6 | LOCATING MODERN POVERTY WITHIN THE CREATION AND DIVISION OF WEALTH: TOWARDS A STRUCTURALIST AND INSTITUTIONALIST POLITICAL ECONOMY APPROACH IN POVERTY STUDIES
In most approaches within the field of poverty studies, the focus has been overwhelmingly on enumeration and descriptions of poverty rather than serious attempts to explain causality, or what has lately come to be popularly referred to as ‘theories of change’. This avoidance of the causal is not necessarily due to an inductivist impulse of building theory from the detailed observation of facts. Rather, poverty studies scholarship, especially the economics scholarship that dominates the field, is primarily deductive, as epitomised by econometric analyses that, in so many cases, do not even involve any direct observation by the researcher of the people studied. Although glorified, these often boil down to fragile edifices of abstracted axiomatic principles and assumptions, superimposed on data and deploying sophisticated techniques to establish authority. So-called econometric tests for causality are usually so narrow (by necessity, in order to have sufficient control for the test) that the insights they produce are rather banal, especially from a broader perspective of development. As argued by Deaton (2009), they also mostly lack external validity beyond the specific contexts studied. Arguably, they are still measuring associations rather than causality in any case, as is the basic principle of positivist deduction. Broader theorisation does lurk in the background of all of this research as the implicit source of assumptions informing choices that ultimately guide and condition results, but not as the explicit object of scholarship. Broader theorisation is instead left to the fads of grand theories, which of late have come to embody quite simplistic and unsatisfactory approaches to understand the political economy of poverty and development.
The causes of poverty and the ways that poverty transforms and is reproduced over time, particularly within a process of development, can only be fully understood from a broader holistic perspective of all that is involved in the creation and division of wealth within and across societies. Theorisation therefore needs to be placed within broader debates on these questions. The ideal starting points for this are classical political economy as well as so-called heterodox economics perspectives, in particular, Marxian, post-Keynesian, structuralist and institutionalist approaches, which have been the main avenues for perpetuating, adapting and updating these classical perspectives into contemporary scholarship.
These intellectual streams of political economy are emphasised here because they are the main theoretical traditions that seek to integrate an understanding of distribution within an understanding of growth, as opposed to neoclassical economic traditions that mostly separate the two. The neoclassical Solow-Swan growth model, for instance, is based on a one-good model with a representative agent and, hence, by design, is unable to deal with issues of distribution. The Stolper-Samuelson and Arrow-Debreu family of models also do not necessarily deal with distribution per se, except by way of collapsing distribution into a theory of exchange and treating profits and wages as factor prices based on relative endowments of capital and labour. However, they do not theorise growth, except by way of the efficiencies gained through a static reallocation of inputs based on comparative advantage. As a more recent example, the theoretical model of Piketty (2014) does not even include labour or capital-labour relations in its attempt to explain inequality through the concept of capital.
In contrast, political economy in the classic sense refers to an understanding that distribution and its interaction with production and hence economic growth are fundamentally political and institutional. The value of labour in particular (and hence of profits in the inverse) is the outcome of power, social relations, and social conflicts (e.g., class, etc.), rather than of market-mediated returns for the marginal productivity of ‘factor inputs’ (i.e., labour and capital). This leads to explicit questions about who creates value (i.e., labour, the organisation of labour by capitalists, or the owners of assets, etc.) and who controls this creation and subsequent circulation of value. ‘Structuralist’ in the sense used here refers to an understanding of the economy in terms of its articulations and patterns of integration and circulation, of where value is occurring and how is it occurring, and with the understanding that sustained economic growth and poverty reduction generally involve processes of structural change, fundamentally driven by rising labour productivity.
Accordingly, this chapter proposes a schematic framework for building an integrated structuralist and institutionalist political economy approach for what we might call critical poverty studies. ‘Critical’ here does not refer to post-structuralism but rather to the more classical principle of political economy that we must critically introspect theoretical, normative and ideological assumptions from the outset and how these are built into our paradigmatic conceptions of – in this case – poverty and development.
The example of inclusive development, as discussed in the previous chapter, can be used to illustrate how different underlying assumptions feed into the conception of policy. If we focus specifically on the case of marginalised groups, an inclusive growth or development perspective would aim for a better ‘inclusion’ of these marginalised groups into growth or development as a way to lessen inequality and poverty, ideally in a manner that does not undermine the agency, autonomy and dignity of the marginalised, as is often the case in state projects aimed at such groups (e.g., see Fischer 2014a). Indeed, this is roughly the current (at the time of writing) generic definition of social inclusion used by the World Bank, as discussed in the previous chapter.
Accordingly, we can think about inclusion in several ways drawing from the classical divisions from political economy (e.g., production, distribution and redistribution). To start with, from a structural perspective, it is important to think about differences in the basic employment structures that frame marginalisation, such as the proportion of the labour force integrated into formal employment, from less than 10 per cent in many poor countries such as India, to over 80 per cent in most so-called Northern countries. The implications of what constitutes ‘marginalisation’ and ‘inclusion’ in each setting would be very different, particularly with respect to contexts where most people are ‘marginalised’ and ‘excluded’ from formal employment, as discussed in the previous chapter. Hence, policies would have very different implications and consequences in each of these types of contexts, particularly given that much of the distributional struggles waged by organised labour in formal employment might be quite disassociated from conditions among more marginalised and/or informalised workers. Transformations in these contexts over time are also crucial considerations.
Following on from this, we can think of three dominant approaches to address marginalisation. The first is the productionist approach, which aims to help marginalised groups to become more productive by giving them productive assets, education, microfinance, land titles and so forth. The idea is that the marginalised people become richer and/or more empowered – and, hence, less marginalised – by becoming more productive, as if, by operating as microentrepreneurs and small businesses, they gradually grow into the norm of the economy. Such approaches are fairly typical in mainstream development policy (besides land reform, which is generally regarded as more radical). When advocated in isolation, they can be characterized as ‘supply-side’, in the sense that policy focuses on providing the inputs (human capital, financial capital, etc.), whereas it is assumed that demand works itself out in due course (as discussed in Chapter 4 and elaborated below).
A redistributive approach might focus on taxing profits and formal wages (and other incomes and assets via consumption or other taxes) and using part of the revenue to spend on services or cash transfers to the marginalised communities. Much of social policy falls into this category, which encompasses both more mainstream approaches (i.e., targeting) as well as less mainstream approaches (i.e., universalistic social policies). Here, the general principle is that you ‘include’ the marginalised through redistributive transfers and social provisioning – giving them education, health, cash, etc. The question with such demand-side approaches is whether such provisioning actually allows for any substantive form of improved inclusion, rather than simply reproducing marginalisation on new terms. The intervention of the state can also involve aspects of top-down impositions or stimulate changes in the regulation or governance of the poor that can also have both positive and negative effects on marginalisation.
A distributive approach would emphasise the integration of marginalised groups into (decent and ideally formal) employment, or to raise their wages and working conditions if employed, with the understanding that most of them will not become microentrepreneurs. Notably, most of the labour force – at least in most rich countries – is composed of employees, not entrepreneurs or business owners. So, policies like affirmative action reservation policies in public employment might actually be considered as distributive, in the sense that they constitute efforts to effect marginalisation through this channel. Redistributive policies can also have an indirect distributive effect. For instance, public spending can have the effect of raising aggregate demand, which presumably raises demand for labour and hence wages. Employment guarantee schemes are probably best characterised as a combination of distributive (increasing employment) and redistributive (because the employment is subsidised and effectively replaces cash transfers).
As discussed in the previous chapter on social exclusion, sociological approaches cut across all three of these political economy approaches – particularly the second and third – by focusing on the processes and pathways that people actually take according to each approach to inclusion. For instance, discrimination might act as an obstacle to asset and productive accumulation, or to effective inclusion through education systems or to achieving decent employment.
Following on from this illustration and as a means of placing these issues within broader debates about how wealth is created and shared within and across societies, I propose a holistic framework for engaging with such perspectives in this chapter. The framework is based on two conceptual dimensions that are implicit in much of the ideas and policies relating to poverty and social needs. Each of the two dimensions is in turn composed of three elements. This two-by-three framework can then be examined in terms of the more conventional discussion of policy realms, such as the division suggested by Cornea (2006) for examining inclusive growth in terms of social policy, developmental policy and macroeconomic policy. The theme of social policy in particular is taken up in the next, penultimate chapter, given its centrality to understanding processes of social integration and ordering.
The first dimension deals with the creation and division of wealth, which can be conceived in classical political economy terms as production, distribution and redistribution (similar but not identical to the idea of spheres of production and circulation). The second deals with the more secondary, indirect and aggregated factors influencing the first. These can be divided into supply-side factors, demand-side factors and factors dealing with terms of trade or wages (or the relative valuing of labour and its fruits). The latter dimension can also be considered in terms of indirect policy means to influence production and distribution, as opposed to direct policy interventions in production or employment, or else direct use of fiscal policy to redistribute resources.
These interacting dimensions help to clarify how different approaches to conceptualising poverty and its reduction, as well as theoretical perspectives in economics and social sciences more generally, usually place selective emphasis on different combinations of elements across these two dimensions, even though all elements are needed to understand the evolution of social needs and the reproduction of poverty. While each approach might have its own insights, the causes of poverty can only be fully understood from a broader holistic perspective of all that is involved in the creation and division of wealth. More specifically, prevailing approaches to understanding poverty, which implicitly rely on neoclassical theoretical approaches, generally overlook several key dimensions. This latter point will be elaborated throughout and in the conclusion.
Production, distribution and redistribution: The classical triad
The conceptual distinction of production, distribution and redistribution is rooted in the foundational division between production, distribution and exchange in classical political economy, or between the spheres of production and circulation in Marxist political economy. These in turn have influenced the conception of twentieth-century national accounting, i.e., GDP, with a similar division between production, income and expenditure approaches to understanding the creation and circulation of monetary value-added in an economy.
Production
Production focuses on the creation of value-added. It is associated with the production approach of national accounting (gross domestic product, or GDP), represented by the primary, secondary and tertiary sectors. The central concern is labour productivity, understood as output per unit of labour, although because output is mostly valued in monetary terms, the more appropriate term is ‘monetary value-added productivity’, in order to keep this point clear. As discussed below with respect to the terms of trade/wages, this actually has problematic implications for our understanding of productivity and hence poverty, particularly as labour increasingly moves out of physical production and into non-physical services.
The importance of a production or productivity focus in poverty studies is that it frames poverty as deficient productivity, and poverty reduction as fundamentally driven by increasing productivity. It therefore emphasises that the primary effort of poverty reduction policy is to increase the productivity of the poor. Indeed, productivity determines the upper possible threshold for incomes and/or wages, regardless of how output is distributed and prior to any redistributions.
This is most easily conceived in agrarian economies, where physical productivity creates a ceiling for how much farmers can earn (or else ceilings on the value that can be appropriated from their labour, as discussed below under distribution). If a farmer harvests 1000 kilograms of grain in a year, net of the inputs of grain used for planting, and has no other source of income, his or her income can be no greater than the value of this harvest, however this value is calculated (e.g., in terms of its monetary value based on market prices, or in terms of its subsistence value if consumed by the farming household, as discussed in Chapter 3).
This easy conceptualisation invariably becomes complicated, however, as soon as nuance is brought into the example. For instance, in the case of a small-holder farmer, labour productivity is constrained by the amount of land, more so than by access to capital and technology, given that there is only so much that can be produced per unit of land even under the best of circumstances. Moreover, strategies to increase land productivity when land is constrained are often predicated on increasing the labour intensity of production and hence on reducing labour productivity. This must then be evaluated relative to other uses of household labour, particularly when households are engaged in off-farm activities and employment (or, conversely, if they experience substantial levels of unemployment or underemployment).
Indeed, many income generation projects devised for poor people by various governmental and non-governmental organisations often receive an unenthusiastic reception from the intended beneficiaries, but not because of behavioural resistance to change, as is often assumed or argued. Rather, lack of enthusiasm is often related to the fact that the projects are very labour-intensive and the implicit returns for labour time have not been well estimated (if at all), with returns that might be well below the effective minimum wages on offer in local labour markets. While those who devise such projects might assume that any rupee or peso earned by a poor person is better than none, poor people are in fact quite sensitive to their own – perhaps qualitative, but nonetheless roughly accurate – evaluations of the relative returns for their labour time.1 In particular, unlike the unemployed in Northern welfare states, poor people in most developing countries are often very time-constrained. This is especially problematic with many interventions that require substantial voluntary time commitments, such as with microfinance or social protection schemes requiring regular meetings and visits.
Despite the rapid proliferation of complexity as one approaches real-world settings, a production focus is nonetheless essential to understand the structural sources of inequality. An important foundation of inequality rests on differences of productivity across sectors or groups of people. The easiest example is the difference between labour productivity in small-scale peasant agriculture versus labour productivity in modern highly capital-intensive large-scale farming, or in modern manufacturing. Productivity measures in this sense can be disaggregated across different sectors of an economy, and even within sectors if both value-added and employment data are available for such disaggregation. With this information, it is possible to identify low- and high-‘productivity’ sectors or even sub-sectors of an economy, again, measured by monetary value-added produced per unit of labour.
This was the basis, for instance, of the well-known pioneering studies by Simon Kuznets on inequality in the process of structural change. His inverted-U hypothesis – that inequality rises during early phases of development, stabilises in the middle and then falls at more advanced phases – specifically referred to the effects of a structural shift between different sectors of an economy with different levels of productivity and inequality. In his classic article, Kuznets (1955, pp. 7–8) posited that rural areas are less unequal than urban areas because there is greater potential for productivity increases in urban industry than in rural agriculture. In other words, rural inequality is more restricted by the upper bound of productivity that can be achieved in agriculture, whereas urban inequality is much less so. He did not argue, as misrepresented by Milanovic (2011, p. 9), that inequality is low in preindustrial societies because almost everybody is equally poor. Rather, he pointed out that, according to the available evidence at the time, ‘underdeveloped’ countries were actually more unequal than developed countries and, on this basis, he already suggested at that time that his insights would probably have limited applicability for developing countries (e.g., see Kuznets 1955, p. 20).
As noted above, measures of productivity face the dilemma that they are generally based on monetary value-added measures, not actual output. Output-based measures of productivity can only be analysed with reference to single homogeneous physical goods (e.g., kilograms of soya beans). However, value-added measures must necessarily be used as soon as productivity is compared across more than one physical good (e.g., kilograms of soya beans versus kilograms of cherries, versus cars produced in a factory, etc.). The problem is that value-added is a combination of output and price, and already includes within it the value of wages. Hence, measuring productivity with value-added data could reflect as much differences in prices and wages as anything specifically related to actual differences in physical output or efficiencies in the use of labour. Indeed, in service sectors where there is no physical output, value-added measures are completely a reflection of wages and prices.
As discussed further below, it is therefore important to exercise caution and a degree of critical distance from ideas that productivity necessarily sets the value of labour, that poverty is due to ‘low productivity’ or that increasing productivity necessarily needs to be at the forefront of poverty reduction strategies. Monetary value-added measures of low productivity might, in fact, be reflecting something else. I have called this a ‘fallacy of productivity reductionism’ (Fischer 2011c), referring to the assumption that monetary valuation can be used as an accurate approximation of productivity in a complex modern economy. The logic might apply in a farm-based economy, although, even then, simple strategies of income generation are notoriously susceptible to the fallacy of composition. For instance, when everyone starts producing handicrafts, this then creates a glut in the market for handicrafts, lowering their price and then eliminating the potential gains of increasing the physical output of handicrafts at the individual level. Indeed, such an outcome occurred in the coffee sector in the 1990s, when the investments of millions of small-scale producers to increase coffee production were cancelled out by collapsing prices, particularly as several large-scale investments came into production in the mid-1990s. In other words, the logic that monetary valuation can represent actual effort or output breaks down in more complex economic interactions or aggregations, even within agrarian settings and where people are owner-producers. Things become even more complex when dealing with wage earners instead, or as we move away from relatively simple agrarian economies to more complex urban-industrial ones.
Distribution in turn deals with how the surplus – or the value-added of output – is divided into various income streams across different sets of people involved in the production process, or who own and/or control the production and distribution process, including assets used in production such as land or loans. In modern economics parlance, these are referred to as payments to the factors of production. According to the income approach of modern national accounting (gross domestic income or GDI), these income streams are generally categorised as wages, profits, rents, interest payments and taxes on production, with some of the value-added produced also deducted for depreciation and for payments to foreign factors of production.
Recently, it has also become common to refer ‘pre-distribution’, made popular by Ed Miliband in 2012, then the Labour Party leader in the UK, among others (e.g., see O’Neill and Williamson 2012). For earlier academic references, see Plotnick (1982); more recently, the term has also been taken up by Wade (2014) – see Kerr (2016) for a good overview. However, in both classical political economy and modern national accounting terms, this runs the risk of conflating the term ‘distribution’ with ‘redistribution’, as is common in the less-attentive uses of the term. The use of the term by O’Neill and Williamson (2012) does imply more than simply the distribution of the economic product, and is used in a way that implies a political strategy of dealing with distributive struggles in property-owning democracies or, as Wade (2014, p. 1079) suggests, ‘changes in institutions and policies to make pre-tax income distribution less unequal, including in laws of corporate governance, employee protection, consumer protection, environmental protection’. Nonetheless, many people simply juxtapose pre-distribution against redistribution, as if implying distribution.
Distribution in this sense refers to the initial division of the product or income, which can then become the object of redistribution. This is also related to the concepts of market wages versus net wages, or pre- and post-tax and transfers. Market wages reflect the amount initially given to different categories of labour, whereas redistribution is generally treated in terms of taxes and transfer payments, which are reflected in net wages. In some Northern welfare states, the differences can be striking, such as in Sweden or Belgium, where inequality before taxes and transfers is among the highest in OECD countries,2 whereas after taxes and transfers is among the lowest (see Brys et al. 2013, p. 30).
In classical political economy terms, wages, profits and rents are related to the abstracted class notions of labour, capital and rentiers, which formed the basis of most classical economic models, from David Ricardo to Karl Marx. As such, this dimension is closely related to and determined by class relations, which is precisely why it is so central to political economy, especially with regard to the social relations and class conflicts that determine the value of labour and of rents. This is also why classical economists referred to themselves as political economists, in addition to the fact that they also understood the study of an economy to be a study about statecraft. As noted in the Introduction of this book, classical economists up to Marx assumed that the value of working-class labour – then the large majority of the work force – would always be pushed down to a level of subsistence, whether due to the forces of population (as per Thomas Malthus and followed by David Ricardo) or due to the nature of class conflict (as per Karl Marx).
It is also useful to clarify that in this classical political economy conception, which also informs modern national accounting, rents are conceived as income earned from ownership and/or control over assets rather than from investment into production, which generates profits. This conception contrasts from the modern neoclassical conception of rents as excess profits earned through the obstruction of markets (relative to a hypothesised ‘market-clearing’ equilibrium). The neoclassical conception places the onus of rent-seeking, so to speak, on state actors or else others holding relatively monopolistic positions over relatively inelastic goods and services, instead of recognising rent-seeking as primarily a private activity endemic to modern capitalism. Rents in the latter sense are literal, i.e. earning rent from property, or earning interest or profits from financial assets. This distinction is often confused in the adoption of rent-seeking as an analytical frame in much of contemporary political economy scholarship, such as with respect to political settlements, probably because some of the lead authors such as Mustaq Khan usually imply a neoclassical conception of rents despite their general Marxist orientation (e.g., Khan 2002, 2006).
Again, distribution (and its national accounting correlates) can be conceived quite intuitively in terms of a simplified farm economy, with a representative agent and abstracting from class relations (as in the conceptual framework of foundational neoclassical economic models, such as in the one-good world of the Solow growth model). If a farmer plants 10 kilograms of grain and harvests 110 kilograms, the value-added is 100 kilograms. The grain that the farmer eats is wages. The grain taken by the landlord is rent. The grain paid to the moneylender is interest, or to the state is tax. The grain that is eaten by mice or that rots is depreciation, and the grain left over is profit, which in this case accrues to the farmer owner-producer (the profit in this sense is the residual). There is effectively nothing else that the grain can be distributed to. The profit can be saved, given or lent to neighbours, or traded, although these activities fall outside of the realm of initial distribution, given that they are subsequent (secondary) uses of the initial product. Some of these secondary uses would also have their own logic of distribution, such as with the value-added generated by trade, etc. It is also clear through this example how the inclusion of class relations creates a fundamental tension given that profits are inversely related to wages, all else held constant.
The importance of distribution for the study of poverty is that it places the focus on wages, or on the net value of own-production in the case of farmers who farm for themselves (either on their own land or on share-cropped or rented land). In other words, it frames poverty as deficient wages, or as a deficient share or value of output. It also identifies the other major source of inequality next to structural sources, which is the functional income distribution, or the distribution of income across different groups, most fundamentally expressed as the profit/wage share. This differs from the household or personal income distribution, depending on how factoral incomes are distributed across individuals and households. Personal income distribution is effectively the outcome of all sources of inequality.
Last, redistribution (or ‘secondary distribution’ in the most recent national accounting terminology) involves shifting value from its initial distribution. This can either be progressive (taking from the richer and giving to the poorer) or regressive (the opposite), thereby either improving or worsening the initial inequalities created by the structural and distributional dimensions (or by the idiosyncratic characteristics of individuals). Redistribution as such is largely constituted by state or collective action, such as fiscal and social policy, or else by community-organised systems of sharing, charity, and so forth. In this sense, it falls into the Polanyian notions of both redistribution and reciprocity.
Redistribution is central to the study of poverty primarily in terms of public policy to address poverty directly through transfers or subsidies, or indirectly through general fiscal spending. While this is perhaps less fundamental in determining poverty than production and distribution, it nonetheless occupies a large part of the attention in poverty studies and policies. The current social protection agenda, which to a large extent refers to cash transfers to the poor, is one of the most obvious current examples. Of course, these must be measured against the tax burden also faced by the poor. In particular, cash transfers are often funded from value-added taxes (VAT), which the poor pay and are generally seen as regressive towards them (meaning that the poor pay a higher tax rate than the rich because a greater proportion of their income is spent on consumption rather than being saved and invested). Indeed, in this sense it is a misnomer to call these programmes ‘non-contributory’, as has become standard, and the net balance between taxes and transfers might in some cases not even be redistributive. Beyond such programmes, other less-explicit forms of redistribution can include all sorts of public spending, such as education spending prioritised towards disadvantaged groups and areas, or regional redistribution, whereby regions receive more spending than they provide to the government in revenue.
A parallel concept that is often referred to in classical (especially Marxist) political economy is the sphere of circulation or exchange, in contrast to the sphere of production (e.g., see Fine and Saad-Filho 1975; Weeks 1977; Passinetti 2007). The distinction might be seen in terms of whether value is created through the initial production of a good or service, or whether value is created through the subsequent circulation of that good or service already produced. Again, a simplified farm economy serves to illustrate this point: the fully specialised function of a doctor, teacher or priest can only be afforded by a community when all other households produce enough surplus and set aside part of this surplus in order to pay for (i.e., feed) such a specialised service within the community. However, this is not the same as redistribution because activities such as finance, trade and commerce are generally conceived as part of the sphere of circulation in this alternative conception (also expressed, for instance, in the distinction between financial and productive capital). Indeed, the example of other services such as health and education are ambiguous within this conception given that they constitute both the original creation of value (of a service) and yet they would typically be included as part of the sphere of circulation, for the reason that they are afforded through a surplus in production, as noted above. It is perhaps due to these ambiguities that this categorisation has not gained much traction in modern national accounting, in which all services are simply included as part of the tertiary sector of production.
The concept of the sphere of circulation nonetheless raises some interesting further questions. For instance, to what degree should finance be considered a redistributive process, in the sense that there are some who gain more than others in the reallocation of savings and/or money, and some who lose quite substantially? Or is it better conceived as simply part of the sphere of circulation, or as an additional service, with its own generation of value added in the economy and hence its own productive, distributive and redistributive implications? As mentioned above, modern national accounting treats it as the latter. Similarly, monetary policy is never neutral, in the sense that it affects debtors and creditors differently, and it can have an important impact on inequalities. Should this role be conceived as one of redistributing primary values created in the economy, or as something more fundamental, particularly through its effect on the valuation of economic activities, prices and so forth? The latter does not fall well into the categories of distribution and redistribution, but is rather part of the regulative functions of states that can also have important impacts on poor people.
Supply, demand and terms of trade and wages
The second dimension deals with the more secondary, indirect and aggregated factors that influence the primary creation and division of wealth. These can be roughly divided into supply-side and demand-side factors, and the terms of trade and wages. Note that the uses of the first two terms can refer to different positions within the demanding and supplying, although here it refers primarily to aggregate processes driving or influencing growth and/or poverty, rather than with respect to service provisioning.
For instance, demand and supply sides are often referred to in the social sector, such as with respect to the demand for education or health by the users of those services (and hence people demanding these services). Similarly, supply-side in the education or health sector refers to how the actual services are provided, in what quantity and quality, by which providers (public or private), with what infrastructure and human resources, etc.
This provisioning usage is common, for instance, in the social protection literature. Cash transfers are seen as ‘demand-side’ interventions given that they augment the ability of those receiving the transfers to pay for education or health services, and to thereby exert ‘demand’ on these services. One of the principle critiques of these strategies (and of the human capital presumptions that come with them as the rationale for conditionalities) is that augmenting demand in such a manner is unlikely to have any significant effect on educational performance or outcomes if and when the supply of education services remains poor (e.g., see this critique in Fischer 2012). Indeed, most of the focus in this cash transfer agenda, of just giving money to the poor (e.g., Hanlon et al. 2010), has been on such demand enhancement rather than on investments in the supply of services and better quality services, or on a better integration and unification of the provisioning systems that service the poor with those that service the middle classes. From a social policy perspective, the latter strategies are arguably the strongest ways to guarantee quality services for the poor, and hence for genuine and broad rather than minimalist universalistic provisioning for the poor. In the absence of attention to such supply-side issues in provisioning or to broader aspects of institutional integration, critics (such as myself) have warned that cash transfers, or similar policy agendas such as basic income schemes, could actually serve as Trojan Horses to advance privatisation agendas in the social sector, whereby demand-side interventions are used to support poor people in their purchase of privatised services (see the next chapter for further discussion of this and related points).
Here, however, the principal reference to demand- and supply-side is in terms of how they have come to be used in macroeconomics, unless otherwise specified. ‘Demand’ refers to effective aggregate demand, as the crucial macroeconomic variable governing demand for labour through its effect on output and economic cycles. ‘Supply’ in turn refers to an economy’s output of goods and services to meet aggregate demand, as well as the supply of labour.
The idea of supply-side economics became famous when it was endorsed and promoted by Ronald Reagan in the 1980s, basically as a free-market ideological attack on Keynesian demand-management, which had been the establishment view up to that time. As discussed in Chapters 2 and 4, it essentially refers to the idea, known as ‘Say’s Law’, that supply creates its own demand. In other words, increasing supplies of goods and services will find buyers, and increasing supplies of labour will find gainful employment, as long as the government does not interfere with the market system that allows the economy to manifest demand for its output or factor inputs. As such, it is usually associated with neoliberalism.
An alternative perspective is that employment transitions and urbanization are primarily driven by social need rather than economic incentives, and that optimal quantities (and qualities) of employment are not guaranteed by the free and flexible operation of labour markets. This requires that governments must actively engage with development and related public policies to mediate the changing supply of labour generated by rapid social structural transformations.3 If not, governments risk exacerbating employment disjunctures in both rural and urban areas, which labour market flexibility might in fact offer little potential to redress.
Continuing with the example of cash transfers, this policy instrument is conceived as demand-side from a service user perspective (the user uses the cash to ‘demand’ services) and also from a macroeconomic perspective (augmenting the aggregate demand of the poor). However, the human capital justification for the policy – which finds its zenith in the promotion of conditionalities or ‘co-responsibilities’ – is better described as deriving from a supply-side logic, with respect to the employment and poverty reduction dynamics beyond the one-off transfer of cash. The cash provided is often seen as just the carrot to induce people to make ‘good’ choices about ‘investing’ in such human capital (that is, school attendance and health checks), rather than as income support. As with human capital theory more generally, the argument is that long-term poverty reduction, beyond the immediate and negligible impact of the cash transferred, is driven by such increases in human capital. The assumption is that the increased supply of human capital, or increased number of people with augmented human capital, creates its own demand, presumably in the form of more employment that corresponds to the increased embodied capital, with higher wages and non-wage conditions, etc.
The problem, as has been recognised since some of the earliest studies of cash transfers in the noughties, even by their proponents, is that these outcomes rarely occur in any significant or systemic manner in the absence of other interventions (this will be discussed further in the next chapter). Few long-term employment impacts are observed among cash transfer recipients that can be attributable to the conditionalities and their effects on enrolments. Arguably, this has been due in large part to the fact that the demand-side factors of employment creation are generally neglected, including the role of government interventions in the labour market, and also because social provisioning to the poor is generally of poor quality, hence unable to produce much transformative punch despite improvements in utilisation indicators. If anything, quality of education or health care might actually deteriorate in the face of increasing usage if insufficiently supported on the infrastructural supply side.
However, if we strip away the human capital supply-side jargon, cash transfers themselves are essentially (albeit fairly marginal) demand-side interventions, arguably best conceived as income support or as a ‘non-contributory’ substitute for social security for those sections of the population who do not have access to social security. As noted above with respect to redistribution, ‘non-contributory’ is of course a misnomer because the poor or informal do contribute significantly through indirect taxes such as VAT, even though they usually do not pay direct taxes or contributions to social security. In other words, they do contribute, just not through the direct institutional channels generally conceived as being appropriate forms of contribution by more privileged parts of the labour force.
Policy obviously needs to consider both supply and demand holistically, as there are good reasons for both supply- and demand-side approaches, from various angles. Indeed, whereas the notion of supply-side accrued a negative association with supply-side economics, the idea of public supply-side interventions is especially relevant for development, in terms of supplying human resources and physical infrastructure, and supporting productive capacity. Without such supply-side interventions, demand-side interventions such as cash transfers can risk running up against supply-side constraints and bottlenecks when scaled up. These constraints could be in food supply (a classical concern), in terms of foreign exchange (depending on the import intensity of the increased consumption) or in terms of an ability to provide quality schooling, in the case of scaling up enrolments. These were some of the classic criticisms of Keynesian demand-management posed by the early structuralists of development economics, i.e. that poor countries are generally supply-constrained, rather than demand-deficient as in the case of rich economies. Conversely, an excessive focus on supply-side accumulation might result in lots of infrastructure or employment but with low rates of utilisation or value, as was a common criticism of the Soviet Union in terms of its neglect of consumer demand. Ironically, however, both the supply-side emphasis of structuralists and socialists as well as the demand-side emphasis of Keynesians came under attack with the rise of neoliberalism from the 1970s onwards, which was adamantly against any form of public intervention from either side. ‘Supply-side economics’ is, in this sense, a misnomer.
Terms of trade and wages
Underlying both supply and demand are the questions of what fundamentally determines the relative value of output and labour in an economy, or what is referred to here as the terms of trade or wages. In other words, if a farmer produces 100 kilograms of surplus grain and decides to trade it, the value of this surplus needs to be measured relative to what it can buy, thereby setting relative prices, such as 1 kilogram of grain for 100 grams of butter. Alternatively, if the farmer decides to work for a wage, then, as classically postulated by Arthur Lewis (1954), the wage must be set slightly above the amount of food the farmer can produce, because otherwise there would be little incentive for him or her to seek wage employment as an alternative to farming (assuming, of course, that labour is free to choose and farmers have land).
The introduction of new products into the mix similarly shakes up these relative price and wage structures, particularly as new products come to be seen as necessities, as has been common with the industrialisation of consumption across the world over the last century. This is why industrialisation is so important even for countries that are effectively non-industrial, because their consumption has, over the course of the century, become increasingly industrialised and hence dependent on the terms of trade of their local non-industrial production in comparison with industrial imports, or else by the prices set by industrial processes in the broader global economy even for their own non-industrial output (such as with grains). Under such circumstances, the value of labour can be easily eroded even while workers and entrepreneurs are making efforts to increase productivity.
This predicament was eloquently elaborated by Arthur Lewis in his classic (and often poorly regurgitated and understood) pioneering contribution to development economics. One of the two key puzzles that he was trying to solve in his 1954 article was the question of ‘why tropical produce is so cheap’.
Take for example the case of sugar. This is an industry in which productivity is extremely high ... [with] a rate of growth of productivity which is unparalleled by any other major industry in the world – certainly not by the wheat industry. Nevertheless workers in the sugar industry continue to walk barefooted and to live in shacks, while workers in wheat enjoy among the highest living standards in the world. (Lewis 1954, p. 442)
To answer this question, he focused on factoral terms of trade in the open economy version of his model of economic growth with unlimited supplies of labour. ‘Factoral’ in this sense means that he was not focused on the terms of trade between the prices of commodities, which is how it is generally treated (such as the price of bananas or sugar compared to the price of wheat or cars). Instead, he was looking at the terms of the price of labour in the respective sectors of these export commodities, or wages in (Southern) sugar production versus wages in (Northern) wheat or car production.
He argued that because wages are set in what he called ‘subsistence sectors’ rather than in capitalist export sectors, and, more importantly, that the supply of labour from these subsistence sectors is unlimited at a wage that is set sufficiently higher than what can be earned in these sectors, the benefits of increasing productivity in the capitalist sector accrue chiefly to the (Northern) importers of these exports by way of lower prices, rather than rising wages for the (Southern) producers. Hence, he contended that ‘the prices of tropical commercial crops will always permit only subsistence wages until, for a change, capital and knowledge are put at the disposal of the subsistence producers to increase the productivity of tropical food production for home consumption.’ The implication of his argument, which is now usually overlooked in deference to a caricature of his theory of growth with unlimited supplies of labour, is that wages are not reflected by productivity, even in physical production, and that increasing productivity in physical production will not necessarily result in rising wages, at least not under the conditions of an open and poor economy with substantial supplies of labour available to work in the so-called ‘capitalist’ sectors.
Later in his life, Lewis (1978, p. 36) similarly predicted that even as developing countries would move into manufacturing exports, these new exports would function in a manner similar to the previous agricultural export commodities, in the sense that increasing productivity would simply reduce the prices of such manufacturing exports, thereby continuing to exhibit declining terms of trade. This prediction appears to have borne true for Southern countries that have moved massively into manufacturing through their integration into international production networks dominated by transnational corporations, including China (for a detailed discussion of this point, see Erten 2011; Fischer 2015, pp. 717–718).
Lewis’ theoretical provocations were pointing towards fundamental debates about what, precisely, determines prices and wages. As discussed above in the section on production, attributes of physical productivity to a certain degree create floors and ceilings for the valuation of labour (or else ceilings on the value that can be appropriated from such labour), especially in relation to rural economies. Economic considerations of supply and demand also operate at the margin, that is, they affect how prices or wages might move higher or lower from a certain position. Indeed, it is in this sense that early neoclassical economics in the late nineteenth century was referred to as the ‘marginal revolution’. However, these marginal effects do not fundamentally determine the price, except perhaps with pure commodities, but labour is definitely not one of these. Change at the margin is arguably not a primary determinant of why wages are set at certain levels and how they change over time. Of course, neoclassical theory did provide an answer (marginal productivity theory, that wages are set by the level at which marginal costs equal marginal revenues), but this does not solve the problem of what the actual wage would then be. It is presumably set by the average productivity of all workers once that equilibrium point has been reached, although because this is also conceived in terms of a system of relative prices, it is again difficult to say, exactly, what that wage might actually be. These questions have been at the core of contestation between different major modern schools of economic thought up to the present (for instance, with reference to theories of price, see Nicholas 2011). Economic science, so to speak, has not solved the question of how labour is fundamentally valued, even though this is fundamental to understand the causes and influences of poverty.
Lewis’ return to the classics was based on his assumption that wages were set by the levels of productivity in the ‘subsistence sector’ rather than by marginal productivities in the capitalist sector. Subsistence productivity set the terms that capitalists had to offer in order to draw people out of this sector in order to work in capitalist production (unless, of course, we assume that these workers were drawn from among the destitute and dispossessed, but he does not seem to accept this idea based on his own scholarship in the economic history of the UK and the Caribbean). He similarly argues that the value of goods produced by such labour must reflect such valuation of labour. Hence, the groundwork of his theorisation was established by the assumption that factoral terms of trade between temperate and tropical commodities are set by the difference in agricultural productivity between Europe and the ‘tropics’. In particular, the great waves of migration of both Europeans and Asians in the late nineteenth century, during a time of unprecedented trade expansion and integration, set the terms of trade for tropical and temperate agricultural commodities respectively. This was because ‘for temperate commodities the market forces set prices that could attract European migrants, while for tropical commodities they set prices that would sustain indentured Indians’ (Lewis 1978, p. 14).
Indeed, the only sector where very clear labour productivity differences can be observed between North and South today – enough to explain the wealth differences between North and South today – is in agriculture (e.g., Canadian farming versus Indian farming). Such labour productivity differences cannot even be seen in manufacturing anymore (e.g., ‘first-world productivity at third world-wages’, whereby Chinese factory workers are as productive in producing the same manufactured good as workers in the US or Europe, despite much lower wages). Rather, value in manufacturing today is accrued more in terms of product differentiation, or component differentiation along value chains, and control over production and distribution networks, including control over the more technological and knowledge intensive parts of the value chain. The point here is that wage differences cannot be discerned through simple comparisons of labour productivity in the physical output of the same or comparable goods, besides perhaps in agriculture.
Part of the controversy in these debates derives from different ways of understanding the nuanced and complex consequences of structural integration, while part derives from the contention that the valuation of labour is fundamentally determined by social and power relations (within structural constraints, as noted above), and not by market intermediation. With respect to the structural, the tendency in most of the contemporary literature on structural change is to treat issues such as productivity and valuation in a rather simplistic manner, more or less assuming or accepting the proposition that the monetary valuation of labour is a roughly accurate measure of labour productivity and that labour productivity is a relatively straightforward function of efficiency and capital accumulation. However, within complex, increasingly integrated economies, the relevant question is about the relative valuation of labour, in which many of the simple linear assumptions about productivity break down, as highlighted by Lewis.
The fallacy of productivity reductionism and development
The idea of what I have called a ‘fallacy of productivity reductionism’ (Fischer 2011c) was already briefly introduced in the section on production. This refers to the assumption that monetary valuation can be used as an accurate approximation of productivity in a complex modern economy. This modern neoclassical assumption contributes to the myth that the rich are rich due to their greater productivity than the less-rich and thus, by implication, that their wealth is a fair and just reward for effort. It is similar to early neoclassical theory in the late nineteenth century, which collapsed the distributional concerns of classical economics into a calculus of market allocation. The assumption arguably lies at the heart of ideological efforts to legitimate the inequalities of the current world economic order, particularly with respect to the increasingly transnationalised networks that regulate the valuation of output produced by poor people working within these networks.
This insight draws in part from the post-Keynesian school of economics. Perhaps the most practical insight that can be distilled from the otherwise-highly abstract ‘Cambridge capital controversies’ of the 1960s is the point that, in a monetary world with heterogeneous goods and services, ‘capital’ is effectively almost impossible to measure in any coherent aggregate sense. In contrast neoclassical assumptions treat capital as substitutable and, hence, more or less identical with consumption goods (such as corn for planting and corn for consumption, if we return again to the highly simplified neoclassical farm economy with only one good, as referred to earlier with respect to the Solow growth model). Accordingly, aggregate ‘productivity’ can only be understood in terms of monetary valuation (i.e., through prices), not in physical terms – this point should be obvious with respect to most services. It should be recalled that Paul Samuelson (1966) implicitly conceded defeat in these debates to Joan Robinson and Piero Sraffa, although this concession was subsequently ignored by most of the profession, particularly from the 1980s onwards.4 Instead, the profession has continued to assume that complex modern industrial economies essentially function in the manner suggested by the one-good (or identical two-good) world imagined in the production functions of standard modern neoclassical theoretical models (and the so-called New Keynesian variants). In these models, capital and productivity are treated as if they can be measured in physical terms, such as in the total factor productivity measures derived from one-good moneyless growth models of the Solow variant. It is in this sense that the bulk of contemporary economics can be described as a faith-based discipline, in the sense that a leap of faith is required to enter its temple of logic.
In consequence, most mainstream approaches to measuring productivity almost always rely on value-added accounting data in various ways as a means to approximate aggregate productivity, whether at industry, sectoral or economy levels. The justification for this was partly based on early econometric work by Kenneth Arrow, Hollis Chenery, Singh Minhas and Robert Solow. Starting with the ‘empirical observation that the value added per unit of labour used within a given industry varies across countries with the wage rate’, these authors concluded that most of ‘the variation in labour productivity is explained by variation in wage rates alone’ (Arrow et al. 1961, pp. 225 and 228). In other words, the implication is that variations in wage rates are mostly due to labour productivity. However, as pointed out by Felipe and McCombie (2001, p. 1222), these results were essentially the product of tautology in that ‘the estimations of production functions [are simply] capturing an underlying accounting identity [that value added equals the wage bill plus profits] … Hence, regressing output on the inputs is bound, almost by definition, to give a very good statistical fit.’ In other words, it is basically saying that there is a very strong correlation between wages and the sum of wages and profits, particularly given that profits are generally a fairly stable and consistent share of the total. To the extent that their contention is true, it undermines the whole edifice of aggregate productivity measures that have since been elaborated on the basis of the results of Arrow et al. (1961), as noted in subsequent work by Felipe and McCombie (e.g., Felipe and McCombie 2003, 2006, 2012, 2013).
The use of value-added as a shorthand for productivity leads to absurd logical implications. For instance, it suggests that a barber in the US is 30 or more times more productive than a barber in India, even though they both ‘produce’ the same number of haircuts per hour (according to the tastes and expectations of their clients), simply because the wage of the barber in the US is 30 or more times higher. Or else, within the US, that the ‘productivity’ of a lawyer is ten times higher than that of a barber. The lawyer’s labour is certainly more valued than the barber’s, whether or not for good reason, but this has little to do with productivity (e.g., what is the output of a lawyer or a bureaucrat?). We might consider that the power to leverage higher value for one’s labour is, in itself, a form of productivity, but this is more accurately represented as power, as discussed further below. Indeed, the rising value of labour in financial services, and hence of ‘productivity’ in this sector, is related to processes of financialization, whereby levels of remuneration in the financial sector are allowed to rise to unheard-of heights (and then might be interpreted as improving ‘productivity’ within that sector as a result). The differences across different types of labour are usually explained in terms of ‘human capital’, referring mostly to years of schooling, and to skills premiums, although similar wage differentials can also be observed between employees with more or less equivalent levels of education, such as teachers versus lawyers, or a computer engineer in California versus one in Bangalore. In other words, much of what we are picking up in most conventional measures of productivity actually amounts to price or wage differences, not actual effort, output or even skill, especially in economies that are increasingly based on services.
The idea that monetary valuation can represent productivity is especially problematic with respect to the tertiary sector, given the notorious difficulties of measuring productivity in services. For instance, while some dimensions of productivity can be measured in some aspects of tertiary work, such as patients treated per doctor, students taught per teacher, numbers of sheets cleaned per domestic worker or merchandise sold in a supermarket per cashier, such productivity measures are often deceptive in that the quality of the good regularly changes through the process of increasing outputs (such as more children per classroom leading to worse learning outcomes, or less care in serving hospital patients). Hence, such measures are not necessarily measuring a consistently comparable good (as would be the case with physical measures of identical goods, such as tonnes of wheat). Even in the case of wholesale and retail trade where productivity measures bear some relevance (i.e., merchandise sold per employee per unit of time) and where the application of technology is also relevant, increases in productivity – especially in retail trade – often imply shifting labour and time costs onto customers, which makes them profitable but not necessarily more productive from a broader holistic perspective that includes all of the labour involved in the process.
It is also somewhat nonsensical and arbitrary to use the standard lexicon of ‘labour intensity’ (versus capital intensity) to explain wage differentials within the tertiary sector or in comparison to other sectors given that service-sector activities are largely – or in many cases entirely – based on labour inputs rather than capital inputs, except again in the activities of trade or transport, where we can conceive of capital inputs, including automation and robotisation. Tertiary-sector professionals such as lawyers are not usually referred to as labour-intensive even though, in capital-labour terms, most if not all of the input of legal services is labour.5 Presumably, they are not referred to as labour-intensive because they generate high value per unit of labour time, or else because their education is taken to be a proxy for capital (unlike nurses or school teachers, who have similar levels of education in terms of degree types but at much lower levels of value-added and, hence, are often referred to as labour-intensive).
The logic of differentiating capital intensity in service activities through the designation of ‘human capital’ is problematic for these and similar reasons. Human capital is usually proxied by years of schooling and hence is basically referring to formal education. However, the evidence that differences in education can predict wage differences – particularly across countries but also across service sectors within a country – is increasingly weak, especially as populations attain higher and broader schooling levels. Much of the earlier assumptions on this were based on cross-country regressions whereby richer countries generally had much higher levels of formal schooling. However, this association becomes progressively weak the more poorer countries (or lower social classes) progress in levels of formal schooling without necessarily experiencing the economic benefits that were associated with equivalent levels in the earlier cases. Hence, like with fertility in demography as discussed in Chapter 4, human development has become increasingly disassociated from economic development. It is a classic case of misattributing causality with association.6
In other words, there is a degree of tautology at work. Labour-intensity is associated with relatively lower wages, and hence when relatively lower wages are observed, the labour is assumed to be more labour-intensive, even though both labour intensity and labour productivity are nonsensical reference points in the measurement of the outputs of intangible services, the value of which is mostly based on the value of the labour inputs.
The mainstream response to the quandary of productivity in services has usually been with reference Baumol’s ‘cost disease’. This refers to the work of Baumol and Bowen (1966), who attempted to explain the apparent paradox of rising salaries in jobs that experience no increase in labour productivity (such as in the performing arts, where the same number of people are required to play a symphony as a century earlier), in comparison to rising salaries in jobs that do experience labour productivity increases (such as in manufacturing). They deduced that this seemingly contradicts the standard assumption that wages are closely associated with changes in labour productivity (as discussed above). Their answer was essentially that wages in sectors without productivity gains increase because they must compete for employees with the sectors that do experience productivity gains. The analysis is notably built on differentiating physical production from services and presumes that productivity in services can be measured using standard total-factor productivity techniques. The very fact that wage equalisation across the economy might be seen as a ‘disease’ is itself indicative of the particular perspective informing this approach.
Triplett and Bosworth (2003) claim that the disease had been cured on the basis of evidence of rising productivity growth in service sectors in the US economy after 1995, although their measurement of such productivity continues to rely on the standard productivity-growth accounting framework originating from the work of Solow (1957), whereby labour productivity is measured as the value-added output per person engaged in production. It also remains bogged down in a variety of theoretical and methodological issues, many of which they admittedly acknowledge. For instance, they exclude social services from their calculations because of difficulties surrounding the treatment of capital in such industries. On the treatment of capital in other service industries, they acknowledge the problems of relying on capital stock data at the industry level as a measure of capital input given that it is dominated by financial wealth, not simply ‘productive’ capital stock that is used as an input to production. They propose a resolution of this problem by combining such data with flow data on capital services by industry, although, as pointed out by Andersen et al. (2009) in their review of methods commonly used to measure capital service flows, a variety of important assumptions are used in constructing such measures and, even in the case of the relatively easy-to-measure capital inputs of agriculture in the USA, they demonstrate that substantial differences in these measures can be computed depending on which assumptions are chosen. In other words, the more one tries to narrow the problem to make it tractable, the more it becomes problematic.
The fallacy of productivity reductionism becomes especially pertinent when considered in light of the increasing transnationalisation of production and distribution in the postwar era (although it was also very pertinent during the colonial era, for similar reasons). For instance, when an increase in the value-added of the tertiary sector is recorded in the US or Europe – such as in the categories of management, finance, marketing, trade or research and development – this is usually interpreted as increasing productivity in those sectors (and increasing economy-wide productivity if this results in overall economic growth). Yet at least part of such an increase represents the inflows of profits and concentration of value in the headquarters of lead corporations (such as Apple, Google or Amazon) from across diverse networks of subsidiaries, affiliates and sub-contractors, many of which are based in the Global South, especially those based in actual physical production. When a transnational corporation practises transfer pricing or equivalent practices that have become standard in international accountancy,7 as a means of transferring profits from a Southern subsidiary or affiliate to a head office in New York or London (usually via various offshore financial centres), the subsidiary appears less productive as a result, whereas the head office appears more productive. Moreover, most of the activities of the head office would be in services, not physical production, which doesn’t seem to factor into the Baumolian concerns of service-sector productivity, presumably because these are seen as ‘business services’ rather than, say, orchestra performers (even though orchestra performers do tend to perform for the beneficiaries of high wages in such business services and might even derive from a similar social class).
Ironically, the actual producers of goods – who are increasingly located in the Global South – might well be accused of being inefficient (in value terms) and in need of extra structural adjusting, even though the appearance of such inefficiency is in many cases the product of accountancy practices that reduce reported measures of value-added to a functional minimum and thereby maintain subsidiaries in a perpetual profitless existence. Such producers might be otherwise working very diligently and investing in a whole myriad of ways in order to keep up with the competition. In such circumstances, the precise meaning of ‘being productive’ is difficult to pin down, except in cases of the physical output of comparable goods. The idea that inter-country differences in wage rates can be explained by obstructions to the free flow of goods, services and capital across borders also becomes increasingly absurd the more the global economy liberalises and becomes increasingly unequal.
These issues also highlight the importance of ownership and/or control over processes of value circulation in monetary economies, and the insidious siphoning of wealth that usually results from a dominance of foreign ownership in the peripheral economies of the Global South.8 Effective outflows of wealth, whether through licit or elicit capital flows or else through subtle processes such as transfer pricing, undermine the monetary aggregate demand in these economies that would otherwise contribute towards enhancing employment generation as well as the fiscal revenues required to finance comprehensive social policies. It is also in this sense that national ownership plays a hugely important role in efforts to retain wealth within national economies, thereby capturing the benefits of productivity increases when and where these take place. For similar reasons, national ownership has been one of the principle targets of ideological attack under the last 30 years of neoliberalism, with the frontline currently targeting China under the guise of accusations of currency manipulation or other unfair trade practices.9
Raising productivity in the Global South through efforts to support farmers or through industrial policies to support fledgling industrialists is obviously an important component of development strategy and of efforts to reduce poverty and inequality, particularly if the resulting benefits are used in ways that genuinely improve wellbeing among the poor. For instance, small-scale farmers would obviously benefit from raising their output on both their existing plots of land and, ideally, on enlarged land holdings (which, by implication, would require land reform), at least so long as the costs of increasing their output would not exceed the benefits. However, such self-evident examples are often used to simplify and legitimise the more generalised patterns of inequality in our world today, which are much more obscure in terms of a direct connection between effort or output and poverty. The underlying fallacy of productivity reductionism is imperative to recognise because, otherwise, an obsession with raising productivity risks being turned into a powerful ethos for disciplining an increasingly Southern global workforce together with nationally based productive capitalists, both subordinated into global networks of production and distribution that increasingly control the most lucrative flows of value in our world economy.
Social relations and structure in the determination of value within development
These issues surrounding the valuing of labour are generally ignored by prevailing mainstream approaches to poverty studies, except when viewed through a particular neoclassical lens drawn from the marginal productivity theory of wages. However, as the fallacy of productivity reductionism reveals, the processes involved in valuing the labour of a poor person are arguably influenced by a variety of broader political economy and sociological factors, both structural and institutional. Correspondingly, the question of modern poverty is not necessarily about how to make poor people work harder and better, but about how their work is valued within a context of development.
In this regard, the other part of the controversy about how labour is relatively valued is the contention that the valuation of labour is fundamentally determined by social and power relations (within structural limits). In other words, especially but not only in the service sector, wage differences are arguably related to social and institutional processes of valuing various categories of labour, which are driven by power relations, ideological factors, socialisation and related processes of social stratification and differentiation, instead of considerations ostensibly related to labour productivities within the production of comparable goods or services.
Real markets and other economic processes, such as exchange operating through administered state channels or reciprocally through family and community channels, are embedded within these social and power relations. Or, as asserted by Furtado (1983), market operations ‘are, as a rule, transactions between agents exercising unequal power’ (p. 15). Competitive markets are always embedded in this sense, even though, as argued by Polanyi (2001 [1944]), non-market forms of social organisation come to be subordinated to the demands of ‘market society’ under capitalism – a process he calls ‘disembedding’, although there has been much debate over the precise meaning of this term (e.g., see Goodwin 2018, Levien 2018). Accordingly, the valuing of labour itself needs to be understood as an intrinsically social process.
This point was poignantly made in the seminal work by Celso Furtado (1983; translated from the Portuguese edition of 1978), one of the founders of the Latin American school of structuralist economics and a colleague of Raul Prebisch. Furtado identified the rise of industrial capitalism with the rise of unified planning – referred to as economic policy – which was achieved through the emergence of modern nation-states from the mid-eighteenth century onwards. This showed two relevant countervailing features: a progressive concentration of economic power and the highly effective organisation of the wage-earning masses (ibid., p. 17). Within such ‘organised capitalism’, which is best conceived as a system of social organisation rather than simply a form of organising production, ‘political power – conceived as capable of changing the behaviour of broad social groups – emerges as a complex structure in which the institutions comprising the State interact with the groups that dominate the accumulation process and with the social organizations capable of intervening in a significant way in the distribution of income’ (ibid., p. 17).
The challenge for contemporary developing countries, he argued, is that the increasing transnationalisation of production weakens the ability of increasingly effective forms of social organisation to exert pressure on the growing power of enterprises for a more immediate distribution of income (ibid., p. 21). While the power of ‘social organizations engaged in the struggle to raise the value of labour, that is, to bring about a more widespread social distribution of the benefits of the growing productivity of labour generated by accumulation’ (ibid., p. 27) was historically enhanced by the tendency of the state to broaden its social bases of support, the same tendency today is undercut by the fact that the power responsible for coordinating increasingly complex international economic relations has shifted, to a large extent, from the national state to the large enterprise (ibid., p. 23). From this somewhat-prophetic perspective, written well in advance of the globalisation literature of the 1990s, the subsequent emphasis on deregulation during the neoliberal years from the 1980s onwards can be understood as a further privatisation of this macro-regulatory capacity rather than deregulation per se.
However, as economies become transnationalised, many of these relations become increasingly opaque and hard to decipher. To a certain degree, this has led to a sort of neo-naïveté in poverty studies, treating poverty reduction as if it is still primarily concerned with poverty traps in traditional subsistence despite decades worth of criticism of such conceptions and even while the conditions governing the poor have long moved beyond such traditional states, if ever they were.
It is important to move beyond such naïveté and towards more systemic understandings of how poor people and regions are integrated into larger economic and social systems. Whether a process whereby people remain poor and do not experience any apparent rising productivity should be called a ‘structural transformation’ or ‘development’ is of course open to question. What is important, however, is that the experiences of poor people who are integrated into broader systems must be considered as integral parts of the transformations of those systems, even if there appears to be nothing particularly transformational in their own particular part (although there usually is, insofar as the reproductions of poverty are driven by the broader transformations).
This preference for a macro-systemic view versus a micro-view is an important antidote for micro-fundamentalism or for methodological nationalism and localism. At sub-national or local levels, the precise combinations of the factors that lead economic development – such as increasing labour productivity and capital accumulation, based on increasing scale and degrees of specialisation – will vary depending on regional patterns of integration into national and global economies. Indeed, some regional economies might possess few or none of the attributes of increasing productivity in agriculture or manufacturing, but benefit nonetheless from a rise in national labour productivity, in particular through the supply of cheap mass-produced goods, the circulation of monetary aggregate demand and the dissemination of wage norms throughout the national economy supported by government fiscal policy. Furtado (1973) referred to this as modernisation without economic development, in the sense that modern goods are inserted into the consumption basket of an economy without the capital accumulation or the adoption of more effective productive processes that are associated with the production of these modern goods. He posited that such new forms of consumption could be afforded by increases of income earned through either the depletion of natural resources or else through static reallocations and specialisations of production according to comparative advantage, both of which do not necessarily entail economic development. We might include aid transfers as another way to raise incomes and hence modern consumption without economic development. From this more systemic perspective, it is difficult to know where to draw the regional boundaries of economic development, that is, the boundaries within which a region can be considered to be developing economically. Notably, there are usually wide swaths of territory and large proportions of population even in advanced industrial countries that are based mostly on the consumption of modern goods rather than their production (increasingly so as employment continues to shift into tertiary activities).
In this respect, remote peripheral rural economies generally transform through consumption-driven integrations into ‘industrial civilisation’ without necessarily being involved in industrial modes of production. Such situations would describe well much of sub-Saharan Africa, outside specific enclave sectors such as mining where industrial processes are definitely in use, although usually under the control and/or ownership of foreign (or non-local) corporations, which subsequently divert much of the value-added out of the national economy. Ignoring the role of fiscal and other transfers, the relative purchasing power of such a peripheral economy would be determined by the terms of trade between its non-industrial, mostly rural or artisanal, labour-intensive output, versus the industrial goods (and related services) that are consumed in the economy, mostly through import. This relative purchasing power then moderates differences in labour productivity, or what Arthur Lewis (1954) called the ‘factoral terms of trade’ (labour being a factor of production), as well as the distributive outcomes of production and circulation.
For instance, even if prices for meat increase relative to mass-consumption goods such as instant noodles or mobile phones, the wealth of a herder is nonetheless limited by the size of the herd he or she is able to sustain, and the disposable income by the number of cattle he or she is able to sell in a year. The number remains limited so long as herding is based on household and non-industrial forms of organisation. Indeed, this is the reason why many governments and development agencies have often been intent on making herders become ranchers as a means of poverty alleviation, although less so after environmental concerns have taken precedence since the late 1990s and small has again become beautiful, at least for the poor. Conversely, a factory worker involved in the production of instant noodles or mobile phones most likely contributes to the production of much more output and value-added in a year than the herder, particularly in a highly mechanised and automated factory. The worker’s income might not be better than the herder’s, not because of productivity considerations but because his or her wages are squeezed through the process of distributing the surplus (i.e., value-added) between profits, taxes, wages and other income payments such as interest payments on the debt used to purchase factory machines or rent. Hence, it is not obvious that the herder would become better-off by becoming a factory worker (or, along the same logic, a rancher), as is often blithely assumed by scholars and policy-makers. However, it is quite likely that, in economic terms, the worker produces more value-added than the herder (i.e., the worker has a higher level of labour productivity), which is then circulated throughout the economy through various processes. (Or else, as noted above, the value-added is circulated out of the economy altogether, thereby not contributing to local process of accumulation, as is the perennial problem of peripheral, externally oriented and externally dependent poorer countries.) Industrialisation is thereby usually seen as a significant driver of economic growth, whereas ‘peasant’ production is not, even though the ‘peasant’ might not necessarily become better-off by becoming a worker within an industrial system of production, at least not until that system starts to produce rising wages. The delicate balance, particularly in peripheral economies, can lead to many paradoxical dynamics within the labour transitions of such peripheral economies, that is, paradoxical from the perspective of our stylised understanding of classical labour transitions in Europe.
In this sense, it is important to break out of simplistic normative binaries regarding development. It is not as if negative attributes (such as persistent poverty) or even a lack of structural transformation in some sub-sections or sub-regions of an economic system (such as the persistence of low-productivity economic activities in rural areas or entire economies) somehow invalidate the existence or possibility of development or structural transformation. Rather, from a broader perspective, it is important to consider how such negative attributes are the dialectical parts of a developmental whole, especially with regard to the integration of peripheral regions within wider economic systems. In particular, the ways that development manifests in peripheral settings can be quite different from the ways it manifests in more central locations. Nonetheless, insofar as they are integrated parts, they need to be considered as part of a holistic conception of structural transformation.
This chapter presented a schematic framework for building an integrated structuralist and institutionalist political economy approach for what we might call critical poverty studies, followed by an elaboration on how this relates to development. The proposed framework is based on two conceptual dimensions that are implicit in much of the ideas and policies relating to poverty and social needs. The first dimension deals the creation and division of wealth, which can be conceived in classical political economy terms as production, distribution and redistribution. The second deals with the more secondary, indirect and aggregated factors influencing the first, which can be divided into supply-side factors, demand-side factors and terms of trade or wages (or the relative valuing of labour and its fruits). The last of this second set was elaborated by placing terms of trade within theoretical debates around productivity, or what I call the fallacy of productivity reductionism.
These interacting dimensions help to clarify how different approaches to conceptualising poverty and its reduction, as well as theoretical perspectives in economics and social sciences more generally, usually place selective emphasis on different combinations of elements across these two dimensions, even though all elements are needed to understand the evolution of social needs and the reproduction of poverty. While each approach might have its own insights, the causes of poverty can only be fully understood from a broader holistic perspective of all that is involved in the creation and division of wealth.
More specifically, prevailing approaches to understanding poverty, which implicitly rely on neoclassical theoretical approaches, generally overlook several key dimensions, such as demand-side factors (with respect to employment and growth) and supply-side factors (with respect to social provisioning). They do emphasise the productivity of individuals, often through discourses of self-discipline and self-responsibility, but without more complex consideration of developmentalist strategies for increasing national productivity, to support individual efforts and to avoid them becoming undermined through fallacies of composition (that is, when everyone does the same thing, everyone’s investment is lost through falling prices). Instead, free markets and private property rights are generally emphasised as being the driving forces of productivity, usually with reference to the theoretically abstract and empirically problematic, but hugely influential New Institutionalist Economics literature, perhaps best represented by North (1990) and Acemoglu et al. (2001). Or else, behavioural approaches, such as that of Banerjee and Duflo (2011), tend to have a narrow conception of the social reproduction of poverty and assume that behavioural fixes are largely sufficient or at least expedient ways to overcome ‘poverty traps’ and launch poor people into virtuous cycles of rising prosperity, which becomes growth in the aggregate. The fact that they rarely deal with the aggregate and how aggregate convolutions might express themselves at the micro-level is perhaps one of the weakest links in the argument for relying exclusively on ‘micro-foundations’, which is invariably made in reference to abstracted neoclassical conceptions of how individuals and firms act.
These two dimensions and their elements also underpin and overlap the more conventional consideration of policy realms as mentioned above. Social policy is often thought of in terms of the spending side of redistribution. However, it also serves important distributive functions and, as seminally highlighted by Mkandawire (2004, 2005), has also played important functions in supporting productive development policies in past cases of late industrialisation. It also needs to be examined from both demand and supply sides, from multiple angles. Development policies, while generally seen as supply-side and productive in nature given their focus on increasing productivity and output, also carry important distributive and redistributive functions that, if ignored, are often their undoing. The general realm of macroeconomic policy identified by Cornia (2006) contains several functions and would be better divided into monetary and fiscal policy, because the two are quite distinct. Fiscal policy has an obvious central role in redistribution and in production, and strong influences over distribution, as well as on demand and supply sides. Monetary policy in turn is never neutral, in the sense that it effects debtors and creditors differently, and in this respect it is fundamentally redistributive (just not in the way that we conventionally conceive of redistribution). It can also have important impacts production (through influencing the cost of borrowing and hence investment) and even distribution through its effects on wages and profits.
Social policy, however, is the realm in which much of poverty alleviation strategy is practised. As such, it reserves special attention in the penultimate chapter of this book. In particular, it is central to understanding processes of social integration and ordering that are so crucial to understanding of the reproduction of poverty and the changing nature of social needs within a context of development, as highlighted in this and in previous chapters.