Economics of poverty, environment and natural resource use (chapter 6).
We review many theoretical predictions that link poverty to deforestation and then examine poverty’s net impact empirically using multiple observations of all of Costa Rica after 1960. Countrywide disaggregate (district-level) data facilitate analysis of both poverty’s location and its impact on forest. If the characteristics of the places the poor live are not controlled for, then poverty’s impact is confounded with differences between poorer and less poor areas and we find no significant effect of poverty. Using our data over space and time to control for effects of locations’ differing characteristics, we find that the poorer are on land whose relative quality discourages forest clearing, such that with these controls the poorer areas are cleared more. The latter result suggests that poverty reduction aids the forest. For the poorest areas, this result is weaker but another effect is found: deforestation responds less to productivity, i.e., the poorest have less ability to expand or to reduce given land quality.
Chapter in “Amazonia and Global Change” book (American Geophysical Union, linked to the NASA LBA project)
We examine the evidence on Amazonian road impacts with a strong emphasis on context. Impacts of a new road, on either deforestation or socioeconomic outcomes, depend upon the conditions into which roads are placed. Conditions that matter include the biophysical setting, such as slope, rainfall, and soil quality, plus externally determined socioeconomic factors like national policies, exchange rates, and the global prices of beef and soybeans. Influential conditions also include all prior infrastructural investments and clearing rates. Where development has already arrived, with significant economic activity and clearing, roads may decrease forest less and raise output more than where development is arriving, while in pristine areas, short-run clearing may be lower than immense long-run impacts. Such differences suggest careful consideration of where to invest further in transport.
Journal of Regional Science 2007 volume 47, no. 1, 2007, pp. 109–123
Understanding the impact of road investments on deforestation is part of a complete evaluation of the expansion of infrastructure for development.We find evidence of spatial spillovers from roads in the Brazilian Amazon: deforestation rises in the census tracts that lack roads but are in the same county as and within 100 km of a tract with a new paved or unpaved road. At greater distances from the new roads the evidence is mixed, including negative coefficients of inconsistent significance between 100 and 300 km, and if anything, higher neighbor deforestation at distances over 300 km.
Conservation Biology 2007 volume 21, number 5, 1165–1173
We evaluated the intention, implementation, and impact of Costa Rica’s program of payments for environmental services (PSA), which was established in the late 1990s. Payments are given to private landowners who own land in forest areas in recognition of the ecosystem services their land provides. To characterize the distribution of PSA in Costa Rica, we combined remote sensing with geographic information system databases and then used econometrics to explore the impacts of payments on deforestation. Payments were distributed broadly across ecological and socioeconomic gradients, but the 1997–2000 deforestation rate was not significantly lower in areas that received payments. Other successful Costa Rican conservation policies, including those prior to the PSA program, may explain the current reduction in deforestation rates. The PSA program is a major advance in the global institutionalization of ecosystem investments because few, if any, other countries have such a conservation history and because much can be learned from Costa Rica’s experiences.
Land Use Policy 24 (2007) 600–610
We review claims linking both payments for carbon and poverty to deforestation. We examine these effects empirically for Costa Rica during the late 20th century using an econometric approach that addresses the irreversibilities in deforestation. We find significant effects of the relative returns to forest on deforestation rates. Thus, carbon payments would induce conservation and also carbon sequestration, and if land users were poor could conserve forest while addressing rural poverty. We note that the poor appear to be marginalized in the sense of living where land profitability is lower. Those areas also have more forest. We find that poorer areas may have a higher supply response to payments, but even without this effect poor areas might be included and benefit more due to higher (per capita) forest area. They might be included less due to transactions costs, though. Unless the Clean Development Mechanism of the Kyoto Protocol is modified in its implementation to allow credits from avoided deforestation, such benefits are likely to be limited.
Resources 2007 volume 165:20-22
Even a perfect measure of the ecosystem services provided by each parcel enrolled in a PES program would be insufficient to measure the overall effectiveness of the program. The simple reason is that if a PES program does not lead to an increase in the provision of ecosystem services compared to what would have happened in the absence of the program—that is, the baseline or “counterfactual”—then it has not accomplished anything. Imagine a PES program focused on forest conservation that makes payments to managers of ecologically rich forest land, who have no incentive to clear the land because it is illsuited for logging, agriculture, or urbanization. Payments to these managers would have little impact on deforestation because the risk of clearing was minimal to begin with. In contrast, payments to managers who have incentives to clear their land would be much more likely to have an impact.
Frontiers of Biodiversity Economics, Cambridge University Press
This chapter is structured as follows. Section 2 describes a simple model of interactions in the context of deforestation, based on an equilibrium in beliefs about the neighbours’ actions. Section 3 discusses empirical issues in measurement of interactions and the benefits of using an instrumental variable approach. Data requirements for analysing neighbours’ interactions in deforestation decisions are discussed in section 4. Finally, results for two regions within Costa Rica, as well as discussion of how to obtain the equilibria once the parameters of the model are estimated, are presented in section 5.
Resource and Energy Economics 26 (2004) 237–254
An index of ‘deforestation pressure’ is suggested as useful for reserve planning alongside the currently used information on the species present at candidate sites. For any location, the index value is correlated with threats to habitat and thus also survival probabilities over time for members of species dependent on that habitat. Threats in the absence of reserves are key information for planning new reserves. The index is estimated using a regression approach derived from a dynamic, micro-economic model of land use, with data on observed clearing of forest over space and time as well as biophysical and socioeconomic factors in land returns. Applying an estimated threat (or probability of clearing) function for Costa Rica to locations of interest yields relevant estimates of sites’ deforestation pressure, which are used to evaluate proposed reserves and to suggest other candidate sites.
Biological Conservation 109 (2003) 123–135
The transformation and degradation of tropical forest is thought to be the primary driving force in the loss of biodiversity worldwide. Developing countries are trying to counter act this massive lost of biodiversity by implementing national parks and biological reserves. Costa Rica is no exception to this rule. National development strategies in Costa Rica, since the early 1970s, have involved the creation of several National Parks and Biological Reserves. This has led to monitoring the integrity of and interactions between these protected areas. Key questions include: ‘‘Are these areas’ boundaries respected?’’; ‘‘Do they create a functioning network?’’; and ‘‘Are they effective conservation tools?’’. This paper quantifies deforestation and secondary growth trends within and around protected areas between 1960 and 1997. We find that inside of national parks and biological reserves, deforestation rates were negligible. For areas outside of National Parks and Biological reserves we report that for 1-km buffer zones around such protected areas, there is a net forest gain for the 1987/1997 time period. Thus, it appears that to this point the boundaries of protected areas are respected. However, in the 10-km buffer zones we find significant forest loss for all study periods. This suggests that increasing isolation of protected areas may prevent them from functioning as an effective network.
Journal of Environmental Management 69 (2003) 25–37
Policy enabling tropical forests to approach their potential contribution to global-climate-change mitigation requires forecasts of land use and carbon storage on a large scale over long periods. In this paper, we present an integrated modeling methodology that addresses these needs. We model the dynamics of the human land-use system and of C pools contained in each ecosystem, as well as their interactions. The model is national scale, and is currently applied in a preliminary way to Costa Rica using data spanning a period of over 50 years. It combines an ecological process model, parameterized using field and other data, with an economic model, estimated using historical data to ensure a close link to actual behavior. These two models are linked so that ecological conditions affect land-use choices and vice versa. The integrated model predicts land use and its consequences for C storage for policy scenarios. These predictions can be used to create baselines, reward sequestration, and estimate the value in both environmental and economic terms of including C sequestration in tropical forests as part of the efforts to mitigate global climate change. The model can also be used to assess the benefits from costly activities to increase accuracy and thus reduce errors and their societal costs.
Central America Project, Environment: Conservation and Competitiveness. HIID 2001. Chapter 15.
In this chapter we consider potential gains derived from preventing deforestation, drawing heavily from information from Chapter 14. It uses the same economic model and econometric technique and the same land use/land cover data. It also uses the carbon stock estimates presented there. The key difference is that, instead of using proxies for land-use returns such as ecological characteristics related to higher productivity, we attempt to directly estimate dollar-valued returns. We use these as an independent variable to explain and predict deforestation patterns. This allows us to simulate the potential supply of carbon sequestration in response to dollar-valued returns for certified emissions reductions. Payments for CERs will reduce deforestation by lowering the net return from forest clearing. The loss of the reward for carbon sequestration will partially offset the positive return from agricultural uses. To estimate the effect of such payments on deforestation, and hence CER supply, we need to estimate the response of deforestation to changes in returns to land use. An increase in agricultural returns is empirically equivalent to a reduction in carbon CER payments. Thus, we construct a variable that estimates the potential return of a plot of land if it is cleared. We construct a variable that varies across space (different crop suitability and yields) and time (changes in export prices, technology, and labor costs). We then use this variable in our econometric estimation. The results are used to calculate a supply curve of CERs. These results are illustrative only. They are produced as part of an ongoing effort at estimation (Kerr, Pfaff, Hughes et al. 2000) and are used to show some underlying features of a dynamic supply curve.
Central America Project, Environment: Conservation and Competitiveness. HIID 2001. Chapter 14.
The chapter is structured as follows. First, below, we begin this analysis of the process influencing land changes with a dynamic model of land-use choices. Such models have often been suggested, but crucial features have often been neglected in application. This model generates testable hypotheses regarding factors underlying patterns of land-use changes in tropical areas. The next section describes the data collected for this project and discusses the quality of land-use data. It also outlines the variables used to test the implications of the model. Following that, we present our results and then discuss the linkage from land-use changes to implied carbon sequestration, and the quality of information currently available on carbon sequestration. Finally, we present some conclusions and lessons learned.
Ecological Economics 35 (2000) 203–221
Protecting tropical forests under the Clean Development Mechanism (CDM) could reduce the cost of emissions limitations set in Kyoto. However, while society must soon decide whether or not to use tropical forest-based offsets, evidence regarding tropical carbon sinks is sparse. This paper presents a general method for constructing an integrated model (based on detailed historical, remote sensing and field data) that can produce land-use and carbon baselines, predict carbon sequestration supply to a carbon-offsets market and also help to evaluate optimal market rules. Creating such integrated models requires close collaboration between social and natural scientists. Our project combines varied disciplinary expertise (in economics, ecology and geography) with local knowledge in order to create high-quality, empirically grounded, integrated models for Costa Rica.
The Forest in the South and North in the Context of Global Warming (book by WIDER).
Historical data from the late 19th to the early 20th century are examined for New England. From the attempt to explain the reforestation that occurred, three main land-use claims arise: 1) population clearly does not fully dictate land use (e.g., de- or re-forestation); while population may well have an independent effect on land use, that effect clearly does not dominate all others; 2) factors that affect relative land-use returns, whether “external” to a region or not, clearly do affect land use; two examples are transport costs and productivity of other regions, which affect trade; and finally, 3) long-run analysis must consider shifts even in overall framework, such as from agriculture to migration and industrialization processes involving different economic dynamics. Support for these claims comes from limited historical data alongside relevant theory concerning optimal allocation of land between the four most relevant land uses: agriculture, manufacturing, forest (for timber or as a result of abandonment), and shelter (or, more generally, land uses other than for production). Supporting the “population” claim, previous New England farm expansion flattened out post-1850 and eventual reversed itself, even as population was increasing. Regarding the “returns” claim, the breakpoint in the 1790-1930 series of within-region measures (based on county-level data) of concentration of population is very clearly at about 1830, precisely the era in which the transportation revolution involving railroads, steamships and canals started to have its effect. Concerning the “long-run” claim, given an interest in land use there are grounds for attention to shifts in regional output, such as towards manufacturing from agriculture, as there is evidence that such shifts involved significant changes, in particular concentration of population within particular counties, along rivers and in particular locations along rivers.
Journal of Environmental Economics and Management 1999 volumne 37, pp. 26-43
While previous empirical analysis of deforestation focused on population, this paper builds from a model of land use which suggests many determinants of deforestation in the Brazilian Amazon. I derive a deforestation equation from this model and test a number of those factors using county-level data for the period 1978-1988. The data include a satellite deforestation measure which allows improved within-country analysis. The major empirical finding is the significance of both land characteristics (such as soil quality and vegetation density) and factors affecting transport costs (such as distance to major markets and both own- and neighboring-county roads). Government development projects also appear to affect clearing, although credit infrastructure does not. However, as such policies themselves may be functions of other factors, estimated effects of policies must be interpreted with some caution. Finally, the population density does not have a significant effect on deforestation when many potential determinants are included. However, a quadratic specification reveals a more robust result: the first migrants to a county have greater impact than later immigrants. This implies that the distribution of population affects its impact.