World Development Vol. 34, No. 8, pp. 14051418, 2006 2006 Elsevier Ltd.

. All rights reserved 0305-750X/$ - see front matter www.elsevier.com/locate/worlddev

doi:10.1016/j.worlddev.2005.10.011

Globalization, Local Ecosystems, and the Rural Poor

RIMJHIM M. AGGARWAL * Southern Methodist University, Dallas, TX, USA
Summary. In this paper, we draw upon models from ecology and New Institutional Economics to examine the various mechanisms through which globalization can lead to loss in resilience of ecosystems and thus increase the vulnerability of poor people who depend on it. To illustrate ecological dynamics, we examine a semi-arid savanna ecosystem that is characterized by nonlinearities and multiple steady states. We discuss how traditional knowledge and institutions aect resource use patterns and resilience of such an ecosystem in the absence of trade. Then we examine the eects of trade liberalization and international technology transfers on institutional and ecological dynamics, and consequently, on poverty. 2006 Elsevier Ltd. All rights reserved. Key words poverty, globalization, environment, ecosystem, resilience, institutions, semi-arid savanna

1. INTRODUCTION As the world economy becomes increasingly integrated through trade liberalization, international technology transfers, and greater mobility of capital and information, there is a growing interest in examining how these changes (referred to as globalization) will impact the wellbeing of the poor in developing countries. Globalization is a multidimensional process, and it may aect poverty though several complex pathways. Among these alternative pathways, an important but relatively less studied pathway works through the eects of globalization on local ecosystems. 1 Local ecosystems constitute a critical link because rural poor in developing countries derive a large part of their incomes from local natural resource based activities such as crop and livestock production, shing, hunting, fuel wood, and minor forest product collection (Dasgupta, 1993). Thus, it becomes natural to ask how globalization, in its various dimensions, aects local ecosystems, and thus the wellbeing of the rural poor who depend on it. A somewhat dierent, but related question, of how trade liberalization aects the environment has been extensively examined in the economics literature. In particular, several theoretical and empirical studies have examined the validity of the Environmental Kuznets

Curve (EKC), which hypothesizes that with rise in per capita incomes, environmental degradation rst increases and then falls. Therefore, in so far as globalization leads to rise in per capita incomes, environmental degradation is likely to increase with increased incomes for poor countries while it falls for the richer countries. In a recent survey of empirical studies on EKC, Copeland and Taylor (2004) found no conclusive evidence to support the existence of a simple and predictable relationship between environmental degradation and per capita incomes. They nd the estimated relationship between per capita incomes and environmental quality to be highly sensitive to the type of pollutant, the time period, and the sample of countries/cities used in the study. Another well-known mechanism relating globalization to environmental degradation works through the export of dirty industries and the creation of pollution havens in developing
* An earlier version of this paper was presented at the UNU-WIDER Conference The Impact of Globalization on the Worlds Poor: First Project Meeting on Conceptual Issues, Helsinki, October 2930, 2004. I would like to thank K.L. Mehra, Machiko Nissanke, Alice Sindzingre, Erik Thorbecke, and an anonymous referee for extensive comments on an earlier version of this paper. Final revision accepted: October 20, 2005.

1405

1406

WORLD DEVELOPMENT

countries due to their weak environmental standards. Empirical evidence suggests that this has not happened on any signicant scale, possibly because the costs imposed by environmental regulations are small relative to the other costs that aect location decisions of Multinational Corporations (MNCs) (Eskeland & Harrison, 2003; World Bank, 2002). An important contribution of this extensive research relating to EKC and pollution haven hypothesis has been that it has led to demolishing the widely held common belief that trade and growth inevitably lead to environmental degradation. This literature has also highlighted the role of institutions and endogenous policymaking, together with conventional factors such as factor endowments, capital accumulation, and technology, in mediating the eects that trade has on the environment. However, an important limitation of this literature is that it has almost exclusively focused on industrial pollutants. As Copeland and Taylor (2004) observe in their survey of this literature, there is very limited empirical work examining the eects of trade on renewable resources such as cropland, pastures, forests, sheries, and water. 2 This is an important shortcoming, particularly if one is interested in examining the eects of trade liberalization on rural poverty, since rural livelihoods are inextricably linked to extraction of these resources. In the limited empirical work that exists on the eects of trade on renewable resources, there is a tendency to focus on a single resource in isolation from the rest of the ecosystem in which it is embedded. In contrast to this, models from ecology often emphasize the need for a system wide approach that pays explicit attention to the feedback and linkages that exist among the key variables that constitute the ecosystem. Until quite recently, ecologists did not pay adequate attention to the working of human systems (in terms of the social, economic, and political dimensions) and its interaction with the ecological system. In recent years, there have been some attempts at analyzing the dynamic interactions of human systems and ecosystems. 3 In this paper, we draw upon some of the ideas put forth in this literature to examine the various mechanisms through which globalization, in its various dimensions, can lead to ecosystem degradation and poverty. As discussed above, an important lesson to emerge from previous research is that there is no simple predictable relation between trade and environment, and that institutionsto-

gether with other national characteristicsplay an important mediating role. Thus, the approach we follow in this paper is to draw upon models on ecosystem dynamics from ecology and link them to models in New Institutional Economics that examine how institutions and technologies evolve. The underlying idea is to rst understand how human and natural systems dynamically interact within a closed community and then to examine the conditions under which integration into the global economy degrades the local ecosystem and its consequences on poverty. The central ideas within the paper are developed as follows. First, we discuss some key conceptual issues related to the dynamics of ecosystems. An ecosystem can be viewed as a system of biotic and abiotic variables that are associated with dierent temporal scales (some change fast while others are slow) and dierent spatial scales (some stretch across several regions while others are relatively localized). Ecologists believe that these cross-scale interactions between constituent variables lie at the heart of some of the most interesting characteristics of ecosystem dynamics (Holling, 1986). In economic models, it is generally assumed that either economic activities do not impact the underlying biophysical environment or that the impact varies smoothly (often linearly) with the amount of stress. Ecologists, on the other hand, have found these dynamics to be largely nonlinear, and often associated with sharp discontinuities, high uncertainties, and alternative stable states. To illustrate these dynamics, we examine a prototypical model of a semi-arid savanna ecosystem that is characterized by multiple steady states and threshold eects. An important ecological concept in this context is that of resilience, which refers to the ability of ecosystems to absorb shocks without changing their essential structure. In economic models, only the productivity of a resource is considered. We discuss why resilience is also an important characteristic and why management institutions that focus only on short-term productivity may lower resilience and contribute to the emergence of crisis. Lower resilience of the ecosystem, in turn, implies greater vulnerability of the poor people who depend on that ecosystem for their livelihood. Thus, studying resilience of ecosystems is critical for poverty analysis. Next, we examine how these ecosystem dynamics interact with the dynamics of human systems. Here we rst consider the case of a tra-

GLOBALIZATION, LOCAL ECOSYSTEMS, AND THE RURAL POOR

1407

Ecosystem state

ditional closed economy and use this discussion as a benchmark to later examine the impact of globalization. Within the context of a traditional closed economy, we discuss how the shared memory of long-term experiences shapes resource-use practices and how these practices fare in terms of productivity and resilience ` vis-a-vis the modern resource management regimes. From an ecological perspective, globalization refers to a process of integration across space of some (not all) constituent variables of the humanecological system. This integration happens at a dierent pace for different variables, and this process transforms the earlier structure of relations across space and time. Some variables, such as institutions, change much slower than other economic variables and we discuss how these become important determinants of the outcomes that emerge. Integration into the world economy may happen through several pathways. We examine, in particular, the eects of trade liberalization and technology transfer on institutional and ecological dynamics, and consequently, on poverty. 2. ECOLOGICAL DYNAMICS Ecosystems have been changing constantly even before humans rst appeared on the planet. Our interest here is to present some stylized versions of how human intervention may impact the ecosystems and how ecosystems respond back to such interventions. Following Scheer, Westley, Brock, and Holmgren (2001), we will use stress as a general term for the eect of human use through harvesting, destroying biomass or aecting abiotic conditions (e.g., groundwater reduction or climate change). (a) Stylized versions of ecosystem response to stress In Figure 1a and b, we present two dierent stylized versions of the relation between stress and state of the ecosystem. 4 In economic theory, it is generally assumed that either human activities do not have any signicant eect on the biophysical environment or that the eect, if signicant, varies smoothly with the extent of stress, as shown in Figure 1a. It is increasingly being recognized by ecologists that this representation is highly simplistic and does not fully capture the complexity of ecosystem

(a)

Ecosystem state

Stress

F2

F1

(b)

Stress

Figure 1. Schematic representation of two possible responses of ecosystems to stress imposed by human use (adapted from Scheer et al., 2001).

dynamics that are often marked with sharp discontinuities and multiple stable states. Very often an ecosystem may not show any signicant visible signs of stress until it ips over to another state when certain thresholds are crossed, as shown in Figure 1b. Scheer et al. (2001, p. 197) describe the case in Figure 1b as one where the ecosystem response line is folded backwards. This is known as a catastrophe fold and implies that the system has two alternative stable states over a range of environmental conditions. When the state of the ecosystem lies in the upper branch of the folded curve, it cannot pass to the lower branch smoothly. Instead, when conditions change suciently to pass the threshold a catastrophic transition to the lower branch occurs. The movement from one stable state to another (also sometimes referred to as a phase transition) may not be reversible, or the

1408

WORLD DEVELOPMENT

reversal may take a very long time or be achieved at a very high cost. An important feature of these systems is that to induce a switch back to the upper branch (in Figure 1), it is not sucient to restore the environmental conditions of before the collapse. Instead, very often, one needs to go back further, beyond the other switch point, where the system recovers by shifting back to the upper branch. This pattern, in which the forward and backward switches occur at dierent critical conditions, is known as hysteresis. A number of ecosystems have been found to exhibit such dynamics. Some widely studied examples include semi-arid savanna grasslands, pest infestations in forest and agricultural ecosystems, and the eutrophication of shallow lakes. In order to further clarify these ecological dynamics and their implications, we describe the case of semi-arid savannas in detail. The semi-arid savannas are home to a large concentration of poor people around the world and the transformations in this ecosystem serve as a useful illustration of the central arguments we propose in this paper. (b) Ecological dynamics in semiarid savanna ecosystems 5 We use the term semi-arid savannas to refer to those regions of the world, which in their natural state have a predominant continuous grass cover with scattered to numerous trees and shrubs (Walker et al., 1981, p. 173). Semiarid savanna grassland systems that were considered to be productive pasturelands have been transformed in to arid shrub lands (often also referred to as wastelands) in several regions such as the Sahel zone of Africa, southern and eastern Africa, and northern India. Semi-arid savanna is a water-limited system. Rainfall is scanty and interrupted by periodic droughts. Thus, a critical ecological variable is the water inltration rate. Biomass of grass is a critical factor in determining the rate and amount of inltration. As grass cover declines (say due to intense grazing), surface pores become sealed and soil erodes, and both of these processes lead to decrease in water inltration rate. The typical vegetation in a semi-arid savanna consists of a mix of grasses (G) and woody vegetation (W). Grasses and woody vegetation both compete for available water on the upper layers of the soil. Grasses are more ecient than woody vegetation in extracting water from upper layers of soil. Thus, in a year

with average rainfall in savanna region, grass grows much more rapidly than woody vegetation. But below the grass root zone (subsoil), woody vegetation has nearly exclusive use of whatever water gets through. Thus, under drought condition, woody vegetation performs better than grass. Cattle and sheep, which are the major herbivores in the region, selectively eat grass and only negligible amounts of woody vegetation. Thus, under high stocking levels, common pattern of range deterioration is as follows. Grasses are aected more adversely than woody vegetation. Short periods of high grazing may reduce grass cover to some very low level. But this may not cause an immediate change in equilibrium conditions, because of slower response of water inltration rate and even slower response of woody vegetation. However, if the grass biomass is kept at a very low rate for a prolonged period then slowly the soil surface, and consequently the water inltration rate, declines. The result is that less water enters the soil, and proportionally more of what does enter penetrates to the subsoil level. So the woody vegetation has relatively more water available, and it begins to dominate. The reduced inltration and greater biomass of woody vegetation combine to prevent the reestablishment of grasses, even if grazing pressure is reduced. It is only if one or both of these factors are changed to allow the grass to develop above the critical unstable equilibrium in Figure 1 will the system revert back to its original state of predominance of grass biomass. As the woody vegetation expands over time, the root system becomes relatively extensive and persistent. Thus, the competitive eect of the root system of the woody vegetation must be removed for long enough to allow for grass root systems to develop to the point where it can take up more water than the tree roots. It is important to note that the initial state with low stocking rates and dominance of grasses and the new state with dominance of woody vegetation are both stable as in Figure 1b. The case of semi-arid savanna ecosystem illustrates several salient features of ecosystem dynamics that we use as a background for discussion in the rest of the paper. Thus, for instance, many ecologists believe that understanding cross-scale interactionsthat is, interactions between fast and slow moving variables, as well as the interaction between variables that have a limited local impact as opposed to those that have a wider impact

GLOBALIZATION, LOCAL ECOSYSTEMS, AND THE RURAL POOR

1409

spatiallyis very important in understanding ecological dynamics. This phenomenon is well illustrated in the above discussed case of semiarid savanna ecosystem through the interaction of grass coverage (a fast changing variable) with woody vegetation (a relatively slow changing variable), and water inltration rate and soil characteristics (the slowest changing variables). Variables that change very slowly are often assumed to be constant and we will discuss how this is an important source of dierence in perceptions of human societies at dierent historical periods. The above discussion also points to the presence of multiple equilibria, threshold eects, and hyteresis in ecological dynamics, which as we explain later become important in understanding the emergence of poverty traps. (c) Resilience of ecosystems and its implications In the presence of multiple equilibria and threshold eects, an important concept that is useful in understanding the impact of stress on ecological dynamics is that of resilience. Resilience is dened as the ability of a system to maintain its structure and pattern of behavior in the face of disturbance (Holling, 1986, p. 297). Ecosystems are constantly changing and adapting to dierent kinds of stresses. Thus, from a policy perspective the relevant question isto what extent can the ecosystem absorb change without changing its basic structure, that is, without ipping from one state to another functionally undesirable state, as in the case of semi-arid savannas described earlier. The concept of resilience is very useful because it emphasizes the qualitative properties of an ecosystem in terms of its structure of relationships between dierent kinds of variables and processes that control system behavior. This also underlies the dierence between the concept of resilience and stability of an equilibrium point. Stability refers to the propensity of populations in an ecosystem to return to an equilibrium point following a disturbance. The dierence between the two concepts lies in the focus of analysis within an ecosystem stability is dened in the context of a micro-focus on a specic population within an ecosystem while resilience refers to a macro focus on the structure of relations between the dierent populations (Common & Perrings, 1992). Dierent structure of relations can produce multiple stable states, each characterized by

its own domain of stability. Sometimes, the changes within an ecosystem are continuous as when the system moves within the same stability domain. At other times, an exogenous event and/or changes induced by internal dynamics can trigger a discontinuous movement across dierent stability states. An important determinant of resilience of ecosystems is biological diversity (Holling, 1986). This is because resilience of an ecosystem depends on the existence of a diverse set of species that are capable of serving key ecological functions under dierent environmental conditions (Perrings, 1995). Thus, loss of diversity implies that the alternate pathways through which stress could earlier be transmitted are no longer available. This in turn implies greater vulnerability. As key variables (e.g., species composition, age structure, spatial distribution) become more homogenousstability domains shrink and perturbations that could earlier be absorbed now cascade upwards to produce more drastic eects. In the next section, we use the above discussion of ecosystem dynamics and the concept of resilience to examine how human systems interact with their local environment. 3. SOCIOECONOMIC AND ECOLOGICAL DYNAMICS IN A CLOSED TRADITIONAL ECONOMY Humans have been active agents in the evolution of ecosystems. However, until quite recently, the study of ecosystems and human systems developed in relative isolation. The dierent elements of social and economic organizationsuch as technology, institutions, values, and culturesco-evolve with ecological variables. Thus, for instance, a change in local environmental conditions (say in the form of decline of certain resources or fall in catch per unit eort) may trigger a response in rules governing resource management. The latter, in turn, may aect resource dynamics in the future. The analysis of these feedbacks from natural to human systems within a dynamic framework is crucial. Unlike other species, which are largely genetically programmed to perform in certain dened ways, humans have the capability of conscious actions. This opens up a myriad of possibilities in their interaction with other humans and with their ecosystems. In this section, we outline some interactions between the ecological system and the socio-economic system in a traditional closed economy.

1410

WORLD DEVELOPMENT

Although trade and markets have been in existence even in pre-modern societies, in this section we discuss a stylized version of a tightly knit closed economy to provide as a benchmark for our discussion on the impact of globalization in the next section. (a) Role of traditional institutions An important channel through which the relation between ecosystems and human systems is mediated is through institutions. Following North (1989), we dene institutions as the constraints that structure repeated human interaction. Customs, conventions, norms, and values that inuence and guide individual behavior are examples of informal institutions. More formal institutions include markets, labor unions, stock markets, and property rights. Our focus here is on institutions that govern the use and management of natural resources. Natural resources in many rural settings can be characterized as common pool resources (CPRs) where costs of exclusion are high and joint use involves subtractability (Ostrom, Gardner, & Walker, 1997). Given these characteristics of CPRs, institutions serve two critical management roles. First is regarding how to govern access to the resource (exclusion problem) and the second relates to instituting rules among users to solve the potential divergence between individual and group rationality (subtractability problem) (Berkes & Folke, 1998). It was believed earlier that users of a CPR would end up in an inescapable tragedy of the commons. However, several studies over the past two decades have drawn attention to the complex set of rules and regulations that exist in many rural communities to deal with the exclusion and subtractability problem discussed above (Berkes, 1989; Ostrom, 1990). Thus, the tragedy of the commons is not inevitable. A diversity of outcomes is possible ranging from a situation where there are no regulations governing access (open access) and the tragedy of the commons does occur to a situation where rights of access and joint use are clearly dened (common property) and the tragedy is averted. Property rights arrangements are often very complex in traditional societies because these involve a bundle of rights including use rights, rights to exclude others, rights to manage, and rights to sell. For instance, among the Barabaigwho are semi-nomadic pastoralists in the semi-arid Hanang district of Tanzaniaa bundle of rights exist for pastures,

trees and water resources (Lane, 1992). Open rangeland is regarded as property of the community and its use is regulated by a set of customary rules. The Barabaig, however, also recognize private property in the form of a homestead and its surrounding areas. Permanent settlements along lake marginswhere the wells provide permanent sources of waterare restricted as this would result in depletion of the pastures near the water source, which is critical for the community in the dry season. Water routes are also closely protected and homesteads are not allowed to be built there. This structure of rights has evolved in close accordance with the underlying ecological conditions and the rotational grazing patterns practiced by this community of pastoralists. Another important point that comes out in Lanes (1992) description of this community is regarding how such traditional rights structures, although more complex than their modern counterparts, are also more adaptive to changing ecological conditions. Particularly in ecosystems characterized by high uncertainties and multiple steady states (as in Figure 1b), adaptability of resource management institutions to changing ecological conditions is very critical. As we discuss later in this paper, many resource crises have been found to occur in cases where these institutions have become rigid and unresponsive to environmental feedbacks. 6 (b) Role of traditional knowledge systems and practices As discussed above, the close dependence of traditional societies on the natural environment led them to be highly responsive to environmental feedbacks. Through a process of trial and error, they learnt how to look for signals of environmental distress and how to respond to it through development of exible institutions. This knowledge was then transmitted through rituals, religious practices, and oral history across generations. The embedded long-term institutional memory of shared experience in these societies forms the basis of customary practices and institutions for use and management of natural resources (Berkes & Folke, 2001). In their extensive research on traditional ecological knowledge systems and practices, Berkes and Folke (2001) found that unlike modern management regimes, which are science-based and draw on a short span of recent

GLOBALIZATION, LOCAL ECOSYSTEMS, AND THE RURAL POOR

1411

experience, the institutional memory embedded in traditional systems leads resource users to be highly aware of long-term trends in resource populations. In particular, they found that traditional knowledge systems have certain similarities and parallels to the theory of complex systems, with emphasis on nonlinear relationships, threshold eects, multiple equilibria, the existence of several stability domains, crossscale linkages in time and space, disturbance, and surprise (2001, p. 124). To support their argument, they give examples of several traditional resource use practices that mimic the behavior of such systems. An important example is that of pulse grazing practiced in several traditional pastoral communities throughout the world. Under this practice, a piece of pasture is grazed intensively for some time and then allowed to rest. Recent ecological studies have found that these pulses of grazing contribute to the capacity of semi-arid grasslands to function under a wider range of climatic conditions and thus contribute to greater resilience (Holling, 1986). The awareness of uncertainties and multiple stable states also led traditional societies to devise various practices that could serve as buers and provide informal insurance. Thus, for instance, under rotational grazing patterns, certain pastures were left to be used only in dry seasons and times of crises (Lane, 1992). Models of resource use based on linear relationships may nd these resource practices to be unproductive over the time horizon generally considered in these models, neglecting their role in enhancing long-term resilience. Global integration has led to a breakdown of many of these traditional knowledge systems, as we discuss in detail in the next section. 4. IMPACT OF GLOBALIZATION We will dene globalization as a process of integration of ows (such as trade, capital, labor, and information) and the policies that facilitate such ows (in the form of reduction of barriers on trade, nancial ows and migration). Individuals have been exposed to global ows ever since the advent of trade in human history. However, as argued in World Bank (2002, p. 23), historically before about 1870 none of these ows was suciently large enough to warrant the term globalization. Since the 1870s, there have been periods of more or less globalization but in this paper, we will ab-

stract away from these details and focus on the long-term impact of globalization on a traditional agrarian economy. In Section 2, we discussed how an ecosystem can be viewed as a system of interrelated variables that operate at dierent scales spatially and temporally. Extending this insight from the ecological literature to economics, the process of globalization can be conceptualized to have the following eect. It leads to an integration of some (not all) of the constituent variables across space. The dierent rhythms of integration and transformation of constituent variable leads to dierent outcomes depending on how the resultant opportunities/tensions are utilized/resolved. Thus, for instance, pests or disease pathogens (as in case of HIV) that earlier had a local eect spread out more across space. However, it is not generally true that their natural predators/host in the original ecosystem would also spread out similarly across space. Similarly, goods and services may be traded globally but institutions, culture, and values remain more strongly embedded locally. As described earlier in Section 2, it is these slowly changing/integrating variables that become an important determinate of the nature of outcomes that emerge. New congurations of variables with dierent spatial and temporal scales emerge and it is dicult to say a priori how the outcomes would aect ecosystem resilience and the welfare of the poor. We outline below some pathways through which the eects could be transmitted. (a) Trade liberalization, specialization, price volatility, and loss of resilience A useful way of studying the impact of trade liberalization on the environment is to decompose the total eects in terms of scale, composition, and technique eects. 7 In terms of the model presented in Section 2, the scale eect measures the change in the state of the ecosystem if the economy were simply scaled up, holding constant the mix of resources extracted and the production techniques. The composition eect measures the eect of change in a mix of dierent resources extracted, keeping constant the scale and production techniques. Finally, the technique eect measures the eect of changes in techniques of resource extraction, keeping the scale and composition constant. In this subsection, we will discuss the scale and composition eect. The technique eect is discussed in the next subsection.

1412

WORLD DEVELOPMENT

In so far as trade and growth bring about greater opportunities and an increase in incomes, there is an upward scale eect, which increases stress on the ecosystem in Figure 1b. This scale eect is further compounded by the composition eect when there is specialization in natural resource extraction activities following trade liberalization. In a widely cited paper on NorthSouth trade, Chichilnisky (1994) begins by observing that developing countries tend to specialize in export of goods, which deplete environmental resources. She argues that this is not necessarily because they are better endowed with such resources. In fact, she shows that even if the two countries are identical in all respects except that one of them (North) has well-dened property rights on natural resources while the other (South) does not, trade will occur between them with the South specializing in production of natural resource based activities. In her model, trade between these two countries increases the overuse of resources and makes the misallocation of resources worse. 8 In cases where the underlying ecosystem exhibits the dynamics represented in Figure 1b, there is a high risk that the sharp hike in resource extraction activities in the postliberalization period may push the ecosystem over the threshold into another less productive state. As opposed to the pre-liberalization period when the scale of production is limited by the size of the domestic market, the need for well-functioning regulatory institutions is very critical in the post-liberalization period as we discuss in Section 4(c). Another important pathway through which specialization in a narrow range of natural resource based activities aects the ecosystem is through the loss in bio-diversity (Brock, Maler, & Perrings, 2001). As discussed before, loss in biodiversity makes the ecosystem less resilient because of a reduced number of pathways through which stress in the environment can be absorbed. This happens, for instance, when forest ecosystems rich in biodiversity are transformed for the purposes of growing a narrow range of crops for exports. This also happened in the earlier discussed case of semi-arid savanna ecosystems in the Sahel region upon opening of external markets. Here the wide diversity of perennial and annual varieties of grasses that occur naturally (with dierent degrees of palatability to herbivores and drought resistance) were replaced by a couple of annual varieties that grow faster but have much lower drought resistance property (Holling, 1986). These

changes, carried out over vast tracts of land, reduced functional diversity and increased spatial uniformity in grassland ecosystems. The overall long-term eect has been an increased vulnerability to external shocks such as droughts. Market integration has also led to higher spatial connectedness and thus the possibility that shocks that earlier had only local impacts are now transmitted more widely (Holling, 1986). Thus, on the one hand, the economy has become exposed to higher risks due to price volatility in international markets. On the other hand, the ecosystem has become more vulnerable and thus less capable of handling any external shocks. (b) International technology transfers and myopia In an interesting historical analysis of the impact of technological development on ecosystems, Westley, Carpenter, Brock, Holling, and Gunderson (2001, p. 117) observe that during most of the twentieth century the goal of technologically based resource management has been to control the external sources of variability in order to seek a singular goal such as maximization of yield (tress, sh). Thus, the focus is on a small set of target variables and there is a tendency to pay insucient attention to the eects of the intervention on other parts of the system. Thus, although the intervention might be successful at rst in its stated objectives, but then slowly as other parts of the system change, new problems often emerge. To support their argument, Westley et al. give the example of industrial nitrogen xation to make fertilizers. This innovation made an important contribution to increasing agricultural production but created unexpected long-term side eects, including toxic levels of nitrate in groundwater and widespread acidication of ecosystems. Another important example comes from modern forest management regimes where frequency of forest res was successfully reduced in order to promote recreational demand. With reduction in frequency of res, more fuel accumulated over time and forests crowns closed. All these changes created a forest system in which whenever a stochastic ignition event such as lightening causes re, the intensity is much higher and the consequences are much more catastrophic than before. Thus, although a longer period elapses before a re occurs, the impact is more drastic because of changes in

GLOBALIZATION, LOCAL ECOSYSTEMS, AND THE RURAL POOR

1413

the structural characteristics of the forest system. 9 Similarly, in many developing countries, transient success in spraying DDT and controlling mosquito populations led over time to human populations with little immunity and mosquito vectors resistant to DDT. As a consequence of these changes, some countries reported a 3040 fold increase in malaria cases in the 1980s compared with 196970 (Holling, 1986). In recent years, there has been a growing recognition of these long-term system wide eects and new technologies have been developed as a consequence. However, there are a number of reasons why resource management techniques in developing countries continue to be plagued by myopia and tend to pay insucient attention to system wide eects. First, with globalization, resource management techniques developed in Western countries in the context of temperate ecosystems are often transferred without much modication to other types of ecosystems worldwide. Thus, although the original technological innovation might have been made with sucient attention to its long-term eects on the local ecosystem, the same level of caution may not carry over as the technique is transferred to other contexts. 10 Second, it is important to keep in mind that MNCs have been responsible for a large part of international technology transfers in recent years. 11 These corporations that are motivated by global output or prot maximization have little interest in long-term domestic impacts of their operations. Finally, there is some limited evidence to show that the burden of large international debts leads governments in developing countries to adopt resource management practices that can quickly earn foreign exchange without adequate attention to their long-term impacts. 12 (c) Institutional dynamics, rent seeking, and poverty In the previous two sections, we have elucidated the eects of trade liberalization and international technology transfer on local ecosystems. However, this story is incomplete without discussing the role of institutions in mediating these eects of global integration on ecosystems and on poverty. In Section 2, we discussed the functioning of institutions in a traditional closed economy. We now examine how globalization aects the evolution of resource management institutions and the consequences of this evolution on poverty.

Recent developments in New Institutional Economics have made a signicant contribution towards our understanding of existing institutions. However, these theories here are still somewhat vague about the mechanisms through which new institutions emerge and why dysfunctional institutions continue to persist over long periods (Akerlof, 1984). As Bardhan (1989) observes, the biological analogy of natural selection in the survival of the ttest institution cannot be applied to understand the evolution of institutions. Based on his long-term study of several ecosystems around the world, Holling (1986) makes an interesting observation. He argues that modern resource management institutions and technologies that became more ecient over time in meeting short-term production goals and controlling variability also became more rigid and less responsive to environmental feedbacks. He believes this rigidity to be an important reason for a fall in long-term resource productivity and eventual crises in a number of ecosystems around the world. This phenomenon is very well illustrated by the long-term impact of the spread of green revolution in the form of monocultures of highyielding varieties of rice and wheat in many developing countries. The initial spread of green revolution technology in areas most favorable to its growth met with phenomenal success over the short term. However, soon with the help of international aid agencies and massive state support, the technology spread to areas that were not very suitable for its adoption. Thus, for instance, the technology was widely adopted even in arid and semi-arid areas as well as regions with marginal soils that were not particularly well adapted for this form of intensive cultivation of a single crop. Several recent ecological studies have compared this technology with more traditional technologies and shown how monocultures have made the local ecosystem more vulnerable to external shocks, such as that induced by weather changes or pest attacks. 13 Recent evidence points to signicant yield increases in diverse cropping systems compared to monocultures. 14 In spite of the growing evidence on the unsustainability of monocultures, this technology continues to spread. Its initial success in some areas has locked it into a groove where further innovation is restricted and more appropriate technologies never get a footing. As Bardhan (1989, p. 1392) argues, this lockin happens dynamically as sequential decisions groove out an advantage from which the system nds it hard to escape.

1414

WORLD DEVELOPMENT

North (1989) points to an interesting reason why new institutions do not automatically develop in order to handle more complex interdependence such as that brought about by globalization. He argues that the breakdown of personal exchange is not just the breakdown of a dense communication network but also the breakdown of communities of common ideologies and of a common set of rules in which we all believe (1989, p. 1321). Institutional change involves an enormous collective action problem, which may be dicult to come about when people have dierent beliefs and aspirations. For instance, consider a closed primitive community of people who are highly dependent on their local natural resource base. Over time, they evolve a set of rules governing the use and management of these resources. These rules may not be equitable but are likely to be consistent with their shared experience and collective learning about resource dynamics. Now as the community opens up and there is greater movement of people and new ideas, this shared belief system may begin to dissolve. The opening up of the community may also provide some people with exit options and thus the old enforcement mechanisms working through threat of punishment or community sanctions may begin to break. Divergence of interests within the community may also imply that the critical mass required for initiating the institutional change might not be present. It is possible that over time, as more and more people adopt the new belief system, the requisite critical mass will be attained and new institutions more suited to the changed circumstances will eventually emerge. 15 However, a distinguishing feature of developing countries today in contrast to the historical experience of developed countries is that the opening of traditional communities is happening at a much greater pace and intensity. Institutions are the relatively slow changing variables in the human system, at least as compared to some of the other changes brought about by trade liberalization (Sindzingre, 2004). Thus, as in the case of ecological dynamics model discussed in Section 2, these slow changing variables lead to threshold eects and multiple stable states. In most cases, the opening up of traditional communities in developing countries has been accompanied with rise in power of the central government and a gradual disempowering of the local communities. North (1989) argues that the rise of the state with unequal coercive power provides the opportunity for individuals

with superior coercive power to enforce rules to their advantage. This happened, for instance, in Thailand during the period when, with trade liberalization, timber became very valuable. State ocials then undermined the local institutions that had been regulating the use and management of this resource. Ross (2001) argues that by undermining the local institutions, they were able to create opportunities for corruption and what he calls rent seizing. 16 This, in turn, led to greater inequality and poverty. It is also generally observed that when large bureaucracies replace local institutions for resource management, there is less sensitivity to environmental feedbacks at the local level. As Pritchard and Sanderson (2001) observe, local communities are connected to their resources in ways that bureaucracies are not. Researchers who have studied resource management institutions in ecosystems characterized by high uncertainties and multiple steady states (as in Figure 1b) have found that adaptability to environmental feedbacks is a critical requirement for sustainability of such systems. 17 As argued in the previous section, this is one respect in which traditional knowledge systems and institutions generally fare better than modern bureaucracies do. However, community management has its own limitations. For instance, as Pritchard and Sanderson (2001) argue, face-to-face communication leads to greater possibility for preference falsication through intimidation or manipulation or the tendency to follow a leader when information is limiting. They also show how community members often loose attention and become disengaged when issues become too complex or the problem at hand poses only a slight threat to the community. Another important problem associated with resource management institutions is that very often their governance structures do not match with the scale of ecosystems. Thus, for instance, watershed management practices may require cooperation among dierent villages or sometimes even dierent countries. In this case, even if management institutions exist at the microlevel (i.e., village or the country level), these alone are not sucient. A reverse problem with scale arises when institutions at the macro level (say at national or international levels) govern diverse local ecosystems. As Costanza, Low, Ostrom, and Wilson (2001) point out, largescale ecosystems are not simply small scale systems grown large, nor are micro-scale ecosystems mere microcosms of large-scale sys-

GLOBALIZATION, LOCAL ECOSYSTEMS, AND THE RURAL POOR

1415

tems. The driving forces and feedback mechanisms in large and small-scale systems operate at dierent levels and exhibit distinct patterns. . . The solution, then, is to match ecosystems and governance systems in order to maximize the compatibility between these two types of systems (2001, p. 7). This matching of ecosystems and governance systems continues to be an important challenge for policy makers in the process of globalization. As discussed earlier, poor people depend heavily on their local natural resource base for their basic needs. The transformation of savanna grasslands to arid shrubland in the Sahel region, as well as parts of eastern Africa and northern India, has led to loss of livelihood options for the poor people who live in these regions. As discussed above, in order for the system to revert back to its original state requires a signicant eort in order to push the system back to the upper branch beyond the point where the initial shift to the new state had occurred, due to hysteresis (see Figure 1b). Moreover, the eort needs to be coordinated and collectively managed among the dierent pastoralists and this may be very dicult once the system is already caught into the lower equilibrium. There are several cases cited in the literature where local communities either on their own or with the help of an external governmental or nongovernmental agency have been able to restore the system to its original state or a dierent more productive state. 18 In other instances, such a restoration may be dicult and the best option for poverty alleviation might be to nd novel ways of dealing with the modied state. These include shift in types of crops grown, changes in stocking densities or types of animals, shift from herding to wildlife hunting, nding o-farm employment or migration. As these changes may take time, there is also a need for creation or revival of old safety nets that can help cushion the eects of external shocks in the short to medium term. Examples of such safety nets include insurance schemes, credit restocking programs and reciprocal exchanges. 19 5. CONCLUSIONS In this paper, we have examined how globalization may aect the wellbeing of the poor through its eects on the local ecosystems on which the poor depend for their livelihood.

Drawing upon models from ecology, we showed how nonlinearities, multiple steady states, and threshold eects characterize many of the ecosystems that the rural poor depend on. An important ecological concept that we discussed in this context is that of ecosystem resilience. Resilience refers to the capacity of an ecosystem to absorb shocks before undergoing a change in the structure of relations between its constituent variables and processes that control behavior. While neoclassical economics has focused almost exclusively on resource productivity, we showed why resilience is also an important factor, particularly when one is trying to understand the long-term impact of dierent resource management regimes. Since the rural poor depend on local ecosystems for their livelihood, a loss in resilience translates into a greater vulnerability to poverty. Resilience is a system-wide property that derives from the linkages and feedback eects that exist been dierent ecological variables. Therefore, when the impact of a specic policy on a single resource is analyzed in isolation from the rest of the ecosystem (as is generally done in studies exploring the link between trade and environment), the impact on resilience is overlooked. For instance, it is often the case that a policy change or adoption of a new technique increases productivity of a single resource but decreases the resilience of the ecosystem in which it is embedded. The prototypical model of a semi-arid savanna ecosystem that we presented in this paper, along with examples drawn from other ecosystems, helps us understand why this happens. This is an important insight that opens up new perspectives for exploring the links between globalization and local ecosystems. For example, we discussed how trade liberalization policies in developing countries often lead to specialization in the production of a narrow range of natural resource based activities. While the static eciency gains that follow from such specialization are well known, other long-term eects on the ecosystem are less widely recognized. An important long-term effect that we discussed in the paper is the loss in bio-diversity that leads, in turn, to a loss in resilience and a consequent increase in poverty. Similarly, we discussed several other examples to show how international technology transfers may increase short-term resource productivity but decrease ecosystem resilience. These examples shed light on some of the mechanisms

1416

WORLD DEVELOPMENT

through which globalization, although leading to new growth opportunities, may increase poverty. The example of the semi-arid savanna ecosystem also helps in understanding the interaction between dierent variables in the ecosystem that operate along dierent temporal and spatial scales (some change fast while others change slow, some have only a local eect while others are more widespread). As ecologists point out, it is these cross-scale interactions that determine system behavior. Drawing upon this insight, we showed how globalization can be viewed as an external shock that brings about rapid change in some variables (like prices or technologies) while other variables (like institutions and culture) remain sluggish. It is this change in structure of relations between dierent constituent variables, operating along dierent scales, which ultimately results in dierent outcomes. Although it may seem that the resulting equilibrium structure is determined by the fast changing variables, the slow changing variablessuch as institutionsare a critical determinant of which among the alternative stable states is ultimately observed.

The linkages between ecosystem and institutional dynamics that we have sketched here, somewhat informally, need to be developed more rigorously and tested empirically. This is likely to be challenging given the cross-disciplinary nature of such an inquiry. 20 Most of the current debates on globalization have focused on the growth of MNCs, the rising inuence of multilateral institutionssuch as the World Bank, the IMF, and the WTOand the loss in national sovereignty. We have not explicitly discussed these factors in this paper. This is not because we do not perceive these factors to be important but because the overwhelming emphasis given to them often distracts attention from our basic understanding of the underlying dynamic evolution of ecosystems and human systems. An important consequence of this mode of thinking is that we have placed far too much attention to control of external shocks and short-term crisis management as opposed to addressing the underlying problems that lead to loss in resilience and chronic poverty.

NOTES
1. Most of the studies on this subject use reduced form estimations and do not shed much light on the causal mechanisms through which globalization aects poverty. 2. An important exception here is Lopez (1997) who examines the eect of trade liberalization on exploitation of biomass through a more than optimal level of land cultivated for the case of Ghana. 3. For a comprehensive collection of papers on the subject, see Clark and Munn (1986), Gunderson and Holling (2001), and Berkes and Folke (1998). 4. It is dicult to capture the changes in ecosystems by a single state variable. However, since many aspects of the ecosystem shift in concert with a few important key state variables, we focus on a single aggregate variable such as plant biomass (see also Scheer et al., 2001). 5. The model presented here is based on Walker, Ludwig, Holling, and Peterman (1981). 6. Gunderson and Holling (2001) and Berkes and Folke (1998) contain several case studies that make this observation. 7. This is a common methodology followed in several papers on trade and environment. See, for instance, Grossman and Krueger (1993) and Copeland and Taylor (2004) for detailed denitions of these eects. 8. Instead of extraction of natural resources, when one looks at pollution from manufacturing plants, then it may be true that composition eect works against the scale eect. This is because more pollution intensive industries also tend to be more capital intensive (Copeland & Taylor, 2004). Thus, it might be the case that developing countries have a comparative disadvantage in these dirty industries and thus with trade liberalization, there is a fall in pollution. Copeland and Taylor (2004) present a simple theoretical model to clarify the conditions under which these dierent outcomes might arise. 9. Holling (1986) gives the example of cycles of ground re experienced prior to re management in the mixedconifer forests of the Sierra Nevada in western United States. He points out that in several areas, these res occurred with a remarkably consistent interval of seven to eight years, and helped maintain conditions of tree regeneration and nutrient cycling. In addition, these light res killed only the young white r thereby

GLOBALIZATION, LOCAL ECOSYSTEMS, AND THE RURAL POOR introducing and maintaining gaps in forest canopy and, in essence, producing natural re breaks. However, if the undersurface is raised because of increased moisture or eective re control practices, more fuel must accumulate before an average ignition event triggers a re. This results in a longer period before a re but also in a more intense re (p. 302). 10. The recent controversies associated with the diusion of Genetically Modied crops in developing countries illustrate this phenomenon very well. In particular, for the case of Bt cotton in India, Sahai and Rahman (2003) discuss how these crops were allowed to be adopted by farmers without adequate investigation of their long-term eects on the local ecosystem and without appropriate biosafety regulations. 11. Young (2004) provides a good collection of papers on the role of MNCs. 12. While environmental groups often make this argument, Pearce, Adger, Maddison, and Moran (1995) nd only limited empirical evidence to suggest that the eect of international debt on environment is signicant. 13. Altieri (2002) provides a comprehensive survey of these studies.

1417

can provide pest control benets, weed control advantages, reduced wind erosion, and improved water inltration. 15. An important caveat here is that we still understand very little about institutional dynamics to make any denitive predictions. See, for instance, Bardhan (1989) for some reections on this theme. 16. However, it is important to bear in mind that an increase in value of a local natural resource may not always lead to breakdown of existing institutions and resource degradation. For instance, there are a number of cases where specic natural resources became major determinants of tourist revenue as in countries like Ecuador, Bhutan, Maldives, and Seychelles (World Bank, 2002). In these cases, since tourist revenue depended on keeping the local ecosystems healthy, environmental regulation was strengthened in order to protect the resources (Wheeler, 2000). 17. See, for instance, the case studies given in Berkes and Folke (1998) and Gunderson and Holling (2001). 18. Ghai and Vivian (1992) and Chopra (2001) cite several such cases. 19. Niamir-Fuller (1998) gives several examples of revival of these traditional safety net mechanisms in the Sahel region. 20. However, it is encouraging to note the eorts being made in this direction, particularly in the emerging eld of ecological economics.

14. As Altieri (2002) points out, enhanced yields in diverse cropping systems may result from a variety of mechanisms such as more ecient use of resources (light, water, nutrients) or reduced pest damage. Intercropping, which breaks down the monoculture structure,

REFERENCES
Akerlof, G. (1984). An economic theorists book of tales. Cambridge: Cambridge University Press. Altieri, M. A. (2002). Agroecology: The science of natural resource management for poor farmers in marginal environments. Agriculture, Ecosystems and Environment, 93, 124. Bardhan, P. (1989). The new institutional economics and development theory: A brief critical assessment. World Development, 17(9), 13891396. Berkes, F. (1989). Common property resources: Ecology and community-based sustainable development. London: Bellhaven. Berkes, F., & Folke, C. (Eds.) (1998). Linking social and ecological systems: Management practices and social mechanisms for building resilience. Cambridge: Cambridge University Press. Berkes, F., & Folke, C. (2001). Back to the future; Ecosystem dynamics and local knowledge. In L. H. Gunderson, & C. S. Holling (Eds.), Panarchy: Understanding transformations in human and natural systems. Washington: Island Press. Brock, W. A., Maler, K., & Perrings, C. (2001). Resilience and sustainability: The economic analysis of non-linear dynamic systems. In L. H. Gunderson, & C. S. Holling (Eds.), Panarchy: Understanding transformations in human and natural systems. Washington: Island Press. Chichilnisky, G. (1994). Global environment and NorthSouth trade. American Economic Review, 84(4), 851874. Chopra, K. (2001). Social capital and development: The role of formal and informal institutions in a developing country. New Delhi: Institute of Economic Growth. Clark, W. C., & Munn, R. E. (Eds.) (1986). Sustainable development of the biosphere. Cambridge: Cambridge University Press.

1418

WORLD DEVELOPMENT Ostrom, E., Gardner, R., & Walker, J. (1997). Rules, games and common-pool resources. Ann Arbor: University of Michigan Press. Pearce, D., Adger, N., Maddison, D., & Moran, D. (1995). Debt and the environment. Scientic American, 272(6), 5256. Perrings, C. (1995). Ecological resilience in the sustainability of economic development. Economic Appli quee, 48(2), 121142. Pritchard, L., & Sanderson, S. E. (2001). The dynamics of political discourse in seeking sustainability. In L. H. Gunderson, & C. S. Holling (Eds.), Panarchy: Understanding transformations in human and natural systems. Washington: Island Press. Ross, M. (2001). Timber booms and institutional breakdown in southeast Asia. New York: Cambridge University Press. Sahai, S., & Rahman, S. (2003). Performance of Bt cotton. Economic and Political Weekly of India, (26 July). Scheer, M., Westley, F., Brock, W. A., & Holmgren, M. (2001). Dynamic interaction of societies and ecosystems; Linking theories from ecology, economy, and sociology. In L. H. Gunderson, & C. S. Holling (Eds.). Panarchy: Understanding transformations in human and natural systems. Washington: Island Press. Sindzingre, A. (2004). Explaining threshold eects of globalization on poverty: An institutional perspective. Paper presented at the UNU-WIDER workshop on the Impact of globalization on the worlds poor, October 2004, Helsinki. Walker, B. H., Ludwig, D., Holling, C. S., & Peterman, R. M. (1981). Stability of semi-arid savanna grazing systems. Journal of Ecology, 69, 473498. Westley, F., Carpenter, S. R., Brock, W. A., Holling, C. S., & Gunderson, L. H. (2001). Why systems of people and nature are not just social and ecological systems. In L. H. Gunderson, & C. S. Holling (Eds.), Panarchy: Understanding transformations in human and natural systems. Washington: Island Press. Wheeler, D. (2000). Growth policy management and environmental institutions: Implications of the World Bank indicator series. Washington, DC: World Bank Development Research Group World Bank. World Bank (2002). Globalization, growth, and poverty: Building an inclusive world economy. New York: Oxford University Press. Young, S. (Ed.) (2004). Multinationals and public policy. Northampton, MA: Elgar.

Common, M., & Perrings, C. (1992). Towards an ecological economics of sustainability. Ecological Economics, 6, 734. Copeland, B. R., & Taylor, M. S. (2004). Trade, growth, and the environment. Journal of Economic Literature, 42(1), 771. Costanza, R., Low, B. S., Ostrom, E., & Wilson, J. (2001). Institutions, ecosystems, and sustainability. New York: Lewis Publishers. Dasgupta, P. (1993). An inquiry into well-being and destitution. Oxford: Clarendon Press. Eskeland, G. S., & Harrison, A. (2003). Moving to green pastures? Multinationals and the pollution haven hypothesis. Journal of Development Economics, 70(1), 123. Ghai, D., & Vivian, J. M. (Eds.) (1992). Grassroots environmental action: Peoples participation in sustainable development. London: Routledge. Grossman, G. M., & Krueger, A. B. (1993). Environmental impacts of a North American free trade agreement. In P. M. Garber (Ed.), The USMexico free trade agreement. Cambridge, MA: MIT Press. Gunderson, L. H., & Holling, C. S. (Eds.) (2001). Panarchy: Understanding transformations in human and natural systems. Washington: Island Press. Holling, C. S. (1986). The resilience of terrestrial ecosystems: Local surprise and global change. In W. C. Clark, & R. E. Munn (Eds.), Sustainable development of the biosphere. Cambridge: Cambridge University Press. Lane, C. (1992). The Barabaiag pastoralists of Tanzania: Sustainable land use in jeopardy. In D. Ghai, & J. M. Vivian (Eds.), Grassroots environmental action: Peoples participation in sustainable development. London: Routledge. Lopez, R. (1997). Environmental externalities in traditional agriculture and the impact of trade liberalization: The case of Ghana. Journal of Development Economics, 3. Niamir-Fuller, M. (1998). The resilience of pastoral herding in Sahelian Africa. In F. Berkes, & C. Folke (Eds.), Linking social and ecological systems: Management practices and social mechanisms for building resilience. Cambridge: Cambridge University Press. North, D. C. (1989). Institutions and economic growth: A historical introduction. World Development, 17(9), 13191332. Ostrom, E. (1990). Governing the commons: The evolution of institutions for collective action. Cambridge: Cambridge University Press.