Chapter 14: Policy Sequencing and the Integrated Mechanism: Rating, Insurance, Catalysis

In the spring of 2016, a provincial development bank in western China launched a green credit program for small and medium-sized manufacturing firms. The mandate was narrow: lend at preferential rates to enterprises that could document verifiable reductions in energy intensity. Within eighteen months the loan book reached 4.2 billion yuan across 340 borrowers. By the spring of 2018, non-performing loans in the program stood at 11.3 percent, roughly four times the bank's overall NPL ratio. An internal review, circulated in draft and never published, named the cause with unusual candor. The bank had screened borrowers on environmental criteria—emissions per unit of output, water discharge, solid waste handling—and had assessed creditworthiness on one dimension only: collateral coverage. When a downturn in the regional construction market hit the borrowers' order books, the environmental metrics stayed satisfactory while the firms slid toward default. The green screen had been thorough. The credit screen had been missing.

This case illustrates the central claim of this book. Support for the green transition is a sequence of institutional functions, and when the sequence is inverted or a function is missing, the others fail. The bank had built the lending channel without the identification channel or the buffering channel. What emerged was a portfolio that looked green on paper and performed badly on the balance sheet.

This chapter turns the three-pillar framework developed in Parts II through IV into a policy sequencing roadmap. It is the most practical chapter in the book and the most provisional. Sequencing claims are conditional by nature: what works in a middle-income economy with an established banking supervisor and a nascent insurance market will not work in a low-income economy with neither. I will state the principles as principles, not as laws, and I will be explicit about where the evidence runs thin.

14.1 Principles of Policy Sequencing

Five principles govern the ordering of interventions. They are not independent. The fourth constrains the first three, and the fifth is a boundary condition on the entire framework.

14.1.1 Identification Precedes Allocation

A financial system cannot allocate capital to green uses unless it can distinguish green from non-green at the level of the individual borrower or project. That sounds trivial. It is not. The distinction requires a methodology—criteria, a data collection protocol, a scoring or rating rule—and the methodology has to sit inside the credit process rather than arrive afterward as a reporting requirement.

Identification comes first because the other two pillars depend on it. Insurance underwriting for eco-agriculture requires an assessment of the underlying risk, which requires a view of the borrower's exposure and vulnerability. Catalytic capital—government-guided venture capital, green bonds, blended finance—requires a way to verify that the recipient is actually engaged in the transition rather than claiming to be. Without identification, insurance premiums are mispriced and catalytic capital is misdirected. The Chinese program described above had identification in the narrow sense (environmental metrics) but not in the credit sense (default probability conditional on environmental performance). The first is a screening device; the second is a pricing device.

What does identification require in practice? A taxonomy that defines what counts as green, with thresholds specific enough to be operational. The European Union's taxonomy, adopted in 2020 and refined through 2023, is the most elaborate attempt to date: technical screening criteria for 107 economic activities across eight sectors, plus a "do no significant harm" principle that disqualifies an activity if it damages other environmental objectives. The taxonomy is not perfect. The inclusion of natural gas and nuclear power in the 2022 complementary delegated act was politically contested and scientifically debatable. But it establishes the principle that green is a defined category rather than a marketing label.

A taxonomy without data is a list of aspirations. The data infrastructure includes emissions reporting standards, verification protocols, and the digital plumbing that lets a lender pull a borrower's environmental performance into the credit file. Chapter 12 treated this infrastructure in detail; the sequencing point is that the data infrastructure must be in place before the rating methodology can be calibrated, and the rating methodology must be calibrated before capital allocation can be risk-sensitive.

Then comes the rating or scoring methodology that integrates environmental performance into the credit assessment. Chapter 3 set out the conceptual case for extending traditional credit rating to include environmental variables. The operational point for sequencing is that this integration is a capability to be built, tested, and refined, not a one-time methodological choice. A supervisor that mandates environmental integration before the underlying data and analytical capacity exist will get compliance theater: boxes ticked, ratings unchanged.

Identification is expensive, and the cost falls disproportionately on small borrowers who lack the staff and systems to produce verifiable environmental data. A sequencing strategy that imposes full taxonomy-aligned reporting on all borrowers at once will ration credit away from precisely the small and medium enterprises that need transition capital most. The practical resolution is tiered reporting: simplified metrics for small borrowers, full taxonomy alignment for large ones, with a phased timetable for tightening the small-borrower requirements as data infrastructure improves.

14.1.2 Risk Buffering Precedes Scaling

Insurance and other risk-transfer instruments must be in place before green lending is scaled to the point where a correlated shock can impair the portfolio. This follows directly from the portfolio theory in Chapter 5. Green lending—particularly to agriculture, forestry, and early-stage clean technology—is exposed to correlated risks: weather, commodity prices, policy changes, technology cost curves. A portfolio of green loans without risk buffering is a concentrated bet on a small number of correlated variables. The bank in the opening anecdote made exactly this bet, and the regional construction downturn was the correlated shock that called it.

Insurance capacity therefore has to be developed in parallel with, or ahead of, the scaling of green credit. Most policy programs follow the opposite order: subsidize green lending, watch the loan book grow, discover the correlation problem when defaults spike, then retrofit insurance. The retrofit is expensive and slow. Weather index insurance, the instrument analyzed in Chapter 6, took roughly two decades to move from pilot to scale in India and Mexico, and it remains commercially unviable without subsidy in most smallholder contexts. A country that waits until its green agricultural loan book is large before investing in index insurance capacity will spend those two decades absorbing weather shocks on its own balance sheet.

A corollary is easy to miss: risk buffering must be priced, not given away. A subsidized premium that does not reflect the underlying risk sends the wrong signal to borrowers and undermines the insurer's solvency. The Chinese agricultural insurance program, discussed in Chapter 7, is a cautionary example. Premium subsidies reached 60 to 80 percent of the actuarially fair premium in some provinces, which achieved high coverage rates but also encouraged cultivation in marginal areas and created a fiscal liability that grew faster than the agricultural output it was meant to protect. The empirical evidence on income stabilization reviewed in Chapter 7 suggests the insurance did stabilize incomes—at a fiscal cost that cannot be replicated at scale without a clear exit strategy.

14.1.3 Catalysis Requires a Functioning Identification and Buffering System

Government-guided venture capital, green bond markets, and blended finance work by changing the risk-return profile of green investment so that private capital will follow. They do not work in the absence of identification and buffering, because the private capital they are meant to catalyze will not follow into a market where it cannot assess risk or transfer it.

The evidence here is reasonably strong, though it comes mostly from the venture capital literature rather than from green finance specifically. Cumming and MacIntosh's 2006 study of Canadian government venture capital found that public funds crowded out private investment when they invested in the same deals on the same terms, but added to total investment when they invested earlier or in different geographies. The mechanism is straightforward: public capital that enters at the same stage and on the same terms as private capital competes with it; public capital that enters earlier absorbs risk that private capital cannot price. The same logic applies to green finance. A green bond market that prices green bonds at the same yield as conventional bonds, because investors cannot distinguish the two, will not redirect capital. A market where the green label is verified and the risk is assessed will.

For sequencing, this means catalytic capital should be deployed after the identification and buffering systems are functional, and into the segments where private capital is absent rather than the segments where it is already active. The temptation to deploy catalytic capital early, to show visible progress, is strong. It is also the most common sequencing error in green finance.

14.1.4 Institutional Complementarity Constrains the Sequence

The three pillars are complements, not substitutes. A country with strong rating capacity but no insurance capacity will have a green credit market that is well-priced in normal times and fragile in stress. A country with strong insurance capacity but no rating capacity will have a market that is resilient to the insured risks but mispriced for everything else. The sequence is therefore not a simple linear ordering; it is a matter of building the three pillars in a balanced way, each pillar's development paced by the others.

Figure 14.1 presents the complementarity as a matrix of outcomes. The rows are the strength of the identification system (rating); the columns are the strength of the buffering system (insurance). The cells describe the resulting green credit market.

flowchart TD
    A[Identification: Rating Capacity] -->|Strong| B[Buffering: Insurance Capacity Strong]
    A -->|Weak| C[Buffering: Insurance Capacity Weak]
    B --> D[Well-priced, resilient green credit market]
    C --> E[Well-priced in normal times, fragile in stress]
    A2[Identification: Weak] -->|Strong| F[Buffering: Insurance Strong]
    A2 -->|Weak| G[Buffering: Weak]
    F --> H[Resilient to insured risks, mispriced otherwise]
    G --> I[Green-labeled lending without risk discipline]

The matrix is a simplification. In practice, identification and buffering are continuous variables, and the catalysis pillar interacts with both. But it captures the central point: the value of each pillar depends on the strength of the others. This is why sequencing is not a matter of picking the easiest pillar first. The easiest pillar is usually catalysis, because it involves deploying money, and deploying money is administratively simpler than building a rating methodology or an insurance market. Catalysis without the other two pillars produces the outcome in the bottom-right cell: green-labeled lending without risk discipline.

14.1.5 Sequencing Is Conditional on Initial Conditions

The fifth principle is a boundary condition on the first four. The optimal sequence depends on the country's initial conditions: the depth of its financial system, the quality of its supervisory capacity, the structure of its economy, and the nature of its green transition challenge.

Consider three stylized initial conditions.

A middle-income economy with a bank-dominated financial system, a competent banking supervisor, a nascent insurance market, and a large agricultural sector. China in the mid-2000s is a reasonable approximation. The binding constraint is identification: the banking system can lend, the supervisor can supervise, but the green taxonomy and the environmental data infrastructure do not yet exist. The sequence should start with taxonomy and data, then move to rating integration, then to insurance for the agricultural segment, then to catalytic capital for the industrial segment. This is roughly the sequence China followed, with the green credit guidelines of 2012 and the green bond guidelines of 2015 preceding the agricultural insurance expansion of 2013–2016 and the green finance pilot zones of 2017.

A low-income economy with a shallow financial system, a weak supervisor, and an economy dominated by smallholder agriculture. The binding constraint is capacity: there is no rating infrastructure to extend, and the insurance market is too thin to absorb the correlated risk. The sequence should start with publicly funded risk buffering—weather index insurance with a strong subsidy and a clear exit path—because the alternative is that the green transition in agriculture simply does not happen. Identification can be simpler: a short list of verifiable practices (cover cropping, agroforestry, water harvesting) rather than a full taxonomy. Catalysis should be limited to concessional finance from development partners until the domestic financial system can absorb it.

A high-income economy with a deep financial system, a strong supervisor, and a decarbonization challenge concentrated in industry and buildings. The binding constraint is coordination: the financial system has the capacity to identify, buffer, and catalyze, but the three functions are siloed across different institutions and different regulators. The sequence should start with the integration of the three functions—a unified green finance framework that links the taxonomy, the disclosure regime, the insurance regulation, and the catalytic capital mandate. The European Union's sustainable finance action plan, launched in 2018, is an attempt at this integration, and its difficulties show how hard coordination is even in a high-capacity setting.

These three cases are not exhaustive. They make one point: the principles are universal; the sequence is local.

14.2 Integrating Rating, Insurance, and Catalysis

The three pillars interact through specific mechanisms. This section describes those mechanisms, with attention to the feedback loops that make the integrated system more than the sum of its parts.

14.2.1 The Rating-Insurance Link: How Risk Assessment Improves Underwriting

A credit rating that incorporates environmental performance produces information an insurer can use to underwrite risk. In the absence of a rating, an insurer underwriting a green agricultural loan must assess the borrower's default probability from scratch, using data it may not have and analytical capacity it may not possess. A well-constructed rating provides a summary measure the insurer can use as an input.

The mechanism runs in both directions. Insurance improves the rating, because a borrower with insurance has a lower probability of default conditional on a weather shock. Chapter 8 modeled this link: the default probability of an insured borrower is lower than that of an uninsured borrower by an amount that depends on the correlation between the insured risk and the borrower's income, and on the basis risk of the insurance contract. The model produces a simple result:

where \(\rho\) is the correlation between the insured risk and the borrower's repayment capacity, and \(c\) is the coverage ratio of the insurance contract. The formula is an approximation—it assumes a linear relationship that holds only under specific distributional assumptions—but it captures the intuition: insurance reduces default probability in proportion to the correlation and the coverage. When the correlation is high, as in rainfed agriculture where a drought destroys both the crop and the borrower's ability to repay, the reduction is large. When the correlation is low, as in a diversified manufacturing firm where a weather shock affects one input among many, the reduction is small.

The practical implication is that the rating methodology should treat insurance coverage as a variable. A borrower with weather index insurance should receive a better rating than an otherwise identical borrower without it, and the improvement should be calibrated to the correlation and coverage. This is not standard practice. Most credit rating methodologies treat insurance as a mitigant at the margin, if they treat it at all. The integrated framework treats it as a structural variable.

There is a second, less obvious link. The rating gives the insurer a way to price the credit risk component of a bundled product. A bundled green loan-plus-insurance product carries two risk components: the weather risk (insured) and the credit risk (uninsured, or partially insured). The insurer offering the bundle needs to price both. The rating gives it a handle on the credit component. Without the rating, the insurer either declines the bundle or prices it conservatively, which reduces take-up.

14.2.2 The Insurance-Catalysis Link: How Risk Buffering Unlocks Private Capital

Risk buffering unlocks private capital by changing the risk-return profile of green investment. The mechanism is easiest to see in project finance. A renewable energy project—a solar farm, a wind farm, an electrolyzer—has a revenue stream that depends on resource availability (sun, wind, electricity prices) and a cost stream that depends on construction and operation. Both are subject to risk. Insurance can transfer some of it: construction all-risk insurance covers the construction phase, business interruption insurance covers operational disruptions, and parametric insurance can cover resource variability.

The effect on the cost of capital is substantial. A project with a comprehensive insurance package can be financed at a lower weighted average cost of capital than an otherwise identical uninsured project, because the lenders' loss given default is lower and the equity investors' cash flow volatility is reduced. The magnitude depends on the project and the market. A 2020 survey by the Climate Policy Initiative of utility-scale renewable projects in Europe and North America found insurance-related risk mitigation reducing the cost of debt by roughly 50 to 150 basis points. On a 100 MW solar project with a capital cost of €80 million, that is €400,000 to €1.2 million per year in debt service savings.

The catalysis link also runs through the rating. A project with a strong rating and comprehensive insurance is a candidate for green bond issuance, because the bond market requires both a credit assessment and a risk mitigation story. The green bond market's growth—from $11 billion in 2013 to over $500 billion in 2023, according to the Climate Bonds Initiative—has been driven in part by the availability of insurance and rating infrastructure that makes green projects bankable at scale.

Insurance is not a substitute for a sound project. A policy cannot make a bad project good; it can only make a good project financeable. The temptation to use insurance as a cosmetic fix—to wrap a weak project in a policy and call it de-risked—is real, and it has produced losses for insurers and investors. The framework's insistence on identification before buffering is meant to prevent this: the rating must come first, and the insurance must be priced against the rated risk.

14.2.3 The Catalysis-Rating Link: How Capital Markets Demand Better Identification

Capital markets demand identification. A green bond that is not rated, or whose green credentials are not verified, trades at a discount or does not trade at all. The demand for identification is a powerful force for improving rating methodology, because issuers who want access to capital markets will invest in the data and analytical capacity the rating requires.

This link explains why green bond markets have developed faster than green bank lending in many countries. A green bond issuer must produce a framework, obtain a second-party opinion, and report on use of proceeds. These requirements are a form of identification, and the market enforces them rather than the regulator. The enforcement is imperfect—greenwashing is a real problem, and the second-party opinion market has its own conflicts of interest—but it is stronger than the enforcement in bank lending, where the green label is often applied by the lender itself.

The catalysis-rating link also runs through the venture capital channel. A government-guided venture capital fund investing in green technology start-ups must assess the technology and the team. The assessment is a form of rating, even if it is not called that. The fund's investment decisions produce information other investors can use, and the information improves as the fund builds a track record. The Israeli experience with Yozma, the government-seeded venture capital program launched in 1993, is instructive: the program's success depended on its ability to attract experienced private fund managers who could assess technology risk. Assessment capacity was the scarce input, not capital.

14.2.4 The Feedback Loop: How the Three Pillars Reinforce Each Other

The three pairwise links combine into a feedback loop. Rating improves insurance underwriting, which improves the risk-return profile of green investment, which attracts catalytic capital, which demands better rating, which improves rating. The loop can be virtuous or vicious depending on where it starts.

Figure 14.2 presents the loop as a diagram.

flowchart LR
    R[Rating: Identification of green credit risk] -->|Informs underwriting| I[Insurance: Risk buffering]
    I -->|Reduces default probability and cost of capital| C[Catalysis: Private and public capital]
    C -->|Demands verification and disclosure| R
    R -->|Better data and methodology| R
    I -->|Lower losses, more capacity| I
    C -->|More deals, more information| C

The loop is not automatic. It requires deliberate institutional design at each node. The rating methodology must be open enough that insurers can use it. The insurance products must be standardized enough that capital markets can price them. The catalytic capital must be patient enough to absorb the learning costs of the first few deals. These are design choices, not market outcomes.

The loop has a dark side. A vicious loop can set in if the first deals perform badly. A green bond that defaults because the project was poorly identified damages the market's confidence in green bonds generally. An insurance product that fails to pay out because the index was badly designed damages the market's confidence in index insurance. The first few deals in any new market carry disproportionate weight, and the principle that identification precedes allocation exists to make sure those first deals are well-identified.

14.3 Implementation Roadmap

The roadmap has four phases. They are sequential in logic and overlapping in time; a country in Phase 2 is still completing Phase 1 tasks. The timeline is indicative, not prescriptive. The phase descriptions are written for a middle-income economy with a bank-dominated financial system, because that is the most common setting for green finance policy; adaptations for other settings are noted where they matter.

14.3.1 Phase 1: Foundation (Years 1–3)

The foundation phase establishes the identification infrastructure. The deliverables are a green taxonomy, a data collection and verification protocol, and a pilot rating methodology.

The taxonomy should be developed by a technical working group that includes the financial regulator, the environmental regulator, industry associations, and academic experts. The working group's first task is to decide scope: which sectors are covered, and at what granularity. A taxonomy that covers everything at a high level of abstraction is not operational; a taxonomy that covers a few sectors in detail is operational but incomplete. The practical resolution is to start with the sectors where green lending is already happening—typically renewable energy, energy efficiency, and sustainable agriculture—and to expand as capacity grows.

The data protocol should specify what borrowers must report, how it is verified, and how it flows into the credit file. Verification is the hardest question. Third-party verification is expensive; self-reporting is unreliable; satellite and remote sensing data are increasingly available but require analytical capacity to use. The protocol should be tiered: large borrowers use third-party verification, medium borrowers use a combination of self-reporting and random audit, small borrowers use simplified metrics with a presumption of compliance for listed practices.

The pilot rating methodology should be developed with a small number of banks and a small number of borrowers. The pilot's purpose is not to produce a final methodology but to learn what data are available, what data are useful, and what the integration costs. It should be evaluated on three criteria: whether the environmental variables have predictive power for default (the statistical question), whether loan officers can use the methodology in practice (the operational question), and whether borrowers can produce the data at reasonable cost (the compliance question). A pilot that fails any of the three is not ready to scale.

The foundation phase also includes the first insurance capacity building. The deliverable is a feasibility study for weather index insurance or another risk-transfer instrument in the priority sector. The study should assess data availability (weather station density, satellite coverage), the actuarial basis (historical loss data, crop models), the delivery channel (banks, cooperatives, mobile money), and the fiscal cost of subsidy. It should not attempt to launch a product; that is Phase 2.

14.3.2 Phase 2: Integration (Years 2–5)

The integration phase links the three pillars in a small number of pilot transactions. The deliverable is a set of integrated green finance deals—bundled loan-plus-insurance products, green bonds with verified use of proceeds and credit ratings, or catalytic capital investments with rating and insurance components—that demonstrate the framework in practice.

The pilot transactions should be selected to test the links. A bundled agricultural loan-plus-insurance product tests the rating-insurance link. A green bond with a credit rating and a second-party opinion tests the catalysis-rating link. A blended finance structure with a first-loss tranche from a development bank tests the insurance-catalysis link. The pilots should be documented in detail, including costs, timelines, and failures. That documentation is the most valuable output of the phase, because it is the input to scaling.

The integration phase also moves insurance capacity from feasibility to product. The product should launch in a limited area with a clear subsidy and a clear evaluation plan. The evaluation should measure take-up, claims experience, and the effect on borrower default rates. The last measure is the one that matters for the integrated framework, and it is the one most often missing from insurance pilots.

The institutional design question in this phase is where the integration sits. Three models are possible. In a lead institution model, one bank or one insurer takes the lead in developing the integrated product and the others follow. In a platform model, a shared infrastructure—a data platform, a rating engine, an insurance pool—serves multiple institutions. In a regulatory model, the regulator mandates the integration through disclosure and capital rules. The choice depends on initial conditions. The lead institution model works when there is a capable institution willing to invest; the platform model works when there is a public good to be provided; the regulatory model works when the market is mature enough to respond to price signals.

14.3.3 Phase 3: Scaling (Years 4–8)

The scaling phase expands the integrated mechanism from pilots to the mainstream of green finance. The deliverables are the integration of environmental performance into the standard credit rating methodology, the expansion of insurance coverage to the majority of green borrowers in the priority sectors, and the deployment of catalytic capital at scale.

Rating integration is the hardest deliverable. It requires the rating agencies, the banks' internal rating units, and the supervisor to agree on a methodology. The agreement is technical and political at once, because the methodology determines which borrowers get cheaper credit. The practical approach is to phase the integration: first, environmental performance is a supplementary factor that adjusts the rating at the margin; then it becomes a full factor with a defined weight; finally it is embedded in the default probability model. The phasing lets the market adjust and the data accumulate.

Insurance expansion requires solving the delivery problem. Pilot products are typically delivered through a single channel—a bank, a cooperative, a mobile money provider. Scaling requires multiple channels and a way to reconcile them. The Indian experience with weather index insurance is instructive: the product was piloted in the 2000s, scaled through the public sector banks in the 2010s, and reached about 30 million farmers by the early 2020s, but the claims ratio has been volatile and basis risk remains a persistent complaint. Scaling insurance is a distribution problem as much as an actuarial one.

Catalytic capital deployment requires a clear mandate and a clear exit strategy. The mandate should specify the segments where catalytic capital is expected to operate—early-stage technology, underserved regions, first-of-a-kind projects—and the segments where it should not. The exit strategy should specify how the capital will be withdrawn as private capital enters, so that it does not become a permanent subsidy. The Israeli Yozma program is the canonical example of a well-designed exit: the government committed $100 million, attracted $150 million in private matching funds, and sold its stakes in the funds after a fixed period, exiting the market by the early 2000s.

14.3.4 Phase 4: Consolidation (Years 7–12)

The consolidation phase embeds the integrated mechanism in the permanent institutional architecture. The deliverables are the revision of the taxonomy and rating methodology based on accumulated data, the integration of the three pillars into the supervisory framework, and the development of a domestic green finance market that does not depend on concessional capital.

Methodology revision is a continuous process, not a one-time event. The taxonomy should be updated every three to five years to reflect technological change and scientific understanding. The rating methodology should be recalibrated as default data accumulate. The insurance products should be repriced as the actuarial basis improves. The consolidation phase institutionalizes this revision process.

Supervisory integration is the most important deliverable. The supervisor should incorporate environmental performance into the capital framework—not necessarily through a green supporting factor, which risks under-pricing the risk, but through a requirement that the rating methodology reflect environmental variables and that insurance mitigants be recognized. The European Banking Authority's 2021 discussion paper on the treatment of environmental risks in the prudential framework is one attempt at this integration; the Basel Committee's work on climate-related financial risks, ongoing since 2021, is another. The integration is technically difficult because the data are sparse and the models are immature, but the direction is clear.

Market development means reducing dependence on concessional capital. A green finance market that requires a government subsidy for every deal is not sustainable. The consolidation phase should aim for a market where the green premium—the additional return green investment earns, or the additional cost non-green investment bears—is sufficient to attract private capital without subsidy. The green premium is the subject of Chapter 10; here the point is that it is a market outcome, not a policy instrument, and it emerges when the three pillars are functioning.

14.3.5 The Roadmap as a Whole

Figure 14.3 presents the roadmap as a Gantt-style chart.

gantt
    title Integrated Green Finance Roadmap
    dateFormat YYYY
    axisFormat %Y
    section Phase 1 Foundation
    Green taxonomy           :a1, 2025, 18M
    Data protocol            :a2, 2025, 24M
    Pilot rating             :a3, 2026, 18M
    Insurance feasibility    :a4, 2026, 12M
    section Phase 2 Integration
    Bundled pilots           :b1, 2027, 24M
    Insurance product launch :b2, 2027, 18M
    Institutional design     :b3, 2027, 12M
    section Phase 3 Scaling
    Rating integration       :c1, 2029, 24M
    Insurance expansion      :c2, 2029, 30M
    Catalytic deployment     :c3, 2029, 36M
    section Phase 4 Consolidation
    Methodology revision     :d1, 2032, 24M
    Supervisory integration  :d2, 2032, 36M
    Market development       :d3, 2033, 36M

The chart is indicative. The phases overlap, and the durations are drawn from the observed timelines of comparable programs: the EU taxonomy took about three years from proposal to adoption, the Indian weather index insurance scale-up took about a decade, and the Israeli Yozma program took about eight years from launch to exit. A country that moves faster than these timelines is either learning from others' experience, which is possible, or cutting corners, which is dangerous.

14.3.6 The Financing of the Roadmap

The roadmap costs money. The public expenditure components are taxonomy development, data infrastructure, insurance subsidy, and catalytic capital. The private expenditure components are rating integration, insurance product development, and deal costs.

Public expenditure is front-loaded. Taxonomy and data infrastructure costs concentrate in Phases 1 and 2, and the insurance subsidy concentrates in Phases 2 and 3. Catalytic capital is deployed in Phase 3 and recovered in Phase 4 if the exit strategy works. Fiscal planning should reflect this profile: a government that commits to the roadmap should budget for the front-loaded costs and design the subsidy with a declining schedule.

Private expenditure is back-loaded. Rating integration and insurance product development costs fall on financial institutions, and they are incurred when the market is large enough to justify them. The principle that identification precedes allocation is also a financing principle: the public sector pays for the identification infrastructure because the private sector will not, and the private sector pays for scaling because the returns are there.

I should be candid about the uncertainty in these estimates. The cost of a green taxonomy depends on scope and consultation process; the EU's taxonomy cost tens of millions of euros, but a simpler national taxonomy could cost a few million. The cost of data infrastructure depends on the existing statistical system; a country with a strong statistical office can build on it, while a country without one must build from scratch. The cost of insurance subsidy depends on the actuarial basis and the coverage target; a program covering 10 million farmers at $10 per farmer per year costs $100 million per year, feasible for a middle-income country but not for a low-income one. The roadmap is a template, not a budget.

14.3.7 What the Roadmap Does Not Claim

The roadmap does not claim that the sequence is optimal in a formal sense. An optimal sequence would require a model of the economy specifying the production function, the financial frictions, and the policy instruments, and it would solve for the sequence that maximizes social welfare. I do not have such a model, and I am skeptical that one can be built with the data available. The roadmap rests on the observed experience of countries that have attempted the transition and on the theoretical logic of the three pillars. It is a reasonable sequence, not a proven one.

The roadmap does not claim that all countries should follow the same sequence. The fifth principle—conditionality on initial conditions—is a genuine constraint, not a rhetorical hedge. A low-income economy that tries to build a full taxonomy before it has a functioning insurance market will spend scarce administrative capacity on the wrong task. A high-income economy that tries to build insurance capacity before it has integrated its rating and disclosure frameworks will duplicate infrastructure it already has. The sequence must be adapted.

The roadmap does not claim that the three pillars are sufficient. They are the financial system's contribution to the green transition, not the whole of the transition. The transition also requires industrial policy, technological innovation, infrastructure investment, and behavioral change. The financial system can accelerate the transition by allocating capital and pricing risk, but it cannot substitute for the real economy's investment. A country that gets the financial sequence right but neglects the real economy will have a well-functioning green finance market that finances too little.

14.4 The Political Economy of Sequencing

The technical roadmap is necessary but not sufficient. The sequence is implemented by institutions, and institutions respond to incentives. This section examines who gains, who loses, and how the losers can be compensated.

14.4.1 The Distributional Effects of Rating Integration

Integrating environmental performance into credit rating redistributes credit access. Borrowers with strong environmental performance get better ratings and cheaper credit; borrowers with weak performance get worse ratings and more expensive credit, or no credit at all. The redistribution is the point of the policy, but it creates losers, and the losers have political voice.

The losers are the brown incumbents: coal-fired power plants, cement kilns, steel mills that have not yet decarbonized. These are often large employers in concentrated regions, and they have both the resources and the motivation to resist. The resistance takes several forms: lobbying against the taxonomy, litigation against the rating methodology, and the threat of stranded assets that would impose losses on the banks that lent to them.

The political economy literature on transition suggests that compensation is necessary for a successful transition. Compensation can take several forms: adjustment assistance for workers, transition finance for firms, regional development funds for communities. The financial system's role in the compensation is to provide transition finance—loans for retrofits, bonds for new equipment, insurance for technology risk—at a price that reflects the transition risk. The integrated framework's rating methodology should distinguish between a brown firm that is transitioning and a brown firm that is not, because the transition risk is different. A steel mill with a credible decarbonization plan is a better credit risk than a steel mill without one, even if both are currently brown.

14.4.2 The Fiscal Cost of Insurance Subsidy

The insurance subsidy is a fiscal transfer from taxpayers to insured borrowers. The transfer is justified by the positive externality of risk buffering—a resilient agricultural sector is a public good—but the magnitude of the transfer is a political choice, constrained by fiscal space.

The politics are complicated by the fact that the beneficiaries are dispersed and the costs are concentrated. The beneficiaries are millions of small farmers, each receiving a small subsidy; the costs are borne by the budget, a single line item. The dispersed beneficiaries have weak incentives to organize in support of the subsidy, while budget officials have strong incentives to cut it. Insurance subsidies are therefore often underfunded relative to need, and the underfunding is a binding constraint on scaling the buffering pillar.

The practical resolution is to design the subsidy with a clear sunset. A subsidy that is explicitly temporary—declining by 10 percentage points per year over a decade, for instance—is easier to defend than a permanent subsidy, because it is framed as an investment in market development rather than a permanent transfer. The sunset also creates an incentive for the insurer to reduce costs and for the borrower to adopt risk-reducing practices, because the subsidy will eventually disappear.

14.4.3 The Governance of Catalytic Capital

Catalytic capital is the most politically exposed of the three pillars, because it involves deploying public money into private markets. The governance challenge is to ensure that the capital is deployed on commercial terms, into the segments where private capital is absent, and withdrawn when it is no longer needed—without becoming a vehicle for patronage or a permanent subsidy to connected firms.

The record is mixed. The Israeli Yozma program is the success case: the government invested in ten funds, each managed by a private partner with a track record, and sold its stakes after five years. The German and Canadian programs of the 1990s and 2000s are closer to the failure case: public funds that invested alongside private funds on the same terms crowded out private investment without adding to the total, and political pressure to keep the funds operating after their mandate expired produced a permanent public presence in a market that no longer needed it. Leleux and Surlemont's 2003 pan-European analysis found evidence of crowding out in several national programs, and the finding has been replicated in subsequent studies.

The governance design that distinguishes success from failure has three features. The public capital is invested through private fund managers with their own capital at risk, so that investment decisions are made by people who bear the downside. The public capital has a defined exit, so that it does not become a permanent fixture. The public capital's mandate is limited to the segments where private capital is absent, so that it does not compete with the capital it is meant to catalyze. These features are not novel; they are the standard prescriptions of the venture capital literature. They are, however, politically inconvenient, because they limit the discretion of the officials who allocate the capital and the rents that the allocation creates.

14.4.4 The Sequencing of Political Feasibility

The technical sequence and the political sequence are not the same. A measure that is technically first—the taxonomy, for instance—may be politically difficult because it imposes costs on organized interests before it delivers benefits to dispersed ones. A measure that is technically last—catalytic capital—may be politically easy because it delivers visible benefits to identifiable recipients.

The practical implication is that the sequence should be designed to build political support as it proceeds. The foundation phase should include the beneficiaries of the later phases—the farmers who will receive insurance, the firms that will receive catalytic capital—so that they have a stake in the program's continuation. The integration phase should produce visible successes, a few well-publicized deals that demonstrate the framework's value. The scaling phase consolidates the coalition by delivering benefits at scale. The consolidation phase institutionalizes the coalition by embedding the framework in the permanent budget and the permanent regulatory architecture.

This is not cynicism. A policy sequence that is technically optimal but politically unsustainable will be abandoned before it delivers its benefits. The Chinese green finance pilot zones, launched in 2017 in five provinces, are an example of a politically designed sequence: the pilots were placed in provinces with strong local leadership and clear economic stakes in the transition, and their results were used to justify national expansion. The sequencing was not purely technical, but it was effective.

14.5 Country Adaptations

The roadmap is a template. This section describes how it adapts to the three stylized initial conditions introduced earlier, and to a fourth case—a small open economy with a specialized financial sector.

14.5.1 The Middle-Income Bank-Dominated Economy

This is the case the roadmap was written for. The binding constraint is identification, and the sequence starts with the taxonomy and the data infrastructure. The key adaptation is the pace of rating integration. A middle-income economy with a competent banking supervisor can move relatively quickly—three to five years from pilot to full integration—because the supervisory capacity exists to enforce the methodology. The insurance pillar is likely to be the slowest, because the actuarial data and the delivery channels take time to build. Catalytic capital is likely to be the smallest component, because the private capital market is deep enough to absorb green investment once identification and buffering are in place.

The risk in this case is that the sequencing is driven by the availability of donor or development finance rather than by domestic constraints. A country that receives a large green bond technical assistance grant may build the bond market before it has the rating methodology, because the grant is earmarked for the bond market. Earmarking is a sequencing hazard, and the practical response is to negotiate the assistance in a sequence-compatible form.

14.5.2 The Low-Income Agricultural Economy

The binding constraint is capacity, and the sequence starts with risk buffering. The key adaptation is the simplification of the identification pillar: a short list of verifiable practices, a simplified data protocol, and a rating methodology that relies on the practices rather than on measured emissions. The insurance pillar is publicly funded, with a strong subsidy and a clear sunset, and it is delivered through the channels that already reach smallholders—cooperatives, mobile money, input suppliers. Catalytic capital is concessional finance from development partners, and its mandate is to demonstrate the model rather than to scale it.

The risk is that the insurance subsidy becomes permanent and crowds out the development of a commercial market. The mitigation is the sunset and the declining schedule, backed by an explicit evaluation that measures the exit criteria. A program that cannot state what would have to be true for the subsidy to end is a program that will not end.

14.5.3 The High-Income Coordinated Economy

The binding constraint is coordination, and the sequence starts with the integration of the three functions. The key adaptation is the use of existing institutions—the rating agencies, the insurance regulators, the development banks—rather than the creation of new ones. The taxonomy is likely to align with the international standard (the ISSB climate disclosure standard, the EU taxonomy) rather than be built from scratch. The insurance pillar is likely to be commercially viable, with the public role limited to the catastrophic layer. Catalytic capital is likely to be targeted at first-of-a-kind technologies—green hydrogen, carbon capture, long-duration storage—where the private market cannot yet price the risk.

The risk is that coordination is achieved on paper but not in practice. The EU's sustainable finance action plan has produced an impressive architecture of regulations, but the integration of the three functions at the level of the individual transaction remains incomplete: a green bond issuer must comply with the taxonomy, the disclosure regulation, and the rating regulation, but the three compliance processes are separate and the data are not shared. The practical response is to build the shared data infrastructure—the European Single Access Point, mandated by the 2023 listing act, is an attempt—so that coordination becomes operational rather than architectural.

14.5.4 The Small Open Economy with a Specialized Financial Sector

This case is exemplified by financial centers such as Singapore, Luxembourg, and Ireland. The binding constraint is scale: the domestic green investment opportunity is too small to support the fixed costs of the three pillars, so the pillars must be built for the export market. The sequence starts with identification, because the export market demands it—a financial center that cannot verify the green credentials of the products it sells will lose the market to a center that can. The insurance and catalysis pillars follow, and they are built to serve the cross-border market rather than the domestic one.

The risk is regulatory arbitrage: a financial center that builds a permissive green taxonomy will attract issuers who want a green label without the substance. The mitigation is alignment with the international standard and participation in the supervisory networks—the Network for Greening the Financial System, the International Platform on Sustainable Finance—that monitor the arbitrage.

14.6 The International Dimension

Sequencing is not purely domestic. The international financial system—the development banks, the climate funds, the standard setters, the rating agencies—shapes the sequence in ways that are often invisible to the domestic policymaker.

14.6.1 The Standard-Setting Sequence

The international standard setters have their own sequence, and it constrains the domestic one. The ISSB's climate disclosure standard, IFRS S2, was issued in 2023 and is being adopted by jurisdictions through 2024 and 2025. A country that builds its taxonomy before the standard is adopted may find that its taxonomy is not interoperable with the international one, which raises the cost of cross-border green finance. The practical response is to build the taxonomy on the international standard's definitions, even if local implementation is phased.

The rating agencies' methodologies are also converging, slowly, on the integration of environmental variables. The convergence is uneven—Moody's, S&P, and Fitch have different approaches—but the direction is clear. A country that builds its rating methodology in isolation may find that its ratings are not comparable with the international agencies' ratings, which limits its borrowers' access to international capital.

14.6.2 The Climate Finance Sequence

The international climate finance architecture has its own sequence, and it is not always aligned with the domestic one. The Green Climate Fund, the Global Environment Facility, and the multilateral development banks have their own project cycles, accreditation requirements, and results frameworks. A country that wants to use international climate finance to support its sequencing must navigate these requirements, and the navigation costs are real.

The practical response is to use international finance for the phases where it is most useful—catalytic capital and the insurance subsidy—and to use domestic resources for the foundation phase, where international finance is least flexible. The foundation phase is a public good, and public goods are best funded by domestic budgets, because international finance is designed for projects rather than for institutional capacity.

14.6.3 The Technology Transfer Sequence

The green transition requires technology transfer, and the transfer is mediated by the financial system. A country that has built its rating and insurance pillars can absorb technology more easily, because the technology risk can be assessed and buffered. A country that has not built them will find that technology transfer is limited to the technologies that can be financed on the balance sheet of a development bank, which is a small subset.

The three pillars are a precondition for technology absorption, not just for capital allocation. The technology transfer literature often misses this point: transfer depends on financial capacity as much as on intellectual property rules or trade policy. The firms that can absorb the technology are the firms that can finance it, and the firms that can finance it are the firms that can be rated and insured. The three pillars are the financial system's contribution to absorption capacity.

14.7 Measuring Success

A sequencing roadmap needs a scorecard. The scorecard should measure the three pillars and their integration, and it should be simple enough to be used by a finance ministry or a central bank without a large analytical staff.

Table 14-1 presents a draft scorecard. The indicators are grouped by pillar, and each group has a small number of measures. The measures are chosen to be observable and comparable across countries, and they are deliberately few, because a scorecard with fifty indicators will not be used.

Table 14-1: Integrated Green Finance Scorecard

The targets are indicative. They are drawn from the observed performance of the leading programs—the EU taxonomy coverage, the Indian insurance claims ratio, the Yozma mobilization ratio—and they are meant to be challenging but achievable. A country that meets all ten targets has a functioning integrated mechanism.

The scorecard has a limitation worth stating openly: it measures the financial system's outputs, not environmental outcomes. A country can score well and still fail to reduce emissions, because the green lending may be financing activities that are labeled green but do not materially contribute to the transition. The scorecard should be complemented by an environmental outcome assessment—emissions per unit of output, renewable share of energy, forest cover—but that assessment is the responsibility of the environmental regulator, not the financial regulator. The financial regulator's job is to ensure that the financial system is doing its part; the environmental regulator's job is to ensure that the part adds up to the target.

14.8 The Limits of the Framework

I have argued that the integrated mechanism—rating, insurance, catalysis—is the financial system's contribution to the green transition, and that the sequence in which the three are built matters. I should now state the limits of the argument, because a framework presented without limits is a framework that will be misapplied.

The first limit is that the framework assumes a functioning financial system. In a country where the banking system is insolvent, the insurance market does not exist, and the capital market is closed, the three pillars cannot be built. The framework's preconditions are a sound banking system, a basic insurance law, and a capital market that can price risk. A country that lacks these preconditions needs financial sector development before it needs green finance sequencing. The two agendas are related—green finance can be a lever for financial development—but they are not the same, and conflating them produces unrealistic expectations.

The second limit is that the framework assumes a credible environmental policy. The financial system allocates capital in response to prices, and the prices are set by policy. A carbon price, a renewable portfolio standard, an efficiency mandate—these are the signals that make green investment profitable. A financial system that is perfectly sequenced but faces no environmental policy will allocate capital to the highest risk-adjusted return, which may be brown. The framework is a transmission mechanism, not a substitute for the policy signal. I made this point in Chapter 2, and it bears repeating: the financial system can accelerate the transition, but it cannot initiate it.

The third limit is that the framework's evidence base is uneven. The rating pillar is supported by a substantial literature on credit scoring and a growing literature on environmental credit risk, but the specific claim that environmental variables improve default prediction is supported by a small number of studies, mostly on listed firms in high-income countries. The insurance pillar is supported by a large literature on index insurance, but the evidence on the income stabilization effect is mixed, as I discussed in Chapter 7. The catalysis pillar is supported by the venture capital literature, but the evidence on green-specific catalytic capital is thin. The framework is a reasonable synthesis of the available evidence, not a proven recipe. The honest statement is that it is the best available guide, and that it should be implemented with evaluation built in, so that the guide can be revised as the evidence accumulates.

The fourth limit is that the framework does not address the distributional question at the global level. The green transition requires investment in the developing world, where emissions growth is fastest and capital is scarcest. The three pillars can be built in a developing country, but the cost is high relative to the country's fiscal capacity, and the international finance meant to cover the cost is insufficient and slow. The framework is silent on global redistribution because that question is outside its scope. A book on the financial system's contribution to the transition cannot resolve the global climate finance negotiation; it can only note that the negotiation is a precondition for the framework's application in the poorest countries.

14.9 Conclusion: The Sequence as a Discipline

The chapter began with a bank that lent green without assessing credit, and lost money. The bank was not foolish. The lesson is that the financial system's support for the green transition is a sequence of functions, and that skipping a function produces a failure that gets attributed to the wrong cause. The bank's failure was attributed to the borrowers' environmental performance, which was fine; the actual cause was the absence of a credit assessment, which was the missing function.

The three-pillar framework is a discipline. It says: identify before you allocate, buffer before you scale, catalyze after identification and buffering are in place. The discipline is inconvenient, because the easy money is in allocation and catalysis, and the hard work is in identification and buffering. The discipline is also provisional, because the evidence is uneven and the sequences are conditional on initial conditions. But it is the best available guide to a problem that will occupy the financial system for the next three decades.

The next chapter steps back from the mechanics and asks what the framework implies for the financial system as a whole. The question is whether the integrated mechanism is a temporary scaffolding for the transition or a permanent feature of the financial system. My answer, developed in Chapter 15, is that it is permanent, because the transition is not a one-time event but a continuous process of adjustment to a changing climate and a changing technology. The financial system that can identify, buffer, and catalyze the transition is the financial system that can survive it. The sequence is not a project; it is a capability.