We develop a forward-looking version of the MIT Emissions Prediction and Policy Analysis (EPPA) model, and apply it to examine the economic implications of proposals in the U.S. Congress to limit greenhouse gas (GHG) emissions. We find that the abatement path and CO2-equivalent (CO2-e) price in the forward-looking model are quite similar to that of the recursive model, implying that the simulation of banking behavior in the recursive model by forcing the CO2-e price to rise at the discount rate approximates fairly well the banking result obtained with the forward-looking model. We find, however, that shocks in consumption path are smoothed out in the forward-looking model and that the lifetime welfare cost of GHG policy is lower than in the recursive model, results we would expect to find given that the forward-looking model can fully optimize over time. The forward-looking model allows us to explore issues for which it is uniquely well-suited, including revenue-recycling, early action crediting, and the role of a technology backstop. We find (1) capital tax recycling to be more welfare-cost reducing than labor tax recycling because of its long term effect on economic growth, (2) potentially substantial incentives for early action credits relative to emission levels in years after a policy is announced but before it is implemented that, however, when spread over the full horizon of the policy do not have a substantial effect on lifetime welfare cost or the CO2-e price, and (3) strong effects on estimates of near-term welfare costs depending on exactly how a backstop technology is represented, indicating the problematic aspects of focusing on short-term welfare costs in a forward-looking model unless there is some confidence that the backstop technology is realistically represented.

The atmospheric composition, temperature and sea level implications out to 2300 of new reference and cost-optimized stabilization emissions scenarios produced using three different Integrated Assessment (IA) models are described and assessed. Stabilization is defined in terms of radiative forcing targets for the sum of gases potentially controlled under the Kyoto Protocol. For the most stringent stabilization case (“Level 1” with CO2 concentration stabilizing at about 450 ppm), peak CO2 emissions occur close to today, implying (in the absence of a substantial CO2 concentration overshoot) a need for immediate CO2 emissions abatement if we wish to stabilize at this level. In the extended reference case, CO2 stabilizes at about 1,000 ppm in 2200—but even to achieve this target requires large and rapid CO2 emissions reductions over the twenty-second century. Future temperature changes for the Level 1 stabilization case differ noticeably between the IA models even when a common set of climate model parameters is used (largely a result of different assumptions for non-Kyoto gases). For the Level 1 stabilization case, there is a probability of approximately 50% that warming from pre-industrial times will be less than (or more than) 2?C. For one of the IA models, warming in the Level 1 case is actually greater out to 2040 than in the reference case due to the effect of decreasing SO2 emissions that occur as a side effect of the policy-driven reduction in CO2 emissions. This effect is less noticeable for the other stabilization cases, but still leads to policies having virtually no effect on global-mean temperatures out to around 2060. Sea level rise uncertainties are very large. For example, for the Level 1 stabilization case, increases range from 8 to 120 cm for changes over 2000 to 2300.

© 2009 Springer

To aid climate policy decisions, accurate quantitative descriptions of the uncertainty in climate outcomes under various possible policies are needed. Here, we apply an earth systems model to describe the uncertainty in climate projections under two different policy scenarios. This study illustrates an internally consistent uncertainty analysis of one climate assessment modeling framework, propagating uncertainties in both economic and climate components, and constraining climate parameter uncertainties based on observation.We find that in the absence of greenhouse gas emissions restrictions, there is a one in forty chance that global mean surface temperature change will exceed 4.9C by the year 2100. A policy case with aggressive emissions reductions over time lowers the temperature change to a one in forty chance of exceeding 3.2C, thus reducing but not eliminating the chance of substantial warming.

The debate over a policy response to global climate change has been and continues to be deadlocked between 1) the view that the impacts of climate change are too uncertain and so any policy response should be delayed until we learn more, and 2) the view that we cannot wait to resolve the uncertainty because climate change is irreversible so we must take precautionary measures now. The objective of this dissertation is to sort out the role of waiting for better information in choosing an appropriate level of emissions abatement activities today under uncertainty. In this dissertation, we construct two-period sequential decision models to represent the choice of a level of emissions abatement over the next decade and another choice for the remainder of this century, both empirical models based on a climate model of intermediate complexity, and analytical dynamic programming models. Using the analytical models, we will show that for learning to have an influence on the decision before the learning occurs, an interaction must be present between strategies in the two decision periods. We define an "interaction" as the dependence of the marginal cost or marginal damage of the future decision on today's decision. When an interaction is present and is uncertain, the ability to learn will introduce a bias in the optimal first period strategy, relative to the optimal strategy if the uncertainty would never be reduced. In general, the bias from learning can be either in the direction of higher or lower emissions, depending on the sign of the interaction and the probability distribution over damage losses relative to abatement costs.
(cont.) We demonstrate using the empirical climate decision models that the difference between optimal emissions abatement today with and without learning is insignificant. The reason is that the IGSM, like most other climate assessment models, omit many of the most important interactions between emissions today and marginal costs or damages in the future. We show that by representing possible interactions, such as induced innovation from policy constraint or the effect of emissions growth on ocean circulation, that learning will have an influence on today's decision, often in the direction of lower emissions if we expect to learn. In general, the "wait-to- learn" is not necessarily a valid argument for delaying a climate policy that constrains emissions.