Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Antoine Dechezleprêtre Sam Fankhauser Matthieu Glachant Jana Stoever Simon Touboul Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Antoine Dechezleprêtre Sam Fankhauser Matthieu Glachant Jana Stoever Simon Touboul © 2020 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org This work is a product of the staff of The World Bank and the Global Facility for Disaster Reduction and Recovery (GFDRR) with external contributions. The findings, analysis and conclusions expressed in this document do not necessarily reflect the views of any individual partner organization of The World Bank, its Board of Directors, or the governments they represent. Although the World Bank and GFDRR make reasonable efforts to ensure all the information presented in this docu- ment is correct, its accuracy and integrity cannot be guaranteed. Use of any data or information from this document is at the user’s own risk and under no circumstances shall the World Bank, GFDRR or any of its partners be liable for any loss, damage, liability or expense incurred or suffered which is claimed to result from reliance on the data contained in this document. The boundaries, colors, denomination, and other information shown in any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. RIGHTS AND PERMISSIONS The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribu- tion to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. Table of Contents Acknowledgments ........................................................................................................................................... 5 Abbreviations..................................................................................................................................................... 5 Key Insights ..................................................................................................................................................... 6 1. Introduction .............................................................................................................................................. 8 2. Data Issues ................................................................................................................................................ 10 3. Invention of Technologies for Climate Change Adaptation................................................... 18 4. International Technology Transfer.................................................................................................... 22 5. Patenting Activity in Relation to Climate Hazards.................................................................... 29 6. Conclusion and Policy Implications ................................................................................................ 34 Appendix A: Supplementary Tables and Figures....................................................................................... 36 Appendix B: Regressions................................................................................................................................. 49 Appendix C: Construction of the Five Hazard Indicators....................................................................... 51 Endnotes. ............................................................................................................................................................ 52 References.......................................................................................................................................................... 53 Figures Figure 2.1 Country-Specific Climate Hazard and Adaptive Capacity Levels, 1995–2015............. 13 Figure 3.1 Climate Adaptation Innovation, as Measured by High-Value Patents, 1995–2015 ... 18 Figure 3.2 Innovation for Climate Change Adaptation, as a Share of Total Innovation, 1995–2015................................................................................................................................... 20 Figure 4.1 Technology Transfer Rates, by Invention Type, 2010–15.................................................. 22 Figure 4.2 Trends in Climate Adaptation Technology Transfer as a Share of Invented Adaptation Technologies, 1995–2015............................................................. 24 Figure 4.3 Transfer Rates of Climate Change Adaptation Technology, by Field, 2010–15............ 25 Figure 4.4 Number of FDI Deals in Climate Change Adaptation Technologies, 2000–15............. 27 Figure 5.1 Relationship between Adaptation Technology Invention and Climate Hazards .............................................................. 29 in High- and Middle-Income Countries, 2010–15. Figure 5.2 Relationship between Climate Hazards and Adaptation Technology Inventions (as a Share of All Technology) in High- and Middle-Income Countries, 2010–15........ 30 Figure 5.3 Relationship between Climate Hazards and Imports of Climate Change Adaptation Technologies in High- and Middle-Income Countries, 2010–15................ 31 Figure 5.4 Correlation between Availability of Adaptation Technologies and Climate Hazard Levels in High- and Middle-Income Countries 2010–15................................................... 33 Figure A.1 Climate Change Mitigation Innovation, as a Share of Total Innovation, 1995–2015 37 4 / Figure A.2 Relationship between Climate Adaptation Technology Invention and Climate Hazards, Middle-Income Countries, 2010–15 Annual Average....................................... 40 Figure A.3 Relationship between Climate Adaptation Technology Imports and Climate Hazards, Middle-Income Countries, 2010–15 Annual Average........................................................ 41 Figure A.4 Correlation between Number of Adaptation Inventions and Climate Hazard Levels in High- and Middle-Income Countries, 2010–15.................................................. 42 Figure A.5 Correlation between Number of Imported Adaptation Inventions and Climate Hazard Levels in High- and Middle-Income Countries, 2010–15.................................... 43 Tables ................................................ Table 2.1 Technology Fields of Y02A Patents included in the Study. 13 Table 2.2 Definitions and Data Sources of the Five Hazard Indicators............................................ 16 Table 3.1 Average Annual Growth of Innovation in Different Fields, as Measured by High-Value Patents, 1995–2015............................................................................................. 19 ... 21 Table 3.2 Top 10 Inventor Countries in Climate Change Adaptation Technologies, 2010–15. Table 4.1 Average Number of Patent Offices Where Internationally Patented Inventions .............................................................................. Are Filed, by Technology Type, 2010–15. 23 Table 4.2 Distribution between Country Income Groups of Patented Inventions .............................................. 26 of Technologies for Climate Change Adaptation, 2010–15. Table 4.3 Flow of FDI Deals in Climate Change Adaptation Technologies between Country Income Groups, 2010–15......................................................................................... 27 Table 4.4 Top Acquirers and Importers of Climate Adaptation Technologies among Low- or Middle-Income Countries, 1995–2015.................................................................. 28 Table A.1 Top 10 Inventor Countries in Climate Change Mitigation Technologies, 2010–15..... 36 Table A.2 Top 10 Inventor Countries in Climate Change Adaptation Technologies, by Sector, 2010–15.................................................................................................................... 37 Table A.3 Transfer between Country Income Groups of Patented Technologies for Climate Change Adaptation and Mitigation, 2010–15...................................................................... 38 Table A.4 Top 10 Importing Countries of Climate Change Adaptation Technologies, 1995–2015................................................................................................................................... 38 Table A.5 Top 15 Acquirer Countries in FDI Deals Related to Climate Change Adaptation ........................................................................................................ Technologies 1995–2015. 39 Table A.6 Economies and Income Groups Included in the Study....................................................... 44 Table B.1 Relationship between Domestic Innovation and Climate Hazard Levels, Population, and Stock of Patented Inventions..................................................................... 50 Table B.2 Relationship between Technology Imports and Climate Hazard levels, ......................................... Population, Stock of Patented Inventions, and Trade Barriers. 50 / 5 Acknowledgments This document is the final report of the project, “Invention and North-South Transfer of Technologies for Climate Change Adaptation,” commissioned by the World Bank. The authors are: • Matthieu Glachant and Simon Touboul, MINES ParisTech, PSL University • Antoine Dechezleprêtre, Organisation for Economic Co-operation and Development and London School of Economics • Sam Fankhauser, Grantham Research Institute on Climate Change and the Environment, London School of Economics • Jana Stoever, Kiel University The authors thank Stéphane Hallegatte, Emma Katherine Phillips, Erika Vargas, and peer reviewers who help to greatly improve the quality of the report: Brenden Jongman, Jean-Louis Racine, Jun Rentschler, and Arame Tall. The authors are also grateful to Mary Anderson and Miki Fernández, who respectively edited and designed this report. Any remaining errors are our own. Finally, the report team is grateful to the Global Facility for Disaster Reduction and Recovery (GFDRR), whose generous funding made this report possible. Contact matthieu.glachant@mines-paristech.fr Abbreviations EPO European Patent Office FDI foreign direct investment IP intellectual property IPCC Intergovernmental Panel on Climate Change PATSTAT World Patent Statistical Database (EPO) R&D research and development Y02A technologies for adaptation to climate change (PATSTAT classification) 6 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Key Insights M any people will adapt to climate change by changing their behavior, perhaps by moving to a new location or changing their occupation. They will also rely on technologies that increase resilience to climate risks and extremes, such as new irrigation systems, advanced weather forecasting tools, and more-resilient crop varieties. The extent to which such technologies are developed and globally available will significantly shape the “new normal” of life—if not sheer survival, for millions—in adapting to climate change. To better grasp the current state and future needs within this sphere, this report relies on patent data to describe and analyze innovation activity in technologies for climate change adaptation. The analysis looks at the pace of innovation; identifies which countries lead and how technologies for climate change adaptation diffuse across countries (international technology transfer); compares trends in adaptation innovation with those in other technology fields; and relates these trends to adaptation needs. Importantly, the reliance on patent data restricts the scope of the analysis to solutions for adaptation that are at the technological frontier and ignores the role of nontechnological forms of innovation and low-tech options. The main findings of the report, summarized below, not only provide the first global snapshot of climate-adaptive technological innovation but also point toward the policy implications of current weaknesses in technology transfer. Invention Invention forfor adaptation relative Globally, mitigation. to number the of patented inventions in technologies for adaptation relative climate change adaptation increased steadily between 1995 and Globally, the number of patented inventions in technologies for climate change adaptation to mitigation. increased 2015. However, this increase in absolute terms does not correspond steadily between 1995 and 2015. However, this increase in absolute terms does not to a proportional rise in innovation for climate adaptation. When correspond to a proportional rise in innovation for climate adaptation. When considering the total considering the total number of inventions across all technologies in all number of inventions across all technologies in all fields, the share of climate adaptation inventions fields, the share of climate adaptation inventions in 2015 was roughly in 2015 was roughly the same as in 1995. This stagnation of research and development (R&D) the same as in 1995. This stagnation of research and development efforts toward adaptation stands in sharp contrast to the trend for climate change mitigation (R&D) efforts toward adaptation stands in sharp contrast to the trend technologies, whose share in total innovation (including non-climate-related) nearly doubled for climate change mitigation technologies, whose share in total during the same period. innovation (including non-climate-related) nearly doubled during the Concentration of innovationsame period. in climate change adaptation. Concentration Technological adapt to climate innovation toTechnological is concentrated change to innovation within adapt to climate a limited change number of is concentrated of innovation in countries. China, Germany, Japan, the Republic of Korea, and the United States together within a limited number of countries. China, Germany, Japan, account the climate for nearlychange high-value two-thirds of all Republic ofinventions Korea, and(inventions the United seeking in more for patents account States together than one nearly adaptation. country) filed globally between 2010 and two-thirds 2015. of all high-value inventions (inventions seeking patents in more than one country) filed globally between 2010 and 2015. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 7 International International diffusion of This concentration patented of innovation activity could in principle be inventions. diffusion of compensated by international technology transfer from the innovating This concentration of innovation patented activity countries. couldthe However, in data principle belimited reveal compensated by international international technology technology inventions.transfer from the innovating countries. However, the data reveal diffusion through the patent system. Few adaptationlimited international inventions technology diffusion through the are patent system. transferred Few across adaptation borders inventions relative are change to climate transferred across mitigation borders relative to climate change mitigation technologies and non-climate-related innovations. technologies and non-climate-related innovations. The international The international diffusion of adaptation diffusion technologies of adaptation related torelated technologies agriculture and coastal to agriculture and and river coastal protection is particularly low. Whether this pattern reveals that technologies for adaptation and river protection is particularly low. Whether this pattern reveals are innovating less applicable outside the that countryfor technologies than other technologies adaptation or that higher are less applicable barriers outside the exist to their international diffusion is an open question. innovating country than other technologies or that higher barriers exist to their international diffusion is an open question. Cross-border transfers. Cross-border Cross-border Cross-border transfers of patented transfers inventions forof patented climate inventions change for climate adaptation change predominantly transfers. adaptation predominantly occur between a small group of countries occur between a small group of countries consisting of high-income economies and China (85 flows).of high-income economies and China (85 percent of global consisting percent of global technological technological flows). Knowledge transfer to low-income countries. Knowledge transfer There is virtually no transfer of patented knowledge to low-income There is virtually no transfer of patented knowledge to low-incomeforcountries. If any, access to to low-income countries. If any, access to technologies climate change adaptation occursadaptation technologies for climate change outsideHowever, occurs system. thissystem. the patent However, this countries. outside the patent situation is not specific adaptation situation is not specific to to technologies. adaptation technologies.The The innovationliterature innovation has shown literature has shown that that on low-tech these countries mainly rely these solutions and organizational innovations. countries mainly rely on low-tech solutions and organizational innovations. Mismatch between countries’ adaptation needs and technological capacity. Mismatch Innovation and technology Innovation diffusion do nottechnology and seem to be diffusion driven bydo not seemneeds adaptation to be but driven by by the between countries’ level of recipient adaptation countries’ needs technological but by the capacities. absorptive level of recipient countries’ This could technological be bad news for adaptation needs absorptive capacities. This could be bad news for adaptation adaptation to climate change because countries with strong technological capacities typically to and technological face lower adaptation needs climate change at present. Thebecause mismatch countries between with strong technological adaptation capacities needs and technology capacity. typically face lower adaptation needs at present. The mismatch availability is particularly serious concerning technologies for mitigating temperature increases. between adaptation needs and technology availability is particularly Ability of market forcesserious to meetconcerning technologies local adaptation for mitigating temperature increases. needs. Economic forces thus seem currently unable transform to seem local adaptation into adapta needs local market Ability of market Economic forces thus currently unable to transform - demand for patented adaptation technologies. Solving this problem and creating the right forces to meet local tion needs into market demand for patented adaptation technologies. incentives for adaptation technologies to spread where they are urgently needed requires a adaptation needs. Solving this problem and creating the right incentives for adaptation better understanding of the market failures technologies that where to spread hinder demand, they whichneeded are urgently is a precondition for requires a bet - designing demand-pull policies in this domain. ter understanding of the market failures that hinder demand, which is a precondition for designing demand-pull policies in this domain. 8 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis 1 Introduction T he Earth’s climate has already begun to change and will inevitably continue to do so. Even if the targets set in the Paris Agreement are met—to keep the global surface temperature increase below 2 degrees Celsius relative to preindustrial levels—many regions will still suffer severely from the consequences of climate change. They will have more frequent extreme weather events, changes in precipitation patterns, rising sea levels, temperature increases, and many other related effects (IPCC 2018). In this context, technology is certainly a major tool to increase societies’ ability to adapt to the adverse effects of climate change (Klein and Tol 1997; Miao 2017; GCA 2019). International technology transfer hence becomes particularly important because a large fraction of the innovation activity in today’s knowledge-based economy takes place in the Global North, while technologies for climate change adaptation are urgently needed in low- and middle-income countries, which are particularly vulnerable to climate shocks (Fankhauser and McDermott 2014) Increasing the availability of technology in vulnerable countries requires knowledge of the current geography of innovation. To that end, this report uses patent data to describe and quantify the invention and global diffusion of technologies for climate change adaptation over recent decades based on a global patent database. Importantly, relying on patent data restricts the scope of the analysis to solutions for adaptation that are at the technological frontier and excludes the role of nontechnological forms of innovation and low-tech options. A particular emphasis is put on the case of low- and middle-income countries, which combine high vulnerability to climate change with low technological resources. The analysis relies on patent data from the World Patent Statistical Database (PATSTAT), maintained by the European Patent Office (EPO), which covers the population of patents filed worldwide. We use the EPO’s new “Y02A” category to identify all patents in PATSTAT pertaining to “technologies for adaptation to climate change.” The classification was released in April 2018 and has so far never been used in empirical analyses. Although innovation scholars and analysts widely use patent data to map technology fields, such data do have some drawbacks, as the report discusses. The patent data are thus complemented with data on foreign direct investment (FDI), which allow us to test the robustness of the results on technology transfer. The literature on the economics of climate-related innovation has focused on mitigation technologies.1 In contrast, few papers so far have analyzed innovation in technologies for adaptation to climate change with an empirical approach (Popp 2019). As in this report, Conway et al. (2015) use a global patent dataset to describe innovation activity and international technology. However, those authors deal only with water-related technologies, while we consider a much larger set of countries and technology fields. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 9 Other papers adopt a different perspective by examining drivers of innovation. Miao and Popp (2014) empirically estimate the impact of historical extreme events (earthquakes, floods, and droughts) on innovation activity in three corresponding risk-mitigating technologies (quakeproof buildings, flood control, and drought-resistant crops). They use patent data from up to 28 countries and find evidence that natural disasters increase risk-mitigating innovations, with a magnitude that differs across types of disaster and technology. Using a similar approach, Hongxiu (2017) and Hu et al. (2018) also observe that past extreme climatic events induced an increase in the number of risk-mitigating technological innovations. The rest of this report proceeds as follows: • Section 2, “Data Issues,” describes our dataset and essential data issues, emphasizing the stren- gths and weaknesses of patent data to measure innovation and technology diffusion. • Section 3, “Invention of Technologies for Climate Change Adaptation,” presents a first set of re- sults showing how global innovative activity in technologies for climate change adaptation has developed over time and space. • Section 4, “International Technology Transfer,” covers the international transfer of adaptation technologies. The latter two subsections relate our findings to results from studies using simi- lar data that have looked at innovation and transfer of climate change mitigation technologies (for example, Dechezleprêtre et al. 2011). • Section 5, “Patenting Activity in Relation to Climate Hazards,” examines the relationship between technology and adaptation needs. • Section 6, “Conclusion and Policy Implications,” summarizes the findings and three important policy implications that follow from them. 10 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis 2 Data Issues Patents as Indicators of Innovation and Technology Transfer Patents are commonly used to measure technological innovation and diffusion. For instance, Dechezleprêtre et al. (2011) adopt a similar approach to examine climate change mitigation technologies. To understand the indicators used below, it is useful to briefly describe how the patent system works. When an individual or organization discovers a new technology, they decide where to market this invention and how to deter imitation by potential competitors. Patenting is a legal way to achieve this, because a patent in a particular country confers the exclusive right to make, use, and sell the protected invention in that country for a maximum period of 20 years. Accordingly, an inventor who plans to market an invention in a particular jurisdiction will patent it there. A set of patents protecting the same invention is called a patent “family.” Most patent families include only one country (often the home country of the inventor, particularly for large countries). A wealth of information is available on patent documents and therefore in the global PATSTAT database. The present study mostly exploits information on the country where the inventor is located, the set of countries in which each invention is patented, the date of the first patent filing within a patent family, and the invention’s technological area. The level of inventive activity is measured by the number of patent families; in other words, the set of countries in which each invention is patented indicates technology transfer from the inventor’s home country to foreign countries. Many articles use this approach to infer innovative activity and international technology diffusion from patent data (Dechezleprêtre et al. 2011; Eaton and Kortum 1996). Using patent data to measure innovation is useful for several reasons: • First, compared with other frequently used proxies, patent data measure the output of the innovation process, while alternative indicators (such as R&D expenditure or the number of researchers employed) measure inputs into this process. • Second, patent data provide not only detailed, disaggregated information on the technology itself but also on many characteristics of its development, such as place of invention, date of filing, names of the inventor and applicant, and sector. • Third, to be considered for patent protection, an invention needs to be marketable—that is, it should potentially have an industrial application. Thus, a patent indicates that an inventor expects some economic benefits from the invention. Because filing a patent is costly (around €30,000 for a European patent), we can safely assume that patents are filed only in places (countries) where inventors see a significant probability that the technology will generate some economic returns. A large fraction of the most economically Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 11 significant innovations in recent years have been patented (Aghion et al. 2016). In a study focusing on Sweden, Svensson (2012) showed that about 61 percent of all patents filed were for commercialized patented technologies. Morgan, Kruytbosch, and Kannankutty (2001) and Griliches (1990) found that, respectively, 47 percent and 55 percent of all patented technologies are commercialized in the United States. However, using patent-related data to measure innovation and technology transfer also features several drawbacks. To start with, inventors may employ alternative methods to protect their innovation—in particular, industrial secrecy or lead-time advantages (Cohen, Nelson, and Walsh 2000). As a result, patents are filed only in countries where intellectual property (IP) protection is sufficiently strong. Otherwise, the inventor may prefer to rely on industrial secrecy or simply to market the innovation without legal protection. These alternative strategies are especially relevant in the context of technology transfer between countries. This is a potentially important limitation for this study because many of the most vulnerable countries also weakly enforce IP rights. The data show few patents in low-income countries. Whether this indicates a measurement problem or low availability of technology is difficult to establish solely with patent data. However, other indicators clearly plead in favor of the second hypothesis. For instance, the median enrollment ratios for tertiary education—a proxy for scientific (and technology) knowledge production—are 6 percent, 28 percent, and 60 percent in the low-income, middle-income, and high-income country groups, respectively (UN DESA 2018). Bernardes and Albuquerque (2003) point out that low-income countries produce far fewer scientific publications than other country income groups. We also test the robustness of the findings with FDI statistics, which show patterns of lower FDI in low-income countries. Another difficulty is that the propensity to patent differs between sectors, depending on the nature of the technology (Cohen, Nelson, and Walsh 2000). We mitigate these issues by looking primarily at the share of a country’s total patent filings that concern technologies for climate change adaptation. This thus accounts for differences in the local enforcement of IP rights and in the propensity to patent across countries, which should apply equally to patents in all technology fields, including adaptation technologies. We also concentrate on time trends, which are immune to this problem if heterogeneities in patenting propensity vary weakly over time. The value of individual patents is also heterogeneous, including across patent offices. For example, inventions filed at the Chinese Patent Office are known to have a lower unit value than inventions filed at the EPO or at the United States Patent and Trademark Office (Boeing and Mueller 2015, 2016). This leads us to restrict parts of our analysis to high-value patented inventions, defined as inventions for which protection has been sought in more than one country.2 Although this is a common solution to the problem of heterogeneity, an alternative option could be to track patent renewal, or the number of times a patent is cited by subsequent patents. However, the geographical scope of the analysis prevents the use of these indicators. Moreover, no global data are available to apply this solution. In particular, there are nearly no data on patent offices in low- and middle-income countries. 12 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis The Y02A Classification As noted earlier, we use PATSTAT as the main database for our analysis. The data, issued biannually by the EPO, include more than 66 million patents that have been filed in 169 national or regional patent offices. In April 2018, the EPO introduced a new classification for patents related to “technologies for adaptation to climate change” (Y02A). Patents classified in this category protect “technologies that allow adapting to the adverse effects of climate change in human, industrial (including agriculture and livestock) and economic activities”,3 which corresponds closely to the UNFCCC (2005) definition of adaptation technologies. Importantly, this classification covers all patent offices included in PATSTAT and was applied retrospectively to all patent applications (not only to new patent filings), ensuring full coverage across space and time. Adaptation patents are divided into six subcategories, covering the main fields of innovation in technologies for climate change adaptation (table 2.1): • Coastal and river protection covers “technologies for adaptation to climate change at coastal zones and river basins,” including technologies for devices that protect homes from flooding, as well as early warning systems. • Water management consists of technologies concerned with “water conservation, efficient wa- ter supply, and efficient water use.” • Infrastructure covers technologies that aim at “adapting or protecting infrastructure (e.g., transport and energy systems) or their operation.” • Agriculture includes “adaptation technologies in agriculture, forestry, livestock, or agroalimen- tary production.” • Health subsumes all technologies concerned with the “adaptation to climate change in human health protection.” • Indirect adaptation refers to “technologies having an indirect contribution to adaptation to cli- mate change” such as climate simulation tools, weather forecasting, and weather surveillance systems. Notably, the technology categories are structured by the economic sector affected, not by climatic threat. Each category is further divided into multiple items. As an illustration, the “coastal and river protection” subcategory includes 36 items. This classification focuses on highly relevant adaptation technologies. However, keep in mind that adaptation also relies on technologies that are not adaptation-specific, such as basic water treatment. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 13 Table 2.1 Technology Fields of Y02A Patents included in the Study Category Description Examples Coastal and Technologies for adaptation to climate Dikes; dams; artificial reefs; groynes; river protection change at coastal zones and river basins real-time flood forecasting Water Technologies related to water Water desalination methods; management conservation, efficient water supply, and saltwater intrusion barriers; water efficient water use filtration systems Infrastructure Technologies for adapting or protecting Floating houses; thermal insulation infrastructure or their operation technologies; passive air cooling Agriculture Technologies for adaption to climate Windbreaks; greenhouse change in agriculture, forestry, livestock technologies; irrigation systems; or agroalimentary production plants tolerant to drought, heat, salinity Health Technologies for adaptation to climate Malaria medical treatment; catalytic change in human health protection converters to control pollutant emission controls Indirect Technologies making an indirect Climate simulation; radar-based adaptation contribution to climate change weather surveillance; real-time adaptation meteorological measurement Source: “Technologies for Adaptation to Climate Change,” Cooperative Patent Classification Subclass Y02A, European Patent Office (EPO). Note: Y02A is a patent classification within the EPO’s World Patent Statistical Database (PATSTAT) for tech- nologies enabling adaptation to climate change. For the full classification, see https://worldwide.espacenet. com/classification?locale=en_EP#!/CPC=Y02A. In evaluating the quality of the Y02A classification, we looked at both inclusion and exclusion errors. First, we checked that all technologies for climate change adaptation listed in a major report on the topic by the United Nations Framework Convention on Climate Change (UNFCCC 2006) were also present in the Y02A classification. Second, to determine whether irrelevant patents had been classified as adaptation technologies, we selected a random sample of 100 patents. For each of these patents, we examined the complete description of the technology, its title, and abstract in PATSTAT. We then checked the relevance of this technology as an “adaptation to climate change” technology by comparing it with the descriptions of technologies listed in UNFCCC (2006). We found that a high number of patents in this random sample (89 percent) were indisputably related to adaptation to climate change as described by the UNFCCC, providing reassurance as to the quality of the Y02A tagging scheme. The fact remains that the boundary between adaptation and nonadaptation technologies is blurry. Most of the solutions that promote economic development also facilitate adaptation to climate shocks. Our solution here was to rely on lists by well-established institutions. 14 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis The full PATSTAT dataset includes all patents filed from 1995 to 2015. This includes 19 million inventions, of which more than 121,000 deal with adaptation. Adaptation inventions thus represent 0.6 percent of all patented inventions in the sample. As a benchmark, we also considered the patents for climate change mitigation technologies, which are covered by another classification.4 The mitigation category is considerably larger (960,187 patents). Note that a patented invention can be classified simultaneously as adaptation and mitigation technology. This is the case for 28 percent (29,019) of all adaptation patents in our sample. An example is a coastal protection system equipped with a wind turbine to pump seawater and store energy and that also includes a desalination station. To compare developments in climate change adaptation technologies with general trends in patented inventions, we built a benchmark for each technology field. We first selected all International Patent Classification (IPC) codes corresponding to adaptation patents, by technology field. Next, we extracted the first four characters of each IPC code, and those with the largest number of observations, until we covered at least 70 percent of the patents in that field and retained all patents with four matching characters in their IPC codes. Foreign Direct Investment Data To challenge the patent-based results, we extracted information on FDI deals for the period 1995–2015 from the Zephyr database provided by Brussels-based business publisher Bureau Van Dijk.5 The objective was to identify foreign investments that could lead to adaptation technology transfers. To identify these deals, we adapted the methodology used in Dussaux, Dechezleprêtre, and Glachant (2018)and foreign direct investments (FDI and applied it to adaptation technologies, as follows: 1. We first selected acquiring firms that have patented at least one adaptation patent in the country where the target firm is located. Using the Y02A PATSTAT classification, we could identify every firm that filed an adaptation patent in a country. We extracted all observations with pairs (applicant firm, application country) specific to adaptation patents and matched these adaptation pairs with the Zephyr database. By doing so, we only retained deals where the acquiring firm had filed at least one adaptation patent in the target country. 2. Second, we used information on target firms’ industrial activity. Using the NACE Rev. 2 classification,6 we identified activities with a potential link to adaptation technologies. We matched the selection of NACE codes linked with technologies for climate change adapta- tion with the industrial activity NACE code of the target firm to only retain target firms in adaptation-related sectors. To identify transfers across countries, we restricted our database to foreign deals, defined as deals where the acquirer and the target country are different. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 15 Measuring Climate Risks To relate the patterns of innovation and technology transfer to the level of climate risk, we needed country-level indicators that measure climate threat. Measuring a country’s vulnerability to climate change is not straightforward owing to both conceptual reasons and data constraints. Before presenting the indicators that we developed, it is first necessary to clarify the concepts used to qualify climate change impacts. The Intergovernmental Panel on Climate Change (IPCC) defines climate change “risk” as “[the] probability of occurrence of hazardous events or trends multiplied by the impacts if these events or trends occur” (IPCC 2014). Risk results from the interaction of three factors: • Natural hazards refer to the possible future occurrence of extreme and nonextreme weather and climate events that may have adverse effects on vulnerable and exposed elements. • Exposure refers to the inventory of elements in an area where hazard events may occur: pres- ence of people; livelihoods; infrastructure; or economic, social, or cultural assets in places that could be adversely affected. • Vulnerability refers to the propensity of exposed elements (such as human beings, their liveli- hoods, and assets) to suffer adverse impacts from hazard events. One of the main vulnerability components is adaptive capacity: “the ability of society and its supporting sectors to adjust to reduce potential damage and to respond to the negative consequences of climate events” (Chen et al. 2015). Access to technology and knowledge is an important component of adap- tive capacity. Others are institutional quality, availability of capital, and so on. For both data and conceptual reasons, the analysis below relies on natural hazard indicators, which are the only unambiguously exogenous factors of climate risk. The degree of vulnerability is influenced by technological capabilities, potentially leading to tautological results: as the size of the losses is influenced by adaptation capacities that include the local availability of technologies, a negative (positive) correlation may simply signal that more (less) innovation increases adaptation capacities. The level of exposure also raises endogeneity concerns (for example, technologies may help people to relocate away from the most-exposed areas) and data availability problems. In contrast, hazards can be measured by physical indicators generated by climate models. No ready-to-use set of indicators quantifies the level of different hazard types at the country level. We therefore combined multiple data sources. We first chose a typology of hazards similar to that of IPCC (2014). We then slightly amended this typology to facilitate the correspondence with the patent classification.7 It includes five threats: sea level rise, temperature extremes, floods, droughts, and storms. (Table 2.2 provides the definition of the corresponding indicators and data sources. Details are provided in appendix C.) Almost all indexes were computed on projections of future climatic threats, based on the idea that innovation prepares for the future. The only exception is storm risk mitigation, which is based on historical events for data availability reasons. This probably did not affect the results, because historic shocks are strongly positively correlated with future ones. 16 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Table 2.2 Definitions and Data Sources of the Five Hazard Indicators Indicator Definition Source Proportion of land areas, adjacent to the ocean, that are lower Sea level rise than the sea level rise and the average height of storm surge ND-GAIN projected by the end of the century Warm spell duration index (WSDI): periods of excessive warmth Temperature using a percentile-based threshold calculated for a five-day Climdex extremes calendar window; projected for 2040–70 under the RCP4.5 scenarioa RX5DAY: monthly maximum precipitation over five consecutive Floods days (mean per year); projected for 2040–70 under the RCP4.5 Climdex scenarioa Change in annual runoff from the baseline projection to the Drought future projection; projected for 2020–40 under the RCP4.5 Aqueduct emission scenarioa Storms Historical number of storms per capita for the period 1900–2015 EM-DAT Note: For details on construction of the five hazard indicators, see appendix C. For the indicators’ correlation coefficients, see appendix A, figures A.4 and A.5. Aqueduct = https:/ /www.wri.org/aqueduct of the World Resources Institute. Climdex = https:/ /www.climdex.org/. EM-DAT = Emergency Events Database of the Centre for Research on the Epidemiology of Disasters. ND-GAIN = University of Notre Dame Global Adap- tation Index. a. RCP4.5 = Representative Concentration Pathway, scenario 4.5, of the Intergovernmental Panel on Cli- mate Change, which refers to a greenhouse gas concentration trajectory for a stabilization scenario (sta- bilizing radiative forcing at 4.5 W m−2 in the year 2100 without ever exceeding that value) assuming the imposition of emissions mitigation policies. In Sections 3 and 4, we relate these indicators to one dimension of a country’s adaptive capacity: technology access. To provide an initial, more general view, figure 2.1 plots the average hazard level versus overall adaptive capacity (technology included, but not exclusively) for 177 countries, using data extracted from the University of Notre Dame Global Adaptation Index (ND-GAIN) database.8 The hazard index features the arithmetic mean values of the five indicators listed in table 2.2, each being normalized so that they range from 0 to 1.9 The graph shows no clear correlation between adaptation capacity and needs, as indicated by a flat regression line. It also splits the countries into three income groups and confirms the observation that low-income countries are highly exposed to projected natural hazards and have low capacities to adapt. This adaptation gap has been frequently reported in the adaptation literature (Barbier and Hochard 2018; Fankhauser and McDermott 2014; Tol 2018). Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 17 Figure 2.1 Country-Specific Climate Hazard and Adaptive Capacity Levels, 1995–2015 1.0 0.8 Overall adaptive capacity score 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1.0 Overall natural hazard score High income Middle income Low income Sources: University of Notre Dame Global Adaptation Index (ND-GAIN) for the overall adaptive capacity score; and calculations of natural hazards based on the following databases: ND-GAIN; Climdex (https:// www.climdex.org/); Aqueduct water risk data, World Resources Institute; and the Emergency Events Data- base (EM-DAT) of the Centre for Research on the Epidemiology of Disasters. Note: Each point describes an individual country. Country income categories use World Bank-defined clas- sifications. The adaptive capacity index ranges from 0 (low capacity) to 6 (high capacity). The hazard index is the arithmetic means of the five hazard indicators listed in table 2.2, each being normalized so that they range from 0 to 1. 18 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis 3 Invention of Technologies for Climate Change Adaptation This section presents the global development and distribution of inventions in climate change technologies for climate change adaptation over the past 20 years. Growth of Climate Adaptation Innovation A first look at the data shows a boom in the number of high-value patented inventions (those patented in at least two countries)—a fourfold increase since 1995 (figure 3.1), which corresponds to an impressive 6.7 percent average annual growth rate (table 3.1). Technologies related to flood protection have experienced the highest growth rates by far. However, these numbers must be put in perspective: they are comparable to the average growth rate for all technologies (5.6 percent) but much lower than the 10.9 percent observed for climate change mitigation technologies over the same period (table 3.1). Figure 3.1 Climate Adaptation Innovation, as Measured by High-Value Patents, 1995–2015 2,000 Number of high-value patented inventions 1,600 1,200 800 400 0 1995 2000 2005 2010 2015 Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Note: “High-value” patented inventions are filed in at least two patent offices. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 19 Table 3.1 Average Annual Growth of Innovation in Different Fields, as Measured by High-Value Patents, 1995–2015 Technology field Average annual growth (%) All climate change adaptation 6.7 Coastal and river protection 17.0 Water management 8.0 Infrastructure 8.0 Agriculture 5.6 Health 7.6 Indirect adaptation 12.1 All technologies in all fields 5.6 All climate change mitigation 10.9 Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Note: “High-value” patented inventions are filed in at least two patent offices. It is common knowledge among patent experts that such upward trends are partly driven by an increase in patenting propensity rather than by a genuine increase in innovation. To control for this factor, and to measure the proportion of R&D efforts directed toward climate-adaptation technologies, figure 3.2 shows the share of patented climate adaptation inventions in total patented innovation. Adaptation inventions, represented by the blue line on the graph, averaged around 0.5 percent of global patenting activity annually during 1995–2015, which is arguably low given the challenges associated with future climate change. In particular, it is considerably lower than the share of climate change mitigation patents over 1995–2015, which averaged 5.7 percent of global inventions annually (see appendix A, figure A.1). More strikingly, this latter percentage has gone down since its peak of about 9 percent in 2012, despite the emerging impact of climate change in many countries. When looking deeper into the data, this reduction appears to mostly concern technology subcategories that facilitate both adaptation to climate change and mitigation of carbon emissions. As noted earlier, more than a quarter of the patents fall into this category. Removing these patents produces a stable percentage (around 0.3 percent of global inventions) throughout the study period (shown by the red line in figure 3.2). This rise and fall in innovation in climate change mitigation technologies over the recent period has been documented in previous studies and has been linked to the evolution of oil prices (Dechezleprêtre et al. 2011). In sum, we observe a constant rate of innovation activity in pure climate change adaptation technologies since 1995. This is bad news because technological progress does not crowd out other forms of innovation. The growing relevance of climate adaptation technologies has seemingly not led to an increase in the proportion of global innovation efforts to develop patented technologies in that field. 20 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Figure 3.2 Innovation for Climate Change Adaptation, as a Share of Total Innovation, 1995–2015 0.6 Share of climate adaptation inventions (%) 0.5 0.4 0.3 0.2 0.1 0 1995 2000 2005 2010 2015 Patent publication year All adaptation Pure adaptation Source: Calculations, based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Note: “Pure adaptation” refers to technologies for climate change adaptation that are not simultaneously classified as mitigation technologies. Geographic Concentration of Climate Adaptation Innovation Turning next to the geographic distribution, innovation appears highly concentrated: the top five inventor countries (China, Germany, Japan, the Republic of Korea, and the United States) account for more than 60 percent of the world’s innovation activity. The United States is by far the world leader, with nearly one-quarter of all high-value adaptation inventions developed between 2010 and 2015 (table 3.2). However, the rapid growth of innovation in China and Korea is notable. These two countries together produced less than 4 percent of all adaptation inventions in 1995 and increased their shares to 8.9 percent and 7.8 percent, respectively, in 2015. Such a high geographic concentration is not specific to this technological area. The same top five countries represent 86.5 percent of the world’s total patented inventions in all technologies and 75.8 percent of high-value climate mitigation innovation (appendix A, table A.1). The fourth column of table 3.2 helps us understand whether these top inventors are “specialized” in adaptation inventions, as defined by the specialization index—the ratio between the share of global adaptation inventions produced by a country and the share of all global inventions produced by that country. The specialization index indicates that the top 10 adaptation technology Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 21 inventors are not relatively specialized in inventing adaptation technologies. Nonetheless, it is worth emphasizing that France, the Netherlands, and Sweden are specialized to some extent in adaptation technologies. The case of China deserves more explanation. We discuss here the number of high-value inventions, which excludes inventions patented in a single country. If we consider all inventions, including those filed in a single country, China becomes by far the most active inventor in the field, filing more than 48 percent of all adaptation patents in the world, because most of its inventions are only filed at home. It is well established that China’s patenting behavior is an outlier in that inventors there file patents of much lower quality than in other countries (Boeing and Mueller 2019; Prud’homme and Zhang 2017). Table 3.2 Top 10 Inventor Countries in Climate Change Adaptation Technologies, 2010–15 Average share of world’s high- value adaptation Specialization Rank Country inventions (%)a indexb Country’s top three technology fields Health, agriculture, indirect 1 United States 23.6 1.02 adaptation Health, agriculture, water 2 Japan 15.8 0.67 management 3 Germany 10.8 1.01 Health, infrastructure, agriculture Health, agriculture, water 4 Korea, Rep. 7.0 0.88 management Agriculture, health, water 5 China 6.5 0.74 management 6 France 5.8 1.59 Health, agriculture, infrastructure Health, water management, 7 United Kingdom 3.8 1.31 agriculture 8 Sweden 2.1 2.05 Agriculture, health, infrastructure  Agriculture, water 9 Canada 2.1 1.23 management, health 10 Netherlands 2.1 1.78 Agriculture, health, infrastructure Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. a. “High-value” patented inventions are defined as patents filed in more than one patent office. b. The “specialization index” is the ratio between the share of global adaptation inventions made by a coun- try and the share of global overall inventions made by that country. A specialization index above 1 indicates that the country represents a higher share of worldwide adaptation inventions than overall inventions and is thus more specialized in adaptation inventions. 22 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis 4International Technology Transfer Given the high geographic concentration of innovation in adaptation technologies, it is of utmost importance to examine whether such technologies diffuse across borders, in particular toward countries with the highest adaptation needs. What ultimately matters for countries is to access technology, whether or not that technology has been developed in the country. This section describes international diffusion patterns. The next section then considers the correlation between patent activity and climate hazard levels. Extent of Cross-Border Technology Transfers A first indicator for measuring international diffusion is the share of patented inventions that are filed in at least two different offices.10 Figure 4.1 compares these shares for three technology groups: climate change adaptation, climate change mitigation, and all technologies. Only 17 percent of adaptation inventions cross at least one border, which is significantly below the average for all technologies (24 percent) and about half that of mitigation technologies (31 percent). Figure 4.1 Technology Transfer Rates, by Invention Type, 2010–15 30 Share of technology transfered (%) 20 10 0 Climate change adaptation All technologies Climate change mitigation Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Note: The technology transfer rate is the share of a country’s technology patents that are also filed in at least one other country. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 23 However, those inventions that do cross at least one border are patented in around five patent offices on average, a figure broadly similar to the rates for climate change mitigation technologies and for technology overall (table 4.1). Although this finding should be interpreted with caution, it suggests that the geographical applicability of climate adaptation inventions is not fundamentally different from that of other technologies. It is also in line with the argument that barriers to technology imports (such as trade barriers and stringent local IP rights) in recipient countries are not specific to adaptation technologies. How, then, can we explain the low transfer rate in figure 4.1? A consistent explanation would be that the average value of individual adaptation patents is low in relative terms, thereby being less likely to warrant foreign patenting. Table 4.1 Average Number of Patent Offices Where Internationally Patented Inventions Are Filed, by Technology Type, 2010–15 Technology field Average number of patent officesa All climate change adaptation 4.88 Coastal and river protection 4.50 Water management 4.54 Infrastructure 4.05 Agriculture 5.46 Health 4.80 Indirect adaptation 5.15 All technologies 4.46 All climate change mitigation 4.51 Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Note: The table includes only inventions patented in at least two countries. a. The number of patent offices is the number of offices where an international patent (a patent filed in at least one foreign country) is filed. The international transfer rate (the share of inventions patented in at least two offices) decreased by half between 2008 and 2015 (figure 4.2). This drastic reduction corresponds to the Chinese patenting boom, which accounts for most domestic patents. However, this too should be interpreted with caution because the boom pertains to all technologies and can thus be observed in all fields. As shown by the dashed line, figure 4.2 indicates no decrease once China is excluded. 24 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Figure 4.2Trends in Climate Adaptation Technology Transfer as a Share of All Adaptation Technology Inventions, 1995–2015 30 Share of adaptation technology transferreds (%) 25 20 15 10 5 0 1995 2000 2005 2010 2015 All countries All countries except China Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Note: Technology transfer refers to the share of technology patents filed in more than one country. This low diffusion rate may be explained by two sets of nonexclusive factors: (a) high barriers to technology transfer (such as tariffs, lax IP enforcement, and limited technological capabilities in potential recipient countries); and (b) lower applicability in the sense that individual innovations are tailored to specific contexts. We return to the interpretation of this result below. Looking at figure 4.3, we find relatively large differences across sectors. Climate adaptation technologies related to agriculture (10.4 percent) and coastal and river protection (9.7 percent) are transferred less often than the average adaptation technology (16 percent). In contrast, indirect adaptation and health adaptation patents are transferred more regularly, each with more than 24 percent of patented inventions filed in more than one country. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 25 Figure 4.3 Transfer Rates of Climate Change Adaptation Technology, by Field, 2010–15 30 Share of technology transfered (%) 20 10 0 n lth t n e re n en ur tio tio tio tu ea ct em ta ta ec ul H ru ric ap ap ot ag st ad ad pr Ag an fra er m ct l In Al riv ire er at d d In W an al ast Co Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Note: The technology transfer rate is the percentage of patents filed in more than one country. “Indirect adaptation” refers to technologies that contribute indirectly to climate change adaptation, such as climate simulation tools, weather forecasting, and weather surveillance systems. Technology Transfers by Country Income Group Table 4.2 gives a more detailed view of international technology flows by considering transfers between different income country groups. It also displays the average score for overall technology transfers as a benchmark (within parentheses). A first notable fact is the overwhelming importance of high-income countries: 93 percent of all exported technologies for climate change adaptation originate from these countries, which also receive 71 percent of all exported inventions. In contrast, low-income countries receive no foreign-patented technologies. As mentioned previously, it could be that foreign inventors protect their technologies transferred toward low-income countries through secrecy (or do not protect them at all). As for middle-income countries, they receive 28 percent of all adaptation transfers, the vast majority of which come from high-income countries, with China accounting for half of these inward transfers. These flows have sharply increased recently: their share was only 7 percent in 1995. Recall, however, that although China and other middle-income countries have become significant recipients, they export few of their patented technologies. 26 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Table 4.2 Distribution between Country Income Groups of Patented Inventions of Technologies for Climate Change Adaptation, 2010–15 (Percent) Destination country Origin country High income Middle income Low income 66 27 0 High income (69) (24) (0) 5 1 0 Middle income (7) (<1) (0) <0.1 <0.01 0 Low income (<0.1) (<0.01) (0) Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Note: Distributions are the percentages of patents filed in both an origin country and at least one destination country. Results for all technologies appear in parentheses. These numbers are comparable with the distribution averages for all technologies (shown within parentheses), except for the slightly higher role of middle-income countries as technology recipients, which is mostly due to China’s particularly high rate as a destination for adaptation patents. A comparison between adaptation and mitigation technologies (appendix A, table A.3) shows that they are equally transferred to middle-income countries (28 percent versus 26 percent). Using FDI Data to Measure Cross-Border Transfers As explained earlier, the analysis of FDI data provides another approach to quantify cross-country technology flows. The data on adaptation-related FDI deals (described in section 2) confirm the picture drawn from patent statistics. The number of deals varies between 26 and 54 per year (figure 4.4) and shows a high variability, probably driven by the small size of the sample (687 deals). The overall trend, however, is not fundamentally different from that displayed in figure 4.2, which shows the evolution of foreign patenting over the same period. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 27 Figure 4.4 Number of FDI Deals in Climate Change Adaptation Technologies, 2000–15 60 Number of adaptation-related deals 40 20 0 2000 2005 2010 2015 Sources: Calculations based on Zephyr database, Bureau Van Dijk, Brussels. Note: N = 687 deals. FDI = foreign direct investment. Table 4.3 compares the share of FDI adaptation deals with the share of international adaptation patent flows (within parentheses) between different country income groups. Nearly all deals originate from firms based in high-income countries. The concentration of FDI deals is thus even stronger than that indicated by patents. Table 4.3 Flow of FDI Deals in Climate Change Adaptation Technologies between Country Income Groups, 2010–15 (Percent) Destination country Origin country High income Middle income Low income 73 27 0 High income (66) (27) (<0.01) 0 0 0 Middle income (5) (1) (0) 0 0 0 Low income (<0.1) (<0.01) (0) Sources: Calculations based on data from Zephyr database (Bureau Van Dijk, Brussels) and World Patent Statistical Database (PATSTAT) (European Patent Office). Note: Results for cross-border transfers of climate change adaptation patents appear in parentheses. N = 243 deals. FDI = foreign direct investment. 28 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Finally, table 4.4 displays the nine largest middle-income countries in terms of inward FDI flows for climate change adaptation, compared with their shares of foreign patents in adaptation technologies. Again, the overall picture is the same for both indicators for most countries, except for China, which attracts a higher share of foreign investments than foreign patents. The case of India deserves particular attention. The absence of Indian patent data in PATSTAT (the only major patent office missing from the database) makes it impossible to measure the contribution of this country to international transfers with this indicator, although FDI data are readily available. Table 4.4 suggests that the country is poorly connected to international technology flows. Table 4.4 Top Acquirers and Importers of Climate Adaptation Technologies among Middle-Income Countries, 1995–2015 FDI Share of world adaptation Share of world adaptation Patent rank Target country deals received (%) patents imported (%) rank 1 China 18.8 9.40 1 2 Brazil 2.6 2.89 2 3 India 1.0 — — 4 South Africa 0.9 1.33 5 5 Turkey 0.6 0.70 6 6 Russian Federation 0.4 1.87 4 7 Peru 0.1 0.18 13 8 Indonesia 0.1 0.11 17 9 Mexico 0.1 2.51 3 Sources: Calculations based on data from Zephyr database (Bureau Van Dijk, Brussels) and World Patent Statistical Database (PATSTAT) (European Patent Office). Note: — = not available. FDI = foreign direct investment. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 29 Patenting Activity in Relation to Climate Hazards 5 How do the above patterns of innovation and technological diffusion correspond to countries’ needs for climate change adaptation technology? Figure 5.1 shows the relationship between a country’s level of patented invention (y-axis) and the average hazard index (x-axis). No clear positive correlation is observed. Focusing on the countries with the highest hazard scores, unsurprisingly, we see that almost all middle-income countries in this group produce little innovation (they are below the regression line), in contrast with highly exposed, high-income countries such as Australia, Japan, Korea, Singapore, and the United States. Figure 5.1 Relationship between Adaptation Technology Invention and Climate Hazards in High- and Middle-Income Countries, 2010–15 6 Log high-value adaptation inventions 4 2 0 –2 –4 0 0.2 0.4 0.6 0.8 Overall natural hazard score High income countries Middle income countries Sources: Calculations based on World Patent Statistical Database (PATSTAT), European Patent Office; Uni- versity of Notre Dame Global Adaptation Index (ND-GAIN); Climdex indexes (https:/ /www.climdex.org/); Aqueduct water risk data, World Resources Institute; and Emergency Events Database (EM-DAT), Centre for Research on the Epidemiology of Disasters. Note: Each point describes an individual country. Country income categories use World Bank-defined classi- fications. Low-income countries are excluded for lack of sufficient patent data. The log number of inventions is the 2010–15 annual average. The hazard index is the arithmetic mean of the five hazard indicators listed in table 2.2, each being normalized to range from 0 to 1. 30 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Availability of Adaptation Technology, by Country Income Level What drives this gap between middle-income and high-income countries? It could simply reflect the differences in general innovative capacity. Figure 5.2 supports this statement, as the gap disappears as soon as innovation for adaptation is expressed as a share of total patents. A quasi- flat regression line however provides no evidence of R&D directed toward adaptation. Figure 5.2 Relationship between Climate Hazards and Adaptation Technology Inventions (as a Share of All Technology) in High- and Middle-Income Countries, 2010–15 –2 Log share climate adaptation in all inventions –3 –4 –5 –6 –7 0 0.2 0.4 0.6 0.8 Overall natural hazard score High income countries Middle income countries Source: Calculations based on World Patent Statistical Database (PATSTAT), European Patent Office; Uni- versity of Notre Dame Global Adaptation Index (ND-GAIN); Climdex indexes (https:/ /www.climdex.org/); Aqueduct water risk data, World Resources Institute; and Emergency Events Database (EM-DAT), Centre for Research on the Epidemiology of Disasters. Note: Each point describes an individual country. Country income categories use World Bank-defined clas- sifications. Low-income countries are excluded for lack of sufficient patent data. The vertical axis shows the number of adaptation inventions made in the country as a share of all technology inventions made in this country (2010–15 annual average). The hazard index is the arithmetic means of the five hazard indicators listed in table 2.2, each being normalized so that they range from 0 to 1. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 31 The availability of technologies for adaptation to climate change is not only a matter of domestic innovation. Countries may also benefit from imports of technologies invented abroad. Do inward flows of foreign technologies compensate for this domestic innovation deficit? Probably not. As shown in figure 5.3, the slope of the regression line is not fundamentally steeper than in figure 5.1, but most middle-income countries again remain below the line.11 Importantly, this technology gap is not specific to adaptation technologies: the average share of a country’s adaptation inventions in its total number of inventions is roughly the same in high-income countries (0.98 percent) as in middle-income ones (0.92 percent). Figure 5.3 Relationship between Climate Hazards and Imports of Climate Change Adaptation Technologies in High- and Middle-Income Countries, 2010–15 6 Log climate foreign adaptation patents 4 2 0 –2 0 0.2 0.4 0.6 0.8 Overall natural hazard score High income countries Middle income countries Sources: Calculations based on World Patent Statistical Database (PATSTAT), European Patent Office; Uni- versity of Notre Dame Global Adaptation Index (ND-GAIN); Climdex indexes (https:/ /www.climdex.org/); Aqueduct water risk data, World Resources Institute; and Emergency Events Database (EM-DAT), Centre for Research on the Epidemiology of Disasters. Note: Each point describes an individual country. Country income categories use World Bank-defined clas- sifications. Low-income countries are excluded for lack of sufficient patent data. The vertical axis shows the log number of foreign climate change adaptation inventions patented in the country (2010–15 annual average). The hazard index is the arithmetic means of the five hazard indicators listed in table 2.2, each being normalized so that they range from 0 to 1. 32 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Availability of Adaptation Technology, by Climate Hazard Type Beyond this general view, the data also allow us to identify the types of hazard with the widest gaps between the need for and access to climate adaptation technology. For each of the five hazard types, figure 5.4 displays the correlation coefficient between technology availability (the sum of high-value domestic inventions and imported patented inventions) and hazard levels in high- and middle-income countries.12 Recall that a correlation is -1 in case of perfect negative correlation and +1 in case of perfect positive correlation. One hazard type shows significant negative correlation: mitigating impacts from temperature extremes. This mostly concerns pollution reduction technologies, greenhouse technologies, plant varieties adapted to hot environments, aquaculture, and air conditioning technologies. Other correlation coefficients are nonsignificantly different from zero. These patterns could reflect the influence not only of adaptation needs but also other factors such as a country’s technological capabilities. In particular, countries with strong technological capabilities (like China, Japan, western European countries, and the United States) are less exposed to heat waves than countries such as Costa Rica, Ecuador, and Peru, which have low technological capabilities. The access to technologies that mitigate climate change impacts may be driven by the existence of a local demand for these technologies or by the availability of technological capabilities to develop them. The local demand for adaptation technologies can be increased, for instance, by international climate adaptation funding. Because financial issues are the principal barrier to accessing technology in low- and middle-income countries (UNFCCC 2018), these funds are intended to create demand for adaptation solutions in the receiving country. The OECD climate-related development finance data allow us to access bilateral climate funding dedicated to adaptation projects for the period 2009–15.13 However, the geographical distribution of this support cannot explain the differences we observe across sectors. We have also tried to disentangle these factors using a regression approach. However, the data are too limited to obtain robust results. (See appendix B for preliminary results of two simple Poisson regressions). Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 33 Figure 5.4Correlation between Availability of Adaptation Technologies and Climate Hazard Levels in High- and Middle-Income Countries, 2010–15 0.6 Correlation of technology availability and climate hazard 0.4 0.2 0 –0.2 –0.4 –0.6 –0.8 Drought Flood Sea level Storms Temperature Sources: Calculations based on World Patent Statistical Database (PATSTAT), European Patent Office; Uni- versity of Notre Dame Global Adaptation Index (ND-GAIN) indicators; Emergency Events Database (EM- DAT), Centre for Research on the Epidemiology of Disasters; Climdex indexes (https://www.climdex.org/); and Aqueduct water risk data, World Resources Institute. Note: The technology variables are averages for the period 2010–15. “Technology availability” is the sum of high-value adaptation inventions (that is, with patents filed in at least two countries) made by a given coun- try and the patents imported by that country. For details on construction of the five hazard indicators, see appendix C. Vertical black lines represent confidence intervals at the 95 percent level. 34 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis 6 Conclusion and Policy Implications T his report is the first global analysis of innovation and the international diffusion of technologies for adaptation to climate change to rely on patent data. It draws a clear picture of a technology gap between low- and middle-income countries and high-income countries. With the exception of China, low- and middle-income countries show low patented innovation outcomes and limited access to foreign-protected technologies for adaptation to climate change. Although the report does not include causal evidence, its analysis suggests that the main driver of this disparity is the weakness of technological capabilities in the developing world—a weakness not specific to adaptation-related technologies. Results also indicate that technological availability and adaptation needs are poorly aligned, particularly regarding technologies for adaption to temperature increases. The observed low innovation and technology transfer activity in adaptation technologies toward middle-income countries—and all the more toward low-income countries—falls below that of climate change mitigation technologies. This is both puzzling and problematic. It is puzzling because, from an economic point of view, climate change mitigation is a pure public good in the sense that it generates benefits for all. In contrast, adaptation benefits are partly private. Hence, from a general economic point of view, it is not obvious that there could be less of a business case for the latter. It is problematic when it comes to low-income countries because they are already highly exposed to climate change impacts, yet transferring climate mitigation solutions toward these countries is less urgent for the time being because they have low emissions so far. It is true that low-income countries may have a greater need for simple, nontechnological solutions given their current technological capacity. Nevertheless, unless such solutions crowd out other forms of innovation, better technologies can only facilitate adaptation and build up technology-absorptive capacities that will be crucial for their future. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 35 From these observations follow three important policy implications: • First, technological capacity building is an essential ingredient to narrow the gap. This mes- sage is not at all specific to technologies for adaptation to climate change. • Second, economic forces seem unable to transform local adaptation needs into demand for adaptation technology on the markets. Solving this problem requires a better understanding of the market failures that hinder demand, a precondition for designing demand-pull policies in the relevant sectors (with public investments, subsidies, and other policy tools). • Third, perhaps less importantly, the data do not suggest that the applicability of individual adaptation inventions to specific national contexts is lower than in other sectors. It follows that relying on foreign technologies is not less relevant than domestic innovation. The pro- motion of technology transfer should thus be a pillar of the policies implemented in this area. This work also suggests several avenues for future research. The first and most important one is to conduct sector-specific studies to identify the factors that hinder the functioning of the markets for these patented technologies. This is the level of analysis required to derive sound policy recommendations. The second is to examine the contribution of these technologies to climate change adaptation in the Global South. How crucial are they? How do they relate to low- tech solutions and other forms of innovation? How do economic incentives shape technology adoption? Third, the data used in this study include very disaggregated information that could be exploited to better identify which patents are relevant in a given sector or national context. Combined with country-level technology needs assessments, it could be exploited to prioritize the transfer of the most useful patents. These are the initial elements of a research agenda to increase the contribution of technical progress to climate change adaptation. • 36 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis A Appendix A | Supplementary Tables and Figures Table A.1 Top 10 Inventor Countries in Climate Change Mitigation Technologies, 2010–15 Average share of world Average share of all world Rank Country high-value mitigation inventions (%)a mitigation inventions (%) 1 Japan 25.8 22.9 2 United States 23.0 11.4 3 Germany 12.2 6.6 4 Korea, Rep. 8.4 12.5 5 China 6.4 33.1 6 France 4.7 2.2 7 United Kingdom 3.1 1.3 8 Italy 1.7 0.9 9 Canada 1.6 0.8 10 Sweden 1.2 0.5 Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. a. “High-value” patented inventions are filed in at least two patent offices. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 37 Figure A.1 Climate Change Mitigation Innovation, as a Share of Total Innovation, 1995–2015 10 Share of climate mitigation inventions (%) 8 6 4 2 0 1995 2000 2005 2010 2015 Patent publication year Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Table A.2 Top 10 Inventor Countries in Climate Change Adaptation Technologies, by Sector, 2010–15 Coastal Water Indirect Rank Country protection management Infrastructure Agriculture Health adaptation 1 United States 1 13 11 29 30 17 2 Japan 2 17 15 17 45 5 3 Germany 1 10 18 14 55 3 4 Korea, Rep. 5 18 14 29 30 5 5 China 3 19 18 36 21 4 6 France 1 9 16 22 45 9 7 United Kingdom 1 22 15 23 31 9 8 Sweden 0 9 13 12 60 7 9 Canada 3 21 9 38 20 10 10 Netherlands 2 10 17 51 21 2 Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Note: For definitions of each sector, see table 2.1. 38 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Table A.3 Transfer between Country Income Groups of Patented Technologies for Climate Change Adaptation and Mitigation, 2010–15 Destination country Origin country High income Middle income Low income 66 27 0 High income (68) (26) (0) 5 1 0 Middle income (5) (1) (0) <0.1 <0.01 0 Low income (<0.1) (<0.1) (0) Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Note: Distributions are the percentages of patents filed in both an origin country and at least one destination country. Results for climate change mitigation technologies appear in parentheses. Table A.4 Top 10 Importing Countries of Climate Change Adaptation Technologies, 1995–2015 Average share of world Rank Country received patents (%) 1 United States 13.3 2 China 9.4 3 Australia 8.0 4 Canada 6.9 5 Japan 5.9 6 Germany 5.0 7 United Kingdom 4.6 8 France 4.5 9 Korea, Rep. 3.8 10 Brazil 2.9 Source: Calculations based on World Patent Statistical Database (PATSTAT) data, European Patent Office. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 39 Table A.5 Top 15 Acquirer Countries in FDI Deals Related to Climate Change Adaptation Technologies 1995–2015 Rank Acquirer country Number of deals Share of worldwide deals made (%) 1 United States 187 24.0 2 Japan 180 23.1 3 Germany 135 17.4 4 France 54 6.9 5 Netherlands 45 5.8 6 Denmark 42 5.4 7 Finland 38 4.9 8 Switzerland 26 3.3 9 Korea, Rep. 18 2.3 10 United Kingdom 12 1.5 11 Norway 10 1.3 12 Taiwan 8 1.0 13 Sweden 7 0.9 14 New Zealand 4 0.5 15 India 3 0.4 Source: Calculations based on Zephyr database, Bureau Van Dijk, Brussels. Note: N = 778 deals. FDI = foreign direct investment. 40 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Figure A.2Relationship between Climate Adaptation Technology Invention and Climate Hazards, Middle-Income Countries, 2010–15 Annual Average 7 Russian Federation Brazil Turkey China South Africa 5 Romania Colombia Log adaptation patent invented (no.) Ukraine Egypt Arab Rep. Indonesia 3 Mexico Bangladesh Morocco Lebanon Malaysia 1 Tunisia Thailand Kazakhstan Philippines –1 Bulgaria Jordan Vietnam –3 Moldova Ecuador Mauritius Belarus Sudan Kenya Sri Lanka Cuba 0 0.1 0.2 0.3 0.4 0.5 0.6 –5 Overall natural hazard score Source: Calculations based on World Patent Statistical Database (PATSTAT), European Patent Office; Univer- sity of Notre Dame Global Adaptation Index (ND-GAIN) indicators; Climdex indexes (https:/ /www.climdex. org/); Aqueduct water risk data, World Resources Institute; and Emergency Events Database (EM-DAT), Centre for Research on the Epidemiology of Disasters. Note: Each point describes an individual country. “Middle-income countries” are defined by World Bank classifications. The log numbers of inventions are the 2010–15 annual averages. The hazard index is the arithmetic means of the five hazard indicators listed in table 2.2, each being normalized so that they range from 0 to 1. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 41 Figure A.3Relationship between Climate Adaptation Technology Imports and Climate Hazards, Middle-Income Countries, 2010–15 Annual Average 7 China Russian Federation South Africa Mexico Turkey 5 Ukraine Log adaptation patent imported (no.) Brazil Morocco Peru Colombia 3 Bulgaria Malaysia Tunisia Philippines Costa Rica Romania 1 Egypt Arab Rep. Ecuador Serbia Cuba –1 North Macedonia Moldova Guatemala Dominican Republic Georgia Albania Honduras –3 0 0.1 0.2 0.3 0.4 0.5 0.6 –5 Overall natural hazard score Source: Calculations based on World Patent Statistical Database (PATSTAT), European Patent Office; Univer- sity of Notre Dame Global Adaptation Index (ND-GAIN) indicators; Climdex indexes (https:/ /www.climdex. org/); Aqueduct water risk data, World Resources Institute; and Emergency Events Database (EM-DAT), Centre for Research on the Epidemiology of Disasters. Note: Each point describes an individual country. “Middle-income countries” are defined by World Bank clas- sifications. The log number of imports are the 2010–15 annual averages. The hazard index is the arithmetic means of the five hazard indicators listed in table 2.2, each being normalized so that they range from 0 to 1. 42 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Figure A.4 Correlation between Number of Adaptation Inventions and Climate Hazard Levels in High- and Middle-Income Countries, 2010–15 0.6 Correlation coefficient of adaptation invention to hazard 0.4 0.2 0 –0.2 –0.4 –0.6 –0.8 Drought Flood Sea level Storms Temperature Source: Calculations based on World Patent Statistical Database (PATSTAT), European Patent Office; Univer- sity of Notre Dame Global Adaptation Index (ND-GAIN) indicators; Climdex indexes (https:/ /www.climdex. org/); Aqueduct water risk data, World Resources Institute; and Emergency Events Database (EM-DAT), Centre for Research on the Epidemiology of Disasters. Note: The technology variables are averages for the period 2010–15. For details on construction of the five hazard indicators, see appendix C. Vertical black lines represent confidence intervals at 95 percent level. “Invention” in this figure represents the number of high-value adaptation inventions (patents filed in at least two countries) made by a given country Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 43 Figure A.5Correlation between Number of Imported Adaptation Inventions and Climate Hazard Levels in High- and Middle-Income Countries, 2010–15 Correlation coefficient of adaptation imports to hazard 0.6 0.4 0.2 0 –0.2 –0.4 –0.6 –0.8 Drought Flood Sea level Storms Temperature Source: Calculations based on World Patent Statistical Database (PATSTAT), European Patent Office; Univer- sity of Notre Dame Global Adaptation Index (ND-GAIN) indicators; Climdex indexes (https:/ /www.climdex. org/); Aqueduct water risk data, World Resources Institute; and Emergency Events Database (EM-DAT), Centre for Research on the Epidemiology of Disasters. Note: The technology variables are averages for the period 2010–15. For details on construction of the five hazard indicators, see appendix C. Vertical black lines represent confidence intervals at 95 percent level. “Imports” in this figure represent the number of foreign adaptation patents (invented by a foreign country) filed in a given country 44 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Table A.6 Economies and Income Groups Included in the Study Economy High income Middle income Low income Afghanistan ■ Albania ■ Algeria ■ Angola ■ Antigua and Barbuda ■ Argentina ■ Armenia ■ Australia ■ Austria ■ Azerbaijan ■ Bahamas, The ■ Bahrain ■ Bangladesh ■ Barbados ■ Belarus ■ Belgium ■ Belize ■ Benin ■ Bhutan ■ Bolivia ■ Bosnia and Herzegovina ■ Botswana ■ Brazil ■ Brunei Darussalam ■ Bulgaria ■ Burkina Faso ■ Burundi ■ Cambodia ■ Cameroon ■ Canada ■ Central African Republic ■ Chad ■ Chile ■ China ■ Colombia ■ Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 45 Table A.6 (Cont.) Economy High income Middle income Low income Comoros ■ Congo, Dem. Rep. ■ Congo, Rep. ■ Costa Rica ■ Côte d’Ivoire ■ Croatia ■ Cuba ■ Cyprus ■ Czech Republic ■ Denmark ■ Djibouti ■ Dominica ■ Dominican Republic ■ Ecuador ■ Egypt, Arab Rep. ■ El Salvador ■ Equatorial Guinea ■ Eritrea ■ Estonia ■ Eswatini ■ Ethiopia ■ Fiji ■ Finland ■ France ■ Gabon ■ Gambia, The ■ Georgia ■ Germany ■ Ghana ■ Greece ■ Grenada ■ Guatemala ■ Guinea ■ Guyana ■ Haiti ■ 46 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Table A.6 (Cont.) Economy High income Middle income Low income Honduras ■ Hungary ■ Iceland ■ Indonesia ■ Iran, Islamic Rep. ■ Iraq ■ Ireland ■ Israel ■ Italy ■ Jamaica ■ Japan ■ Jordan ■ Kazakhstan ■ Kenya ■ Korea, Rep. ■ Kuwait ■ Kyrgyz Republic ■ Lao PDR ■ Latvia ■ Lebanon ■ Lesotho ■ Liberia ■ Libya ■ Lithuania ■ Luxembourg ■ Madagascar ■ Malawi ■ Malaysia ■ Maldives ■ Mali ■ Malta ■ Mauritania ■ Mauritius ■ Mexico ■ Moldova ■ Mongolia ■ Montenegro ■ Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 47 Table A.6 (Cont.) Economy High income Middle income Low income Morocco ■ Mozambique ■ Myanmar ■ Namibia ■ Nepal ■ Netherlands ■ New Zealand ■ Nicaragua ■ Niger ■ Nigeria ■ North Macedonia ■ Norway ■ Oman ■ Pakistan ■ Panama ■ Papua New Guinea ■ Paraguay ■ Peru ■ Philippines ■ Poland ■ Portugal ■ Qatar ■ Romania ■ Russian Federation ■ Rwanda ■ Samoa ■ São Tomé and Príncipe ■ Saudi Arabia ■ Senegal ■ Serbia ■ Seychelles ■ Sierra Leone ■ Singapore ■ Slovak Republic ■ Slovenia ■ Solomon Islands ■ Somalia ■ 48 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Table A.6 (Cont.) Economy High income Middle income Low income South Africa ■ Spain ■ Sri Lanka ■ St. Kitts and Nevis ■ St. Lucia ■ St. Vincent and the Grenadines ■ Sudan ■ Suriname ■ Sweden ■ Switzerland ■ Syrian Arab Republic ■ Tajikistan ■ Tanzania ■ Thailand ■ Togo ■ Tonga ■ Trinidad and Tobago ■ Tunisia ■ Turkey ■ Turkmenistan ■ Uganda ■ Ukraine ■ United Arab Emirates ■ United Kingdom ■ United States ■ Uruguay ■ Uzbekistan ■ Vanuatu ■ Venezuela, RB ■ Vietnam ■ Yemen, Rep. ■ Zambia ■ Zimbabwe ■ Note: Country income categories are based on World Bank-defined classifications. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 49 Appendix B | Regressions B W e provide the results of two simple Poisson regressions in tables B.1 and B.2. The dependent variables are inventioni (the number of patented inventions of technologies for climate change adaptation invented in country i), and (the number of foreign inventions of technologies for climate change adaptation patented in country i in the period 2005–10). More specifically, we estimate the following count models: inventioni = exp(α1 hazardi + β1 invention_stocki + γ1 populationi + ε1i ) and (B.1) importsi = exp(α1 hazardi + β2 invention_stocki + γ2 populationi + ε2i ) and (B.2) which include three explanatory factors: • hazardi , which captures the local demand measured by the country’s hazard level in country • invention_stocki , which captures the level of technological capabilities proxied by the (dis- counted) stock of patented inventions in all technologies filed in the country • populationi , which represents the country size proxied by its population. ε1i and ε2i are two error terms. The results confirm that country’s technological capability is a strong predictor of innovation and technology imports in all fields, as is the country’s size. The influence of local demand is difficult to analyze. Some corresponding coefficients are negative (floods for innovation, droughts and floods for imports) in these two regressions. When we change or add control variables, the signification of the hazard-related coefficient is modified among hazards, but the puzzle remains as we still get a negative coefficient for some categories. No theory can explain a significant and negative coefficient relative to the hazard variable, indicating that we have not found an appropriate model. 50 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Table B.1 Relationship between Domestic Innovation and Climate Hazard Levels, Population, and Stock of Patented Inventions Dependent Temperature variable All Droughts Floods Sea level rise Storms extremes -1.519* 0.235 -1.393*** -0.584 -0.571 -0.755 Hazard (0.813) (0.744) (0.442) (0.675) (0.780) (1.390) 0.135*** 0.186* 0.140* 0.185 0.209** 0.111 Population (0.040) (0.100) (0.076) (0.162) (0.104) (0.088) Stock of 0.763*** 0.651*** 0.648*** 0.581*** 0.687*** 0.787*** patented (0.026) (0.051) (0.040) (0.080) (0.056) (0.057) inventions # observations 118 95 99 87 91 104 Note: The dependent variable is the number of patented inventions of technologies for climate change adaptation invented in the country in the period 2005–10. The coefficients are estimated from a Poisson regression. Robust standard errors in parentheses. ***p < 0.01 **p < 0.05 *p < 0.1 Table B.2 Relationship between Technology Imports and Climate Hazard levels, Population, Stock of Patented Inventions, and Trade Barriers Dependent Temperature variable All Droughts Floods Sea-level rise Storms extreme -0.389 -1.741*** -1.277** -0.634 0.548*** -0.606 Hazard (0.798) (0.507) (0.640) (0.514) (0.209) (0.649) 0.450*** 0.451*** 0.451*** 0.279*** 0.374*** 0.459*** Population (0.077) (0.117) (0.075) (0.086) (0.077) (0.069) Stock of 0.356*** 0.329*** 0.335*** 0.334*** 0.338*** 0.382*** patented (0.050) (0.071) (0.048) (0.049) (0.054) (0.045) inventions 0.009 0.103 0.264* 0.150 0.159 -0.101 Trade barriers (0.156) (0.195) (0.136) (0.174) (0.140) (0.105) # observations 72 65 66 59 62 68 Note: The dependent variable is the number of foreign inventions of technologies for climate change adap- tation patented in the country in the period 2005–10. The coefficients are estimated from a Poisson regres- sion. Robust standard errors in parentheses. ***p < 0.01 **p < 0.05 * p< 0.1 Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 51 Appendix C | Construction of the C Five Hazard Indicators T he sea level rise indicator is the original University of Notre Dame Global Adaptation Index (ND-GAIN) indicator whose definition is provided in table 2.2. Importantly, it corresponds to projections by the end of the century. Temperature and flood indicators rely on Climdex (https:/ /www.climdex.org/) simulations for the period 2040–70 under the IPCC’s RCP4.5 scenario. The final score for both indicators is the 14 average of the resulting indicator computed in 20 climate models as part of the CMIP5 program.15 The drought indicator is the projected change in annual runoff under the RCP4.5 scenario from the Aqueduct water stress database of the World Resources Institute. For each country, we compute the mean of the basin level’s score for the water supply indicator under the RCP4.5 scenario for the years 2020, 2030, and 2040. We then average these four indicators at the country level. Last, the storms indicator only includes past events because we could not find any relevant projections. It is the country’s average annual number of storms per capita for the period 1900– 2015. The historical number of storms per country for the period 1900–2015 is extracted from the Centre for Research on the Epidemiology of Disasters (CRED) Emergency Events Database (EM-DAT). The assumption here is that past events are strongly positively correlated with future threats. 52 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Endnotes For a recent survey of the research, see Popp (2019). 1 A commonly used alternative control for differences between patent offices would be to include only triadic pa- 2 tents in our sample (for example, patents filed at US, European, and Japanese patent offices). However, because we are specifically addressing technology transfer between countries, this selection would make a meaningful global analysis impossible. “Technologies for Adaptation to Climate Change,” Cooperative Patent Classification (CPC) Subclass Y02A, 3 Espacenet website, European Patent Office: https://worldwide.espacenet.com/classification?locale=en_EP#!/ CPC=Y02A. Specifically, all the mitigation technologies are classified as Y02B, Y02C, Y02D, Y02E, Y02P, Y02T, or Y02W. 4 For more information on the Zephyr database, see the Bureau Van Dijck website: https://www.bvdinfo.com/en-gb/ 5 our-products/data/specialist/zephyr. “NACE Rev. 2” refers to Statistical Classification of Economic Activities in the European Community (NACE, for the French 6 “nomenclature statistique des activités économiques dans la Communauté européenne”). It is the industry standard classification system used in the European Union. The current version is revision 2. We have excluded ocean acidification because the IPCC does not mention any technology mitigating this pro- 7 blem. Note that the IPCC concepts of “risk” and “hazard” are referred to, respectively, as “vulnerability” and “exposure” 8 in ND-GAIN. The present report exclusively relies on IPCC definitions. Choosing an arithmetic mean is conventional. Using a more sophisticated weighting rule would require more 9 data and be less transparent. Moreover, we do not think that it would modify the overall messages conveyed by figures 5.1–5.3 (in section 5). Note that the most important figure, figure 5.4, presents hazard-specific results. 10 The structure of the patent data distinguishes between the inventing (home) and receiving (foreign) countries of a patented invention. The transfer rate is then calculated as the share of a country’s inventions that are also filed in at least one foreign country. 11 In appendix A, figure A.2 and figure A.3 are variants of figure 5.1 and figure 5.3, restricted to middle-income countries and displaying country names for readers interested in particular cases). 12 In appendix A, two graphs (figure A.4 and figure A.5) give these correlation coefficients, distinguishing between domestic and imported adaptation technology innovations 13 For these data, see the Excel files for “Climate-related development finance at the activity level, recipient pers- pective” on the OECD web page, “Climate Change: OECD DAC External Development Finance Statistics”: http:// www.oecd.org/dac/financing-sustainable-development/development-finance-topics/climate-change.htm. 14 A Representative Concentration Pathway (RCP) scenario refers to a greenhouse gas concentration trajectory adopted by the IPCC. Four pathways were used for climate modeling and research for the IPCC Fifth Assessment Report in 2014. RCP4.5 is a stabilization scenario (stabilizing radiative forcing at 4.5 W m−2 in the year 2100 without ever exceeding that value) that assumes the imposition of emissions mitigation policies. 15 The Climate Model Intercomparison Project, Phase 5 (CMIP5) refers to a standard experimental protocol for studying the output of coupled atmosphere-ocean general circulation models (AOGCMs). It includes 35 climate model experiments to evaluate the relevance of such models in simulating the recent past climate, to produce projections of future climate conditions, and to analyze the factors leading to differences among the models’ pro- jections. For more information about CMIP5, see the Program for Climate Model Diagnosis & Intercomparison website: https://pcmdi.llnl.gov/mips/cmip5/. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 53 References Aghion, Philippe, Antoine Dechezleprêtre, David Hémous, Ralf Martin, and John van Reenen. 2016. “Carbon Taxes, Path Dependency, and Directed Technical Change: Evidence from the Auto Industry.” Journal of Political Economy 124 (1): 1–51. https://doi.org/10.1086/684581. Barbier, Edward B., and Jacob P. Hochard. 2018. “The Impacts of Climate Change on the Poor in Disadvantaged Regions.” Review of Environmental Economics and Policy 12 (1): 26–47. https://doi.org/10.1093/reep/rex023. Bernardes, Américo Tristão, and Eduardo da Motta e Albuquerque. 2003. “Cross-Over, Thresholds, and Interactions between Science and Technology: Lessons for Less-Developed Countries.” Research Policy 32 (5): 865–85. https://doi.org/10.1016/S0048-7333(02)00089-6. Boeing, Philipp, and Elisabeth Mueller. 2015. “Measuring Patent Quality in International Comparison: Index Development and Application to China.” Discussion Paper No. 15-051, Leibniz Centre for European Economic Research (ZEW), Mannheim, Germany. Boeing, Philipp, and Elisabeth Mueller. 2016. “Measuring Patent Quality in Cross-Country Comparison.” Economics Letters 149 (December): 145–47. https:/ /doi.org/10.1016/j. econlet.2016.10.039. Boeing, Philipp, and Elisabeth Mueller. 2019. “Measuring China’s Patent Quality: Development and Validation of ISR Indices.” China Economic Review 57 (October): 101331. https://doi. org/10.1016/j.chieco.2019.101331. Chen, C., I. Noble, J. Hellmann, J. Coffee, M. Murillo, and N. Chawla. 2015. “University of Notre Dame Global Adaptation Index: Country Index Technical Report.” University of Notre Dame, South Bend, IN. Cohen, Wesley, Richard Nelson, and John Walsh. 2000. “Protecting Their Intellectual Assets: Appropriability Conditions and Why U.S. Manufacturing Firms Patent (or Not).” Working Paper No. 7552, National Bureau of Economic Research, Cambridge, MA. https:/ /doi. org/10.3386/w7552. Conway, Declan, Antoine Dechezleprêtre, Ivan Haščič, and Nick Johnstone. 2015. “Invention and Diffusion of Water Supply and Water Efficiency Technologies: Insights from a Global Patent Dataset.” Water Economics and Policy 1 (4): 1–34. https://doi.org/10.1142/ S2382624X15500101. Dechezleprêtre, Antoine, Matthieu Glachant, Ivan Haščič, Nick Johnstone, and Yann Ménière. 2011. “Invention and Transfer of Climate Change-Mitigation Technologies: A Global Analysis.” Review of Environmental Economics and Policy 5 (1): 109–30. https:/ /doi. org/10.1093/reep/req023. 54 / Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis Dussaux, Damien, Antoine Dechezleprêtre, and Matthieu Glachant. 2018. “Intellectual Property Rights Protection and the International Transfer of Low-Carbon Technologies.” Working Paper No. 323, Centre for Climate Change Economics and Policy; Working Paper No. 288, Grantham Research Institute on Climate Change and the Environment, London School of Economics and Political Science. Eaton, Jonathan, and Samuel Kortum. 1996. “Trade in Ideas Patenting and Productivity in the OECD.” Journal of International Economics 40 (3–4): 251–78. https://doi.org/10.1016/0022- 1996(95)01407-1. Fankhauser, Samuel, and Thomas K. J. McDermott. 2014. “Understanding the Adaptation Deficit: Why Are Poor Countries More Vulnerable to Climate Events than Rich Countries?” Global Environmental Change 27 (1): 9–18. https://doi.org/10.1016/j.gloenvcha.2014.04.014. GCA (Global Commission on Adaptation). 2019. “Adapt Now: A Global Call for Leadership on Climate Resilience.” Report, GCA, Rotterdam; World Resources Institute, Washington, DC. Griliches, Zvi. 1990. “Patent Statistics as Economic Indicators: A Survey.” Journal of Economic Literature 28 (4): 1661–1707. Hongxiu, Li. 2017. “Innovation as Adaptation to Natural Disasters.” Working Paper No. 1709, Department of Economics, University of Waterloo, Ontario. Hu, Hui, Ting Lei, Jie Hu, Songlai Zhang, and Philip Kavan. 2018. “Disaster-Mitigating and General Innovative Responses to Climate Disasters: Evidence from Modern and Historical China.” International Journal of Disaster Risk Reduction 28: 664–73. https:/ /doi.org/10.1016/j. ijdrr.2018.01.022. IPCC (Intergovernmental Panel On Climate Change). 2014. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Edited by Core Writing Team, R. K. Pachauri, and L. A. Meyer. Geneva: IPCC. IPCC. 2018. “Summary for Policymakers.” In Global Warming of 1.5°C. An IPCC Special Report on the Impacts of Global Warming of 1.5°C above Pre-Industrial Levels and Related Global Greenhouse Gas Emission Pathways, in the Context of Strengthening the Global Response to the Threat of Climate Change, Sustainable Development, and Efforts to Eradicate Poverty, edited by V. Masson-Delmotte, H.-O. Pörtner, P. Zhai, D. Roberts, J. Skea, P. R. Shukla, et al, 3–24. Geneva: IPCC. Klein, Richard J. T., and Richard S. J. Tol. 1997. “Adaptation to Climate Change: Options and Technologies: An Overview Paper.” Technical Paper No. FCCC/TP/1997/3, United Nations Framework on Climate Change Secretariat, Bonn. Miao, Qing. 2017. “Technological Innovation, Social Learning and Natural Hazard Mitigation: Evidence on Earthquake Fatalities.” Environment and Development Economics 22 (3): 249– 73. https://doi.org/10.1017/S1355770X1700002X. Invention and Global Diffusion of Technologies for Climate Change Adaptation: A Patent Analysis / 55 Miao, Qing, and David Popp. 2014. “Necessity as the Mother of Invention: Innovative Responses to Natural Disasters.” Journal of Environmental Economics and Management 68 (2): 289–95. https://doi.org/10.1016/j.jeem.2014.06.003. Morgan, Robert P., Carlos Kruytbosch, and Nirmala Kannankutty. 2001. “Patenting and Invention Activity of U.S. Scientists and Engineers in the Academic Sector: Comparisons with Industry.” Journal of Technology Transfer 26 (1–2): 173–83. https://doi.org/https://doi. org/10.1023/A:1007856800497. Popp, David. 2019. “Environmental Policy and Innovation: A Decade of Research.” Working Paper No. 25631, National Bureau of Economic Research, Cambridge, MA. Prud’homme, Dan, and Taolue Zhang. 2017. “Evaluation of China’s Intellectual Property Regime for Innovation: Summary Report.” Background paper for Innovative China: New Drivers of Growth. Washington, DC: World Bank. https://doi.org/10.13140/RG.2.2.25427.48167/7. Svensson, Roger. 2012. “Commercialization, Renewal, and Quality of Patents.” Economics of Innovation and New Technology 21 (2): 175–201. https://doi.org/10.1080/10438599.201 1.561996. Tol, Richard S. J. 2018. “The Economic Impacts of Climate Change.” Review of Environmental Economics and Policy 12 (1): 4–25. UN DESA (United Nations Department of Economic and Social Affairs). 2018. World Economic and Social Survey 2018: Frontier Technologies for Sustainable Development. New York: United Nations. UNFCCC (United Nations Framework Convention on Climate Change). 2005. “Report on the Seminar on the Development and Transfer of Technologies for Adaptation to Climate Change.” Note by the Secretariat, FCCC/SBSTA/2005/8, UNFCCC, Bonn. UNFCCC. 2006. Technologies for Adaptation to Climate Change. Bonn: UNFCCC. UNFCCC. 2018. Summary of Country Priorities: Technology Needs Assessments 2015–2018. Bonn: UNFCCC; Copenhagen: United Nations Environment Programme–DTU Partnership.