imperfect competition) onto a core model still rooted in the same infertile soil, that is, the methodologically individualist, equilibrium-based theory of neoclassical economics.
By the 2000s, some of the protagonists of orthodox economic growth theory finally began accepting the obvious: that it is extremely difficult to understand the finer processes of economic growth. William Easterly, for example, who had previously argued that high scores on an index of ethno-linguistic fractionalization was a key reason behind ‘Africa’s growth tragedy’ wrote: ‘in reality, high-growth countries follow a bewildering array of paths to development’.17 Other economists aimed ‘to be comprehensive in including potential explanatory variables: firm characteristics, geography, infrastructure, access to finance … We offer a menu of facts and explanations’ in the hope of explaining ‘Africa’s disadvantage’.18 Growth, then, becomes something of a hodgepodge recipe; a bit of market openness with a dash of good governance thrown in along with various other ingredients. This approach does at least accept the importance of investment, with a recent analysis by International Monetary Fund (IMF) economists of the long-run data for sub-Saharan Africa stating that ‘investment seems to play a strong role in warding off growth stops and sustaining growth spells, as it not only supports aggregate demand but also expands productive capacity’.19
Growth is about ‘accumulation and innovation’, with innovation—technical change—itself embodied in the capital stock (for example, investment and innovation combine in a new and faster pattern-cutting machinery in the textile industry). The long-run evidence analysed by Maddison shows an ‘impressive growth of non-residential gross capital stock per person employed’ in France, Germany, Japan, the Netherlands, the UK, and the USA between 1890 and 1987.20 What is especially striking in his historical data is how the ‘followers’—those countries, above all Japan, who were catching up to the UK (the first economic leader of the modern capitalist era)—had the fastest growth over time in capital (p.82) stock per employee. Such historical experience itself embodies what is called ‘the advantages of backwardness’.21
Maddison focused on the proximate causes of growth experiences, which he distinguished from the less easily quantifiable and generally more debatable ultimate causes (including institutions, ideologies, pressures from interest groups, historical accidents, and variations in national policy). His explanation of episodes of growth, convergence on the leader, and slow-down across these advanced capitalist economies from the late nineteenth to the late twentieth centuries emphasizes that ‘the movement of capital stock provides a much more powerful degree of explanation for growth acceleration and slow-down than … the analysis of labour input’.22
The post-Second World War history of ‘catching up’ development confirms the point. A comprehensive review of the most rapid experiences of economic development found: ‘Capital accumulation holds the central place in East Asia’s virtuous growth regime.’23 Accumulation—the share of investment in GDP—rose from 10 per cent in Korea and from 16 per cent in Taiwan in the 1950s to more than 30 per cent in the 1980s; Southeast Asian economies such as Indonesia, Malaysia, and Thailand began to raise their investment-to-GDP ratios in the 1970s and kept them high; while China sustained a rise from an already significant 25 per cent ratio in the early 1970s to more than 35 per cent in 1995–2000.
One argument in favour of pushing for a high investment rate is that ‘even a skilful cook cannot make food out of nothing’.24 All too often, however—in many Latin American and African countries, for example—investment has either tailed off too soon or gone into speculative urban construction (rather than productive infrastructure). Furthermore, in a context characterized by an acute balance of payments constraint, investment has often failed to promote a sufficiently fast rate of growth of export earnings. Thus, a fundamental issue is what alternative returns are available to those allocating investment finance—whether it be in tax havens abroad, real estate, the financial sector, or in manufacturing—and how policy affects these relative returns.
Savings and Investment: The Mainstream View and the Fate of Financial Liberalization
How, though, are low- and middle-income countries to raise and sustain a high rate of investment? How are they to finance investment? In this section, we argue (p.83) that demand is a far more important driver of sustained investment than is often acknowledged, and that levels and patterns of demand are shaped by institutions, as well as historically specific social relations and politics. This is contrary to the typical emphasis of economics texts, which emphasize supply-side sources of investment and growth.
Given the assumptions underpinning growth accounting exercises and the measurement of TFP (discussed earlier in this section), it is unsurprising that the standard policy recommendations are designed to: bring economies ever closer to the ideal of perfect competition (market deregulation, privatization); encourage a proliferation of small and medium-sized, price-taking enterprises (entrepreneurship programmes, access to finance schemes, micro-credit facilities); and encourage frictionless switching between capital-intensive and labour-intensive production in moves towards full employment (greater ‘flexibility’ of labour markets without minimum wages, legal constraints on hiring and firing, and/or powerful trade unions). A key market that, if sufficiently liberalized, is supposed to funnel household savings into investible funds is the market for finance.
Mainstream economists argue that economies need to build up both public and private sector savings in order to create a pool of investible surpluses available to private enterprises through the financial system. If the banking system lends these savings at interest rates set by market transactions and conditions, this will help to weed out unproductive borrowers incapable of generating high enough returns to meet repayment obligations. In other words, market-determined interest rates will lead to efficient investment, while institutionally regulated (or repressed) interest rates may lead to indiscriminate and wasteful investment. According to this view, it is best to let market prices determine interest rates and stimulate a gradual increase in savings, which will in turn lead to a gradual rise in efficient investment. As one summary of the influential McKinnon–Shaw argument for financial liberalization claims:
introducing market principles and competition in the banking sector increases interest rates on deposits. These higher interest rates lead to higher saving and investment rates, ultimately contributing to higher growth rates … increasing competition puts pressure on the profit margins of banks, in particular on the interest rates demanded for loans. This reduces the cost of debt, leading to a rise in investment and growth.25
The empirical evidence does not confirm a causal relationship between financial liberalization and growth.26 Rather, data on GDP per capita growth rates in sub-Saharan African countries from 1950 to 2016 show that financial liberalization is (p.84) linked to episodes of growth deceleration. Neither is there any evidence linking fewer financial restrictions with growth acceleration elsewhere in the world,27 with financial liberalization in several Latin American countries, for example, followed by a string of domestic financial crises, including the Mexican ‘Tequila Crisis’ of 1994–5.28 Additional costs associated with financial liberalization include a higher frequency of financial crises, more volatile flows of capital, and the social costs associated with ‘financialization’.29
In the wake of the global financial crisis, some of the problems of overenthusiastic ‘financial globalization’ were reluctantly recognized even by economists employed by the IMF.30 As a result, a new literature developed in largely mainstream quarters, arguing that while the direct benefits of financial globalization were ‘elusive’, there were indirect benefits. However, these indirect benefits only tended to kick in when economies had passed certain ‘thresholds’ of financial depth, institutional preconditions, and level of per capita income.
Investment and Savings: A More Keynesian/Kaleckian View
There is another view, usually associated with Keynes and Kalecki, that investment generates the savings needed to cover its financing over time.31 Productive investment creates wage incomes and profits, some of which are saved—by employees depositing savings in banks, by firms retaining profits or banking surpluses.
Again, the East Asian experience, where a high and rising investment-to-GDP ratio preceded rising savings rates, is instructive. In terms of macroeconomic balances, there was an ‘ex ante excess of investment over savings’.32 That investment was anything but gradual and market-led, instead being the product of determined state intervention that amounted almost to ‘forced investment’. East Asia may be exceptional in many ways, but in the basic fact of investment surges driving rising savings there is nothing unusual about the region’s experience. Elsewhere, for example in Mexico between the 1940s and late 1970s, firms made investment decisions based on judgements of expanding market size and ‘independently of the availability of finance’.33
Even if there are investible savings generated domestically, they are often not invested, becoming what Albert Hirschman called ‘frustrated savings’.34 After (p.85) financial markets and the capital account were deregulated in South Africa, for example, there was a dramatic acceleration of purchases of short-term financial assets (by both domestic and foreign speculators) to achieve quick capital gains, while flows into productive investments dwindled.35 Enormous sums of African savings have been squirreled away abroad and in tax havens, or held by large corporations but not invested productively within African economies. Between 1970 and 2015, capital flight from 30 sub-Saharan African economies reached huge proportions, according to one reckoning much larger than the capital inflows of official development assistance and foreign direct investment; and capital flight has been accelerating.36 These authors suggest that the main motive for capital flight from Africa was to evade taxation or the seizure of illicitly acquired assets; but this suggestion risks underestimating the significance of demand. If savings are to be invested—and corporate profits reinvested in Africa—there must be confidence in expanding demand.
Angola and Ethiopia provide examples of this dynamic. In Angola, following the end of the country’s war in the early 2000s, rising oil prices and Chinese demand did create opportunities for illicit asset accumulation and capital flight.37 But there were also incentives to invest domestically because of oil-fuelled expectations of rising middle-income consumption; and high oil prices also sustained the government’s ability to spend on urban housing and infrastructure projects as part of its strategy for post-war reconstruction and political consolidation. The infrastructure and housing boom had clear linkage effects, leading to the emergence of a building materials and cement production industry (and associated employment) within Angola. Rising Angolan middle-class incomes in the wake of rising export earnings prompted investment in, for example, the brewing and beverages sector. This in turn led to linkages to bottling and canning facilities in Angola, some of which grew large enough to export to other African economies.38 In Ethiopia, a large-scale government programme to construct middle-income urban housing was the demand spur that led to a domestic cement industry (and then later expansion of foreign-owned cement production as well) and to the proliferation of domestic building materials and services firms.39
These experiences confirm the views of economists working in the tradition of Keynes and Kalecki. One of these, King, for example, highlights the proposition that:
The relationship between aggregate investment and aggregate savings is fundamental to macroeconomic theory, and causation runs from investment to saving, and not vice versa.40
(p.86) It bears repeating that there is a great rift between those who think savings ‘lead to’ investment and those who think investment can lead to and be matched by savings. Most policy advice to developing countries is based on the former (neoclassical) view and ignores the intellectual revolution in economics during the 1930s:
Keynes’s intellectual revolution was to shift economists from thinking normally in terms of a model of reality in which a dog called savings wagged his tail called investment to thinking in terms of a model in which a dog called investment wagged his tail labelled savings.41
However, the econometric evidence on the relationship between savings and investment in 10 African economies suggests that this is complex, varying from country to country.42 It may be difficult to make simple and universal statements about causation for several reasons: first, there is a ‘chicken and egg’ problem: a virtuous cycle of investment generating and being matched by savings (in macroeconomic accounting, savings must equal investment), which in turn allows for more investment, has to start somewhere. This often requires high borrowing (including borrowing abroad), directed credit, and government spending, and can involve rapidly creating surpluses out of exploitation, wage repression, and forced or deferred savings. Second, savings do not instantly and automatically appear to rescue macroeconomic balances and match investment. Policy measures are important in mediating this. And, third, the impetus to sustain investment can peter out. Initial sources of financing may evaporate if the balance of political forces changes such that profits are squeezed by high wages and demand for more consumer goods is asphyxiated by high costs of living, or if broad consumer demand fails to grow because of extreme income inequality, as has been argued is the case with regard to sluggish manufacturing growth in Angola.43 Flows of concessional finance from abroad are often volatile and driven by geostrategic and political forces.44 The animal spirits (as Keynes called them) of both African and foreign investors flicker and fade for various (licit and illicit) reasons, but often result in massive shifts in the volume of capital flight. This third issue is obviously political, involving struggles over the distribution of surpluses between profits and wages.
The State and Uncertainty
How, then, does it come about that the investment ratio is raised, increasing aggregate demand and setting off a cumulative process of growth? Firms, both (p.87) domestic and foreign, will be eager to invest if they can identify opportunities in a large and growing market: demand propels investment. In the prevailing uncertainty of capitalism, strong evidence of demand and an expanding market helps create ‘animal spirits’—the herd-like behaviour characteristic of investors, expressed in this case in the confident expectation that market demand will be buoyant. Keynes emphasized the role that hunches, anxiety about the actions of others, and emotions play in investment decisions. More recent and fashionable work in ‘behavioural’ economics explores bounded rationality and systematic ‘biases’ and heuristics (what earlier economists might have called rules of thumb) in economic behaviour, but much of this work effectively makes modest tweaks for ‘irrationality’ on largely unaltered neoclassical models and principles.45
However, the whole point about low-income economies is not only that the supply of factories and new machines is low, but also that the absolute level of aggregate demand is low. Furthermore, growth in demand is often faltering. In middle-income countries the market for many goods and services might be larger, but all too often the momentum of growth in aggregate demand stalls. Investment decisions in developing countries rely not only on the choices of individual capitalists or firms, but also a on a number of other factors, such as how liberalized capital markets are, the pattern of income distribution, and how institutions and infrastructure function. Additionally, how societies are structured—for example, marked spatial, racial/ethnic, or gendered inequalities and the strength of trade unions—shapes patterns of effective demand that has consequences for investment.46 In these circumstances, sustaining high investment ratios has typically involved a leading role for the state.
States can also have a powerful role in sustaining investment in order to avoid the sudden collapses in growth that some economists have come to emphasize as being a critical threat to sustained structural change. Economic historians argue that
The improvement of economic performance over the long run has occurred primarily because the frequency and rate of shrinking have both declined, rather than because the growing rate has increased.47
If public sector investment fails to grow, as in Latin American economies after the 1982 debt crisis, then both investment-to-GDP ratios and growth will stagnate.48 In several East African economies, policy-induced declines in demand between the 1980s and mid-1990s, signalled by massive reductions in real wages in the public sector, were also associated with collapses in growth. Across a wider group (p.88) of sub-Saharan African countries it has been argued that lower government expenditures (after signing agreements with the IMF) had the effect of reducing growth and employment while widening inequality. Where policies were introduced to restrain domestic demand without the direct intervention of the IMF, as in South Africa since the mid-1990s, the economy remained ‘stuck in low gear’.49
If the state is the central economic actor in the drama of late development, even more so than in advanced economies, one important role it must play is in stimulating and coordinating investment.50 Aside from the issue of using state-owned enterprises to achieve these ends, this includes managing monetary policy, fiscal policy, and development finance: that is, organizing macroeconomic policies and a financial system around the goals of sustained long-run economic growth and structural change. These goals have to be defined in terms of growth, productivity, employment, and the balance of payments. That, however, is completely contrary to the advice that has been meted out to the central banks and governments of developing countries since the late 1970s or so. African governments who wish to avoid growth collapses are still subjected to conventional policy exhortations—made to adopt monetary policy geared towards low inflation, reduce market distortions, and improve the business climate—by some IMF economists.51 Such economists appear to shrug off the doubts that the IMF’s own Chief Economist expressed in 2011 about this skewed advice, that mainstream economists and policymakers had ‘convinced [themselves] that there was one target, inflation. There was one instrument, the [interest] policy rate. And that was basically enough to get things done.’52
Development Finance: Central Banks, Development Banks, and Government Spending
The common-sense view dismissing development banks has been so powerful that it can easily seem as if such banks are an historical oddity, a diversion from the norm of prudent and sensible banking. A review of the history of debates on financial systems from the early twentieth century onwards finds that the neoliberal way of thinking about banking, central banks, and monetary policy is actually the exception.53 A sensible approach is now said to involve keeping central banks out of a government’s hands (‘independent’) so that the former can focus on holding inflation as low as possible, and this approach is part and parcel of a belief in the virtues of austerity and a commitment to curbing demand. It also involves deregulating capital markets, making it harder for governments to kick-start a higher investment ratio thanks to the increased difficulty of financing investment (p.89) projects through central bank policy. In this scenario, there is no incentive for central banks to limit the flow of investment into shopping malls and luxury condominiums and instead promote investments targeting structural change, increased employment, and rising productivity.
The range of instruments that central banks can use, and have often used, is in fact far wider than simply tinkering with short-term interest rates in the service of low inflation. Rather, central banks can play many other roles, both through their links to development banks and through their own policies and practices. Many developing countries, and in the earlier part of the twentieth century also the UK, US, and European governments, took a very different view of central—and development—banks. Particularly after the Great Depression, central banks came to be major agents of economic development, and there is a long and richly varied history of state-owned ‘development’ banks. This evidence largely dispels more recent myths claiming that state-owned banks are inevitably inefficient.54
There are examples of lax and corrupt central bank regulation of the banking sector that have undermined investment efficiency and political stability, including in countries where reforms have been introduced to make the central bank more independent.55 On the other hand, central banks in many countries have, among other things, successfully introduced capital controls to reduce exposure to volatile surges in inflows and outflows of capital; made low-cost capital available to specific development institutions; used differential discount rates to allocate credit to large, priority capital projects; and imposed reserve requirements that fund government bonds. Development banks, meanwhile, have been used to provide concessionary loans, often with lengthy repayment schemes, to stimulate investment in new manufacturing activities.56
Central banks have also frequently got prices wrong—anathema to the neoclassical economist—by engineering negative real interest rates, again to stimulate borrowing for investment and to make it easier for governments to use deficit financing to push for higher investment rates.57 Governments often have to spend more than they receive in tax revenue in order to bring about higher investment rates, the aim being to sustain a high level of demand rather than restrict it. To re-emphasize a key point: what might make sound economic sense for a household does not make sense for an entire economy (see Chapter 3).
Of course, government deficit funding can also result in wanton spending and wasteful use of extremely scarce resources, for example in accommodating military adventures or providing for handouts to special interest groups in order to (p.90) preserve political power. An obvious example is the post-independence history of Zimbabwe, a country whose sizeable formal sector, well-developed civil service, and initially successful record of post-independence service delivery seemed to mark it out from hyperbolic generalizations about weak states in Africa.58 Ultimately, though, it suffered catastrophic economic decline, rapid deindustrialization, shrinking agricultural output and exports, currency collapse, and periods of hyperinflation.
There have been different phases of the Zimbabwean crisis but what unites them is a political dynamic: the ‘violent intolerance’ of the leadership of the ruling Zimbabwe African National Union—Patriotic Front (ZANU–PF) in attempting to remain in power.59 This commitment was pursued through institutional and policy measures that made large numbers of people with access to (but not title over) land, minimum-wage protected jobs, and civil service positions dependent on the central state. Institutions were increasingly politicized and in many cases militarized.60 Meanwhile, an increasingly dominant group of rentiers accumulated private wealth through the state while failing to play the role expected of a productive capitalist class.61 In the 1990s, a poorly designed and implemented enhanced structural adjustment programme (ESAP) was championed by domestic elites (not imposed by the IMF and World Bank, which, whatever their failings, insisted on the inclusion of a social programme, against the instincts of the government). This paved the way for even faster deindustrialization and further opportunities for personal wealth and rent accumulation. By the end of the 1990s it had become obvious the government could or would not control its deficits and confidence in the economy plummeted, with the currency collapsing in late 1997. Further economic mayhem continued throughout the 2000s, encapsulated by, among other things, unbudgeted pay-outs to war veterans, an economically misjudged land reform, and a military excursion into the conflicts of the Democratic Republic of the Congo (DRC).62
But if—by contrast with the Zimbabwean experience—‘excess’ demand represents investments in activities and capital equipment that raises productivity (and foreign exchange earnings), then it makes no sense to rein in demand. A striking example of a government refusing to repress demand and presiding over a sharp increase in net borrowing from the rest of the world is Norway in the 1970s. Norway encouraged a rapid surge in investments in capital equipment to facilitate increased productivity in the country’s new oil sector. Thus, policy officials can choose either to follow Norway’s example by provoking a supply response when (p.91) macroeconomic deficits develop; or they can react how the orthodoxy of the past forty years advises, by repressing demand.63
The most important point here is how policy officials can bring about a rapid increase in investment levels. One view is that positive high real interest rates will make it possible for the more efficient private sector investors to borrow and invest. If this leads to an external imbalance (because undirected new investment may lead to imports rising faster than export earnings), then foreign investment should be encouraged to cover the external gap or exchange rate adjustments should be made to send more appropriate signals to investors. That is largely the approach the South African government embarked on after the first post-apartheid elections, when it set out to (and for a while managed to) reduce the fiscal deficit, meaning the government no longer competed for savings with the private sector and inflation could slow down. However, real interest rates remained very high, growth was slow, and private investment—both domestic and foreign—dismally failed to respond in the manner predicted by a large chorus of mainstream economists.
The alternative is to argue that government deficit spending is central to boosting the investment ratio, especially because private sector enterprises are so reluctant to invest in low-income and slow-growing economies. Their reluctance is perfectly reasonable. The market is relatively small and often stagnant, and, while growth episodes happen, they are typically short-lived. Uncertainty is pervasive and the risks investors face are very high. Additionally, the costs of investment are high as it is difficult to get inputs and costly to install equipment; infrastructure is either absent or creaky; and there are insufficient skills and technical know-how in the economy. Particularly in more open economies, it is likely to take some time before domestic enterprises can compete effectively in international markets, and in some situations private investors may be wary of bearing the cost of making production more profitable if others can cash in on their efforts without absorbing some of the cost, for example, when providing irrigation or graded roads. Thus, infrastructure becomes a critical issue: transport and power infrastructure shortcomings undermine investor confidence, but the large financing commitments required to improve them are usually off-putting for private investors.
While the private sector makes only a small contribution to financing investment in African transport infrastructure, investment expenditures on road transport absorb a relatively high proportion of total public sector investment in low-income African economies.64 Road construction costs per kilometre for activities undertaken between 2005 and 2007 are said to have been have been two or three times higher in some African economies than in other developing areas. Unfortunately, the costs of irrigation investments in sub-Saharan Africa are also (p.92) high compared to costs in South Asia and South East Asia.65 In such circumstances, political risks loom larger too, with governments having to absorb a great deal of the risk in order to provide collectively enjoyed goods such as rural electrification and irrigation.
An example of these dynamics can be seen in Ethiopia’s Upper Awash Valley, in which there is an ungraded road strung with sizeable farms on privatized land capable of extremely high-value output: fruits, flowers, and wine grapes. None of these enterprises has been able (or willing) to finance expenditures to improve the road, and despite repeated pledges by the regional government, nothing was done for years. This directly hits profitability and earnings, with a farm manager for a very large fruit growing and processing enterprise estimating that this business loses 12 per cent of the export value of packaged passion fruit puree because of damage done when the cartons are shaken up on the bad road.66
Investment, Savings, and the Question of ‘Crowding Out’
Governments do not compete against the private sector, shrinking a pool of savings that would otherwise be used to fund efficient private sector investments. Rather, the government may build up deficits through productivity-enhancing expenditures that generate incomes (and savings), raising profitability across all sectors and creating further scope for private sector investment, as well as incomes and savings, to rise. The World Bank’s most recent discussion of the relationship between public and private investment in Africa may be rather unsophisticated—based on crude cross-country statistical evidence, but they conclude that there is a significant positive association across the 26 countries studied—that is, public investment crowds in private investment.67
The same grudging acceptance that large-scale public expenditure can drive growth in low-income countries characterizes an assessment of the recent record in Ethiopia. Assuming unquestioningly that Ethiopia’s heterodox macroeconomic policies must have perverse, growth-reducing effects, a standard neoclassical growth accounting model reveals that ‘the negative growth effects of heterodox macro policies were quantitatively much less important than the positive growth drivers they helped to achieve’.68 It is tempting to rewrite this statement: the evidence suggests that heterodox macro policies have driven rapid growth in Ethiopia. Critics of a dominant role for the public sector, such as Western donors of development assistance, have responded to Ethiopia’s recent growth record (p.93) with an implausible counterfactual assertion: privatization, more foreign direct investment, and a smaller state ‘would have’ resulted in even faster growth.69
Ethiopia’s heterodox policies have included financial repression, directing credit at negative real interest rates towards public investment in infrastructure, inflationary finance, and an overvalued exchange rate. These policies, combined with a fiscal revenue-raising effort and public consumption restraint in the wake of the Ethiopian–Eritrean War (1998–2000), enabled dramatic public investment in areas such as roads and rail, energy, education, and health. These types of public infrastructure investment (rising to 18.6 per cent of GDP by 2011) were the main factor in Ethiopia’s growth acceleration, with real per capita GDP growth of 8.3 per cent per year between 2004 and 2014 higher than had historically been the case in the country, and higher than regional and low-income averages in the same period. However, despite noting the increasing consensus that infrastructure is central to growth and development, World Bank economists have insisted on the short-term view that Ethiopia’s public sector-led strategy has ‘led to a crowding out of many private investment projects owing to a lack of financing’.70 They also believe that whatever the possible short-run contribution of heterodox economic policies, these are unsustainable:
[W]e anticipate that macroeconomic policies face increasing trade-offs between public infrastructure improvements, private sector access to credit, and limits to restraining government consumption … Ethiopia’s growth rate is likely to decelerate in the future …71
A still cautious but rather different and more convincing assessment of Ethiopia’s macroeconomic balances suggests that the core approach may well be sustainable, as long as there is a greater commitment to generating rapid export growth.72
Mobilizing Savings, Managing the Growth of Consumption
If an investment push evaporates without generating incomes, profits, and savings, the balancing act will then have to involve contractions elsewhere or sharp increases in foreign borrowing (with strings attached). Thus, a government engineering a sharp increase in the investment level has to adopt policies which simultaneously ensure that investments are productive and that they result in rising real wages (as a source of savings as well as a source of demand). This may involve intervening to reduce some forms of consumption (for example, on luxury (p.94) goods or foreign holidays by the elite and their children) and/or introducing incentives to save.
Restricting consumption (and domestic production) of luxury goods may be more difficult to do today than it was, say, in South Korea or Taiwan in the 1960s and 1970s. There is, for example, a sharp contrast between levels of car ownership in South Korea and Japan, on the one hand, and South Africa on the other (Figure 4.1).73 At a per capita income level of around $3,000, car ownership in South Korea and Japan was about 5 per 1,000 people; at a similar level of income per capita in South Africa, car ownership is very much higher, about 65 per 1,000 people. Nonetheless, just as capital controls and development banking were anathema in international development circles for many years but have become more acceptable in the wake of the global financial crisis, so direct controls on consumption may yet come to play some role. These need not involve outright bans, but could include steep duties on luxury consumption as part of a broader effort to promote a national developmentalist ethos.
Other policies that would encourage savings to catch up with rising investment levels include, among other things, post office savings schemes, keeping import duties low on investment goods, using tax breaks judiciously to encourage investment, and directing the banking system to increase some forms of long-term credit on favourable terms. The key to many of these incentives to profitability is tying them to performance. This is what Alice Amsden called the ‘reciprocal control mechanism’ and is what has typically distinguished countries that have succeeded with state-led development strategies from those which (p.95) have had a similar array of incentives but have failed to sustain the momentum of structural change.74
The history of Mozambique since late 1992, to take just one African example, illustrates a failure to monitor and control state resources. Carlos Cardoso, an investigative journalist, was assassinated in Maputo in 2000 while investigating corruption in a bank controlled by the state.75 More recently, the travesty that is industrial policy in Mozambique reached astonishing proportions with revelations of a scandal involving loans contracted to support the development of a national fishing fleet and maritime industry: three privately incorporated but effectively state-owned enterprises failed to account for at least a quarter of the $2 billion they had received in loans, as well as paying millions more for purchased equipment than it was worth. The loans had been hidden from donors and, once revealed, pitched Mozambique into a protracted conflict with the IMF and other donors.76
There is a different—less overtly corrupt and unpleasant—method of financing investment. Many governments have financed a deficit-led investment programme through inflationary finance (and the forced savings it generates, as rising prices erode people’s spending power and reduce consumption). When governments spend more than they receive in revenue, they usually have to secure the funds to do this from the central bank, which has to print money with potentially inflationary implications for the overall economy. The resulting dynamic takes the form of distributional (class) conflicts: inflation reduces real wages and raises profits, at least in the short term. This shift of resources from wage workers to capitalists releases resources for investment, given that capitalists are more likely to save out of profits than workers are out of wages.77 It can also shift resources from wage workers to the state.
At the heart of many of the dynamics of investment, demand, government spending, and related policies, is one issue that gets very little attention in most economic texts and in advice to policy officials: what happens to the prices of basic wage goods.
Wage Goods: The Fundamental Constraint on Financing Investment
The most serious constraint on financing a cumulative dynamic of investment, productivity growth, employment, and savings, stems from the price inflation of (p.96) basic wage goods. This compromises both savings by workers and the profitability of investment, and can also threaten political stability. This is the insight of Michał Kalecki’s development economics.
African wage workers, especially relatively unskilled workers in elementary occupations—such as domestic servants, cleaners, and construction workers—spend a large proportion of what they earn on a limited range of basic goods and services. Spending on food takes up a large share (about 75 per cent) of their disposable income, together with basic wage goods such as clothing, kerosene, transport, and rent.78 If the prices of these necessities rise relative to nominal wages, there are a number of implications. Employers can face pressure to raise wages in order that their employees stay in work and can work productively. This is a very real issue and if the pressure does not come through organized labour in the form of trade unions it may come through other mechanisms, including a wastefully high turnover of labour.79 The increasing cost of wage goods has been one of the main reasons why flower firms and textile exporters in Ethiopia have found it difficult to retain workers; the result being not only that employers face recurring training costs, but also that labour productivity in Ethiopian floriculture remains low compared to levels achieved on farms in other countries.80 If firms have to raise wages then their profit rate may be hit, negatively affecting the momentum of new investment. Even if firms avoid this pressure and protect profits, they may face insufficient domestic demand for the goods they produce, and wage workers will find it difficult to sustain, let alone expand, their rate of saving. Finally, sustained wage goods inflation may lead to a political reaction, frightening existing and potential investors.
While inflationary financing of government deficit spending is one possible source of wage goods inflation, the broader problem is one of how to finance economic development, with the core challenge for policymakers in developing countries being how to ensure an adequate, non-inflationary supply of these goods. Kalecki realized there were often rigidities (social and economic) that held back supply of agricultural commodities, leading to inflationary pressure as the number of urban industrial workers expanded. These supply rigidities lead not only to higher costs for wage workers, but also higher costs of intermediate goods such as sugar or cereals used by manufacturing firms in agro-processing industries. So, for industrialization and structural change to proceed, ‘there must be no inflationary price increase of necessities’.81 This is one reason why policies targeting investment to raise agricultural productivity are not an alternative to industrial policy, but are integral to it (see Chapter 9).
(p.97) The domestic food supply may be a less binding constraint in a world of rising food productivity and, often, internationally available cheap foods. Even so, the volatility of global food prices means that carefully designed government interventions to protect the real wages of poor wage workers remain a priority. Additionally, the broader point that structural change and growth are constrained by the non-inflationary availability of basic wage goods holds. It is therefore surprising that Kalecki’s insights have been largely ‘forgotten or discarded’.82
Policy officials need to understand the importance of monitoring food prices and money wages much more closely. If financing investment ignores basic wage good inflation, then the distributional consequences may constrain the further growth of investment. Such a strategy is not economically or politically feasible over a long period of time, as it hits the lowest-paid worst, who have the least bargaining power to increase their money wages rapidly in response to food price spikes. And it is not—as some may think—a good thing for the rural population. Most people in rural Africa—and most definitely the poorest, who as we show in Chapter 8 depend on wage employment opportunities—are net food purchasers, and thus food price inflation may hit them harder than urban workers. In fact, food prices in difficult-to-reach rural areas and areas affected by violent conflict are often even higher than in large towns.83
Inflationary shocks may be unavoidable. Violent conflicts, global commodity speculation, and climatic shocks can provoke or accelerate wage good scarcity, as well as sharp price increases. In such circumstances, it becomes important to address price fluctuations with appropriate policies. Another ‘lesson’ from the historical record in East (and South East) Asia is that governments intervened to manage basic goods prices and got them ‘wrong’. Two conclusions arise. First, long-term success in addressing rural poverty and reducing vulnerability to fluctuations in food prices requires public investment in rural infrastructure, above all in irrigation—an argument made in more detail in Chapter 9. The key to a secure non-inflationary supply of food is rising agricultural labour productivity. Agricultural growth also makes it possible for rural wages to rise, without which poverty cannot fall. Second, protecting low-income classes from periodic food price spikes requires state intervention to stabilize prices. The very mention of this is enough to alarm most economists, going as it does against established ‘best practice’, and raising the spectre of ineffective and long-lasting subsidies that distort key price signals in agricultural markets.
That said, policymakers would be remiss not to pay some attention to these warnings. There have been plenty of examples—in Africa, Latin America, and (p.98) elsewhere—of disastrous state management and control of agricultural markets that has resulted in the transfer of income from farmers to predatory bureaucrats.84 However, as the relative success of state intervention in cotton marketing in some West African countries shows very clearly, such outcomes are far from inevitable.85 It is important not to ignore the fact that it has been possible for other countries to successfully intervene to stabilize food prices against volatility on world and domestic markets. Food price stabilization in Indonesia and the Philippines, for example, has been expensive, but the benefits have outweighed the costs. What marks out rice price stabilization in Indonesia and some other East and Southeast Asian countries has been limited but sufficient government intervention in markets to ensure adequate stocks, as well as a government monopoly on importing rice to protect against huge swings in international prices.86
4.3 Sectoral Policies and Structural Change
In the first part of this chapter we discussed the challenge of raising investment levels, emphasizing how investment is constrained by the supply of food and other basic wage goods. The productivity of wage good production, in other words, sets a limit on an economy’s ability to push investment to high levels and keep it there, in order to sustain the long-run capital accumulation that has historically been at the heart of growth, development, and catching up. Should resource allocation to agriculture, then, be the major priority? More broadly, if raising investment levels is key, does it matter which sectors attract investment?
The history of modern economic growth shows that contributions made by particular sectors to output growth matter a great deal. Specifically, what matters most is investment in manufacturing. However, a narrow, ‘traditional’ notion of manufacturing would be misleading, and could lead to a strategy that ignores gains from a range of activities that have not previously fitted neatly into this category. Later in this section, we discuss an expanded and policy-relevant definition of manufacturing.
Kaldor’s Growth Laws and Structural Change
Economic development involves shifting resources out of low-productivity activities into higher-productivity economic activities. This is what we mean when we talk about ‘structural change’. Structural change can also mean, broadly, a shift in (p.99) the relative weight of different sectors. So, for example, many advanced capitalist economies have gone through phases during which they changed from predominantly agricultural economies to industrial economies; then, in a process known as ‘deindustrialization’, have become dominated by service sector activities. Sectoral shifts are central to our sense of structural change, but their relevance lies in what they imply for changes in productivity across the economy, as well as the growth of employment and export revenue.
In low-income countries, large numbers of people languish in low-productivity agricultural (and related rural) activities. They live and labour on small farms, and however hard they work, grind out only modest outputs, often of low value and for little material reward. Many live and work far from graded roads or railway links, walking for hours to weekly markets, weighed down by bananas or tomatoes, only to trudge back to farmsteads that are not irrigated and do not have any fertilizers to increase yields. In the dry terms of economics, the returns to labour are low. Their children often get insufficient nutritious food, no or at best poor-quality schooling, and are poorly served by health posts and safe drinking water.
Arthur Lewis provided the classic way of thinking about changes in the structure of employment during the process of development. It is frequently alleged that he focused on the importance of rural–urban migration or the shift out of agriculture into industry, but this is a misinterpretation—his point was always more about the shift into higher-productivity activities than specifically migration to cities. In his model of how accumulation might take place in a poor economy, more and more people begin to work in activities with dramatically higher productivity and returns to labour.87 More capitalist enterprises are established in the economy and more people work for larger employers, earning higher wages. Historically, this takes the form of both a rising share of manufacturing industry in GDP and a rising share of total employment in manufacturing. Higher labour productivity comes about from a new combination of human labour and other inputs, such as machines, the organization of farms and firms, and logistical services.88
There is a fundamental process by which the more people focus on specific activities, the better they become at doing them, and then the higher the returns they typically earn. This is the terrain of Adam Smith’s seminal discussion of how the ‘extent of the market’ (i.e. market size) and ‘specialization’ play off and support each other. The larger the market, the greater the scope for people specializing in particular parts of a production process; the more people are employed in these specialized roles, and the cheaper these efficiently produced products become, the further the market for them expands. When this is (p.100) repeated across different goods, a cumulative causation process will kick in and underpin sustained economic growth.
A corollary of the cumulative causation process is new opportunities for production to benefit from economies of scale and scope. One reason that the unit cost of producing outputs falls as more and more of them are produced is precisely that the firm, and its workers, become increasingly efficient in producing them: there is ‘learning by doing’. Adam Smith naturally saw this too, with economies of scale playing a significant role in the early chapters of An Inquiry into the Wealth of Nations. Increasing returns, however, faded from economic view for a long time.
If the assumption of diminishing returns is a founding axiom of neoclassical economists, for other economists, identifying where there is scope for increasing returns is a key principle of analysis and policy design. What has become clear to many is that the scope for increasing returns is significantly greater in manufacturing than in either agricultural or service activities (there is also a relatively high scope for economies of scale in some other industrial activities, such as mining, utilities, and construction).
Manufacturing seems also to be associated with higher productivity and swifter improvements in productivity than other sectoral activities, with evidence showing that ‘value added per worker is much higher in manufacturing than in agriculture’. Shifting resources from agriculture to manufacturing, in line with this evidence, results in a one-off economic gain. There may be more dynamic gains too, if the rate of growth of productivity is faster in manufacturing than in other sectors. In most developing countries, the historical as well as the most up-to-date evidence shows this to be the case.89
It was this kind of evidence that led Kaldor to craft his three ‘growth laws’, though, as we show, these need to be expanded to four in order to take account of the balance of payments constraint on growth. Kaldor’s original propositions are as follows. First, there is a strong positive relationship between the rate of growth of output in manufacturing and the rate of growth of GDP. This is not just an ‘association’ but a causal relationship, for which there is a huge amount of evidence.90 It is this evidence that is behind the rediscovery of industrial policy—active interventions by governments to accelerate and promote manufacturing expansion—among many mainstream economists, and that has led to rather misleading claims of a new consensus among economists. Unusually, a recent contribution by IMF economists expressly referenced some of the well-known and previously ignored heterodox work on industrial policy.91 Actually, there is no such consensus; the mainstream case for industrial policy only justifies very (p.101) limited forms of policy ‘correcting’ market failures rather than creating new markets.92
Kaldor’s second proposition is that there is a strong positive relationship between the rate of growth of manufacturing output and the growth of manufacturing labour productivity. This is also known as Verdoorn’s Law and captures dynamic economies of scale. It flows from the greater scope for learning by doing, specialization, and economies of scale in manufacturing; as well as from rising investment in industrial equipment and machinery that itself embodies productivity-raising technical change.
There are implications for what happens to employment. If a firm can produce the same output with fewer workers, or if firms are increasing output faster than they take on new employees, then the employment elasticity of manufacturing will decline. Tregenna finds that industrial upgrading and new labour-dislodging machinery were associated in South Africa, between 1980 and 2005, with a falling employment multiplier in manufacturing. Manufacturing is still the engine of growth but it may not always carry quite so many people with it (at least, not directly).93
Policy officials in Africa might see a dilemma: why should they promote manufacturing if it is not helping to address the fundamental challenge of creating sufficient employment to absorb the (rapidly growing) number of new entrants to the labour force? But there is no dilemma. Even within manufacturing, many kinds of production generate large amounts of new employment, and even if manufacturing has directly a low employment elasticity of output, it is hugely important indirectly to the employment-creation capacity of the economy as a whole. Evidence suggests that for every manufacturing job created two or three jobs are also created outside the sector as a result of manufacturing expansion.94 The contribution of manufacturing to employment is even stronger once we accept a broader definition of manufacturing that takes account of agricultural production, which is, on closer inspection, clearly industrial (see later in this section). However, the most appropriate distinctions are not those between manufacturing and other activities, but between those activities with a high potential for increasing returns and employment multipliers, and those without such potential.
Kaldor’s third proposition is that there is a strong positive relationship between the growth of manufacturing productivity and productivity in the rest of the economy. This is partly because, as manufacturing draws labour out of, say, agriculture, it leaves behind a smaller agricultural labour force without shrinking (p.102) overall agricultural output. But it is also the outcome of a richer process, whereby productivity gains in manufacturing are diffused to other sectors through the supply of cheaper equipment and inputs to agricultural producers—for example, agro-chemicals and irrigation pumps. They are also spread through industrial demand for agricultural inputs—for cotton inputs into in textile production, for example—and propagated by the spread of organizational know-how, though this too is a matter for policy rather than just an automatic process.
Empirical work has largely confirmed the significance of these propositions. For example, Wells and Thirlwall tested for and found clear evidence of Kaldor’s growth laws across a number of African economies.95 Elsewhere, another recent study found support for Kaldor’s first and second laws in a group of 63 middle- and upper-income countries.96
The implication is that there is a very strong case for governments designing and implementing industrial policy to promote manufacturing investment, as this results in general economy-wide benefits. Later in this section we show how important it is to think carefully about what kinds of activity to include within the range of industrial policy, as well as highlighting some of its key features. Before we do that, though, we focus on the significance of the balance of payments constraint.
The Balance of Payments Constraint
One thing that can bring momentum to a standstill is, as previously discussed, inflation in the prices of food and other necessity wage goods. Another equally important factor is a shortage of foreign exchange. Developing countries have a thirst for imports. Partly this is a function of rising incomes: as the demand for basic goods starts to take up less of people’s incomes, there is rising demand for a range of imported consumer goods, whether this be globally fashionable second-hand jeans and T-shirts, or smartphone apps. Even more importantly, almost by definition an expanding economy must invest in technology-embodying machines and inputs, which creates a powerful need for imports. A developing economy does not have to reinvent the wheel of industry—rather, one of the advantages of backwardness is that expertly designed wheels can be imported. To some extent, it may be possible to limit the rate of growth of luxury consumer good imports, but it is impossible to successfully sustain rapid development if imported capital and intermediate goods dry up. The building of the Grand Ethiopian Renaissance Dam, the Kandadji Dam in Niger, and the infrastructure renewal programme in Côte d’Ivoire are all associated with sharp spikes in (p.103) imports.97 Across sub-Saharan Africa, electric and non-electric generating equipment imports increased from $2.3 billion in 2000–3 to $11.6 billion in 2008–11.98
The greater the demand for imports, the greater the challenge of securing enough foreign exchange to cover the import bill. It is very difficult for an economy to sustain a growing balance of payments deficit indefinitely. If the size of the deficit is accelerating and will have to be financed from additional foreign sources, then there is a real danger that a government may be forced to abandon any kind of independent policy initiatives: arguably, this has sometimes affected policy direction in countries such as Ghana and Mozambique.
An economy can acquire the foreign exchange needed to finance its import bill by attempting to attract foreign capital flows, including concessional loans and grants (i.e. foreign aid or overseas development assistance); or by earning foreign exchange through exporting goods and services. All these sources of foreign exchange involve costs and uncertainties. Commercial capital flows are regarded as more beneficial to Africa than aid, especially by employees of Goldman Sachs, such as economist Dambisa Moyo.99 But ‘hot money’ funds shifted into African emerging market bonds and stock exchanges, as opposed to foreign direct investment in farms and factories, are fickle, responding unpredictably to international changes in interest rates, policy environments, and bandwagon hunches about the next global bonanza or calamity. Those directing these flows to emerging markets are said to be involved in ‘stampedes by intrinsically restless (and often underinformed) fund-managers—so prone to oscillating between manias and panics’.100 Additionally, lines of credit from China (secured by export commodities or escrow accounts) are now an extremely important, but often poorly recorded and administratively opaque, source of finance for new energy infrastructure projects in several African countries.101 At the same time, it has been the vogue in recent years to use public–private partnerships (PPPs) as a means of financing infrastructure investment across the world. PPPs have been promoted in Africa by the African Development Bank, United Nations Economic Commission for Africa (UNECA), and the African Union, with the support of the World Bank and Organisation for Economic Co-operation and Development (OECD) among others.102 Supporters argue that risk is transferred to the private sector and that PPPs are good value for public money. But their record globally, including in the UK, which was one of the pioneers of PPPs, has been unimpressive.103 Many people, including research staff at the IMF, warn how difficult it is to put in place effective institutional mechanisms to protect public finances against spiralling costs and fiscal unsustainability, (p.104) ‘particularly when governments ignore or are unaware of their deferred costs and associated fiscal risks’.104
The volatility of foreign aid flows has created very serious problems for policymakers and economic performance in Africa. Donor agencies can swiftly change the activities to which they commit resources depending on the latest ideas in vogue.105 The administrative costs of experienced public officials having to deal with a proliferating number of donors, donor fads, and demands are high.106 Foreign aid also comes with strings and narrows economic policy debates, providing material and intellectual support for African elites arguing for demand restriction and market liberalization, and reinforcing flawed ideas through repetition of rhetorical commonplaces. As has been pointed out, IMF programmes
still incorporate a considerable number of structural conditions, and the total number of such conditions still far exceeds that observed in the pre-1994 period. The emphatic return of structural conditionality in recent years calls into question the IMF’s ‘we won’t do that any more’ rhetoric.107
It is possible to negotiate important improvements in the terms by which foreign investment and foreign aid inflows become available, as shown by the variety of outcomes and negotiating strategies across African countries. Foreign aid ‘essentially means an improvement of external conditions of growth’ and may be compared to ‘a positive shift in terms of trade to the extent which both increase the capacity to import’.108 However, negotiations over concessional and non-concessional flows might not always be successful—or even be conducted in the national interest—so an over-reliance on these sources of foreign exchange is unwise. Under the pressure of demand for imports, it may be a far more effective strategy for sustaining growth to promote a rapid expansion in foreign exchange earnings through exports. Thirlwall’s Law, capturing this insight, states that a country’s growth rate over the long run is set by the ratio of the rate of export revenue growth to the income elasticity of demand. The faster imports expand as an economy grows, the faster export earnings have to rise if growth is to be sustained.109
There are two dimensions to why exports have to be at the heart of a development strategy. One is that, as explained, the momentum of accumulation and structural change cannot be sustained without a swift rise in export revenue. The other is that exports—especially manufactured exports—are at the heart of the dynamics captured in Kaldor’s propositions. Exploiting the economies of scale (p.105) that manufacturing has a special propensity for depends, just as Adam Smith claimed, on the ‘extent of the market’. Given the extent of the market (i.e. demand) is limited in developing countries and within the African region, the rest of the world must ergo be the major source of demand. This is just one of the ways in which manufactured exports matter to industrialization-led development. More formally, there are two linked relationships that have been observed internationally. First, growth in manufacturing raises the rate of growth of export earnings. Second, the faster export earnings grow, the higher the overall growth rate of the economy.
What Counts as Manufacturing?
Auto assembly plants, furniture workshops, a tomato canning facility, industrial parks exporting garments, a blueberry producer: all except one are classified in standard statistics as manufacturing. The obvious exception is the blueberry producer, but arguably blueberry producers and exporters ought to be regarded as industrial producers. If a blueberry producer in South Africa is highly capitalized and has invested in a range of physical components, inputs, and processes, and if that producer has organized the production process into a long string of distinct and specialized operations, are they merely a large farmer? Expanding the idea of what is included in industrial production has important policy implications.
If policy officials want to take seriously the dynamics of structural change and promote investment, they need to recognize the shifting boundaries around the ‘industrial’, classifying economic activities according to their role in an economy: above all, their potential to generate increasing returns of scale and scope; accelerate productivity gains; lead directly and indirectly to employment growth; and produce foreign exchange earnings.
Changing the understanding of what constitutes manufacturing is even more important when we consider that much agro-industry is in fact more labour-intensive than a great deal of the manufacturing industry located in export-processing zones or industrial parks. In Ethiopia, for example, the total number of workers employed by agribusiness in floriculture was, by 2012, already considerably larger than total employment in the manufacture of textiles or in the basic metal and engineering industries. In Kenya, meanwhile, 125,000 workers were employed in the cut-flower sector in 2015, which was about 3 times the number of people employed in textiles, and 20 times larger than the number of people employed in the motor vehicles subsector.110
(p.106) Farms nowadays increasingly require massive investment in complex technology and equipment, and employ highly trained and productive workers. People may (and do) migrate in search of employment opportunities with higher returns, not simply from rural areas to towns and cities, but also from one rural area to another. A development strategy designed to raise investment levels and increase profitability in the economy has to take into account the huge scope for productivity increases (and employment growth) within rural areas, as well as in urban and peri-urban factories.
Premature Deindustrialization: The Latest Version of Fracasomanía?
Rumours of the demise of industrialization as the engine of development are greatly exaggerated. Many neoclassical economists traditionally argue that it matters little which sector economic activities take place in, and that therefore it is market ‘distorting’ to try and encourage manufacturing specifically. Additionally, there are old populist arguments that it is possible to raise income per capita without having to ‘pass through’ an (unhappy) industrial stage.111 A similar tradition of pessimism about the feasibility of industrial policy in Latin America led Albert Hirschman to coin the phrase fracasomanía—an obsessional fascination with failure.
More recently another theory of failure has gained currency: the risk of ‘premature deindustrialization’, which has shown that there is a global trend for economies to experience, for structural reasons, a shrinking share of manufacturing in GDP at much lower income per capita levels than has historically been the case. An optimistic slant on this issue, tinged with the enduring appeal of economic populism, is that services are nowadays more productive than manufacturing, and therefore it may be possible to have a service-driven development strategy without having to grind through the socially, environmentally, and technically challenging process of industrialization.112
However, the evidence across developing countries suggests a far more positive interpretation of the ‘premature deindustrialization’ narrative, even taking into account a traditional classification system that—we and others argue—is increasingly unfit for purpose. In real terms, global MVA increased in every year between 1995 and 2015.113 In the developing countries as a group, aggregate MVA/GDP has in fact risen rather rapidly since about 2002, as shown in Figure 4.2. Disappointing shares of MVA in GDP and weak manufacturing employment, (p.107) with the UK economy as an extreme example of the latter, are attributable to policy failure in some countries rather than the structural impossibility of industrialization.114 The growth of MVA has been especially impressive in East Asia, where China in particular has shown the enduring scope for manufacturing expansion and manufacturing employment growth.
4.4 Conclusions: Criteria for Policy Priorities
We have argued in Section 4.3 of this chapter that manufacturing, suitably (re)defined, plays a leading role in economic development. There are good reasons for this, which are captured by Kaldor’s growth laws and supported by consistent evidence internationally. Despite claims that ‘premature deindustrialization’ means that manufacturing is no longer an effective engine of growth and employment generation, the evidence demonstrates this is not true. What has been the case in many countries in the recent past is not a guide to what governments can and ought to do to promote economic development in the future. Industrialization (p.108) remains central to economic development, and industrial policy remains central to making industrialization happen.
The scope of industrial policy is wider than before. Manufacturing has become increasingly imbricated with other activities traditionally labelled as services or agriculture. This has resulted in the servicification of manufacturing and the industrialization of freshness, meaning industrial policy has to address these trends. The instruments of industrial policy have, for example, been shown to be useful in supporting the expansion and competitiveness of high-value agricultural production in some African economies.
Finally, industrial policy must focus on manufactured exports. The association between manufactured exports and sustained higher rates of growth is a strong one. Just as more diversified and higher-productivity manufacturing is likely to help protect against the frequency and intensity of growth collapses, so too is an effective manufacturing export sector. This is because, among other benefits, it can help overcome the balance of payments constraint that is often a cause of stalled growth.
Both industrial policy and linked policies promoting agribusiness should follow four criteria guiding the allocation of all government support (other chapters pick up on these criteria, so we will highlight them only briefly here). First, does an activity or sector generate foreign exchange earnings (i.e. will it help relax the balance of payments constraint on growth)? Second, is it characterized by scope for rising labour productivity vis-à-vis other sectors and activities? Note that this is not the same as asking whether or not it is manufacturing in the traditional sense. Third, will it create employment (directly and indirectly), especially for women? In particular, will it generate an increase in higher-productivity job opportunities relative to other projects that might benefit from state support? Fourth, will it help address the need for a non-inflationary supply of basic wage goods? Even if resource allocation addresses these criteria, investment or investment support can only work if it is supported by other policies, for example reducing the risk of export revenues flowing into tax havens abroad.
These criteria need to be combined with policy design that takes sequencing into account. The question that should be asked when considering a potential incentive or infrastructure investment is: how fast will this or that activity bring foreign exchange earnings onstream? In some circumstances, there may be a case for supporting an activity that does not meet all the criteria set out, but that does promise to relax one binding constraint (e.g. foreign exchange or wage goods supply) particularly quickly, creating greater room to pursue longer-term objectives. This is relevant for debates about the role of primary commodity exports, with many governments and development economists having neglected investment in this area. In neglecting these activities for fear of, for example, declining terms of trade, governments may miss out on high international prices (p.109) for particular commodities, thereby foregoing short-run foreign exchange earning opportunities that could be fundamental to deeper prospects for structural change.
To conclude this chapter with an example, a large-scale investment in Ethiopia promoting a rapid expansion in high-quality, premium-priced, wet-processed coffee production capacity may address several of our key criteria for selective intervention by government. It would involve a large ‘lumpy’ investment in infrastructure (e.g. rural roads and electrification) that without government spending would not happen nearly fast enough. It would generate additional foreign exchange, without a long lead time to develop international competitiveness. It would also generate a substantial increase in rural wage employment for women. The output of each of these workers will be lower than the output per worker in, say, factories producing sports shoes, but their labour productivity would nevertheless be much higher than that of most workers in rural Ethiopia.115 An effective policy promoting the rapid expansion of high-yield, higher-quality coffee would require exactly the kind of ‘reciprocal control mechanism’ characteristic of effective industrial policy and for the same reasons; namely, the need to provide selective incentives to those with a demonstrated capacity for and track record of generating foreign exchange and employment in a sector with the potential for generating linkages to other activities, which are also likely to be producers with scope for scale economies. It would make industrial policy for other activities—for example, packaging, cement, and machine tools—much more feasible due to its contribution to the balance of payments.
These kinds of criteria for policy design, as well as the broader significance of setting export growth as a high priority, are developed in Chapter 5 on trade.
(3) Marx (1993: 265).
(8) Sato (2005: 636).
(10) See Pollen (2018), critically assessing Mwanawina and Mulungushi (2008).
(19) Arizala et al. (2017: 17); see also Ghazanchyan and Stotsky (2013: 3). Earlier cross-country regressions produced fragile results on the determinants of growth. But Levine and Renelt (1992) did find that across many different statistical studies the clouds parted to reveal a clear and strong relationship between the investment ratio and growth.
(20) Maddison (1991: 65).
(22) Maddison (1991: 139).
(37) Ferreira and Soares de Oliveira (2018: 17–18).
(47) Broadberry and Wallis (2017: 3–4).
(52) Blanchard (2012: 1).
(55) Nyambura-Mwaura and Genga (2016) on Kenya.
(56) Lazzarini et al. (2015). On BNDES (Banco Nacional de Desenvolvimento Económico e Social) see: Tavares de Araujo Jr. (2013) de Aghion (1999); UNCTAD (2016); Di John (2016). Weiss (2015) emphasizes the role development banks can play in counter-cyclical spending, boosting demand at critical points in the economic cycle.
(63) Sandbu (2017).
(68) Moller and Wacker (2017: 199).
(71) ibid (209).
(77) Joan Robinson (1964: 341) attributed to Kalecki the pithy formulation that in a capitalist economy ‘workers spend what they get, capitalists get what they spend’, though Kalecki’s writings do not contain this direct statement.
(83) For instance, prices of maize are often much higher in rural than in urban South Africa, according to data in the National Agricultural Marketing Council’s regular Food Price Monitor (http://www.namc.co.za). On the impact of violent conflict on food prices see Raleigh (2015).
(88) Differences in labour productivity (and in the growth of labour productivity) between low-income and advanced capitalist countries are discussed in Schaffner (2001).
(91) Cherif and Hasanov (2019).
(98) Co (2014: 11–12).