Climate Change Economics

We analyse the effects of rainfall and temperature on yields of paddy and millets (pearl millet and sorghum) in India for the period 1966–1999, at the district level. Unlike other studies, we control for fertiliser use and irrigation. We find that paddy (India's leading food crop) is sensitive to the climate variables but also to fertiliser use and irrigation. Millets are less affected by climate variables although sorghum shows some sensitivity to temperature. Our results have important implications for how India's agriculture will adapt to climate change.

Agriculture is a source of greenhouse gas emissions that could be included as offsets in a cap-and-trade system. This paper describes offset rules that could be applied to reduced nitrogen fertilizer application on wheat, a source of nitrous oxide emissions. Unlike other papers that have examined offset rules, we do not assume that farmers' business-as-usual nitrogen emissions can be perfectly predicted. We construct a structural model of wheat production and use this, along with Agricultural Resource Management Survey data, to model or estimate participation in the offset market, offset supply conditional on participation, and business-as-usual emissions. We find that roughly two-thirds of the supplied offsets would be non-additional at an allowance price of $15/tonne CO₂-e. Under assumptions about the social damages from greenhouse gas emissions, we find that allowing nitrogen offsets would marginally pass a standard benefit-cost test.

The release of the Global Change Assessment Model (GCAM) version 3.0 represents a major revision in the treatment of agriculture and land-use activities in a model of long-term, global human and physical Earth systems. GCAM 3.0 incorporates greater spatial and temporal resolution compared to GCAM 2.0. In this paper, we document the methods embodied in the new release, describe the motivation for the changes, compare GCAM 3.0 methods to those of other long-term, global agriculture-economy models and apply GCAM 3.0 to explore the impact of changes in agricultural crop yields on global land use and terrestrial carbon. In the absence of continued crop yield improvements throughout the century, not only are cumulative carbon emissions a major source of CO₂ emissions to the atmosphere, but bioenergy production remains trivial — land is needed for food. In contrast, the high crop yield improvement scenario cuts terrestrial carbon emissions dramatically and facilitates both food and energy production.

The Social Costs of Carbon (SCC) equals the marginal welfare loss associated with one unit of emitted CO₂, divided by the marginal welfare gain associated with one unit of consumption. In stochastic assessments, both the nominator and denominator can depend on uncertain parameters; specifically they depend on the (implicit) scaling of the welfare function with the parameters. I discuss some pitfalls when calculating the expected value or the certainty equivalent of the SCC, and show that a mistaken procedure easily leads to very high or very low estimates for the SCC. I use the paper by Newbold et al. (2013) as an illustration.

In this reply to the comment by Gerlagh, we confirm an error in our estimate of the certainty-equivalent social cost of carbon (SCC) reported in Newbold et al. (2013), and we discuss the underlying conceptual difficulties that arise in conducting a social welfare analysis when preferences are heterogeneous or uncertain. The certainty-equivalent SCC depends crucially on the reference level of per capita consumption used to normalize marginal utility across possible preference parameters, and our estimate of the certainty-equivalent SCC was driven largely by an arbitrary choice of measurement units. All other results from our rapid assessment model are based on the deterministic SCC or its simulated probability distribution, which does not depend on the reference level of per capita consumption.