In This Issue
Climate Change
September 15, 2010 Volume 40 Issue 3

Risk Assessment and Risk Management for Infrastructure Planning and Investment

Wednesday, September 15, 2010

Author: Gary Yohe

Managing risks associated with climate change is essential for planning and policy decisions.1

In Climate Change 2007: Impacts, Adaptation and Vulnerability, the contribution of Working Group II to the Fourth Assessment Report (AR4) of the Intergovernmental Panel on Climate Change (IPCC, 2007a), authors from around the world focused attention on the impacts of climate change that we either cannot avoid or choose not to avoid. In other words, IPCC made the case that adaptation to climate change should no longer be considered as giving up on the problem. This message is reinforced in the panel report on adaptation to climate change released in May by the National Research Council (NRC, 2010a).

Many impacts of climate change become apparent through increasingly intense and/or more frequent extreme weather events (e.g., heavier precipitation, more intense coastal storms, and severe droughts, floods, wildfires, and heat waves). Both the IPPC and NRC reports describe changes attributed to anthropogenic sources that have already been observed and are threatening some unique social and natural systems. The magnitude of these changes will very likely be exacerbated over the near and more distant future as natural climate variability is distributed around increasingly worrisome central tendencies. Indeed, because temperature increases driven by higher greenhouse-gas concentrations reflect only 50 percent of the corresponding equilibrium warming, near-term decisions to mitigate climate change modestly (or not at all) may actually commit the planet to sudden, irreversible changes by the end of the century (NRC, 2010d; Solomon et al., 2009).

Given the evidence, climate is changing, and absent significant reductions in emissions of greenhouse gases, it will continue to do so at an accelerating pace. IPCC, its client governments, the New York Panel on Climate Change (NPCC, 2010b), the National Academies (NRC, 2010a,b,c,d), and many other international assessments have all turned to risk management as a framework for constructing responses to climate challenges. Indeed, the unanimously approved “Summary for Policymakers” in the IPCC Synthesis Report closed with a statement on the importance of considering risk in all deliberations: “Responding to climate change involves an iterative risk management process that includes both adaptation and mitigation and takes into account climate change damages, co-benefits, sustainability, equity and attitudes to risk” (IPCC, 2007b, emphasis added).

Governments throughout the world have thereby clearly stated their understanding that managing risks associated with climate change must be the central theme in present and future planning and policy decisions. Moreover, now that all four NRC panels of the America’s Climate Choices initiative have accepted this tenet, we can count the United States among those governments.

This article begins by covering some critical definitions and fundamental insights about applying the risk-management paradigm to climate adaptation and mitigation and a brief description of a specific application for infrastructure investment in urban Boston designed explicitly to respond to potential changes in climate driven by natural cycles. This is followed by a description of New York City’s decision to include climate change in its planning processes to protect both public and private infrastructure. The article ends with some observations about the context and applicability of risk management approaches to adaptations to climate change.

Definitions and Fundamentals

Our understanding of some of the aspects of climate change is well established. For example, IPCC (2007b) concluded that it is “virtually certain” that global mean temperatures are rising, and NRC (2010a) confirmed this conclusion. Both assessments also concluded that we know with “very high confidence” that anthropogenic emissions are the cause of this temperature rise.

Thus, even though substantial uncertainties persist about specific sources of risk from specific manifestations of climate change at specific locations, IPCC and NRC agree that near-term action, including adaptation, should be taken immediately to minimize the costs of reducing the rate and magnitude of climate change impacts driven largely by increases in global mean temperature. This means that local decision makers must take action in the face of substantial uncertainties and associated risks, particularly when making decisions about major investments in infrastructure.

All risk management techniques are based on the same statistical definition of risk—the probability that an event will occur multiplied by a measure of its consequences (e.g., Raiffa and Schlaiffer, 2000). Many decision makers favor risk-based approaches because they are based on the same theoretical underpinnings that support other kinds of economic analyses and because they can be applied to situations characterized by significant uncertainty.

Finance directors, government officials, and infrastructure managers, all of whom deal with risk and associated best practices on a daily basis, understand that spreading risk can improve social and/or private welfare. Even though risk diversification does not eliminate risk in most cases, spreading risk does lower net exposure for all participants.

On a fundamental level, first principles of economic efficiency in an uncertain world lead to robust responses that work reasonably well for a wide range of possible outcomes even though they may not be optimal for any particular outcome.2 Because uncertainty is inherent in our understanding of climate change and climate impacts, particularly impacts driven by changes in the distributions of climate variability, it is entirely appropriate that risk has become the “currency of the realm.”

Take, for example, a recent analysis of a large public investment to provide substantial protection along a developed coastline in an urban area of Boston; the area is subject to future coastal storms whose intensities will be amplified by sea level rise (Yohe et al., forthcoming). Figure 1 shows distributions of underlying damages that could be suffered for selected times for a 1.0 meter sea level rise by the end of the century.

Figure 1

Decisions about whether and when to make this investment (and thereby commit to ongoing maintenance expenditures that will last for decades) involved determining when the present value of benefits (i.e., reductions in damages calibrated to include attitudes toward risk) would exceed the present costs. Specifically, the analysis confirmed that damages attributed to sea level rise—the source of value for this adaptation—would increase as risk aversion increased. This is illustrated in Figure 2, which shows the internal rates of return (IRRs) for undertaking the Boston investment at various times between now and 2040 as aversion to risk increases.

To understand Figure 2, relative risk aversion (RRA) must be set at 0 to indicate complete risk neutrality. In other words, RRA = 0 means that decision makers are agreed that a dollar of damage is the same regardless of whether it is the result of a catastrophically large storm or an unusually small storm that might be inconsequential from a societal perspective. Allowing the RRA value to rise above zero indicates that decision makers feel that the consequences of coastal storms increase with their intensities, and at an increasing rate (because simple dollar metrics do not reflect the magnitude of social disruption and human pain caused by larger storms).

Figure 2

For reference, the IRR of an investment indicates the discount rate for which the present value of net benefits is zero. Investments may be made with IRR values greater than the going rate of interest (i.e., the rate by which decision makers discount future costs and benefits), but investments with IRR values below the applicable interest rate are either deferred or discarded entirely.

Therefore, because the IRR increases with risk aversion, investments that reduce risk become increasingly appealing as decision makers become increasingly averse to risk. Because the IRR increases over time, the adaptation investment has a predictably greater chance of being above the implementation threshold with the passage of time.

Several general hypotheses can be derived from this analysis for cases in which the manifestations of climate change cause economic damage stochastically correlated with long-term trends. First, the choice of a baseline against which to gauge the values of various responses to external stress is not just an academic exercise. Differences in baselines, which can be framed in terms of the degree to which economic risk can be spread across a population, and therefore the degree to which RRA approaches zero, can easily change the value of an adaptation and influence its optimal timing.

Second, the economic value of an adaptation should be expressed in terms of differences in expected outcomes (damages with and without the adaptation) only if the affected community has access to efficient risk-spreading mechanisms or reflects risk neutrality in its decision making procedures. Otherwise, increases in decision makers’ aversion to risk will increase the economic value of adaptations that reduce expected damages and diminish the variance of their inter-annual variability.

Finally, for engineering and other adaptations that involve significant up-front expenses followed by annual operational costs for the foreseeable future, increases in decision makers’ aversion will increase the value of that adaptation and, therefore, move the date of economically efficient implementation closer to the present.

The New York City Approach to Adapting Infrastructure

Although in theory a risk-based approach can be applied to many types of adaptation decisions (e.g., retrofitting existing infrastructure, changing the design of new infrastructure, or initiating new infrastructure projects), the requisite data may not always be available. Thus the need to identify information requirements and gaps in knowledge is one reason to begin planning for and prioritizing adaptation options as soon as possible.

New York City adopted a risk-based approach of the kind described above to protect its enormous private and public infrastructure from increasing vulnerability to climate change and associated climate variability. The city was motivated by abstract, sometimes academic constructions of risk which were turned into practical, transferable decision-support tools that could be applied in situations where information was scarce.

From the beginning, the research and policy communities understood that setting climate policy for an entire century would not be possible. For example, based on our current understanding of climate sensitivity, the likely range of temperature rise is from 2oC to more than 4.5oC, but it could also be much higher (IPCC, 2007b). In addition, it is now widely accepted that even advances in fundamental scientific understanding are not likely to lead to substantial decreases in this temperature range.3 Roe and Baker (2007) showed, for example, that “the probability of large temperature increases” is “relatively insensitive to decreases in uncertainties associated with the underlying climate processes.” Allen and Frame (2007) further argued that it is pointless for policy makers to count on narrowing this fundamental uncertainty.

Thus decision makers and resource managers must accept that inflexible, long-term, climate-change policies that can predictably limit greenhouse gas emissions and associated risks will not be put into place in the near future. Therefore, there must be a process by which interim targets and objectives for both mitigation and adaptation can be informed by long-term goals to enable appropriate adjustments to be made as efficiently and transparently as possible (Yohe et al., 2004).

Although this simple conclusion makes sense, problems arise as soon as one begins thinking about how to make it operational, especially for infrastructure investments with lifetimes that can last for many decades or longer. Figure 3 is a schematic portrait of one approach showing a threshold level of acceptable risk (represented by a horizontal wave) that would be breached around 2035 if climate change continues unabated. Incremental adaptation alone, reflected by a “saw-toothed” trajectory, would involve a sequence of responses for keeping risk below the acceptable limit. Since this trajectory approaches the threshold of tolerable risk more quickly and more frequently with the passage of time, Figure 3 illustrates why IPCC (2007a,b) concluded that unabated climate change could easily overwhelm the capacity to adapt by 2100 even in developed countries.

Figure 3

Figure 3 also suggests that mitigation could slow this process—producing a lower risk profile that would not cross the acceptable-risk threshold until 2065. Even though mitigation would provide a delay of only 15 years, this would mean that adaptation responses could be pursued with a more leisurely and presumably less expensive investment program. Schematically, then, it is not difficult to see how mitigation and adaptation might complement each other.

Based on this context, New York City planners worked with the New York Panel on Climate Change (NPCC) to develop a multi-step process to help stakeholders create an inventory of their at-risk infrastructure and develop adaptation strategies to address the risks. Each step, illustrated schematically in Figure 4, became an integral part of ongoing infrastructure maintenance and operation programs, as well as part of a priority-setting planning process for the city agencies and private actors who manage and operate critical infrastructure.

Figure 4

At the start of the project, NPCC (2010b) reported ranges of possible futures (Step 1 at top of the circle). The task force was also provided with tables and other materials showing ranges of broad indicators of climate change through the turn of the century. Perhaps most important to decision makers was a table (Table 1) showing the frequency of extreme weather events that would create risks to particular infrastructure sectors.

Table 1

Building on this and other information reported by NPCC, the city authorized the newly created Climate Change Adaptation Task Force to apply three risk-based decision-support tools that had been designed in consultation with NPCC (2010b): (1) sector-specific infrastructure questionnaires to help stakeholders create inventories of infrastructure vulnerable to the impacts of climate change, especially impacts driven by dynamic climate variability (Step 2 in Figure 4); (2) risk matrices to help stakeholders categorize at-risk infrastructure based on the likelihood of impact and the magnitude of consequences as the climate changes and the parameters of variability mutate (Step 3 in Figure 4); and (3) strategy frameworks to assist stakeholders in developing and prioritizing adaptation strategies based on criteria related to effectiveness, cost, timing, feasibility, co-benefits, and other factors (Step 4 in Figure 4).

These process-based tools ultimately provided a foundation for the development of climate-change adaptation plans for critical infrastructure in the New York City region as part of an overall planning process described in Adaptation Assessment Guidebook (NPCC, 2010a).

The task force was divided into four working groups representing categories of infrastructure important to the city: communications, energy, transportation, and water/waste.4 In addition, a policy group was convened to review the codes, rules, and regulations governing infrastructure in New York City and to identify historical standards that might have to be altered in the future to maintain the current level of acceptable risk. Each group provided participants interested in a particular critical sector of city infrastructure (either public or private) a context within which they could identify common vulnerabilities, share best practices, take advantage of potential synergies, and develop coordinated adaptation plans.

The working groups met from time to time as necessary to ensure consistency, identify opportunities for coordination, investigate impacts of adaptation strategies on other sectors, and develop cross-sector adaptation strategies. In addition, the policy group led reviews for each working group of regulations that embodied common attitudes toward tolerable risk and searched for synergies and inconsistencies across sectors. Informed by these interactions, working groups put forward adaptation strategies that went beyond the scope of their specific interests and could become part of city-wide initiatives.

NPCC’s (2010b) projections of climate change were essential to the process. The projections (reflected in Table 1) provided all stakeholders with a common understanding of climate science, initial potential impacts, and uncertainties. As a result, the inventories and adaptation alternatives developed by stakeholders and proposed to wider audiences were based on the same state-of-the-art climate-change projections. In addition, decisions about allocations of scarce resources were more productive and less contentious than they might have been otherwise.

New York City stakeholders then completed automated templates in which risks for particular “pieces” of infrastructure were determined to be the product of the likelihood an impact would occur and the magnitude of the consequences. Because both terms were described qualitatively, the lack of precise information about distributions and consequences did not impede the process. The risks were then placed in a matrix (see Figure 5).

Figure 5

Perhaps most important, the risk-management process included monitoring and assessment exercises (Step 8 in Figure 4) designed to feed directly into subsequent iterations of the same process. In other words, the city envisioned a dynamic cycle of analysis and action, followed by re-analysis and possible adjustments to or continuations of previous actions (i.e., learn, then act, then learn some more). Thus, the approach was based on an understanding of the need for flexible adaptation pathways of the sort shown schematically in Figure 3 that would evolve over time as knowledge about the climate and local, national, and global economies improved. Indeed, as of summer 2009, the Bloomberg administration intended to push the City Council to pass a law requiring subsequent administrations to submit progress reports and revised climate adaptation plans, just as they are now required to submit updates on progress toward achieving sustainability goals.

Figure 4 also shows that no part of an adaptation plan can be considered in isolation (Step 7). Investments in adaptation programs must be integrated into budget decisions based on a myriad of competing demands for scarce resources (Step 5). New York City has concluded that the urgency and cost of any proposed adaptation response must be compared with other options so its place in the long-term sustainability planning of the city can be determined and supported (Step 6).

Notice that the monitoring function in Step 8 of Figure 4 must include not only adaptations that have been implemented, but also adaptations that have been deferred. In that way, the flexible, iterative program can adjust the urgency of options for the next round of decisions. In summary, New York City’s planning for climate change is a fully integrated component of its ongoing plans for managing growth, infrastructure, and environmental sustainability.

Concluding Remarks

Major conclusions that can be drawn from this brief review of current thinking about addressing risks associated with climate change have been succinctly summarized by the IPCC (2007b). To paraphrase, responding to climate change requires a risk-management approach by which adaptation and mitigation are understood as part of an iterative process that explicitly takes into consideration changes over time and the need for mid-course corrections as knowledge of the underlying science of climate and its translation into climate variability evolves. It follows, then, that governments and other institutions must establish cooperative mechanisms by which they can track, analyze, and project key manifestations of climate change, their associated impacts, the degree to which responses might reduce both exposure and sensitivity to those impacts, and the inevitable interactions of these responses with other private and public initiatives.

It is important to emphasize in closing, however, that the decision to base targeted climate-protection levels over time on socially tolerable levels of risk inferred from existing design standards and codes involves informed, precautionary value judgments that other jurisdictions, with different values and resources, may not find acceptable. Although different values may therefore be adopted in different locations, IPCC (2007b) and NRC (2010a,b,c,d) agree that the risk-management approach to coping with climate change can accommodate this diversity.


Allen, M.R., and D.J. Frame. 2007. Call off the quest. Science 318(5850): 582–583.

IPCC (Intergovernmental Panel on Climate Change). 2007a. Climate Change 2007: Impacts, Adaptation and Vulnerability. Cambridge, U.K.: Cambridge University Press.

IPCC. 2007b. Climate Change 2007: Synthesis Report. Cambridge: Cambridge University Press.

NPCC (New York Panel on Climate Change). 2010a. Adaptation Assessment Guidebook. NPCC Workbook. Available online at

NPCC. 2010b. Climate Change Adaptation in New York City: Building a Risk-Management Response. Annals of the New York Academy of Sciences 1196. Available online at

NRC (National Research Council). 2010a. Advancing the Science of Climate Change. Prepublication. Washington, D.C.: National Academies Press. Available online at

NRC. 2010b. Adapting to the Impacts of Climate Change. Prepublication. Washington, D.C.: National Academies Press. Available online at

NRC. 2010c. Limiting the Magnitude of Climate Change. Prepublication. Washington, D.C.: National Academies Press. Available online at

NRC. 2010d. Climate Stabilization Targets: Emissions, Concentrations, and Impacts of Decades to Millennia. Prepublication. Washington, D.C.: National Academies Press. Available online at

Raiffa, H., and R. Schlaiffer. 2000. Applied Statistical Decision Theory. New York: Wiley Classics.

Roe, G.H., and M.B. Baker. 2007. Why is climate sensitivity so unpredictable? Science 318(5850): 629–632.

Solomon, S., G-K. Plattner, R. Knutti, and P. Friedlingstein. 2009. Irreversible climate change due to carbon dioxide emissions. Proceedings of the National Academies of Science 106(6): 1704–1709.

Yohe, G., N. Andronova, and M. Schlesinger. 2004. To hedge or not against an uncertain climate future. Science 306(5695): 416–417.

Yohe, G., K. Knee, and P. Kirshen. Forthcoming. On the economics of coastal adaptation solutions in an uncertain world. Climatic Change.


 1 This article relies heavily on material in Chapter 2 of NPCC (2010b) and the section on the New York City approach in Chapter 5 of NRC (2010b).

 2 Risk analyses have demonstrated that decisions are critically dependent on subjective prior distributions with which we weight the relative likelihoods of future outcomes, thereby demonstrating how aversion to risk influences the value of information. Economic efficiency establishes criteria by which maximal welfare could be achieved with limited resources by allocating them effectively to meet a wide range of competing demands.

 3 Climate sensitivity is the increase in equilibrium global mean temperature associated with a doubling of atmospheric concentrations of greenhouse gases from pre-industrial levels. IPCC (2007a) reports, for example, that “the equilibrium climate sensitivity is a measure of the climate system response to sustained radiative forcing. It is not a projection but is defined as the global average surface warming following a doubling of carbon dioxide concentrations. It is likely to be in the range 2°C to 4.5°C with a best estimate of about 3°C, and is very unlikely to be less than 1.5°C.  Values substantially higher than 4.5°C cannot be excluded, but agreement of models with observations is not as good for those values.” 

4 The Waste and Water Working Group includes agencies that handle the city’s solid waste and wastewater as well as entities responsible for the natural environment and a significant portion of waterfront properties.


About the Author:Gary Yohe is Woodhouse/Sysco Professor of Economics, Department of Economics, Wesleyan University.