1. Introduction
The 2015 Paris Agreement sought to limit global warming to a 2 °C increase above pre-industrial levels by the end of the century, with an aspirational increase of 1.5 °C. The Intergovernmental Panel on Climate Change (IPCC) [
1] has predicted a greater than 50% likelihood that global warming will reach or exceed 1.5 °C much sooner, between 2030 and 2052, even for a greenhouse gas emissions scenario that increases at less than the current rate. In its Adaptation Gap Report 2021 launched at COP26, the United Nations Environment Programme (UNEP) [
2] warned that mitigation efforts to cut greenhouse gas emissions are “…still not anywhere near strong enough…”, with countries currently on track to experiencing a 2.7 °C increase. UNEP [
3] predicted that implementation of current mitigation pledges implied a 50% chance of keeping global warming to 2.5 °C (with a range of 2.0 °C to 2.9 °C) by the end of the century.
An editorial in Nature Climate Change [
4] warned that mitigation actions and pledges to reduce greenhouse gas emissions have been insufficient to meet either Paris Agreement target, making adaptation increasingly urgent. The IPCC [
5] Sixth Assessment Report (AR6 WG II) emphasised the significance of adaptation by noting that near-term actions to mitigate emissions that would limit global warming would reduce projected losses and damages in both human systems and ecosystems but would not eliminate them. As in previous Assessment Reports, it stressed the inequity of the distribution of the adverse impacts of global warming, with the most vulnerable people and systems being disproportionately affected. Unlike previous Assessment Reports, it highlighted adaptation actions that are effective, equitable, and just. These views echo those expressed so eloquently in The Stern Review [
6]. Stern argued that the climate will continue changing through the near future, however successful mitigation efforts are, and without early and strong mitigation the costs of adaptation will rise exponentially. Moreover, the communities affected first and most severely are often those that are least able to adapt and contribute little to the problem, creating a double inequity and warranting increased assistance from developed countries. This study examines the effectiveness and, following Stern, the equity of adaptation to climate change.
A range of constraints can prevent adaptation from reaching its potential. Some nations adapt better than others, perhaps because they face fewer constraints to adaptation than others, inadequate resourcing and financing being prominent constraints, limited governance competence another. Adger and Barnett [
7] and Massetti and Mendelsohn [
8] claimed that while adaptation has the potential to greatly reduce climate change damage, there is no guarantee this potential will be reached. The widely proclaimed distinction between adaptive capacity and adaptation suggests an application of the concepts of best practice and dominance to evaluate nations’ adaptation performance. Analytical techniques have been developed to implement the concepts of best practice and dominance, with empirical performance evaluation applications ranging widely, from hospitals and schools to industries and nations, and even to the environmental performance of businesses (Trinks et al., [
9]) and nations (Bosetti and Buchner [
10], Matsumoto et al., [
11]). This study applies these concepts to assess nations’ climate change adaptation performance.
Adaptation, particularly transformative adaptation, requires resources and financing often unavailable in developing countries. UNEP [
2] noted that while mitigation is the preferred way to lower the impacts and costs of climate change, and recent climate change financing has gone primarily to mitigation, support for adaptation is critical to keep existing adaptation gaps between implemented adaptation and societally set adaptation goals from widening. It documented that planning, financing, and implementation of adaptation remain weak, and only a small portion of the fiscal stimulus to combat the COVID-19 pandemic has targeted climate finance, and a small portion of that has gone to adaptation. IPCC [
5] AR6 WG II documented that recent adaptation progress has been distributed unevenly across regions, with observed adaptation focused more on planning than implementation. In his statement on the release of IPCC [
12], United Nations Secretary-General António Guterres called adaptation “the neglected half of the climate equation” and urged the public and private sectors to work together to ensure a just and rapid transformation to a net zero emissions global economy.
The potential gains to adaptation can be substantial. The Global Commission on Adaptation [
13] has estimated that a
$1.8 trillion investment in adaptation, including early warning systems, climate-resilient infrastructure, improved agricultural practices, mangrove protection along coastlines, and resilient water resources, could generate
$7.1 trillion in benefits through a combination of avoided costs and a variety of social and environmental benefits. It also estimated that universal access to early warning systems could deliver benefits up to ten times the initial cost.
With this background, the study has three objectives. The first is to obtain a set of adaptive capacity and adaptation readiness indicators and to aggregate these indicators into a composite adaptation index for a large number of nations; this appears in
Section 2,
Section 3 and
Section 4, with guidance from the IPCC and related literatures. The second is to provide an analytical framework for modelling and quantifying the composite adaptation performance of nations; this is the subject of
Section 5, in which a pair of complementary performance evaluation techniques is developed. The third and most significant objective is to explore the distribution of the composite adaptation performance of nations, and to illustrate its double inequity; this is the subject of the empirical analysis in
Section 6.
The paper proceeds as follows.
Section 2 and
Section 3 survey the literature seeking to define adaptive capacity and adaptation readiness, the activities that contribute to them and the factors that constrain them, and the inequity of their distributions.
Section 4 introduces and evaluates the data used to assess the adaptation performance of nations. These data are obtained from the University of Notre Dame Global Adaptation Initiative [
14] and are used in a manner guided by the IPCC and related literatures.
Section 5 describes the distance to frontier and dominance techniques used to implement an assessment of nations’ adaptation performance.
Section 6 presents the empirical findings of the study, which provide strong support for the growing emphasis the IPCC has placed in successive Assessment Reports, and strongly urged by Secretary-General Guterres at COP26, for the need to increase climate change funding, and to reallocate funds from mitigation to adaptation and from developed to developing nations.
Section 7 concludes with a summary of the findings and their implications for the climate justice movement, and notes two limitations of the study that may spur additional research into climate change adaptation.
2. The IPCC on Adaptation
The IPCC [
15] Third Assessment Report AR3 of Working Group II (AR3 WG II) expressed five reasons for concern about vulnerability to climate impacts, concerns which it has re-evaluated in subsequent Assessment Reports. The reasons involve relationships between global mean temperature increase and five projected adverse impacts: risk of damage to or irreparable loss of unique and threatened systems, the distribution of its impacts, the magnitude of aggregate impacts, the risk of extreme weather events, and the risk of large-scale singular events. These concerns led it to consider the role of adaptation, which it defined as adjustment in natural and human systems to the actual or expected impacts and risks of climate change, and it noted that adaptive capacity is more limited in natural systems than in human systems. It distinguished adaptation from adaptive capacity, which it defined as the ability (emphasis added) of a system to adjust to climate change, to moderate potential damages, or to cope with the consequences. It viewed adaptation as a necessary strategy to complement climate change mitigation efforts and noted that the ability to adapt to and cope with climate change depends on wealth, technology, education, information, human and social capital, infrastructure, institutions, management capabilities, and access to resources. Subsequent Assessment Reports have expanded on these drivers, emphasising the roles of technology, citing new and possibly disruptive technologies and enhanced climate-driven innovation, and a supportive institutional framework.
AR3 WG II attributed the difference between actual adaptation and adaptive capacity to maladaptation and constraints to achieving potential adaptation. This distinction was strengthened in IPCC [
16] AR4 WG II, which stated “The message from the literature is clear: adaptive capacity signals potential but does not guarantee adaptive action”, and warned that more extensive adaptation than was currently occurring would be required to reduce vulnerability to future impacts of global warming. It cited regulations and policies, limited governance capacity, availability and distribution of finance, violent conflict, the spread of infectious diseases, and urbanisation as factors that may facilitate or constrain adaptation. This list of constraints has changed little through successive Assessment Reports.
Of particular relevance to this study, the IPCC [
15] AR 3 WG II asserted that nations with the least resources have the least capacity to adapt and are most vulnerable to climate change impacts. It also asserted that the projected distribution of impacts would increase the disparity in well-being between developed and developing nations, with the disparity growing for higher projected temperature increases. The IPCC [
17] AR5 WG II noted that differences in vulnerability and exposure that require adaptation arise from non-climatic factors and from multidimensional inequalities associated with uneven development processes, with people who are socially, economically, culturally, politically, institutionally, or otherwise marginalised being especially vulnerable. The IPCC [
1] Report on Global Warming of 1.5 °C identified populations and regions at disproportionately higher risk of adverse impacts, including disadvantaged and vulnerable populations, local communities dependent on agricultural or coastal livelihoods, small island developing states (SIDS), and Least Developed Countries (LDCs). It suggested that social justice and equity are core aspects of climate-resilient pathways, that a consideration of ethics and equity can address the uneven distribution of adverse impacts and concluded that international cooperation is critical for developing countries and vulnerable regions. The IPCC [
5] AR6 WG II stated that poor and otherwise disadvantaged groups are especially vulnerable because they have fewer assets and less access to funding, technologies, and political influence. People in the most vulnerable situations and regions are also highly exposed to climate change impacts. The most vulnerable regions include South Asia, Micronesia and Melanesia, Central America, and most of Africa.
These definitions of adaptation, adaptive capacity, and adaptation constraints, and the proclaimed inequity in the distribution of adaptive capacity, have changed little through successive Assessment Reports. A second consistency is the lack of a definition of adaptation readiness, although some of its drivers appear often (e.g., innovation, education, governance) (The IPCC [
5] AR6 WG II has introduced the concept and extolled the significance of enabling conditions, including political commitment, institutional frameworks, knowledge, and monitoring and evaluation. These conditions overlap with the adaptation readiness indicators used in this study.) A third consistency involves the growing acknowledgement of an association between the greenhouse gas emissions of developed countries and the climate change impacts that disproportionately affect developing countries responsible for few emissions. What has changed in successive Reports is the knowledge base, consisting of advances in science and increases in the quantity and quality of evidence in databases and in the scientific, technical, and socioeconomic literature. This has allowed the IPCC to increase the confidence it attaches to its assessments of the relationship between global warming and its impacts in each successive Assessment Report.
4. The Data
Many of the variables cited above as influencing adaptation to climate change appear in the ND-GAIN country data from the University of Notre Dame Global Adaptation Initiative [
14]. These data are therefore used, in a manner guided by the IPCC and related literatures. The ND-GAIN country index is constructed from 36 vulnerability indicators and nine readiness indicators for up to 192 nations over varying time periods concluding in 2019. The vulnerability indicators consist of 12 exposure indicators, 12 sensitivity indicators and 12 adaptive capacity indicators, each measured on [0, 1] with low (high) values indicating low (high) vulnerability. (The ND-GAIN indicators have been transformed from raw data. The University of Notre Dame [
14] provides raw data and derived indicators, and The University of Notre Dame [
52] provides details of the transformation procedures. The indicators have been used often to study climate change vulnerability and adaptation; among recent studies are Edmonds et al. [
53], Halkos et al. [
54], Amegavi et al. [
28], and Ripple et al. [
55]).
The selection of data is guided by the observation that ND-GAIN defines adaptive capacity as “the ability of society and its supporting sectors to adjust to reduce potential damage and to respond to the negative consequences of climate events…”. The 12 adaptive capacity indicators “…seek to capture a collection of means, readily deployable to deal with sector-specific climate change impacts”. ND-GAIN defines adaptation readiness as preparedness “…to make effective use of investments for adaptation actions thanks to a safe and efficient business environment…”, and it measures adaptation readiness with three components: economic readiness, governance readiness and social readiness. These interpretations and definitions suggest a strong complementarity between adaptive capacity and adaptation readiness and are consistent with the views expressed in the IPCC and related literatures reviewed in
Section 2 and
Section 3. They also support the creation of a composite adaptation index combining the two concepts, since adaptive capacity itself is insufficient for successful adaptation without the political, social, and institutional support provided by adaptation readiness. (Amegavi et al. [
28] used a subset of the database we use to show that adaptation readiness is significantly and negatively related to vulnerability to climate change in 51 African nations. Our results support this finding and point to the significance of adaptation readiness in a larger sample of nations).
Consequently, it is hypothesised that the overall adaptation performance of nations is a function of their adaptive capacity and features of their institutional environment that enhance or constrain their adaptive capacity. These features are called enabling conditions in the IPCC [
5] AR6 WG II and captured by the adaptation readiness indicators in this study. To test this hypothesis 12 adaptive capacity indicators and nine adaptation readiness indicators are extracted from the ND-GAIN database for the terminal year 2019. The adaptive capacity indicators are augmented with the adaptation readiness indicators because the IPCC and other literatures reviewed in
Section 2 and
Section 3 consistently refer to various enabling conditions (e.g., regulatory quality, innovation, and education) as being important elements in the performance of nations to adapt to unmitigated climate change.
Table 1 lists the ND-GAIN adaptive capacity and adaptation readiness indicators. Each of the 21 ND-GAIN indicators is designed to capture both capacity and access characteristics. Detailed descriptions of and rationale for each indicator appear in University of Notre Dame [
52].
Three adjustments have been made to the ND-GAIN data. The 12 adaptive capacity indicators have been transformed because ND-GAIN associates high vulnerability indicators with high vulnerability, and adaptive capacity is one of three components of vulnerability. Since adaptive capacity reduces vulnerability, each adaptive capacity indicator is redefined so that high values of each adaptive capacity indicator are associated with high adaptive capacity, thereby retaining their [0, 1] range. Two of the transformed ND-GAIN adaptive capacity indicators, “improved water source (% of population with access)” and “improved sanitation facilities (% of population with access)” have missing values for 94 and 103 nations, respectively, and these indicators have been deleted, leaving 10 adaptive capacity indicators. A new water indicator was adopted, the geometric mean of “dam capacity” from ND-GAIN and “average precipitation in depth (mm per year)” from the World Bank’s World Development Indicators. This new water indicator combines rainfall with water storage capacity and provides a nearly necessary condition for the original water indicator “improved water source (% of population with access)”, while greatly increasing coverage from 98 to 134 nations, leaving 11 adaptive capacity indicators.
These three adjustments generate a pair of data matrices, one consisting of 11 adaptive capacity indicators for up to 192 nations, and the other consisting of nine adaptation readiness indicators for up to 192 nations. However, these two matrices contain many missing observations. One adaptive capacity indicator, disaster preparedness, is missing for 56 nations, and two adaptation readiness indicators, social inequality and innovation, are missing for 43 and 44 nations. These three indicators have been deleted, leaving nine adaptive capacity indicators and seven adaptation readiness indicators. If a nation is missing one or more of the nine remaining adaptive capacity indicators, that nation is deleted from the adaptive capacity matrix, and similarly for the adaptation readiness matrix. This leaves an adaptive capacity matrix consisting of nine indicators for 143 nations and an adaptation readiness matrix consisting of seven indicators for 172 nations. In order to merge information on adaptive capacity with information on adaptation readiness into a composite adaptation index, the sample is restricted to nations having values for all 16 indicators. This leaves an adaptive capacity matrix consisting of nine indicators and an adaptation readiness matrix consisting of seven indicators, both for the same 134 nations. Summary statistics for the 16 indicators appear in
Appendix A Table A1.
These two data matrices reflect the difficult trade-off between coherence and comprehensiveness of indicators and comparability and inclusiveness of nations. (For a conceptual treatment of this trade-off, see Ford and Berrang-Ford [
56], who proposed four requirements for successful adaptation tracking: (1) a consistent definition for monitoring panel data, (2) observed units must be comparable, (3) sample size must be large enough to be comprehensive, and (4) indicators must be coherent with our understanding of adaptation.) A preference for inclusiveness reflects our desire to retain as many developing nations as possible. The data set contains 32 Least Developed Countries (LDCs) identified by the UN and 11 SIDS identified by the UN and includes 27 sub-Saharan African nations and 10 North African nations. These nations have been singled out by the IPCC and at COP26 as being victims of climate change caused largely by developed nations, who have lagged in both their mitigation efforts and their financial support to developing nations to enhance their adaptation performance through National Determined Contributions. This dichotomy has been labelled an equity and ethical issue in consecutive IPCC Assessment Reports, and a justice issue by many, including by Robinson and Shine [
57], Simmons [
50], and Klinsky [
51].
5. The Analytical Techniques
A pair of complementary analytical techniques are used to assess the relative adaptive capacity and adaptation readiness of nations. Each technique is illustrated using adaptive capacity, and the same analysis applies to adaptation readiness and composite adaptation. Both techniques identify leading and lagging nations. The first identifies leaders and laggards by using Data Envelopment Analysis (DEA) to exploit the “distance to frontier” concept of Acemoglu et al., [
58] and applied to OECD productivity dispersion by Andrews et al. [
59] and Berlingieri et al. [
60,
61]. The second identifies leaders and laggards by using dominance analysis to identify nations that are structurally similar but perform better than other nations, regardless of their distance to the best practice frontier.
The first technique, DEA, is a linear programming technique developed by Charnes et al. [
62] to assess the relative performance of observations in a sample. Rather than fitting a regression through the data, as most statistical techniques do, DEA constructs a frontier that envelops the data, from above if the objective is to maximise and from below if the objective is to minimise. The frontier consists of best practice observations, and with a maximisation orientation, all observations beneath the frontier lag best practice by varying degrees. In the current setting higher adaptation indicator values are preferred, and DEA constructs an adaptive capacity frontier that bounds an adaptive capacity set from above. The adaptive capacity frontier consists of best practice nations, those that adapt best, and the interior of the adaptation set contains all nations whose adaptation performance lags best practice, or the “best” and the “rest” in the OECD productivity literature. DEA simultaneously identifies adaptation leaders on the best practice frontier and measures the radial distance to the frontier of the adaptation laggards. Distance to the frontier provides a new measure of the adaptation gap.
Let nations be indexed by i = 1, …, I, and let a nation’s adaptive capacity be tracked across N sectors with sectoral adaptive capacity indicators labelled y
n and indexed by n = 1, …, N. In the current application I = 134 and N = 9. The DEA program that evaluates the aggregate adaptive capacity to cope with climate change of nation “o” is given by the dual pair of linear programs in
Figure 1. These programs calculate an endogenously weighted adaptive capacity index ACI for each nation. This index aggregates N individual adaptive capacity indicators y
n into a single adaptive capacity index ACI.
The envelopment and multiplier programs contain sectoral adaptive capacity indicators y
n but no additional variables that might influence adaptive capacity such as resource availability. This abbreviation of a conventional DEA program is the contribution of Adolphson et al., [
63], and Lovell and Pastor [
64]. Unlike most models of business or economic behaviour that contain variables to be maximised, such as business revenues or educational outcomes, and constraining variables, such as business expenses or resource availability, this adaptive capacity model is restricted to variables to be maximised, the N sectoral adaptive capacity indicators. The envelopment program in
Figure 1 envelops nations’ adaptive capacity data from above and calculates the potential of nation “o” to expand its vector of sectoral adaptive capacity indicators y
o as much as possible, subject to N constraints, one for each sectoral indicator. These constraints bound the expanded vector
y
o above by a nonnegative combination of the most capable nations in the sample.
The optimal value of ∈ [1, +). A value = 1 indicates best practice adaptation on the adaptive capacity frontier, with larger values of indicating the degree to which a nation must improve its adaptation performance to reach the best practice frontier. also forms the basis for a measure of a nation’s adaptive capacity gap, the difference between (or ratio of) its actual adaptive capacity yo and its potential adaptive capacity yo. Deviations beneath this frontier capture an alternative representation of nations’ adaptation gaps to the UNEP Adaptation Gap Reports of the same name by replacing a vague “societally set goal” with a best practice that can be estimated empirically.
The reciprocal −1 ∈ (0, 1] is a nation’s adaptive capacity index ACI. A value = 1 indicates best practice adaptation, and lower values of −1 indicating reduced levels of adaptive capacity. −1 provides a ranking of nations based on their overall adaptive capacity to cope with climate change, independently of any other national characteristics, which are ignored in the present analysis.
The multiplier program in
Figure 1 calculates for nation “o” a vector of endogenous weights ν
n ∈ (0, +
) with which to aggregate its sectoral adaptive capacity indicators into its ACI. By the duality theorem of linear programming, at optimum
= μ, and ACI can therefore be expressed as an endogenously weighted sum of its sectoral adaptive capacity indicators,
−1 = (
)
−1 for any nation i = 1, …, o, …, I. (These endogenous weights are also known as “benefit of the doubt” weights, a term introduced by Melyn and Meusen [
65]. Cherchye et al. [
66] provide details on benefit of the doubt composite indices.)
Endogeneity of weights is central to the analysis, having the virtue of not forcing nations to value sectoral adaptive capacities equally. For a nation with an abundance of sectoral adaptive capacity indicator y
n the program implicitly attaches a large weight to this indicator to maximise its ACI. Conversely, for a nation with a paucity of sectoral adaptive capacity indicator y
n the program implicitly attaches a small weight to this indicator to maximise its ACI. These endogenous weights provide a considerable improvement over the fixed weights used in most composite indices, including the popular UNDP [
46] Human Development Index and the ND-GAIN indices. Fixed weights impose perfect substitutability among component indicators, with rates of substitution constant across nations. The endogenous weights generated by DEA also impose perfect substitutability among component indicators, but with the important advantage that weights, and rates of substitution among component indicators, are allowed to differ across nations according to their circumstances. Weight flexibility is particularly important in the construction of an adaptive capacity index, since nations differ in their exposure, sensitivity, and vulnerability to climate change across sectors. Endogeneity of weights allows Pacific Island nations to value adaptation indicators differently than sub-Saharan African nations. New Zealand has ample rainfall, and Mauritania is arid, leading to the expectation that New Zealand assigns a relatively high weight and Mauritania assigns a relatively low weight to a water indicator. By reflecting different adaptive capacities across sectors that in turn reflect different national circumstances, these weights have the potential to assist in the design of policies intended to allocate climate finance to enhance adaptive capacity in an equitable manner, as noted in
Section 1 with reference to Mendelsohn [
36] and Anderson et al. [
38].
However, endogeneity of weights has a potential drawback. As the number of choice variables relative to the sample size increases, estimation becomes exponentially more difficult, a situation referred to as the curse of dimensionality. In our setting the number of adaptive capacity indicators relative to the number of nations in the sample N/I = 9/134 is sufficiently large to hinder evaluation of the adaptation performance of nations. In effect, having nine adaptive capacity indicators gives nations excessive freedom to choose weights in creating their ACIs, resulting in many nations receiving ACI = 1, even though their index is the consequence of being different rather than excelling. The curse is less severe in the case of adaptation readiness, where N/I = 7/134.
Summarising, the DEA methodology makes three contributions to the construction of an adaptive capacity index. It exploits the ability to generate endogenous weights with which to aggregate sectoral indicators that respect nations’ varying circumstances. Nations’ endogenously weighted adaptive capacity indices provide an analytically sound way of identifying leaders and laggards and quantifying adaptation gaps. These endogenous weights have the potential to guide policy intended to lower the cost of enhancing adaptive capacity in an efficient, i.e., resource-saving, and equitable manner.
The second technique, dominance analysis, provides information complementary to that provided by DEA. The basics of dominance analysis are extracted from a much more detailed presentation in Tulkens [
67]. Consider two nations with adaptive capacity vectors y
j and y
k. Nation j dominates nation k if nation j has at least as much adaptive capacity as nation k for all N indicators, that is if y
nj y
nk,
n = 1, …, N. Aggregating the inequality over all k = 1, …, I nations generates the number of nations nation j dominates. Reversing the inequality generates the number of nations that dominate nation j. This strategy can be extended by deleting d
1 adaptive capacity indicators at a time, with replacement, to evaluate dominance with N-d indicators. This provides a way of determining the indicators for which a nation is most or least dominant.
Dominance analysis is independent of the notions of best practice adaptation and an adaptive capacity index. Rather, it identifies leaders as the most frequently dominating nations and laggards as the most frequently dominated nations. In doing so it identifies role models for dominated nations. These role model nations are relevant because they have similar mixes of adaptive capacity indicators, but with larger indicator values. It is important to note that a nation can dominate other nations by being similar to them and without being best practice, and a nation can be best practice by being different from other nations and without dominating any of them. This distinguishes dominance analysis from DEA and highlights the complementarity between the two techniques.
This exposition of DEA and dominance analysis has been illustrated with application to adaptive capacity, and the analysis applies equally to adaptation readiness and composite adaptation, with only the number of variables and their definitions changing. The joint contribution of these two complementary techniques is to refocus the analysis of adaptation from a global concept, or from a developed nations vs. developing nations concept, to a performance analysis specific to each individual nation. Importantly, these techniques identify leading and lagging nations, and quantify the three adaptation gaps for each lagging nation. Finally, they provide a rigorous foundation for a nation-focused investigation into the double inequity of composite adaptation to climate change.
7. Conclusions
The introduction set three objectives for this study: to create a database of indicators conforming to the IPCC concept of adaptation, to propose analytical techniques with which to assess the adaptation performance of nations, and to explore the distribution of composite adaptation performance among nations and empirically assess its inequities.
The database is created in
Section 4 and incorporates the adaptive capacity and adaptation readiness indicators proposed in successive IPCC Assessment Reports. These indicators reflect a belief that a supportive institutional environment provides the readiness essential to the success of adaptation efforts. The database is drawn from the ND-GAIN database, although it is not equivalent to it, and it satisfies the essence of the Ford and Berrang-Ford [
56] requirements for successful adaptation tracking.
In
Section 5 a linear programming distance to frontier technique, DEA, is augmented with a dominance analysis, providing complementary insights into nations’ composite adaptation performance, by identifying leaders and laggards according to different criteria, and by identifying the indicators at which they perform relatively well or relatively poorly. Dominance analysis adds value to DEA by dispensing with the frontier concept and evaluating nations’ adaptation performance relative to other nations, rather than relative to an adaptation frontier. Together the two techniques provide a rigorous analytical foundation for subsequent empirical analysis of the composite adaptation performance of nations.
The third objective is achieved in two stages. The empirical analysis in
Section 6 identifies leading and lagging nations in terms of their relative adaptive capacity and their relative adaptation readiness separately, and in terms of their relative composite adaptation performance. The overriding impression gained is one of very large dispersion in nations’ adaptation performance. The composite adaptation gap between leading and lagging nations is large, with lagging nations’ adaptation performance on the order of 57% of that of leading nations. The gap is attributable primarily to inadequate adaptation readiness of institutional environments that plagues nearly all nations and is particularly severe among lagging nations. This impression of large dispersion is reinforced when DEA is used to aggregate adaptive capacity and adaptation readiness, with most nations weighting the former more heavily than the latter. When the distance to frontier analysis is augmented with a dominance analysis on adaptive capacity and adaptation readiness criteria separately, the significance of adaptation readiness is strengthened. Dominance relationships are roughly twice as frequent with adaptation readiness as with adaptive capacity, attesting further to the importance of a supportive institutional environment. These findings highlight the empirical significance of the complementarity between adaptive capacity and adaptation readiness, and quantify the magnitudes of the three adaptation gaps, two results that have received insufficient attention in the literature.
In the second stage of the empirical analysis composite adaptation leaders and laggards are identified geographically. In terms of both distance to frontier and dominance analyses, composite adaptation leaders are overwhelmingly European nations and their Western Offshoots located in higher latitudes in the northern and southern hemispheres, and laggards are equally overwhelmingly least developed countries, most of them sub-Saharan African and South Asian, located in lower latitudes close to the equator. The distance to the equator principle of economic development applies equally well to climate change adaptation performance.
When an income dimension is added to the characterisation, leaders have approximately 15 times the GDP per capita as laggards have. This relationship applies to the entire distribution of nations, not just to the leading and lagging tails; the correlation between income and composite adaptation performance is 0.75. National composite adaptation performance varies positively and strongly with national income, as the IPCC asserts. This finding illustrates one of Stern’s double inequities of adaptation; the poorest nations are the least able to adapt to climate change impacts.
When responsibility for climate change is added to the characterisation, leaders generate more than three times the amount of GHG emissions per capita as laggards do. This relationship also holds for the entire distribution of nations; national composite adaptation performance varies positively, although not strongly due to a few prominent outliers, with responsibility for climate change. This finding illustrates the other of Stern’s double inequities of adaptation; nations least responsible for causal greenhouse gas emissions are least able to adapt to their impacts.
When a combination of income and responsibility is added to the characterisation, Stern’s double inequity is clearly revealed. The correlation between a combination of income and responsibility and composite adaptation performance is 0.68. National income and responsibility for climate change vary positively and strongly with composite adaptation performance. Those nations having weak composite adaptation lack the resources to adapt to climate change attributable largely to those nations having relatively abundant composite adaptation. Stern’s double inequity is portrayed graphically in
Figure 2 and
Figure 3, which are barely distinguishable. These findings provide analytically based empirical results confirming the well-known but inadequately documented double inequity of climate change. They quantify each inequity gap, and a generic double inequity gap, for each nation. Each of these gaps is large on average, and enormous for some nations.
These findings also reveal two limitations of the research. One is illustrated by the white gaps in
Figure 2 and
Figure 3 representing Bolivia in the western Amazon basin and the Democratic Republic of Congo and other nations in central Africa. These regions contain major portions of the two largest rainforests in the world. Given the importance of the ecosystems in these two regions, and their vulnerability to climate change, it would have been desirable to include these nations in the empirical analysis. Despite our efforts to retain as many nations as possible, insufficient data are available for these nations to allow their inclusion. A second limitation involves the scope of adaptation. Although the underlying data provide an adequate basis for assessing adaptation in the human environment, they provide a limited basis for assessing adaptation in the natural/ecological environment. The International Union for the Conservation of Nature (IUCN) has conducted studies and amassed data on ecosystem-based adaptation that complement our knowledge of adaptation focused on the human environment and expand the elements related to natural/ecosystem adaptation to climate change. (See, for example, IUCN [
73] and Keith et al. [
74] for details on adaptation in the natural/ecosystem environment.) It would be worthwhile in subsequent research to determine if it is possible to merge the IUCN natural/ecosystem adaptation data with the ND-GAIN largely human adaptation data to gain a more complete picture of adaptation.
A lively literature has emerged that regards climate change as a justice issue. Although it is not among the 17 United Nations Sustainable Development Goals, climate justice “…looks at the climate crisis through a human rights lens…” (
https://www.un.org/sustainabledevelopment/blog/2019/05/climate-justice/, accessed on 15 October 2022), thereby providing a holistic but loosely defined notion of the (in)ability to achieve these goals. This study has addressed climate change as an equity issue by providing a rigorous analytically based confirmation of Stern’s double inequity assertion, a positive assertion that can be and has been tested empirically against a measurable alternative of adaptation equality. This empirical approach to climate change as an equity issue contrasts with the popular assertion that treats climate change as a justice issue. The latter is a normative assertion that can be debated but cannot be tested empirically until a benchmark is developed against which climate justice can be measured.