globalchange  > 气候变化事实与影响
DOI: doi:10.1038/nclimate2481
论文题名:
Temperature impacts on economic growth warrant stringent mitigation policy
作者: Frances C. Moore
刊名: Nature Climate Change
ISSN: 1758-1059X
EISSN: 1758-7179
出版年: 2015-01-12
卷: Volume:5, 页码:Pages:127;131 (2015)
语种: 英语
英文关键词: Climate-change policy ; Environmental economics ; Climate-change impacts ; Climate sciences
英文摘要:

Integrated assessment models compare the costs of greenhouse gas mitigation with damages from climate change to evaluate the social welfare implications of climate policy proposals and inform optimal emissions reduction trajectories. However, these models have been criticized for lacking a strong empirical basis for their damage functions, which do little to alter assumptions of sustained gross domestic product (GDP) growth, even under extreme temperature scenarios1, 2, 3. We implement empirical estimates of temperature effects on GDP growth rates in the DICE model through two pathways, total factor productivity growth and capital depreciation4, 5. This damage specification, even under optimistic adaptation assumptions, substantially slows GDP growth in poor regions but has more modest effects in rich countries. Optimal climate policy in this model stabilizes global temperature change below 2 °C by eliminating emissions in the near future and implies a social cost of carbon several times larger than previous estimates6. A sensitivity analysis shows that the magnitude of climate change impacts on economic growth, the rate of adaptation, and the dynamic interaction between damages and GDP are three critical uncertainties requiring further research. In particular, optimal mitigation rates are much lower if countries become less sensitive to climate change impacts as they develop, making this a major source of uncertainty and an important subject for future research.

Integrated assessment models (IAMs) have traditionally captured the negative impacts of climate change with a damage function that relates global temperature change to a loss of current economic output. This formulation captures the transient effects of climate on the economy such as lost agricultural output, increased cooling demand, or lower worker productivity due to hotter temperatures7, 8, 9. Factors of production, namely labour and capital, and their total factor productivity (TFP) are not directly impacted, meaning that climate change has no effect, or only a very weak effect, on GDP growth. Two IAMs recently used for the US government social cost of carbon (SCC) estimate, FUND and PAGE, assume that GDP growth is entirely exogenous10, 11. In the DICE model, labour and TFP are specified exogenously and capital formation is determined through endogenous investment decisions5; temperature shocks can therefore alter economic growth through capital stock reductions, but this effect is small and indirect12.

Damages from climate change that directly affect growth rates have the potential to markedly increase the SCC because each temperature shock has a persistent effect that permanently lowers GDP below what it would otherwise be (Supplementary Fig. 1). Continued warming therefore has a compounding effect over time, so that even very small growth effects result in much larger impacts than the traditional damage formulation12. Examples of pathways by which temperature could affect the growth rate of GDP include damage to capital stocks from extreme events, reductions in TFP because of a change in the environment that investments were originally designed for, or slower growth in TFP because of the diversion of resources away from research and development and towards climate threats1. Empirical evidence that these impacts exist is mounting. Two studies have found a reduced-form relationship between temperature shocks and GDP growth4, 13, and other studies have demonstrated plausible pathways including increasing conflict risk14 and changes in labour supply15. Previous work has demonstrated that DICE results are sensitive to the inclusion of growth impacts12, 16, but no previous studies have calibrated these damages using empirically grounded results from the econometrics literature. Given the potentially first-order impacts of these growth effects, understanding their implications for climate policy is of critical importance.

Here we examine alternative formulations of the DICE damage function based on empirical estimates of the impact of inter-annual temperature variability on national economic output and growth rates by Dell and colleagues4. They find large, statistically significant negative effects of hot temperatures on growth rates in poor countries, smaller effects in rich countries, and mixed effects on output (Table 1). To implement these parameters in an IAM, we develop a two-region version of DICE (ref. 17; DICE-2R). We then modify the damage pathway so that warming affects either TFP growth or capital depreciation as per results in ref. 4 (gro-DICE) and investigate sensitivities to the parameters used by Dell et al.4 (Methods). We present results of the TFP pathway here, but the capital pathway gives quantitatively similar results and is discussed further in the Methods and Supplementary Information.

Table 1: Parameters used to calibrate the gro-DICE damage functions, reported in Dell et al. Table 3, column 4 (ref. 4).
Full table

To study the growth effects as presented in Dell et al.4 (DJO in this section) we created a two-region version of DICE (DICE-2R). The rich and poor regions are parameterized on the basis of output-weighted regional values from the 2010 RICE model5, 17 (Supplementary Table 1). DICE-2R chooses mitigation and savings so as to maximize the discounted sum of utility in both regions, weighted by regional Negishi weights28. We also altered DICE by fixing emissions in 2005 and 2010, making 2015 the first year when mitigation is possible. As the parameterization of the rich and poor regions in DICE-2R, although consistent with RICE2010, differs from the DICE-2013R aggregate, DICE-2R does not exactly reproduce the most recent DICE results5. Specifically, the slightly faster TFP growth in DICE-2R means that incomes and emissions are higher in DICE-2R than in DICE-2013R in the second half of the twenty-first century.

We investigate two alternative pathways by which warming could affect economic growth: slowing the growth of TFP or accelerating depreciation of the capital stock. For the first pathway, climate damages impact the growth rate of TFP, reflecting the fact that climate change could affect the productivity of the research sector or existing investments12:

where Aj, t is TFP in region j in time period t, rTFP is the exogenous annual TFP growth rate, T is the global temperature change from pre-industrial, Δt is the model time step, and is the regional growth-rate sensitivity to temperature, calibrated to reproduce the DJO result (Table 1). Calibration is necessary because economic growth is not completely exogenous in DICE but is partly determined by an endogenous capital stock, meaning that reductions in TFP affect economic growth both through lower productivity and through lower capital. Details on the calibration are given in the Supplementary Information. The gro-DICE model also includes transient impacts of temperature on regional output estimated by DJO (β0jTt, Table 1), but this effect is small compared with the growth-rate damages.

The second pathway assumes climate damages fall on the capital depreciation rate. This simulates the impact of climate change on physical infrastructure through more frequent or larger extreme events or on institutional capital through, for example, increased risk of civil conflict14. We calibrate the relationship between temperature change and depreciation rate for the DJO results for values of capital stock, investment, TFP and labour in the reference run for a range of temperatures up to 6 °C (calibration details in Supplementary Information and Supplementary Fig. 9). This gives a concave, quadratic function relating warming and depreciation rate (Supplementary Fig. 10). We find comparable implications for climate policy along both the TFP and depreciation pathways. In reality, both impact pathways (as well as others) are likely to be important in determining climate change damages, but we present them separately here for clarity and because of the lack of empirical studies on their relative roles.

We model adaptation in gro-DICE using an exponential decay curve in which the initial impact of a change in temperature (determined by parameters calibrated to the DJO results) declines over time at the rate of adaptation. We introduce a new variable, the effective temperature, which is the sum of all residual temperature shocks:

where ETt is the effective temperature at time t, Ti is the temperature in year i, and a is the rate of adaptation. For runs with a positive adaptation rate, ETt replaces Tt in the calculation of damages (equation (1)). As there is a very limited empirical basis for the rate of adaptation, we use a value of 10% per year and vary it between 0 and 20% per year in a robustness check. Ten per cent per year is equivalent to a 95% reduction in the impact of a temperature shock after a 30-year adjustment period (Supplementary Fig. 2). The contribution to effective temperature of temperature change before the start of the model time horizon is based on the global surface temperature record since 1850 (ref. 29). The effective temperature rather than absolute temperature is then used to define damages on output and TFP or capital. This formulation means that impacts depend both on the magnitude and the rate of temperature change because faster warming results in larger disequilibrium and therefore higher adjustment costs.

The temperature and resilience mechanisms are implemented such that the growth-rate damage parameters are a function of either temperature or per-capita GDP, respectively. In the temperature mechanism, sensitivity in poor regions remains constant but increases with warming in rich regions, not exceeding the sensitivity observed at present in poor regions (Supplementary Fig. 11). The resilience mechanism causes sensitivity in poor regions to decrease until they reach the per-capita GDP of rich regions today, reducing damages from warming over time as poor regions develop (Supplementary Fig. 12).

The effect of parametric uncertainty in five factors is investigated by independently varying each parameter from its reference value to a high or low value using one-at-a-time sensitivity analysis (Fig. 4). The uncertainties captured and not captured by this approach are discussed more fully in the Supplementary Information.

Corrected online 28 January 2015
In the version of this Letter originally published, in equation (1) and in the explanatory sentence following the equation, jTFP should have read rTFP. In the second line of the equation, jDJOj,t should have read rDJOj,t. These errors have been corrected in the online versions of the Letter.
  1. Pindyck, R. S. Uncertain outcomes and climate change policy. J. Environ. Econ. Manage. 63, 289–303 (2012).
  2. Stern, N. The structure of economic modeling of the potential impacts of climate change: Grafting gross underestimation of risk onto already narrow science models. J. Econ. Lit. 51, 838–859 (2013).
  3. Revesz, R. L. et al. Global warming: Improve economic models of climate change. Nature 508, 173–175 (2014).
  4. Dell, M., Jones, B. F. & Olken, B. A. Temperature shocks and economic growth: Evidence from the last half century. Am. Econ. J. Macroecon. 4, 66–95 (2012).
  5. Nordhaus, W. D. & Sztorc, P. DICE 2013R: Introduction and User’s Manual 1–102 (2013); http://www.econ.yale.edu/˜nordhaus/homepage/documents/DICE_Manual_103113r2.pdf
  6. IAWG, U. Technical support document: Technical update of the social cost of carbon for regulatory impact analysis under executive order 12866. 1–22 (US government, 2013)
  7. Deschênes, O. & Greenstone, M. Climate change, mortality, and adaptation: Evidence from annual fluctuations in weather in the US. Am. Econ. J. Appl. Econ. 3, 152–185 (2011).
  8. Schlenker, W. & Roberts, D. L. Nonlinear temperature effects indicate severe damages to US corn yields under climate change. Proc. Natl Acad. Sci. 106, 15594–15598 (2009).
  9. Dell, M., Jones, B. F. & Olken, B. A. What do we learn from the weather? The new climate-economy literature. J. Econ. Lit. 52, 740–798 (2014).
  10. Hope, C. W. The marginal impact of CO2 from PAGE2002: An integrated assessment model incorporating the IPCC’s five reasons for concern. Integr. Assess. J. 6, 19–56 (2006).
  11. Anthoff, D.<
URL: http://www.nature.com/nclimate/journal/v5/n2/full/nclimate2481.html
Citation statistics:
资源类型: 期刊论文
标识符: http://119.78.100.158/handle/2HF3EXSE/4887
Appears in Collections:气候变化事实与影响
科学计划与规划
气候变化与战略

Files in This Item:
File Name/ File Size Content Type Version Access License
nclimate2481.pdf（516KB）期刊论文作者接受稿开放获取
CC BY-NC-SA
View Download

Recommended Citation:
Frances C. Moore. Temperature impacts on economic growth warrant stringent mitigation policy[J]. Nature Climate Change,2015-01-12,Volume:5:Pages:127;131 (2015).
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[Frances C. Moore]'s Articles
百度学术
Similar articles in Baidu Scholar
[Frances C. Moore]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[Frances C. Moore]‘s Articles
Related Copyright Policies
Null
收藏/分享
文件名： nclimate2481.pdf
格式： Adobe PDF
此文件暂不支持浏览
所有评论 (0)
暂无评论
 

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.