Structural Transformation and Inclusive Growth
Structural Transformation and Inclusive Growth
Conceptual Points of Departure
Abstract and Keywords
This chapter sets out some conceptual points of departure for the book in terms of structural transformation and inclusive growth. It revisits the Lewis model of economic development and proposes it as a heuristic device to connect structural transformation and inclusive growth. The chapter argues first, that both structural transformation and inclusive growth have tended to be defined in a reductionist sense, in a way that disconnects the two concepts. It is contended that this matters because the relationship between structural transformation and inclusive growth is embedded in—rather than separated from—the modality of late capitalism pursued. Second, that the work of pioneering development economist, W. Arthur Lewis and the Lewis dual economy model provides a useful heuristic device for thinking about the relationship between structural transformation and inclusive growth.
Chapter 1 outlined the ‘developer’s dilemma’. To recap, developing countries are seeking economic development, which requires structural transformation and productivity growth via intra- and inter-sectoral reallocation of economic activity. At the same time, developing countries are also seeking some form of broad-based or ‘inclusive growth’ which can be defined in various ways, typically in terms of who benefits from growth and by how much. The dilemma is the trade-off between structural transformation and inclusive growth. There is a ‘developer’s dilemma’, in that structural transformation has historically been thought to have a tendency to push up inequality, as Arthur Lewis and Simon Kuznets originally posited, though they differed to some extent on whether this is inevitable and what the countervailing forces are. On the other hand, inclusive growth implies a need for steady or even falling inequality to spread the benefits of growth as broadly as possible. In short, the ‘developer’s dilemma’ is a distribution dilemma or tension at the heart of economic development. Empirical support for this potential trade-off or developer’s dilemma—that structural change is associated with rising inequality—was originally offered in Sumner (2016a, 2016b), and then offered in greater detail by Baymul and Sen (2017).1 Given the empirical support for (p.14) the developer’s dilemma, how are countries to manage the tension and what are the experiences of countries who have managed the tension?
The purpose of this chapter is three-fold. First, to outline the components of the developer’s dilemma in terms of the conceptual points of departure for the book in terms of structural transformation (ST) and inclusive growth (IG). Second, to situate those concepts of ST and IG in a model of economic development appropriate for the study of the developer’s dilemma. Third, to contextualize Lewis and Kuznets to contemporary times to take account of multiple forms and dimensions of structural change.
The following chapter (Chapter 3), then focuses on the stock of theory relating to the facilitators and impediments to structural transformation and industrialization, and the distributional questions arising. Much of this theory is focused on the nature of the adverse incorporation of developing countries into the global economy and extends the discussion of the Lewis model of economic development. This is followed by a discussion of the contemporary literature on the heterogeneity or differentiated regimes of late capitalism and culminates in the rationale for, and outline of, the ‘Social Structure of Capital Accumulation’ approach as the analytical approach of the book. Chapters 4, 5, and 6 then consider how the developer’s dilemma played out in South East Asia in three distinct ‘waves’, each culminating in crisis, and all of which, were, it is argued, triggered by changes in the global economy and by the nature of countries’ incorporation into the global economy.
The core arguments of this chapter are as follows: first, that both ST and IG have tended to be defined in a reductionist sense, in a way that disconnects the two concepts (in fact those who study one would, most likely, not study the other). It is posited that, by defining ST solely by shares of gross domestic product (GDP) or employment, more important aspects of ST are neglected. In a similar vein, by defining IG as the reduction of absolute poverty (or the consumption or income of the poorest only) and—sometimes—by changes in individual income inequality, some other aspects of IG, notably capabilities, inequality of opportunity and importantly, employment, have been neglected. It is argued that this matters because the relationship between ST and IG is embedded in—rather than separated from—the modality of late capitalism pursued.
The second part of the thesis is that both the work of pioneering development economist, W. Arthur Lewis, and the Lewis dual economy model, provide a useful heuristic device for thinking about the developer’s dilemma and the relationship between ST and IG. Further, although Lewis did not ignore inequality, the seminal work of Simon Kuznets, who used a dual economy model, and those writing in the Kuznetsian tradition, are useful to revisit in order to understand the inequality dynamics of the transition from a traditional to a modern economy though Kuznets and Lewis need to be (p.15) contextualized because of multiple forms of contemporary structural change, not least ‘premature deindustrialization’ and circular (back-and-forth) or ‘commuting’ labour movements; non-farm rural income; the growth of inter-sectoral resource flows and the multiple dimensions of inclusive growth.
Contributions since Lewis and Kuznets seminal work have placed emphasis on the modality of incorporation into the global economy.2 For instance, much of the theory discussed in the following chapter takes the view that the modality of incorporation into the global economy and the intermittent tectonic plate-type changes in the global economy, such as changes in global oil prices, are important for understanding national economic development. At the same time, others have focused on domestic factors and/or the interplay of both.
The current chapter is structured as follows: Sections 2.2 and 2.3 discuss the conceptual components of the developer’s dilemma—the concepts of structural transformation and inclusive growth, and their meaning. The argument of this section is that important aspects have been neglected in the reductionist definitions of ST and IG. Further, that ST and IG have been considered separately, rather than as interconnected phenomena. Section 2.4 situates the concepts within the transition as outlined by Arthur Lewis, and the application by Simon Kuznets of income inequality dynamics within the transition. The Lewis model postulates that the transfer of labour, from low productivity to higher productivity activities and sectors, is the key driver of economic development.3 The Kuznets hypothesis on rising inequality during structural transformation implies a pattern of growth that constrains income/consumption gains at the poorest end of the distribution. Empirically, the Kuznets hypothesis has been generally rejected in cross-sectional data, though it has seen an empirical resurgence in recent time-series data for fast-growing developing countries where growth is accompanied by structural transformation (see Sumner 2016a, 2016b). Various contemporary authors, writing in the Lewis/Kuznets tradition (for example, Acemoglu and Robinson 2002; Aghion and Bolton 1997; Gallup 2012; Lindert 1986; Lindert and Williamson 2001; Oyvat 2016; Roine and Waldenström 2014; Williamson 1985) are discussed. Finally, Section 2.5 concludes and sets the scene for the next chapter.
The purpose of this section is to discuss, first, what might be called the dominant definitions of both core concepts, ST and IG. This section argues that such definitions neglect deeper aspects of ST and IG, and that, in doing so, they disconnect the concepts of ST and IG, thus delinking poverty, inequality, and capital accumulation. This section is structured as follows: this overview is followed by a discussion of the meanings of ST and IG. This is succeeded by a proposal for alternative definitions of ST and IG that link the two conceptually. In short, the argument of this section is that the two conceptual aspects of the ‘developer’s dilemma’ have been historically approached in a more reductionist sense and treated as analytically separate. Consequentially, the relationship between ST and IG has not been teased out in greater depth.
2.2.2 Structural Transformation: Definitions
The purpose of this section is to critically discuss different definitions of ST. In considering the conceptualization of ST, one is pushed to consider ST from what, between what, and in what direction, and the possibility of multiple and co-existing/co-evolving transitions beyond that of a shift from the ‘traditional’ or agriculture or rural or informal sector to the ‘modern’, manufacturing, urban, or formal sector. This section argues that dominant approaches to defining ST have generated understandings based largely on sectoral shares of GDP and employment. Yet, the broader set of structures related to the factoral composition of growth, labour productivity, and international trade are just as important in garnering a deeper understanding of ST. One could potentially cast the net even wider to include political and social institutional change.
First, we consider the neglect of ST in developing countries (relative to the attention to IG). It is reasonable to say that structural transformation—in the first instance meaning the reallocation of economic activity not only between, but also within, sectors towards higher productivity—has been neglected at least relative to the attention to IG with reference to developing countries.4 Although the importance of the shift to higher productivity is not disputed in neoclassical economics, a one-sector model of economic growth has become (p.17) standard in macroeconomics. In this one-sector model of economic growth there is no account of the process of inter-sectoral reallocation of economic activity or structural transformation. This is because, in the neoclassical growth model (of Solow 1956), growth is driven by incentives to save, accumulate physical and human capital, and innovate. In contrast, for most heterodox economists, growth is due to reallocation of economic activity to more productive sectors or more productive sub-sectors.
In some ways, this debate is embedded in historic discussion on the need, or not, for ‘balanced growth’ across sectors.5 Regardless of whether growth should be balanced or not, ST in itself is an important driver of growth because of (often dramatically) differing productivity levels between, and even within, sectors. Although whether productivity can be accurately measured is contested by Fischer (2011) who refers to a ‘fallacy of productivity reductionism’, which is the assumption that productivity can be measured in a complex economy. Measuring productivity relies on value-added account data, but such data is a combination of output and prices/wages. So, most measurements for productivity show price or wage differentials not actual effort, output, or skill. This is an even bigger problem in the service sector as the comparability of services is more problematic because they are not physical goods that can be compared. Fischer (2014) also notes another problem that, because transnational companies (TNCs)—who dominate production and its coordination in global value chains—conduct practices such as transfer pricing and the transferring of profits from Southern subsidiaries to Northern HQs (for example, low-interest loans from subsidiary to parent company), such actions could make the subsidiary look less productive. These are clearly important issues that, although not easily resolved, should not be forgotten.
If one assumes, for the sake of continuing further, that productivity can be measured albeit imperfectly, McMillan and Rodrik (2011, p. 1), in taking sectoral and aggregate labour ‘productivity’ data empirically, show that the (p.18) transfer of labour and other inputs to higher productive activity is a driver of economic development, as Lewis hypothesized.6 They note:
One of the earliest and most central insights of the literature on economic development is that development entails structural change. The countries that manage to pull out of poverty and get richer are those that are able to diversify away from agriculture and other traditional products. As labour and other resources move from agriculture into modern economic activities, overall productivity rises and incomes expand. The speed with which this structural transformation takes place is the key factor that differentiates successful countries from unsuccessful ones.
They go on to note that ST can in fact be growth-enhancing or growth-reducing depending on the reallocation of labour.7 This is an important point and relates to the multiple modes of ST and direction between sectors, which may be regressive as well as progressive in the sense of productivity gains or losses. They show how ST had been growth-enhancing in Asia because labour has transferred from low to higher productivity sectors. However, the converse is the case for sub-Saharan Africa and Latin America because labour has been transferred from higher to lower productivity sectors and this has reduced growth rates.8
Is manufacturing ‘special’? Rodrik (2013) shows that unconditional convergence is evident in manufacturing, meaning faster productivity growth the further away from the labour productivity frontier. Furthermore, returns to scale imply that as costs fall, demand rises for manufacturing foods (high-income elasticities of demand), triggering more manufacturing and higher incomes, more demand, and cost reductions. Many such as Rodrik (2015) argue that most services are (i) non-tradable, and (ii) not technologically (p.19) dynamic, and that (iii) some sectors are tradable and dynamic, but they do not have the capacity to absorb labour. Similar shortcomings can be observed about the manufacturing sector. A significant share of manufacturing is (i) non-traded (even though it is tradable) and (ii) much of manufacturing in developing countries is not technologically advanced (at least in relative terms to other modern sectors), and (iii) where some manufacturing sectors are technologically dynamic, they may not create much employment, as some service sectors do. This is especially true now that it is robots and machines that perform more factory work in the electronics sector. Developing the manufacturing industry is important, but one should be careful of ‘industrial fundamentalism’. To achieve the goal of upgrading the economy and creating jobs, one should not overlook opportunities in other sectors. In short, there is a need to go beyond blue-collar jobs and manufacturing investment, and a need for policies that boost the service sector (see discussion of Kim et al., 2017).
Others, such as Herrendorf et al. (2013), concur empirically with the argument that the sectoral composition of economic activity is key to understanding not only economic development but also regional income convergence, productivity trends, business cycles, and inequality in wages. This echoes Kuznets (1971, p. 1), who, writing almost fifty years ago, listed a high rate of structural transformation as one of the six main features of ‘modern economic growth’ about which he said,
[t]he rate of structural transformation of the economy is high…[and] [m]ajor aspects of structural change include the shift away from agriculture to non-agricultural pursuits and, recently, away from industry to services; [also included is] a change in the scale of productive units, and a related shift from personal enterprise to impersonal organisation of economic firms, with a corresponding change in the occupational status of labor.
Many of the other features of ‘modern economic growth’ outlined by Kuznets relate to a broader conceptualization of ST beyond GDP and employment shares alone. Changes in societal structures, many of which are related to ST, such as urbanization and technological progress, were also included in Kuznets’ conceptualization. This raises the question of definitional scope noted at the outset of this section, and the structural change of what, between what, and in what direction, to which we next turn.
Kuznets (1971, p. 348) made a case that ST was a much broader societal process as it was embedded in societal change beyond economic structure alone. ST could be extended to population, legal, political, and social institutions, social ideology, and beliefs, and such shifts were prerequisites for the ‘modern economic growth’ he outlined. This argument also resonates with the ‘special case’ outlined by Dudley Seers (1963) when he referred to the ‘highly special case’ of advanced countries. One could call such countries (p.20) the ‘arrived’ cases of ST (with caveats for the socio-economic problems in advanced countries). Seers went beyond a definition of ST based solely on shares of GDP and employment, and the characteristics set out by Seers of the ‘special case’ remain an enduring set of features that define an advanced economy or an ‘arrived’ case in terms of ST (and indeed, IG, given the aspects of poverty and inequality noted).
In terms of the structural transformation of the economy, foreign trade, and other matters, Seers (1963, pp. 81–3) identified the following list (summarized here) to demonstrate how one might differentiate developed nations from developing nations: by factors of production (a literate and mobile labour force who are mostly in employment; substantial quantities of skilled labour; most available land cultivated; all sectors heavily capitalized, with spare capacity; comprehensive transport and power systems; a favourable climate for enterprise; firm legal basis for companies); by sectors of the economy (for example, agriculture wholly commercial; mining of limited size; manufacturing diversified and much larger than either agriculture or mining); by public finance (for example, reliance on direct taxes; tax laws enforceable; big outlays on social security and agricultural subsidies); by foreign trade (exports that have a large internal market and are sold to many countries with high price and income elasticities; imports largely of primary products and income elasticity of demand not high); by household consumption (for example, very few people below subsistence level and a moderately equal distribution of income post-tax; food not the overwhelming majority of household expenditure); by savings and investment (for example, well-developed financial intermediaries; significant personal savings and high investment); and by ‘dynamic influences’ (no chronic tendency to deficits; slow population growth and high urbanization). In short, this resonates with Kuznets (and Polanyi) and an embedding of ST in broader societal social, political, and economic change. Seers, however, argued that the ST or development of the advanced countries was a special case and not a general one that could simply be reproduced.9
More recently, echoing somewhat Kuznets’ and Seers’ thinking on ST, Pritchett et al. (2010, pp. 3–4) conceptualized ‘development’ in a broader way, as follows:
When people speak of the ‘development’ of societies most people refer, implicitly or explicitly, to a cumulative historical process whereby economies grow through enhanced productivity, prevailing political systems represent the aggregate (p.21) preferences of citizens, rights and opportunities are extended to all social groups, and organisations function according to meritocratic standards and professional norms (thereby becoming capable of administering larger numbers of more complex tasks). A given society undergoes a four-fold transformation in its functional capacity to manage its economy, polity, society and public administration, becoming, in time, developed.
Of course, one cannot deny that ST is a much broader process and economic ST is interwoven into socio-political institutions. However, there is the usual trade-off of excessively broad definitions (everything is ST) versus narrower definitions (ST is economic in nature), which are manageable for analysis.
Even if one decides to focus on economic ST alone, a set of methodological questions arises. Syrquin (2007) briefly identifies such questions and they include defining what is meant by ‘sectors’ and thus what ST means (inter- or intra- depends on the breadth of definitions of sectors) and the blurring between ‘services’ and ‘manufacturing’ due to technological advances and outsourcing (see later discussion on ‘premature deindustrialization’). Furthermore, it is necessary to discuss the related changing nature of industrialization itself in developing countries and globally. Industrialization used to mean vertically integrated national economic development to be realized in domestic industry building, but over time industrialization has come to mean horizontal integration into global value chains via foreign direct investment or as domestic suppliers. Thus, the meaning of what is to be achieved in industrialization has changed, which is not inconsequential. The changing nature of industrialization is of relevance as assessments of ST tend to overemphasize shares of GDP and employment and pay limited attention to the questions of ownership and the national capture of economic value added.
In sum, one could say that a conceptualization of ST has three discernible dimensions framed around a shift towards higher productivity activities. These are sectoral, factoral, and integrative. The first dimension—the sectoral aspects of ST—is about the inter- and intra-reallocation of sectoral activity towards higher productivity. This includes the common measures of ST, notably shares of GDP and employment. However, ST may not only be agriculture to industry or manufacturing, but agriculture to services or other sectors (see later on the modes of sectoral ST). The second dimension is the factoral aspects of ST and is about the composition or drivers of economic growth in terms of a shift of factors of production towards higher productivity activities. Underlying this are questions of demography too with reference to labour. Third are the integrative aspects of ST. This is the extent of integration in terms of the global economy and a shift from forms of incorporation—trade deficits and capital inflows that come with liabilities (for example, profit repatriation or debt repayment)—towards trade surpluses (p.22) and outward FDI. Here the changing nature of global capital accumulation is important.10
2.3 Inclusive Growth: Conceptual Points of Departure
The purpose of this section is to critically assess different meanings of IG. The argument of this section is that common approaches to defining IG have generated limited understandings based largely on absolute poverty lines and—sometimes—on reductions in inequality of outcomes. Instead, it is argued that a broader set of measures is important in garnering a deeper understanding of IG and conceptually linking IG to ST. In this section we discuss first the emphasis given to IG: we discuss first the relative overemphasis given to IG in developing countries (relative to the neglect of ST), albeit defined in a narrow way. Research has tended to show almost all growth to be, at least minimally, inclusive, although there are emerging signs that some are empirically questioning this once again (for example, Shaffer 2017; Sen 2014). Second, we discuss the basic dimensions for conceptualizing IG in terms of who to include, how much, and in what way (though this latter dimension is implicitly material- or consumption-based now or at some point in the future). There is the further question of who decides who to ‘include’ and on what terms, which relates to the social structure supporting and sustaining capital accumulation (see Chapter 3).
The substantial attention to IG in developing countries (certainly relative to ST) demonstrated by the sheer volume of literature, has tended to show growth in general is inclusive at least in a minimal way. In contrast to ST, IG has received a considerable amount of attention in academic (and policy) literature, albeit among a different scholarly community (one who works on poverty and inequality issues rather than on economic development, ST, productivity, and growth per se). IG, like ST has a lengthy genealogy with an evolutionary path into the contemporary concept of IG. IG is a term (p.23) which some associate with the World Bank, though the use of the term need not be, as we shall come to see shortly.
There has, without a doubt, been a substantial amount of attention given to IG in developing countries. Historically, the interest in the broad area—defined as who benefits from growth and by how much—grew from debates in the early 1970s that were critical of the then distribution of the benefits of growth (for example, Adelman and Morris 1973; Chenery et al. 1974). Such issues received a lot of attention in the late 1990s through to the mid 2000s under different but related terms. For example, ‘growth with equity’, drew on critical debates on East Asian development (see for example, Fei et al. 1979; Jomo 2006; World Bank 1993). Then ‘growth with equity’ was supplanted by the label of ‘pro-poor’ growth (see for example, Besley and Cord 2006; Grimm et al. 2007; Shorrocks and van der Hoeven 2004) that in turn was supplanted by the term ‘inclusive growth’ (see Ali and Zhang 2007; Klasen 2010; McKinley 2010; Rauniyar and Kanbur 2010) which became the umbrella term for considering who benefited from growth.11
This range of labels entails some differences. For example, ‘growth with equity’ was typically defined as growth where inequality does not rise or may even fall (and is a term associated with World Bank 1993). In contrast, ‘pro-poor growth’ was taken to be absolute pro-poor growth if growth was accompanied by a falling poverty headcount (or rising incomes of the poor by a poverty line or fractile line), or relative pro-poor growth if that fall in poverty headcount was accompanied by falling inequality of outcome (see discussions in Bourguignon 2003; Kakwani and Pernia 2000; Ravallion 2004).
In contrast, ‘inclusive growth’ was framed as poverty reduction both in monetary and non-monetary terms. This entailed the participation of the poor or a broader group beyond just the poor in growth processes, via employment and the expansion of capabilities (in terms of public good access), which reduced poverty and potentially reduced the inequality of opportunity and/or inequality of outcomes.
The body of literature discussed above provides the empirical basis for a generally accepted notion that economic growth is inclusive in a general sense: on average, the poverty headcount falls and the incomes of the poorest rise in line with average income growth (see Dollar and Kraay 2002; Kraay 2006; Dollar et al. 2013). However, two recent contributions have reopened this debate. First, Shaffer (2017) notes that in 15–35 per cent of episodes of growth, absolute poverty actually rises with per capita growth and he connects this with historical debates on ‘immiserizing growth’, a term which has been (p.24) used (see Bhagwati 1958) to demonstrate that export-led growth could lead to a country being worse off due to changes in the terms of trade. However, for Shaffer it refers to growth accompanied by no change in, or even increasing poverty.
Second, Sen (2014) concurs with this, finding that there is a surprising number of growth episodes that are not inclusive based on ‘traditional’ interpretations of IG (that of falling poverty with or without non-rising inequality). Sen separates types of growth episodes between ‘growth acceleration’ and ‘growth maintenance’ and finds that the former is much less likely to benefit the poor than the latter. Sen argues that this is because the institutional factors that lead to growth accelerations are different from those that lead to growth maintenance.12 This suggests that during periods of growth, potentially when ST is most rapid—a growth acceleration episode—IG may suffer and only recover when or if that growth acceleration becomes a growth maintenance episode and institutional arrangements change. This points towards a well-noted argument, even in the traditional IG studies, that the average inclusivity of growth can be misleading as it is subject to enormous variation across countries and highly sensitive to where the poverty line is set (see also the discussion of Edward and Sumner 2015; Sumner 2016a, 2016b). Much of the debate turns on whether inequality is high or rising, as high and rising inequality can hamper not only poverty reduction but also future growth prospects, which can impact future poverty reduction.13 This brings us to our second question and a taxonomy of IG.
If we turn next to the basic dimensions for conceptualizing IG in terms of who to include, how much they are ‘included’ relative to others, and in what way they are ‘included’ (implicitly current or future consumption), it is immediately evident that the debate about the distributional pattern of growth raises a number of normative issues: should the poor (by whatever poverty line) see their standards of monetary and non-monetary living improve more than the non-poor? If so, where does one draw the poverty line? Or should the line be at median consumption or even, for example, at say $10 per day, which is a line associated with permanent escape from the risk of falling back into poverty (see López-Calva and Ortiz-Juarez 2014)? That level of daily consumption would capture most of the population of many developing countries. Then, once the line is taken and a normative decision is made to favour those (p.25) below the line more so than those above the line, what are the intra-poor weightings? Again, this is a normative question. If the poor are 90 per cent of the population, their incomes could rise at a faster rate than the top 10 per cent across the 90 per cent, but within the 90 per cent should there be progressive weighting with the strongest weighting on the poorest? These questions point towards the complexity of defining precisely what an ‘inclusive growth’ episode should look like. Further complicating matters is the relative simplicity of the relationship of growth to expenditure or monetary poverty versus the more complex relationship of growth to multidimensional poverty.14
One could say the contemporary concept of IG has several discernible dimensions. All of these IG types have something to say on who is ‘included’ and how much they are included. To recap, the assumption is that the ‘in what way included’ dimension is answered as consumption broadly defined now and/or in the near future. The first dimension or type of IG entails a focus on the benefits of growth to those under a poverty or population fractile line and their absolute or relative income or consumption. There are also sub-types. This definition includes IG as poverty reduction or absolute pro-poor growth. This refers to growth whereby those under a specific poverty line or fractile of consumption (e.g. the poorest 40 per cent) benefit in absolute terms and the proportion of population under that line diminishes. The definition also includes poverty reduction with falling inequality of outcome or relative pro-poor growth. This refers to growth whereby those under a specific poverty line or fractile of consumption (that is, the poorest 40 per cent) benefit more than those above that line (meaning inequality falls).
In contrast, a further dimension or type of IG entails a focus on the benefits of growth to the entire population, or at least a much broader group than those under a specific poverty or population fractile line, and their capabilities in terms of the expansion of education levels or employment prospects. A type II IG is growth that focuses on ‘equalizing opportunities’ ex ante or redistributing opportunity in the future. This is more in keeping with the recent evolution of the field of enquiry. This refers to growth whereby the education attainment expansion is prioritized (and one could add health access policies (p.26) and social protection access) and is configured to ensure an equalizing of opportunity across the whole population. In short, the divergence between skilled and unskilled workers is constrained by expanding the education levels of the lesser skilled. During growth, it is likely that skilled workers will benefit first if growth opportunities are focused in higher skilled employment or sectors. The expansion of education attainment thus provides for a means to constrain the divergence of the skilled–unskilled worker pay differentials by expanding the supply of skilled labour and reducing the supply of unskilled labour. A third dimension or type III IG is full or fuller employment or employment-generating or labour-intensive growth or productivity that translates into employment growth. This refers to a situation whereby growth is accompanied by substantial employment generation as an outcome of productivity growth. This ensures that the opportunities presented by growth are broad-based and productivity gains are not entirely captured in real-wage increases but spread widely across a larger section of workers due to the extent of employment growth.
In sum, defining IG as a reduction in poverty by an absolute poverty line or the income of the poorest fractiles may miss important issues. Next, we consider the relationship between ST and IG. In the following section, we revisit the Lewis model and its distributional consequences for IG as outlined in Kuznets’ seminal work and of others’ since.
2.4 The Relationship Between Structural Transformation and Inclusive Growth
The discussion thus far has alluded to a relationship between ST and IG during the process of economic development without yet specifying a model. The purpose of this section is thus to situate ST and IG in a model of economic development. We do this to understand the relationship between ST and IG. We focus on the dual economy of the Lewis model as a heuristic device or an ‘ideal type’. The model, developed by W. Arthur Lewis, envisaged capital accumulation through a transition driven by the labour movement from the traditional to the modern sector. The writing of Simon Kuznets and others following in the Kuznets tradition and using the Lewis model added the IG dimension by focusing on what happens to inequality during ST.
Why do we take the Lewis model of economic development? Why not another model of economic development? The answer to these questions is largely because the Lewis model of economic development is based on economic development by ST and thus ideally fits with the study of ST and (p.27) further how IG interacts with ST. The Lewis model is appealing for the study within this book not only because of its focus on ‘sectors’ but because of its heterodox history and critique of neoclassical approaches that are indifferent to sectors and activities. It also covers the three dimensions of ST that are of interest (factoral, sectoral, and integrative ST). Palma (2005) refers to three schools of economic development theory.15 The group within which Lewis sits is that which is premised on the notion that economic growth is both ‘sector-specific’ and ‘activity-specific’. What binds this group together is that growth dynamics are dependent on the activities being developed and the capital accumulation effects of manufacturing. Thus, issues such as technology, externalities, balance of payment sustainability, and convergence with advanced countries are a function of the size, strength, and depth of manufacturing. The importance of manufacturing is predicated on the work of Kaldor (1967) who sought to explain the economic development of Western Europe through the development of manufacturing, which he argued was the engine of growth for every country at every stage of economic development. In terms of empirical support for the importance of manufacturing see Duarte and Restuccia (2010). Kaldor posited that: economic development requires industrialization because increasing returns in the manufacturing sector mean faster growth of manufacturing output which is associated with faster economic growth.16
The argument of this section is as follows: first, that the vision of Lewis provides the model of economic development that brings together both ST and IG because other models of growth place little importance on sector analysis. Second, that the Lewis model is best utilized as a heuristic device or ‘ideal type’ on the transition within which Kuznets’ work on distribution dynamics can sit.17
2.4.2 The Lewis Model Revisited
Arthur Lewis (see notably, 1954, 1958, 1969, 1972, 1976, 1979) provided one of the best-known and optimistic models of economic development in developing countries. Although sixty years old in its earliest iteration, the model remains relevant today to developing countries (see for contemporary discussion, Gollin 2014). Since Lewis’s original work on the labour transition between sectors, much literature has been concerned with labour push-and-pull factors leading to various extensions of the model.18 The dual model provides an ideal type, in the Weberian sense, for thinking about structural transformation and economic development with an emphasis on labour, which is the factor of production that dominates most developing countries.
Lewis argued that the driver of capital accumulation was a sectoral movement of the factor of production abundant in developing countries, labour, from the ‘traditional’ or ‘non-capitalist’ sector (of low productivity, low wage, priced to average product not marginal product, and thus with widespread disguised unemployment) to the ‘modern’ or ‘capitalist’ sector (of higher productivity, and where wages are set by productivity in the ‘subsistence sector’.). Crucial is the existence of surplus labour in the traditional or non-capitalist sector.19 Because of this wages are set just above subsistence across the whole economy, leading to the transfer of labour over time from traditional or non-capitalist to modern or capitalist sectors and the capture of labour productivity gains to capitalists as profits as these are the source of growth via reinvestment. The floor for wages is institutionally set at subsistence. Lewis (1954, pp. 151–2) posited that the transition of labour from the traditional to the modern sector was to be understood as follows:
The key to the process is the use which is made of the capitalist surplus. In so far as this is reinvested in creating new capital, the capitalist sector expands, taking more people into capitalist employment out of the subsistence sector. The surplus is then larger still, capital formation is still greater, and so the process continues until the surplus labour disappears.
(p.29) When the surplus labour disappears an integrated labour market and economy emerge and wages will then start to rise. After which marginal productivity determines wages.20 Lewis (1979, p. 211) later noted the ‘wide range of specifications’ to which his dual economy model had been characterized, which led him to reiterate the core elements as his saw them:
The version I am using here has three characteristics. First, there are two sectors, hereinafter called ‘modern’ and ‘traditional’, such that the modern sector grows by recruiting labour from the traditional. Second, unskilled labour is paid more in the modern sector than in the traditional sector for the same quantity and quality of work. And thirdly, unskilled labour is initially abundant in the sense that at the current wage much more labour is offered to the modern sector than that sector wishes to hire.
The Lewis model was intended as a critique of the neoclassical approach in that labour is available to the modern or capitalist sector of an economy not in a perfectly elastic supply but upward sloping rather than flat, and with a distinction between surplus-producing labour and subsistence labour (the latter of which was a negligible source of net profits for reinvestment, which Lewis saw as the driver for growth).21 Lewis also rejected the assumptions of neoclassical economists of perfect competition, market clearing and full employment and Lewis (see 1958, pp. 8, 18) made the distinction between productive labour, which produced a surplus, and unproductive labour, which did not:
A transfer of workers from the latter [unproductive labour sector] to the former [productive labour sector] raises the national income, increases the total surplus over wages, and so makes possible further expansion…So long as unlimited labour is available at a fixed real wage, the share of profits in national income will increase. There are two reasons for this. First the share of profits in the capitalist sector may increase. And secondly the capitalist sector will expand relatively to national income.
In sum, in the Lewis model, growth is sustained by sectoral ST and the transfer of the factor of production abundant in developing countries. That is labour from low productivity to higher productivity sectors.22 There were two sectors: (i) a modern or capitalist sector which is not necessarily synonymous with (p.30) urban or industrial or the private sector though these may be part of this sector, and (ii) a traditional or non-capitalist sector which is not necessarily synonymous with agriculture or the non-capitalist or public sector, though these may be part of this sector.23 The sectors are not necessarily unified geographically.24
Those two sectors are different in terms of the ‘rules’ that apply in each: the rules of the modern or capitalist sector are competition and profit maximization and marginal productivity determines resource allocation; the rules of the subsistence sector are set by social conventions. The former is ‘fructified’ by capital or ‘that part of the economy which uses reproducible capital, and pays capitalists for the use thereof’ and the latter is ‘all that part of the economy which is not using reproducible capital’ (1954, p. 146). The former has higher wages and higher marginal productivity of labour than the latter and is capital intensive. The latter is lower wage, labour intensive, and has an oversupply of labour. Labour migrates and this migration sustains growth and economic development by raising productivity per worker. The capitalist surplus is reinvested in new capital and this expands the sector, which in turn provides more employment. This process continues until the surplus labour is used, or when capital accumulation is faster than population growth, or when the terms of trade between the sectors changes, or when new technology raises production in the subsistence sector. Wages in the subsistence sector start to rise as the Lewis ‘turning point’ is reached, at which point the supply of surplus labour is exhausted. This impacts on profits, the rate of reinvestment, and the capital stock.
Lewis did not ignore inequality. Indeed, Lewis (1954, p. 147) highlighted the high visibility of inequality, ‘between the few highly westernized, trousered, natives, educated in western universities, speaking western languages, and glorying Beethoven, Mill, Marx, or Einstein, and the great mass of their countrymen who live in quite other worlds’.25 Inequality was of importance to Lewis to the extent that it had an impact on output growth. Lewis generally discussed rising inequality in terms of functional distribution rather than (p.31) household or individual income distribution, as that was central to the model, and a rise in the share of capital drives growth but does not necessarily imply a change in the within-labour share. If there were a larger share of profits in national income this would mean more resources for capital formation but as Lewis (1954, pp. 157, 158) noted
[t]he central fact of economic development is that the distribution of incomes is altered in favour of the saving class…All that the workers get out of the expansion is that more of them are employed at a wage above the subsistence earnings.
When the Lewis turning point is reached and surplus labour is exhausted wages would rise and the functional distribution of income would move in favour of labour though this would slow or end the transition. At which point labour markets would be unified not dualistic.
Lewis posited that the state could substitute for a capitalist class (where no such class was of a significant size) in order to deal with what he referred to as ‘the sociological problem of the emergence of a capitalist class’ (1954, p. 159). Further, the state needs to play a strong role in disciplining the capitalist class when it does exist, or guiding the emergence of such a class.26 The role of the state was also important because during the Lewis transition inequality may rise and public policy intervention is needed because,
the Gini coefficient may actually show a rise in inequality, since the share of national output accruing to the bottom 50 per cent may fall…To tax its developed sectors and subsidise its underdeveloped sectors is one of the most powerful ways that a government can use to ensure the benefits of development…The moral for policy makers is of course not to rely on trickle down to benefit the traditional sector, but to attack the problems of that sector directly (1979, pp. 212, 216).
Lewis (1976) presents an explicit framework to consider this relationship between growth and distribution, noting a starting point that,
growth takes place in enclaves, surrounded by traditional activities…Development must be inegalitarian because it does not start in every part of an economy at the same time. Somebody develops a mine, and employs a thousand people. Or farmers in one province start planting cocoa, which will grow only in 10% of the country. Or the Green Revolution arrives, to benefit those farmers who have plenty of rain or access to irrigation, while offering nothing to the other 50% in drier regions
(p.32) Lewis (1976) discusses the relationship between economic development and distribution as one based on within and between sector inequality. He argues that the growth of the modern or capitalist sector, or the ‘enclave sector’ as he calls it in that paper, has good and bad impacts on the traditional sector (p. 27). Notably, the enclave may enrich the traditional sector by buying commodities and services from it; providing employment to those in the traditional sector; sending remittances; selling goods and services cheaper; and by developing infrastructure, public goods and, through an example of new ideas and institutions, the enclave sector can modernize the traditional sector. Whether development leads to widening inequality depends, he argued on whether the enclave is able to respond to the new economic opportunities (e.g. price changes or the demand for labour). In short, inegalitarian development is not the failure of ‘trickle down’ vertically from rich to poor but the failure to trickle along or spread horizontally the benefits from enclave to traditional sectors.27
Again, the role of the state is highlighted by Lewis (pp. 30–5) who posited that distribution in the enclave depends on the pattern of growth and a set of factors, many of which are ‘susceptible to public control’ (p. 35), notably the distribution of property, economic structure (in terms of firm size and the capital intensity of production and dependence on foreign resources) and the speed of growth which has the potential to alter ‘the relative quantities of the factors of production, and the derived demands, and therefore the distribution of income’. Further, the traditional sector may see income stagnate because the enclave may be predatory (e.g. driving people off their land); products may compete with traditional trades; the wage rate in the enclave may be too high and raise the price of labour above its marginal productivity; because of geographical polarization (the enclave attracts ‘best brains’ and capital); because population growth accelerates due to improved public health reducing the death rate; and/or excessive migration from the countryside. Lewis concludes (1976, p. 29) that whether the enclave enriches or not the traditional sector ‘probably depends most on whether the government coerces or helps the traditional sector, and on the nature of the enclaves’ (meaning the modern or capitalist sector).
There have been various critiques of the Lewis model, many of which are of a ‘red herring’ variety as Ranis (2004, p. 716) puts it, meaning they are easily (p.33) responded to or actually criticisms of Lewisians rather than the writing of Lewis himself. Many relate to the assumption of labour abundance in the subsistence sector (and thus the dominance of the wage from that sector across the economy), and the emergence of the urban informal sector, although Lewis’s conception of surplus labour explicitly included the urban informal sector.28
It is clear Lewis (1954, p. 141) did not ignore the urban informal sector in the unlimited supply of labour concept:
The phenomenon is not, however, by any means confined to the countryside. Another large sector to which it applies is the whole range of casual jobs—the workers on the docks, the young men who rush forward asking to carry your bag as you appear, the jobbing gardener, and the like. These occupations usually have a multiple of the number they need, each of them earning very small sums from occasional employment; frequently their number could be halved without reducing output in this sector.
Informality was taken a step further in Ranis and Stewart (1999) who developed a model of dualism within the urban informal sector between a dynamic sub-sector linked to the formal sector and a less dynamic ‘sponge’ (meaning highly labour absorbing) sub-sector.
There are other critiques of the Lewis model (see discussion in Fei and Ranis 1964; Harris and Todaro 1970; Minami 1973; Schultz 1964; Rosenzweig 1988; Todaro 1969). One relates to the oversimplicity of a two-sector model. However, as Basu (1997, pp. 151–2) notes ‘[T]he assumption of duality is merely for analytical convenience. If fragmentation—irrespective of the number of parts—in itself causes some problems and we wish to examine these, then the simplest assumption to make is that of dualism.’
Neoclassical economists would not accept that wages were set by any other mechanisms other than demand and supply. Contentions arising were related to labour use and agricultural productivity and the assumption of zero marginal productivity in agriculture (that said the Lewis model actually rests upon the elastic supply of labour itself).
Finally, there has been an incorrect view that the Lewis model takes little account of the integrative aspects of ST and open economies and thus contemporary globalization and global economic integration. This point is absolutely a misperception and based on the notion that the seminal Lewis (1954) piece outlining the basic Lewis closed economy model only contained a small section on an open economy. However, the role of external trade, and investment and finance are discussed in the 1954 paper and are highly evident in many (p.34) other writings of Lewis, given his interest in primary commodity-exporting countries. The main focus in the original (1954) work related to integrative ST is where it is argued that the Lewis turning point of surplus labour exhaustion may be delayed by international migration and capital export. The closed economy versions of the Lewis model (the first and the second) were building blocks to get to the third model (the open economy model), which Lewis believed represented most developing countries. It is the third model, the one that explains the tendency for declining factoral terms of trade, which was a major concern for Lewis. Furthermore, an entire section is dedicated to critiquing comparative advantage based on his open economy model.29 Elsewhere (Lewis 1976) a real concern of Lewis was—with great foresight—that primary exports dependency would eventually become a new dependency on a handful of manufacturing exports which would fall in value relative to import cost. Thus dependency on a few primary commodities whose relative price was falling vis-à-vis import needs, would be replaced with dependency on a few manufactures whose relative price was falling vis-à-vis import needs.
A set of contemporary challenges throws up greater levels of complexity. First, is that domestic labour migration may not be permanent but circular (back-and-forth) or ‘commuting’. This means a worker may be active in both ‘traditional’ and ‘modern’ sectors. For example, non-farm rural income is generally estimated to be a substantial part of rural incomes, suggesting too that workers are active in and beyond the ‘traditional economy’ at a point in time (see Booth 2016; Wee and Jomo 2014). Second, that the contemporary scale of inter-sectoral resource flows via the growth of remittances further blurs the line between sectors with economic development and distributional impacts. Finally, that the Lewis transition can take a variety of forms beyond the anticipated one by Lewis and it is by no means guaranteed that the transfer will be from low to high productivity activities as flagged by MacMillan and Rodrik (2011). A transfer from low productivity agriculture to low productivity services has been the experience of many developing countries and a reversing of the Lewis transition has also been a phenomenon noted in a number of developing countries in ‘premature deindustrialization’. In short, multiple pathways of ST are possible and not all are progressive.
If one focuses on four economic sectors alone, there are six potential modes of inter-sectoral ST: agriculture to non-manufacturing industry, agriculture to (p.35) manufacturing, agriculture to services, non-manufacturing industry to manufacturing, non-manufacturing industry to services, and manufacturing to services. To this one could add four modes of intra-sectoral ST.
Nevertheless, the Lewis model does provide a framework to situate the developer’s dilemma and the inequality dynamics or tension between structural transformation and inclusive growth if one considers that structural transformation can take place in multiple ways. Lewis considered distributional dynamics in terms of the functional distribution of income and also the relationship between sectors as the discussion above highlights. Each mode of structural transformation is likely to have its own distribution dynamics as labour moves between or within sectors. The distribution dynamics of such movements were also of concern to Simon Kuznets whose seminal work we revisit next.
2.4.3 Kuznets Revisited
We next consider the contribution of Simon Kuznets, and those writing in the Kuznets tradition, in order to introduce inequality dynamics into a Lewis-type two-sector model. Kuznets focused, like Lewis, on the factor abundant to developing countries, that of labour and its sectoral transfer from rural to urban areas.
Often, one tends to find that in public policy discourse Kuznets’ thinking is reduced to his famous curve alone. However, there is much more depth and many more nuances and caveats than are often recognized (see for more discussion, Kanbur 2012, 2017). Kuznets (1955) took a dual economy model somewhat like Lewis (though with a focus more so on urban–rural rather than modern and traditional sectors) though he posited that a labour transition from a rural sector to an urban sector would be accompanied by rising inequality in the early stages of development. In contrast, neoclassical economic theory would posit that the relative scarcity of factors determines inequality outcomes. The most important point to note regarding Kuztnets’ seminal work is how tentative it is, and the number of disclaimers due to the scarcity of data available at that time. Kuznets builds theory through an abstract arithmetic model and then draws what he notes are very tentative conclusions. Kuznets’ original thesis was based on time-series data for three countries (US, UK, and two states in Germany) plus point estimates for inequality in India, Puerto Rico, and Ceylon. He argued that inequality would rise in an ‘upswing’ and then fall later in the ‘downswing’ of what became known as the inverted-U or Kuznets curve.
Kuznets (1955, pp. 6–7) outlined two forces that would increase pre-tax and transfer inequality in the long term:
The first group relates to the concentration of savings in the upper-income brackets…The second source…lies in the industrial structure of the income (p.36) distribution. An invariable accompaniment of growth in developed countries is the shift away from agriculture, a process usually referred to as industrialization and urbanization.
The former of these two forces resonates with Lewis’s view on the savings class and the latter to Piketty’s (2014) focus on capital accumulation as the driver of rising inequality.30
Kuznets (1955, pp. 7–8) argued that inequality would rise as the inter-sectoral shift away from agriculture leads to income differences between sectors, and changes within each sector, as some of the population leave the more equal sector.31 Inequality, Kuznets posited, is composed of inequality between and within sectors (urban and rural), and inequality tends to be lower in the rural sector (relative to the urban sector).32 Thus, as the size of the more unequal urban sector increases, this will further add upward pressure on inequality and, given that during economic growth, productivity in urban areas is likely to increase faster than in rural areas, this will add even more upward pressure on inequality.
In short, inequality in the dual sector economy is an aggregation of (i) inequality in each sector (be that urban and rural or traditional and modern ‘sectors’); (ii) the mean of each sector; and (iii) the population shares in each sector (Kanbur 2017). Thus, even the population shift itself could raise inequality as Kuznets (1955, pp. 14–15) himself noted:
[E]ven if the differential in per capita income between the two sectors remains constant and the intra-sector distributions are identical for the two sectors, the mere shift in the proportions of numbers produces slight but significant changes in the distribution for the country as a whole.
Kanbur and Zhuang (2013) show how the between urban–rural and within urban–rural components of overall inequality can differ considerably between countries, and how the contribution of urbanization itself to inequality changes at the national level can also differ considerably between countries.
Although inequality may rise as a result of movement between sectors, that occurrence may be balanced or outweighed by what happens to the (p.37) within-sector components and the shares of each sector. Initial inequality between and within sectors will also play a significant role.
Although largely rejected as a universal law in studies in the 1990s, the Kuznets hypothesis remains relevant to ST and IG as a backdrop to any discussion.33 Piketty (2006) is broadly illustrative of critiques of Kuznets. He argues the following:
The reasons why inequality declined in rich countries…do not have much to do with the migration process described by Kuznets…Inequality dynamics depend primarily on the policies and institutions adopted by governments and societies as a whole.
(pp. 2, 11)
As Kanbur (2017) notes, this is unfair to Kuznets. Not only does Kuznets discuss countervailing forces (for example, the forces noted previously) but also institutional aspects are prominent for Kuznets (1955). For example,
One group of factors counteracting the cumulative effect of concentration of savings upon upper-income shares is legislative interference and ‘political’ decisions. These may be aimed at limiting the capital accumulation of property directly through inheritance taxes and other explicit capital levies. They may produce similar effects indirectly […]. All these interventions, even when not directly aimed at limiting the effects of capital accumulation of past savings in the hands of the few, do reflect the view of society on the long-term utility of wide income inequalities. This view is a vital force that would operate in democratic societies even if there were no other counteracting factors…Furthermore, in democratic societies the growing political power of the urban lower-income groups led to a variety of protective and supporting legislation.
(pp. 8–9, 16–17)
There is resonance here with Polanyi’s (1957) ‘double movement’ or the dialectical process of the expansion of marketization and the push back against marketization in a ‘counter-movement’. Polanyi argued that liberal reforms seek to establish and expand a market society where all things are commodified, the economy is ‘disembedded’, and society is subordinate to the market economy. This triggers a reaction of ‘counter-movement’ whereby the creation of institutions such as labour law seeks to re-embed the economy in society.
Kuznets, in his seminal piece, said little about the integrative ST dimensions. There has been a set of contemporary scholars building new theory in (p.38) the Kuznetsian tradition, many of whom have focused on such integrative aspects of ST. Such scholars have developed theory with a focus on open economies and agrarian liberalization, the role of technology, as well as national aspects of domestic political economy and land distribution. For example, Galbraith (2011) argues that it is global forces that have driven the changes in national inequality since 1970. The key drivers of the changes in national inequality are world interest rates and commodity prices (and between-sector terms of trade). He argues that a commodity boom reduces inequality in countries with a dominant agricultural sector as it raises the relative income of farmers, and higher rates of interest are bad for debtor countries and this increases inequality. Galbraith presents an ‘augmented Kuznets curve’, or S-curve, whereby the curve rises, then falls, and then rises again.34
In a somewhat similar vein, at least in the sense of a focus on open economies, Lindert and Williamson (2001) argue that it is the shift towards market orientation (domestic to export) of agriculture and not the shift from agriculture to manufacturing and services that causes inequality to rise. Lindert and Williamson predict an initial rise in inequality. However, while Lewis and Kuznets envisaged a downswing, Lindert and Williamson argue that inequality continues to rise because income in the urban sector outpaces rises in income in the rural sector as agriculture shifts to market orientation.
In contrast, Roine and Waldenström (2014) suggest a new Kuznets curve based on technological developments starting not a sectoral shift of agriculture to industry but a shift from traditional industry to technologically intensive industry. If a given technology makes skilled workers more productive and there is an increase in the relative demand for those workers, the rewards accrue to a small proportion of the population who are skilled workers. Based on Tinbergen’s (1974, 1975) hypothesis that the returns to skills are a competition between education and technology, the supply of skilled workers then determines whether or not their wages rise. Roine and Waldenström (2014) argue that the drivers of the Kuznets downturn were political and exogenous shocks.35
(p.39) Oyvat (2016) argues that it is agrarian structures—land inequality—that are deterministic. Consistent with Kuznets, he argues that migration is driven by higher urban incomes and this suppresses wages in the urban sector. If land inequality is higher, more people will migrate for lower wages as they do not own land or own small plots, and rural incomes are lower which will further depress urban wages.36 Empirically, Oyvat argues that the level of land inequality has a significant impact on urbanization, intra-urban inequality, and overall inequality. The results suggest that land reforms or subsidies to rural smallholders would thus reduce urban inequality.
Acemoglu and Robinson (2002) discuss the political economy of the Kuznets curve in two models of late capitalism. The first is a high inequality, low output model that they call ‘autocratic disaster’. In this model, inequality does not rise and political mobilization is too limited to address existing inequality. A second model is the ‘East Asian miracle’ of low inequality and high output where inequality does not rise in order to ensure political stability and avoid the discipline of democracy being forced on elites. They argue that when the process of industrialization does increase inequality, this leads to the political mobilization of the masses that are concentrated in urban areas and factories. Political elites thus undertake reform to ensure their continued position at the top. The extension of the franchise is the best option for elites as it acts as a commitment to future redistribution and thus prevents unrest.37
In sum, there have been various attempts at new theory-building on how income inequality evolves with economic development. As with Lewis, a set of contemporary challenges throw up greater levels of complexity. First and foremost is the experience of premature deindustrialization in developing countries.
ST has been associated with deindustrialization in the developed world. This typically means that there is a focus on the shrinking proportion of industrial or manufacturing activity in GDP, employment, or exports. A key question is what drives such processes and also if they matter (see for discussion, Rowthorn and Ramaswamy 1999; Singh 1977)? Such discussions were extended to developing countries in Palma (2005) and Rodrik (2015), though Singh (1977) and (p.40) others have noted that shrinking shares are not necessarily a cause, but rather more a symptom of underlying structural economic problems.
This phenomenon, or what Palma (2005) and Rodrik (2015) labelled as ‘premature deindustrialization’ with reference to developing countries, is that developing countries have reached ‘peak manufacturing’ in employment and value-added shares at a much earlier point than advanced nations.38
Premature deindustrialization has two components. The inverted-U pattern of manufacturing shares versus GDP per capita shifts down and leftwards over time, making it harder for late developers to attain the benefits of industrialization that earlier developers saw. The first component is that ‘peak manufacturing’, in employment or GDP shares (or export shares), has been reached and the inverted-U curve is now on the plateau or even downswing of the curve. The second component is that that inverted-U curve is moving leftward over time. This means the point at which the inverted-U turns is, on average, lower in per capita income terms now than in the 1990s, which was already lower than in the 1980s (see Palma 2005).
What are the causes of the phenomenon? There are differing views. Rodrik (2015) links the phenomenon to trade liberalization over time and the impact of China’s entry into manufacturing. One could also potentially add automation and technological change. Felipe et al. (2015) argue that premature deindustrialization is caused by the fact that large national increases in labour productivity were counteracted by a shift of manufacturing jobs to lower productivity economies.39 They note that the global shares of employment and GDP in manufacturing have changed very little in the last forty years. What has happened is that international competition has spread what manufacturing there is across more countries.
Palma (2005) argues that there are several other potential hypotheses (which are not mutually exclusive) that could explain the phenomenon observed: (i) it is due to a statistical illusion caused by contracting out of manufacturing jobs to services (for example, cleaning or catering); (ii) it is due to a fall in the income elasticity of manufactures; (iii) it is due to higher (p.41) productivity growth in manufacturing; or (iv) it is due to outsourcing globally whereby manufacturing employment has fallen in OECD (Organisation for Economic Co-operation and Development) countries; (v) it is due to the change in policy regimes in OECD countries away from Keynesianism; or (vi) it is due to technological progress.40
Whatever the causes of premature deindustrialization, it is empirically visible though a question remains as to how or why it matters and if the service sector really is inferior to manufacturing output, employment, and exports. That said, in light of its visibility alone, the distribution dynamics of premature deindustrialization warrant further exploration.
The purpose of this chapter was to set out the conceptual points of departure in terms of ST and IG, and to situate the two in a model of economic development. We have thus discussed the concepts of ST and IG. We have revisited the Lewis model of economic development and the contribution of Kuznets and recent iterations.
To recap, this chapter has argued the following: (i) that ST and IG have tended to be approached in a reductionist sense and disconnected from each other. By defining down each to a more minimalist definition, one loses important aspects of each and of their interrelationship; (ii) that the vision of Lewis and Kuznets provides a two-sector model for connecting ST and IG that can be related to the sectoral, factoral, and integrative aspects of ST and the income, inequality, education, and employment aspects of IG. However, Lewis and Kuznets require contextualizing to the contemporary period in which a set of ‘new’ issues has gained prominence. Notably, multiple forms and dimensions of structural change, ‘premature deindustrialization’ and circular (back-and-forth) or ‘commuting’ labour movements; as well as non-farm rural income; the growth of inter-sectoral resource flows via remittances and the multiple dimensions of inclusive growth.
We have noted that Lewis focused on the inter-sectoral labour reallocation and related it to the functional distribution of income and inequality dynamics in terms of the relationship between sectors. Kuznets was concerned largely with the labour share in terms of the relationship between his sectors. Both were concerned with the sectoral aspects of ST as well as factoral ST.
The essence of Kuznets was that the labour transition could, unless governments intervene, be unequalizing if people move from the more equal sector (p.42) (which Kuznets took to be rural in developing countries) to the less equal sector (which Kuznets took to be urban) because of the differences in relative income and the relative population weighting. All else being equal, the Kuznets hypothesis will only hold if the rural sector is more equal or if governments do not intervene to neutralize the forces of inequality that were unleashed by ST.
On the other hand, the transition could well be equalizing if the rural sector is more unequal (or if governments intervene). However, it will depend on (i) where in the distribution of each sector the migrant starts out from and ends up in and (ii) on how equalizing the wage gain they get from the urban sector is overall, and furthermore, (iii) the extent of counterbalancing forces due to the movement of surplus labour from the traditional sector, which would lead the average income in that sector to rise and generate a pattern that could be equalizing.
In short, if developing countries seek economic development via ST and seek to make that inclusive form of growth then the Kuznetsian forces unleashed by ST require neutralizing. ST and capital accumulation tend to generate distribution tensions, these can be managed to some considerable extent by a focus on expanding public policy intervention. Specifically, addressing the urban–rural income divergence; the skilled–unskilled worker wage divergence; and the capture of productivity gains vis-à-vis capital and labour shares and the within-labour shares, and the extent to which productivity growth is translated into real-wage growth or employment growth. This would suggest ST with IG requires—counter-intuitively—highly activist rural and agricultural policy and public investments to constrain the urban–rural income divergence; large investments in public education to constraint the skilled–unskilled worker wage divergence; and state–capital–labour pacts to ensure substantial employment growth is the consequence of productivity growth.
In conclusion, this chapter has characterized ST as shifts towards higher productivity activities with aspects of that transformation that are: (a) sectoral, (b) factoral, and (c) integrative. The chapter has defined IG by income/consumption and capabilities, notably as (a) absolute poverty reduction with or without inequality of outcome reduction; (b) inequality of opportunity reduction and raised education levels; and (c) the direct inclusion in growth via employment growth. The chapter has also outlined the relationship between ST and IG using the Lewis model as a heuristic device/‘ideal type’, and revisited Kuznets and those writing since in the Kuznets tradition who have sought to include in their thinking such factors as agrarian structure, global integration, government regimes/policies, and technology/skills.
The discussion takes us to the essence of the developer’s dilemma: If developing countries seek economic development via ST and seek to make that (p.43) inclusive form of growth then the Kuznetsian forces unleashed by ST require neutralizing. There is the divergence between capital and labour and who captures productivity gains, the divergence within labour shares of workers in ‘sectors’ (urban versus rural for simplicity) and by skills (unskilled and skilled labour for simplicity). If developing countries seek to address the distributional tension between ST and IG then countervailing interventions—many of which Lewis noted in his own writing—are required.
Chapter 3 expands on these points and discusses theories of the broader drivers and impediments to the Lewis transition, their distributional consequences, and in particular why it is that developing countries have not followed the ideal-type pathway or transition that the Lewis model envisaged. The chapter also draws upon recent literature on the heterogeneity of late capitalism and provides a rationale for the analytical framework of this book, which is outlined and discussed as the culmination of the chapter.
(1) Baymul and Sen (2017) make use of the Timmer et al. (2015) dataset and the United Nations University World Institute for Development Economics Research (UNU-WIDER) World Income Inequality Database (WIID) (UNU-WIDER forthcoming). They find that a movement of workers away from agriculture is unambiguously associated with rising income inequality. However, in most cases, the shift of employment is from agriculture to services, not agriculture to manufacturing.
(2) Though there is only one section explicitly on the open economy in the original exposition of the Lewis (1954) model, more is evident implicitly and Lewis, himself, elucidated more in later works. Specifically, an open economy would delay the Lewis turning point where surplus labour was exhausted and wages would rise. Lewis (1954) argued, using an open economy and factoral terms of trade, that the benefits of productivity growth accrue to Northern importers of exports of developing countries by way of lower prices because wages are set in the subsistence sector where there is an unlimited supply of labour.
(3) It is not given that structural change will always be growth-enhancing as Diao et al. (2016) note, and Baymul and Sen (2017) concur.
(4) See Herrendorf et al. (2013), and earlier reviews of structural transformation, notably Matsuyama (2008), Ray (2010), and Greenwood and Seshadri (2005). One can also note the works of Clark (1957), Chenery (1960), Kuznets (1966), and Syrquin (1986).
(5) On balanced growth, the key seminal (and first) promoter of this idea was Rosenstein-Rodan. Scholars are/were divided on the need to which growth should be balanced. For example, Singer and Hirschman, for example, both argued that growth need not take place in an inter-sectorally balanced way. In contrast, Kalecki (1954, 1955, 1967) argued that ‘balanced development’ in the agriculture and industrial sectors is required. Kalecki argued that the inter-sectoral terms of the trade between the agriculture and industrial sectors need managing so that the terms of trade do not undermine the agriculture that provides food, and that failure to do so would slow down industrialization as it would depress real wages in industry and profits. Lewis also noted,
In unenlightened circles agriculture and industry are often considered as alternatives to each other. The truth is that industrialization for a home market can make little progress unless agriculture is progressing vigorously at the same time, to provide both the market for industry, and industry’s labour supply. If agriculture is stagnant, industry cannot grow.
(Lewis 1953, cited in Kanbur 2016, p. 6)
(6) For a contrarian view see Roy (1951). Specifically, the causality could run from growth to ST, meaning ST is an outcome of growth not a cause of growth; what appear to be sectoral differences in productivity are not necessarily due to sectors but could be skills, for example; further that inter-sectoral relocations may occur without changes in productivity gaps between sectors. One needs to show not that incomes are higher in urban than in rural areas (or different sectors) but that the income of the same individual would be higher in an urban area than in a rural area if they migrated. Importantly, if Roy (1951) is correct then the transition of labour from low productivity to higher productivity sectors will not reduce inequality; rather, it will increase it, because the same person would earn less in the high productivity sector.
(7) McMillan and Rodrik (2011) find that countries with a large share of exports in natural resources tend to experience growth-reducing structural transformation and, even if they have higher productivity, cannot absorb surplus labour from agriculture. In a similar vein, Gollin et al. (2016), too, argued that natural resource exports drive urbanization without structural transformation because natural resources generate considerable surplus which is spent on urban goods and services, and urban employment tends to be in non-traded services. McMillan and Rodrik (2011) also find that an undervalued (competitive) exchange rate, which operates effectively as a subsidy on industry and labour market characteristics (so labour can move across sectors and firms easily), leads to growth-enhancing structural transformation.
(8) In a similar vein, Diao et al. (2017) argue that the most recent growth accelerations in the developing world, unlike East Asia’s historical experience, have not been driven by industrialization but by within-sector productivity growth (in Latin America) and growth-increasing structural transformation, but this has been accompanied by negative labour productivity growth within non-agricultural sectors (in Ethiopia, Malawi, Senegal, and Tanzania).
(9) Seers (1963, pp. 79–80) noted the ‘special case’ of ‘First World’ or developed industrial economies as, ‘by no means typical’ and developing countries as having a ‘largely unindustrialised economy, the foreign trade of which consists essentially in selling primary products for manufactures’. He argued that orthodox economics pays too little attention to the role of the export sector and that developing countries cannot be understood without reference to the global economy, and how the public sector automatically compensates for fluctuations in the private economy in industrial economies, which is much harder or impossible for developing countries.
given the normal levels of the other components of the balance of payments, we may define an efficient manufacturing sector as one which…not only satisfies the demands of consumers at home, but is also able to sell enough of its products abroad to pay for the nation’s import requirements. This is, however, subject to the important restriction that an ‘efficient’ manufacturing sector must be able to achieve these objectives at socially acceptable levels of output, employment, and the exchange rate…In operational terms, a structural problem can arise in this sense, if the manufacturing sector, without losing price or cost competitiveness, is unable to export enough to pay for the full-employment level of imports.
(11) The term ‘shared prosperity’ is one more label sitting under the umbrella of ‘inclusive growth’. It is a term that is again associated with the World Bank (for example, World Bank 2016) though it need not necessarily be.
(12) In fact, the average relationship during a growth acceleration episode was negative for the poorest quintile (and, on average, the Gini coefficient rises) and the poorest 20 per cent were worse off in a range of countries including Bangladesh, China, and Nigeria, each home to large populations of the world’s poorest. In contrast, during a growth maintenance episode, the income of the poorest quintile, on average, rises and the Gini falls.
(13) The debate in the literature on the relationship between inequality and growth received a detailed review in Cunha Neves and Tavares Silva (2014). Although numerous methodological issues remain, inequality may support growth at low levels of average income, but rising or high inequality can hamper growth at middle-income levels.
(14) In this book, focus is on consumption/income/monetary poverty (though not exclusively). This is not to say poverty is not multidimensional. Clearly it is. The link, however, between monetary poverty and ST is clearer than that between multidimensional poverty and ST. The relationship between multidimensional poverty and growth is more complex than the mathematical identity for monetary poverty. Santos et al. (2016) find the relationship between multidimensional poverty and growth much weaker than monetary poverty and growth. Rising incomes among the monetary or multidimensional poor can lead to improved nutrition intake and outcomes, or improved access to education and health and outcomes, but public spending may be important in terms of the provision of free or subsidized public education and health. Social policy, such as redistributive transfers, can further support the reduction of both monetary and multidimensional poverty (see, for a discussion of countries with multidimensional poverty data over time, Alkire et al. 2015).
(15) Palma (2005) identifies the neoclassical school, which is indifferent to both sectors and activities. The Solow convergence models (traditional and augmented), ‘endogenous’ models based on increasing returns, are examples of this school. Additionally, models based on market imperfections in technological change are a result of the production function. A second school is ‘sector-indifferent’ but ‘activity-specific’, which Palma associates with Roemer and the neo-Schumpeterians who argue that research and development matter but that there is nothing special about manufacturing in terms of Kaldor-effects. A third school is activity and sector specific and includes Lewis. Diao et al. (2017, pp. 3–4) seek to link the structural dualism of Lewis with the neoclassical model by arguing that the neoclassical model shows the growth process within the modern sector and the dual model shows the relationship among sectors.
(16) This is because backward and forward input–output linkages are strongest in manufacturing, and the scope for capital accumulation, technological progress, economies of scale, and knowledge spillover is strong. Further, there is a strong causal relationship between manufacturing output growth and labour productivity because of a deepening division of labour, specialization and learning-by-doing, and the scope for productivity gains is large due to economies of scale.
(17) Kirkpatrick and Barrientos (2004, p. 688) for example, referred to the Lewis model as ‘an illuminating framework within which to discuss the reality of the process of development, not taking the homogeneity of its sectors literally, but looking behind this to uncover their internal workings and heterogeneity’.
(19) Lewis believed in contrast to Asia that Africa had a labour shortage due to agricultural land availability. The constraint to growth in Africa was low agriculture productivity rather than manufacturing growth and required government intervention in agriculture (See Kanbur, 2016, p. 7).
(20) Lewis later (1979, p. 211) dropped the term ‘surplus labour’ for fear of the ‘emotional distress’ it causes to some and reiterated that ‘the idea intended to be conveyed is that of an infinitely elastic supply of labour to the modern sector at the current wage’.
(21) As Lewis (1984, p. 132) put it,
[A]ll one needed to do was to drop the assumption—then usually (but not necessarily) made by neoclassical macroeconomists—that the supply of labour was fixed. Assume instead that it was infinitely elastic, add that productivity was increasing in the capitalist sector, and one got a rising profits share.
(22) Lewis’s thinking was influenced by his predecessors, notably Rosenstein-Rodan (1943, p. 202) who was concerned with similar themes such as the agrarian ‘excess population’ or ‘disguised unemployment’ in agriculture and the transfer of this low- or zero-productivity population to the industrial sector, some years prior to Lewis’s own writing.
(23) Fields (2004) expresses a preference for the terms ‘formal’ and ‘informal’ sectors.
What we have is not one island of expanding capitalist employment, surrounded by a vast sea of subsistence workers, but rather a number of such tiny islands…We find a few industries highly capitalized, such as mining or electric power, side by side with the most primitive techniques; a few high class shops, surrounded by masses of old style traders; a few highly capitalized plantations, surrounded by a sea of peasants.
(25) Lewis also wrote on horizontal inequality in terms of class, gender, and ethnicity (see Mosley et al. 2004, pp. 758–9 for discussion).
(26) Lewis (1955, p. 408) was critical of the state too, listing nine ways in which the state can harm the process of economic development through excessive laissez-faire as well as excessive control or excessive spending and failing to maintain order, plundering citizens, exploiting one class over another, constraining trade, neglecting public goods, and expensive wars.
(27) Lewis (1979, pp. 212–15), too, drew attention to several possible ways the modern sector might benefit the traditional sector: provision of employment; sharing physical facilities; modernization of ideas and institutions (Lewis cites new technologies introduced, girls attending school, land tenure systems changing, for example); and through trade (if the modern sector depends on the traditional sector for part of its needs, for example, for food or raw materials, the expansion of the modern sector will rely on the expansion of commodities in the traditional sector, but the traditional sector could be damaged by buying imports from the modern sector or abroad.
(28) This critique is really a critique of the Todaro model, which introduced the necessity of being physically in an urban area to be hired in a formal sector job.
(29) Kaldor (1967) also took the two-sector model to be applicable to trade between developing and developed countries through the export of agriculture products from the former, and import of manufactured goods from the latter. He argued that international trade could make developing countries poorer because liberalization would increase agriculture exports, which are produced at decreasing returns. These returns are not sufficient to compensate for the loss of manufacturing exports, which is a sector that produces increasing returns (this is essentially the same as the Prebisch-Singer hypothesis which is discussed in detail in Chapter 3).
(30) More specifically Piketty’s theory is that rising inequality in the contemporary period is explained by slow growth (and hence a diverging of wages and returns to capital).
(31) Kuznets argued that the early benefits of growth go to those with capital and education but, as more people move out of the traditional sector, real wages rise in the modern sector and inequality falls. He argued that the poorest lost out more rapidly than other groups as income-expanding opportunities arose away from agriculture.
Kuznets argued that the only way to offset this was for the share of lower non-agriculture income groups to rise. He further contended that, in democracies, urban migrants would become politically organized, leading to redistribution.
(33) Anand and Kanbur (1993a, 1993b) dismissed the Kuznets curve. Deininger and Squire (1988) found evidence of the curve in some countries and not others. Most recently, Gallup (2012) empirically posits an anti-Kuznets curve or non-inverted-U: inequality declines and then rises. Alvaredo and Gasparini (2015) do, very cautiously, identify a U-shaped curve using 2005 PPP data but the upswing of the curve is entirely in sub-Saharan Africa and the downswing is entirely in HICs.
Indeed, Kuznets himself only found a rising trend or upswing in the data he used, not an inverted-U (his downswing of the inverted-U was based on theory and a data simulation).
(34) Galbraith (2011) argues that national inequality tends to follow similar trends around the world and that there have been four phases over time: (i) a first period, from 1963 to about 1971, of relative stability in national inequality; (i) a second period from 1972 to 1980 when inequality declined slightly in much of the world due to the post-Bretton Woods inflationary boom based on extensive lending at negative real interest rates; (iii) a third period dated from 1982 to about 2000, consisting of sharply rising inequality due to the debt crisis, the collapse of many communist countries, and the liberalization of the 1990s which led to a fiscal squeeze and public sector retrenchment. There are some exceptions. China and India did not see the rise in national inequality until the 1990s and liberalization, because they did not liberalize their financial markets; (iv) a fourth period which began in 2000 and is one of modest declines in national inequality due to the slowing down of liberalization and the commodity boom.
(36) Oyvat argues that agrarian structures in Asia tend to have more owner-cultivators and tenants and thus small- and median-scale family farms rather than Latin America or sub-Saharan Africa, which tend to have high land inequality and large plantation structures that hire wage labour, and very small family farms. Thus, in Latin America and sub-Saharan Africa, the lack of sufficient formal employment opportunities for migrants generates an urban reserve army of labour in the urban subsistence sector, which depresses wages in the urban modern/capitalist sector.
(37) Examples given are those of the UK and France where extension of franchise led to changes in labour market institutions, mass education, and reduced inequality. The thesis of Acemoglu and Robinson (2002) is that capitalist industrialization increases inequality but that this induces a change in political regime towards more redistribution.
the surest way to run into trouble is to have ‘de-industrialization’ (industrial employment growing more slowly than the labour force), since this means that the reservoir of cheap labour will be filling instead of emptying. The political and social health of the community, no less its economic health, requires a continual transfer from the reservoir to the more productive sectors, rather than the relative expansion of the reservoir.
(39) So the average employment share in manufacturing that could be achieved has fallen over time, and countries have experienced deindustrialization earlier than they used to. In short, the changes in supply chains and the shift to lower productivity economies (for example, China) has spread manufacturing jobs more thinly, making it harder for individual countries to sustain high levels of manufacturing employment.
(40) Palma goes on to argue countries that have a commodity export surge or policy shift away from Keynesianism have an ‘additional degree’ of deindustrialization.