Chapter 1: Introduction: The Financial System and the Green Transition

In the spring of 2011, I stood in a computer-controlled greenhouse in Wuliming Town, Zhaodong City, Heilongjiang Province, watching a corn crop that would eventually yield one ton per mu. Soil-moisture sensors timed the irrigation. Leaf-color analysis metered the fertilizer. A single technician monitored four hundred hectares from a screen. What held my attention was not the technology but the financing behind it. The equipment had been bought on credit. The land had been consolidated through a lease backed by a rural credit cooperative. The whole operation carried an insurance policy whose payout was tied not to measured damage but to a rainfall index. Three financial instruments, each doing a different job, had been stitched together to make one farm viable. None of them alone would have been enough.

That morning has stayed with me through fifteen years of research on credit rating, weather insurance, and the financing of technology firms. It explains the shape of this book. Industrial green transition is not a single financing problem. It is at least three problems wearing the same coat. A steel mill that wants to install hydrogen-based direct reduction needs someone to judge whether its environmental plan is credible. That is a rating problem. A forestry company planting a thirty-year rotation needs someone to absorb the risk that drought or fire wipes out a decade of growth. That is an insurance problem. A start-up with a promising electrolyzer design needs capital before it has revenue, collateral, or a credit history. That is a catalysis problem. Credit alone cannot solve all three, and neither can insurance, and neither can venture capital. My thesis is that they work only as an integrated mechanism, and that the financial system's failure to integrate them is a principal reason green investment remains slower than the physical urgency demands.

1.1 The Green Transition Imperative

1.1.1 What Decarbonization Asks of Industry

Start with the scale. The International Energy Agency's net-zero scenario requires annual clean-energy investment to reach roughly USD 4 trillion by 2030, up from about USD 1.8 trillion in recent years. That gap is a structural shortfall in capital allocation. Industry accounts for roughly a quarter of global CO₂ emissions, and the abatement technologies for steel, cement, chemicals, and aluminum are capital-intensive, commercially unproven at scale, and long-lived. A blast furnace relined today will still be operating in 2045. A cement kiln built without carbon capture locks in emissions for forty years. When investment decisions get made matters as much as how much gets invested.

Conventional project finance handles this badly. Green industrial projects carry high upfront capital costs and uncertain operating savings. The savings depend on carbon prices that do not yet exist at the required level, on technology learning curves that are guessed rather than observed, and on policy regimes that can change with elections. A lender evaluating a green hydrogen plant is therefore evaluating a chain of conditional probabilities, most of which have no historical frequency to draw on. There is no twenty-year track record of default rates for green hydrogen projects, because there were no green hydrogen projects twenty years ago.

The financial system's standard tools were built for a world where the past predicts the future. Credit ratings are calibrated on historical default experience. Insurance premiums are calibrated on historical loss experience. Venture capital tolerates uncertainty but prices it through diversification across many bets, most of which fail. None of these tools was designed for a transition in which the relevant uncertainty is structural rather than cyclical: a one-time reallocation of the entire capital stock toward technologies whose risk profiles are still being discovered.

1.1.2 Why Markets Underprice Transition Risk

The standard explanation for underinvestment in green assets is externalities. Carbon is underpriced, so green projects appear less profitable than they are. True enough, and carbon pricing is necessary. But it is not the whole story. Even where carbon prices exist—the EU Emissions Trading System has traded above EUR 80 per tonne for extended periods—capital does not flow at the required rate. Something else is happening inside the financial system itself.

Three frictions stand out. Information comes first. A bank's loan officer can read a balance sheet. Reading an environmental performance record is harder. Is a company's reported 30 percent reduction in emissions intensity real, or is it an artifact of outsourcing the dirtiest production step to a supplier? Without standardized, auditable environmental data, green claims are cheap to make and expensive to verify. When verification costs exceed the value of the information, lenders rationally discount all claims, including true ones. The market for green credit then suffers the lemons problem: good green borrowers are pooled with pretenders, and the pool clears at a discount that drives out the best borrowers first.

Risk follows. Green projects are exposed to transition risk—the risk that policy, technology, or consumer preferences shift in ways that strand assets. A gas-fired power plant built as a "bridge" fuel may become a stranded asset if storage costs fall faster than expected. A bank holding such exposure faces a loss correlated across its entire portfolio, because a single policy decision (a carbon tax, a subsidy removal) affects all its carbon-intensive borrowers at once. Diversification cannot eliminate this. It is a systematic factor, and standard credit risk models that treat default correlation as a function of business cycles do not capture it well. The European Systemic Risk Board's 2016 report Too Late, Too Sudden warned that a late and abrupt transition could produce exactly this kind of correlated loss.

Maturity mismatch compounds both. Green infrastructure has long payback periods. A forest takes twenty-five to forty years to mature. A transmission grid takes a decade to permit and build. Bank deposits have maturities measured in months. Maturity mismatch is inherent to banking, but it becomes acute when the assets being financed have both long lives and uncertain terminal values. Insurance can bridge part of this gap by transferring tail risk to a party with a longer horizon, but insurance markets for green assets are thin.

1.1.3 The Policy Response So Far

Regulators have not been idle. The Network of Central Banks and Supervisors for Greening the Financial System, founded in 2017, now includes more than 130 central banks and supervisors. The European Central Bank published its Guide on Climate-Related and Environmental Risks in November 2020, setting supervisory expectations for risk management and disclosure. The European Banking Authority followed with a report on management and supervision of ESG risks in June 2021. The Task Force on Climate-related Financial Disclosures produced a framework now widely adopted, and the EU's Corporate Sustainability Reporting Directive and Taxonomy Regulation have begun to standardize what companies must disclose and what counts as sustainable.

These are necessary steps. Disclosure is the precondition for pricing. But disclosure alone does not create the institutional capacity to act on the disclosed information. A bank that knows its borrower's emissions profile still needs a method for incorporating that profile into a credit decision. An insurer that knows a region's drought history still needs a product that pays out when the drought occurs. A venture fund that knows a technology's potential still needs a mechanism for exiting the investment. The regulatory agenda has focused on transparency. The harder work is building the instruments that turn transparency into capital allocation.

Here the three-pillar framework enters. The pillars do not substitute for carbon pricing or disclosure. They are the transmission mechanism that converts policy signals into financial decisions.

1.2 Three Pillars: Rating, Insurance, Catalysis

1.2.1 The Logic of Integration

Consider what happens when the three functions are separated. A bank lends to a green project on the strength of a rating that does not account for environmental performance. The rating is accurate as far as it goes, but it systematically misprices transition risk: a borrower with high emissions faces future carbon costs that the rating ignores, so the loan is underpriced relative to its true risk. The bank discovers this when the carbon price rises, and it responds by withdrawing from the sector entirely—a cliff effect more damaging than a gradual repricing would have been.

Now add insurance, but keep it separate from credit. The insurer offers a weather index policy to an agricultural borrower. The policy pays out when rainfall falls below a threshold. The bank does not know about the policy, or does not adjust its lending terms to reflect it. The borrower's default probability is genuinely lower because the insurance stabilizes income, but the bank's pricing does not reflect this. The borrower pays a credit spread that overstates its risk, and the insurance premium is a cost without a financing benefit. Each instrument is sound in isolation; jointly they are inefficient.

Add catalysis, but keep it separate from both. A government-guided venture fund invests in a green technology start-up. The fund's money is patient and tolerant of risk. But the start-up still cannot obtain bank credit for working capital because it has no collateral and no rating history. The venture fund's equity is diluted by the high cost of the debt it can obtain, or the start-up simply cannot scale. Catalysis worked at the seed stage and failed at the growth stage.

The integration logic is straightforward once stated: each pillar addresses a different failure, and the failures interact. Rating without insurance misprices correlated tail risk. Insurance without rating fails to transmit its risk-reduction benefit to the cost of capital. Catalysis without either cannot bridge the gap between early-stage equity and bankable debt. The integrated mechanism is designed so that the output of one pillar becomes an input to the next.

flowchart TD
    A[Green industrial project] --> B[Pillar 1: Rating]
    B -->|Environmental performance scored| C[Adjusted credit assessment]
    C --> D[Pillar 2: Insurance]
    D -->|Weather / forest / transition risk transferred| E[Stabilized cash flow]
    E --> F[Lower default probability]
    F --> G[Pillar 3: Sci-tech finance]
    G -->|Patient capital and guarantees| H[Scaled, bankable asset]
    H --> I[Green transition financed]
    C -->|Rating signal| G
    F -->|Risk-adjusted return| G
    G -->|Guarantee and co-investment| C

The diagram shows the intended flow, but it also shows the feedback loops that make the mechanism more than the sum of its parts. The rating signal informs the catalysis decision: a technology firm with a strong environmental score is a better candidate for policy-guided venture capital. The risk-adjusted return, once insurance has stabilized cash flow, informs the rating: a borrower whose income is insured against drought has a lower probability of default, which should be reflected in its score. And the guarantee provided by the catalysis pillar can improve the rating of a project that would otherwise be too risky for a commercial bank to consider.

1.2.2 Pillar One: Rating as Identification

The first pillar identifies green performance. A credit rating is, at bottom, a forecast of default probability. Traditional ratings incorporate financial ratios, industry position, management quality, and macroeconomic conditions. They do not, as a rule, incorporate environmental performance, because the historical data linking environmental performance to default has been sparse and the causal channels poorly specified.

That is changing. There are now plausible mechanisms by which environmental performance affects credit risk. A borrower with high emissions faces direct costs from carbon pricing, indirect costs from energy price volatility, and regulatory costs from tightening standards. A borrower with strong environmental performance may face lower transition risk, better access to green-labeled capital, and lower insurance costs. These mechanisms are measurable, and the data to measure them is improving. The task of the first pillar is to build scoring models that translate environmental performance into a rating adjustment, and to validate those models against observed default experience as it accumulates.

This is not a matter of adding an ESG score to a credit score and averaging them. The two are not commensurable. Environmental performance affects default probability through specific channels—cost, revenue, asset value, litigation—and the rating model must specify those channels. A steel company's emissions intensity affects its cost structure through the carbon price. A real estate company's energy efficiency affects its asset value through tenant demand and building regulations. A utility's generation mix affects its revenue through the merit order. The modeling work is sector-specific, and the book treats it that way.

1.2.3 Pillar Two: Insurance as Risk Buffering

The second pillar is the transfer of risk that credit markets cannot absorb. Insurance does something lending cannot: it converts a variable cash flow into a more stable one. A farmer whose income depends on rainfall faces a distribution of outcomes, some of them catastrophic. A lender can diversify across many farmers, but if the drought is regional, diversification fails—all the farmers default together. An insurer that writes policies across regions can diversify the risk, provided the regions' weather is not perfectly correlated. The insurer's capacity to absorb the risk depends on its own access to reinsurance and capital markets, which is why the insurance pillar is ultimately connected to the broader financial system.

The book focuses on two insurance instruments: weather index insurance and forest insurance. Weather index insurance pays out on a meteorological trigger—rainfall below a threshold, temperature above a threshold—rather than on verified loss. This solves the moral hazard and verification cost problems that plague traditional crop insurance, at the cost of basis risk: the policy may not pay out when the farmer actually suffers a loss, because the index and the farm's experience diverge. Forest insurance covers fire, pest, and disease risk for timber assets, which is essential for the long-rotation investments that carbon sequestration requires. Both instruments illustrate a general principle: insurance can make a risky green asset bankable by truncating the left tail of its return distribution.

The empirical evidence in this book, drawn from agricultural insurance in China, shows that insurance affects the level of farmer income as well as its variance. Insured farmers invest more, because they are less exposed to catastrophic loss. Income stabilization is therefore a productivity effect, not only a welfare transfer. The same logic applies to green industrial investment: an insured forestry company can afford to plant more trees, because it is not self-insuring against fire by holding excess capital.

1.2.4 Pillar Three: Sci-Tech Finance as Catalysis

The third pillar is the provision of capital to green technology firms that cannot yet access conventional credit. The term "sci-tech finance" (科技金融) refers to the set of institutions and instruments designed to finance technological innovation: venture capital, government-guided funds, intellectual property pledges, science and technology banks, and guarantee schemes. The book's prior research on credit rating for technology enterprises provides the foundation for this pillar. Technology firms have unusual risk profiles: their assets are intangible, their revenue is uncertain, and their survival depends on continued innovation. Standard credit models handle them poorly.

The catalysis function is to provide capital at the stage where the risk is highest and the collateral is thinnest, and to do so in a way that prepares the firm for later-stage conventional financing. Government-guided venture capital is one instrument. It can absorb early-stage risk that private capital will not, and it can signal quality to later investors. It can also crowd out private capital, distort incentives, and allocate capital politically rather than commercially. The design of the incentive structure matters, and the book treats it in detail in later chapters.

The connection between catalysis and the other two pillars is the key to the integrated mechanism. A green technology firm that receives a strong environmental rating is a better candidate for catalysis. A firm whose revenue risk is buffered by insurance is a better candidate for bank lending after the catalysis stage. Rating identifies green performance, while insurance buffers risk and catalysis accelerates deployment. The pipeline is only as strong as its weakest connection.

1.2.5 What the Integrated Mechanism Is Not

It is worth being clear about what the framework does not claim. It does not claim that the three pillars can substitute for a carbon price. Without a price signal, green projects remain less profitable than brown ones, and no amount of financial engineering can change that. It does not claim that the pillars can eliminate the need for public investment in early-stage research, which has public-good characteristics that private capital will not fund. And it does not claim that the integrated mechanism is sufficient on its own; it is a necessary complement to fiscal policy, industrial policy, and international coordination.

What it does claim is that the financial system's current fragmentation is a binding constraint. Even with a carbon price, even with public research funding, the transition will be slower than necessary if rating, insurance, and catalysis operate in separate silos. Without integration, capital flows at half the required rate.

1.3 Structure and Contribution of the Book

1.3.1 The Roadmap

The book is organized in four parts, corresponding to the three pillars and their integration.

Part I, comprising Chapters 2 through 5, develops the rating pillar. Chapter 2 lays out the theoretical foundation: information asymmetry, credit rationing, and the rationale for integrated finance. It draws on the classic literature on credit rationing—Stiglitz and Weiss (1981) showed that adverse selection and moral hazard can cause banks to ration credit rather than raise rates—and extends it to the green context. Chapter 3 examines how environmental performance can be incorporated into credit rating, reviewing existing ESG rating methodologies and their limitations. Chapter 4 develops the modeling framework: scoring models, default probability estimation, and loss given default for green credit. Chapter 5 addresses the portfolio level, where dependence and concentration create risks that single-borrower models miss.

Part II, Chapters 6 through 8, develops the insurance pillar. Chapter 6 introduces weather index insurance and forest insurance, explaining their design and the basis risk problem. Chapter 7 presents empirical evidence on agricultural insurance and farmer income stability, drawing on field research in China. Chapter 8 connects insurance to credit, showing how insurance can mitigate default risk in eco-agriculture and, by extension, in other green sectors.

Part III, Chapters 9 through 11, develops the catalysis pillar. Chapter 9 examines sci-tech finance and its role in green innovation. Chapter 10 analyzes government-guided venture capital and the green premium—the question of whether green firms can earn a return premium that compensates for their risk. Chapter 11 addresses performance evaluation and incentive design for policy-guided green finance, a topic often neglected but decisive for whether the catalysis actually works.

Part IV, Chapters 12 through 15, addresses the integration. Chapter 12 examines the digital infrastructure and data systems that the integrated mechanism requires. Chapter 13 addresses international cooperation, technology transfer, and climate finance, which are essential for the transition in developing economies. Chapter 14 lays out the policy sequencing for the integrated mechanism, addressing the question of what to do first. Chapter 15 concludes with the vision of a resilient and inclusive green financial system.

1.3.2 The Intellectual Foundations

The book draws on three strands of my prior research, each corresponding to a pillar.

The first is credit rating for technology enterprises. In earlier work, I developed scoring models for technology firms that incorporated intangible assets, innovation capacity, and growth potential into credit assessment. The models were calibrated on Chinese technology firms, where the problem of lending to asset-light companies is particularly acute. The green rating framework in this book extends that work by adding environmental performance as a dimension of assessment, and by specifying the channels through which environmental performance affects default probability.

The second is weather index insurance. My research on weather index insurance in China examined the design of index products, the magnitude of basis risk, and the determinants of farmer participation. The key finding, which recurs in Chapter 7, is that insurance affects investment behavior, not just consumption smoothing. Insured farmers invest more in productivity-enhancing inputs, because they are less exposed to catastrophic loss. This finding is the empirical basis for the claim that insurance can make green assets bankable.

The third is the impact of agricultural insurance on farmer income. The empirical work in Chapter 7 uses household survey data to estimate the effect of insurance participation on income levels and income variance. The results show that insurance reduces income variance and, in some specifications, raises mean income. The mechanism is investment: insurance allows farmers to take on more productive but riskier activities. The same mechanism applies to green industrial investment, where insurance allows firms to take on the risk of new technologies.

The book's contribution is to integrate these three strands into a single framework. The integration is not merely additive. The three pillars interact in ways that neither the rating literature nor the insurance literature nor the sci-tech finance literature has fully explored. The book's central argument—that the interaction effects are large enough to justify treating the three as a single mechanism—is developed through the modeling in Parts I and II and the institutional analysis in Parts III and IV.

1.3.3 Who This Book Is For

The book is written for three audiences. The first is academic researchers in finance, economics, and environmental policy who are interested in the microfoundations of green finance. The modeling in Chapters 4 and 5 is technical, but the book is not a textbook; it develops the models necessary to make the argument, and it refers the reader to the underlying literature for the full derivations.

The second audience is practitioners: bankers, insurers, and investors who are building green finance products. The book's treatment of rating methodologies, insurance design, and incentive structures is intended to be directly useful. The case material, drawn from China and Europe, is chosen to illustrate design choices rather than to provide a comprehensive survey.

The third audience is policymakers and regulators who are designing the institutional framework for green finance. The book's argument that the three pillars must be integrated has direct implications for policy: disclosure rules, capital requirements, and guarantee schemes should be designed with the integration in mind, not pillar by pillar. Chapter 14 develops these implications in detail.

1.3.4 A Note on Method and Geography

The book's empirical base is primarily Chinese, with European and international comparisons where relevant. This is a deliberate choice. China is the world's largest emitter and the world's largest source of green investment. Its financial system is bank-dominated, which makes the credit rating pillar particularly important. Its agricultural sector is exposed to climate risk, which makes the insurance pillar particularly relevant. And its government has made green finance a policy priority, which makes the catalysis pillar particularly active.

But the book is not a study of China alone. The three-pillar framework is general, and the chapters draw on European and American experience where it illuminates the argument. The EU's sustainable finance action plan, the ECB's supervisory guidance, and the US Commodity Futures Trading Commission's 2020 report on climate risk all inform the analysis. The framework is intended to apply across financial systems, with the caveat that the relative importance of the pillars varies with the structure of the financial system. In a market-based system like the United States, catalysis through venture capital is more developed and rating through public bond markets is more important. In a bank-based system like China's, rating through bank credit assessment and insurance through agricultural schemes are more central.

1.3.5 The Argument in Brief

The green transition requires a reallocation of capital at a scale and speed the financial system has not yet achieved. The shortfall is not primarily due to a lack of savings; global savings are ample. It is due to a failure in the intermediation mechanism—the institutions and instruments that convert savings into investment.

That failure has three distinct sources: information asymmetry about environmental performance, unhedged transition and physical risk, and the inability of conventional credit to finance early-stage innovation. These correspond to the three pillars.

The three sources interact. Rating without insurance misprices correlated risk. Insurance without rating fails to transmit its benefit to the cost of capital. Catalysis without either cannot bridge the gap between early-stage equity and bankable debt. Integration is therefore a matter of necessity rather than convenience.

The integrated mechanism requires specific institutional designs: scoring models that specify the channels through which environmental performance affects default, insurance products that address basis risk and connect to credit terms, and catalysis instruments with incentive structures that avoid crowding out private capital. The book develops these designs and evaluates them against the available evidence.

The stakes are concrete. A steel mill in Hebei that relines a blast furnace today will emit for another twenty years. A forestry company in Yunnan that cannot insure against fire will underplant. A start-up with a better electrolyzer will relocate to a jurisdiction with more patient capital. These are not abstractions; they are decisions being made now, with the financial system's current tools. The book is an argument that the tools can be better, and a specification of how.

1.3.6 The Plan of the Chapter

The remainder of this chapter previews the three pillars in more detail, then describes the book's structure and contribution. Section 1 has laid out the imperative: why industrial decarbonization needs financial system reform, and why the reform must address the three frictions of information, risk, and maturity. Section 2 has introduced the three pillars and their integration logic, and has been explicit about what the framework does not claim. Section 3, which follows, provides the chapter roadmap and states the key arguments of each chapter. The chapter closes with a brief note on the evidence base and the limits of the analysis.

1.4 The Three Pillars in Greater Detail

1.4.1 Rating: From Financial Ratios to Environmental Channels

The traditional credit rating model is a weighted combination of financial and non-financial factors. The financial factors—leverage, coverage, liquidity, profitability—are well understood and have been validated against decades of default data. The non-financial factors—industry position, management quality, corporate governance—are harder to quantify but are incorporated through analyst judgment and industry-specific scorecards.

Adding environmental performance to this framework is not simply a matter of adding a fourth category. The question is how environmental performance affects the financial factors. A company's emissions intensity affects its cost structure through energy prices and carbon costs. Its environmental liabilities affect its leverage through contingent claims. Its environmental reputation affects its revenue through customer preferences and regulatory risk. The rating model must specify these channels, estimate their magnitudes, and validate them against default experience.

Chapter 3 reviews the existing ESG rating methodologies and finds them wanting for credit purposes. Most ESG scores are designed for equity investors, who care about long-term value creation, rather than for creditors, who care about default probability over a one-to-five-year horizon. The two are related but not identical. A company with a strong long-term environmental strategy may still default next year because of a liquidity crisis. A company with a weak environmental strategy may survive for years because its transition costs are deferred. The credit rating model must be calibrated to the credit horizon, not the equity horizon.

Chapter 4 develops the modeling framework. The scoring model is a logistic regression of default on financial and environmental variables, with the environmental variables chosen to capture the specific channels identified in Chapter 3. The default probability is then combined with loss given default to produce an expected loss, which is the basis for pricing and capital allocation. The model is calibrated on a panel of Chinese firms, with the caveat that the data on environmental performance is incomplete and the default experience is limited. The calibration is therefore provisional, and the chapter is explicit about the uncertainty.

Chapter 5 addresses the portfolio level. The single-borrower model assumes that defaults are independent conditional on the observed variables. In reality, defaults are correlated, and the correlation is higher for green borrowers than for the general population, because transition risk is a systematic factor. A carbon price shock affects all high-emission borrowers at once. A technology breakthrough affects all firms in the affected sector at once. The portfolio model must capture this dependence, using factor models or copulas, and must account for concentration risk—the risk that a single large green project dominates the portfolio. The chapter shows that ignoring dependence leads to a systematic underestimation of portfolio risk, which is precisely the error that the European Systemic Risk Board warned about in 2016.

1.4.2 Insurance: From Indemnity to Index

Traditional insurance indemnifies verified loss. The insurer sends an adjuster, verifies the damage, and pays the claim. This works for fire and theft, where the loss is observable and the moral hazard is manageable. It works poorly for agricultural losses, where the loss is dispersed, the verification is costly, and the moral hazard is severe—a farmer who is fully insured against crop loss has little incentive to tend the crop.

Weather index insurance solves these problems by paying on a trigger that is observable and exogenous. If rainfall at the local weather station falls below 50 millimeters during the growing season, the policy pays a fixed amount per hectare. The farmer's actual loss is not verified. The moral hazard is eliminated, because the farmer cannot influence the rainfall. The verification cost is eliminated, because the weather station data is public. The remaining problem is basis risk: the farmer may suffer a loss even if the index does not trigger, or may receive a payout even if the loss is small.

Chapter 6 explains the design of index products and the magnitude of basis risk. The basis risk depends on the correlation between the index and the farmer's yield, which varies with the distance from the weather station, the topography, and the farming practices. The chapter reviews the evidence on basis risk in China and finds that it is substantial but manageable, particularly when the index is designed at a fine spatial scale and when the product is bundled with other risk management tools.

Forest insurance is a different product with a different risk profile. The risk is fire, pest, and disease, and the loss is the value of the timber, which accumulates over decades. The insurance is essential for long-rotation investments, because a single fire can destroy the entire investment. The chapter reviews the design of forest insurance in China and Europe and finds that the coverage is often inadequate, because the timber value is hard to assess and the risk is correlated across large areas.

Chapter 7 presents the empirical evidence on agricultural insurance and farmer income. The data come from a household survey in Heilongjiang and Henan provinces, covering 2,400 households over five years. The results show that insurance participation reduces income variance by 15 to 25 percent and raises mean income by 5 to 8 percent. The mechanism is investment: insured farmers spend more on fertilizer, seeds, and machinery, because they are less exposed to catastrophic loss. The finding is robust to a variety of specifications, including instrumental variable estimates that address the endogeneity of insurance participation.

Chapter 8 connects insurance to credit. The argument is that insurance reduces default probability by stabilizing cash flow, and that this reduction should be reflected in credit terms. The chapter develops a model in which the lender observes the insurance policy and adjusts the interest rate accordingly. The model shows that the adjustment can be substantial—a 50 to 100 basis point reduction in the spread for insured borrowers—and that the reduction is larger when the insurance covers the tail risk that the lender cannot diversify. The chapter then examines the obstacles to linking insurance and credit in practice: the lender may not trust the insurer, the insurance contract may not be assignable, and the regulatory capital treatment of insured loans may not recognize the risk reduction. These obstacles are institutional, not technical, and they are the subject of Chapter 14's policy recommendations.

1.4.3 Catalysis: From Venture Capital to Guarantee

The catalysis pillar is the least standardized of the three. Venture capital is a well-understood instrument, but its application to green technology is complicated by the long time horizons, the capital intensity, and the policy dependence of green projects. Government-guided venture capital adds another layer of complexity, because the government's objectives are not purely commercial.

Chapter 9 reviews the landscape of sci-tech finance in China and compares it with the US and European models. The Chinese model is characterized by government guidance funds, science and technology banks, and intellectual property pledge financing. The US model is characterized by private venture capital, university technology transfer, and the Small Business Innovation Research program. The European model is characterized by a mix of public and private funds, with a strong emphasis on the European Investment Bank and the European Investment Fund.

Chapter 10 examines government-guided venture capital and the green premium. The green premium is the excess return that green investments must earn to compensate for their risk. If the premium is positive, private capital should flow to green investments without government intervention. If the premium is negative or zero, government intervention may be necessary. The chapter reviews the evidence on the green premium and finds that it is positive for some green technologies and negative for others, depending on the maturity of the technology and the policy environment. The chapter then examines the design of government-guided funds and finds that their performance depends critically on the incentive structure: funds that are evaluated on financial returns perform better than funds that are evaluated on political criteria.

Chapter 11 addresses performance evaluation and incentive design. The evaluation of policy-guided green finance is complicated by the fact that the objectives are multiple: financial return, environmental impact, and innovation. The chapter develops a multi-objective evaluation framework and applies it to a sample of Chinese government guidance funds. The results show that funds with clear environmental mandates and independent management teams perform better on both financial and environmental metrics than funds with vague mandates and political management. The chapter concludes with recommendations for incentive design: performance-based compensation, independent evaluation, and transparency in reporting.

1.5 The Structure of the Book

1.5.1 Part I: Rating (Chapters 2–5)

Chapter 2 develops the theoretical foundation. It begins with the classic problem of information asymmetry in credit markets and reviews the Stiglitz-Weiss model of credit rationing. It then extends the model to the green context, showing that environmental information asymmetry creates an additional source of rationing: lenders cannot distinguish green borrowers from brown ones, so they ration credit to both. The chapter argues that the solution requires more than disclosure; it requires a rating technology that converts environmental information into a credit-relevant signal.

Chapter 3 examines the existing ESG rating methodologies and their limitations for credit purposes. It reviews the major rating agencies' approaches—MSCI, Sustainalytics, S&P Global—and finds that they are designed for equity investors, not creditors. The chapter then develops a credit-specific framework that specifies the channels through which environmental performance affects default probability: cost, revenue, asset value, and litigation. The framework is applied to four sectors: steel, cement, power, and agriculture.

Chapter 4 develops the modeling framework. It presents the scoring model, the default probability model, and the loss given default model. The scoring model is a logistic regression with financial and environmental variables. The default probability model uses the scoring model's output as an input to a hazard rate model. The loss given default model uses collateral values, seniority, and recovery rates. The chapter includes a worked example of a green steel project and shows how the model produces a rating that differs from the conventional rating.

Chapter 5 addresses the portfolio level. It presents factor models and copula models for default dependence, with the environmental factor as an additional systematic factor. The chapter shows that ignoring the environmental factor leads to an underestimation of portfolio risk, particularly in scenarios with a sharp carbon price increase. It then examines concentration risk and shows that a portfolio with a large green project can be more risky than a diversified portfolio, even if the green project has a lower standalone default probability.

1.5.2 Part II: Insurance (Chapters 6–8)

Chapter 6 introduces weather index insurance and forest insurance. It explains the design of index products, the basis risk problem, and the evidence on basis risk in China. It then examines forest insurance, with a focus on the long-rotation problem and the role of insurance in enabling carbon sequestration. The chapter concludes with a discussion of the conditions under which insurance can make a green asset bankable.

Chapter 7 presents the empirical evidence on agricultural insurance and farmer income. It describes the survey data, the identification strategy, and the results. The chapter finds that insurance reduces income variance and raises mean income, with the mechanism being increased investment. It then examines the heterogeneity of the effect: the income effect is larger for smaller farmers and for farmers in more risky regions.

Chapter 8 connects insurance to credit. It develops a model of the lender's decision with and without insurance, and shows that insurance reduces the optimal interest rate and increases the optimal loan size. The chapter then examines the obstacles to linking insurance and credit in practice and proposes institutional solutions: insurance-linked loans, credit guarantees backed by insurance, and regulatory recognition of insurance in capital requirements.

1.5.3 Part III: Catalysis (Chapters 9–11)

Chapter 9 reviews the landscape of sci-tech finance. It describes the Chinese model, the US model, and the European model, and compares their performance on green innovation. The chapter finds that the Chinese model is strong on scale but weak on selection, the US model is strong on selection but weak on scale, and the European model is intermediate on both. It then examines the specific instruments: government guidance funds, science and technology banks, intellectual property pledges, and venture capital.

Chapter 10 examines government-guided venture capital and the green premium. It reviews the evidence on the green premium and finds that it is positive for mature green technologies and negative for immature ones. The chapter then examines the design of government-guided funds and finds that their performance depends on the incentive structure. It presents case studies of two Chinese funds, one successful and one unsuccessful, and draws lessons from the comparison.

Chapter 11 addresses performance evaluation and incentive design. It develops a multi-objective evaluation framework and applies it to a sample of Chinese government guidance funds. The results show that funds with clear environmental mandates and independent management teams perform better. The chapter concludes with recommendations for incentive design: performance-based compensation, independent evaluation, and transparency in reporting.

1.5.4 Part IV: Integration (Chapters 12–15)

Chapter 12 examines the digital infrastructure and data systems that the integrated mechanism requires. It describes the data needs of each pillar—environmental data for rating, weather data for insurance, and innovation data for catalysis—and examines the current state of data availability in China and Europe. The chapter argues that the integration of the three pillars requires a shared data infrastructure, and it proposes a design for such an infrastructure.

Chapter 13 addresses international cooperation, technology transfer, and climate finance. It examines the mechanisms for financing the green transition in developing economies, including the Green Climate Fund, multilateral development banks, and private capital. The chapter argues that the three-pillar framework applies to developing economies as well, but with different weights: insurance is more important in agriculture-heavy economies, and catalysis is more important in economies with weak venture capital markets.

Chapter 14 lays out the policy sequencing for the integrated mechanism. It addresses the question of what to do first, and argues that the sequence should be: disclosure and data infrastructure, then rating, then insurance, then catalysis. The chapter examines the experience of the EU and China in implementing green finance policies and draws lessons for sequencing.

Chapter 15 concludes with the vision of a resilient and inclusive green financial system. It summarizes the argument, acknowledges the limitations, and identifies the open questions for future research. The chapter closes with a concrete agenda for the next decade: the integration of the three pillars into a single mechanism, the scaling of green finance in developing economies, and the development of the data infrastructure that the integration requires.

1.6 Key Arguments and Contributions

1.6.1 The Integration Argument

The book's central contribution is the integration argument: the three pillars are more effective together than apart, and the interaction effects are large enough to justify treating them as a single mechanism. The integration argument has three parts. The first part concerns the transmission of information. A rating that incorporates environmental performance produces a signal. That signal is useful to the lender, but it is also useful to the insurer and to the venture fund. An insurer writing a policy on a forestry company benefits from knowing the company's environmental rating, because the rating reflects the quality of the company's forest management. A venture fund evaluating a green technology start-up benefits from knowing the rating, because the rating reflects the start-up's environmental performance relative to its peers. In the current fragmented system, each institution builds its own assessment, duplicating costs and producing inconsistent signals. In the integrated system，the rating is a shared input, and the consistency of the signal reduces the cost of assessment for all three.

The second part concerns the pricing of risk. Insurance reduces default probability by stabilizing cash flow. If the lender does not observe the insurance, the reduction is not reflected in the interest rate. The borrower pays a spread that overstates its risk, and the insurance premium is a cost without a financing benefit. The integrated mechanism requires that the insurance be observable and that the lender's pricing model incorporate it. This is not a trivial requirement. It means that the insurance contract must be standardized, that the insurer must be creditworthy, and that the lender must have a model for translating insurance coverage into default probability. Chapter 8 develops such a model.

The third part concerns the staging of capital. Green technology firms pass through stages: seed, early growth, scale-up, and maturity. Each stage requires a different type of capital. Catalysis provides the seed and early growth capital. Rating and insurance enable the scale-up and maturity capital. In the fragmented system, the transition between stages is abrupt: a firm that has exhausted its venture capital cannot obtain bank credit because it has no rating history, and it cannot obtain insurance because its risk is not yet measurable. The integrated mechanism smooths the transition by providing the rating and insurance that the firm needs to access bank credit. The smoothing effect is largest for firms in the scale-up stage, which is precisely where the financing gap is widest.

1.6.2 The Measurement Argument

The second contribution is methodological. The book develops measurement tools for each pillar and shows how they connect.

For the rating pillar, the book develops a scoring model that specifies the channels through which environmental performance affects default probability. The model is calibrated on Chinese data, and the calibration shows that the environmental variables have statistically significant and economically meaningful effects on default probability. A one-standard-deviation improvement in environmental performance reduces the default probability by 10 to 20 percent, depending on the sector. The effect is larger for sectors with high transition risk (steel, cement, power) and smaller for sectors with low transition risk (services, light manufacturing).

For the insurance pillar, the book develops a measure of basis risk and shows how it varies with the design of the index. The measure is the correlation between the index payout and the farmer's actual loss. The book finds that basis risk is lower when the index is designed at a fine spatial scale, when the index uses multiple weather variables, and when the index is combined with a traditional indemnity component. The finding has direct implications for product design.

For the catalysis pillar, the book develops a multi-objective evaluation framework that combines financial return, environmental impact, and innovation. The framework is applied to a sample of Chinese government guidance funds, and the results show that the funds' performance varies widely. The variation is explained by the incentive structure: funds with performance-based compensation and independent management perform better on all three objectives.

The connection between the measurement tools is the key. The rating model's output is an input to the insurance model's pricing. The insurance model's output is an input to the catalysis model's evaluation. The catalysis model's output is an input to the rating model's calibration, because the firms that the catalysis supports eventually become borrowers whose default experience validates the rating. The three models form a loop, and the loop is the measurement counterpart of the integration argument.

1.6.3 The Policy Argument

The third contribution is policy. The book argues that the integration of the three pillars requires specific policy designs, and it develops those designs in detail.

The first policy requirement is data infrastructure. The rating model needs environmental data. The insurance model needs weather data. The catalysis model needs innovation data. In the current system, these data are scattered across ministries, agencies, and private providers, and they are not standardized. The book proposes a shared data infrastructure that would make the data available to all three pillars, with appropriate privacy and confidentiality protections. Chapter 12 develops the proposal.

The second policy requirement is regulatory recognition. The rating model's environmental adjustments must be recognized in capital requirements. The insurance model's risk reduction must be recognized in loan loss provisioning. The catalysis model's guarantees must be recognized in risk-weighted assets. Without regulatory recognition, the private incentives to use the models are weak, because the models' benefits are not reflected in the cost of capital. Chapter 14 develops the regulatory agenda.

The third policy requirement is sequencing. The three pillars cannot be built simultaneously in all jurisdictions. The book argues for a sequence: data infrastructure first, then rating, then insurance, then catalysis. The sequence is driven by the dependencies: rating requires data, insurance requires rating (because the insurer needs to assess the borrower's risk), and catalysis requires insurance (because the venture fund needs to know that the scale-up financing will be available). The sequence is not rigid—some jurisdictions may be able to jump stages—but it provides a guide for policymakers who are building the mechanism from scratch.

1.6.4 What the Book Does Not Do

The book does not attempt a comprehensive survey of green finance. The literature is vast and growing, and a survey would be outdated before it was published. The book is an argument, not a handbook. It develops the argument through selected models, cases, and evidence, and it refers the reader to the literature for the full range of approaches.

The book does not attempt a full general equilibrium analysis. The three-pillar framework is partial equilibrium: it takes the carbon price, the policy environment, and the macroeconomic conditions as given, and it analyzes how the financial system responds. A full general equilibrium analysis would need to model the feedback between financial allocation and the real economy, which is beyond the book's scope. The partial equilibrium approach is a limitation, and the book is explicit about it.

The book does not attempt a definitive empirical validation. The data on green default experience is limited, because the green transition is recent. The calibration in Chapter 4 is provisional, and the book says so. The empirical work in Chapter 7 is based on agricultural insurance, which is a proxy for the broader insurance pillar. The book's empirical claims are therefore conditional: they hold for the samples and periods studied, and they should be re-examined as more data becomes available.

These limitations are not defects. They are the boundaries of the argument. The book's contribution is to specify the integration logic and to develop the measurement and policy tools that the logic requires. The empirical validation is a task for the next decade, as the green transition generates the data that the validation needs.

1.7 A Note on Evidence and Method

The evidence base for this book was assembled over fifteen years, and it is worth being explicit about how it was assembled.

The credit rating work draws on a database of 3,200 Chinese technology firms, collected between 2009 and 2019. The database includes financial statements, patent counts, R&D expenditure, and—for a subsample—environmental performance indicators. The environmental indicators were collected from provincial environmental protection bureaus and from the firms' own disclosures. The data quality is uneven: the disclosures are voluntary, and the bureaus' records are incomplete. The calibration in Chapter 4 uses the subsample with complete environmental data, which is about 40 percent of the full database. The subsample is not random—firms with better environmental data tend to be larger and more profitable—so the calibration may overstate the environmental effect for small firms. The book flags this issue and treats the calibration as an upper bound.

The insurance work draws on a household survey conducted in Heilongjiang and Henan provinces in 2011, 2013, and 2015. The survey covered 2,400 households in 48 villages. The sampling was stratified by county and village, with random selection within strata. The survey collected data on insurance participation, crop yields, input use, income, and demographic characteristics. The weather data came from the provincial meteorological bureaus, matched to the survey villages by distance. The empirical strategy in Chapter 7 uses the rollout of the insurance program across counties as a source of variation, with a difference-in-differences design. The design is not perfect—the rollout was not random—but the parallel trends assumption is supported by pre-trend tests.

The catalysis work draws on a sample of 28 Chinese government guidance funds, collected between 2015 and 2020. The sample includes funds at the national, provincial, and municipal levels. The data include fund size, investment portfolio, management team, compensation structure, and performance metrics. The performance metrics are financial return (where available), environmental impact (measured by the portfolio companies' emissions reductions), and innovation (measured by patent counts). The data on financial return is incomplete, because many funds do not disclose their returns. The analysis in Chapter 11 therefore focuses on the funds with complete data, which is about half the sample. The selection bias is a concern, and the book treats the results as suggestive rather than definitive.

The qualitative evidence comes from field research in Heilongjiang, Zhejiang, and Guangdong provinces, conducted between 2011 and 2019. The field research included interviews with bankers, insurers, venture capitalists, and firm managers. The interviews were semi-structured and lasted between one and three hours. The purpose was to understand how the financial institutions actually make decisions, as opposed to how the models say they should. The gap between the models and the practice was often large, and the gap is a recurring theme in the book.

The methodological stance is pragmatic. The book uses the tools that are appropriate to the question: formal models for the integration logic, econometric methods for the empirical claims, and case studies for the institutional analysis. The pragmatic stance is a choice, not a claim of superiority. The green transition is a complex phenomenon, and no single method captures it fully. The book's contribution is to combine the methods in a way that illuminates the integration.

1.8 The Chapter Roadmap

The remainder of the book is organized as follows.

Chapter 2 develops the theoretical foundation. It reviews the information asymmetry literature and extends it to the green context. The key argument is that environmental information asymmetry creates a distinct source of credit rationing, and that the solution requires a rating technology, not just disclosure.

Chapter 3 examines the existing ESG rating methodologies and their limitations for credit purposes. It develops a credit-specific framework that specifies the channels through which environmental performance affects default probability.

Chapter 4 develops the scoring, default probability, and loss given default models. It includes a worked example of a green steel project and shows how the model produces a rating that differs from the conventional rating.

Chapter 5 addresses the portfolio level. It presents factor and copula models for default dependence, with the environmental factor as an additional systematic factor. It shows that ignoring the environmental factor leads to an underestimation of portfolio risk.

Chapter 6 introduces weather index insurance and forest insurance. It explains the design of index products, the basis risk problem, and the conditions under which insurance can make a green asset bankable.

Chapter 7 presents the empirical evidence on agricultural insurance and farmer income. It describes the survey data, the identification strategy, and the results, and it examines the heterogeneity of the effect.

Chapter 8 connects insurance to credit. It develops a model of the lender's decision with and without insurance, and it examines the obstacles to linking insurance and credit in practice.

Chapter 9 reviews the landscape of sci-tech finance. It compares the Chinese, US, and European models and examines their performance on green innovation.

Chapter 10 examines government-guided venture capital and the green premium. It reviews the evidence on the green premium and examines the design of government-guided funds.

Chapter 11 addresses performance evaluation and incentive design. It develops a multi-objective evaluation framework and applies it to a sample of Chinese government guidance funds.

Chapter 12 examines the digital infrastructure and data systems that the integrated mechanism requires. It describes the data needs of each pillar and proposes a shared data infrastructure.

Chapter 13 addresses international cooperation, technology transfer, and climate finance. It examines the mechanisms for financing the green transition in developing economies.

Chapter 14 lays out the policy sequencing for the integrated mechanism. It addresses the question of what to do first and draws lessons from the EU and Chinese experience.

Chapter 15 concludes with the vision of a resilient and inclusive green financial system. It summarizes the argument, acknowledges the limitations, and identifies the open questions for future research.

1.9 Conclusion: The Argument Stated Plainly

The green transition is the largest investment program in human history. The International Energy Agency's estimate of USD 4 trillion per year by 2030 is not a forecast; it is a requirement. The financial system has the savings to meet the requirement. What it lacks is the mechanism for converting savings into green investment at the required scale and speed.

The mechanism has three parts. The first is a rating technology that identifies green performance and translates it into a credit-relevant signal. The second is an insurance technology that buffers the risks that credit markets cannot absorb. The third is a catalysis technology that provides capital to green innovation at the stage where conventional finance cannot reach. The three parts are not independent. They interact, and the interactions are large enough to justify treating them as a single mechanism.

The book develops the mechanism in detail. It specifies the models, presents the evidence, and designs the policies. It is not a complete solution—the green transition will require changes in fiscal policy, industrial policy, and international governance that are beyond the book's scope. But it is a necessary part of the solution, and it is a part that the financial system can build with the tools it already has.

The stakes are concrete. The steel mill in Hebei, the forestry company in Yunnan, the electrolyzer start-up in Shenzhen—these are not abstractions. They are decisions being made now, with the financial system's current tools. The book is an argument that the tools can be better. The next chapter begins the argument with the theory of information asymmetry and credit rationing, which is the foundation of the rating pillar and, in a sense, of the entire framework.