Policy Research Working Paper 10983 The Potential Cascading Impacts of Climate Change in Cambodia Hector Pollitt Migle Petrauskaite Macroeconomics, Trade and Investment Global Practice November 2024 Policy Research Working Paper 10983 Abstract This paper develops a “plausible worst-case” scenario for financial contagion remains, especially due to the growing Cambodia to illustrate how a severe, 1-in-10-year flood sovereign-bank nexus. The paper highlights the importance could trigger cascading impacts, including widespread of integrating climate risks into Cambodia’s broader risk disease outbreaks and financial instability. The analysis management strategies and suggests preemptive interven- shifts from forecasting to risk management, focusing on tions, such as improving flood forecasting, health care the economic consequences at each stage of this disas- infrastructure, and exploring disaster risk finance instru- ter chain. As climate change increases the frequency and ments. These measures could help mitigate the cascading severity of extreme weather events, Cambodia’s vulnera- impacts of climate-induced disasters and build long-term bilities are likely to intensify, with severe floods leading to resilience. The paper concludes that a shift from reactive disruptions in health care, declines in labor productivity, crisis management to proactive preparedness and adapta- and risks to financial stability. Although Cambodia’s cur- tion will be crucial for Cambodia’s ability to manage future rent financial position provides some resilience, the risk of climate risks and ensure economic and social stability. This paper is a product of the Macroeconomics, Trade and Investment Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at hpollitt@worldbank.org and mpetrauskaite@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team The Potential Cascading Impacts of Climate Change in Cambodia Hector Pollitt*, Migle Petrauskaite* * World Bank Group, Washington DC Keywords: climate risk, cascading disasters, scenario modeling JEL classification: D81; E17; Q51; Q54 This paper is a product of the Macroeconomics, Trade, and Investment (MTI) Global Practice, East Asia Pacific region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at hpollitt@worldbank.org and mpetrauskaite@worldbank.org 1. Introduction Cambodia is facing multiple crises. The lingering effects of the COVID-19 pandemic, growing impacts of climate change, economic imbalances, and rising debt distress make the country vulnerable to compound shocks. The unique nature of compound risks – characterized by non- linear1 and hard-to-predict effects – defies simple summation of individual shocks. These complexities amplify the consequences of climate-related shocks and introduce formidable challenges to financial stability and monetary policy. When compound shocks intersect with pre- existing vulnerabilities, they may unleash cascading effects, further magnifying impacts and increasing the risk of enduring consequences (Pescaroli and Alexander, 2018). This paper considers a ‘worst-case plausible’ scenario in which a chain of ‘cascading’ disaster events is triggered. It draws on the principle of ‘risk thinking’ laid out in Dembo (2021) and builds on the possibility of climate-induced cascading risks described in Pescaroli and Alexander (2018) and Kemp and others (2022).2 Although the definition of ‘plausible’ is necessarily ambiguous to reflect the underlying uncertainty about the events discussed, the scenario laid out in this paper resembles the situation in Pakistan in 2023, whereby the country experienced disease outbreaks and severe financial instability following the floods in 2022. The scenario is not a prediction; instead, it aims to depict potential outcomes in the absence of sufficient preparations to protect against future losses. This paper highlights the range of spatial and temporal scales that can ultimately lead to extreme impacts: from long-term global warming making climate and weather patterns more volatile, through to the increased occurrence and severity of flooding events along the Mekong River Basin, and finally to the localized effects in terms of disease outbreaks and devastating damage to the region’s economies. While each individual disaster event may be relatively unlikely, the chances of it happening may increase both due to climate change and due to previous events in the chain occurring (Zscheischler and others, 2018). The probability of the full scenario occurring (if it could be measured) is thus higher than multiplying the probabilities of each individual element. Climate change increases the likelihood and severity of disaster at every stage in the chain. The aim of this exercise is also to identify key vulnerabilities within Cambodia, which may be exacerbated by climate change. Having identified these vulnerabilities, it becomes possible to assess pre-emptive responses that could reduce future risks within Cambodia. Given that Cambodia cannot much influence the rate of global climate change, these responses could play a key role in the country’s long-term strategy for developing a resilient economy. 1 In the context of compound shocks, non-linearity implies that the combined impacts of events have the potential to diverge from what would be expected by simply adding up the impacts of individual shocks. 2 For clarity, we adopt the Intergovernmental Panel on Climate Change (IPCC) definition of ‘risk,’ which focuses on the potential for adverse consequences. Unlike economic definitions that consider both positive and negative outcomes (Keynes, 1921; Knight, 1921), the IPCC’s definition applies only to adverse events. While the probabilities of these events cannot be measured, this analysis focuses on downside risks. A similar definitional issue arises with the use of the term ‘conservative,’ which has different connotations in economic and environmental contexts. In economics, ‘conservative’ often refers to estimates that are likely milder than the actual outcome, while natural scientists may use ‘conservative’ to indicate a worst-case scenario. Throughout this paper, we avoid using the term ‘conservative’ without clarifying its intended meaning. The design of the scenario is summarized in Figure 1. It requires a highly interdisciplinary approach to assess adequately. The following sections discuss each part of the chain sequentially, highlighting the different elements in the chain: severe flooding, disease outbreak, economic impact and financial crisis. Section 5 then discusses how to build a strategy to protect against the risk of such cascading failures in Cambodia. The final section draws broader conclusions from the exercise. Figure 1: Summary of the worst-case plausible scenario Climate change more likely more likely ly m ike or el Outbreak of disease el or ike m ly Severe floods Financial crisis ly m Compromised healthcare ike or el el or ike m ly One factor that is not considered in this paper is possible feedback mechanisms in the opposite direction to those in the figure. For example, a financial crisis or an outbreak of disease could hamper measures to reduce the impacts of a flood. There is therefore a potential self-reinforcing effect between the different events that would push impacts towards the higher end of expected outcomes. 2. The potential for severe flooding 2.1. Historical data and recent analysis The risk of severe flooding in Cambodia is already well known. Table 1 summarizes some of the previous large floods that have been recorded, over approximately a 35-year timeframe. Each of these floods impacted at least 10 percent of the population of Cambodia. The figures show the possibility for at least some of the country to remain submerged for extended periods of time, suggesting a prolonged loss of economic production in some areas. The scale of damages from the floods varies substantially, depending on where the flood takes place. However, as Cambodia develops and builds new infrastructure, the level of damages in monetary terms is expected to increase. Table 1: Impacts of large floods Year Duration (days) No. people affected (m) Damages ($m 2022 price) 1996 35 1.30 2.8 2000 31 3.45 272.0 2001 97 1.67 24.8 2002 101 1.47 0.2 2011 84 1.64 677.8 2013 21 1.50 628.1 Notes: Table shows floods with more than 1m people affected in Cambodia since 1990. Source: EM-DAT database. In the period since 1990, four floods have impacted more than 50,000 km2 in Cambodia, representing more than a quarter of the country’s total land mass. Davies and others (2014) report that 18 of the 24 provinces were impacted in the 2011 floods. The 2011 floods affected more than 10 percent of the population and caused over $677 million in damages (5.3 percent of GDP). Even if the floods do not hit major urban centers (meaning that monetary damages are much reduced), the impacts on agriculture and transport infrastructure can be substantial. Despite the 2011 floods having the highest recorded cost in terms of damages, ADB (2012) suggests that more than half of the damages could be attributed to agriculture (particularly rice paddy) and transport infrastructure. Furthermore, damage to transport infrastructure can have important implications for providing disaster relief and assistance in recovery (Davies and others, 2014), which increases the possibility of cascading disaster effects (see next section). The impacts of disasters can be difficult to identify in macroeconomic statistics because efforts to rebuild can create a stimulus effect that counterbalances production losses during the disaster. However, ADB (2012, p.2) reported a noticeable long-term impact on economic development in Cambodia from the 2011 floods: “Despite the positive overall economic performance of Cambodia in 2011, the flood slowed the potential for sustained economic growth due to the devastating impact at the household and macroeconomic levels.” Even in cases where resources are diverted toward recovery efforts, there could be displacement effects, for example if public services are cut to balance budgets. Alternatively, an increase in debt levels could impact financial stability, as discussed in Section 4. Table 1 shows that large floods with substantial human and economic cost already occur in Cambodia on a frequent basis, in part because of the intensity of rainfall in the rainy season (when 80 percent of annual rainfall occurs). Paltan and others (2018) report that climate change will increase the frequency further; even under a scenario with limited climate change, what was previously a 1-in-100-year flood could occur once every 25 years or even once every five years. World Bank Group (2021) summarizes the recent literature and finds that 25 percent of Cambodia’s population is expected to be exposed to floods between 2035 and 2044. However, there remains substantial uncertainty about both the scale and potential impacts of future floods. Phy and others (2022) note that Cambodia lacks both flood hazard mapping and risk and damage assessment. The authors recommend further data collection and model-based analysis, combined with predictions about future land use change and infrastructure development, to identify key future risks in Cambodia. 2.2. Modeling the economic impacts of floods The risks from flooding are confounded by the high degree of uncertainty about the scale and location of future floods. There is further uncertainty about the economic impact of these floods. This uncertainty is not well reflected in standard modeling approaches, which typically adopt assumptions about perfect knowledge. In this section we attempt to quantify potential losses of production from flooding. A set of eight possible scenarios is assessed, each one based on the likelihood of the flood occurring. Using the probabilistic risk assessment approach outlined in the World Bank Group’s (2023) Cambodia Country Climate and Development Report (CCDR), we consider both low-frequency, high-impact events and high-frequency, lower-impact events to capture a realistic range of outcomes, including those that may arise under projected climate conditions. The data taken from the floods analysis are: i) the potential losses of production from manufacturing and services sectors; and ii) the expected duration of the loss of production. The loss of production in manufacturing sectors was disaggregated further, using the split of impacts in the 2011 Thailand floods. This disaggregation is important for informing analysis of possible bottlenecks in production supply chains (see below). Although the pattern of impacts across sectors in Cambodia is likely to be different to that in Thailand, the aim of the exercise is to give a sense of the scale of the heterogeneity of the impacts and how these could in turn affect complex supply chains. The total loss of production is estimated as being equal to the total duration of the flood (defined as time the business is inundated with water) plus the time to rebuild production capacity. Rebuilding is assumed to be a gradual process that starts immediately once flood drainage has been completed. We draw on Hallegatte (2014) and use the same approach as Tanoue and others (2020) in assuming a linear trajectory of recovery. The time to full recovery is assumed to be ten times the duration of the flood itself, which compares to an estimate of 11.5 times for Thailand (ibid). Given that rebuilding is expected to be faster in more advanced economies, our assumption for total duration of lost production may therefore understate the true value. Because production recovers linearly during the rebuilding period, with companies resuming production when ready, the average recovery time is half the value of ten. If the duration of the floods is expressed in days, the total loss of annual production is therefore: 10 7  + 2  =   = %   ∗ 365 These losses in themselves could be substantial. However, there may also be two types of indirect economic impacts. First, supply shortages may prevent companies from producing, even if their own facilities are undamaged. Second, a loss of income may lead to further losses of production across the economy due to demand deficiency. Both types of effect are typically missing from standard equilibrium-based analysis. Computable General Equilibrium (CGE) models usually assume that the economy can immediately adjust to its new conditions and find a new optimal position. In this way supply shortages are restricted to where they cause least economic damage. Losses of demand are similarly largely absent because market clearing mechanisms ensure that supply and demand can match exactly. In reality, however, there is little time for any such adjustment in a disaster situation. The potential for further losses of production from both types of effect is therefore real. Here we use a non-equilibrium modeling approach to demonstrate the potential impact of secondary effects of floods on annual production in an arbitrary year around 2030. This approach reflects the impacts both of a loss of factor inputs, and from the disruption that the flood causes. A critical variable in the analysis is how well prepared businesses are for potential supply-chain disruptions. Options for supply-chain resilience include holding stocks and keeping the flexibility to change suppliers if necessary. Figure 2 shows a range of potential outcomes, varied by the degree of supply-chain resilience measured as the number of weeks it is possible to maintain production in the event of a loss of inputs (ranging from zero to six weeks). Figure 2: Potential loss of production from floods Note: The red cross represents the flood severity discussed in this note, given a medium degree of supply-chain flexibility. Source: Authors’ calculations based on MINDSET model data from the GLORIA database. The scenarios in the figure assume no new adaptation measures to manage the impacts of stronger floods resulting from climate change. The figure shows that the impacts of common floods can be mostly avoided through supply-chain flexibility. To some extent we see this already in the historical data, although, as noted above, the data may mask the loss of production in the stimulus caused by rebuilding efforts. In addition, the impact of future floods is expected to be more severe because of climate change. Even for larger floods, supply chain flexibility has benefits for reducing damages. However, for low-probability, high-impact floods, measures to prevent the floods from impacting on businesses in the first place (i.e., climate change adaptation measures like flood walls and channels), would be required to prevent widespread loss of production in the range of 20 percent. It is important to stress that the results above are intended to illustrate worst-case outcomes. They are not predictions of future impacts but estimates of what could happen if adequate preparations are not made to offset future losses. These impacts would happen in the single year in which the flood occurred, representing a sudden loss of production. The story that emerges of what is the worst that could happen is much different to the standard question of what we expect to happen. Nevertheless, it seems implausible that some measures will not be taken to offset the worst of the damages and therefore we take forward the results of the 1-in-10-year scenario as the benchmark figure for the following sections. However, this figure must not be interpreted as an ‘absolute worst case’ outcome. The model results show that the risk of large-scale loss of output is real. In the absence of suitable response measures, climate change will make it more likely. 3. The potential for disease outbreaks The effects of a large flood will be felt across all economic sectors. The model results from the previous section show that, with a medium degree of preparedness, aggregate impacts of 10 percent of GDP are possible in a future scenario with a medium-likelihood flood (see Figure 2), with larger impacts possible if no preparations are made. Leaving aside the loss of production, the results also show the cost of repairs could be up to $940 million. The repair bill would include $100 million of damages to homes, potentially displacing 264,000 people (1.6 percent of the population). The inundation of the agricultural sector, the main source of income for 80 percent of the flood-affected population, could have knock-on effects to a wide range of livelihoods that depend on crop production. Given the damages to dwellings and other buildings in rural areas, health care needs could increase at the same time as capacity falls because of flooding to health care facilities. Modeling estimates indicate that approximately 25 percent of the sector could experience partial incapacitation with floods lasting at least a week, and a recovery period several times longer. The provinces along the Mekong River are expected to bear the brunt of this impact. The authorities’ ability to provide adequate health care services may be further limited by the difficulty of reaching people in heavily flooded areas. Disruptions to health care could be especially important because, as floods intensify with climate change, different types of disease outbreaks will become more likely. Climate change will already have negative health impacts, including more common outbreaks of disease, and flooding events exacerbate the effects of pre-existing diseases, as well as bringing new ones through contaminated floodwaters. Water-borne illnesses are of particular concern in Cambodia, given the high pre-existing burden of diseases and the lack of advanced sanitation infrastructure in rural areas. Over 80 percent of the population lives and works on flood plains, and a large proportion of the rural population lacks access to improved water sources, further increasing the risk of disease transmission. As a result, consumption of contaminated water during flooding events is common, increasing the risk of disease transmission. Several studies have established a connection between floods and disease outbreaks, with transmission occurring through a variety of pathways (see, for example, Ahern and others, 2005; de Man and others, 2017). The pathways include waterborne diseases such as diarrheal disease and cholera, as well as vector-borne diseases like dengue, malaria, and leptospirosis. Additionally, floods can contribute to the spread of diseases such as skin infections, dermatitis, conjunctivitis, and others, particularly in densely populated areas where human–animal interactions are frequent (McIver and others, 2016). The standing pools of water created by more frequent flooding provide breeding grounds for mosquitoes that carry malaria and dengue virus, increasing the severity of outbreaks. Children are especially vulnerable: in Cambodia, diarrhea and malaria remain one of the primary causes of child mortality. The two biggest dengue outbreaks were recorded in 1998 and 2007, affecting 15,000 and 17,000 people, respectively. Beauté and Vong (2010) highlight the high societal and individual family burden of dengue, with total costs ranging between 0.03 percent and 0.17 percent of GDP. These figures may substantially underestimate the real costs caused by endemic dengue, given that it primarily strikes children. According to the World Health Organization (WHO), climate-sensitive diseases cause up to 400,000 deaths per year in Cambodia, and this figure is projected to increase by 10 percent because of climate change. By 2050, the extra health burden from climate change on climate-sensitive diseases would therefore be about $120 million, or 0.85 percent of GDP. The Asian Development Bank’s 2009 climate change impact study cites a higher figure, suggesting that climate change could increase the burden of some climate-sensitive diseases by as much as 18 percent over the next 20-30 years. Although the dangers of compromised health care are clear, health centers in flood-affected provinces have so far demonstrated resilience in past flooding events, including the 2011 floods. According to the ADB (2012), most health centers were able to resume their activities within a few days of the 2011 floods clearing, and most were overall in reasonably sound condition. Despite severe damage to rural road networks, health centers were able to overcome logistical and operational challenges by relocating to evacuation sites and safer areas away from flood zones. Furthermore, the WHO (2019) reported that in the aftermath of storm-induced flooding, health workers used boats to provide outreach services to the poorest and most vulnerable populations, running mobile clinics. Such initiatives underscore the critical role of health care in disaster response and the importance of directing investments in the sector to improve its resilience and preparedness for future flooding events. Like the incidence of floods, outbreaks of disease will not be uniform across the period to 2050, especially for diseases related to floods. A substantial economic burden could therefore arise from disease outbreaks linked to floods in the years that the floods occur. It is difficult to predict the overall economic impact of disease outbreaks because the effects would be non-linear. Such an estimate would need to include the loss of a productive labor force due to direct incidence of disease, the loss of productive labor force because of caregiving requirements for the sick, the redirection or suppression of economic activity to prevent further spread of disease, and the indirect multiplier effects that would restrict economic demand. The economic costs would include the same factors, regardless of whether the disease outbreak resulted from a flood, although the incidence of a flood makes the occurrence more likely. It is important to note that disease outbreaks could also occur indirectly from flooding, especially if health care services are compromised. For example, routine treatments and vaccinations would likely be suspended in flooded areas. In 2020, measures to control COVID-19 caused economic growth in Cambodia to fall from 7 percent to –3 percent (i.e., a net loss of 10 percentage points), with further subdued growth in 2021 (source: WDI database). A loss of 5 percent of production in Cambodia would therefore suggest a serious outbreak that stops production across large parts of the economy for an extended period. It is noted that there may be some degree of double counting if we add together the effects of floods and disease outbreaks (e.g., sick agricultural workers may not have fields to tend to anyway). However, the approach is appropriate for assessing a plausible worst-case outcome and any interaction effects could make the outcomes worse. Figure 3 thus summarizes the combined potential impact of the floods and disease outbreak. Figure 3: Potential loss of production from floods and disease outbreak combined Source: Authors’ estimates. As with the floods, a range of outcomes is possible. The 10 percent loss of output from COVID- 19 could perhaps be considered an absolute worst-case outcome (on the basis that authorities are now better prepared). The impacts could obviously be smaller as well. Still, it is worth noting that if we took the absolute worst-case outcomes for both floods and disease outbreak, the combined impact could be a 30 percent loss of production rather than a 15 percent loss. 4. The risk of a financial crisis It is already well established that natural disasters can push countries toward insolvency. Mallucci’s (2020) analysis of recent cases of loan default highlights that severe weather conditions have frequently been an important contributing factor. Weather conditions are particularly salient for small countries with economies that are heavily reliant on agriculture, because a large share of the country’s land may be impacted in a single event. Countries in which economic activity is concentrated in a small geographical area may also be vulnerable. There are many examples of countries that have defaulted because of climate conditions. Moldova and Suriname experienced defaults in 1992 and 1998, respectively, due to severe droughts that negatively impacted the production of agricultural goods intended for export (IMF, 1999; van Dijck and others, 2000). El Niño floods in Ecuador in late 1997 and early 1998 led to thousands of deaths, loss of household assets and crops, and caused extensive damage to vital infrastructure. The floods destroyed large agricultural areas, reducing exports, and impairing several banks’ assets mainly from the country’s coastal region. Ecuador defaulted several months after severe floods caused widespread power outages (Sturzenegger and Zettelmeyer, 2007). Natural disaster risk also reduces the government’s ability to issue debt. In Pakistan, the 2022 floods worsened the country’s risk of defaulting on external debt. The floods affected 33 million people, causing substantial damage across 81 districts, and exacerbated existing vulnerabilities. Health systems were strained, with over 1,460 facilities impacted, while waterborne diseases surged (ReliefWeb, 2022). The floods led to a 2.2 percent direct GDP loss, with indirect effects causing GDP growth to fall from 6 percent to below 2 percent for 2022-23. Inflation reached 24 percent in 2023, worsening poverty. Damaged infrastructure, reduced agricultural output, and increased unemployment further hit key sectors, pushing the fiscal deficit to 5.5 percent of GDP. By late 2023, Pakistan faced insolvency, with foreign exchange reserves critically low and the value of sovereign bonds falling by over 60 percent (EIU, 2023). Whether climate/disaster related or not, financial crises lead to substantial and enduring costs. These costs may manifest as elevated levels of unemployment, decreased economic growth, and heightened social and political instability. Reinhard and Rogoff (2009) document how during the financial crisis-induced downturn phase, which lasts an average of more than four years, the unemployment rate tends to rise by around 7 percentage points. Output also declines substantially, falling by, on average, more than 9 percent from peak to trough. Public debt tends to increase substantially, driven largely by the sharp decline in tax revenues that governments experience as a result of deep and long-lasting output contractions. In more recent analysis, Laeven and Valencia (2020) systematically show that banking crises impose large costs on the sovereign and find that sovereign debt and currency crises tend to follow banking crises. The external shocks described in this paper could be sufficient to trigger a financial crisis, under certain conditions, with potential losses of around 15 percent of GDP. Typically, the root causes of financial crises are initially small; for example, in July 2008 the value of defaulting sub-prime mortgages was valued at $273 billion, which led to $8 trillion of equity losses.3 The ensuing bank failures and loan pullbacks exacerbate the recession, leading to more loan defaults and further bank failures. Given the crucial role that financial systems play in modern economies, including Cambodia, a freeze-up in the banking system could exacerbate or even multiply the direct effects of the crises. Although Cambodia was relatively unaffected by the Asian Financial Crisis in 1998, Thailand and Malaysia recorded GDP losses above 7 percent, while Indonesian GDP fell by more than 13 percent. A higher share of agriculture in the economy means that the impacts in Cambodia could be smaller than those in other ASEAN countries, but the high rate of dollarization could increase potential costs. If we include the lost baseline growth and subsequent weaker growth, net losses 3 Mayer and others (2008) reported that 21 percent of the reported $1.3 trillion of sub-prime loans was delinquent. of 10 percent of GDP (slightly higher than Reinhard and Rogoff’s (2009) mean estimate) appear plausible. Higher losses could be realized in an absolute worst-case outcome scenario. The central flood scenario, accompanied by disease outbreak, could lead to fiscal vulnerability in Cambodia; but under current financial conditions, it is unlikely to trigger a sovereign default. A one-off 15 percent reduction in GDP and the corresponding drop in tax revenue would dramatically widen Cambodia’s fiscal deficit, given lower economic growth expectations and increased expenditure commitments from the floods. The bulk of the financial needs will be in the period up to two years after the flood has receded, when most of the reconstruction would take place. To finance this additional expenditure, the government would need both to prioritize reconstruction spending and borrow additional resources to fund the required measures. Current debt and deficit levels remain moderate and sustainable, while broad-based economic growth has helped boost tax receipts, bringing the fiscal deficit down to 4.4 percent. The total level of public debt, estimated to be at 37 percent of GDP, compares favorably to the median of emerging markets, indicating that the government is in a good position to borrow to finance the deficit if a large flood did hit the country. But as the severity of natural disasters increases, so will sovereign borrowing costs. Research by Beirne and others (2021) emphasizes that vulnerability to the direct consequences of climate change has larger impacts on sovereign borrowing costs than climate risk resilience. The impact on bond yields becomes particularly pronounced for countries like Cambodia that are highly susceptible to climate change. There is a range of transmission channels through which climate change can amplify sovereign risk and thus worsen a sovereign’s standing, including through the depletion of natural capital, affecting fiscal sustainability and sovereign risk pricing (e.g., Pinzón and others, 2020), and via fiscal consequences of climate-related disasters, which disrupt economic activity, reducing tax revenue and increasing social transfers (e.g., Schuler and others, 2019). These transmission channels are not independent of each other, with climate impacts magnifying the transmission of risk through multiple channels. Climate change could further impact sovereign borrowing costs through broader macroeconomic implications, negatively affecting credit ratings. It is well established that climate risks can lead to lower investment, financial losses, and asset valuation disruptions (NGFS, 2018). Several studies also reveal that short-term supply and demand shocks from extreme weather events can have long-lasting effects on growth and public finance sustainability (Acevedo, 2014; Botzen and others, 2019; Klomp & Valckx, 2014). Additionally, research by Cevik and Jalles (2020) indicates that greater climate change vulnerability leads to lower sovereign credit ratings. They show that a one percentage point increase in climate change vulnerability results in a 0.69 percent reduction in creditworthiness for emerging market economies. Conversely, climate change resilience is positively associated with improved sovereign credit ratings, although this effect is more pronounced in advanced economies than in developing ones. In short, the implications of increased floods due to climate change could make borrowing on favorable terms more challenging for Cambodia. Today, Cambodia’s strong external position would allow the government to absorb the shock. Its current account balance would deteriorate rapidly in the scenario assessed in this paper, because of reduced manufacturing production and a substantial drop in revenues from tourism. However, if export revenues fall, the central bank could use some of its reserves to compensate for any sudden loss. External indicators suggest that Cambodia’s current reserves position is strong, covering over four years of external debt service and seven months of imports, although the requirement to hold reserves to maintain a fixed exchange rate in a largely dollarized economy should not be discounted. Total external debt, at 30 percent of GDP, is moderate and does not pose much of a risk to sustainability. But in an event where a severe flood triggers a financial crisis, additional costs could arise that might impact Cambodia’s currently sound fiscal position. Stabilizing measures to address the financial crisis, such as supporting struggling financial institutions or providing stimulus packages, could strain the government’s fiscal resources. These additional financial obligations, combined with the immediate costs of flood recovery and reconstruction, would put pressure on public budgets. In such a scenario, Cambodia’s fiscal resilience would face a complex challenge from multiple economic shocks. Incorporating disaster clauses or other flexible lending instruments into Cambodia’s financial strategy could help to address the immediate impacts of climate-related disasters and to safeguard its fiscal stability amid cascading risks (see Section 5.3). Sovereign risk presents an important channel through which the likelihood of a financial crisis could be amplified in Cambodia. The sovereign-bank nexus links the fiscal health of the sovereign with the stability of the banking system, where stress in either sector could initiate and amplify turmoil in the other. Empirical evidence underscores that the nexus has grown markedly in emerging markets in recent years (Feyen and Zuccardi, 2019). Notably, banks have increased exposure to their sovereign government, which is positively correlated with the likelihood of financial institutions defaulting. Moreover, the assets of banking systems and their credit to the private sector have grown substantially, and this growth could potentially constrain the sovereign’s ability to manage a banking crisis. Quantitative estimates reveal that banks’ exposure to the public sector is linked to a higher probability of default (ibid). The deepening sovereign-bank nexus thus poses further risk of an adverse feedback loop that could potentially jeopardize macro-financial stability in Cambodia in the case of severe flooding. While the sovereign’s role is vital, climate risks also substantially impact the financial system. In Cambodia, severe floods may threaten financial stability primarily through physical risks, which result from escalating climate-related events and long-term climate shifts, impacting firms and wealth allocation. Non-insured losses from natural disasters can threaten solvency, and insured losses can strain insurers and reinsurers, affecting capital adequacy (Bolton and others, 2020). The high private credit to GDP ratio for Cambodia suggests that a large share of the country’s economic activity relies on borrowing from the private sector. High private debt levels could be a concern if climate-related disruptions lead to economic shocks, making it more challenging for borrowers to meet their debt obligations. Increased risks of default could affect the stability of the banking sector. In summary, it is essential to consider the broader economic implications of climate change, as disruptions in economic activity can indirectly impact both the private and public sectors. While public sector borrowing may not be a major contributor to the banking system’s stability, other factors mean that resilience to climate change will be a critical factor to monitor in future. The local currency, the riel, is operating under a managed peg with the US dollar, threatening to undermine exchange rate stability during climate-related disruptions. 5. Strategies to avoid cascading failure The analysis in this paper has found that the economic fallout from a plausible worst-case scenario in Cambodia could be substantial. Table 2 summarizes the findings from the previous sections, with rough estimates of the impact and likelihood of large floods on Cambodia’s economy, health of the population, and the financial sector. If we stretch the definition of plausibility, however, it would be possible for the scale of impacts to increase. In the absence of any preventative measures, an absolute worst-case outcome could have double the magnitude of impacts shown in the table at each step. Table 2: Summary of outcomes in a worst-case plausible scenario Crisis Potential Likelihood Does Does the Likelihood GDP impact climate previous given (%) change crisis previous increase increase crisis likelihood? likelihood? Flood 10 High Yes n/a 0.1 Disease 5 Medium Yes Yes 0.2 outbreak Financial crisis 10 Low No Yes 0.5 Note: Likelihoods are illustrative (see below). Source: Authors’ estimations. Achieving resilience to climate-driven crises in Cambodia will require multi-faceted interventions. This section locates key entry points for strategies to avoid cascading failures. Shifting the mentality from response to one that is geared towards preparedness, adaptation, and multi- sectoral collaboration will be key. 5.1. Floods: Strengthen disaster preparedness Flood forecasting and early warning systems are integral to contingency plans that involve emergency measures. The lack of flood forecasting on top of weak adaptive capacity, poor infrastructure, and limited institutions exacerbates Cambodia’s vulnerability to climate variability and change. The establishment of the early warning system EWS 1294, which uses smart sensing water level gauges, is a positive development towards minimizing the effects of floods in potentially affected regions (Phy and others, 2022). Furthermore, given Cambodia’s geographical vulnerability, a multi-scaled hydro-meteorological modeling approach would be necessary to enhance forecasting capabilities, connected with transboundary initiatives both upstream and downstream. Upgrading automatic weather and hydro stations across the country could be more effective if linked to community-based preparedness plans. With advance warning, people can relocate to safer areas and take necessary precautions. Early warning systems must be linked to disaster preparedness and contingency plans. A real-time forecasting system set-up by a technical institution is a powerful tool that will help translate climate information for early warning and effective contingency planning. Early warning systems can help to ensure that hospitals are prepared in advance for surges in climate-sensitive diseases, such as dengue. Moreover, it could facilitate the capacity building in post-disaster needs assessments and recovery. Risk analysis should guide priority actions and areas. Failing to improve Cambodia’s adaptive capacity could lead to a quarter of the population, approximately 4 million people, being exposed to extreme river floods by 2040. In addition to disaster preparedness measures and early warning systems, investing in flood protection—particularly in urban areas where risks are concentrated—will be crucial to reducing future impacts. Flood-proofing critical infrastructure, such as transportation networks, hospitals, and power facilities, could help to ensure continuity of essential services during extreme events. Resilience could be strengthened by aligning infrastructure design standards with climate risk levels, particularly for critical infrastructure, and prioritizing regular maintenance of flood defenses and water storage systems. Developing a comprehensive flood management strategy that combines physical infrastructure improvements with risk-informed planning and nature-based solutions would further strengthen Cambodia’s climate resilience.4 The value of flood prevention greatly increases when cascading risks are considered, even taking mean effects. For example, the payback time of adaptation policies that focus on early flood prevention becomes shorter once knock-on effects, such as disease outbreaks and financial crisis, are accounted for (Figure 4). In the example in the figure, annual investments in adaptation starting in 2023 would begin to yield benefits exceeding the costs within four years, by 2027. This accelerated payback period shaves off a full year of potential flood damages and related impacts compared to scenarios where these effects are not accounted for. If we assign rough probabilities that a severe flood increases the likelihood of a disease outbreak by 20 percent, and both make a financial crisis 50 percent more likely (Table 2), flood prevention offers high value outcomes, providing quicker and substantial returns on investment by mitigating cascading risks. Figure 4: Hypothetical cumulated costs and benefits of flood prevention ($bn) Source: Authors’ calculations 5.2. Health: Strengthen individual and community resilience Building resilience is crucial to minimizing the health risks associated with climate change impacts on water. The potential measures fall into two categories: reducing the likelihood of disease outbreaks and ensuring that adequate treatment is available when outbreaks do occur. An 4 See the Cambodia CCDR (2023) for a more comprehensive analysis of disaster preparedness. existing focus on improving water, sanitation, and hygiene (WASH) in health centers and communities will contribute to reducing the likelihood of future disease outbreaks. Essential to these efforts is making sure that drinking water is available and safe to use in both the rainy and the dry seasons, when most health centers experience a shortage of drinking water for staff and patients. Although Cambodia has not previously had to deal with disease outbreaks from floods, the experience of COVID-19 could be used as a template for handling future disease outbreaks. Cambodia’s Hotline-115 system, launched in 2016, proved critical during the pandemic by accelerating outbreak detection and reducing data collection costs. Its success, particularly in automating contact tracing and minimizing the need for human intervention, demonstrated the country’s capacity to manage health crises effectively with limited resources, making it a valuable tool for addressing future disease outbreaks. In addition, ensuring that patients in flooded areas receive treatment in other parts of the country could dampen the reinforcing effects of disease outbreak and compromised health care. This would require additional resilience or flexibility in transport networks, expanding on the current boat services. Community education programs prior to and during floods may also be effective in reducing the impact of potential disease outbreaks. Studies suggest that educational factors, such as school attendance and adult literacy, may play a more significant role in mitigating climate-related risks like diarrheal disease than access to improved water, sanitation, and hygiene facilities (McIver and others, 2016). Furthermore, providing public health system support and involving communities in decision-making could enhance their resilience and strengthen the health care system’s ability to handle repeated extreme weather events (Saulnier, 2020). 5.3. Finance: Keep a strong external balance and explore disaster risk finance options To mitigate the risk of a financial crisis in the event of an external shock and minimize adverse effects from the sovereign-bank nexus, Cambodia will need to maintain strong reserves and careful fiscal management. Presently, Cambodia possesses sufficient foreign exchange reserves to cushion the economic blow resulting from a large flood, even with its commitment to a fixed exchange rate. Public debt levels are low, which could provide scope for offering support in a crisis without risking financial instability. Banks must also maintain strong and transparent capital and liquidity buffers. The current risk of financial contagion is therefore low and will remain so if underlying conditions do not change. Nevertheless, the space for monetary policy will remain limited because of Cambodia’s managed exchange rate and extensive dollarization. Structural risks will thus remain in the financial system. With climate-induced crises becoming more common in the coming decades, there is a strong case for exploring the role of insurance that is built into disaster-sensitive financial instruments. Mechanisms like the Southeast Asia Disaster Risk Insurance Facility (SEADRIF) could provide rapid liquidity following a disaster, helping Cambodia to respond swiftly and maintain stability in the face of the cascading crises described in this paper. The range of disaster risk finance options is growing – catastrophe bonds, parametric insurance, and risk pools could provide rapid access to funds following disasters, enabling quicker recovery. Cambodia could also explore options like the World Bank’s Catastrophe Deferred Drawdown Option (Cat DDO), insurance mechanisms, or establishing contingent credit facilities. These tools, when combined with risk layering principles, could enhance the country’s resilience by providing a comprehensive financial safety net for disaster response and recovery. Aligned with this need for increased resilience, Cambodia is taking important steps in fortifying its financial defenses against climate disasters. The newly adopted Disaster Risk Financing Strategy (DRFS) presents a critical pathway for enhancing economic resilience. The DRFS aims to strengthen Cambodia’s financial capacity to withstand climate shocks and natural disasters, with measures designed to protect the national budget, households, and firms against climate-induced disasters. Among the instruments advocated in the DRFS is the establishment of a national Reserve Fund, which could provide Cambodia with a dedicated source of funding to respond quickly to climate-related disasters, reducing the need for disruptive reallocations of the national budget during emergencies. Having such tools in place could give Cambodia a stronger backup for managing disaster risks and bouncing back more quickly when crises hit. 6. Conclusions This paper shifts the focus of climate impact analysis from prediction to risk management. It changes the question from ‘what do we think will happen?’ to ‘what might happen?’ In doing so, it developed a ‘plausible worst-case’ scenario in which a climate-induced extreme weather event leads to a secondary crisis, with potential resulting financial instability. Although such a cascade of impacts has not yet been seen in Cambodia, large floods have occurred previously, and such floods have led to outbreaks of disease in other countries. At present, financial contagion appears unlikely because of sound fiscal management within Cambodia, but risks from the sovereign-bank nexus remain and fiscal resilience could be further improved. The analysis in this paper has found that the costs to GDP could be substantial at each step in the chain. Given that Cambodia is largely powerless to stop climate change itself (and even that would make large floods less likely, not impossible), it is necessary to consider interventions further down the causal chain. Some possible options are introduced briefly in Section 5; however, it is important to consider each option as part of the wider system. For example, adaptation measures not only reduce flood damages, but they also reduce the potential for disease outbreaks and resulting financial instability. Some of the suggested measures are already development priorities (e.g., WASH and safe drinking water), and this analysis reinforces the case for early investment. The same principles could be applied to measures to improve resilience in other policy areas too. Climate change will increase risks throughout Cambodian society, including in unexpected and unpredictable ways. Shifting from a response-focused crisis mentality to one that prioritizes preparedness, adaptation, and multi-sectoral collaboration from a systems perspective will be critical in achieving these goals. Most importantly, a change in mindset is required. There will be other disaster risks from climate change (and otherwise) that were not assessed in this paper. These other risks could lead to cascade effects too, again with potential substantial impacts on socio-economic welfare. Each country in the world faces its own unique set of disaster risks and will be impacted by climate change in different ways. It seems clear, however, that climate must be integrated into existing risk management procedures. This paper has started to develop such an understanding of the key vulnerabilities for Cambodia. References Acevedo, S (2014) ‘Debt, growth and natural disasters: A Caribbean trilogy’, IMF Working Paper No. 14/125. ADB (Asian Development Bank) (2012) ‘Flood Damage Emergency Reconstruction Project: Preliminary Damage and Loss Assessment’, Flood Damage Emergency Reconstruction Project (RRP CAM 46009), see https://www.adb.org/sites/default/files/linked-documents/46009-001- cam-oth-01.pdf Ahern, M, R Kovats, P Wilkinson, R Few and F Matthies (2005) ‘Global Health Impacts of Floods: Epidemiologic Evidence’, Epidemiologic Reviews, Volume 27, Issue 1, pp. 36–46. Beauté, J and S Vong (2010) ‘Cost and disease burden of Dengue in Cambodia.’ BMC Public Health, Volume 10, Article 521. Beirne, J, N Renzhi and U Volz (2021) ‘Feeling the Heat: Climate Risks and the Cost of Sovereign Borrowing’ International Review of Economics & Finance, Volume 76, November 2021, pp 920-936. Bolton, P, M Despres, L Pereira Da Silva, F Samama and R Svartzman (2020) ‘The Green Swan: Central banking and financial stability in the age of climate change’, BIS and Banque de France, January. Botzen, W, O Deschenes and M Sanders (2019) ‘The economic impacts of natural disasters: A review of models and empirical studies’, Review of Environmental Economics and Policy, Volume 13, Number 2, pp 167-188. Burke, M, SM Hsiang and E Miguel (2015) ‘Global non-linear effect of temperature on economic production’, Nature, Volume 527, pp 235-239. Cevik, S and JT Jalles (2020) ‘Feeling the Heat: Climate Shocks and Credit Ratings’, IMF Working Paper No. 2020/286. Choi et al (2016) ‘Effects of weather factors on dengue fever incidence and implications for interventions in Cambodia’, Volume 16, Article 241. Davies, GI, L McIver, Y Kim, M Hashizume, S Iddings and V Chan (2014) ‘Water-borne diseases and extreme weather events in Cambodia: review of impacts and implications of climate change’, International Journal of Environmental Research and Public Health, Volume 12, Issue 1, pp 191-213. de Man, H, L Mughini-Gras, B Schimmer, IHMFriesema, AM de Roda Husman and W van Pelt (2014) ‘Quantitative assessment of infection risk from exposure to waterborne pathogens in urban floodwater.’ Water Research, Volume 48, pp 90–99. Dembo, RS (2021) ‘Risk Taking… in an uncertain world’, Archway Publishing: Bloomington, IN. Feyen, Erik HB, H Zuccardi and I Esteban (2019) ‘The Sovereign-Bank Nexus in EMDEs: What is it, is it Rising, and What are the Policy Implications?’ World Bank Policy Research Working Paper Number 8950, see https://ssrn.com/abstract=3430565 Hallegatte, S (2014) ‘Modeling the Role of Inventories and Heterogeneity in the Assessment of the Economic Costs of Natural Disasters’, Risk Analysis, Volume 34, Issue 1, pp 152-167. IMF (1999) ‘Moldova: Recent Economic Developments’, see https://www.elibrary.imf.org/view/journals/002/1999/110/002.1999.issue-110- en.xml?Tabs=citedby-102778 IPCC (2020) ‘The concept of risk in the IPCC Sixth Assessment Report: a summary of cross- Working Group discussions’, IPCC: Geneva, Switzerland. See https://www.ipcc.ch/site/assets/uploads/2021/02/Risk-guidance-FINAL_15Feb2021.pdf Kemp, L, C Xu, J Depledge, KL Ebi, G Gibbins, TA Kohler, J Rockström, M Scheffer, HJ Schellnhuber, W Steffen and TM Lenton (2022) ‘Climate Endgame: Exploring catastrophic climate change scenarios;, PNAS, Volume 119, Issue 34, article e 2108146119. Keynes, JM (1921) ‘A Treatise on Probability’, Macmillan: London, UK. Klomp, J and K Valckx (2014) ‘Natural disasters and economic growth: A meta-analysis’, Global Environmental Change, Volume 26, pp 183-195. Knight, FH (1921) ‘Risk Uncertainty and Profit’, Hart, Schaffner & Marx: Boston, MA. Laeven, LA and FV Valencia (2020) ‘Systemic Banking Crises Database: A Timely Update in Covid-19 Times.’ CEPR Discussion Paper Number DP14569, see https://ssrn.com/abstract=3594190 Mallucci, E (2020) ‘Natural Disasters, Climate Change, and Sovereign Risk,’ International Finance Discussion Papers 1291. Washington: Board of Governors of the Federal Reserve System, see https://doi.org/10.17016/IFDP.2020.1291 Mayer, CJ, KM Pence and SH Sherlund (2008) ‘The Rise in Mortgage Defaults’, Finance and Economics Discussion Series, 2008-59, Divisions of Research & Statistics and Monetary Affairs, Federal Reserve Board, Washington, D.C, see https://www.federalreserve.gov/pubs/feds/2008/200859/200859pap.pdf McIver, LJ, VS Chan, KJ Bowen, SN Iddings, K Hero and PP Raingsey (2016) ‘Review of Climate Change and Water-Related Diseases in Cambodia and Findings From Stakeholder Knowledge Assessments’, Asia Pacific Journal of Public Health, Volume 28, Issue 2, pp 49S- 58S. National Council for Sustainable Development (NCSD) (2019) ‘Addressing climate change impacts on economic growth in Cambodia (CEGIM)’. See https://ncsd.moe.gov.kh/resources/document/addressing-climate-change-impacts- economic- growth-0 Network for Greening the Financial System (NGFS) (2018) ‘First Progress Report, Network for Greening the Financial System’, NGFS:Paris, France. Paltan, H, M Allen, K Haustein, L Fuldauer and S Dadson (2018) ‘Global implications of 1.5 °C and 2 °C warmer worlds on extreme river flows’, Environmental Research Letters, Volume 13, Article 094003. Pescaroli, G and D Alexander (2018) ‘Understanding Compound, Interconnected, Interacting, and Cascading Risks: A Holistic Framework’, Risk Analysis, Volume 38, Number 11, pp2245- 2257. Phy, SR, T Sok, S Try, R Chan, S Uk, C Hen and C Oeurng (2022) ‘Flood Hazard and Management in Cambodia: A Review of Activities, Knowledge Gaps, and Research Direction’, Climate, Volume 10, Issue 11, Article 162. Pinzón, A, N Robins, M McLuckie and G Thoumi (2020) ‘The Sovereign Transition to Sustainability: Understanding the Dependence of Sovereign Debt on Nature’, Grantham Research Institute on Climate Change and the Environment, London School of Economics and Political Science, and Planet Tracker, London. Saulnier, DD, C Hanson, P Ir, HM Alvesson and J von Schreeb (2018) ‘The Effect of Seasonal Floods on Health: Analysis of Six Years of National Health Data and Flood Maps’, International Journal of Environmental Research and Public Health, Volume 15, Issue 4, Article 665. Schuler, P, LE Oliveira, G Mele and M Antonio (2019) ‘Managing the Fiscal Risks Associated with Natural Disasters’, in Pigato, MA (eds) ‘Fiscal Policies for Development and Climate Action’, World Bank: Washington, DC, pp 133-153. Sturzenegger, F and J Zettelmeyer (2007) ‘Debt Defaults and Lessons from a Decade of Crises’, MIT Press: Cambridge, MA. Tanoue, M, R Taguchi, S Nakata, S Watanabe, S Fujimori and Y Hirabayashi (2020) ‘Estimation of Direct and Indirect Economic Losses Caused by a Flood With Long-Lasting Inundation: Application to the 2011 Thailand Flood’, Water Resources Research, Volume 56, Issue 5, e2019WR026092. van Dijck, P, G Dijkstra, N de Jong, D Martin and R Vos (2000) ‘The Suriname economy: experiences of the 1990s and challenges ahead’, see https://www.cedla.nl/_files/ugd/52820e_fc4862596945459193e5eb671857bf50.pdf?index=true World Bank Group (2023) ‘Country Climate and Development Report (CCDR): Cambodia’, World Bank Group: Washington, DC, see http://documents1.worldbank.org/curated/en/099092823045083987/pdf/P17887106c6c2d0e909 aa1090f3e10505c1.pdf World Bank Group (2021) ‘Climate Risk Country Profile: Cambodia’, World Bank Group: Washington, DC, see https://climateknowledgeportal.worldbank.org/sites/default/files/2021- 08/15849-WB_Cambodia%20Country%20Profile-WEB.pdf WHO (2019) ‘Preparedness crucial to protect health and well being of Cambodians affected by floods’, see https://www.who.int/cambodia/news/detail/16-10-2019-preparedness-crucial-to- protect-health-and-well-being-of-cambodians-affected-by-floods Zscheischler, J, S Westra, BJJM van Den Hurk, SI Seneviratne, PJ Ward, A Pitman, A Aghakouchak, DN Bresch, M Leonard, T Wahl and X Zhang (2018) ‘Future climate risk from compound events.’ Nature Climate Change, Volume 8, Issue 6, pp 469-477.