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.

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.