Macro- and Micro-Perspectives on Economic Development and ... · Draft: Do not cite without authors’ permission 1 Macro- and Micro-Perspectives on Economic Development and Biodiversity

Draft: Do not cite without authors’ permission

1

Macro- and Micro-Perspectives on Economic Development and Biodiversity

Dale S. Rothman* Senior Researcher

International Institute for Sustainable Development 161 Portage Avenue East, 6th Floor

Winnipeg, Manitoba R3B 0Y4 Canada

Phone: +1 204-958-7731 Fax: +1 204-958-7710

Email: [email protected]

and

Neha Khanna Associate Professor

Economics and Environmental Studies Binghamton University (LT 1004)

P.O. Box 6000 Binghamton, NY 13902-6000

Phone:+1 607-777-2689 Fax: +1 607-777-2681

Email: [email protected]

Binghamton University Economics Department Working Paper 0801

This paper is forthcoming in a special section of Conservation Biology which will contain the only peer reviewed version of this paper.

* corresponding author Running Head: Macro- and Micro-Perspectives Keywords: Environmental Kuznets Curve; Biodiversity; Poverty; Conservation Word Count: 5748 Acknowledgements B. Zhang (Binghamton University), for able research assistance.


2

Abstract What is the relationship between economic development and biodiversity? Is this

relationship characterized primarily by compatibilities or conflicts, does the nature of the

relationship change as development proceeds? Two distinct dialogues on these issues can be

found, one operating at a macro-scale and the other at a micro-scale. The former, focusing on

the environmental Kuznets curve hypothesis, explores the connection between levels of

economic development and biodiversity at the national and international scale. The latter, part of

the broader dialogue on poverty and the environment, examines the relationships between

poverty and biodiversity, most commonly at a much more local scale. To date the two dialogues

have occurred almost completely independently. Both suffer from a lack of consistent, strong

empirical data, which prevents clear conclusions from being drawn about the nature of the

relationships between economic development and biodiversity at either the micro- or macro-

level, much less across them. Both, however, point to the need to go beyond a narrow focus on

income levels, be it at the level of an individual or a society as a whole, in order to understand

the complexities in these relationships. Further and better designed integrated research is

required to not only improve our understanding of these relationships at both the micro- and

macro-scales, but also across these scales.

I. INTRODUCTION Since it first appeared as a formal concept in the early 1990s, the environmental Kuznets

curve (EKC) hypothesis has engendered significant debate within academic and policy literature.

The hypothesis states that environmental degradation will increase with economic development

up to a point, whereupon additional development will lead to a decline in environmental

degradation. Under this hypothesis, when the measure of environmental degradation is plotted

on the vertical axis against a measure of economic development on the horizontal axis, an


3

inverse-U shaped curve is generated; thus the other common name for this hypothesized

relationship is the inverse-U hypothesis.

The EKC hypothesis (and the studies it has spawned) has received critical examination

from a number of perspectives - its theoretical underpinnings, the presence and/or absence of

empirical support, and the methods of analysis used to test it. The original EKC studies tended

to consider local and regional air and water pollutants, although other indicators, notably

deforestation were also considered. More recently a number of studies have more directly

examined the relationship between economic development and biodiversity and conservation.

These studies consider a variety of measures of biodiversity, e.g. species diversity, species

richness, threats to particular species, and the multivariate National Biodiversity Risk

Assessment Index (NBRAI), and have explored a variety of theoretical explanations for the

observed relationships.

Closely related to, but not necessarily overlapping with this debate, has been a growing

examination of the relationships between poverty and biodiversity. From a scientific

perspective, there is the question of the degree of importance of biodiversity, and the ecosystem

goods and services resulting from it, to the income and livelihoods of the poor. From a policy

perspective, there are questions surrounding the relationship between poverty alleviation and

biodiversity conservation. This literature has taken a more micro-perspective, emphasizing the

complex web of interactions between people and the environment, sometimes referred to more

broadly as the poverty-environment nexus.

This paper will build a bridge between these presently disconnected discussions on the

relationships between economic development, biodiversity, and conservation. It will review the

literature on the EKC hypothesis, with an emphasis on those studies with a particular focus on


4

biodiversity. This will be followed by a review of the poverty, biodiversity, and conservation

literature. The connections between the two dialogues will then be discussed followed by a set

of summary remarks.

II. A MACRO-PERSPECTIVE: APPLYING THE EKC HYPOTHESIS TO BIODIVERSITY

II.A. What is the EKC Hypothesis? The EKC hypothesis refers to the relationship between income and environmental quality

as an economy develops, where economic development is defined and measured by increasing

per capita income. According to this hypothesis, as incomes increase, environmental

degradation, i.e. losses in environmental quality, also increases, but only up to a point,

whereupon additional increases in income lead to declines in environmental degradation, i.e.

improvements in environmental quality. When the measure of environmental degradation is

plotted on the vertical axis against income on the horizontal axis, this generates an inverse-U

shape; thus the other common name for this hypothesized relationship is the inverse-U

hypothesis. However, the hypothesis itself does not say anything about the symmetry of the

rising and falling sections of the curve. Even if environmental degradation was rapid and the

eventual improvements in environmental quality occurred at a significantly slower pace, it would

be consistent with the hypothesis. Nor does the EKC speak to the issue of sustainability per se.

The idea behind the EKC hypothesis starts with the assumption that negative changes in

environmental quality, e.g. pollution or the loss of biodiversity, are by-products, i.e. involuntary

and unplanned consequences, of consumption and production activities intended to improve

human well-being. As the scale of these activities increases as the economy grows, the level of

environmental degradation will also increase, unless: 1) the negative changes in environmental

quality associated with specific activities decreases per unit of activity or 2) the mix of activities


5

changes such that the share of activities associated with fewer negative changes increases, and 3)

the degree to which these changes occurs is large enough to compensate for the overall increase

in levels of activity. The first set of changes is generally associated with technological advances;

the latter with changes in the structure of the economy. A further important supposition here is

that the degradation in environmental quality is reversible and that the knowledge exists on how

to affect the reversal.

Based upon these basic ideas, a number of reasons have been proposed to support the

EKC hypothesis. First, many technological advances, particularly those that lead to more

efficient use of resources, do indeed reduce the negative changes in environmental quality per

unit of activity. As economies develop there are more financial resources available to invest in

these technologies. Second, as economies develop, there is also a general shift in emphasis from

agriculture and resource-based sectors to industry and then to services. Many types of

environmental degradation, e.g. local air pollution, are most associated with industrial processes.

Thus, as the share of industry increases so does the level of environmental degradation; as this

share falls with the increase in services, the level of environmental degradation follows. The

third, and perhaps most important reason, is the assumption that a consumer’s welfare is

determined to some extent by the environmental quality that she faces or, more simply, that she

must ‘care’ about the environmental quality, because of a link between environmental quality

and her state of well-being. This link could be fairly direct, for example, through the relationship

between pollution and health, or indirect, e.g. through the relationship between pollution and

agricultural yields. As incomes rise, the ability of a consumer to act upon this desire for

environmental quality by undertaking efforts to improve environmental quality increases, be this

through working to improve environmental quality (for example, exerting political and social


6

pressure on firms to reduce the effect of their production activities on the environment,

pressuring governments to bring a greater amount of land area under forestation and

conservation) or by changing the nature of her consumption (example, using more renewable

energy or biodegradable products) (Andreoni & Levinson 2001).

Assuming that the structure of an economy, i.e. the relative size of economic sectors is

ultimately driven by consumer demand, the driving force underlying the an EKC is thus the

interaction between technology and consumer preferences (Lieb 2002; Plassmann & Khanna

2006). From a theoretical point of view, it can be shown that suitable preferences can always

lead to an EKC, but assuming that some environmental degradation is an inevitable by-product

of consumption and production activities, then there is no technology that yields an EKC for all

types of preferences (Plassmann & Khanna 2006). In other words, presuming that environmental

degradation is reversible, if the consumer cares sufficiently about the environmental quality that

she faces, we will always get an EKC, regardless of the nature of the technology that links

consumption and production to pollution.

II.B. The EKC for Biodiversity – Theoretical Considerations In order for an EKC to be possible, the conditions identified in the previous section must

be satisfied. Using biodiversity as the indicator of environmental quality, as opposed to e.g. air

quality, presents some particular theoretical problems. This begins with how biodiversity is

defined and measured. For example, if it is measured in terms of the land area under

conservation, it is possible to both decrease and increase biodiversity. But if, for example, we

measure biodiversity in terms of the number of animal or plant species, an increase may not be

possible, at least not on a timescale of interest. This basic asymmetry is further complicated by

an incomplete understanding of biodiversity itself and its relationship to human welfare.


7

It is not a stretch to argue that human activities cause negative changes in biodiversity.

For example, increasing industrialization and urbanization can result in the destruction of native

habitats. It is also plausible to argue that technologies, e.g. improvements in crop yields, and

shifting consumer preferences, e.g. taking photographs of birds rather than shooting them, can

slow down the rate of a negative biodiversity change. For a review of the negative impacts of

economic activity on biodiversity, see Czech et al. (2000) and the Millennium Ecosystem

Assessment (2003; 2005). But is the reverse also true? In other words, is it possible to increase

biodiversity? This depends very much on how biodiversity is measured, as noted above.

Furthermore, the knowledge of how to obtain an increase in biodiversity may not be available.

A second question relates to the links between changes in biodiversity and human

welfare. What are these and do they operate on a timescale that is relevant to human decision

making? These links have been explored in a number of recent studies, most notably the

Millennium Ecosystem Assessment (2003; 2005). Because the complex ecological interactions

in the biosphere are largely unknown, the pathways through which biodiversity affects human

welfare are often unclear. Still, it is reasonable to argue that changes in biodiversity can indeed

affect human welfare. For example, a loss in species diversity can impact human health through

changes in soil and water quality, reduction in agricultural yields, loss in recreational

possibilities, loss in medicinal and pharmaceutical resources, etc, and each of these changes may

occur with sufficient speed to impact human welfare in a relatively short timeframe. This may

be especially true if there are threshold and/or ecological multiplier effects. Increases in

biodiversity, even where possible, are likely to take place on an ecological timeframe, rather than

on a timeframe within which much human decision-making takes place, making the link between

increases in biodiversity and increases in human welfare even more tenuous. The only plausible


8

link seems to be via an insurance policy type mechanism. Because of the possibility of

catastrophic losses, it may be rational for the consumer to allocate resources so as to increase

biodiversity even if the possibility of a catastrophic loss is a very low probability outcome.

II.C. The EKC for Biodiversity – The Empirical Evidence Table 1 summarizes the currently published empirical literature on the EKC for

biodiversity. As is typical in the broader EKC literature, most authors estimate a regression

model with some measure of biodiversity as the dependent variable and per capita income (or a

higher order polynomial) as the independent variable, in addition to some additional covariates

that might be relevant. The most common additional covariates are population density or some

related measure such as fraction of urban population that is likely to be correlated with

population density, economic aggregates such as measures of trade intensity or the share of

agriculture, and measures of civil and political liberties and/or political institutions. The

empirical evidence is assessed in terms of the statistical significance of the estimated coefficients

on the income terms and the location of the ‘turning point,’ that is, the level of income at which

biodiversity-income relationship changes from decreasing biodiversity with increases in income

to increasing biodiversity with increases in income.

Given the ambiguity about how to quantify changes in biodiversity, it is not surprising

that different authors use very different measures of biodiversity. In fact, the literature may be

classified into two broad categories based on the measure of biodiversity used. The first

category includes those studies where the authors argue that their measures of biodiversity are

such that the loss in biodiversity is irreversible and cannot be recovered so that the relationship

between their measure of biodiversity and per capita income is expected to be monotonic.

Asafu-Adjaye (2003) and Dietz and Adger (2003) measure biodiversity in terms of species count

(species density, number of threatened species, percentage change in the number of known


9

species within a taxa, and species richness). Both sets of authors argue that speciation occurs so

slowly that neither study allows for the possibility of an inverted U-shaped relationship in their

empirical analyses. Rather, they look for evidence than the rate at which biodiversity is lost

slows down with increasing income. Dietz and Adger (2003, p.27) also consider the relationship

of income and protected areas, but they explicitly assume that “countries are expected to

supplement the area of protection as development proceeds.” Thus, they also treat this as a

monotonic relationship.

The second group of studies follows the traditional EKC hypothesis and allows for the

possibility of a non-monotonic relationship between biodiversity and per capita income. All

other studies reviewed in Table 1 fall into this category and employ a measure of biodiversity

can be expected to decline as well as increase within a reasonable timeframe. These measures

range from very broad aggregates such as the National Biodiversity Risk Assessment Index

(Mozumder et al. 2006) to more specific measures such as the percentage of threatened bird or

animal species (McPherson & Nieswiadomy 2005).

The evidence regarding the existence of an EKC for biodiversity is inconclusive at best.

Even though a majority of the studies report statistically significant coefficients on income and

its squared value, the results are not always robust to sample or model specification. Naidoo and

Adamowicz (2001)), who estimate separate regression models for several different taxanomical

groups, find that the coefficients on the income terms are not statistically significant in all cases.

Also, for fish species the statistical significance of the coefficients on the income terms gets

eroded when they use a smaller sample, excluding the data on Mexico. Furthermore, they find

that when they change the model specification to include the squared terms for non-economic

variables on the right hand side of their regression equation, the results change yielding a


10

positive monotonic relationship in the case of fish and invertebrate species, and the anticipated

inverted U shape for mammals. Similarly, Koop and Tole (1999) find that their results change as

they use different estimation procedures – ordinary least squares, panel data methods (fixed and

random effects estimation) and a random coefficients specification which allows for cross-

country hetrogeniety in the estimated income coefficients.

Even where the coefficients obtained are statistically significant and the authors report a

turning point (TP) estimate, it is not clear whether the TP is real or simply a result of having fit

the data to a polynomial of income. If the TP lies well outside the data range, it is clear the TP is

not real. However, even if it lies within the sample data range but is close to the upper or lower

tail it is likely that the relationship is monotonic, either steadily increasing or steadily decreasing

depending on whether the coefficients on income and its square are positive or negative,

respectively, rather than non-monotonic. Because the turning point is simply a point estimate, it

is does not provide much information about the nature of the biodiversity-income relationship by

itself. Just as with any point estimate, we need to construct the confidence bands in order to

determine whether the biodiversity-income relationship is truly non-monotonic (see Plassmann

& Khanna 2006). Unfortunately, none of the studies reviewed here provide sufficient

information to make a conclusive assessment about the precision of their turning point estimates.

Based on the limited information on the mean and standard deviation for per capita income

reported in the papers, it is our guess that in all cases the turning point is not real.

Yet, it is important that we do not conclude that evidence does not support the existence

of an EKC. All we can say is that the income level at which the TP occurs, presuming that it

does occur, has not been reached by a sufficiently large number of countries in the data sets that

have been used to empirically test this hypothesis. This means that an improvement in


11

biodiversity is unlikely to occur at currently observed income levels; it does not mean that it will

not occur at some higher income level that might be achieved in the future.

One key insight from this literature is the important role of non-income variables in

explaining biodiversity change. It is clear that changes in biodiversity are associated with other

socio-economic and political factors such as the structure of the economy, population density and

urbanization, and the degree of civil and political liberties. This finding is very much in line

with the wider literature on the EKC hypothesis. The EKC hypothesis itself, however, is silent

on the role of many of these factors other than the structural composition of the economy.

III. A MICRO-PERSPECTIVE: POVERTY AND BIODIVERSITY

III.A. What are the Linkages Between Poverty and Biodiversity? Spurred on, in part, by the political imperatives of the United Nations’ Convention on

Biological Diversity (CBD) and the Millennium Development Goals (MDGs), there has been an

increasing effort to understand better the relationships between poverty alleviation and

biodiversity conservation. This has, in turn, forced greater attention to be paid to the

fundamental linkages between biodiversity and the income and livelihoods of the poor,

particularly in rural areas of developing countries. The importance of these issues has been

embodied in institutions both to increase our knowledge, e.g. the Poverty and Conservation

Learning Group (http://www.povertyandconservation.info/) and to act upon this knowledge, e.g.

the United Nations’ Equator Initiative (http://www.undp.org/equatorinitiative/) and Millennium

Villages Project (http://www.unmillenniumproject.org/mv/).

A number of authors attempt to summarize the linkages between poverty and biodiversity

(Adams et al. 2004; Fisher 2004; Millennium Ecosystem Assessment 2003; Roe & Elliott 2005;

Scherr 2003; UK Department for International Development 2002; World Resources Institute


12

2000; World Resources Institute et al. 2007). Roe & Elliott (2005) differentiate three categories

of linkages and associate a number of hypotheses with each of these:

• Biodiversity and Poor People – how poor people affect and are affected by the availability or loss of biodiversity

o There is a geographical overlap between biodiversity and poverty o Poor people depend on biodiversity o Poor people are responsible for biodiversity loss

• Conservation and Poor People – the impact that conservation activities have on poor people and the role that poor people play in conservation activities

o Conservation activities hurt poor people o Poor people can undermine conservation

• Biodiversity and Poverty Reduction – the role that biodiversity plays in poverty reduction efforts and the impact that poverty reduction activities can have on biodiversity

o Biodiversity is irrelevant to poverty reduction o Poverty reduction activities can cause biodiversity loss

Before turning to the empirical evidence supporting or rejecting these hypotheses, it is

useful to consider some of their implications. Here we will discuss just a few of these, both

based on the notion of poverty traps.

Assume that poor people are disproportionately dependent upon biodiversity, but also

disproportionately responsible for its loss. This sets up a classic dilemma - can the poor improve

their wellbeing faster than they destroy the resource base upon which they depend? If so, then

we have the basis for the ‘grow first, clean up later’ philosophy that underpins the EKC. If the

reverse is true, then we would expect to see downward spirals, where deepening poverty and

biodiversity loss feed upon each other. In between, there may exist so called ‘environmental

poverty traps’, in which individuals work ever harder just to stay in place (see (Barrett &

Swallow 2006; Carter et al. 2007; Durning 1989; Perrings 1989; Scherr 2000) for discussions of

poverty traps and downward spirals).

If we further assume that poverty reduction activities cause biodiversity losses, then,

depending upon their relative effects on poverty and biodiversity, they may push the system from


13

one state to the other, or simply accelerate the existing process. A similar sort of dilemma can be

associated with conservation activities. If conservation activities hurt poor people and poor

people undermine conservation, it is not hard to imagine a vicious spiral leading to failed

conservation and increased poverty. This has lead to a particularly vigorous debate over the

issue of ‘fortress conservation’ and displacement from protected areas. By somehow combining

poverty reduction and conservation activities, the stage may be set for a virtuous cycle of

enhanced conservation and reduced poverty. Still, opinions differ over whether there are

inherent synergies or tradeoffs between efforts to alleviate poverty and to conserve biodiversity

(Agrawal & Redford 2006). (For a recent introduction to the debate on displacement from

protected areas, see the special section in Conservation and Society volume 4, number 3, 2006;

for the more general debate on poverty alleviation and conservation, see the special sections in

Oryx, volume 38, number 2, 2004 and WorldWatch January/February 2005).

III.B. What do the Empirical Analyses Show? Numerous studies have attempted to test the hypotheses described above. The following

are just a few of the questions that have appeared in the titles of recent journal articles and

reports:

• Is “local support” Necessary for Sustainable Protected Areas? (Brockington 2002) • Can Protected Areas Contribute to Poverty Reduction? (Scherl et al. 2004) • Conservation in Africa: But for Whom? (Githiru 2007) • Can Biodiversity Conservation be Reconciled with Development? (Inogwabini 2007) • Will Alleviating Poverty Solve the Bushmeat Crisis? (Robinson & Bennett 2002) • Who Bears the Burden of Our Environmental Efforts? (Robalino 2007) • Poverty Alleviation and Tropical Forests – What Scope for Synergies? (Wunder 2001)

Roe and Elliot (2005) and Agrawal and Redford (2006) provide overviews, albeit incomplete,

and meta-analyses of much of this literature. Given the ongoing debates noted above, it is not

surprising that they find conflicting evidence in the literature on most of the questions noted


14

above and the underlying hypotheses about the links between poverty and biodiversity. In

almost every case, researchers find that income measures alone are not adequate to explain the

relationships between poverty and biodiversity. Poor people may be more dependent upon

biodiversity, particularly as a form of insurance, but this is hard to quantify and is strongly

influenced by other factors such as access to common property resources (Baland & Francois

2005; de Sherbinin et al. 2007; Sjaasted et al. 2005). The question of access and tenure rights is

also fundamental to understanding the relationship between conservation activities and poor

people (Agrawal & Redford 2006; Brockington et al. 2006; Redford & Richter 1999; Scherl et

al. 2004), including the role of agriculture (Lee & Barrett 2001, McNeely, 2002 #90). This is

reflected in the variety of categories of protected areas (Upton et al. forthcoming). Poverty can

contribute to biodiversity loss, but it is by no means the only or primary factor (Czech et al.

2000; Weber 2006). Particular concern has also been raised about the lack of attention to

biodiversity in efforts to alleviate poverty (Sanderson 2005; Sanderson & Redford 2003).

More troublesome is that much of the evidence that does exist is not necessarily useful

for answering these questions.

A general lack of coherent data, information, knowledge and informed debate underlies the recurrent disconnect and/or discord between those working on poverty reduction and those working on biodiversity conservation (as well as those observing conservation – poverty dynamics) (Roe & Elliott 2005, p. 12)

34 of the 37 identified studies share two common features: a focus on processes and outcomes in a single case and single time period, and a drastic simplification of the complex concepts of poverty and biodiversity. In addition, the cases we examine are relatively inattentive to the relationships between observed outcomes and the contextual features of programmatic interventions. As a result of these shared features, the mass of scholarly work on the subject does not permit systematic and context-sensitive generalizations about the conditions under which it may be possible to achieve poverty alleviation and biodiversity conservation simultaneously (Agrawal & Redford 2006, p. ii)


15

What is clear, though, is the importance of non-economic factors. Perhaps the best way to

summarize the conclusions presented by these and most other studies addressing these questions

is to simply state that context, particularly governance matters. Needless to say, many studies

point to the need for further and better designed integrated research (Agrawal & Redford 2006),

and to the need for separating the empirical (what is) from the moral (what ought to be)

(Brockington (2002).

IV. LINKING THE DIALOGUES For the most part, the macro- and micro- dialogues discussed above have taken place in

isolation from each other. The studies focusing on poverty and biodiversity may mention the

Environmental Kuznets’ Curve (see for example (Roe & Elliott 2005)), but usually only in

passing. The studies testing the Environmental Kuznets’ Curve hypothesis primarily consider

the relationship between economic development and biodiversity in the aggregate, with little

specific consideration of distributional aspects within countries and poverty.

This leaves open two related questions – do the macro and micro approaches tell a

consistent story and can the macro story can be seen to emerge from the micro story. With

respect to the first of these, the premises underlying the two perspectives are closely related.

Both share the notion that there may be tradeoffs between the alleviation of poverty, at least by

increasing incomes, and conserving biodiversity. The macro-perspective has less to say about

the link from biodiversity to poverty, but there is the sense that the individual poor, or at least

less developed nations, may be more directly dependent upon biodiversity.

The lack of good consistent empirical data in both dialogues limits our ability to go

further to say definitively whether the evidence supports these and other hypothesis associated

with the different perspectives. Even still, there is a certain amount of consensus on the


16

complexities in the relationships between poverty and biodiversity. A narrow focus on income

levels, be it at the level of an individual or a society as a whole, will miss many of the most

important factors involved in these relationships.

Upton et al (forthcoming) provide one of the few examples of an effort to bridge the

micro- and macro-perspectives. Their analysis focuses on protected areas and finds that, while

‘the distribution, size and category of protected areas vary between counties of different income

groups. . . . there is no clear relationship between the area of land in protected areas and national

levels of poverty or wealth.’ (pp. 4-5) Furthermore, ‘(i)n no income-based category of countries

do the economic and social effects (whether positive or negative) of conservation in the form of

protected areas have a substantial effect on aggregate, national poverty statistics’ (p.5). In their

discussion, they go further to consider why there may be a disconnect between the relationships

between poverty and biodiversity at local and national scales. Among the factors that they note

are: the relatively small percentage of poor people in any single country affected by protected

areas; the proportion of urban versus rural poor; and the aggregate impacts of migration within

countries.

V. SUMMARY In this paper, we have explored two presently distinct discussions on the relationships

between economic development, biodiversity, and conservation. At the macro level is the debate

over the Environmental Kuznets’s Curve hypothesis; at the micro level are the various threads of

the poverty, biodiversity, and conservation dialogue. A common question addressed by both is

whether economic development, particularly for the poor, is inherently associated with

biodiversity loss or whether there are inherent synergies between poverty alleviation and

biodiversity conservation.


17

Although our analysis confirmed the relative absence of overlap in the two dialogues, we

did find some commonalities. On a fairly negative note, both suffer from a lack of consistent,

strong empirical data, which prevents clear conclusions from being drawn about the nature of the

relationships between economic development and biodiversity at either the micro- or macro-

level, much less across them. Both, however, do point to the need to go beyond a narrow focus

on income levels, be it at the level of an individual or a society as a whole, in order to understand

the complexities in these relationships.

As further research and policy is pursued around the issues of economic development,

biodiversity, and conservation, we feel that the two perspectives have much to learn from each

other. This will improve our understanding not only within, but also across scales.

References

Adams, W. M., R. Aveling, D. Brockington, B. Dickson, J. Elliott, J. Hutton, D. Roe, B. Vira,

and W. Wolmer. 2004. Biodiversity Conservation and the Eradication of Poverty.

Science 306:1146-1149.

Agrawal, A., and K. Redford 2006. Poverty, Development, and Biodiversity Conservation:

Shooting in the Dark? Wildlife Conservation Society, New York.

Andreoni, J., and A. Levinson. 2001. The simple analytics of the environmental Kuznets curve.

Journal of Public Economics 80:269-286.

Asafu-Adjaye, J. 2003. Biodiversity Loss and Economic Growth: A Cross-Country Analysis.

Contemporary Economic Policy 21:173-185.

Baland, J.-M., and P. Francois. 2005. Commons as Insurance and the Welfare Impact of

Privatization. Journal of Public Economics 89:211-231.


18

Barrett, C. B., and B. M. Swallow. 2006. Fractal Poverty Traps. World Development 34:1-15.

Bhattarai, M., and M. Hammig. 2001. Institutions and the Environmental Kuznets Curve for

Deforestation: A Crosscountry Analysis for Latin America, Africa and Asia. World

Development 29:995-1010.

Brockington, D. 2002. Injustice and Conservation: Is Local Support Necessary for Sustainable

Protected Areas? Policy Matters 12:22-30.

Brockington, D., J. Igoe, and K. Schmidt-Soltau. 2006. Conservation, Human Rights, and

Poverty Reduction. Conservation Biology 20:250-252.

Carter, M. R., P. D. Little, T. Mogues, and W. Negatu. 2007. Poverty Traps and Natural

Disasters in Ethiopia and Honduras. World Development 35:835-856.

Czech, B., P. R. Krausman, and P. K. Devers. 2000. Economic Associations among Causes of

Species Endagerment in the United States. BioScience 50:593-601.

de Sherbinin, A., L. VanWey, K. McSweeney, R. Aggarwal, A. Barbieri, S. Henry, L. Hunter,

W. Twine, and R. Walker 2007. Rural Household Micro-Demographics, Livelihoods and

the Environment. Institute of Beharioral Science, University of Colorado at Boulder,

Boulder, Colorado.

Dietz, S., and W. N. Adger. 2003. Economic Growth, Biodiversity Loss and Conservation Effort.

Journal of Environmental Management 68:23-35.

Durning, A. B. 1989. Poverty and the Environment: Reversing the Downward Spiral.

Worldwatch Paper. Worldwatch Institute, Washington, DC.

Ehrhardt-Martinez, K., E. M. Crenshaw, and J. C. Jenkins. 2002. Deforestation and the

Environmental Kuznets Curve: A Cross-National Investigation of Intervening

Mechanisms. Social Science Quarterly 83:226-243.


19

Fisher, M. 2004. Links Between Poverty Alleviation and Conservation: The Debate Continues.

Oryx 38:119-120.

Githiru, M. 2007. Conservation in Africa: But for Whom? Oryx 41:119-120.

Inogwabini, B.-I. 2007. Can Biodiversity Conservation be Reconciled with Development? Oryx

41:136-137.

Koop, G., and L. Tole. 1999. Is There an Environmental Kuznets Curve for Deforestation?

Journal of Development Economics 58:231-244.

Lee, D. R., and C. B. Barrett, editors. 2001. Tradeoffs or Synergies? Agricultural Intensification,

Economic Development and the Environment. CAB Publishing, New York.

Lieb, C. M. 2002. The Environmental Kuznets Curve and Satiation: A Simple Static Model.

Environment and Development Economics 7:429-448.

McPherson, M. A., and M. L. Nieswiadomy. 2005. Environmental Kuznets Curve: Threatened

Species and Spatial Effects. Ecological Economics 55:395-407.

Millennium Ecosystem Assessment 2003. People and Ecosystems: A Framework for Assessment

and Action. Island Press, Washington, DC.

Millennium Ecosystem Assessment 2005. Ecosystems and Human Well-being: Synthesis. Island

Press, Washington, DC.

Mozumder, P., R. P. Berrens, and A. K. Bohara. 2006. Is There an Environmental Kuznets Curve

for the Risk of Biodiversity Loss? The Journal of Developing Areas 39:175-190.

Naidoo, R., and W. L. Adamowicz. 2001. Effects of Economic Prosperity on Numbers of

Threatened Species. Conservation Biology 15:1021-1029.

Perrings, C. 1989. An Optimal Path to Extinction? Poverty and Resource Degradation in the

Open Agrarian Economy. Journal of Development Economics 30:1-24.


20

Plassmann, F., and N. Khanna. 2006. Preferences, Technology, and the Environment:

Understanding the Environmental Kuznets Curve Hypothesis. American Journal of

Agricultural Economics 88:632-643.

Redford, K. H., and B. D. Richter. 1999. Conservation of Biodiversity in a World of Use.

Conservation Biology 13:1246-1256.

Robalino, J. A. 2007. Land Conservation Policies and Income Distribution: Who Bears the

Burden of Our Environmental Efforts? Environment and Development Economics

12:521-533.

Robinson, J. G., and E. L. Bennett. 2002. Will Alleviating Poverty Solve the Bushmeat Crisis?

Oryx 36:332.

Roe, D., and J. Elliott. 2005. Poverty-Conservation Linkages: A Conceptual Framework. Page

12. Poverty and Conservation Learning Group, London.

Sanderson, S. 2005. Poverty and Conservation: The New Century's "Peasant Question?" World


Sanderson, S. E., and K. H. Redford. 2003. Contested Relationships Between Biodiversity

Conservation and Poverty Alleviation. Oryx 37:389-390.

Scherl, L. M., A. Wilson, R. Wild, J. Blockhus, P. Franks, J. A. McNeely, and T. O. McShane.

2004. Can Protected Areas Contibute to Poverty Reduction? Opportunities and

Limitations. Page 60. International Union for Conservation of Nature and Natural

Resources, Gland, Switzerland.

Scherr, S. J. 2000. A Downward Spiral? Research Evidence on the Relationship between Poverty

and Natural Resource Degradation. Food Policy 25:479-498.


21

Scherr, S. J. 2003. Hunger, Poverty and Biodiversity in Developing Countries. Mexico Action

Summit, Mexico City, Mexico.

Sjaasted, E., A. Angelsen, P. Vedeld, and J. Bojö. 2005. What is Environmental Income?

Ecological Economics 55:37-46.

UK Department for International Development 2002. Wildlife and Poverty Study. UK

Department for International Development, London.

Upton, C., R. Ladle, D. Hulme, T. Jiang, D. Brockington, and W. M. Adams. forthcoming. Are

Poverty and Protected Area Establishment Linked at a National Scale? Oryx.

Weber, J. 2006. The "Guilty" are not the Poor! Policy Matters 14:27-36.

World Resources Institute 2000. World Resources 2000-2001: People and Ecosystems - The

Fraying Web of Life. World Resources Institute, Washington, DC.

World Resources Institute, M. o. E. N. R. Department of Resource Surveys and Rmote Sensing,

Kenya, M. o. P. a. N. D. Central Bureau of Statistics, Kenya, and I. L. R. Institute 2007.

Nature's Benefits in Kenya, An Atlas of Ecosystems and Human Well-Being. World

Resources Institute, Washington, DC and Nairobi.

Wunder, S. 2001. Poverty Alleviation and Tropical Forests - What Scope for Synergies? World



Table 1: Summary of Empirical Literature on the Existence of an EKC for Biodiversity Author Dependent Variable Independent Variables Data Methodology Results & Authors’ Conclusions*

(Mozumder et al. 2006)

• National Biodiversity Risk Assessment Index – standard, adjusted, and upgraded

• GDP per capita • Trade as % of GDP • Number of Tourists • Gross Foreign Direct

Investment as % of GDP • Per-capita Foreign Aid • Fertilizer Consumption • Urban Population (% of total) • Average Circulation of Daily

Newspapers per 1000 people

Cross-country data: 62 countries for NABRAI; 104 countries for Adjusted and Upgraded NABRAI

OLS regression The estimated coefficients on the GDP related variables are not significant for all of the biodiversity risk indices. No evidence in support of an EKC relationship

(McPherson & Nieswiadomy 2005)

• % of threatened bird and mammal species (as defined by the IUCN) in each country

• GDP per capita • % threatened birds (2000) • % threatened mammals

(2000) • % endemic mammals • % endemic birds • Island dummy • persons per square kilometer • Political rights and civil

liberties • Antigovernment

demonstrations, per year • Civil law dummy • Common law dummy • Communist law dummy • Muslim law dummy

Cross-country Data: 113 countries

Spatial lag model (also known as the mixed regressive-spatial autoregressive model) Four models are presented for each species. Model 1 includes all of the variables. Since there is some multicollinearity among political liberties, demonstration, and the legal system dummies, they also present models that include only some of these variables to see if some of these variables are significant.

For mammals: linear and squared income variables are significant. The turning point is at approximately $12,000. For birds: linear and squared income variables are significant. The turning point is in the $12,000 to $14,000 range. The cubed income term was not significant in any of the models. For both mammals and birds, results indicate a possible EKC curve.

(Asafu-Adjaye 2003)

• Number of known mammal species/10000 km

• Number of known bird species/10000 km

• Number of known higher plant species/10000 km

• % of bird and mammal species threatened with extinction

• GDP per capita; • Agricultural value added as

percentage of GDP; • Economic freedom index; • Black market premium on

exchange rates; • Population density; • Urban population growth; • % of land developed for

agriculture and other uses; • % of protected land area; • Climate dummy

Cross-section data for 100 countries, including 50 low-income, 25 middle-income and 25 high-income countries.

OLS regression No GDP square in the model.

GDP has a significant negative effect on species density for mammals and birds, but not for higher plants. It appears to have adverse effect on indicator 5. The proxy for the composition of economic output is highly significant for indicator 5. Although economic growth has an adverse effect on biodiversity, the type or composition of this growth is also significant for biodiversity loss.


• Average annual % change in the number of known mammal species 1989-1999

(Dietz & Adger 2003)

• Predicted species richness in any year compared to the reference year 1970

• National parks and protected areas as a percentage of national land territory

• Percentage of CITES reports submitted relative to those expected in 1999

• GDP per capita; • Population change; • Population density; • Linear time trend; • Forest area; • Democracy - the sum of

political rights and civil liberty indices;

Panel data for first two indicators; cross section data for third indicator

Hyperbolic and linear in income equations for indicators (1) and (2) Linear in income equation for indicator (3) OLS, fixed effects and random effects

Income terms are significant, but the signs are opposite of what we would expect. Economic development is related to the area of state protected land but it is not the overriding determinant of the rate of designation

(Naidoo & Adamowicz 2001)

• Number of threatened species (plants, mammals, birds, amphibians, reptiles, fish and invertebrates)

• GDP per capita; • Number of species; • Number of endemic species; • Country area; • Percent area domesticated; • Percent area protected; • Percent original forest left

Global, cross-sectional data: 157 countries, 1996

Classified threatened species into several taxonomic groups and analyzed per-capita income relationships separately for each group. Negative-binomial regression for count data Three regression analyses: all countries, outlier removed, nonlinear effects of noneconomic variables.

When all countries were included in the analysis, income variables were significant predictors in all taxonomic groups except for mammals. The number of threatened plants increased linearly with log(GNP). The number of threatened amphibians, reptiles, fishes, and invertebrates all had a negative coefficient on log (GNP) and a positive quadratic term, indicating a general U-shaped relationship. The number of threatened bird species exhibited an inverted-U relationship with increasing log(GNP).

(Ehrhardt-Martinez et al. 2002)

• Average annual rate of deforestation

• GDP per capita; • Forest stock 1980; • Data reliability; • % urban; • Population pressure; • Rural-urban migration 1970-

1990; • Labor in services 1980; • Secondary education 1980;

All developing countries with available forest cover estimates

OLS. The analysis proceeds in three steps: Structural modernization models, political modernization models and international political economy models. Structural modernization models:

Structural modernization models: the operative modernization variable is urbanization rather than GDP. Urbanization is the more central dynamic producing the EKC. Political modernization models: the urban polynomial remains significant throughout this series, indicating the stability of the


• Protected areas 1991; • Government scope 1980; • Democracy 1980; • Debt level/GDP 1980; • Change in debt 1980-1990; • Forest exports/GDP 1980; • Forest export/Global forest

exports; • Forest import/Global forest

imports; • Imports/Exports 1980; • Semiperiphery dummy

Forest stock, data reliability, log GDP per capita and squared, % urban, population pressure, R/U migration, Population*Migration, Labor in services, services*urban; Political modernization models: Forest stock, data reliability, % urban, urban squared, secondary education, protected areas, government scope, democracy, scope*democracy; International political economy models: Forest stock, data reliability, % urban, urban squared, Debt level/GDP, change in debt, Forest exports/GDP, Forest export/Global forest exports, Forest import/Global forest imports, Imports/Exports

structural EKC. International political economy models: dependency and world systems theory have little net impact on deforestation once the ecological modernization theory is appropriately specified.

(Bhattarai & Hammig 2001)

• Annual % change in forest and woodlands area

• GDP per capita; • Time; • Political institutions; • Black market premium on

foreign exchange; • External debt as % of GDP; • Change in Cereal yield; • Population growth; • Rural population density

Panel data for 66 countries in Latin America, Africa, and Asia, 1972-1991

Fixed effects model estimated by weighted least squares

The results confirm the existence of an EKC for Latin America and Africa. Quadratic income term is negative and cubed term is positive. The first turning point is around $6,000 for Latin America; $1,300 for Africa. EKC for Asia follows a different pattern from LA and Africa (opposite signs).

(Koop & Tole 1999)

• Changes in forest cover (deforestation)

• GDP per capita; • Change in GDP; • Population density; • Change in population

Panel data Begin with a simple regression model, then consider fixed and random effects models, and a random coefficients model.

A significant environmental Kuznets curve exists in the simple regression, but it is gradually lost when the specification is freed up.

*The final column summarizes the conclusions made by the authors of each study. These do not always represent our conclusions.

Documents

Macro- and Micro-Perspectives on Economic Development and ... · Draft: Do not cite without authors’ permission 1 Macro- and Micro-Perspectives on Economic Development and Biodiversity