The Role of Substitution in Understanding the Costs of Climate Change Policy
Prepared for the Pew Center on Global Climate Change
Dale W. Jorgenson, Harvard University
Richard J. Goettle, Northeastern University
Peter J. Wilcoxen, University of Texas at Austin
Mun Sing Ho, Harvard University
Press Release 
Eileen Claussen, President, Pew Center on Global Climate Change
The U.S. economy has proven both resilient and adaptive over the past century. From the "bust" of the Great Depression to the current "boom" associated with information technology, the economy's ability to adapt stems largely from the substitution possibilities within it — that is, how businesses and households alter their behavior when a major economic change occurs.
Reducing greenhouse gases could alter future economic conditions, largely through increased energy prices. While these changes could be significant, the economy is rapidly becoming even more flexible and responsive as technology changes the way things are invented, produced, and distributed. Accordingly, the damages to the economy might be less. Yet, many economic models used in predicting the future costs of climate change policies do not adequately capture the economy's full range of substitution possibilities. A recent Pew Center report entitled, An Introduction to the Economics of Climate Change Policy, identified substitution assumptions as one of five key factors having the largest influence on modeling results. The other factors are: how baseline greenhouse gas (GHG) projections are measured; what policy regime is considered; how technological progress is represented; and whether GHG reduction benefits are included.
This analysis by Dale Jorgenson, Richard Goettle, Peter Wilcoxen and Mun Sing Ho explores the role of substitution in adapting to economic change. It begins with what is considered a "flexible" model (a top-down, computable general equilibrium model of the U.S. economy) and then constrains the flexibility parameters within this model to observe its new results. In essence, the authors use the same model to behave both "flexibly" and "inflexibly" in order to observe the effect of this pivotal assumption on model outcomes.
The most striking conclusion of this work is that the failure to depict the full range of historically-observed substitution possibilities (as many economic models do) can lead to as much as a doubling of the estimated costs of a climate change policy, an overestimate that is wholly attributable to this one pivotal assumption. This overestimation may be even more pronounced since the economy appears more flexible today than in the post-war period when these observations were made. Another interesting finding is that v a rying the flexibility households have in choosing to work more or fewer hours can be as important in predicting carbon prices and economic outcomes as the assumptions about flexibility in all of production. In summary, economic models of climate change must represent the full range of flexibility that is achievable or risk significant errors in estimating economic benefits and costs.
This paper would not have been possible without the comments and support from several individuals. The Pew Center and authors would like to thank Larry Goulder, Jeffrey Frankel, and Hadi Dowlatabadi for their thoughtful comments on early drafts of this report. Special thanks are due to Ev Ehrlich for serving as a consultant on this project, and to Judi Greenwald for her editorial assistance.
The U.S. economy's reaction both to climate change itself and to the policies designed to avoid climate change depends largely upon the abilities of consumers and producers to adapt to these changes and move forw a rd under new conditions. In turn, these abilities depend on the ease with which consumers and producers can alter their purchasing behavior without sacrificing welfare, income, and production. This ease is reflected largely in the economy's "substitution possibilities" — the options available to con-sumers and producers to change what they buy and sell in response to changes in the prices of particular goods and services. If the cost of economic substitution is low, and the range of substitution possibilities is wide, then mitigation costs — the damages to welfare, income and production — are likely to be low and the burden on the economy is likely to be small. If the cost of substitution is high, and range of substitutability is narrow, then mitigation costs are likely to be high. The purpose of this paper is to examine the economy-wide impacts of reduced substitution opportunities when the economy must adjust to a constraint on carbon emissions.
This analysis uses an economic model that, compared to other models, depicts a relatively complete set of substitution possibilities for consumers and producers. Simulation results from the model portray the economy's response to an emissions reduction schedule that is implemented through a system of tradable emissions permits. The first model simulation used substitution possibilities that were estimated from historical data. Next, the authors systematically replaced key parameters (i.e., coefficients or multipliers of selected mathematical relationships embedded in the model) in a manner that drastically reduced the substitution possibilities of producers and consumers. Each of these simulations defined a different world or economy. The authors then simulated each economy's reaction to proportionally identical emissions constraints. In this manner, the model produced measures of the economic responsiveness both with and without flexibility and the analysis quantified the benefits and costs of substitution.
Three areas of substitution are most important to the overall economic reaction to climate change. These are:
The principal conclusions emerging from this analysis are:
1. When allow able substitutions reflect the observed behavior of the past , constraining carbon emissions to around 70 percent of their projected base-case levels costs the economy about a one and one-quarter percent loss in real Gross Domestic Product (GDP) and a one-tenth of one percent loss in economic welfare. For perspective, at current levels, this loss in GDP corresponds to an annual loss in income of $430 per person living in the United States and the welfare loss is equivalent to a tax, payable today, of $3,175 per person.
2. Constraining carbon emissions is generally more costly when substitutability in consumption or production is restricted. Thus, flexibility within the economy significantly reduces the adverse impacts of climate change and climate change policies. Real GDP losses are slightly larger when consumption is less flexible, and are doubled when production is less flexible. Failing to account for the full range of substitution possibilities in consumption and production will lead to overestimation of the negative effects of climate change policy.
3. Just as "rigidity" magnifies economic costs, it can also magnify economic benefits under certain circumstances. For example, inflexibility in consumption or production is beneficial to economic performance when: (a) climate change policies lead to additional tax revenues, and (b) the tax policy for reusing these additional revenues is economically advantageous.
In fact, the benefit is magnified the more inflexibility is introduced (assuming a and b hold).
4. Differences among models' treatment of the substitutability between consumption and leisure are likely to be every bit as important in predicting emissions permit prices and economic outcomes as are the models' underlying details of technology, consumption, or production. The more inflexible households are with respect to their consumption-leisure tradeoff, the lower the costs of reducing emissions. Contrary to what occurs when substitution is constrained in production or consumption, rigidity in this instance appears beneficial. However, this rigidity can also prove harmful. The combination of an emissions constraint, inflexibility in consumption and production, and more favorable tax treatment leads to economic benefits (point 3 above). Add inflexibility in consumption and leisure, and the combination leads to economic costs. Rigidities in household choices between consumption and leisure substantially limit the observed economic outcomes from climate change policy: either the adverse impacts are smaller or the potential benefits never materialize.
This analysis is important not only because of its results, but also because it explores this topic in a detailed and systematic manner within a single methodology. It is among but a few efforts to fundamentally change the character of a model in developing a sensitivity analysis. The numerous and well-documented outcomes of other policy experiments have informed the policy process. There now are fewer surprises when a particular policy design is subjected to the scrutiny of a broad range of models. But the quest for understanding does not end here. A model's outcomes depend on interactions among the various components that govern its behavior, and thus analysts need to identify and examine these components in both isolation and combination. The intent of this exercise is to increase understanding of the nature and magnitude of the benefits and costs of substitution by exploring the key features of one particular model and the economy it can portray. The hope is that this exercise makes a modest contribution to the formulation of environmental and economic policies that are beneficially robust over the broadest possible range of economic circumstances.