Costs and Benefits of Climate Change Impacts and Adaptation

Climate change already results in economic impacts and will do so increasingly in the future. These impacts affect different aspects of the economy, public health and the natural environment. Assessing the economic impacts of climate change is a complex undertaking, with considerable uncertainties surrounding the magnitude of future biophysical impacts and the monetary value of those impacts. Notwithstanding these difficulties, economists have been examining the relationship between climate change and economic impacts for over 20 years. In 2011, for example, the National Round Table on the Environment and the Economy (NRTEE) estimated the average future cost of a high climate change‒rapid growth scenario for Canada at $35‒$62 billion (2019 dollars) annually by 2050, with a 5% chance that costs could exceed $72‒$131 billion per year (NRTEE, 2011).

Information on the economic consequences of climate change, as well as on the costs and benefits of alternative courses of action, is increasingly being demanded by a wide range of private and public sector actors. This information is needed to inform resource allocation decisions in response to actual and projected climate change risks (National Research Council, 2010; 2009). Two generic response options are available: greenhouse gas (GHG) emissions reduction and adaptation measures (see Box 1.2 in Canada’s Changing Climate Report)—an effective and efficient policy response will require a mix of both options. Indeed, from an economic perspective, the total costs associated with climate change can only be minimized through a combination of GHG emissions reduction and adaptation actions (e.g., Agrawala et al., 2011; de Bruin et al., 2009a).

The economics profession has historically been more focused on GHG emissions reduction (Fankhauser, 2017), although the number of studies on adaptation costs and benefits is increasing. The Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) dedicated an entire chapter to the economics of adaptation (Chambwera et al., 2014) and several other recent reviews have also focused on this subject (e.g., Kahn, 2016; Rouillard et al., 2016a; Markandya et al., 2014).

This chapter assesses the state of knowledge and practice on climate change impacts and adaptation economics in Canada. It focuses on answering the following questions: What do we know about the economic costs of climate change for Canada? What is the distribution of these costs across different regions, sectors and population centres? What are the costs and benefits of actions taken to moderate potential damages or to seize beneficial opportunities? And what economic tools and methodologies can be used by practitioners to address these questions? Decision makers need answers to these questions in order to allocate scarce public and private resources for climate change adaptation, and to ensure that these resources are directed towards the most efficient actions. This chapter will be of interest to a wide range of decision makers, economists and practitioners at all levels of government, and to businesses operating in climate-sensitive sectors.

Over the last several decades, extreme weather events—such as wildfires, flooding, heat waves and storms—have caused billions of dollars in economic damages annually worldwide (Aon, 2020; Swiss Re Institute, 2020). Since 1980, cumulative damages worldwide have surpassed $4.9 trillion (2019 dollars)¹ (Munich RE, 2020); the U.S. alone has sustained about $2.2 trillion (2019 dollars) in damages resulting from 265 weather and climate disasters over the last 40 years (National Centers for Environmental Information, 2020). Over a similar period, damages in Canada totalled about $31 billion (2019 dollars) (Public Safety Canada, 2020). Inflation-adjusted damages have also been trending upwards—globally, regionally and in Canada (see Section 6.4.2; Aon, 2020; National Centers for Environmental Information, 2020; Insurance Bureau of Canada, 2019).

Climate change has increased the likelihood of certain types of extreme climate and weather events occurring (Zhang et al., 2019; National Academies of Sciences, Engineering and Medicine, 2016) and is expected to intensify some events in the future (Bush and Lemmen, 2019). Unabated climate change is projected to result in hundreds of trillions of dollars in economic damages globally in 2100 (Warren et al., 2018)—both from the intensification of certain climate extremes and from the impacts of slow-onset climate trends (e.g., processes like sea-level rise and melting of permafrost). A recent study, for example, suggests that a persistent increase in average global temperature of 0.04°C per year (consistent with a scenario of no major policy changes and continued GHG emissions) will reduce global economic output per capita by about 7.2% below where it would otherwise be in 2100; projected declines in per capita output in the U.S. and Canada are higher still, at 10.5% and 13.1%, respectively (Kahn et al., 2019).

Adaptation can significantly reduce the projected costs of climate change by billions of dollars per year (U.S. Global Change Research Program, 2018), though it is unlikely to entirely offset economic damages (see Section 6.8.2). Ambitious policies to reduce global GHG emissions are also needed to limit the negative impacts of climate change (OECD, 2015; Agrawala et al., 2011; de Bruin and Dellink, 2011; Wang and McCarl, 2011). However, adaptation is not costless. Globally, it is estimated that investment needs for climate change adaptation in industrialized countries will reach US $29–$138 billion (2019 dollars) per year by 2030 (UNFCCC, 2007). Adapting coastlines and water, transportation and energy infrastructure in the United States could cost tens to hundreds of billions of dollars annually by 2050 (Sussman et al., 2014). In Canada, an investment of just over $5 billion (2019 CAD dollars) will be needed annually, on average, over the next 50 years to adapt municipal infrastructure (buildings, facilities, roads, etc.) to climate change (Insurance Bureau of Canada and Federation of Canadian Municipalities, 2020). The exponential shape of adaptation cost curves suggests that initial levels of adaptation can be achieved at relatively low cost, but that costs could be substantially higher in the long term as increasingly less cost-effective actions are required to achieve greater levels of adaptation (Agrawala et al., 2011). Nevertheless, judicious adaptation decisions can yield benefits—in the form of avoided damages—that far exceed costs (Global Commission on Adaptation, 2019; Lempert et al., 2018).

Given the potential magnitude of investment costs for climate change adaptation in the short and long terms, there is a need to provide decision makers with reliable economic information on costs and associated benefits to support adaptation investment decisions. Decision makers—whether in the public or private sector—face limited human and financial resources. They will not be able to pursue every prospective program or policy, and so must justify and set priorities for allocating available resources, including for climate change adaptation strategies and actions. In this regard, the field of economics can be of assistance, as it encompasses the study of how to efficiently allocate resources to meet desired goals. Specifically, economic analysis can help decision makers to weigh the costs of acting vs. the costs of inaction (i.e., continuing with a business-as-usual approach); to choose how much to invest in relation to competing priorities that are not climate-related; to decide which types of adaptation options, sectors and locations should receive resources; to balance near-term and long-term objectives; and, relatedly, to consider the impacts for future generations (Chambwera et al., 2014; National Research Council, 2010). Many adaptation benefits (i.e., avoided damages) will also deliver impacts in other areas, including health and safety, cultural heritage, ecosystem services and equity. Failure to include these types of non-market considerations in the decision-making process leads to underinvestments in adaptation. Economic analysis can be helpful in this regard as well, offering specialist techniques for capturing non-market climate change impacts in decision making.

As decision makers become increasingly aware of the risks of climate change, there is growing demand for more effective ways to support adaptation decisions (Moss et al., 2014; National Research Council, 2010). The framing for adaptation planning has changed to meet these demands. With increased recognition of the need to manage uncertainty and develop practical early actions, there has been a shift towards a more policy-centric approach. Such an approach has the starting objective of climate change adaptation, as well as increased interest in the timing and sequencing of adaptation options, and the use of adaptive risk management frameworks for decision making (Rouillard et al., 2016a; Watkiss, 2015). These changes have important consequences for the use of economic analysis in informing adaptation decisions.

Decision support in the context of climate change presents unparalleled challenges. Uncertainties associated with climate change—relating to how future social and economic systems will evolve; time lags between human activities and the response of the climate system; the dynamics of climate and biophysical systems; the diverse mix of potentially affected stakeholders; and autonomous adaptation by natural and human systems—make it hugely difficult to predict when and where climate change impacts will occur, as well as their relative importance (Chambwera et al., 2014; Jones et al., 2014; Heal and Millner, 2013). Regarding adaptation, these uncertainties are exacerbated at the regional and local levels, where many adaptation options are implemented. Adaptation decisions are complicated by further uncertainties relating to different stakeholder perspectives, multiple and competing objectives, long decision time frames, the choice of monetary values, and the broad range of adaptation options to select from (e.g., private or public, reactive or planned, stand-alone or integrated (“mainstreamed”) (Rouillard et al., 2016a; Jones et al., 2014; Li et al., 2014; Randall et al., 2012).

Consideration of uncertainty is fundamental to adaptation decisions and related economic analyses. Given the multifaceted and uncertain nature of adaptation decisions, the consensus view is that such decisions are best considered in an adaptive (i.e., iterative) risk management framework (Lempert et al., 2018; Jones et al., 2014; Moss et al., 2014; IPCC, 2012; National Research Council, 2010).

Adaptive risk management provides a framework in which potentially significant, but uncertain, consequences of current and future climate change and adaptation actions are continually identified, assessed, prioritized, managed and revised; this framework includes monitoring, which takes into account new information, experience and stakeholder input (Lempert et al., 2018; National Research Council, 2010). It entails an ongoing cycle of assessment, action, reassessment and response that will continue in perpetuity, rather than informing one-off decisions at a single point in time (Lempert et al., 2018; Willows and Connell, 2003). The National Research Council (2010) draws an analogy with decisions in a chess game, where pieces are repositioned and risk is reassessed in response to the opponent’s moves.

From an economic analysis perspective, adaptive risk management provides a useful framework for adaptation decision making. It allows for the use of a broad range of concepts, processes and decision support tools—including traditional tools like cost-benefit analysis (CBA), cost-effectiveness analysis (CEA) and multi-criteria decision analysis (MCDA) (see Table 6.4), as well as tools that are more adept at accommodating deep uncertainties, such as robust analysis and dynamic adaptation pathways (see Figure 6.3; Table 6.4; Moss et al., 2014). Importantly, adaptive risk management allows decision makers to consider a broad range of criteria (e.g., costs, benefits, co-benefits and co-impacts (see Section 6.3.4), equity, affordability, flexibility, robustness, etc.) when formulating adaptation strategies in the face of uncertainty (see Section 6.6.1).

The awareness, assessment and planning stages of a generalized adaptive risk management framework provide specific entry points for economic information, analysis and decision support (see Figure 6.1). During the awareness raising stage, information on the costs of inaction (i.e., the net cost reflecting the difference between economic damages and any beneficial opportunities arising from climate change) can be used to persuade decision makers of the need and urgency to allocate resources to adaptation planning. This information may include estimates of the scale of climate-related costs; the distribution of those costs across locations, sectors, population groups, etc.; and the time frame over which they are projected to become significant, if they are not already so. The same information can also be used by analysts and stakeholders to inform the prioritization of current and future climate risks and vulnerabilities during the assessment stage. Economic analysis also plays an important role during the planning stage, where it can be used to inform the overall scale of investment in adaptation; the selection, timing and sequencing of specific adaptation options; as well as the distribution of adaptation costs and benefits.

The assessment and planning stages of an adaptive risk management framework are typically navigated following one of two generic analytical processes (Jones et al., 2014). Historically, the predominant approach is based on a “science-first” (also known as “top-down”, “scenario-led” and “predict-then-act”) impact assessment-driven process (Gregory et al., 2012; Wilby, 2012; Ranger et al., 2010). This involves first defining impact pathways—climate change projections combined with socioeconomic information to assess future risks and costs—using impact models or damage functions (i.e., an empirical relationship characterizing the predicted change in monetary damages that is attributable to a change in a climate variable or index). The range of estimated risks is then used to frame the selection of adaptation options. Identified options are appraised as a final step in the process to determine the desired adaptation level, which is informed by adaptation costs, benefits and residual costs (i.e., the monetary damages attributable to climate change that will remain after adaptation) (Watkiss, 2015). With this approach, however, uncertainty is compounded at each stage of the analytical process (see Figure 6.2) and is seldom adequately characterized (Wilby, 2012). In the presence of such ballooning uncertainties, the range of plausible impacts and adaptation responses can become unworkable, rendering the “science-first” approach impractical (Dessai et al., 2009; Dessai et al., 2005). Further issues with the “science-first” process include the following: it does not adequately address non-climate drivers of impacts and risks; its long-term focus does not align with immediate policy needs to inform near-term adaptation decisions; it fails to consider the adaptation process itself and ignores potential barriers, transaction costs and baseline policies; and it tends to emphasize “hard” (i.e., engineered) adaptation options over “soft” options, like building adaptive capacity (Rouillard et al., 2016a; Watkiss, 2015; Patt et al., 2010).

Given the shortcomings of the “science-first” approach, there has been a shift in adaptation practice towards “policy-first” analytical processes (Rouillard et al., 2016a; Watkiss, 2015; Watkiss et al., 2015a; Downing, 2012). “Policy-first” approaches—also known as “bottom-up”, “assess-risk-of-policy” and “decision-centric” approaches (Pielke et al., 2012; Brown et al., 2011; Ranger et al., 2010; Dessai and Hulme, 2007)—place greater emphasis on adaptation as the starting objective, rather than considering it as the final step, which is usually the case in a traditional “science-first” impact assessment. With the “policy-first” approach, a significant amount of effort is devoted at the outset to characterizing the decision problem (e.g., a flood risk management plan). This includes first identifying relevant objectives, current practices, constraints and drivers of change, as well as stakeholder preferences and related decision criteria, all of which frame subsequent analyses. Next, the process involves assessing the vulnerability of the defined system to current climate, socioeconomic and policy conditions, before considering sensitivities to future stressors, including climate and non-climate related stressors (Ranger et al., 2010). Once the limitations of current practices are understood, alternative options are identified if necessary and assessed with respect to achieving the stated objectives across a range of plausible future scenarios. For instance, the Thames Estuary 2100 Project in London, U.K., (see Case Story 6.4) was one of the first large-scale infrastructure projects to adopt a “policy-first” approach to adaptation planning (Ranger et al., 2013).

Compared to the classic “science-first” approach, the “policy-first” approach has a number of advantages (Gregory et al., 2012; Pielke et al., 2012; Brown et al., 2011; Ranger et al., 2010; Wilby and Dessai, 2010; Dessai et al., 2009). For example, because the “policy-first” approach requires only climate information pertinent to the decision problem at hand and focuses the assessment on adaptation options that are acceptable given the objectives and constraints of a particular decision, the analysis is streamlined and targeted from the outset, which makes it less resource- and data-intensive, as well as less sensitive to ballooning uncertainties. Furthermore, because the analysis is context-driven and not unduly influenced by scientific modelling, it emphasizes “big picture thinking” and encourages decision makers to consider interactions with broader policy priorities and to seek adaptation options that deliver co-benefits with other policy areas.

Alongside the shift towards a “policy-first” approach and the use of adaptive risk management frameworks, increased consideration is also being given to the timing and sequencing of adaptation as a further means to manage uncertainties (Wise et al., 2014). There has been a move away from viewing adaptation as comprising one single response (e.g., a wall to reduce future flood risk). Instead, there has been a shift towards viewing adaptation as a coherent package of options to address current vulnerabilities and to prepare for medium- and long-term climate change risks, with a focus on practical early implementation (Rouillard et al., 2016a; Watkiss, 2015). Packages of options will typically comprise three types of activities or building blocks for early action (Rouillard et al., 2016a; Watkiss, 2015; Watkiss et al., 2014):

1. Actions to address the existing adaptation deficit: Immediate adaptation options that address risks arising today from current weather and climate extremes and, in so doing, also build resilience to future climate change. This would include “win-win”, “no-regret” and “low-regret” adaptation options that provide clear, immediate benefits or co-benefits.
2. Actions to adapt decisions with long lifespans: Adaptations that are mainstreamed into near-term decisions that have long lifetimes (e.g., decisions relating to climate-sensitive infrastructure or land-use planning), and thus will be influenced by future climate conditions and future risks, in addition to current conditions. In contrast to the previous category, uncertainty over future benefits is a much bigger concern. As a result, greater emphasis is placed on using robust options (i.e., actions that provide future benefits under a range of plausible future scenarios) and flexible strategies that provide opportunities for learning, with options that can be delayed or brought forward, and/or scaled up or down as new information emerges over time.
3. Actions to support long-term adaptation: This includes activities such as monitoring, surveillance, research and engagement, which immediately start building the capacity needed to support future actions to manage long-term climate impacts and risks. Examples include generating better information—which is required to inform later decisions about managing major, highly uncertain long-term risks—and actions that are needed to create a suitable policy and socio-cultural environment to enable and ensure that future options are still possible.

When viewed collectively as an integrated adaptation strategy, these three building blocks form a dynamic adaptation pathway (Haasnoot et al., 2018; 2013), as shown in Figure 6.3.

The shift towards a “policy-first” assessment process, coupled with greater emphasis on the timing and sequencing of adaptation, and use of adaptive risk management frameworks, has had significant implications for the economic analysis of adaptation strategies. Mainly, these changing practices have necessitated the development, application and refinement of alternative decision support tools. Traditional economic decision support tools (e.g., CBA, CEA) are adequate for appraising early actions to address the existing adaptation deficit, where uncertainty over future impacts is less of a concern. However, for mainstreaming adaptation into decisions with long lifespans—where consideration of climate and non-climate-related uncertainties over future drivers of change is much more important, and where decision makers are thus looking for robust options or flexible strategies—alternative economic decision support tools like real options analysis, robust decision making, portfolio analysis and dynamic adaptation pathways are more appropriate for appraising options (see Table 6.5). Use of these tools globally in a climate change adaptation context is still in its infancy, and they have yet to be formally applied in Canada (see Section 6.7.1).

Further consequences of these changing practices for the economic analysis of adaptation include:

• With greater importance being placed on 1) mainstreaming and understanding the process of adaptation (including barriers to action), and 2) capacity building to ensure that long-term adaptation options remain possible (such as research, monitoring and institutional strengthening), there is an increasing need to assess the costs and benefits of non-technical options, including behavioural interventions to overcome barriers to change, and the value of information collated through monitoring systems. The characteristics of these options are different from those of outcome-based or engineered actions, with costs and benefits that are more challenging to measure and to include in economic analysis.
• Likewise, the increased emphasis on no-regret and low-regret adaptation options places greater importance on the need to fully capture co-benefits in the economic analysis, which requires the monetization of a broader range of non-climate impacts, in addition to avoided climate-related damages.
• Viewing adaptation strategies as a set of time-sequenced activities presents challenges for economic analysis, since each building block is unique and may require different information and methods for the quantification and valuation of physical impacts, and may entail resource implications for the analysis.
• All economic analyses typically need to consider trade-offs between early costs and future benefits, rendering the results sensitive to the discounting process and choice of discount rate (see Section 6.6.3.2). Analysis of actions to address the current adaptation deficit will generally be less sensitive to discounting assumptions. However, for early actions to adapt long-life decisions and keep long-term options open over many decades, outcomes will be more sensitive to the discounting of future benefits, making it essential to consider alternative and inter-generational discounting practices.

Typologies of the economic consequences of climate change impacts—specifically impacts arising from extreme events—often distinguish between direct and indirect impacts, similar to the literature on natural disaster impacts. Direct and indirect impacts can be negative or positive, giving rise to costs (i.e., from losses or damages) or benefits (i.e., gains), respectively. This section refers solely to costs, although it applies equally to benefits.

Direct costs arise from the physical impacts of climate hazards, such as damage or disruption to tangible goods and services that can be traded in a market and thus have an observed price (e.g., costs incurred to repair or replace damaged homes, medical treatment costs for heat stress, lost revenue from reduced crop yields, etc.). Direct costs also arise from physical impacts to intangible items not bought or sold in a traditional market and thus having no readily observable price (e.g., ecosystem services, stress or pain levels, and general quality of life). Economists have developed multiple techniques to allocate a “shadow price”—an estimated price for a good or service whose market price does not accurately reflect its actual value or for which no market price exists—to these intangible items, which are referred to as non-market impacts. When non-market impacts are rendered equivalent to market impacts using shadow prices, they can be substantial—perhaps larger than market costs (Nordhaus and Boyer, 2000). Omitting relevant non-market impacts from the economic analysis of adaptation strategies could substantially bias the outcomes.

Indirect costs stem from direct climate change impacts. When infrastructure, a building or a park is damaged or destroyed, this can interrupt normal use or service flows (e.g., a flooded shop may have to temporarily close for repairs). Damaged infrastructure may result in disruption to the delivery of critical services (e.g., electricity, water, sanitation), which may interrupt the operations of businesses that are not directly affected by climate hazards. Workers may also not be able to get to work if road networks or transit infrastructure have been affected. These impacts are referred to as business interruption costs (Kousky, 2012). Interactions between businesses may in turn result in secondary or multiplier impacts through the economy (e.g., a flooded shop that closed its doors for repairs will not need to purchase supplies until it reopens). Like direct costs, indirect losses can also be divided into market costs (e.g., business interruption costs) and non-market costs (e.g., delayed illnesses and mental health disorders, increased inequality, etc.). In contrast to direct costs, indirect costs often span a longer time period and take place over a wider spatial scale than the site of the direct physical impacts of climate change (Hallegatte, 2013).

The sum of all relevant direct and indirect, market and non-market costs provides one measure of the total economic impact of climate change. Interest often focuses on the overall net result—the sum of potentially positive and negative impacts—and whether climate change produces net costs or net benefits.

If the sum of direct and indirect market costs is sufficiently significant, it may impact macroeconomic indicators, such as consumer and producer price inflation, the unemployment rate and GDP. GDP measures the value of output in an economy, part of which reflects investment and part of which reflects consumption. The GDP impacts of climate change can be estimated directly using computable general equilibrium (CGE) models—large-scale numerical models that simulate the main economic interactions (e.g., those between different product markets) in an economy—or through supply and use tables available from Statistics Canada that capture all relevant direct and indirect market costs. Projected changes in macroeconomic indicators, like GDP, should be used only as a supplementary lens through which to view the economic consequences of climate change. Macroeconomic indicators capture the aggregated direct and indirect climate change impacts on the economy. Macroeconomic impacts, if estimated directly, should not be added to other estimates of direct and indirect market costs, as this would entail double counting (Ratti, 2017; Kousky, 2012). At the same time, focusing solely on aggregate macroeconomic indicators like GDP can be misleading from a distributional perspective. The spatial scale of an extreme weather event can be different from the scale over which GDP is measured. Significant losses for local populations may have no visible impact on national GDP, or even on provincial or territorial GDP. However, this does not imply that the impacts are negligible for the people who are affected. This particularly applies to disadvantaged populations or locations, whose economic output is generally invisible in aggregate macroeconomic indicators. A further distributional issue regarding the spatial scale of climate change impacts and the use of aggregate macroeconomic indicators is that losses at one location can be offset by gains at another.

The theoretically correct measure of the economic consequences of climate change is the resultant change in the welfare of affected populations (Kousky, 2012; Stern, 2006; Nordhaus and Boyer, 2000). Estimating changes in a theoretical metric like welfare is nonetheless difficult in practice. Consequently, GDP is often used as a practical, though far from perfect, proxy for welfare (Diaz and Moore, 2017a, b; Jones and Klenow, 2016). In addition to the aforementioned problems with using GDP to measure climate change costs, GDP only captures the value of impacts to market goods and services. As noted above, non-market impacts can be substantial—perhaps larger than market costs. Failure to account for non-market impacts will lead to seriously underestimating welfare losses. The output of the economy, as measured by GDP, also does not directly affect the welfare of individuals—what matters most to people is consumption and the loss of consumer surplus (Hallegatte, 2013). In the aftermath of extreme events, GDP can increase as the amount of investment increases to repair damaged assets—while this might suggest an increase in welfare based on the above definition, welfare will actually fall since households are foregoing consumption that they would otherwise have enjoyed in favour of investment. As a result, a more appropriate proxy for welfare costs resulting from climate change is a measure of consumption loss, rather than output loss as measured by GDP. Notwithstanding these concerns, the welfare costs of climate change in monetary terms are sometimes expressed as a percentage of projected GDP, referred to as a GDP-equivalent impact (Vivid Economic, 2013).

When making adaptation decisions, it is essential to consider another category of impacts that are important from an economic perspective and are referred to as “co-impacts”—these are more commonly referred to as “co-benefits” when the impacts are positive. In addition to applicable lifecycle costs and avoided climate-related damages, adaptation options can give rise to various ancillary impacts of potential significance, known as co-impacts (Chambwera et al., 2014). Recognizing co-impacts in adaptation decisions is important, as evidence suggests that people are more likely to act on climate change if the related impacts associated with specific actions are highlighted (Bain et al., 2015).

Many different terms are used with reference to co-impacts, depending on whether they are positive or negative, and intentional or unintentional (Floater et al., 2016; Urge-Vorsatz et al., 2014). Intentionality refers to the degree to which co-benefits are explicitly pursued by the decision maker, as early no-regret and low-regret adaptation options are prioritized to manage uncertainty (see Section 6.2.3). In addition to avoiding climate-related damages, adaptation options can contribute to GHG emissions reduction and other non-climate policy objectives related to issues such as economic development, public health, sustainability and equity. Avoiding climate-related damages can be the secondary objective of GHG emissions reduction or non-climate policies, or can serve as one of a number of objectives to be pursued simultaneously as part of a coherent, integrated package of policies (Floater et al., 2016). For example, the use of green roofs in cities as a strategy for reducing urban heat also helps to manage storm water, sequester carbon and improve urban biodiversity. In this case, using green infrastructure to address the adverse health effects of heat waves also contributes to the co-benefits of reducing GHG emissions, flood management, and the delivery of ecological services (see Case Story 6.1).

A range of terms are used to describe negative co-impacts—which are treated as unintentional—including co-costs, ancillary costs, adverse side-effects and externalities (Urge-Vorsatz et al., 2014). Examples of negative co-impacts generated by adaptation options would include increasing GHG emissions, increasing risks to other groups or sectors that are not targeted by the option, or limiting future adaptation choices.

The final set of economic terms commonly encountered in the literature relates to the perspective adopted by the decision maker for appraising adaptation options. The costs and benefits of adaptation options can be assessed from a social, as well as a private, perspective (Halsnæs et al., 2007). From a social perspective, where a public policymaker is looking for a socially optimal allocation of resources to climate change adaptation, the appraisal of adaptation options should consider co-benefits, as well as potential negative consequences alongside estimated lifecycle costs and avoided damages (Floater et al., 2016; Chambwera et al., 2014). In contrast, households and businesses will be interested in a narrower set of private costs and benefits when making adaptation decisions—specifically, those costs and benefits that accrue to the individual decision maker. These private costs and benefits (sometimes referred to as financial impacts) are typically based on actual market prices. To understand the importance of the difference between the two perspectives, consider, for example, a home damaged by a flood event, where some direct costs are reimbursed through the Government of Canada’s Disaster Financial Assistance Arrangements (DFAA) program. The private cost of the event to the homeowner is the difference between the repair costs incurred (not covered by private insurance) and the amount of aid received from the government. However, from the perspective of society, the aid represents a transfer payment from one taxpayer (a loss) to another (an equivalent gain); the loss and gain cancel each other out, leaving the full cost of repairs as a measure of the social cost of the flood event.

Prior to reviewing evidence of the projected economic consequences of climate change for Canada, information on the costs of past severe climate and weather events is presented for context. Extreme events—such as heat waves, drought, flooding or strong storms—have the potential to cause extensive damage and impacts to people, buildings, infrastructure and the natural environment; severe weather causes tens of billions of dollars of damage each year worldwide (Aon, 2020; Swiss Re Institute, 2020). It is anticipated that climate change will intensify some types of extreme weather events in the future (Bush and Lemmen, 2019) and will contribute to rising damages in the coming decades. As a result, an appreciation of current vulnerabilities and gaps in preparedness in Canada is a good starting point for building a robust case for early action for climate change adaptation (see Section 6.2.3).

This section focuses on a single line of evidence—damages associated with weather extremes in Canada, as documented by the insurance industry. While extreme events might be the face of climate change, gradual trends in Canada’s climate (e.g., rising mean annual and seasonal temperatures, rising sea levels, melting glaciers and permafrost, etc.) may also be leading to impacts with important economic consequences, including recent problems with mountain pine beetle infestations in B.C. (Withey et al., 2015) and the spread of Lyme disease vectors (Ebi et al., 2017). Compared to the impacts of extreme weather events, evidence of the economic consequences of these slow-onset impacts is sparse. The information presented in this section provides only a partial picture of the economic costs of past climate-related hazards in Canada, recognizing that current risks from weather extremes are significant and rising, and warrant early action.

The insurance industry is a key source of information on the economic consequences of weather extremes. Large reinsurers, such as Munich RE and Swiss Re, monitor and record information on losses from natural catastrophes globally to evaluate the capacity of national and international reinsurance markets to absorb losses (see Box 6.1 for a description of key insurance industry terminology).

In 2018, weather-related natural disasters globally caused total economic losses of about $215 billion (2018 dollars USD), of which private insurers paid out a record $100 billion in losses (Munich RE, 2020). The global protection gap—the difference between insured losses and total losses—was therefore $115 billion (54% of total economic losses). These figures are similar to those produced by the Swiss Re Institute for 2018; estimated total economic losses and insured losses from weather-related natural disasters globally were, respectively, about $201 billion and $98 billion (2018 dollars USD), making the protection gap about $103 billion (or 51% of total economic losses) (Swiss Re Institute, 2019a). Both economic losses and insured losses in 2018 were higher than the corresponding inflation-adjusted annual average for the last ten years (2008–2018), which are $174 billion and $65 billion, respectively (Munich RE, 2020).

Globally, economic losses from natural disasters are rising. Overall losses and insured losses have been trending upward over the last several decades; this is evident in Figure 6.4, which shows worldwide loss data for the period 1980–2018 as recorded by Munich RE, the world’s largest reinsurance company. Loss data recorded by the Swiss Re Institute also shows rising economic damages from weather-related natural disasters worldwide (Swiss Re Institute, 2019a). In terms of five-year moving averages, overall losses recorded by Munich RE grew by 5.1% annually between 1980 and 2018, and insured losses grew by 4.3% annually. With growth in overall economic losses outpacing insured losses, the protection gap has risen in absolute dollar terms over time, although the gap is falling in percentage terms. Despite increasing penetration of relevant insurance products with a greater proportion of damages covered by insurance (Swiss Re Institute, 2019a), society is absorbing increasing residual losses from weather-related natural disasters.

Economic losses from natural disasters affecting the U.S. are also rising (e.g., National Centers for Environmental Information, 2020). For example, the frequency of billion-dollar disasters between 1980 and 2011 increased at about 5% per year (Smith and Katz, 2013). During the period of 1980–2019, the U.S. experienced, on average, 6.6 events annually; over the most recent 5 year period (2015–2019), the annual average number of billion-dollar disasters was roughly double, at 13.8 events (National Centers for Environmental Information, 2020).

Public Safety Canada’s Canadian Disaster Database (CDD) monitors overall economic losses from significant meteorological and hydrological disasters, including payments made under the DFAA program (see below) and those made by private insurers. The Insurance Bureau of Canada (IBC) also tracks private insurance payouts for extreme weather events dating back to 1983. However, data on overall losses in the CDD seems incomplete considering that, in over half of the years since 1983, insured losses recorded by the IBC exceeded total economic losses in the CDD. Due to the incomplete nature of the economic loss data, the narrative below focuses only on insured losses. If the U.S. can be considered to provide a reasonable analogy for Canada, overall losses from weather extremes are roughly double the amount of the insured losses (Aon, 2020).

Insured losses in Canada have been rising since 1983, as is evident from the trend line in Figure 6.5. Between 1983 and 2007, annual losses averaged about $0.4 billion (2018 dollars); in contrast, over the most recent decade, losses have averaged about $1.9 billion per year (Insurance Bureau of Canada, 2018). The largest insured loss in a single year on record was $5.3 billion (2018 dollars) in 2016, with the wildfire in Fort McMurray and the surrounding area resulting in insurance payouts totalling $3.9 billion (Insurance Bureau of Canada, 2019).

Like insured losses, the number of extreme weather events has been increasing over time—the five-year moving average grew by about 7% annually between 1983 and 2018. In terms of the distribution of extreme weather events in Canada, Alberta was affected the most, with 55 events impacting the province over the period 1983‒2018, followed closely by Ontario, with 52 events. The Maritime provinces experienced the fewest events. Alberta is the epicentre of extreme weather events when it comes to losses—six of the ten largest insured loss events in Canada since 1983 occurred in this province (see Table 6.1).

As with payments from private insurers, the annual cost of the federal DFAA program has been rising since the 1970s (Office of the Auditor General Canada, 2016; Parliamentary Budget Officer, 2016). In the event of a disaster—and if the response and recovery costs exceed certain thresholds deemed acceptable for a province or territory to bear on its own—the Government of Canada provides financial assistance on a sliding scale through the DFAA program. Payments are made directly to provinces, which then distribute funds to individuals, businesses, non-profit organizations and local governments. Between 1970 and 1994, annual average DFAA payouts for hurricanes, convective storms, floods and winter storms averaged $56 million (2018 dollars) (Parliamentary Budget Officer, 2016). In contrast, annual average payouts averaged $303 and $427 million (2018 dollars) over the periods of 1995‒2004 and 2005‒2014, respectively (Parliamentary Budget Officer, 2016).

Long-term trends in losses have been interpreted as indicative of a contemporary adaptation deficit (Burton, 2009) and of increasing future climate-related risks (Hallegatte, 2014; Schipper and Pelling, 2006). There are many reasons for the observed adaptation deficit, including a range of market, behavioural and policy failures (see Section 6.8.1). The deficit has been increasing and is anticipated to widen with climate change (Burton, 2009), strengthening the case for early adaptation efforts.

The observed increase in losses from weather-related disasters has led to questions about whether climate change is contributing to the trend (Bouwer, 2011). The IPCC, for example, suggested that the upward trend in historic losses provided indirect evidence of a potential climate change signal. However, some scholars argued that to make reliable comparisons between the losses of past and more recent weather-related natural disasters, it is necessary to control for changes in various socioeconomic factors that influence the magnitude of losses; otherwise, one is comparing apples to oranges (e.g, Pielke, 2007). The process of making adjustments for relevant socioeconomic and non-climate related factors is known as “loss normalization” (Swiss Re Institute, 2020; Pielke et al., 2003). Normalization helps to address the following question: “What would the magnitude of losses be if present-day assets and values were exposed to a historic event?” Analyses of normalized losses help to clarify the extent to which socioeconomic factors contribute to observed rising damages over time and, by inference, the role of other factors in shaping loss trends. One likely factor contributing to the trend of rising losses is improved and more comprehensive data collection over time; similarly, lower observed losses in the early 1980s may partly be explained by a lack of available data.

Studies that analyze time series of normalized economic and insured losses from weather-related disasters—whether they occur at a global or regional level—generate mixed results. Some studies find no significant upward trends, despite substantial increases in nominal losses (e.g., Bouwer, 2011; Neumayer and Barthel, 2011). Other studies find statistically significant long-term trends in losses (e.g., Gall et al., 2011; Schmidt et al., 2009). A recent study of normalized natural disaster losses globally uncovered strong evidence of a rightward skewing of the loss distribution and a corresponding increasing trend in the most extreme damages, but weaker evidence for an increasing trend in mean losses (Coronesea et al., 2019).

Figure 6.6 presents normalized insured losses for the same extreme weather events in Canada as shown in Figure 6.5. Losses are normalized using a conventional approach (e.g., Miller et al., 2008; Pielke et al., 2008; Pielke et al., 2003): the original nominal values, not adjusted for inflation, are modified by three multipliers to account for changes in producer prices, population and wealth, which are measured in terms of GDP per capita over time. A significant, though slightly weaker, positive trend is still observed in normalized losses from weather-related disasters in Canada. This upward trend is also evident when considering the five-year moving average of normalized losses, which produced an annual growth rate of 3.5% between 1983 and 2018; this rate was still increasing, but at a much slower rate than that of nominal losses (10.6%) and real losses (8.2%), adjusted for inflation, over the same period. Since 1983, the increasing trend in insured losses associated with extreme weather disasters in Canada has primarily been due to an accumulation of value (e.g., people, assets, wealth) year-on-year. But rising losses cannot entirely be explained by growing exposures, asset values and general price inflation—climate change may be playing a role. While a rise in normalized losses is not “proof” of climate change, it is certainly consistent with the anticipated intensification of some weather extremes with rising temperatures (Swiss Re Institute, 2020; Coronesea et al., 2019; IPCC, 2013).

Research on event attribution in Canada—which assesses how the likelihood of extreme weather events is altered by GHG emissions from human activity—found that climate change increased the likelihood of the 2016 Fort McMurray wildfire and the extreme rainfall that contributed to the 2013 flooding in southern Alberta (Zhang et al., 2019), two of the most costly weather disasters on record in Canada (see Prairie Provinces chapter).

A key piece of economic information supporting climate-related decisions is the future cost of inaction—the economic consequences that result from allowing climate change to continue unabated and without further planned adaptation (Ackerman and Stanton, 2011; European Environment Agency, 2007). While the magnitude of the projected cost of inaction is uncertain, judiciously caveated estimates can be used in tandem with information on the current costs of climate and weather extremes (see Section 6.4) to persuade decision makers of the urgent need to allocate resources for adaptation, and to prioritize the allocation of such resources to address key climate risks. Projected costs also provide a baseline for weighing the cost of adaptation projects, programs and policies. To supplement the observed cost information for historic extreme weather events presented in Section 6.4, this section reviews available evidence relating to the projected future costs of climate change for Canada: looking at aggregate, multi-sector and national-level cost assessments; cost assessments for single climate-sensitive sectors at the national, regional and provincial level; and cost assessments for specific municipalities.

Aggregate estimates of the economic consequences of climate change strictly for Canada and across multiple sectors are scarce. The frequently cited study by the NRTEE (2011) remains the benchmark multi-sector national assessment of the economic costs of climate change for Canada. Using the integrated assessment model, PAGE09 (Hope, 2011), the future economic cost of climate change for Canada was estimated for two climate and two socioeconomic scenarios, producing four plausible futures: 1) “low climate change–slow growth”, 2) “low climate change–rapid growth”, 3) “high climate change–slow growth” and 4) “high climate change–rapid growth”.² Projected annual costs for Canada in 2050—assuming no new adaptations—range from $30 billion (2019 dollars) under the “low climate change–slow growth” scenario to $62 billion under the “high climate change–rapid growth” scenario. Under the “high climate change–rapid growth” scenario, more people, assets and wealth are exposed to a larger temperature change than under the “low climate change–slow growth” scenario, resulting in larger projected costs. The PAGE09 model explicitly captures uncertainty in its parameters, which generates a frequency distribution of estimated annual costs. The distributions of possible costs across all scenarios suggest a small chance that costs could be much higher—there is a 5% chance that the annual cost of climate change in 2050 could exceed $131 billion under the “high climate change–rapid growth” scenario. By 2075, under the same four plausible scenario combinations, annual costs are projected to range from $74 to $319 billion, with a 5% chance that they could exceed $1,185 billion annually under the “high climate change–rapid growth” scenario. These projections reflect the undiscounted expected costs to traditional economic sectors (e.g., construction, manufacturing, retail trade, educational services, etc.) and non-economic sectors (e.g., impacts to health and ecosystems) from warming, the expected costs of sea-level rise and the expected costs of “fat-tail” catastrophic events³ (e.g., from rapid melting of the Greenland and West Antarctic ice sheets) (NRTEE, 2011).

The only other multi-sector national estimates of the aggregate impact of climate change on Canada come from large-scale global macroeconomic studies. Results for Canada from three of these studies are presented in Table 6.2. All three studies use similar approaches, which involve integrating information about biophysical impacts and economic valuation derived from independent damage assessments for specific sectors (like agriculture) into a multi-regional, multi-sector model of the global economy to estimate the impact of climate change on economic output (GDP). The studies include similar sector-specific impact categories (see Table 6.2) and draw damage information to inform the magnitude of projected impacts largely from the same set of primary studies (e.g., Lafakis et al., 2019; Kompass et al., 2018; Roson and Sartori, 2016; Kjellstrom et al., 2009; Bosello et al., 2006; Hamilton et al., 2005). Despite these similarities, the results are not strictly comparable due to, among other things, different time horizons and modelled temperature changes, and differences in how the independently estimated climate impacts are integrated into each macroeconomic model.

All three studies suggest that projected climate change impacts on annual real Canadian GDP will be less than (plus or minus) 1%. Though small in percentage terms from a national perspective, this still equates to very large dollar amounts, which will be significant for affected Canadian sectors and regions. Both the Moody’s Analytics study and the OECD study project small net gains for the Canadian economy. In each case, the dominant driver behind the result is projected increases in tourism flows, with fewer domestic departures and more international arrivals expected as Canada warms. These gains more than offset output losses attributable to the other climate impacts that were analyzed. However, the projected increases in tourism flows for Canada should be viewed with caution, due to the simplified and aggregate nature of the underlying impact model, in which tourist flows only depend on mean annual temperature, per capita incomes, an attractiveness index, and the distance between the capitals of origin and destination countries (Hamilton et al., 2005). Other key determinants of future tourism flows are ignored, including the influence of changing precipitation patterns, supply-side constraints on the availability of tourist infrastructure, and feedback effects on international arrivals due to reduced incomes in origin countries related to climate change impacts on other sectors of the global economy. Kompass et al. (2018) deliberately excluded climate change damage functions for tourism from their analysis because of these concerns, and they projected small net losses for the Canadian economy related to climate change.

Estimates of the aggregate impact of climate change on the Canadian economy are available from another global macroeconomic study. Kahn et al. (2019) used a “top-down” empirical approach—which estimates a relationship between a climate variable (e.g., temperature) and an aggregate measure of economic output for the whole economy (e.g., national GDP)—to investigate the impact of climate change on long-term economic growth across 174 countries, including Canada. To put the results of this study into perspective, the PAGE09 integrated assessment model used by the NRTEE measures climate change impacts on GDP levels, not the GDP growth rate (i.e., it measures short-term growth effects). However, climate change can cause lasting damage to capital stocks and productivity in most sectors of the economy, and is likely to impact long-term growth rates (Revesz et al., 2014; Stern, 2013). In this case, output and consumption at some future date will not depend solely on the temperature at that date, but is more likely to be affected by the entire path of temperature, output and consumption up to that date. Studies that have investigated the impact of climate change on GDP growth rates have found substantially larger losses than studies that measured impacts on the annual level of GDP (Diaz and Moore, 2017a, b). The assessments summarized in Appendix 6.1 incorporated a mix of both effects—some of the sector-specific impacts that were analyzed affected GDP growth rates (e.g., damage from sea-level rise to capital stocks), while others affected output levels (e.g., changes to crop yields).

Kahn et al. (2019) used a model linking mean annual temperature deviations from historic norms over the period 1960–2014 to changes in labour productivity (see Case Story 6.2), and in turn to real GDP per capita. The model was used to investigate the cumulative effect of changes in labour productivity resulting from persistent increases in mean annual temperature under RCP8.5. In contrast to the projections presented in Appendix 6.1, Kahn et al. (2019) found that climate change would substantially reduce real GDP per capita in Canada in 2050 and 2100 by 4.4% and 13.1%, respectively.

The global macroeconomic studies discussed above have several limitations, meaning that the projected impacts of climate change for Canada are likely overwhelmingly negative and much higher than shown. For a start, many important impacts were omitted, including impacts on livestock and aquaculture, changes to forestry yields, impacts associated with the spread of invasive species and pests, impacts of water stress on electricity production and the availability of potable water for end users, impacts on human security (e.g., conflict and migration) and impacts on the provision of ecosystem services. Furthermore, macroeconomic studies cannot capture non-market impacts, such as welfare losses from impacts to cultural ecosystem services, premature mortality, or pain and suffering from illness or injury. The economic consequences of extreme weather events were also not covered, with the exception of hurricanes in the OECD study. The costs of these disasters can be considerable (see Sections 6.4, 6.5.2 and 6.5.3). Finally, large scale disruptive or “fat-tail” catastrophic events were not captured.

While only a few studies provide aggregate cost estimates across multiple sectors for Canada, many more studies have investigated the economic consequences of climate change for individual climate-sensitive sectors (e.g., forestry, agriculture, coastal areas). Appendix 6.1 provides a summary of these studies and results, organized by climate-sensitive sector.

It is nearly impossible to compare the relative magnitude and significance of estimated economic consequences between climate-sensitive sectors, or even within a sector, due to differences in assumptions and methodologies across studies. In addition to differences in geographical scope, key differences between studies that influence the results relate to:

• The choice of emissions scenario(s) driving the biophysical impacts and, relatedly, the future time period(s) and climate norms used to measure changes in relevant climate variables. Looking at coastal areas, for example, some studies use the old IPCC Special Report on Emissions Scenarios (SRES), whereas others use the Representative Concentration Pathways (RCPs). This affects the time horizon over which streams of losses or gains are aggregated (e.g., the studies in Appendix 6.1 assessing the same coastal area sites use three different time periods: 2009–2054, 2015–2064 and 2011–2100).
• Assumptions about future socioeconomic developments, which will influence both the quantity and monetary valuation of buildings, infrastructure, crops, etc. that are affected under the assumed baseline and against which the impacts of projected climate change are assessed. Some of the studies examined the impact of future climate change on the sector today (based on current or historical information), while others—such as the NRTEE (2011) forestry and coastal area studies—investigated the impact of future climate change on future projections for the sector.
• Whether one or more biophysical impacts are considered. For example, some studies that focused on coastal areas included erosion and flooding from sea-level rise and storm surge, while others only considered impacts from flooding.
• The types of economic consequences resulting from the biophysical impacts included in the analysis. Most studies considered only direct impacts (e.g., damage-related costs, changes in agricultural land values, foregone ski revenues, increases in ski resort operating costs or fire suppression costs), while some studies also assessed indirect and macroeconomic impacts. The more recent studies measured economic consequences in terms of changes in projected GDP or welfare, which were estimated using CGE models. Furthermore, some studies that measured only direct impacts considered both market and non-market impacts (e.g., coastal area studies for Quebec), while others included only the former.
• The choice of economic modelling tool, specifically with respect to agriculture. Most agricultural studies estimate using a Ricardian model (an approach that estimates an empirical relationship between land values and climate variables) to measure the economic consequences of climate change on farmland values, while some use CGE models. As noted above, CGE models measure direct and indirect impacts, and account for market-driven behavioural (price) responses throughout the economy; Ricardian models capture only direct impacts and typically assume that prices are fixed. Results from both modelling approaches applied to agriculture are not comparable.
• Whether the economic consequences are measured in current (nominal) dollars or constant (real) dollars, and which base year is selected in the latter case (e.g., coastal area studies measured costs in 2000, 2008 and 2012 constant dollars). This is less of an issue, though, since it is possible to express all results in the prices of a common base year.
• The choice of discount rate, where results are reported as the present value of cumulative losses or gains over a defined time horizon (e.g., some studies use a real annual discount rate of 3%, whereas others use 4%). A higher discount rate will produce lower present-value costs or benefits, and a lower discount rate will produce the opposite. For reference, the value of a $1 cost (in 2020) incurred in the 2080s is $0.15 and $0.08 at a discount rate of 3% and 4%, respectively. Some studies also present undiscounted costs or benefits. Discounting and discount rates are discussed in Section 6.6.3.2.

While the above factors make it difficult to arrive at firm conclusions about the magnitude of economic impacts, it is possible to draw conclusions about the direction of projected impacts for all sectors studied, except for agriculture. The available evidence (based on the studies listed in Appendix 6.1) suggests that climate change will have predominantly negative economic consequences for forestry, coastal areas, human health in Quebec, low water levels in the Great Lakes and St. Lawrence River region, ski resorts in British Columbia and Quebec, and ice-based access roads in the Northwest Territories. Below are key observations across this series of studies, by climate-sensitive sector.

Projections of the economic consequences of climate change have also been made for individual cities in Canada. Appendix 6.2 provides a summary of key city-specific studies, which vary markedly in scope, methods, emissions scenarios, socioeconomic assumptions, time horizons considered and measures of economic impact, making comparisons difficult. The small number of studies also makes it difficult to draw assured conclusions. Nevertheless, initial observations from the available evidence suggest that the net economic consequences of climate change for cities are projected to be negative.

It is also evident from the studies in Appendix 6.2 that the scope of the analysis—the number of climate hazards, biophysical impacts and exposures considered, and whether the economic impacts include both direct and indirect impacts, as well as market and non-market impacts—is an important determinant of the magnitude of projected costs. For example, non-market impacts accounted for 23‒42% of total flooding costs in Fredericton due to climate change (Lantz et al., 2012). The omission of non-market impacts from these analyses would result in a significant underestimate of projected climate-related costs. The scope of the City of Edmonton analysis (see Case Story 6.3) was more comprehensive than the other city studies and included many more climate-related impacts (17 rapid-onset impacts alongside changes in heating and cooling degree days) on a broader inventory of exposed buildings, assets, infrastructure and services, as well as non-market economic impacts on human health and the natural environment. This more expansive scope also included direct and indirect impacts, explaining why the projected climate-related costs found by the study are relatively high; expected annual GDP costs are $1.6 billion in 2055 and $3.5 billion in 2085 (both in 2016 dollars), which is equivalent to 1.6% and 1.9%, respectively, of the city’s projected GDP.

All studies employed bottom-up, process-based modelling approaches. The City of Edmonton and the Cities of Halifax and Mississauga studies also adopted good practices, determining incremental losses attributable to climate change with and without socioeconomic development relative to today. This allows for an analysis of how socioeconomic development in relation to climate change contributes to projected economic costs, as well as an analysis that isolates the fraction attributable solely to climate change. Consistent with observations from the multi-sector regional studies discussed in Section 6.5.3, growth in the “stock at risk” to climate change and rising valuations of that stock can be an equally important determinant of future economic costs as climate change itself (see Case Story 6.3).

In addition to seeking evidence on the economic consequences of climate change, decision makers are increasingly requesting information on the costs, benefits and key trade-offs of actions to support their adaptation decisions. This section provides a review of economic analysis tools to support adaptation decision making and will provide context for the next section, which reviews the application of these tools in Canada. Prior to an examination of the economic decision support tools, common evaluation criteria are introduced to highlight the trade-offs that decision makers often consider in making adaptation decisions.

Decision makers bring diverse objectives, interests, knowledge and values to climate change adaptation decisions. This results in a diverse range of decision criteria to consider as various courses of action are weighed. The literature contains many decision criteria and groupings of those criteria to support the appraisal of adaptation actions and their implementation (e.g., Rouillard et al., 2016b; Weiland and Tröltzsch, 2015; PROVIA, 2013; United Nations Environment Programme, 2011). These reviews conclude that the appraisal of adaptation actions should ideally capture trade-offs between all relevant outputs (benefits) and all relevant inputs (costs) that are needed to deliver those outputs. Two further important considerations relate to uncertainty surrounding the anticipated outputs and the ease with which an adaptation action can be successfully implemented, which will also affect costs. Based on these reviews, the main decision criteria typically used to assess the relative merit of investment in adaptation actions are described in Table 6.3.

There are multiple methods for appraising adaptation actions. The standard analytical technique used for the economic appraisal of policies, programs and projects is cost-benefit analysis (CBA) (see Figure 6.9). CBA is a suitable method for economic appraisal when the adaptation goal is to minimize the economic costs of a climate-related threat or to maximize the economic benefits of a climate-related opportunity. In some decision contexts, however, the adaptation goal might be to achieve a given level of risk reduction or to cancel out all adverse climate-related impacts (i.e., to maintain baseline conditions) (Chambwera et al., 2014). In this context, the other main economic appraisal method, cost-effectiveness analysis (CEA), could be used to identify the action or portfolio of actions necessary to achieve this goal at the lowest cost or the greatest level of risk reduction for a fixed investment budget.

A third traditional method that can be used to appraise adaptation actions is multi-criteria decision analysis (MCDA). Though it is not technically an economic appraisal tool, it can accommodate monetized costs and benefits information in the decision calculus, alongside a range of other decision criteria. A decision maker may want to use MCDA when economic efficiency is not the sole decision criterion of interest or when important inputs to, or outcomes of, the adaptation action cannot be valued in monetary terms. Given the multiple criteria now being considered when making adaptation decisions, increasing emphasis is being placed on such “multi-metric” appraisal tools to provide support for decision makers (see Table 6.4; Chambwera et al., 2014).⁴ Only CBA, however, has been applied in the available literature for Canada (see Section 6.7).

Key methodological challenges related to the economic appraisal of adaptation actions include how to handle uncertainty in economic appraisal, discounting choices and distributional considerations. Because of these challenges, the literature is skeptical about relying mainly on traditional economic appraisal tools to rank adaptation actions (e.g., Dennig, 2018; Lempert, 2014; Li et al, 2014).

The body of literature on the appraisal of adaptation costs and benefits in Canada is limited to the public sector and a few climate-sensitive sectors; it therefore covers a narrow range of climate change impacts, regions and potential adaptation actions. This makes it difficult to form widespread generalizations about the costs and economic attractiveness of adaptation actions in all contexts. This section reviews the application in Canada of the methods discussed in Section 6.6.

To help draw conclusions regarding the economic case for climate change adaptation from the available studies listed in Appendix 6.4, adaptation actions where a benefit-cost ratio (BCR) was reported or could be derived (60 actions in total) are rank-ordered and presented in Figure 6.16. A BCR is given based on the present value of benefits of an adaptation action, divided by its present value costs. As such, it controls for the scale of adaptation actions, thereby facilitating comparisons across actions of different size (other factors limiting comparisons are discussed below). A BCR greater than one indicates that an adaptation action’s benefits exceed the costs incurred to generate those benefits—such an action would typically be a justifiable investment on economic efficiency grounds. However, not all actions with a BCR greater than one are necessarily implemented, as their implementation depends on a range of factors, including available resources (see Section 6.8.1).

Figure 6.16 shows that, across the 60 adaptation actions, 75% pass a cost-benefit test. The unweighted average BCR is 5.6 (i.e., every dollar invested in climate change adaptation actions generates, on average, $5.60 in benefits). The average BCR is heavily skewed by a handful of extremely high values, however. The unweighted median BCR is 1.5, and half of the values lie between 0.9 and 2.7. These values are consistent with international experience—for example, in a review of a statistical sample of the nearly 5,500 Federal Emergency Management Agency (FEMA) grants in the U.S. awarded between 1993 and 2003 for addressing earthquake, flood and wind hazards, Rose et al. (2007) found that the overall average BCR ratio was about 4.0, whereas the average BCR for flood reduction measures was 5.1. Likewise, the Global Commission on Adaptation found that $1 judiciously invested in climate change adaptation could generate $2–10 in economic benefits (Global Commission on Adaptation, 2019).

The available examples from the 60 adaptation actions suggest that “soft” adaptation actions (with a mean BCR of more than 10:1) represent more economically efficient investments than “hard engineering” adaptation actions (with a mean BCR of about 3:1). This is largely due to the higher upfront investment expenditures needed for the latter set of actions. It is also partly due to the inclusion of monetized co-benefits generated by some of the “soft” actions (specifically in the appraisal of coastal adaptation actions in Quebec) and to the inclusion of both direct and indirect benefits (i.e., avoided costs) in a couple of studies (e.g., timber supply) that examined only “soft” adaptation actions. As noted in Section 6.6.1, the economic performance of a project is only one of several important criteria for selecting adaptation options. In some cases, “soft” options may not offer an acceptable level of risk reduction, thus necessitating the adoption of “hard” options, which may have less attractive BCRs.

In general, the diversity of methodological choices makes comparing results across available appraisals of adaptation costs and benefits difficult. Studies use different time horizons, emissions scenarios and norms; they make different assumptions about socioeconomic development, and monetize different combinations of market and non-market impacts, co-benefits, and direct and indirect impacts (see Case Story 6.6). They also use different discount rates, though this is less of an issue regarding the comparability of studies, since nearly all apply constant rates of 3–4% per annum, which is indicative of a prescriptive rather than a descriptive approach to discounting (see Section 6.6.3.2).

It is evident from the NRTEE (2011) studies in Appendix 6.4 that, even with adaptation actions in place, residual costs from damage due to climate change remain. For instance, the cost of adopting a portfolio of adaptation actions to address the impacts of climate change on timber supply is $2.3–3.6 billion (in present value terms, at a 3% constant discount rate over the period 2010‒2080) and reduces economic losses by $19.9‒137.9 billion, but residual losses of $4.6‒37.1 billion remain. The total costs of climate change in this case are therefore $6.9‒40.7 billion. While not explicitly reported by the other studies listed in Appendix 6.4, a quick comparison of the results in Appendix 6.1 and Appendix 6.2 with those in Appendix 6.4 reveals the presence of residual damage costs in most cases, suggesting potential—though undefined—limits to what adaptation can achieve (see Section 6.8.1). The presence of residual costs does not indicate that an action is poorly designed or that an insufficient level of adaptation was implemented; it may simply indicate that aiming for zero residual damages is not feasible or that its costs would exceed the dollar value of avoided damages.

Both theory and evidence indicate that adaptation cannot cancel out all negative climate change impacts, nor can it capture all positive impacts (Chambwera et al., 2014; Dow et al., 2013). Earlier sections in this chapter highlight potential instances of insufficient or ineffective adaptation (e.g., the current adaptation deficit with respect to extreme weather events in Canada) (see Section 6.4.2). This section examines barriers and limits to adaptation from an economic perspective. A barrier refers to any type of challenge, obstacle or constraint that can impede or stop the adoption of certain adaptation actions by businesses or households, but that is surmountable with concerted effort; a limit is a constraint that cannot be overcome without incurring unreasonable costs or taking unreasonable action (Eisenack et al., 2014; Productivity Commission, 2012).

From an economic perspective, private actors such as businesses and households are expected to undertake a significant amount of adaptation as they modify decisions and behaviours in response to climate signals to maximize their profit or welfare (Mendelsohn, 2012). Such behavioural reactions to climate stimuli form the premise underlying what is referred to as autonomous adaptation (Fankhauser, 2017). For example, people adjust their vacation destinations or travel dates in response to climate (Hamilton et al., 2005), and farmers adjust crops or use different harvest or seeding dates in response to changing precipitation patterns (Food and Agriculture Organization, 2007).

However, there is evidence that autonomous adaptation by businesses and individuals is not always adequate or efficient (Eisenack et al., 2014; Klein et al., 2014; Porter et al., 2014; de Bruin et al., 2011; Agrawala et al., 2010). There are multiple barriers and limits to private adaptation (Biesbroek et al., 2013), which means that only a subset of adaptation needs may be met in practice (see Figure 6.17). Limits can be technological (e.g., snow-making equipment may not be able to sustain adequate snow cover at lower altitude ski resorts as the climate warms), ecological (i.e., ecosystems and species may be unable to adapt at higher rates of warming), economic (i.e., the level of adaptation that is justified on economic grounds once the lifecycle costs of actions have been considered, in relation to the projected benefits), and institutional (e.g., available funding and capacity) (Chambwera et al., 2014). The gap between the level of adaptation required to cancel out all negative impacts (or capture benefits from all opportunities) and the maximum potential for adaptation after taking into account technological and ecological limits is referred to as the “unavoidable impacts” (Chambwera et al., 2014). However, not all actions for overcoming avoidable impacts will pass a basic cost-benefit test, which is indicative of the economic potential for adaptation. The lifecycle costs of some adaptation actions will exceed the economic costs averted, indicating that alternative investments offer better value for money.

The presence of market, behavioural and policy failures means that the economic potential for adaptation is not fully realized. This creates a key role for government (Fankhauser, 2017):

• Firstly, to remove policy distortions that impede economically efficient adaptation choices by private actors: for example, to reform (e.g., reduce, restructure or eliminate) the subsidies that fuel the self-reinforcing cycle of continued growth in coastal or riverfront areas that are prone to flooding (Boyd et al., 2015).
• Secondly, to use regulatory and economic instruments to overcome market and behavioural failures, and to provide incentives for efficient private adaptation (e.g., Boyd et al., 2015; Hotte and Nelson, 2015). Regarding the use of economic instruments to incentivize adaptation, it is important for the design of these instruments to account for common behavioural failures that have the potential to undermine their effectiveness (Boyd et al. 2015).
• Thirdly, to provide public goods and services dedicated to adaptation, like the production and dissemination of climate information, spending on research and monitoring programs, investment in large-scale flood protection, early warning systems for communities, improvements to emergency planning and preparedness, and the development of policies to enhance the resilience of ecosystems.

However, not all forms of government intervention will make sense from an economic perspective. It is also necessary to demonstrate that the benefits arising from such interventions exceed the costs of implementation for private actors and government (Productivity Commission, 2012). Only a certain level of adaptation is achievable after accounting for the effectiveness of regulatory and economic instruments to redress barriers to efficient adaptation, with these instruments themselves having passed a cost-benefit test (see Figure 6.11).

Some individuals, businesses or communities may be unable to afford or finance the required investment in planned adaptation actions, even though they know it is in their best interest to do so (Lecocq and Shalizi, 2007).⁹ Another role for government is to aid vulnerable and disadvantaged groups and communities that do not have access to the necessary resources to adapt sufficiently (Fankhauser, 2017). At the same time, governments too will face financial and capacity constraints, and must allocate resources among competing needs. When an economically efficient level of adaptation is achieved after allowing for technical, social and ecological constraints, residual damages may well occur. The fact that some level of residual damages may be unavoidable gives rise to a range of important ethical and social justice issues that are at the core of the “loss and damage” discourse at the international level—referring to unavoidable impacts beyond the limits of adaptation (van der Geest and Warner, 2015). While a discussion of these issues is outside the scope of this chapter (see Wallimann-Helmer et al., 2019 for an overview of the main ethical and justice challenges), government may also have a role in defining what is an acceptable level of residual damage and how best to reconcile the welfare effects of these unavoidable impacts.

There is much that is yet to be known about the costs of climate change for Canada, both in aggregate and for specific sectors, regions, communities and vulnerable populations. Future projections of the total economic consequences of climate change for Canada are highly uncertain. Some simplified, highly aggregate modelling exercises project net gains for Canada’s economy, while others project net losses. Further study is needed to resolve uncertainty around the aggregate cost of climate change for Canada.

Adaptation decisions are largely made at the local or provincial level, where the current state of knowledge regarding the cost of inaction is highly fragmented. There are large knowledge gaps when it comes to the Prairie provinces, the Northwest Territories, the Yukon and Nunavut, the Interior of British Columbia, Ontario, as well as First Nations, Inuit and Métis peoples. Furthermore, high-quality estimates of economic consequences exist for only a few Canadian cities. Given that most adaptation decision making takes place at the local level, a priority for future research should be not only to resolve uncertainty around the total economic consequences for Canada in aggregate, but to improve the geographical coverage and scope of damage estimates for municipalities, as well as the level of disaggregation by sectors, assets and services, and climate hazards. This implies the need for a bottom-up, multi-sector approach that addresses several cross-cutting gaps in the current literature. Recommendations for new economic studies include the following:

Studies considering a broader range of climate hazards: Most of the available national-level aggregate projections (and most regional projections) are focused on slow-onset climate impacts (i.e., gradual changes in temperature and precipitation, and select biophysical impacts that result from these changes). Future investigations of economic consequences would benefit from increased attention on extreme events and catastrophes (i.e., low-probability and high-consequence events).

Studies considering a broader range of climate-sensitive sectors: Some sectors are better represented in the current economic literature than others. A range of estimates are available for coastal zones, agriculture and forestry. For other sectors—namely, tourism, labour, water resources and public health—only a few incomplete estimates are available. There are also major gaps in our understanding of the economic consequences of climate change for public health. Other sectors are not yet represented in the literature, such as ecosystems, fisheries, energy infrastructure (including oil and gas), transportation infrastructure (including rail, air and ports), water quality and security (e.g., crime, migration, conflict).

Studies considering a broader range of economic impacts: A comprehensive assessment of economic consequences would capture both market and non-market impacts. An important consequence of climate change for welfare is the loss of goods and services that are not traded in markets and therefore cannot be valued using market prices or captured by CGE models. Examples of broader economic impacts worthy of study include species loss, pain and discomfort, loss of cultural heritage, conflict and forced migration. These welfare losses can be sizeable, even though they are largely omitted from current estimates. Research is needed to ensure that they are better represented in future estimates of economic impacts.

Studies considering inter-sectoral impacts: There is a broad range of potentially important inter-sectoral impacts that are not well captured, especially within a bottom-up, multi-sector approach. For example, water is used to produce electricity (e.g., for thermal cooling), and electricity is used to supply water (e.g., to operate pump stations). These linkages are typically omitted from estimates. Some non-biophysical interactions occur through market mechanisms and can be captured using CGE models, for instance. Other interactions do not function in this manner, such as when damage to ecosystems amplifies other impacts. Research is needed to understand which inter-sectoral linkages are economically significant at a local or regional level, and should therefore be captured in the next generation of estimates.

Studies considering socioeconomic developments: An important conclusion from the current literature concerns the importance of future socioeconomic change (e.g., growth in populations, assets and wealth) as a key driver of the absolute magnitude of projected economic costs. Despite the demonstrable role of such change as a determinant of the cost of inaction, socioeconomic futures are either incompletely addressed or not addressed at all in many current studies.

Knowledge relating to the appraisal of adaptation costs and benefits in Canada is currently restricted in scope to a few climate-sensitive sectors, which in turn means that only a narrow range of adaptations to a limited set of climate impacts in specific regions has been considered to date. Also, existing studies are almost exclusively focused on the public sector. Consequently, despite the promising results from existing studies (see Section 6.7.1), it is not possible to make widespread generalizations about the economic attractiveness of adaptation actions in all contexts. There is a lot to learn about the costs and benefits of the full range of adaptation that is likely needed to manage the impacts of climate change regarding tolerable levels. Research is needed to understand more about the economic efficiency of capacity building actions and public policy interventions to overcome barriers to adaptation. This includes understanding how lessons from behavioural economics can be used to improve the design and effectiveness of policies to provide incentive for implementing desirable levels of private adaptation. At the same time, a better understanding of current public policies that promote maladaptation is needed; removing prevailing policy failures is crucial if interventions to incent adaptation are to be effective.

While theory favours short-lived, flexible and relatively inexpensive “soft” adaptation measures over long-lived, capital-intensive, “hard” adaptation measures in the face of deep uncertainties, it has yet to be demonstrated in Canada through practical applications which adaptations have the greatest merits, and under what circumstances. Case studies are needed to better understand the economic merits of sequencing adaptation decisions over time under multiple futures, rather than making a single, seemingly optimal decision now. All of the current “proof of principle” examples are international.

Current economic appraisals pay scant attention to distributional issues and to the political economy of adaptation (i.e., how adaptation decisions are made, taking into account political, cultural and economic factors). Adaptation, like any form of intervention, will typically have winners and losers, although none of the economic studies that have been formally reviewed considered the distribution of costs and benefits across actors. Since distributional impacts are a major talking point in local, provincial and national debates about climate policy, a sounder understanding of these impacts would aid in both the design of adaptation actions and in moving towards implementation.

Finally, awareness of the cost of climate change adaptation in Canada is only starting to develop, helped by two recent studies of the level of investment needed to adapt public infrastructure to climate change at the national level (Insurance Bureau of Canada and Federation of Canadian Municipalities, 2020) and in Quebec (AGECO Group, 2019). However, many knowledge gaps remain. For instance, there is no information on the aggregate level of investment needed to adapt other economic sectors for anticipated climate change impacts. Even regarding public infrastructure, there is poor understanding of adaptation investment needs for certain parts of the country (e.g., British Columbia and Nunavut) and for larger population centres. These knowledge gaps make it difficult to characterize the scale of the required adaptation effort, how it should be financed and—in conjunction with estimates of adaptation benefits—how available funds should be deployed.

The framing of adaptation decision making is changing, with implications for adaptation economics. Whereas the predominant approach to navigating the assessment and planning stages of an adaptive risk management framework was historically based on a “science-first” (or “top-down”) approach, there has been a recent shift in the economic literature towards a “policy-first” (or “bottom-up”) analytical process, with a focus on early action (see Section 6.2.3).

This shift has significant implications for the economic analysis of adaptation actions and necessitates the development and application of alternative decision support tools. Where consideration of deep uncertainties over future impacts is important—and where decision makers are looking for flexible or robust options—new economic decision support tools like adaptation pathways, real options analysis, robust decision making and portfolio analysis are more appropriate for economic appraisal than conventional tools like cost-benefit analysis (see Section 6.2.5).

The greater importance placed on capacity building, behavioural interventions and the value of information under the “policy-first” approach also creates challenges for the monetization of costs and benefits, requiring different approaches to the quantification of physical impacts and their subsequent valuation. Increased consideration of the adaptation process also places greater emphasis on understanding barriers and economic limits to efficient adaptation (e.g., market, behavioural and policy failures), and on the costs and benefits of government interventions designed to overcome these barriers (see Section 6.8). Designing effective policy interventions requires an understanding of behavioural responses to different incentives. In short, economic decision support is itself adapting to meet the evolving needs of decision makers.

There is growing recognition that an efficient level of adaptation is being hampered by more than issues of affordability. A combination of market failures (e.g., lack of quality, accessible information on relevant risks and adaptation responses, or the presence of public goods or externalities), behavioural anomalies (e.g., cognitive capacity, inertia, high discount rates) and prevailing policy distortions (e.g., subsidies that ultimately promote maladaptation) limit the potential for adaptation (see Section 6.8). There is a greater role for government at all levels to do more than provide financial assistance and invest in public goods (such as climate information services). Other important steps to take include removing prevailing policy distortions, and designing and implementing regulations and economic instruments to overcome relevant market imperfections and behavioural failures. Equally important is the need for governments to reflect on what would be an acceptable level of residual damage and how best to address the welfare effects of unavoidable impacts, given the potential for significant ethical and social justice concerns.

Another talking point in the economic literature is the extent to which the economic consequences of climate change could be much higher than current projections suggest—not because of the limitations of emissions and climate change impact models, which omit important risks, but because of how economic models treat damages and growth. The issue is whether the level of economic output is either reduced by a climate shock or stress, but with the underlying rate of economic growth being unaffected, or whether climate change has a more persistent, cumulative impact on the growth rate itself. Until recently, most estimates of the cost of climate change were based on static losses of annual economic output. However, if climate change causes lasting damage to capital stock, land and the efficiency at which these factors and labour are turned into economic output, as some scholars suggest, then the annual growth rate will be affected in addition to the output level, leading to much deeper and longer-lasting impacts on economic output, due to the compounding effects of reduced growth. The debate remains unsettled in the literature.