Ignore At Your Own Risk: A Risk Management Perspective on Climate Change

Gallons of digital ink has been spilled discussing and debating the legitimacy of concerns over climate change and global warming. Unfortunately, like many internet debates, both sides of the argument at times rely on emotional arguments, unrealistic assumptions, and overly simplistic outlooks, with a noticeable lack of nuance. While there is overwhelming scientific consensus that global climate is definitively warming, this alone provides only a small part the story. The Earth is indeed warming, but so what? Does it matter, and if so what are possible solutions?

Global warming has environmental, economic, and societal impacts. The projections of the specific timing and severity of those impacts result in a wide range of values, however. Given this, How can we pinpoint the required industrial and societal changes necessary for reducing those impacts if we cannot precisely quantify those impacts in the first place? An effective climate change policy must account for numerous scientific and socioeconomic uncertainties that are likely to substantially affect both the costs and benefits of any policy action.

Scientific Background

Climate change science is in strong agreement on several important points: While there are several unknowns concerning the specific timing and severity of climate change impacts, robust evidence supports the overall concept that the Earth is significantly warming to a dangerous level and that human activity through the release of greenhouse gases (GHGs) is primarily responsible for this effect. Examples of some potential impacts include:

       ·         Increased human mortality due to extreme heat, starvation, and economic stress

       ·    Property damage in coastal and flood plains communities from sea level rise and greater precipitation and storm intensity

       ·    Increased water insecurity due to drier conditions, greater energy demands, and rising sea levels

       ·    Decreased worker productivity and increased mortality due to extreme heat

       ·    Ecosystem destruction and possible extinction of various species

       ·    Decreased regional agriculture yield due to drought

       ·    Greater incidence of infectious diseases due to increased range of insects and water pollution

       ·    Diminished global economy due to reduced productivity and increased costs

Risk and Uncertainty

Climate change cannot be “solved”, but we can attempt to reduce its risks. Risk (in the context of climate change) is defined by the National Academies of Sciences as “a combination of the magnitude of a potential consequence of a hazard or hazards attributable to climate change and the likelihood that the consequence will occur. It may refer to physical, biological, or socioeconomic consequences.” In other words, risk refers to negative impacts of any damages multiplied by the overall likelihood of their occurrence. Risk can be described either quantitatively or qualitatively. Risk analysis includes the identification and assessment of risks (risk assessment), along with development and execution of strategies to balance the costs and benefits of reducing those risks (risk management).

A particular challenge in analyzing risks, especially those associated with climate change, is uncertainty. If decision makers knew exactly what would happen when, and how bad the outcome would be, devising responses to balance mitigating the damage against the cost of doing so could be fairly straightforward. A now-infamous quote from former Secretary of Defense Donald Rumsfeld highlights the various types of uncertainty:

Indeed, in addition to the factors we are aware of, there always may be variables that we have not even yet thought of. While uncertainties should be sufficiently acknowledged, their presence does not preclude the possibility of effective risk management.

Climate change may be more complex than other issues, but it is certainly not unique as a policy issue (environmental or otherwise) that involves substantial uncertainty. If we avoid treating opinions on climate change as merely a binary yes-or-no political purity test, perhaps risk analysis strategies could be used to reduce climate change risk in a way that will be more agreeable to the majority of the country.

Uncertainty in Climate Change
Inherent variability within a system incorporates imprecision into any model. It differs from uncertainty in the traditional sense and can be considered fully distinct in the context of climate change. For example, there is large variability in the weather patterns of Omaha, NE over the course of a month. The known range of variability for that particular time and location is part of the overall long-term climate however, so the imprecision of specific measurements is already accounted for in the climate pattern. Figure 1 demonstrates how the probability of individual weather outcomes changes as climate shifts over time.

Figure 1. The Relationship Between Climate Variability and Weather

Climate includes a wide distribution of potential daily temperatures. Although there may be temporary periods when multiyear temperatures remain stable or fall, the overall increase in warming will result in elevated average temperature and increased probability of previously rare extreme heat events.

Source: Gordon, Kate et al. “Risky Business: The Economic Risks of Climate Change in the United States - A Climate Risk Assessment for the United States.” Risky Business Project, June 2014.

The phenomenon of climate change involves numerous unknowns of differing magnitudes. Models of future global warming and associated climate impacts rely on various assumptions and estimates. Climate change analyses are therefore faced with both scientific and socioeconomic uncertainties. These analyses are extremely difficult, as climate change involves what is known as deep uncertainty, in which numerous uncertainties and conflicting individual opinions make it impossible to even fully define what factors are important.  This means that there will always be a substantial subjective component to any climate change analysis. Acknowledging and accounting for potentially-irreducible uncertainties is necessary for presenting an honest and accurate analysis. Some examples of these uncertainties are below:

Scientific	Socioeconomic
Statistical relationship between GHGs, warming, and climate impacts	True cost vs benefits of different policy options
Accuracy of climate models	Future technological advances that could reduce costs
Timing and severity of particular climate change impacts	Subjective importance of the disproportional effects of climate impacts on different groups
Potential for rare but catastrophic outcomes	Quantifying the value of non-financial factors (e.g. the environment)

These factors affect scientific models of climate change and its impacts, economic projections of various policy options, and the immediacy of the need for action on climate change.

Understanding Climate Change within the Realm of Uncertainty
Deep uncertainty about the severity of impacts, the timing of their effects, specific regional differences, or the proper economic valuation of climate change does not detract from the overall evidence suggesting that climate change increases the probability of detrimental impacts. This probability only increases as warming continues. Additionally, there is not only the possibility of impacts being more mild than predicted. Unforeseen major risks may exist that have not even been accounted for in current climate models, making the actual damages much greater than estimated. 

Issues of Timescales

Given both what we know, and what we don’t know, does the evidence favor policy intervention immediately, or is it better to collect more information before implementing any significant changes? The below chart includes some considerations in favor of each side:

Act Sooner	Wait for More  Information
Uncertainty is unlikely to be reduced within a reasonable timeframe and may even increase as new parameters emerge	If climate damage is linear over time, greater-than-normal benefits/cost ratios are required for justifying expedited action
There is a non-zero probability of unlikely yet catastrophic impacts (“the tails” of the distribution)	Policies aimed at drastically reducing emissions drastically may require or induce dramatic technological and industrial shifts that could have significant economic effects
Positive feedback (such as melting ice caps releasing more methane) will exponentially increase warming over time	Early investments in climate protection without certainty concerning their benefits may represent large sunk costs
Certain climate impacts may be irreversible within human timescales if not prevented

Applying risk analysis to climate change can help gauge both the severity and likelihood of various risks while also evaluating options for managing those risks. It is important to note however that risk determinations are probabilistic. Therefore, any policy decision may ultimately turn out to be suboptimal even if arrived at through the best available practices. Additionally, there is no magic target for emissions or other policy action. Even the best policies can only hope to reduce risk, not completely eliminate it.

Factors Influencing Risk Management Decisions

Dealing with uncertainty in any policy area usually requires at least some aspect of subjective decision making. In determining how stringent a policy should be, decision makers must consider their risk attitude – a value judgement made in the face of a risky situation. In other words, when assessment results are uncertain, how much are you willing to risk?

Your risk attitude is influenced by various factors that affect your risk perception, or how you personally recognize underlying risks and their consequences. Risk perception can be affected by both subconscious biases and emotional influences, as well as conscious rational assessment. These factors differ among every group and individual, resulting in a potentially wide range of risk attitudes. For example, public health agencies and organizations might choose to be quite risk averse, instituting more stringent regulations in order to ensure civilian safety. Financial investments on the other hand may range from risk averse to risk tolerant or even risk seeking, depending on the chosen strategy.

Environmental risk analysis, especially for climate change, is particularly complex because it involves ethical and moral issues in additional to standard economic analysis. It is difficult to apply economic values to human suffering or ecological damage, especially when negative impacts will be disproportionally distributed. Climate change also involves issues of environmental justice, because economically vulnerable groups are most likely to suffer from its effects. Moral considerations become especially prominent when considering how to value impacts of ongoing U.S. emissions on both non-U.S. populations and future generations. Similarly, climate change risk management may require prioritizing which risks are most damaging, and which are acceptable. How one feels about all of these issues will influence their personal risk attitude toward climate change policy.

Climate Change Risk Management Strategies Under Uncertainty

The Congressional Budget Office lays out three potential responses for dealing with climate change: mitigation of GHGs, adaptation of human and ecological systems to an increasingly warmer climate, and continued research into climate impacts and how to minimize them. Mitigation involves actions aimed at reducing the rate of climate change through reduction of atmospheric GHG levels. Mitigation practices may take the form of promoting low and zero-carbon energy sources, improving energy efficiency, new technologies to remove GHGs from ambient air, and improved land management practices. Adaptation refers to actions taken to limit the damage to people, communities, and infrastructure from climate change impacts. The term is often used interchangeably with resiliency. Adaptation can include methods such as building protective infrastructure, providing compensation for loss, emergency planning, and promoting best practices for risk reduction. Successful risk management strategies will likely involve a mix of both mechanisms. Research could be directed towards reducing important uncertainties, developing new technologies to aid mitigation and adaptation, improving resolution of climate models down to the local/regional level, and studying risk management strategies.

Cost-benefit analysis (CBA) is a common strategy for evaluating the net economic outcomes of particular policy decisions in monetary terms. CBA attempts to determine whether the benefits of a proposed policy or decision outweigh the costs, and by how much, given certain assumptions. CBA attempts to monetize the value of all considerations, including effects on quality of life and mortality. The U.S. federal government developed a metric in an attempt to quantify the “costs” of climate change – equivalent to the benefits of mitigation – for use in CBA. This metric is known as the Social Cost of Carbon (SCC) and is expressed as U.S. Dollars per additional ton of CO2 (or other greenhouse gas).

When calculating costs and benefits for CBA, discount rates are often required. Discount rates account for inflation, since future benefits are worth less in today’s value per dollar. Therefore, $10 spent today on a policy must provide greater than $10 in future benefits in order to be economically worthwhile. As the chart indicates, discount rate can have a very significant effect on the evaluation of relative costs and benefits. Rational arguments exist for discount rates between ~1% and 7%, based on how one interprets the potential for future investment earnings or destructive losses. This decision ties back once again to risk attitude.

CBA provides a probability distribution of outcomes, however only the average or “expected” value is typically used for analysis. It is therefore not ideal for dealing with very large uncertainties, and it does not sufficiently account for the extremes at the “tails” of the model. Despite its previous use in climate change policy analysis, it therefore may not be the ideal strategy for dealing with all of the associated uncertainties and complexities. The academic literature highlights several methodologies (OECD 2015, see summary on p. 132) for decision making and risk management under uncertainty in the context of climate change. The two major strategies are described below:

Adaptive Strategies: Adaptive strategies involve selecting policies that allow refinement over time through evaluation, monitoring, and future learning. This strategy requires the ability to identify important thresholds or “trigger points” that would prompt an update to the policy. Specific examples include:

·      Iterative Risk Assessment – A commitment to continuously updating policy over time based on what is learned from the previous period.

·      Real Options Analysis – A strategy borrowed from financial markets, real options analysis involves paying a small premium upfront to greatly reduce costs of potential future adjustment (e.g. build levee walls against expected future flooding, but with a structure easily allowing expansion of flooding exceeds expectations).

Robust Strategies: Robust strategies attempt to identify policy options that produce a positive net benefit under a variety of scenarios. The most robust option may not maximize potential benefit, but will provide positive value over a wide variety of circumstances. Specific examples include:

·      Portfolio Analysis – Making broad and diverse investments, ensuring that they properly account for risk and work cohesively together to accomplish a goal.

·   Robust Decision Making – Use computer modeling to simultaneously test the effect of various model parameter changes on the outcomes of a policy decision. In addition to identifying "robust" policy decisions, it can determine which parameters have the greatest influence on the economic outcome.

Robust Decision Making (RDM) has shown its utility in real-world examples. In 2007, Congress was deciding whether to reauthorize the Terrorism Risk Insurance Act (TRIA), which was established after 9/11 to compensate private insurers for massive losses from large terrorist attacks. In determining whether the bill would be likely to net save or cost money, the RAND Corporation utilized RDM to identify the most important factors affecting the bill’s relative cost. While the Congressional Budget Office and Treasury Department provided a single-value estimate predicting that the TRIA would result in higher costs, RDM provided a more complete picture of possible outcomes that accounted for important uncertainties. Analysis found that estimated costs were heavily dependent on both the assessed likelihood of a large terrorist attack and how much compensation Congress would choose to provide for losses (Figure 2). RDM demonstrated that the TRIA would result in higher costs only under relatively unlikely scenarios (high likelihood of attack and very low compensation). This analysis aided Congress in its eventual decision to re-authorize the TRIA.

Figure 2. RDM Output for Deciding on TRIA Reauthorization

RDM identified that TRIA would save money under the majority of likely scenarios, with the likelihood of a terrorist attack and the compensation offered by Congress proving to be the most influential variables.

Source: Lempert, Robert J., Steven W. Popper, David Groves, Nidhi Kalra, Jordan R. Fischbach, Steven C. Bankes, Benjamin P. Bryant, et al. “Making Good Decisions Without Predictions,” 2013.

 

Conclusions

Climate change clearly presents an existential threat to the world. It is important however, to take reasoned, economically realistic approaches toward a solution. Fixing such a “wicked problem” that doesn’t have a true “correct” answer is MUCH easier said than done, and beneficial strategies will likely involve some necessary trade-offs. It is impossible to know future damages of climate change, or the costs of any policy aimed at reducing those damages. By acknowledging and not avoiding the uncertainty inherent in any decision making, we can come up with an approach that is at least most likely to work.