Paul Krugman on Bill Nordhaus and Climate Change: Insights not Answers

by Fran Sussman

Paul Krugman’s review of Bill Nordhaus’s recent book, The Climate Casino: Risk, Uncertainty, and Economics for a Warming World, is thoughtful and informative.  But it unintentionally raises some difficult questions about the role of benefit-cost analysis in assessing the consequences of climate policy, and whether focusing on this type of analysis draws resources and attention away from other, more important topics in economics.

Benefit-cost analysis of global climate change is fundamentally either an academic exercise or one intended to shape public opinion.  The Climate Casino is a little of both.

The Climate Casino is an argument for adopting moderately aggressive policies to address climate change. At the argument’s core are the results of Integrated Assessment Models, or IAMs, including  Nordhaus’s DICE model. [DICE stands for Dynamic Integrated model of Climate and the Economy.] An IAM, in turn, is a framework that integrates models of  the physical climate system, natural systems, and human processes and the economy,  as depicted in this figure from Nordhaus’s guide to DICE.

Integrated assessment and benefit-cost analysis

IAMs—such as DICE—are used to develop benefit-cost assessments of alternative levels of action (or inaction) on climate change. The principle of benefit-cost analysis is simple: If we expend resources on reducing greenhouse gas emissions, we will both slow the rate of climate change and reduce the damages of climate change.  The benefit-cost assessment balances the costs of reducing emissions against the effects that reducing emissions has on the rate of climate change and, therefore, on the resulting damages. The optimal rate of climate change (i.e., the optimal level of mitigation) is reached when the marginal cost of further reducing emissions just equals the marginal reduction in damages.  If we reduce emissions more than that, the costs of mitigation will mount.  If we reduce too little and allow the climate to change more, the damages will mount.

Both Krugman and Nordhaus tacitly accept that benefit-cost analysis, as embodied in IAMs like DICE, provides the “rational analysis” that we should use to guide decisions about global climate change—specifically, that we should use to answer to the question, “What is the optimal rate of climate change?”  This is, arguably, a leap of faith.

A tool for rational decision making?

An ideal tool for rational decision making should  possess a few key characteristics.  It should be based on good information—preferably information that represents a solid forecast about the future.  It should yield a conclusion that is desirable even in the event that the “solid information” turns out to be wrong. It should incorporate information on all the factors that are important to the decision, and weight them in a way that reflects their  relative importance to the decision-maker.  It should deal with uncertainty in a satisfying way. Finally, it should be capable of adapting to new information and updating the decision as needed.

Consider, for example, a personal decision about how much and where to save.  In making a decision I weigh my best guess about the returns from different types of investments, how long I expect to earn a salary (and at what level), what I think the earnings of my dependents will be, and how long I expect to live.  My decision uses information derived from  my own experience, reflects my preferences (about risk and about tradeoffs between my consumption and that of my dependents),  and incorporates my beliefs (hopefully based on some evidence) about the long term stability of the stock market,  my life expectancy, and even how sure I am about my assessment of everything else.  While there is certainly uncertainty about the future, I have relatively good information about each of these pieces of the puzzle—enough that I can have confidence in my ability to make a rational decision about the future.

Unfortunately, the benefit-cost analysis embodied in IAMs  does  not have the characteristics we would ask of a tool for rational decision making where global climate change is concerned.

Gaps in information

One  problem is missing information. We have, for example,  lots of gaps in our estimates of the economic damages of climate change—and thus the benefits of reducing the rate of change. Very  little has been done to estimate the costs of loss of life and illnesses, or the value of losses or changes to unmanaged natural ecosystems (although many researchers believe these to be large), or the damages associated with extreme events, such as hurricanes, flooding, or drought.  We know little about how to evaluate effects such as the loss of culture or livelihood, feelings of insecurity, or the damages to mental health that arise after disasters—and still less how to place a monetary value on those effects.

An unconvincing framework

More fundamentally, however, IAMs  cannot accommodate situations  in which there is considerable scientific uncertainty, where there is a possibility of catastrophic effects, and where preferences are not well-defined.  Climate change is the poster child for an environmental problem that challenges the fundamental principles of benefit-cost analysis.

For example, the potential for tipping points—in climate or in the responses of natural systems to climatic changes—is often cited as a reason for action that goes beyond the results of a benefit-cost analysis.  Differences in impacts on developing countries with much lower incomes per capita, compared with their more developed counterparts, raise questions of how to address equity effects, and “whose preferences” to use to weight costs and benefits borne by different countries. Much discussion also centers around how the welfare of future vs. current generations should be treated.

The question of how to address these difficult aspects of climate change in a benefit cost analysis is vigorously debated by economists, and unlikely to be resolved satisfactorily any time soon (see, for example, see the debate on “fat tails” in the Summer 2011 issue of the Review  of Environmental Economics and Policy.

Out of date assumptions and data

Even if we are willing to ignore the limitations imposed by the framework of benefit-cost analysis and by gaps in our understanding of key relationships and data underlying the models, we cannot ignore the problem that the assumptions—and hence the results—of IAMs are almost always out of sync with the latest research.

The differences between the assumptions in the IAMs and the latest scientific findings can be significant, with newer research critically altering  our understanding of the likely future emissions path, the relationship between that path and current international policies, the effects of emissions on future climate, and the damages likely to be connected with projected climatic changes.  Such collective understanding  is updated and expanded far more rapidly than IAMs can take advantage of.  Consequently, there is a lag—sometimes of many years—between developments in the research community and when the new  information appears in the IAMs.

For example, the agricultural sector in DICE is based on the findings reported in the Fourth Assessment Report released by the United Nation Intergovernmental Panel on Climate Change (the so-called AR4) in 2007.  In AR4, the impacts on agriculture, globally, were found to be modest at low levels of temperature rise.  However, the more recent literature, as synthesized in the Fifth Assessment—a copy of which was leaked recently—departs  from the earlier assessment and finds that, globally, crops will find it more difficult to thrive, and production may be reduced by as much as 2 percent each decade [NYT article]. Given the importance of food supply in health and illness, as well as local economies, the changes to the balance of costs and benefits in the IAMs could be significant.

Objectivity and belief systems

Modeling is always described as an “art as well as science” but the climate change problem goes well beyond the demands of typical model building.

Many of the conclusions of any IAM that is built will depend on a host of decisions that are made by the developer.  The discount rate is one of the most prominent in discussions by economists (and is discussed by Nordhaus in his book), but others are equally important.  How rapidly do we think prices of fossil fuels will rise?  How quickly will technology develop to  bring down the cost of alternative energy and so reduce the costs of controlling greenhouse gases?  How resilient do we think natural systems will be?  How easily will the economies of different countries adapt to climate changes?

While we like to think that these assumptions and  inputs to the models are developed objectively, the gaps in data and the long time period over which the models must make predictions provides much room—as well as need—for judgment.  Climate change pushes the limits of researchers’ ability to claim objectivity.

In 1993, the late Lester Lave and Hadi Dowlatabadi published a seminal article on climate change decision making in Environmental Science and Technology.  The article, entitled, “Climate change: The effects of personal belief and scientific uncertainty,” examined how an individual’s belief affected their perception of climate policy.  The authors identified viewpoints—which they called scenarios—that represented combinations of optimism or pessimism about the cost of abatement (optimism meaning that these costs will be low) and the damages of climate change (optimism meaning that these damages will be low). They gave caricature-like names to these, several of which have always stuck with me:  Dr. Pangloss (the eternal optimist about both abatement costs and damages) and her counterpart, Dr. Doom (the eternal pessimist);  the individual they called the environmentalist (pessimistic about damages but optimistic about abatement costs); and the individual termed the  industrialist (pessimistic about abatement costs but optimistic  about damages).

Our beliefs clearly influence how we view what needs to be done about climate change. And while a modeler’s  view is informed by the literature, it is not determined solely by what other researchers have found.  Of necessity, interpretation must fill in the gaps.  Moreover,  the messages that researchers find in the literature are overlaid on existing belief systems and opinions.

In many respects, benefit-cost analysis of climate change involves picking a rabbit out of a hat. The color and size of the rabbit we pull out will be determined by what we put into the hat.  And what we choose to put in will be determined in large part by our beliefs.

Whither now?

Certainly, there are clearly economic elements to the question of what we should do about climate change: we want to have information on what technologies are out there that reduce emissions and what they cost; we want to understand what the damages will be—for example, what infrastructure will be damaged or destroyed and how expensive it might be to repair or rebuild.  But these economic elements—and our understanding of them—are not sufficient to conduct a benefit-cost analysis of global climate issues with any degree of confidence.   Where global climate change is concerned, benefit-cost analysis should be used to inform—not shape—decisions. Or, to quote a colleague:  benefit-cost analysis provides “insights, not answers.”

There are things that economists are quite good at.  Economics can be far more nimble—and useful—than the rather weighty analyses undertaken by the IAMs might suggest.  While perhaps not as sexy as models with miles of code, information on costs and damages can be used to inform decision making at the sectoral level. For example, state and local governments exploring alternative approaches to coastal protection, or different land use patterns, can use information on the benefits and costs to identify the most cost-effective methods.  Information on costs and damages can be used to assist in short and long term budgetary planning by all levels of government. Businesses deciding where to locate or how to build can take into account the costs of alternative locations, and the potential damages associated with climate change.

But while economic information can assist in specific “project-like” decisions that arise out of climate change, it is at its weakest when asked what we should do about climate change at a global level.  We want to make rational decisions, but that doesn’t mean we want to use benefit-cost analysis to tell us what they are.

Leave a comment

So, what did you think?

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: