Substantial heat and in some ways limited light surrounds the linkages between globalisation and poverty.
Despite the allegations and concerns that globalisation adversely affects the poor (Khor 2002; Watkins 2002) and especially so in poorer countries, and counter-claims that it has “supported” poverty reduction (World Bank 2002), conclusive evidence on the linkages and the magnitude of effects is difficult to find.There is extensive literature discussing the possible channels through which globalisation and poverty are linked and a body of theoretical and empirical literature seeming to provide evidence either in one direction or the other (O’Rourke 2001; McKay et al. 2000; Dollar and Kraay 2001; and Weller and Hersh 2002), so the outcomes are, at best, ambiguous.
Precise definitions of globalisation are elusive, but it is usually perceived as a process of increased integration between and within countries,manifested through an increase in the movement of commodities, labour, capital(financial and physical capital), and technology. The wide-ranging nature of these globalisation processes and the inherent difficulties in identifying and measuring them, in terms of either their initial shocks or their impacts on various parts of the economic system (especially their impact on welfare and poverty) creates a number of analytical and empirical challenges. Even if we focus on just one kind of globalisation shock, trade liberalisation, and consider the consequences analytically (via a modelling framework), the shock might be seen to have quite different effects in different models, under different configurations of otherwise similar models, and based on different experiments (e.g. using different replacement taxes to make tariff reductions revenue neutral).
Equal care is required when clarifying the poverty side of the globalisation and poverty link. It can be interpreted as “money-metric”-based poverty or expressed in terms of more broadly based social indicators (Ravallion 1993). Clearly, the measure of poverty used matters because different poverty indices do not correlate perfectly. Monetary-based measures also vary according to the choice of relative or absolute numbers below poverty lines, on nominal or real incomes, the reference unit (households, individuals, household subgroups); or they may rely on measures of relative poverty based on skilled/unskilled wage differentials, or on the relative incomes of population subgroups.Data on poverty is often fragmentary and where available data may be inconsistent one source compared with another, so inferences are quite difficult to make.
Many of these analytical difficulties have been addressed in an attempt to identify some linkages between globalisation and poverty in four South Asian economies in a project involving researchers from Bangladesh, India, Pakistan and Sri Lanka. Each researcher executed modelling work aimed at quantitatively evaluating the sign and significance of key elements of linkage; for example, how capital flows and changes in foreign remittances occurring simultaneously with tariff changes influence the results of tariff/inequality studies; how significant export surges in garment industries have been in reducing gender inequality; how the separate influences of trade and technical change occurring under globalisation can be measured; and other such targeted analyses.
In the course of executing this work, broader questions have also been considered. Just what has been the record on poverty and inequality change in these countries as globalisation processes have occurred? When did major globalisation shocks (such as trade liberalisation) occur and how did inequality measures seemingly respond and when? What does a crude data-based analysis of linkage suggest? What are some of the pitfalls in using data in model-based counterfactual analyses in trying to unearth directions and size of linkage mechanisms? These issues are addressed here by considering some of the broader themes and results that emerge from the project.
In assessing the role of various elements of linkage, a number of points can be made. First, the choice of poverty measure matters. This is not simply a matter of distinguishing between relative and absolute poverty. Outcomes may differ as between measures based on income, consumption, health and education, and other dimensions of human well-being. Second, the structure of models used to unearth linkage is critical. Models with specific immobile factors have localised rents that change in a narrow and prescribed way with trade liberalisation; models with mobile factors do not have this feature. Quotas (if unauctioned) confer rents which liberalisation takes away; quotas which are auctioned or sought (rent-seeking) do not have these features. Third, trade policies that raise revenues (tariffs), if replaced by similar revenue raising instruments (a progressive income tax, or a VAT), may see their perceived inequality effects largely determined by the replacement policies. These are not just abstract, analytical issues, although they can be addressed using models. They reflect differences in economic circumstances that will influence the impact of trade liberalisation on poverty.
Many other pitfalls exist in such analyses, to the point that one can argue that meaningful discussion of the globalisation–poverty linkage can only take place if very precise contours for the discussion exist. Is globalisation inequality-worsening in a particular model of a particular economy using certain assumptions and conducting a precise experiment? Without such specificity, precise answers to the linkage conundrums cannot be given; with any small change in setup the answers could change, often dramatically. On top of this, globalisation is likely to impact on economies in different ways because the appropriate “contours” differ between countries; and because of other factors, such as the influence of institutions and of policy impacts other than those that are directed at liberalisation, which affect the “other things being equal” assumptions modellers are obliged to make.
In taking this perspective, this project differs from the Department of International Development (DFID) handbook on trade liberalisation and poverty (McCulloch et al. 2001) that provided the background framework for the DFID White Paper on globalisation. The authors argued that ‘in general, trade liberalisation is an ally in the fight against poverty’ (McCulloch et al. 2001: 3). It did not draw a clear distinction between relative and absolute poverty, although the impacts might be quite different. While it focused on impacts on goods prices, wages and employment, and outlined alternative detailed pathways through which trade liberalisation might have a direct effect on poverty, the study acknowledges that the total effect might be ambiguous, and this accords with our own experience. McCulloch et al. also suggested that agriculture and services were key sectors for poverty alleviation, although they provided limited intersectoral analysis and did not stress the many pitfalls in assessing the nature and magnitude of such a linkage.
The rest of this article is organised into three sections. Section 2 briefly considers the evidence and some of the difficulties (both conceptual and practical) in determining when liberalisation occurs and at what speed, and the apparent changes in poverty during the period of liberalisation. In two of the countries (Sri Lanka and Pakistan) liberalisation was gradual, whereas in India and Bangladesh, liberalisation was much more rapid. Section 3 then draws together some experience from the simulation exercises. There is considerable variety in terms of both the experiments and the modelling approaches. We purposely do not use one single generic model and apply it to each country. Instead, the models range from small, stylised, and more transparent models to much larger, computable general equilibrium (CGE) models embracing macro and micro closures. The final section, Section 4, draws some conclusions about the broad effects of globalisation on poverty from these experiments.
This article comes from the IDS Bulletin 35.1 (2004) Globalisation and Poverty: Implications of South Asian Experience for the Wider Debate