Modeling Environment-Improving Technological Innovations under Uncertainty

Alexander Golub, Anil Markandya

• Publisher: Routledge
• Year: 2008
• ISBN:
  9781134041190
• Edition: 1
• Topic: Commerce
• Subtopic: Commerce Général

The issues of technology and uncertainty are very much at the heart of the policy debate of how much to control greenhouse gas emissions. The costs of doing so are present and high while the benefits are very much in the future and, most importantly, they are highly uncertain. Whilst there is broad consensus on the key elements of climate change science and agreement that near-term actions are needed to prevent dangerous anthropogenic interference with the climate system, there is little agreement on the costs and benefits of climate policy. The book looks at different ways of reconciling the needs for sustainability and equity with the costs of action now.

Presenting a compendium of methodologies for evaluating the economic impact of technological innovation upon climate-change policy, this book describes mathematical models and their predictions. The goal is to provide a practitioner's guide for doing the science of economics and climate change. Because the assumptions motivating different problems in the economics of climate change have different complexities, a number of models are presented with varying levels of difficulty: reduced-form and structural, partial- and general-equilibrium, closed-form and computational. A unifying theme of these models is the incorporation of a number of price and quantity instruments and an analysis of their respective efficacies. This book presents models that contain structural uncertainty, i.e., uncertainty that economic agents respond to via their risk attitudes. The novelty of this book is to relate the effects of risk and risk attitudes to environment-improving technological innovation.

1 Cost and benefits of climate policy under uncertainty

Alexander Golub, Elena Strukova and
James S. Wang

There is a broad consensus on the key elements of climate change science and agreement that near-term actions are needed to prevent dangerous anthropogenic interference with the climate system. However, there is little agreement on the costs and benefits of climate policy. Any policy will result in an irreversible but environment-improving investment in alternative technologies; this change will generate immediately realized costs but significantly delayed benefits. Hence, a critical element in policy selection is the inherent uncertainty in the climate and economy that can be expected over time.

Climate policy, obviously, generates benefits expressed in terms of avoided damage and savings on adaptation costs, but, at the same time, climate policy generates mitigation costs. All parameters of the climate puzzle mentioned above are uncertain. Nevertheless, decisions on climate policy should be made regardless of uncertainties, but realizing the risks of sunk costs or irreversible climate change. Insufficient climate policy may create significant barriers for sustainable economic growth in the future; at the same time, introduction of excessively tight emissions target may impose unreasonable burden on the economy and interfere with short-term economic growth. In this regard new technology has a major role to prevent climate change. Deploying new technologies that cut long-term mitigation costs will bring alternative energy, transportation, etc.

Climate policy induces these changes, but may create imbalance in the allocation of limited resources available for R&D and capital investment. So first we start by selecting an environmental target under the shadow of two major uncertainties: 1) the uncertain benefits of climate policy; and 2) the uncertain mitigation costs needed to meet policy targets.

1 Conventional methodology and uncertainties

An application of conventional methodology that focuses on expected value approach (also known in the literature as “expected utility”) may not be the best way to understand the benefits of climate policy, which imposes a burden on the economy now and has direct positive effects on the economy only in the distant future. Expected value or even expected utility approaches average out various climate outcomes; therefore they may not be an appropriate tool for accurate analysis of irreversible processes with outcomes described by log-normal or β-distributions. In the context of irreversibility, flexibility has an economic value that cannot be captured by the expected value approach. The fat tail distribution of damage attributed to global climate change suggests: there is a good chance that flexibility in terms of future climate policy could be more valuable than flexibility in terms of irreversible investment into sunk cost needed to ensure flexibility of climate policy.

Despite abundant literature on the economics of climate change, it is hard to disagree with William Nordhaus that the key questions about climate change policy, namely “how much, how fast and how closely,” remain open (Nordhaus, 2007). On the one hand, publications like the Stern Review (Stern, 2006) and the Intergovernmental Panel on Climate Change Fourth Assessment Report (IPCC AR4, 2007) urge immediate action to drastically curb carbon emissions based on newly available results in the field of climate science. On the other hand, applications of conventional economic analyses (see for example Nordhaus, 2007 and 2007a) suggest that such urgent and sharp greenhouse gas (GHG) reductions do not pay back.

As Nordhaus states: “The [Stern] Review’s radical revision of the economics of climate change does not arise from any new economics, science, or modeling. Rather, it depends decisively on the assumption of a near-zero time discount rate combined with specific utility function” (Nordhaus, 2007, p. 701). The discount rate of 0.1 or 0.15 percent considered in the Stern Review is simply too low. This attempt to “correct” the conventional net present value (NPV)-based expected value approach makes the Review vulnerable to critics, and diminishes its important results. In this chapter we look beyond traditional NPV and offer a different tool for integrated assessment of climate policy that may provide a common ground for resolving the dispute between Stern’s and Nordhaus’s proponents.

The IPCC AR4 (IPCC, 2007) adds fuel to the debates on climate policy. Immediate action to curb carbon emissions appear even more urgent in the context of new and increasingly more negative information regarding human-induced climate change. While a number of scientists and politicians call for mitigation policy that will limit global warming within 2°C above the pre-industrial level (see Stern, 2006), a temperature target that should prevent major irreversible changes in the climatic system, some economists express concern that such an ambitious environmental goal could be very expensive. For example, the “optimal” emissions target computed by DICE is well above the 2°C (above pre-industrial era) stabilization pathway (see Nordhaus, 2007). In order to meet the 2°C target, emissions must start declining not later than 2035. In contrast, the “optimal” target allows emissions growth until the beginning of the next century. Indeed this “optimal” target results in a greater temperature increase (more then 2.5°C by the end of this century and around 3.5°C by 2200). Selecting a particular emissions trajectory may result in irreversible climatic change. According to IPCC AR4, continuous growth of GHG emissions after around 2020 may preclude temperature stabilization within 2°C limits. Even temporarily overshooting a particular threshold may lead to irreversible changes in ecosystems, well described in AR4. In this chapter we are focusing on the irreversible character of climate policy and economic valuation of irreversible decisions under uncertainties (including uncertain climate sensitivity, economic damage function, mitigation costs, etc.).

The specifics of climate policy selection can be described, to some extent, as a choice between different alternatives with little future flexibility to change that choice unless current targets are sufficiently strict. For example, if during the next 20–30 years world GHG emissions grow according to the projected business-as-usual trajectory, there will be little chance of keeping the temperature rise within 2°C above the pre-industrial era. However, regulators can impose GHG emission limits that prevent an irreversible temperature increase. Regulators can select an interim emission target that can later be revised dependent upon the growing knowledge of the climate response to anthropogenic interference as well as the corresponding response of the socio-economic system to climate change. Let us say, a major revision of climate policy will be conducted in 2050. At present, regulators should select an interim policy that accounts for the irreversibility of climate change and maintains flexibility in the 2050 decision, rather than an interim policy that reduces flexibility in 2050. In other words, regulators can maintain their future GHG-policy options only if an interim GHG emission target that gives a reasonable likelihood of keeping long-term temperature increases within 2°C is selected.

Thus, the cost–benefit analysis based on the application of net present value methodology would suggest rejecting a climate policy sufficient to prevent risks that will dangerously interfere with the climatic system. The net present value methodology is based on estimations of expected values in combination with discounting. Any attempts to address these shortcomings by applying an artificially low discount rate leads to even more confusion. The Stern review was heavily criticized by economists for using an extremely low discount rate of 0.15 percent. In our view, the problem is not a particular discount rate, but an entire concept of rational expectation applied to an irreversible decision making process characterized by uncertainties. The expected value approach neglects the most important benefit of current climate policy: the preservation of flexibility for the future, when new knowledge regarding the climate system and new mitigation technologies will become available. As Dixit and Pindyck (1995) have noted, “The NPV rule is easy, but it makes the false assumption that the investment is … reversible …” They go on, “An action to create an option should be valued more highly than a naive NPV approach would suggest.” In other words, flexibility has an economic value. We demonstrate how this value can be derived, and the implications of this for the choice of an optimal climate policy.

In the literature there are several attempts to correct the expected value methodology. A full literature review is beyond the scope of this chapter—however, some relevant work must be mentioned. As has been noted in the literature, the conventional approach does not account for the presence of thresholds in the concentration–response function, and the risk of global catastrophic climate events that, though characterized by a low probability of occurrence, would lead to significant economic damage. To address this issue Mastrandrea and Schneider (2004) proposed a probabilistic approach to modeling the uncertainties of climate change and determining an optimal policy with regard to these uncertainties and the risk of reaching dangerous anthropogenic interferences (DAI). They used the DICE model with an explicit introduction of uncertainties and then applied Monte Carlo simulation. They focused on three uncertain parameters: climate sensitivity, climate damage and discount rate. As a result, Mastrandrea and Schneider computed probabilities of outcomes that exceed different DAIs and linked these probabilities with particular levels of carbon tax. They demonstrate that when considering the probabilities of exceeding dangerous thresholds, the estimated carbon tax is significantly higher than that suggested by the conventional approach.

Most agents typically strive to eliminate risk and are often willing to pay a price to do so. Martin Weitzman (2007) studied the influence of deep structural uncertainties related to climate sensitivity, and the damage response to climate impacts. Weitzman proposes new theoretical principles that apply under strict relative risk aversion and potentially unlimited exposure described by a fat tailed distribution. He demonstrates that conventional expected utility theory is not applicable in the economic analysis of catastrophes. To address this, Weitzman proposes a “tail-slimming weight-loss program”.¹ In this case, climate policy is a hedge against the risk that greenhouse gas emissions will lead to catastrophe. The purchase of insurance to hedge risks related to home, health, and life is a useful analogy. Hedging risk, therefore, is the major benefit of climate policy.

So how does one decide how much to spend to hedge a risk? All approaches encompass the notion of how probable the risk is and the cost of damages if it occurs. The insurance industry uses an actuarial approach. Economists tend to use models that run sensitivities and come up with optimal answers at different levels of risk. Financial portfolios look at the worst things that have happened in the past and calculate the loss if they occur today (referred to as value-at-risk). All approaches narrow the probabilities and costs to reach one answer—or a series of answers, depending on the level of risk remaining after the hedge (analogous to picking a deductible amount for an insurance policy).

Mitigation costs are also uncertain, but, in contrast to negative climate outcomes, uncertain costs can be described by a right truncated distribution. Back-stop technology may be expensive, but there is a numbe...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• List of figures
• List of tables
• Acknowledgments
• Introduction
• 1 Cost and benefits of climate policy under uncertainty
• 2 Trade-offs between expectations and uncertainties: applying real options methodology to climate policy analysis
• 3 Learning about climate change and its implications for near-term policy
• 4 Structural uncertainty in the DICE model
• 5 Abatement cost uncertainty and policy instrument selection under a stringent climate policy
• 6 The effects of climate policy on the energy—technology mix: an integrated CVaR and real options approach
• 7 Risk-averse firms and new technologies
• 8 The evolution of technological complexity: an agent-based simulation model of the global energy system
• 9 Does the Kyoto Protocol cost too much and create unbreakable barriers for economic growth?
• 10 Improving the contribution of economic models in evaluating energy and climate change mitigation policies
• 11 The economic benefits of an energy-efficiency and onsite renewable energy strategy in meeting growing electricity needs in Texas
• 12 Low-cost offsets and incentives for new technologies