FINANCE FINANCE EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT A climate transition risk assessment for the banking sector of the Philippines By Owen Nie, Nepomuk Dunz and Hector Pollitt © 2024 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 with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, and other information shown on 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. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. 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 attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Photo Credits Cover, pages 1 & 40: Alexpunker / iStock Page 2: Danilo Pinzon / World Bank Page 5: Kevin Izorce / iStock Page 11: Michael Edwards / iStock Page 22: Valeriy Volkonskiy / iStock Page 25: maska82 / iStock Page 32–33: tobiasjo / iStock Page 38: Dominic Chavez / World Bank >>> Acknowledgments Produced by a World Bank (WB technical team led by Owen Nie, Nepomuk Dunz, and Hector Pollitt with inputs from Henk Jan Reinders, Aileen Laurent Amor Bautista, Hasan Dudu, Dieter Wang, Nimarjit Singh, and Biying Zhu. The team is grateful for excellent collaboration with a Bangko Sentral ng Pilipinas (BSP) technical team consisting of Rhodora Brazil-De Vera, Richie Legaspi, Cherry Wyle Layaoen, Rafael Augusto Cachuela, and Dennis Bautista. The team acknowledges overall guidance from Uzma Khalil and Radu Tatucu (WB). Insightful peer review and comments from Kevin Carey, Robert Utz, Rekha Reddy and Ezio Caruso are gratefully acknowledged. We appreciate funding support provided by the International Financial Corporation’s 30by30 zero program under Paul Xavier Espinosa’s supervision. 2 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES >>> Contents Preamble 4 1. Introduction 6 2. Methodological approach 9 3. Climate transition risk scenarios for the Philippines 11 A. Overview of the MANAGE CGE model 14 B. Model results for the overall economy 16 C. Sectoral results 17 4. Macroeconomic and sectoral economic impact of climate transition risk 14 5. Linking economic impact with financial sector risk 19 A. Data 19 B. Empirical specification 20 C. Results and interpretation 20 6. Banking sector stress testing 26 A. Reverse stress testing results 26 Basic banking stress testing model to gauge impact on other banking sector variables B.  27 Plausibility check of climate stress testing results C.  30 7. Conclusion and Policy Implications 31 References 34 Appendix 36 Appendix I – List of MANAGE model sectors 36 Appendix II – Additional figures 37 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 3 >>> Preamble This report describes a transition risk assessment for the Philippines banking sector conducted by a World Bank technical team. As a follow-up work to the World Bank Philippines Country Climate and Development Report (CCDR) and supported under the Philippines Programmatic Advisory Services and Analytics (PASA), the analysis incorporates and builds on previous and ongoing work, including the Technical Note (TN) on Climate Change and Environmental Risks and Opportunities that was prepared as part of the 2019- 2020 Philippines Financial Sector Assessment Program (FSAP), a comprehensive and in-depth analysis of a country’s financial sector carried out jointly between the World Bank and the International Monetary Fund (IMF) every couple of years. The next steps identified in the FSAP TN included performing an in-depth climate and environmental risk assessment, improving information collection, monitoring of relevant risks, and developing capacity to stress test the financial prudential impacts of climate change. The second operation of the Philippines financial sector development policy financing (DPF), WB financing support through a program of policy and institutional actions, supports numerous policy actions in greening the financial sector, including regulatory guidelines in stress testing. This work seeks to implement some of the recommendations of the FSAP and informs policy actions supported under the DPF. The BSP has undertaken various supervisory initiatives in the climate and sustainability- related risks and opportunities space since the 2019-2020 FSAP mission and has implemented various recommendations. On the risks side, the BSP has been prioritizing the issuance of granular guidelines on climate and environmental risk management, stress testing, reporting, and disclosure, in alignment with Financial Stability Board’s Task Force on Climate-Related Financial Disclosures (FSB TCFD) recommendations. The BSP issued the Sustainable Finance Framework (Circular No. 1085, series of 2020) which sets out the expectations for banks to integrate sustainability principles in their corporate governance, risk management systems, strategies, and daily operations. Building on this, the BSP released the Environmental and Social Risk Management (ESRM) Framework in October 2021 (Circular No. 1128, series of 2021), which spells out the actions required by banks for the integration of environmental and social (E&S) risk factors in lending activities and business continuity and calls for the development and implementation of an ESRM framework in alignment with international principles, standards, and practices. Moreover, the BSP issued supplemental guidance on the implementation of the ESRM System and climate stress testing in September 2022. On the opportunities side, the BSP issued a regulation (Circular No. 1149, series of 2023) requiring banks to integrate sustainability standards into their capital market activities to promote green 4 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES investments. Likewise, to address the lack of clear standards for financial institutions in conducting assessment for climate and definitions of green assets, the Financial Sector Forum and environmental-related risks, this paper, together with (FSF)1 is developing sustainable finance taxonomy guidelines the earlier physical risk stress testing conducted during the that can help to better define green or sustainable activities. last FSAP, are intended to serve as important references The BSP has likewise issued regulations to incentivize banks for both the regulatory community and financial institutions to finance green or sustainable projects, including transition interested in conducting similar exercises. Moreover, this activities. report might also be of interest to others (e.g., academia and other observers) interested in the implications of climate- This report is intended to provide a source of reference related risks for the economies and financial systems of for financial regulators and financial institutions in the the Philippines and the Southeast Asia region. This report Philippines that are interested in conducting climate focuses solely on the impact of climate transition risk, while transition risk assessments. Following the issuance of it should be noted that both physical risk and transition Memorandum No. M-2022-042 on 29 September 2022, in risk could potentially interact and affect financial stability.2 which BSP sets out supervisory expectations and guidelines 1 The Financial Sector Forum is a voluntary interagency body comprised of the BSP, the Securities and Exchange Commission (SEC), the Insurance Commission (IC), and the Philippine Deposit Insurance Corporation (PDIC). The FSF is chaired by the BSP Governor. 2 For instance, a disorderly transition could negatively affect financial stability and hence firms’ capacity to invest in climate adaptation that could increase their resilience to climate physical risks. See The double materiality of climate physical and transition risks in the euro area. European Central Bank Working Paper No. 2665/May 2022. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 5 1 > > > F I G U R E 1 - GHG emissions by sector up to 2019 250.0 200.0 150.0 100.0 >>> Introduction The Philippines’ notable economic growth has been driven by industries that are heavily reliant on fossil fuels and are therefore substantial greenhouse gas (GHG) emitters, despite comparatively low energy and emissions per capita by global standards.3 As of 2021, the Philippines is the world’s 38th largest emitter of GHG, representing 0.48 percent of the global total, with per capita emissions at 1.27 tons of CO2 equivalent, well below the global average (6.9 tCO2e).4 However, emission levels are currently growing strongly, driven by electricity and heat (30.8 percent), agriculture (25.5 percent), and the transport sector (16.0 percent). The upward trajectory of emissions and the overall share of fossil fuels in the primary energy supply are driven by gross domestic product (GDP) growth, population growth, and increased energy consumption. This trend would need to be reversed to achieve the objectives of the Paris agreement. The Climate Transparency Report 2020 estimates that the Philippines would need to reduce its emissions to below 132 metric tons of carbon dioxide equivalent (MtCO2e) by 2030 and to below 198 MtCO2e by 2050, to be within a 1.5°C ‘fair share’ pathway.”5 Others Electricity and heat Transport Manufacturing and construction Industry 50.0 Waste Agriculture 0.0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Source: Ourworldindata.org 3 Manufacturing and wholesale and retail trade contributed sizably to output growth for the past two decades. The average contribution to GDP growth of manufacturing and wholesale and retail trade for the period 2001 to 2022 (excluding the pandemic year 2020) was at 1.5 percentage points (pps) and 1.7 pps, respectively. These are followed by finance and insurance activities (1.0 pps); professional and business services (0.8 pps); construction (0.7 pps); and agricultural, forestry, and fishing (0.7 pps). Philippine Statistics Authority (n.d.). National Accounts Data Series. Retrieved July 5, 2023, from https://psa.gov.ph/national-accounts/base-2018/data-series 4 See Our World in Data Philippines GHG emissions (https://ourworldindata.org/co2/country/philippines). 5 https://www.climate-transparency.org/wp-content/uploads/2021/01/Philippines-CT-2020.pdf. 6 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES Although the Philippines has started to build renewable Calabarzon, and Northern Mindanao) all have relatively young energy capacity, its electricity generation is currently fleets of coal power plants.7 Retrofitting or replacing these still dominated by coal-fired power plants. As of 2020, plants would be necessary to achieve the emissions targets coal accounted for 57 percent of power generation in the set by the Philippines. The high financing cost of moving Philippines. Renewables, mainly hydro, contribute only 21 away from coal presents a major challenge, but there is also percent of total generation (Figure 2). The Philippines currently a substantial risk of coal plants becoming stranded fossil fuel operates a total coal plant capacity of 12GW, with an additional assets if the transition is not managed effectively. Nonetheless, 2.3GW either announced, permitted, or under construction.6 it is important to note that the Philippine Department of Energy The median age of the coal plants is only eight years, which (DOE) issued a moratorium on greenfield coal-fired power is lower than the global median of 17 years. The regions with projects in 2020. the highest coal capacity in the Philippines (Central Luzon, > > > F I G U R E 2 - Electricity supply by sector up to 2020 120,000 Waste 100,000 Wind Solar PV 80,000 Natural gas GWh 60,000 Geothermal Hydro 40,000 Biofuels 20,000 Oil 0 Coal 1990 1995 2000 2005 2010 2015 2020 Year Source: IEA, WB staff calculation The domestic decarbonization and the global shift toward Unanticipated transition costs could lead to financial risks a green economy present potential financial risks in for households, businesses, and investors, potentially the Philippines, particularly if they are not adequately affecting financial stability. These risks are linked to the anticipated by the financial sector. For example, climate shift toward a low-carbon economy and can manifest as credit transition risks could arise from abrupt shifts in policy, or market risks for banks and underwriting risks for insurers, emerging technologies, or changes in consumer preferences. as depicted in Figure 3. Philippine banks are exposed to The design, timing, and targets of domestic climate policies transition risks through their holdings in polluting and GHG- will have substantial implications for GHG-intensive sectors intensive industries; for example, the recent FSAP estimates of the economy. International climate policy ambitions may that coal-fired power generation accounts for 8 percent of also impact these sectors because trade and international the banking sector’s total loan portfolio. An abrupt transition financial flows are becoming increasingly sensitive to the toward a greener and carbon-neutral economy could lead to embedded GHG emissions associated with products and rapid revaluations of underlying financial assets, and hence investments. Delayed action on climate change could expose to market and credit losses. In the short run these exposures the Philippines to transition risks induced by foreign countries, pose increasing operational and reputational risks, while in the firms, and investors who are pursuing aggressive supply chain long run these assets run the risk of becoming stranded. decarbonization. 6 Global Coal Plant Tracker, Global Energy Monitor, February 2023 release. 7 Global Coal Plant Tracker, Global Energy Monitor, February 2023 release. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 7 2 At the same time, financial institutions are vital in establish new production facilities with lower emissions.8 By facilitating the Philippines’ low-carbon transition and they doing so, they could reduce reliance on public sector support need guidance from supervisors to understand transition- and help meet the Philippines’ contribution to achieving related financial risks. There are regulatory incentives and global temperature targets. Such measures would also requirements for financial institutions to better understand and enable financial institutions to enhance their readiness for the assess climate-related transition risks and to offer favorable transition while minimizing associated risks in their lending financing conditions that incentivize companies to reduce their and investment portfolios. carbon footprints, invest in research and development, and > > > F I G U R E 3 - Climate-related risks for the financial sector Physical Risks Transition Risks (Extreme weather events and (Policy, technology, gradual changes in climate) consumer prepences) Increase Lower value Business Asset Reconstruction/ in energy Economy Migration of stranded disruption destruction replacement prices with assets dislocations Lower property Lower Lower corporate Lower growth and Nagative and corporate household profits, more productivity affecting feedback asset value wealth litigation financial conditions from tighter financial conditions Market losses Credit losses Operational risk Financial (equities, bonds, (residential and Underwriting losses (including liability system commodities) corporate loans) risk) Source: Grippa et al. 2019 The rest of this paper is organized as follows. We begin by macro-sectoral level and financial implications, respectively, discussing the methodological approach (Section 2), and and Section 6 assesses specific banking sector risks. We then introduce the climate-transition scenarios that were conclude with policy implications for the central bank and assessed (Section 3). Sections 4 and 5 present results at the financial regulators, as well as for financial institutions. 8 The BSP regulations include Circular No. 1085 dated 29 April 2020, Circular No. 1128 dated 26 October 2021, Circular No. 1149 dated 23 August 2022, Memorandum No. M-2022-42 dated 29 September 2022, and Circular No. 1159 dated 4 November 2022. 8 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 2 >>> Methodological approach The domain of climate-related stress-testing is still in its early stages and is still developing, given the multiple uncertainties stemming from climate change and the transition toward a low-carbon economy, with regard to their potential impact on economies and financial systems. Several central banks and international financial institutions have undertaken preliminary climate risk assessments to evaluate the financial system’s exposure to climate transition.9 Simultaneously, academic research has started delving into this subject matter.10 Assessing climate-related financial risks entails the use of diverse methodological approaches, ranging from simple exposure assessments to intricate stress tests.11 These approaches entail the integration of various models and data sources to pinpoint potential risks and vulnerabilities within the economy and financial sector. The selection of the scenario time horizon also varies. Some assessments adopt longer time frames, such as up to 2050 (Network for Greening the Financial System (NGFS), 2021), to offer a comprehensive understanding of the climate-economy interplay, encompassing the transition costs and long-term benefits of mitigating physical risks. Conversely, shorter time horizons (NGFS, 2023) align with the traditional stress-testing timelines of central banks and exercises focused on financial stability. The shorter time horizon facilitates the establishment of a more realistic baseline, the inclusion of immediate adverse shocks, and a more precise application of constant balance sheet or loan portfolio assumptions. Despite significant strides in the development of climate risk assessments, considerable challenges and deficiencies persist, especially for emerging market and developing economies (EMDEs). These challenges primarily revolve around: (i) the scarcity, specificity, and reliability of data for risk assessment, and the capability to effectively utilize this data; (ii) the breadth of the risks taken into consideration; (iii) the methodologies and assumptions employed for risk assessment (e.g., Monasterolo, 2020); and (iv) the capacity requirements for conducting analyses. Due to the substantial uncertainties inherent in the methodology, current climate risk assessments diverge from conventional stress-testing and instead serve as an exploratory and learning exercise, intending to raise awareness about climate transition risk among supervisors, central banks, and commercial banks. 9 See for instance ECB 2021, 2023 and Reinders et al. 2021. 10 See Acharya et al. 2023 for an overview. 11 See Reinders et al. 2023 and Bingler et al. 2022 for an overview of current approaches of climate risk assessment. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 9 3 Nevertheless, current generations of climate stress tests Our methodological framework follows Reinders et play a pivotal role in establishing a prudent approach al. (2021) and integrates various key components, and enhancing capacity for financial institutions, encompassing the formulation of transition risk scenarios, enabling them to identify and evaluate potential risks and macroeconomic modeling, financial risk modeling, and opportunities linked with climate change. Despite existing stress testing within the banking sector (Figure 4). We limitations, these tests can aid supervisors in quantifying initiate the process by defining the transition scenarios of climate risks to financial stability and incentivize institutions interest and procuring carbon price trajectories over time for to devise strategies for their management. Additionally, the each of these designated scenarios. The devised transition implementation of climate stress tests fosters capacity building, scenario framework directly informs our macroeconomic transparency, and the disclosure of climate-related risks and modeling, allowing us to scrutinize both the overall and sector- opportunities. Supervisors can leverage their own climate-risk specific trajectories of production across diverse scenarios. analysis to inform both micro- and macro-prudential activities, Subsequently, we employ econometric tools for analyzing the encompassing the continuous monitoring of macro-prudential potential repercussions of the shifts in sectoral output on the indicators and the establishment of climate risk management quality of bank assets. Lastly, we employ a banking stress and stress testing guidelines for financial institutions. test model to comprehend the implications for critical banking sector indicators such as capital adequacy. As with all current approaches, our approach also contains certain shortcomings. > > > F I G U R E 4 - Transition risk stress testing approach Scenario design Real economy modeling Financial risk modeling Banking sector Impact Focus on transition risk: Obtain a model that can Develop a financial Set-up a simplified obtain carbon price paths integrate carbon price risk model that can link top-down stress test over time in different paths and preferably has macroeconomic outcomes model based on scenarios, based on the a sectoral disaggregation to financial sector bank’s balance sheet updated NGFS scenarios variables and P&L data, which Short term option: or NDC. is disaggregated for Current approach: • World Bank MANAGE sectors (as used by BSP (CGE) model • Develop regression or more granular) and model to establish link provides capital adequacy Medium/long term between credit risk outcomes for individual option: variables and economic banks and at aggregated • BSP in-house model variables levels Source: WB staff 10 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 3 >>> Climate transition risk scenarios for the Philippines We define seven different climate policy scenarios for the Philippines up to 2040, which differ in terms of policy timing and degree of climate ambition. We closely follow the scenario design of the NGFS database for the first four scenarios.12 We design three additional scenarios based on an international carbon price floor, as outlined by the IMF, with a carbon tax of around US$75 in 2030. The scenarios reflect the emission pathway that the Philippines would need to follow to achieve global climate targets (e.g., Net Zero by 2050). Other transition scenarios are available, and our choice is motivated by cross-country comparability. The climate physical damages based on different emission pathways (covered under the recent FSAP) as well as foreign-induced transition impacts in these scenarios and the subsequent analysis were not considered in this exercise. 12 These NGFS Scenarios specifically employ the trajectories established by the REMIND-MagPie model under the NGFS phase II and developed by the Potsdam Institute for Climate Impact Research (PIK) (Hilaire and Bertram, 2020). The REMIND-MagPie model employs an iterative process involving a macroeconomic Ramsey model and a cost-minimizing energy technology choice model to evaluate economic and energy technology trajectories. The macroeconomic model component determines energy demand, while the energy model component calculates energy supply and associated input costs based on a target emission level and corresponding carbon price. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 11 > > > T A B L E D 1 - Description of transition scenarios SCENARIO SCENARIO DESCRIPTION Ambitious scenario that aims to limit global warming to 1.5°C through stringent Net Zero 2050 (NGFS) climate policies and innovation, reaching net zero CO2 emissions around 2050. Targets net-zero emissions by 2050, but incurs higher costs due to divergent Divergent Net Zero (NGFS) policies across sectors and a faster phase-out of fossil fuels. Gradually increases the stringency of climate policies, providing a 67% chance of Below 2 Degrees (NGFS) limiting global warming to below 2°C. Physical risks associated with this scenario are relatively low. Includes all pledged policies, even if they have not been implemented yet. While Nationally Determined Contributions (NDCs) (NGFS) emissions decline, this scenario still leads to a 2.6°C warming, resulting in more severe physical risks. The IMF International Carbon Pricing Floor (baseline) The IMF International Carbon Pricing Floor with increased Introduces measures equivalent to a global carbon tax of around $75 per ton of public climate investment (Variant 1) CO2 by 2030, with further increases beyond 2030. This policy is negotiated among a small number of large emitter countries. The IMF International Carbon Pricing Floor with labor market frictions and slower growth of renewables (Variant 2) Note: The NGFS scenarios follow the carbon price trajectory given by REMIND-MAgPIE 2.1-4.2. The carbon price in the NGFS scenarios varies based on are provided on a five-year basis and are interpolated to align the level of ambition and timing of climate policy. Stringent with the yearly time intervals of the CGE model used later in policies result in higher carbon prices (Figure 5), while the analysis. In addition to the baseline IMF scenario with a assumptions regarding the availability and cost-effectiveness carbon tax of around US$75 in 2030, two variants are also of green technologies influence the carbon price (lower prices considered. The first variant recycles carbon tax revenues for cheaper green technologies). Furthermore, disorderly for additional public climate investment. The second variant policy introductions, such as under the Divergent Net Zero resembles a less smooth transition, with labor market frictions trajectory, lead to quicker and steeper required increases in and slower growth of renewables making the carbon price carbon prices to achieve climate targets. Transition trajectories less efficient. > > > F I G U R E 5 - Carbon price in 2010 USD per ton of CO2 450 400 Below 2°C 350 Nationally Determined 300 Contributions (NDCs) 250 200 Divergent Net Zero 150 Net Zero 2050 100 IMF USD 75 / t CO2 50 carbon tax 0 2020 2025 2030 2035 2040 Source: WB staff calculation 12 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES The carbon price trajectories within the various climate carbon price scenarios while incorporating certain limits on the policy scenarios play a crucial role in determining the maximum speed of technological change. To ensure a smoother required trajectory for the power sector, specifically in uptake we consider the averages over a three-year period. reducing the share of non-renewable sources such as The result is a projected change in fuel mix within the power coal and increasing the share of renewable sources like sector under the different carbon price trajectories within the solar photovoltaic (solar PV). We use estimates by the climate policy scenarios (Figures 6a and 6b respectively show recent Philippine CCDR using the PLEXOS power sector the fuel mix in the power sector under the Net Zero 2050 and model13 for the sensitivity of power-generating sectors to the NDC scenarios). Those fuel mixes are used as inputs for one-unit changes in carbon price. Investment costs related to the MANAGE model, as the model has a production structure the various technologies/power sector trajectories are inputs that allows for an endogenous energy intensity of production to this simulation. However, the scenarios presented in this that fluctuates with carbon pricing policies (see Section 4 for study are more ambitious than those in the CCDR, requiring details). The model then determines the macroeconomic and an extrapolation of sensitivities with the more ambitious sectoral impacts from those climate policies. > > > F I G U R E 6 A - Fuel mix of the power sector under a Net Zero 2050 scenario 70% COAL 60% GAS FUEL_OIL 50% NUCLEAR 40% GEO 30% WATER 20% WIND WIND_OFF 10% PV 0% BIOMASS 28 26 29 33 35 31 34 23 25 32 21 24 22 37 27 30 40 38 36 39 STORAGE 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 F I G U R E 6 B - Fuel mix of the power sector under an NDC scenario 70% COAL 60% GAS FUEL_OIL 50% NUCLEAR 40% GEO 30% WATER 20% WIND WIND_OFF 10% PV 0% BIOMASS 28 26 29 33 35 31 34 23 25 32 21 24 22 37 27 30 40 38 36 39 STORAGE 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Note: Figure 6 presents projections under two different scenarios by WB staff, using available data and power sector models. WIND_ OFF: offshore wind. STORAGE: carbon capture and storage. PV: Photovoltaic. BIOMASS: renewable energy from plants and animals. Source: WB staff calculation 13 PLEXOS is a power system planning software platform and was used to understand the implications of different levels of emissions reductions on the capacity and generation mix given assumptions about demand growth and available technologies. The results should not be interpreted as forecasts, but as projections of the scale and speed of necessary interventions. See https://www.energyexemplar.com/plexos for details. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 13 4 >>> Macroeconomic and sectoral economic impact of climate transition risk A. OV ERV IE W O F T H E M ANAGE C GE M OD E L Computable General Equilibrium (CGE) models, a class of macroeconomic models which the MANAGE model belongs to, are widely used to study the long-term effects of policy and external shocks on the economy. CGE models are economy-wide models that are designed to capture the interactions between economic agents (e.g., households, firms, governments, and the external sector) in a simplified way and rely on detailed economic data reconciled in a Social Accounting Matrix (SAM). SAMs are consistent with system of National Accounts and give a snapshot of the flows between economic agents in a given (base) year by combining: the Input-Output table, government finances, balance of payments, trade, and micro data such as household, labor, agriculture, industrial, and surveys.14 The modeling framework allows users to analyze policy shocks (e.g., tax changes, new trade policies) and external shocks (climate, natural disasters) and assess how they affect different economic activities, households, and production factors. The CGE model is based on assumptions about agent optimization and flexible markets. Households and firms engage in optimizing behaviors in the model: households maximize their utility from goods and services given their prices and the budget constraint, and firms maximize profits given the prices of production factors and subject to the current technological possibilities. Prices adjust to clear all product and factor markets to achieve equilibrium in all markets, i.e., general equilibrium. All domestic and external transactions and transfers also must be balanced: total investment equals total savings; trade balances equal net capital inflows (balance of payments), and government finances are balanced (budget is equal to net savings). 14 A SAM is usually constructed by building a macro-SAM using the National Accounts and various other sources, and then disaggregating macro-SAM to the desired degree of detail. A variety of data sources are needed to construct a SAM, including 1) National Accounts (aggregated and by institutional sectors), 2) full Input-output tables, 3) household surveys including information on consumption and expenditure patterns of households and the source of household income, 4) labor force surveys, i.e. statistics on the labor market and its composition, especially regarding skilling, employment, wages, sectors, etc., 5) specific surveys on certain sectors and on specific socioeconomic aspects, such as population census, agricultural census, surveys of manufacturing and services, 6) accounts of the public administrations, i.e. tax data and statistics, and 7) debt statistics, and trade data, i.e. imports and exports. 14 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES > > > F I G U R E 7 - An example of a stylized representation of a CGE model (flow of payment) domestic wages and rents private investment financing Factor Markers Households dir taxes lending interest trnsfr+ factor private trnsfr- demand consumption indir taxes interest gov cons and inv Government intermediate input demand lending interest trnsfr+ Private Capital Account Activities imports FDI lending Rest of World Domestic domestic Commodity demand Markets exports foreign wages and rents private investment Source: WB staff The strength of the CGE model is its consistent uncertainty that could lead to a misallocation of resources, representation of the whole economy as well as its high e.g., through ‘stranded assets,’ (Semieniuk et al., 202215) but sectoral granularity. The model captures both direct and this is only accounted for indirectly through change in value- indirect effects, provides consistency over the use of resources added, and there is currently no feedback from stranded and the constraints of the economy, and includes detailed assets to the model. Optimization has been shown to be information on economic activities and production factors. a poor approximation of real-world behavior (Trutnevyte, Hence, the model allows users to estimate quantitatively the 201616) and the representation of the financial sector lacks ex-ante impact of policy shocks on a range of macroeconomic empirical basis (Pollitt and Mercure, 201817; Werner, 201418). indicators. At the same time, CGE models usually have a The model assumes a fixed set of technologies, which misses high sectoral granularity, especially useful to identify sectoral important transition dynamics. By assuming immediate climate policy impacts to provide guidance on the winners and adjustment to equilibrium, the models cannot address the losers of policy reforms. short-term impacts that typically drive financial behavior. Thus, while the CGE model remains the standard tool for Nevertheless, it is important to be aware of the application, alternative approaches under development (e.g., limitations of the CGE modelling approach, some of Mercure et al., 201819; Burns et al., 201920) could substantially which are particularly relevant to climate stress testing. improve the analysis in future. A key feature of the low-carbon transition is an increase in 15 Semieniuk, G., PB Holden, et al. 2022 “Stranded fossil-fuel assets translate to major losses for investors in advanced economies.” Nature Climate Change, Volume 12, pp 532–538. 16 Trutnevyte, E. 2016. “Does cost optimization approximate the real-world energy transition?” Energy, Volume 106, 1 July 2016, pp 182-193. 17 Pollitt, H. and J-F Mercure. 2018. “The role of money and the financial sector in energy-economy models used for assessing climate and energy policy.” Climate Policy, Volume 18, Issue 2, pp 184-197. 18 Werner, RA. 2014 “Can banks individually create money out of nothing? — The theories and the empirical evidence.” International Review of Financial Analysis, Volume 36, pp 1-19. 19 Mercure, Jean-Francois, Hector Pollitt, et al. 2018b. “Environmental impact assessment for climate change policy with the simulation-based integrated assessment model E3ME-FTT-GENIE.” Energy Strategy Reviews, Volume 20, April 2018, pp 195–208. 20 Burns, Andrew, Benoit Campagne, et al. 2019. “The World Bank Macro-Fiscal Model Technical Description.” https://documents1.worldbank.org/curated/ en/294311565103938951/pdf/The-World-Bank-Macro-Fiscal-Model-Technical-Description.pdf. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 15 For assessing macroeconomic impacts of climate allows for an endogenous energy intensity of production, transition scenarios, we use the World Bank’s Mitigation, which fluctuates with carbon pricing policies. We do not Adaptation and New Technologies Applied General go into the specification of the model, referring interested Equilibrium (MANAGE) CGE model, which is designed readers to the two references cited above. Also, we realize specifically to look at climate issues. As outlined in the that for many government agencies in EMDEs, capacity- MANAGE model documentation (2017) and Hallegatte constraint related to CGE modeling may be an impediment et al. (2023), most features of the MANAGE model match to the proper assessment of climate-related risks. the standard CGE approach based on neoclassical theory. For our purposes, we calibrate the MANAGE model to Firms are profit maximizers under technologies with constant Philippines data. The calibration was completed for the returns to scale. All markets in the model are perfectly Philippines CCDR published in 2022. The model is calibrated competitive, implying that prices are equal to marginal costs to replicate the 2018 SAM for the Philippines, which consists of in equilibrium. A nested production function with constant the 2018 macro-aggregates, the 2018 Input-Output (I-O) table, elasticity of substitution (CES) technology describes how and household surveys. It includes 90 sectors and products the economic system transforms inputs into outputs and (aggregated to 53 for the CCDR), four factors of production, added values. The supplies of labor and land are determined six tax types, one trade partner, and ten types of households by functions that are sensitive to average real wages and defined by income decile. The SAM also distinguishes between land prices, respectively. In addition, the MANAGE model includes a detailed energy specification that allows for capital/ public and private investment demand and includes seven electricity activities that produce a homogeneous electricity labor/energy substitution in production, intra-fuel energy product generated from coal, gas, nuclear, hydropower, wind, substitution across all demand agents, and a multi-output, solar and other sources (e.g., biomass). multi-input production structure. This production structure B . MO DEL RESULTS FOR T H E OV E RAL L EC O NO M Y To provide context, we present aggregate results transition leads to larger output losses: in the extreme case, that demonstrate impacts of the climate transition on for the Divergent Net Zero scenario, the most disorderly economic output in the Philippines; the size of such transition under the seven scenarios we considered, output impacts depends on the nature of the transition scenario losses relative to baseline are estimated to reach 7.9 percent considered. The MANAGE model produces annual simulation by 2040. In contrast, under the most orderly transition, i.e., IMF results for the years 2022-2040 for the path of aggregate International Carbon Pricing Floor with carbon tax revenue economic output, as well as for each sector.21 Figure 8 used for investment, the output loss is estimated to be only presents our results for the aggregate economy expressed as 0.5 percent by 2040. Output losses come mainly from higher percentage change relative to the baseline business as usual energy (in particular, electricity) prices and investment needs (BAU) scenario (cumulative emissions of 3,340 MtCO2e). that divert investment from the rest of the economy. These Under most scenarios, output losses are projected to gradually results highlight the importance of properly managing the increase over the years and reach anywhere between 3 to 6 economic adjustment costs during the transition to a greener, percent relative to the baseline in year 2040. More disorderly carbon-neutral economy. 21 There is no specific modeling adjustment related to COVID-19 related shocks. For this purpose, the model does not consider significant one-off shocks to the economy and focuses on the long-run steady state. 16 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES > > > F I G U R E 8 - Impact of climate transition on aggregate output under different scenarios 2 Below 2°C Percentage Difference from baseline scenario Nationally 0 determined Contributions (Business as usual) -2 Divergent Net Zero Net Zero 2050 -4 IMF USD 75 / t CO2 carbon tax -6 IMF USD 75 / t CO2 carbon tax w/ slower -8 renewable growth IMF USD 75 / t CO2 -10 carbon tax w/ 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 carbon tax revenue to investments Source: MANAGE model simulation C . SEC TO R AL RESULTS Aggregate results mask differences at the sectoral level For instance, public administration and defense, real estate as there are clear winners and losers from the low-carbon activities, and coconut production are sectors projected to transition. Certain sectors are heavily exposed to transition have similar output paths relative to the baseline. However, risks and are therefore expected to suffer large output losses, most sectors related to transportation, energy generation, and while others are crucial for clean energy transition and hence some sectors in agriculture (i.e., livestock) could experience are expected to benefit from such a process.22 Figure 9 shows significant positive or negative impacts in the low-carbon output paths for two sectors that are most strongly affected transition.24 by the transition to illustrate this point: electricity generation by wind, a clean energy sector, is expected to be 13 times Our economic modeling does not feature a financial larger relative to the baseline by 2040 under most scenarios; sector able to respond to the climate transition, and hence by contrast, electricity generation by coal is expected to shrink steps are needed to link the macroeconomic modeling by 99 percent by 2040. A core difference across scenarios results to impacts on the financial sector. Ideally, in a is the speed of the transition and respective output changes climate macroeconomic model like the MANAGE model, the for certain high- and low-carbon sectors, which can have financial sector would react endogenously to energy intensity implications for financial sector risk, as faster transitions and carbon emission pathways. For instance, if banks expect might come as a surprise for investors.23 However, not all the coal sector to shrink significantly, they will anticipate such sectors are strongly impacted by the low-carbon transition. a reduction by adjusting their lending to the sector. In turn, 22 Appendix I lists all 53 sectors included in the MANAGE model simulations. 23 A time-series path under different scenarios is available for all 53 sectors. 24 Appendix II ranks the ten sectors with the most output gains and losses across the entire simulation period, i.e., between 2022 and 2040 (instead of just 2040): sectors that expanded the most throughout the 19-year period include electricity generation from wind, solar, hydro, as well as more general sectors like accommodation and food service activities; sectors that contracted the most include electricity generation from coal, non-metallic mineral products, gas extraction, electricity generation from gas, water transport, oil and coal extraction. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 17 5 this would increase the cost of financing for the coal sector hence overall transition impacts. There are a few examples and could accelerate energy transition. Another effect could of macroeconomic models to date that endogenously be that banks would lend to, and institutional and retail incorporate macro-financial interactions,25 but research gaps investors would invest more in, the clean energy sectors such remain. Hence, we currently abstract from financial feedbacks as electricity generation by solar and wind, thereby reducing onto the macroeconomy, focusing on macroeconomic impacts their financing costs. If such financing decisions would be on the financial sector only in our methodological approach featured endogenously in a macroeconomic model, they could and leave an assessment of those feedback effects for further have implications for the output paths of different sectors and analyses. > > > F I G U R E 9 A - Impact on output for “Electricity from Wind” sector under different transition scenarios 1800 Percentage Difference from baseline scenario Below 2°C 1600 1400 Nationally determined Contributions 1200 (Business as usual) Divergent Net Zero 1000 800 Net Zero 2050 600 IMF USD 75 / t CO2 carbon tax 400 IMF USD 75 / t CO2 carbon tax 200 w/ slower renewable growth 0 IMF USD 75 / t CO2 carbon tax -200 w/ carbon tax revenue to 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 investments F I G U R E 9 B - Impact on output for “Electricity from Coal” sector under different transition scenarios 20 Percentage Difference from baseline scenario Below 2°C 0 Nationally determined Contributions -20 (Business as usual) Divergent Net Zero -40 Net Zero 2050 -60 IMF USD 75 / t CO2 carbon tax -80 IMF USD 75 / t CO2 carbon tax w/ slower renewable growth -100 IMF USD 75 / t CO2 carbon tax -120 w/ carbon tax revenue to 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 investments Source: MANAGE model simulation 25 See for instance Gourdel et al. 2023 for an application of a Stock-Flow Consistent macro-model of climate physical and transition risk in the Euro area. 18 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 5 >>> Linking economic impact with financial sector risk In this section we use statistical analysis based on historical data to establish an empirical relationship between the economic impact of climate transition and the impact on bank asset quality, both at the aggregate level and at the sectoral level. We hypothesize that economic growth correlates negatively with default risks and hence bank asset quality. Several studies in the academic and policy literature (i.e., IMF (2006)26, Nkusu (2011)27, Beck et al. (2013)28, Lee and Rosenkranz (2019)29, Ari et al. (2019)30, Alnabusi (2022)31) have examined similar research questions or have adopted a similar empirical approach. If firm-level data are available, a more granular modeling approach could link sectoral economic impacts to nonfinancial corporate balance sheet indicators, and then through corporate credit risk paths to bank exposures. A. DATA Our empirical analysis combines several data sources. Bank balance-sheet data were provided by the BSP for 60 banks across three categories (universal and commercial banks or UKBs, thrift banks or TBs and rural and cooperative banks or RCBs with 20 banks in each category) on nonperforming loans (NPLs), total volume of lending as well as other indicators (i.e., profitability and liquidity) at a quarterly frequency between 2010Q1 and 2021Q4 by sector (note that BSP sectoral classification differs from MANAGE model sector classification). This is complemented by sectoral output data and country-specific macroeconomic variables such as unemployment, inflation, and GDP growth. Exchange rate and global macro-financial variables such as the VIX, an index measure of financial market volatility, are also added. For the analysis, we compile a panel dataset at the bank-quarter level for 2010Q1 and 2021Q4. Due to data availability, the panel is unbalanced.32 26 International Monetary Fund. 2006. “Spain: Financial Sector Assessment Program - Technical Note: Stress Testing Methodology and Results.” IMF Staff Country Reports, 06(216), https://doi.org/10.5089/9781451812190.002 27 Nkusu, M. 2011. “Nonperforming Loans and Macro-financial Vulnerabilities in Advanced Economies.” IMF Working Paper WP/11/161. International Monetary Fund. 28 Beck, Roland, Petr Jakubik, Anamaria Piloiu. 2013. “Non-performing loans: what matters in addition to the economic cycle?” Working Paper Series 1515, European Central Bank. 29 Lee, Junkyu and Peter Rosenkranz. 2020. “Nonperforming Loans in Asia: Determinants and Macro financial Linkages.” International Finance Review, Volume 21: Emerging Market Finance: New Challenges and Opportunities, edited by Bang Nam Jeon and Ji Wu, 33–53. Bingley, England: Emerald Publishing Limited. https://doi.org/10.1108/S1569-3767202021. 30 Ari, A., Sophia Chen, and Lev Ratnovski. 2019. “The Dynamics of Non-Performing Loans during Banking Crises: A New Database.” IMF Working Paper WP/19/272). International Monetary Fund. https://www.imf.org/en/Publications/WP/ Issues/2019/12/06/The-Dynamics-of-Non-Performing-Loans-during-Banking-Crises-A-New-Database-48839. 31 Alnabulsi, K., Emira Kozarević, and Abdelaziz Hakimi. 2022. “Assessing the determinants of non-performing loans under financial crisis and health crisis: Evidence from the MENA banks.” Cogent Economics & Finance, 10(1), 2124665. https://doi. org/10.1080/23322039.2022.2124665. 32 Most panel regressions, including our empirical specification, can be used for unbalanced panels. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 19 B . E M P I R ICAL SPECIFICAT IO N We use a panel autoregressive distributed lag (PADL) model orders of NPL and sectoral output, are selected based on the with fixed effects to link the NPL ratio with output at the sectoral Akaike information criterion (AIC) and the Bayesian (Schwarz) level while controlling for other factors that could affect bank information criterion (BIC). asset quality. The empirical specification is: W represents control variables including 1) bank-level controls such as return on assets, return on equity, cost to income, earning assets yield, funding cost, interest spread, net interest margin, cash and due from banks to deposits, liquid assets to where the subscript b is bank, t is quarter, and s denote deposits, loans to deposits, and loan growth 2) macroeconomic sectors. y, the dependent variable, is NPL ratio (henceforth variables such as changes in unemployment, GDP growth, referred to as NPL) for each bank at each quarter and its lags inflation, level and slope of the government yield curve, the real are included as independent variables to control for serial effective exchange rate, and the VIX index. To avoid estimating correlation, and b is bank fixed effects. X are sectoral outputs too many coefficients for the controls using a relatively small and their lags, and βs are the coefficients of interest. Both dataset, we extract the first three principal components of the NPLs and sectoral output are specified as log differences control variables, which explain 90 percent of the variations in to remove time trend and have a percentage change these variables. interpretation for the coefficients estimated. P and D, the lag C . RES U LTS AN D IN T ERPRE TAT IO N Figure 10 below shows the coefficient of interest for each qualitative relationship between economic activity and NPLs sector for the lag order we chose to use. We also estimate is true for all but two sectors in the economy (real estate and the overall sensitivity across all sectors, i.e., for the aggregate education). An economic downturn of the same magnitude economy. triggers a stronger deterioration in bank asset quality in some sectors, i.e., financial sector and transportation and Overall, we find that changes in NPLs are negatively communication, health and social activities, manufacturing; associated with changes in economic conditions for and weaker reactions in some other sectors, i.e., construction, both the overall economy and for most specific sectors, accommodation, agriculture, etc. For real estate activities and with a few exceptions. Our hypothesis that an economic education, we find a positive relationship between economic downturn (expansion) is associated with higher (lower) levels activities and NPLs. Although we do not have sector-specific of nonperforming loans in the banking system is largely explanations, we think that it is not unreasonable that bank confirmed by two separate regressions: one using data from all asset quality for some sectors will behave in a countercyclical sectors to provide estimates for the overall economy while the way. Nevertheless, inferences about the sensitivity of the NPL other estimating an effect for each sector. Also, most sectoral ratio to GDP should be drawn with caution. These estimates estimates are statistically significant at the 95 percent level. provide inputs for linking economic impacts to financial sector We find that a 1 percent increase in GDP is associated with a risks, although they are on the lower end of empirical estimates 0.14 percent decrease in NPLs for the overall economy. The in the literature. 20 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES > > > F I G U R E 1 0 - Empirical relationship between economic activity and bank asset quality, by sector , ly s nt pp ra Su au n g io st in at on re ic iti d un an nd m ls Co m s te e Co r- iti Ho r ai Ai iv d p ct an e, d re es an ic A 1 n nd iti rv k io er or iv Se ,a at at lW t de m Ac od ,W 0.678 or a ia Fo e tr am nf oc nc 0.473 ng nd il ,I S ra a te yi ge et nd a su ,S r n ra 0.5 R ar a In io g as o nd Qu th rin at St nd n ,G al a tio od tu d la , ity He le rt an m ac uc sa ia po ic m an g nc uf tr le tr in s co ns an m ho an na ec in Ac Hu Co M W Tr M Fi El 0 -0.057 -0.101 -0.0825 -0.202 -0.151 -0.131 -0.224 -0.203 es g es n in io -0.304 iti iti sh at -0.5 tiv tiv uc Fi ac Ac Ed nd ce te ,a vi ta ry er Es st rS re al Fo he Re -1 Ot g, tin un ,H -1.218 re tu -1.275 -1.5 ul ric Ag Source: WB staff calculation These sensitivity estimates are then combined with the electricity generation would be 20 percent higher than that economic impact estimates from the MANAGE model to in the baseline by the mid-2020s under most scenarios. form our estimates of the impact of climate transition For sectors that are projected to expand in the low-carbon risks on NPL ratios under different scenarios.33 Figures 11 transition, e.g., renewable energy (RE), the impact on NPLs and 12 show the extrapolated time-series paths of NPLs, like will likely be muted because there is arithmetically a lower Figures 8 and 9 for output. As expected, system wide NPLs for bound to how much NPLs can decrease from baseline, i.e., the economy increase under different transition scenarios, but 100 percent.34 the increase is more pronounced for more disorderly transitions Several caveats should be noted for this approach. (such as the divergent net-zero scenario) and less so for First, just like the extrapolation for GDP, as we extrapolate the more orderly ones (such as the IMF carbon tax revenue further into the future, we are less certain about the recycling scenario). In the median scenario, NPLs would be estimates, especially beyond five years. Second, we use around 0.5 percent higher than in the baseline scenario by a point estimate from historical data to construct these 2040. Also, sectors that suffer higher output losses during the estimates, and the pattern from historical data may not climate transition will also suffer more pronounced increases hold for the projection period, which is especially relevant in NPL ratios over time. For example, in the coal extraction for an economy-wide low-carbon transition with no historical sector, NPLs would be around 6 to 10 percent higher than in precedence. Finally, we do not account for banks’ reactions the baseline scenario by 2040, a much larger impact than that to these risks, including rebalancing their lending portfolios found for the overall economy. NPLs for lending to coal-fired away from high-risk sectors. 33 Economic impact estimates multiplied by the sensitivity would yield impact on NPL ratio. 34 Although the RE sector is projected to grow in general, there are also notable cases around the world where firms in this new sector have failed due to commercially unsuccessful technologies and/or rapid price reductions and hence thin profit margins. In these cases, NPLs would increase in the RE sector. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 21 > > > F I G U R E 1 1 - Impact of climate transition on system wide NPLs under different scenarios 1.4 Percentage Difference from baseline scenario Below 2°C 1.2 Nationally determined 1 Contributions (Business as usual) 0.8 Divergent Net Zero 0.6 Net Zero 2050 0.4 IMF USD 75 / t CO2 carbon tax 0.2 IMF USD 75 / t CO2 carbon tax w/ slower renewable growth 0 IMF USD 75 / t CO2 carbon tax -0.2 w/ carbon tax revenue to 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 investments Source: WB staff calculation 22 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES > > > F I G U R E 1 2 A - Impact on NPLs for “Coal Extraction” sector under different transition scenarios 14.00 Percentage Difference from baseline scenario (Business as usual) 12.00 Below 2°C 10.00 Nationally determined Contributions 8.00 Divergent Net Zero 6.00 Net Zero 2050 IMF USD 75 / t CO2 carbon tax 4.00 IMF USD 75 / t CO2 carbon tax w/ slower renewable growth 2.00 IMF USD 75 / t CO2 carbon tax w/ carbon tax revenue to 0.00 investments -2.00 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 F I G U R E 1 2 B - Impact on NPLs for “Electricity from Coal” sector under different transition scenarios 25.00 Percentage Difference from baseline scenario (Business as usual) 20.00 Below 2°C Nationally determined 15.00 Contributions Divergent Net Zero 10.00 Net Zero 2050 IMF USD 75 / t CO2 carbon tax 5.00 IMF USD 75 / t CO2 carbon tax w/ slower renewable growth IMF USD 75 / t CO2 carbon tax 0.00 w/ carbon tax revenue to investments -5.00 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 Source: MANAGE model simulation A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 23 Our analysis shows that the low-carbon transition (3.7 percent) to translate the impact on NPLs into change in scenarios seem to have a limited effect on NPL ratios for percent. The impact of transition risk is very small under an orderly transition scenarios, with more pronounced effects orderly transition. Even under the Divergent Net Zero scenario, for disorderly transition scenarios. Table 2 presents the NPLs would increase by only 0.1 percent in 2040 relative to results for the overall economy and the financial system under the baseline. This suggests little financial stability concern for each scenario for the year 2040, using two different estimates the overall financial sector. for NPL-output elasticities (one estimated in this paper and However, the aggregate assessment masks large the other from the literature for a group of European countries, heterogeneity across sectors, with some carbon intense e.g. (Beck et al. 2013). We find it necessary to include results sectors such as coal extraction (Table 3) facing large using another estimate because the new estimates are from increases in NPL ratios based on our analysis. Table 3 a short and unbalanced panel dataset and hence suffer from shows the impact of the low-carbon transition on the coal data limitations. Although the economic impact of climate extraction sector for the year 2040. Although sectoral NPL- transition is substantial under some disorderly scenarios, the output elasticities and the NPL ratio are similar, large output overall impact on system-wide NPL ratios is low because of a contractions in the sector lead to NPL increases for 2040 at moderate NPL-output elasticity and the generally low level of around 0.5 to 1 percent for most scenarios. This impact is NPL (as of early 2023). The second column of the table shows sizable if it accumulates over a period of several years without the impact on aggregate GDP for four different scenarios. Column 3 uses our empirical estimate of the NPL-output action on the part of the banks holding large portfolio in the sector. Furthermore, this analysis indicates the relevance of elasticity to translate the GDP impact into NPL impact, while a sectoral approach, as can also be seen by the subsequent column 4 does the same for a different elasticity estimate analysis. from Beck et al. (2013). Column 5 uses end-of-2022 NPL ratio > > > T A B L E 2 - Transition risks for the overall economy and financial sector for different transition scenarios, year 2040 Impact on NPL ratio Impact on NPL ratio Impact on NPL ratio (percentage deviation (percentage deviation (change in percent, with Impact on GDP from baseline in 2040, from baseline in 2040, NPL-output elasticity of (percentage deviation with NPL-output with NPL-output 0.5, using end 2022 NPL Transition scenario from baseline in 2040) elasticity of 0.14) elasticity of 0.5) level) IMF International carbon pricing floor – (carbon -0.54 0.08 0.27 0.01 tax revenue recycled as investment) IMF International carbon pricing floor – -3.08 0.43 1.54 0.05 (Baseline) IMF International carbon pricing floor – (Labor market frictions -3.27 0.46 1.63 0.06 and slower growth of renewables) Nationally Determined -3.46 0.48 1.73 0.06 Contribution (NDC) Below Two degrees -3.53 0.49 1.76 0.06 Net Zero by 2050 -5.27 0.74 2.63 0.08 Divergent Net Zero -7.86 1.1 3.93 0.12 24 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES > > > T A B L E 3 - Transition risks for the coal extraction sector for different transition scenarios, year 2040 Impact on NPL ratio Impact on NPL ratio Impact on NPL ratio (percentage deviation (percentage deviation (change in percent, with Impact on Sectoral GDP from baseline in 2040, from baseline in 2040, NPL-output elasticity of (percentage deviation with sectoral NPL-output with sectoral NPL-output 0.5, using end 2022 NPL Transition scenario from baseline in 2040) elasticity of 0.2) elasticity of 0.5) level) IMF International carbon pricing floor – (carbon -21.06 4.21 10.53 0.33 tax revenue recycled as investment) IMF International carbon pricing floor – (Labor market frictions -27.64 5.52 13.82 0.43 and slower growth of renewables) IMF International carbon pricing floor – -31.14 6.23 15.57 0.48 (Baseline) Nationally Determined -33.14 6.62 16.57 0.51 Contribution (NDC) Below Two degrees -35.18 7.02 17.59 0.56 Net Zero by 2050 -47.20 9.44 23.6 0.75 Divergent Net Zero -62.12 12.42 31.06 0.99 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 25 6 >>> Banking sector stress testing A. REV ERSE ST RESS T EST ING RESULTS In this section we gauge the short-term resilience of Philippine banks to credit shocks, indicating the required size of a transition risk shock to trigger bank undercapitalization. Following the methodology proposed in Feyen and Mare (2021)35 we capture short-term resilience using a simple bottom-up reverse stress test that focuses only on credit risk and allows us to compute the consolidated distance to break point (CDBP). The CDBP is the minimum percentage-point increase in nonperforming loans (as a percentage of gross loans) necessary to trigger undercapitalization for banks representing at least 20 percent of banking system assets, a level commonly associated with a systemic banking crisis. We use bank balance-sheet data from Fitch in 202236 for the analysis. According to this analysis, the banking sector in the Philippines seems relatively resilient against credit risk and could sustain NPL increases associated with climate transition risk. The NPL ratio should reach a level of close to 14 percent, i.e., a CDBP of 10.7 percent or a 10.7 percentage point increase, before 20 percent of the banking system would breach the minimum capital adequacy ratio (CAR) due to depleted capital buffers. The CDBP of the Philippines banking system is higher than the EMDE average of 9.9 percent but lower than the East Asia and Pacific (EAP) average of 18.2 percent. This analysis suggests that the findings from Tables 2 and 3, are unlikely to pose systemic risk for the overall Philippines banking sector. Yet caveats of this analysis remain, which are outlined above. 35 For details, please refer to Figure 18 in the Appendix. 36 Fitch data was used to facilitate cross-country comparison with other EMDE banking systems. The Philippine sample includes the 30 largest commercial and foreign banks reporting to Fitch. 26 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES  AS I C BANKING STRESS T EST ING M O DE L TO GAUGE B. B IM PAC T ON OT HER BANK ING SEC TO R VA RIA B L ES To assess the impact of climate policies on bank balance and the coal sector has a higher share than might be expected sheets, we employ an accounting-based stress test in 2030. The most severe scenario is Divergent Net Zero, where framework developed by Čihák (2007) and link it to the UKB banks with a higher lending exposure to manufacturing estimated increases in nonperforming loans. The Čihák and transport experience more significant losses. RWA are model utilizes balance sheet and profit and loss information affected to a lesser extent ranging from a drop by 1.6 percent in from banks to evaluate the effect of conventional shocks in the Divergent Net Zero scenario to a drop of 0.6 percent under the financial sector on the banks’ CAR and risk weighted the IMF carbon tax scenario (Figure 14). However, overall assets (RWA). To incorporate sectoral credit exposure data, impacts are still relatively contained, and all bank types remain we modify this model.37 It is important to note that our analysis well-capitalized under each scenario. assumes static bank balance sheets as of the end of 2021, The most affected banks could face a drop in CAR in 2030 providing a baseline for evaluating the potential impact of of 25 percent (compared to an 8 percent drop for the overall climate policies on banks’ financial positions. banking system) in the Divergent Net Zero scenario (Figure The banking sector could face reductions in CAR by up 15), indicating the substantial heterogeneity across banks to 8 percent in 2030 under a Divergent Net Zero scenario; based on their sectoral lending portfolio. The bank level results however, overall impacts are still relatively contained, and are sorted by the Divergent Net Zero scenario, and the CAR all bank types remain well-capitalized under each scenario exhibits substantial impacts and variation between banks. The (Figure 13).38 The banking stress-test results on CAR and findings suggest that banks with a higher lending exposure to RWA suggest that sectoral shocks have a significantly stronger manufacturing and transport are more vulnerable to sector- impact than that estimated by the historical relationship of GDP specific shocks.39 At the same time, banks with exposure growth to NPL growth for the overall banking sector. This finding to sectors that might benefit from the low-carbon transition highlights the importance of considering sectoral portfolio (e.g., renewable electricity), might even face a reduction in composition when assessing the impact of stress scenarios NPLs and hence positive CAR effects from the transition. on banks; some sectors expand (e.g., wind energy) and other This information is useful for supervisors to understand the sectors (e.g., coal) decline substantially under ambitious climate potential risks associated with banks that have a significant policy scenarios. UKBs with a large portfolio share of transport lending exposure to specific sectors. However, the individual and manufacturing are particularly affected. The impacts are on bank analysis again indicates that banks can withstand low- the upper bound because NPL weights are taken from 2021, carbon transition impacts. 37 From the analysis above we estimate nonperforming loan (NPL) changes in response to changes in sectoral valued added for specific sectors, such as livestock, textile, and electricity from gas. However, the Čihák model uses a more comprehensive sector that groups detailed sectors. To address this, we conducted subsector-to-sector mapping, where we categorized subsectors such as livestock and crops under Agriculture, while textiles and clothing were classified under Manufacturing. We determined sector shocks by applying weights based on the value added of each subsector in 2021. The sector shock was then calculated as the sum of the weighted coefficients of each subsector that belongs to the stress test custom sector. We calculated the percentage change in NPLs for each stress test custom sector using the formula: The model then generates post-shock RWA and CAR values based on the new NPLs derived from each scenario. 38 The percentage of provisioning applied for new NPLs is assumed to be 25 percent. 39 See the Appendix section for a plausibility check comparing the portfolio composition of the most and least impacted banks in the sample. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 27 > > > F I G U R E 1 3 - CAR percent change by 2030 for different bank types across policy scenarios 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 Below 2°C, Nationally Divergent Net Zero 2050, IMF IMF USD 75 IMF USD 75 2030 determined Net Zero, 2030 USD 75 / t CO2 carbon tax carbon tax w/ Contributions, 2030 carbon tax, w/ slower carbon tax 2030 2030 renewable revenue to growth, investments, 2030 2030 All Banks UKBs TBs RCBs Source: Authors’ calculation. UKBs refer to universal and commercial banks, TBs refer to thrift banks, and RCBs refer to rural and cooperative banks. > > > F I G U R E 1 4 - RWA percent change by 2030 for different bank types across policy scenarios 0 -0.2 -0.4 -0.6 -0.8 -1 -1.2 -1.4 -1.6 -1.8 Below 2°C, Nationally Divergent Net Zero 2050, IMF IMF USD 75 IMF USD 75 2030 determined Net Zero, 2030 USD 75 / t CO2 carbon tax carbon tax w/ Contributions, 2030 carbon tax, w/ slower carbon tax 2030 2030 renewable revenue to growth, investments, 2030 2030 All Banks UKBs TBs RCBs Source: Authors’ calculations. UKBs refer to universal and commercial banks, TBs refer to thrift banks and RCBs refer to rural and cooperative banks. 28 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES > > > F I G U R E 1 5 - CAR percent change by 2030 for top 15 most affected banks across policy scenarios 0 -5 -10 -15 -20 -25 -30 UKB20 TB12 UKB19 UKB3 UKB17 UKB5 UKB7 RCB20 TB17 UKB10 UKB12 UKB14 UKB16 TB15 TB11 Below 2°C IMF USD 75 / t CO2 carbon tax Nationally determined IMF USD 75 / t CO2 carbon tax Contributions w/ slower renewable growth Divergent Net Zero IMF USD 75 / t CO2 carbon tax Net Zero 2050 w/ carbon tax revenue to investments Source: Authors’ calculations. UKBs refer to universal and commercial banks, TBs refer to thrift banks and RCBs refer to rural and cooperative banks. > > > F I G U R E 1 6 - RWA percent change by 2030 for top 15 most affected banks across policy scenarios 0 -0.5 -1 -1.5 -2 -2.5 -3 UKB19 UKB3 TB12 UKB12 UKB10 UKB17 UKB5 TB1 UKB7 UKB18 TB17 UKB16 TB15 UKB20 UKB14 Below 2°C IMF USD 75 / t CO2 carbon tax Nationally determined IMF USD 75 / t CO2 carbon tax Contributions w/ slower renewable growth Divergent Net Zero IMF USD 75 / t CO2 carbon tax Net Zero 2050 w/ carbon tax revenue to investments Source: Authors’ calculations. UKBs refer to universal and commercial banks, TBs refer to thrift banks and RCBs refer to rural and cooperative banks. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 29 7  LAUSIBILITY CHECK OF CLIMATE STRESS TESTING RESULTS C. P A plausibility check of the results compares sectoral storage. These sectors are particularly affected by the low- lending exposure to most transition affected sectors. We carbon transition assessment conducted by the MANAGE conducted a plausibility check of the analysis by examining model (refer to Figure 17). Banks with lower RWA and CAR the top and bottom banks at risk based on RWA and CAR losses tend to have a higher concentration of lending activities losses. Our analysis indicates that banks with higher RWA in the financial and other sectors. This information can be and CAR losses have a greater exposure to lending in sectors valuable for investors and stakeholders as it provides insights such as electricity and gas, manufacturing, transport, and into the lending structure of banks. > > > F I G U R E 1 7 - Plausibility Check: Bank Lending by Sector 100 Agriculture, Hunting, Forestry, and Fishing Manufacturing Electricity, Gas, Steam, Water and Air-Conditioning Supply, water supply, sewage, wastemanagement and 80 remediation activities Other Other Service activities Arts, Entertainment and Recreation Education 60 Human Health and Social Work Activities Public Administration and Defense, Compulsory Social Security, Administrative and Support service activities Professional, Scientific and Technical Activities 40 Real Estate Activities Financial and Insurance Activities Transport, Storage, Information and Communication Accommodation and Food Service 20 Activities, Hotels and restaurants Wholesale and Retail trade, repair of motor vehicles, Motorcycles, and Personal and Household Goods Construction Mining and Quarrying 0 UCB20 UCB19 TH12 UCB3 UCB12 UCB10 UCB7 TH20 TH10 RC7 RC5 TH5 TH13 TH14 TH6 TH9 TH3 Source: Authors’ calculations. UKBs refer to universal and commercial banks, TBs refer to thrift banks and RCBs refer to rural and cooperative banks. 30 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 7 >>> Conclusion and Policy Implications The Philippines is experiencing a rapid increase in GHG emissions, a trend that stands in contrast to the ambitious nationally determined contribution targets set by the country. Consequently, this dissonance poses the risk of a potential transition in which certain economic sectors may encounter challenges such as dwindling profits and revenues, with a subsequent impact on the stability of the financial sector. Our methodology adopts a modular framework, building on Reinders et al. 2021, which establishes a linkage between climate transition scenarios and a macroeconomic CGE model. Subsequently, the sectoral outputs derived from the CGE model are connected to impacts on banks’ NPLs through econometric analysis. Finally, we employ a consolidated approach that incorporates the distance to default and results from bank balance sheet stress tests, enabling the evaluation of the potential effects on banks’ risk-weighted assets. The analysis suggests that while the Philippines banking sector might experience limited financial stability effects due to climate transition risk, banks with significant exposure to high-emission sectors could face heightened vulnerabilities. As per the CDBP analysis, the banking sector in the Philippines demonstrates relative resilience against credit risk and appears capable of withstanding potential increases in NPLs associated with climate transition risk. The NPL ratio is projected to approach approximately 14 percent, translating to a CDBP of 10.7 percent or an increase of 10.7 percentage points before 20 percent of the banking system breaches the minimum CAR, as a result of diminished capital buffers. Under a Divergent Net-Zero scenario, the banking sector’s CAR could potentially decrease by up to 8 percent by 2030. Despite these impacts, the overall effects remain manageable, with all types of banks maintaining well-capitalized positions across the outlined scenarios. Notably, the most affected banks might encounter a 25 percent decline in CAR by 2030 (compared to the overall banking system’s 8 percent decline) in the Divergent Net-Zero scenario (refer to Figure 15), indicating significant divergence among banks based on their respective sectoral lending portfolios. It is important to acknowledge that the results presented in our analysis are contingent upon the assumptions and scenarios used in the modeling process. It is possible that additional factors not considered in our analysis could result in either larger or smaller transition risks. It is worth noting that our analysis only focuses on corporate lending within the banking sector. Other financial market segments, such as securities and equity impacts, are not currently included. This could be significant because larger fossil fuel-related firms may rely on these financial instruments, which could potentially amplify the risks of transition. Additionally, A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 31 our analysis does not account for interbank lending effects, of nonfinancial corporate balance sheet data is important, which could lead to second-round effects in the financial since it will allow the linking of economic impact and banking sector. This means that the interconnectedness between sector exposures through firm-level analysis, helping to banks and the potential ripple effects of transition risks are not better capture the heterogeneous drivers of transition risk currently considered. Furthermore, it is important to recognize at the sectoral level. Ensuring data completeness, quality, that transition scenarios do not incorporate climate physical and consistency, including sectoral definitions, is critical for risks. These physical risks, such as extreme weather events accurate risk assessments. Finally, as discussed in Section 4, or rising sea levels, could impose additional strains on the current macroeconomic models have important limitations, but financial sector, as indicated in the Philippines physical risk new tools are addressing some of the shortcomings. Hence, analysis,40 particularly under high-carbon trajectories on a this exercise and results should be interpreted with caution global scale. However, it is also worth noting that financial given the cited uncertainties or challenges. sector participants are not passive actors and may adjust their Our approach reflects the current state of knowledge and lending behavior in response to transition risks. This proactive capacity of financial regulators and financial institutions adjustment could potentially reduce transition risk. in climate risk assessment, which calls for extensive Our analysis faces several challenges related to capacity building efforts over time. First, financial methodology, modeling, and data, all of which highlight regulators and financial institutions with an interest in climate the uncertainties surrounding the quantitative results risk assessment should work closely with academia and the while pointing to important next steps in advancing relevant government agencies such as the Climate Change such assessments. First, given the long time horizon of Commission and the Department of Energy to develop and climate transition and high uncertainties related to the policy continuously update various scenarios, since this component and technology paths that could be undertaken, our results of the model falls outside the expertise of the financial sector. over the medium term come with significant uncertainties. Second, in the medium term, central bank and financial Second, current modeling infrastructure does not allow for regulators should develop internal macroeconomic modeling the endogenous responses of the financial sector to projected capacity, so that a set of in-house tools designed to analyze economic impact, so extrapolating for financial risks beyond climate related risks is available for use and a climate risk three to five years remains a challenge. Third, the availability assessment could be self-contained. Although this paper 40 See Hallegatte et al. 2022 for details. 32 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES focuses on the banking sector, it is also helpful to include the necessary disclosure and reporting requirements in place is insurance sector in future exercises due to its important role a precondition for successful monitoring and assessment of in risk diversification and its exposure to climate-related risks, climate-related risks for the financial sector. Communicating particularly through the agricultural sector. Finally, having the stress test results also improves awareness of climate risks capacity to incorporate the financial sector in central bank and may encourage banks to strengthen their risk management macroeconomic modeling is also important and could reduce practices and foster further research. Thus authorities may the need to have distinct blocks for financial risk modeling. also leverage the content of this report to raise awareness and understanding of climate and environment-related risks Nonetheless, this report provides a source of reference for among financial sector participants as well as broader civil financial institutions interested in undertaking a climate society. Also, capacity building and data collection initiatives risk assessment. As high-level supervisory guidance for are not limited to financial sector authorities and may involve financial institutions to incorporate climate and environmental other relevant line agencies. risks into their risk management and governance framework is in place, financial institutions looking to conduct pilot climate Relatedly, it is important for central banks to continue risk stress testing could benefit from guidance on methodology, collecting data and monitoring climate-related risks. such as scenario development and economic modeling. Central banks play a crucial role in providing supervisory Sectoral analysis could provide insights to banks looking at guidance to banks. The establishment of a climate data the transition risk of their individual portfolio allocation. catalogue is essential, encompassing not only transition risks but also nature-related risks. This catalogue should be The result of this report has important implications for expanded to include the capital market and insurance sectors, relevant government authorities, especially on reporting ensuring a comprehensive understanding of climate and and disclosure. Relevant authorities in charge of setting environment-related financial risks. By taking these steps, climate economic and financial policies could use information central banks can contribute to building a more resilient in this report to inform the development of policy frameworks financial system in the face of climate change. and initiatives. Specifically, this report points to the importance of requiring climate related disclosure from financial institutions as well as corporates and other stakeholders. Having the A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 33 References Bingler, Julia, Mathias Kraus, and Markus Leippold. 2022. “Cheap Talk and Cherry-Picking: What ClimateBert has to say on Corporate Climate Risk Disclosures.” Finance Research Letters, Volume 47, Part B, 102776. https://papers.ssrn.com/sol3/ papers.cfm?abstract_id=3796152. Monasterolo, Irene. 2020. “Climate Change and the Financial System,” Available at SSRN: https://ssrn.com/ abstract=3479380 or http://dx.doi.org/10.2139/ssrn.3479380. Network for Greening the Financial System. 2020. “The Macroeconomic and Financial Stability Impacts of Climate Change Research Priorities.” www.ngfs.net/sites/default/files/medias/documents/ngfs_research_priorities_final.pdf. Vermeulen, Robert, Edo Schets, Melanie Lohuis, Barbara Kölbl, David-Jan Jansen, and Willem Heeringa. 2021. “The heat is on: A framework for measuring financial stress under disruptive energy transition scenarios.” Ecological Economics, Volume 190, 107205. https://doi.org/10.1016/j.ecolecon.2021.107205. Reinders, Henk Jan, Dirk Schoenmaker, Mathijs Van Dijk. 2023. “DP17921 Climate Risk Stress Testing: A Conceptual Review.” CEPR Discussion Paper No. 17921. https://cepr.org/publications/dp17921 Regelink, Martijn G.J.. 2019. “Philippines Financial Sector Assessment Program: Climate Change and Environmental Risks and Opportunities.” World Bank. http://hdl.handle.net/10986/36182. Semieniuk, Gregor, Philip B. Holden, et al. 2022. “Stranded fossil-fuel assets translate to major losses for investors in advanced economies.” Natural Climate Change. 12, 532–538. https://doi.org/10.1038/s41558-022-01356-y. Trutnevyte, Evelina. 2016. “Does cost optimization approximate the real-world energy transition.” Energy, Volume 106, 1 July, pp 182-193. https://doi.org/10.1016/j.energy.2016.03.038. Pollitt, Hector and Jean-Francois Mercure. 2018. “The role of money and the financial sector in energy-economy models used for assessing climate and energy policy.” Climate Policy, Volume 18, Issue 2, pp 184-197. www.tandfonline.com/doi/full/10.10 80/14693062.2016.1277685. Werner, Richard A. 2014 “Can banks individually create money out of nothing? — The theories and the empirical evidence.” International Review of Financial Analysis, Volume 36, pp 1-19. https://doi.org/10.1016/j.irfa.2014.07.015. Mercure, Jean-Francois, Hector Pollitt, et al. 2018b. “Environmental impact assessment for climate change policy with the simulation-based integrated assessment model E3ME-FTT-GENIE.” Energy Strategy Reviews, Volume 20, April 2018, pp 195–208. doi.org/10.1016/j.esr.2018.03.003. Burns, Andrew, Benoit Campagne, et al. 2019. “The World Bank Macro-Fiscal Model Technical Description.” Available at https://documents1.worldbank.org/curated/en/294311565103938951/pdf/The-World-Bank-Macro-Fiscal-Model-Technical- Description.pdf. Hallegatte, Stephane., Floren McIsaac, et al. 2023. “The Macroeconomic Implications of a Transition to Zero Net Emissions: A Modeling Framework.” Policy Research Working Papers; 10367. World Bank, Washington, D.C. https://openknowledge. worldbank.org/entities/publication/659cc0e8-38a4-4657-b8bd-6aa4373a9dc2. International Monetary Fund. 2006. “Spain: Financial Sector Assessment Program - Technical Note: Stress Testing Methodology and Results.” IMF Staff Country Reports, 06(216), https://doi.org/10.5089/9781451812190.002. 34 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES Nkusu, Mwanza. 2011. “Nonperforming Loans and Macro-financial Vulnerabilities in Advanced Economies.” IMF Working Paper WP/11/161, International Monetary Fund. https://www.imf.org/external/pubs/ft/wp/2011/wp11161.pdf. Beck, Roland, Petr Jakubik, and Anamaria Piloiu. 2013. “Non-performing loans: what matters in addition to the economic cycle?” Working Paper Series 1515, European Central Bank. https://www.ecb.europa.eu/pub/pdf/scpwps/ecbwp1515.pdf Lee, Junkyu and Peter Rosenkranz. 2020. “Nonperforming Loans in Asia: Determinants and Macro financial Linkages.” International Finance Review, Volume 21, 33–53. https://doi.org/10.1108/S1569-3767202021. Ari, Anil, Sophia Chen, and Lev Ratnovski. 2019. “The Dynamics of Non-Performing Loans during Banking Crises: A New Database.” IMF Working Paper WP/19/272, International Monetary Fund. https://www.imf.org/en/Publications/WP/ Issues/2019/12/06/The-Dynamics-of-Non-Performing-Loans-during-Banking-Crises-A-New-Database-48839. Alnabulsi, Khalil, Emira Kozarević, and Abdelaziz Hakimi. 2022. “Assessing the determinants of non-performing loans under financial crisis and health crisis: Evidence from the MENA banks.” Cogent Economics & Finance, 10(1), 2124665. https://doi. org/10.1080/23322039.2022.2124665. Feyen, Erik and David Salvatore Mare. 2021. “Measuring systemic banking resilience, a simple reverse stress testing approach.” Policy Research Working Papers; 9864. World Bank, Washington, D.C. https://econpapers.repec.org/paper/ wbkwbrwps/9864.htm. Cihak, Martin. 2007. “Introduction to Applied Stress Testing.” IMF Working Paper 07/59, International Monetary Fund, Washington, DC. https://www.imf.org/en/Publications/WP/Issues/2016/12/31/Introduction-to-Applied-Stress-Testing-20222. A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 35 Appendix A P P END IX I – LIST OF MANAGE M O DE L SEC TO RS MANAGE MODEL SECTOR CODE SECTOR NAME a-pal Palay a-cor Corn a-coc Coconut a-sug Sugarcane a-ban Banana a-crp Other crops a-lvs Livestock a-frs Forestry and fishing a-coa Coal extraction a-oil Oil extraction a-gas Gas extraction a-oxt Other mining and quarrying a-ofd Food manufactures a-b_t Beverage and tobacco industries a-tex Textiles and clothing a-lum Wood, bamboo, cane and rattan articles a-ppp Paper and paper products, printing and publishing a-p_c Petroleum and other fuel products a-chm Chemical and chemical products a-bph Basic pharmaceutical products and pharmaceutical preparations a-rpp Rubber and plastic products a-nmm Non-metallic mineral products a-met Basic metal industries a-fmp Fabricated metal products a-ele Computer, electronic, and optical products a-eeq Electrical equipment a-ome Machinery and equipment except electrical a-mvh Transport equipment a-omf Other manufacturing a-TnD Electricity transmission and distribution a-eCo Electricity from coal a-eGa Electricity from gas a-eHy Electricity from hydro a-eWi Electricity from wind a-eSo Electricity from solar a-eOt Electricity from other sources continued on the next page 36 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES continued from previous page MANAGE MODEL SECTOR CODE SECTOR NAME a-wtr Water a-wst Sewerage and wastewater remediation activities a-cns Construction a-trd Distribution and retail a-otp Land transport a-wtp Water transport a-atp Air transport a-cmn Communication activities a-afs Accommodation and food service activities a-ofi Financial services a-rsa Real estate activities a-obs Other business services a-osg Public administration and defense a-edu Education a-ros Arts, entertainment, and recreation a-hht Human health and social work activities a-osr Other service activities, nec A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 37 A P P END IX II – ADDIT ION A L FIGURES > > > F I G U R E 1 8 - Outline of Reserve Stress Testing Framework Credit Shock  NPLs  Provisioning  Capitalization Bank Break Point (BP) Bank Break Point (BP) NPLs (% of gross loans) that NPLs (% of gross loans) that BANK LEVEL would deplete regulatory would deplete regulatory capital buffers capital buffers Impact Most fragile bank representing at least 20% of a country’s banking system COUNTRY LEVEL assets (“Banks at Risk”) fall at or below the minimum capital to assets ratio (min CAR) Source: Feyen and Mare (2021) 38 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES > > > F I G U R E 1 9 - Winning and Losing Sectors Due to Climate Transition Ten sectors with most gains SECTOR GRAND TOTAL 44 Human Health and Social Work Activities -3.33 45 Other Service Activities, nec 0.77 46 Coconut 3.81 47 Education 8.11 48 Accommodation and Food Service Activities 22.24 49 Other business services 61.49 50 Elctricity from other sources 153.52 51 Electricity from hydro 868.89 52 Electricity from solar 2877.12 53 Electricity from wind 16251.92 Ten sectors with least gains SECTOR GRAND TOTAL 1 Electricity from coal -1348.51 2 Non-metallic mineral products -959.87 3 Gas extraction -836.01 4 Sewerage and waste water remediation activities -753.48 5 Electricity from gas -624.67 6 Water transport -520.38 7 Oil extraction -424.73 8 Coal extraction -408.73 9 Electrical equipment -395.64 10 Machinery and equipment except electricall -389.87 Note: units are in percentage change from baseline, cumulative through 2021-2040 A CLIMATE TRANSITION RISK ASSESSMENT FOR THE BANKING SECTOR OF THE PHILIPPINES 39