Chapter 3: From Green Finance to Transition Finance: A Conceptual Map of the Financial Ecosystem

A steel mill in Hebei province wants to replace a blast furnace with electric-arc capacity. The project costs 1.4 billion yuan, cuts emissions by 40 percent, and pays back over nine years. The company's bank holds a green credit quota it must fill by December. The loan officer reads the file and declines: under the bank's taxonomy, electric-arc steel is less brown, not green, and the credit committee has no box to tick. The finance director tries a green bond. The framework requires proceeds to fund "eligible green projects," and a steel retrofit appears on no eligible list. The project stalls with a positive net present value at any reasonable discount rate.

The ecosystem prices credit risk, not transition risk. Those are different risks, requiring different data and different committees. Green finance, built over fifteen years, moves capital efficiently to wind farms and solar parks. It has no comparable machinery for retrofitting an existing industrial asset that is not yet green but could become significantly less brown. Most of the transition's capital needs sit in that gap.

This chapter maps the gap. It defines green and transition finance and locates the boundary between them, describes the institutional actors who must fund the transition, and argues that no single instrument or actor can do the job. The argument is structural. It is about what the system was built to price.

3.1 Defining Green and Transition Finance: Clarifying Concepts and Boundaries

3.1.1 The definitional problem

"Green finance" has no single binding definition, and the disagreement is substantive rather than editorial. It concerns what the category is for.

The narrowest definition is project-based: capital raised for projects that are themselves environmentally beneficial. A solar farm, a reforestation program, a wastewater treatment plant. The European Union's Taxonomy Regulation, in force since July 2020, formalizes this. An activity qualifies as environmentally sustainable only if it makes a substantial contribution to one of six environmental objectives, does no significant harm to the others, and meets minimum social safeguards. The taxonomy is a list, and a long one, but it remains a list. Something is on it or it is not.

The broadest definition is flow-based: any financial flow that supports the transition to a low-carbon economy, whether or not the underlying asset is green today. The People's Bank of China's green bond catalogue, first issued in 2015 and revised in 2021, takes a wider view than the EU taxonomy. It includes "clean coal" and "energy-saving" categories the EU would reject. The mismatch had a concrete consequence: a bond issued under the Chinese standard and marketed to European investors could not be labeled green in Europe. The China-EU Common Ground Taxonomy, published in 2021, reconciled the two on 55 percent of covered activities.

The definition determines which projects get cheaper capital. A green bond typically prices 2 to 8 basis points tighter than a comparable conventional bond. Small, but for a large issuer it runs into millions of dollars over the bond's life. The definition of green is a pricing mechanism.

Transition finance sits awkwardly against both definitions. The term emerged around 2019, largely from Asian financial centers—Singapore, Tokyo, Hong Kong—that recognized a political problem. Their economies are full of heavy industry that needs capital to decarbonize but cannot access green finance because it is not green. The Monetary Authority of Singapore's 2023 transition finance framework defines transition activities as those "not currently aligned with net-zero pathways but on a credible trajectory to align." That phrase, "credible trajectory," is the crux and the operational difficulty.

3.1.2 A working distinction

I use a three-way distinction that is more operational than the official taxonomies.

Green finance funds assets that are already low- or zero-carbon. The test is the asset's current emissions profile. A solar farm, a wind turbine, a green hydrogen electrolyzer, an electric bus fleet. The financing is green because the thing financed is green.

Transition finance funds the decarbonization of assets that are currently high-carbon. The test is the rate of change, not the level. A steel mill that cuts emissions intensity from 2.0 to 1.2 tons of CO₂ per ton of steel is a transition project. So is a cement plant installing carbon capture, a shipping company retrofitting vessels for methanol, a chemicals producer switching from naphtha to bio-based feedstock. The asset is not green at the end of the financing period. It is less brown than it was, and it is on a path.

Brown finance funds high-carbon assets with no credible decarbonization pathway. The test is the absence of a transition plan, or the presence of one that does not survive scrutiny. A new coal-fired power plant with no retirement date. A greenfield oil refinery. These are the assets transition finance is meant to render unnecessary, and no relabeling makes them acceptable.

The boundary between transition and green finance is a matter of trajectory. The boundary between transition and brown finance is a matter of credibility. Both are contested.

Take natural gas. A gas-fired plant emits roughly half the CO₂ of a coal plant per unit of electricity. If it replaces coal and runs for 20 years, is it transition finance? If it runs for 40 years, is it? The EU taxonomy proposed to include gas, then removed it, then included it under a separate "transitional activity" category after political negotiation in 2022. The final compromise requires an emissions threshold of 100 grams of CO₂ per kilowatt-hour—which most gas plants cannot meet without carbon capture—and a plan to switch to low-carbon fuels by 2035. The threshold does the work; the label does not.

Or take China Baowu Steel Group, which in 2021 issued a 5 billion yuan transition bond to fund a hydrogen-based direct reduction pilot. The bond was labeled transition under a framework the company developed with a third-party verifier. Verification relied on the company's own emissions baseline and its own projection of future reductions. No external standard defined what counted as a credible transition plan. The bond sold. Whether it should have carried the label is a question no regulator could answer at the time, because the standard did not exist.

Green finance has a taxonomy. Transition finance has a set of frameworks, most of them voluntary, most developed by the institutions that use them. They differ on the basics: what counts as a baseline, what counts as a credible plan, what happens if the plan fails.

3.1.3 Why the boundary is hard to draw

The difficulty is informational and institutional, not conceptual. Everyone agrees a steel mill cutting emissions is doing something useful.

Informational. To classify a project as transition, a lender needs the asset's current emissions, the projected emissions under the plan, and an assessment of the plan's credibility. The first is measurable, though often not measured. The second is a forecast, and forecasts of industrial emissions are unreliable. The third is a judgment about management competence and commitment—the kind of judgment credit analysts make about financial performance but not about decarbonization.

The information problem compounds because the transition plan is typically the borrower's own document. The lender has no independent way to verify the baseline or the projection. Third-party verifiers exist—the Climate Bonds Initiative, Sustainalytics, the large accounting firms—but their methodologies differ and their fees shut out smaller borrowers. A 2022 survey by the International Capital Market Association found that only 38 percent of transition-labeled bonds included a quantified baseline and a quantified target. The rest relied on qualitative descriptions.

Institutional. The actors holding the capital and the actors holding the technical knowledge are different people. A bank's credit committee understands leverage ratios and cash flow coverage. It does not understand the difference between a blast furnace and a basic oxygen furnace, or why a direct reduction plant running on green hydrogen is a different asset class from one running on natural gas. That knowledge sits with the engineering firm, the equipment vendor, the industry consultant. The capital sits with the bank. The translation between them is the missing piece—the same missing piece Chapter 2 described in technology lending.

The taxonomy tries to solve the institutional problem by making classification rule-based. If the list is clear enough, a credit officer can check a box without understanding the metallurgy. The list is never clear enough, because the technology keeps changing and the boundary cases keep multiplying. The EU taxonomy has been amended repeatedly since 2020, and each amendment was a political fight. The fights are not about science. They are about which industries get cheaper capital.

3.1.4 The transition finance spectrum

The boundary is better understood as a spectrum with four zones.

Zone 2 is where the action is, and it has the weakest institutional infrastructure. Zone 1 has a taxonomy, a bond market, and an established investor base. Zone 3 has exclusion lists and divestment campaigns. Zone 4 has emerged only as physical climate risk has become priced. Zone 2 has frameworks, but they are not yet standard, and capital is not flowing at scale.

The scale of the gap deserves plain numbers. The International Energy Agency estimates that reaching net-zero by 2050 requires annual clean energy investment in emerging market and developing economies to rise from roughly $150 billion in 2020 to $1 trillion by 2030. Most of that increase must come from the private sector, and most must fund Zone 2 projects—retrofits, fuel switching, efficiency upgrades—because that is where the emissions are. Industry accounts for about 31 percent of global emissions, and most industrial assets operating today will still be operating in 2050. The transition is less about building new green things than about changing the things that already exist. That is a different financial problem from the one green finance was designed to solve.

3.2 The Financial Ecosystem for Transition: Banks, Insurers, Investors, and Intermediaries

3.2.1 The ecosystem as a set of functions

The financial system does not fund the transition as a single actor. It funds it through five functions, each performed by a different type of institution, each with its own constraints.

1. Identification. Assessing which projects are transition projects, and how much capital they need.
2. Risk absorption. Bearing the risk that the project fails or the transition plan does not deliver.
3. Maturity transformation. Converting short-term savings into long-term capital.
4. Price discovery. Setting the cost of capital for transition projects relative to brown projects.
5. Monitoring. Tracking whether the transition plan is being executed.

No single institution performs all five well. Banks are good at identification and monitoring, but constrained by deposit maturity and capital requirements. Insurers are good at risk absorption, but constrained by solvency regulation and by the difficulty of pricing transition risk. Capital markets are good at maturity transformation and price discovery, weak at monitoring. Intermediaries—rating agencies, verifiers, consultants—are good at identification, but paid by the issuers they assess, which creates a conflict of interest.

The ecosystem works when the functions connect. It fails when they do not.

3.2.2 Banks

Banks are the largest source of external finance for firms in most economies, and the natural starting point for transition finance. A bank already has a relationship with the steel company, its financial statements, and a credit officer who visits the plant. Adding an environmental assessment to that relationship is, in principle, cheaper than building a new channel from scratch.

Three constraints get in the way.

Capital. A corporate loan to a steel company carries a 100 percent risk weight under the Basel III standardized approach, or whatever the internal ratings-based model produces. A green loan to a solar developer may qualify for a lower risk weight under some jurisdictions' green supporting factors, though the justification is contested; the European Banking Authority's 2021 report on ESG risk management warned that green supporting factors could create a capital shortfall if they understate actual risk. Banks get little capital benefit from green lending and none at all from transition lending. The capital rules treat a transition loan like any other corporate loan.

Maturity. A steel retrofit pays back over seven to ten years. A bank's deposit base matures in months. The bank can lend long, but it must hold capital against the mismatch and fund the loan with wholesale funding when deposits fall short. In a tightening cycle, wholesale funding gets expensive and long-dated transition loans lose their appeal. Banks therefore prefer to originate transition loans and distribute them to institutional investors—but the distribution market for transition loans is thin.

Classification. The credit committee needs a rule to decide whether a loan is a transition loan. If the rule is the EU taxonomy, the steel retrofit does not qualify. If the rule is the bank's own framework, the committee needs training to apply it, and most banks have not trained their credit committees to assess decarbonization plans. A 2023 survey by the Institute of International Finance found that 62 percent of banks had a climate risk assessment process, but only 29 percent had a process for assessing the credibility of a corporate transition plan.

Some banks have built the capability. ING's Terra approach, launched in 2022, uses sector-specific pathways to assess whether a corporate client's strategy aligns with a 1.5°C scenario. It requires the bank to collect emissions data from the client and model future emissions under different scenarios. It is expensive, and it works best for large clients in heavy industry—exactly where the transition capital is needed. Whether it scales to smaller clients is open.

3.2.3 Insurers

Insurers perform two functions in the transition ecosystem: they absorb risk, and they price it.

Risk absorption. A transition project carries technology risk (the new process does not work as expected), construction risk (the retrofit is delayed or over budget), and performance risk (the emissions reduction falls short of projection). Insurance can cover some of this. Performance guarantees from equipment vendors cover technology risk. Construction all-risk policies cover construction risk. Parametric insurance—which pays when a measured variable crosses a threshold—can cover performance risk if the variable is measurable and the threshold well-defined.

The industry has been slow to develop products for transition projects, for a structural reason. An insurer prices risk by pooling many similar exposures. A transition project is, by definition, one that has not been done at scale before. There is no pool. The insurer must price from first principles, which means a high premium, which raises the project's cost of capital. The insurance meant to enable the project becomes a barrier to it.

Risk pricing. Insurers are also large institutional investors, holding corporate bonds, infrastructure debt, and equity. When an insurer prices a corporate bond, it is implicitly pricing the issuer's transition risk. A steel company with a credible transition plan should carry a lower spread than one without. In practice, insurers have been slow to build transition risk into credit assessments, partly because the data is poor and partly because regulators have not required it. The European Insurance and Occupational Pensions Authority's 2021 opinion on climate risk did require insurers to assess transition risk, but the methodologies are still being developed.

The insurance function is most developed in agriculture and forestry, the subjects of Chapters 8 through 10. Weather index insurance, forest insurance, and mutual risk-sharing arrangements are the prototypes of the risk-buffering function the transition needs. The lesson from those markets: insurance works when the risk is measurable, the pool is large enough, and the government provides a backstop. For industrial transition, none of the three conditions is yet met.

3.2.4 Institutional investors

Institutional investors—pension funds, sovereign wealth funds, asset managers, insurance companies—hold the long-term capital transition projects need, and they have the longest time horizons. They are the natural holders of transition risk.

They are not set up to monitor transition plans. A pension fund allocates across asset classes and hires external managers to select securities. The manager is evaluated on quarterly performance against a benchmark. A transition project with a seven-year payback and an uncertain emissions outcome does not fit that cycle. The manager who buys the transition bond and holds it will underperform the benchmark for the first three years and may be fired before the project delivers.

The European Banking Authority's 2019 report on undue short-term pressure from the financial sector on corporations found that quarterly reporting and short-term remuneration structures did create pressure on companies to prioritize short-term results. It recommended that regulators consider whether the pressure was excessive. It did not recommend abolishing quarterly reporting, because the evidence that short-termism reduces long-term investment is mixed.

For transition finance the sharper problem is that the benchmarks themselves are brown. A corporate bond index is weighted by market value, and the market value of high-carbon issuers is large. A manager tracking the index is structurally overweight brown assets. Going green means deviating from the benchmark, which creates tracking error, which creates career risk. The partial solution that has emerged is the transition index—an index that includes brown companies with credible transition plans, weighted by the pace of their decarbonization. The first such indices launched in 2021. Their track records are too short to evaluate.

3.2.5 Intermediaries

The intermediaries—rating agencies, verification bodies, consultants, data providers—are the connective tissue. They perform the identification function, and they are paid by the issuers they identify. That is the same conflict of interest the credit rating agencies carry, with the same potential for failure.

The major agencies have moved into transition assessment. Moody's, S&P, and Fitch all offer second-party opinions on green and transition bond frameworks, assessing whether the framework is credible and whether the issuer's transition plan is consistent with it. The opinion is not a credit rating. It does not affect the bond's risk weight or its eligibility as central bank collateral. It affects marketability, because some investors require a second-party opinion before buying a green or transition bond.

The verification bodies use different methodologies. The Climate Bonds Initiative's standard is the most widely used for green bonds: a sector-specific criteria document for each eligible sector, plus mandatory third-party verification. Its transition criteria are thinner. The 2023 version of the standard includes a transition designation for issuers not yet green but with a plan, and the criteria for the plan are qualitative.

A concrete example shows how much the methodology matters. When the Climate Bonds Initiative assessed a Southeast Asian utility's transition framework in 2022, its reviewers worked from the utility's own emissions inventory, checked the inventory against the utility's fuel-consumption records, and then tested the retirement schedule for the utility's coal units against the IEA's Southeast Asia net-zero scenario. The framework passed on the retirement schedule and failed on scope: the utility had excluded emissions from two jointly owned plants. The revised framework added them, and the bond's projected proceeds fell by roughly 15 percent. That is what verification looks like when it is done at the asset level—and it is expensive enough that few issuers below investment grade can afford it.

The data providers are the weakest link. Transition assessment requires emissions data at the asset level. Most companies report at the corporate level, and the corporate figures are often estimated rather than measured. The Science Based Targets initiative's 2023 progress report found that only 38 percent of companies with approved targets had reported progress against them. The data gap is a governance problem, not a software problem: companies do not want to report asset-level emissions because the numbers are embarrassing, and regulators have not required them to.

3.2.6 The ecosystem as a chain

The functions connect in a chain. Identification produces a classification. The classification determines the risk weight and the eligible investor base. Those determine the cost of capital. The cost of capital determines whether the project proceeds. Monitoring feeds back into identification, because actual emissions performance updates the assessment.

The chain is only as strong as its weakest link, and the weakest link is identification. The taxonomies are incomplete, the transition frameworks voluntary, the data poor. Transition projects are therefore mispriced. Some are priced as brown, which makes them unviable. Some are priced as green, which means the investor is not compensated for transition risk.

flowchart TD
    A[Project sponsor] -->|seeks capital| B[Banks]
    A -->|seeks capital| C[Capital markets]
    B -->|originates loan| D[Identification: taxonomy + transition framework]
    C -->|issues bond| D
    D -->|classification| E[Risk weight / investor eligibility]
    E -->|cost of capital| F[Project proceeds or stalls]
    F -->|emissions outcome| G[Monitoring]
    G -->|feedback| D
    B -->|risk transfer| H[Insurers]
    C -->|risk pricing| H
    H -->|risk absorption| F
    I[Intermediaries: raters, verifiers, data] -->|assessment| D

Figure 3.1: The transition finance chain. What the diagram adds to the text is the feedback edge from monitoring back to identification: a green bond is assessed once, at issuance, and never again, which is why the loop is drawn as weak. A transition loan assessed at origination and not updated leaves the lender holding a stale classification for the life of the loan.

3.3 The Need for an Integrated Approach: Why Single Instruments Fail and Integration Is Key

3.3.1 The single-instrument fallacy

Policy discussion about green finance is organized around instruments: green bonds, green loans, green insurance, green venture capital. Each has its own advocates, standards, and evaluation literature. The implicit assumption is that scaling each instrument up will finance the transition.

Each instrument addresses one function in the chain, and the functions are complementary. Scaling one without the others produces a bottleneck.

Green bonds reached $500 billion in annual issuance in 2023, up from $11 billion in 2013. The market is concentrated. Energy, buildings, and transport account for about 80 percent of issuance; heavy industry for less than 5 percent. Heavy industry does not lack capital needs. Its projects do not fit the green bond framework, because they are transition projects. Scaling green bonds does not help the steel mill.

Green credit in China reached 30 trillion yuan in 2023, the largest balance in the world, concentrated in green transportation, green energy, and energy efficiency. Industrial transition is a small share, for the same reason: the green credit catalogue does not classify transition projects as green, and credit committees lack the tools to assess them.

Weather index insurance in China covered 120 million hectares in 2022, with payouts of 40 billion yuan. The instrument works for agriculture, where risk is measurable and the pool is large. It does not work for industrial transition, where risk is project-specific and the pool is small.

Each instrument works for the projects that fit its design. The projects left out are disproportionately transition projects—retrofits, fuel switches, efficiency upgrades—because those are hardest to classify.

3.3.2 The integration argument

The instruments must be connected because the risks they address are connected. A transition project carries technology risk, market risk, construction risk, and policy risk. No single instrument covers all four. The project needs a package: a loan to fund construction, an insurance policy for construction risk, a performance guarantee for technology risk, a long-term offtake agreement for market risk. The package is what makes the project bankable; the individual instruments are components.

The same logic holds at the level of the financial system. Identification (rating) produces the information risk absorption (insurance) needs to price the risk. Risk absorption reduces the risk the catalysis function (venture capital, government funds) must bear. Catalysis provides the equity that makes the debt financeable. The three functions form a chain, and the chain is the unit of analysis.

3.3.3 What integration looks like in practice

A financed project makes the abstract argument concrete. In Wanzhou District, Chongqing, 234 specialized cooperative economic organizations had 78,500 members by the end of 2009—a farmer participation rate of 21.5 percent—and drove 107,400 farmer households, 29.4 percent of the district total. The cooperatives grew citrus and faced a familiar problem: farmers could not get credit because they had no collateral and no credit history.

The solution had three parts. The district government built a credit rating system for the cooperatives from production data, sales records, and member repayment history—not a conventional credit score but a translation of the cooperative's private information into a form a bank could use. It introduced a weather index insurance product for citrus that paid out when rainfall or temperature crossed a threshold, protecting farmer revenue against weather shocks and reducing the bank's risk. And it set up a risk compensation fund that guaranteed a portion of the bank's loss on default, absorbing the residual risk the insurance did not cover.

The parts worked together. Rating without insurance would have identified creditworthy farmers who still could not repay after a drought. Insurance without rating would have protected farmers without helping them borrow. The guarantee fund alone would have absorbed losses without improving underlying credit quality.

Wanzhou is small and agricultural, not industrial. The structure is the same one industrial transition needs: rating identifies the project, insurance buffers the risk, catalysis supplies the equity or guarantee that makes debt financeable.

3.3.4 The counter-argument

Integration is expensive, and the transaction costs may outweigh the benefits. A large project can afford a rating, an insurance policy, and a guarantee. A small one cannot. The fixed cost of the package falls hardest on small and medium enterprises, the majority of firms in most economies.

The objection has force. Wanzhou worked because the district government subsidized the rating and the insurance; without the subsidy the fixed cost would have exceeded the benefit for most cooperatives. A small steel fabricator cannot afford a bespoke transition assessment whose cost is a significant fraction of the project.

The fixed cost is a function of current technology. Rating models, insurance products, and guarantee structures are all bespoke, built project by project, which is why they are expensive. The path to scale is standardization: standard rating models per sector, standard insurance products per risk, standard guarantee structures per project type. Standardization lowers the marginal cost and makes the package available to smaller projects.

Standardization is happening slowly. The EU taxonomy is a standard. The Climate Bonds Initiative's sector criteria are standards. The Science Based Targets initiative's sector pathways are standards. They are incomplete and contested, but they are the raw material for a standardized package.

3.3.5 The policy implication

Instruments should be evaluated as a portfolio, not individually. A government that subsidizes green bonds but not transition insurance subsidizes one link and leaves the others weak. A government that sets a green taxonomy but does not train credit officers to use it sets a standard no one applies. A government that guarantees green loans without requiring emissions monitoring guarantees a loan that may not deliver the transition.

A portfolio approach needs a coordinating body. In China, the People's Bank of China has played that role through its green finance evaluation scheme and its carbon emission reduction facility. In the EU, coordination is split among the Commission (taxonomy), the EBA (prudential rules), and the ECB (supervision), which makes coordination harder; the EU's transition finance framework has developed more slowly than China's despite the EU's earlier start on taxonomy.

This is the standard problem of financial regulation, organized by sector and by instrument. The transition requires a cross-sector approach, and the regulatory architecture is not built for it. The next chapter takes up the measurement problem underneath the coordination problem: the green premium, and why it is the diagnostic metric the chain needs.

3.4 The Conceptual Map

3.4.1 A map, not a taxonomy

Two axes. The horizontal axis is the asset's current emissions intensity, from zero (green) to high (brown). The vertical axis is the credibility of the transition plan, from none to high—a plan with a quantified baseline, a quantified target, and independent verification. Four quadrants follow.

The map is diagnostic, not prescriptive. It does not tell you which projects to finance. It tells you where a project sits and therefore which instruments it needs. A Zone 1 project needs a green bond and a standard green loan. A Zone 2 project needs a transition loan, a transition bond, an insurance policy, and a guarantee. A Zone 3 project should be excluded or priced at a penalty reflecting its risk. A Zone 4 project needs a green loan with a transition-risk premium, because the asset is green today but may not be tomorrow.

The map's value is that it makes the instrument choice explicit. The policy conversation often treats "green finance" as one category and asks how to scale it. The map shows four different problems inside that category, with instruments that are not interchangeable.

3.4.2 The map applied

The Hebei mill has a current emissions intensity of 2.0 tons of CO₂ per ton of steel—high—and a transition plan to retrofit to electric-arc capacity, with a quantified baseline (2.0), a quantified target (1.2), and third-party verification. It sits in Zone 2 and needs transition finance.

The bank's green credit quota does not help, because the mill is not green. The green bond framework does not help, because the mill is not an eligible green project. A transition bond framework could help, if the bank has one. The insurance market could help, if a performance guarantee exists for electric-arc technology. A guarantee fund could help, if the government has one for industrial retrofits. The mill is bankable in principle. The Zone 2 instruments are underdeveloped: the transition bond market is small, performance guarantees are bespoke, guarantee funds are rare.

A solar farm sits in Zone 1. Its emissions intensity is zero, its transition plan is not applicable, and it can raise capital at a lower cost than the steel mill. Its constraints are grid connection and land availability, which are not financial problems.

The financial system has built a deep, liquid, standardized market for Zone 1 projects and no comparable market for Zone 2. The transition's capital needs are mostly in Zone 2.

3.4.3 The map and the chain

Zone 1 projects need identification (a green taxonomy) and price discovery (a green bond market). They do not need the risk-buffering function, because technology risk is low and policy risk manageable. Zone 2 projects need all three functions, because technology risk, policy risk, and market risk are all high.

That is why the chain organizes the book. The chain is the set of functions Zone 2 projects need: rating identifies, insurance buffers, catalysis supplies equity. The three links are the minimum viable package for a transition project.

The map also shows why single-instrument scaling fails. A larger green bond market does not help Zone 2, because Zone 2 projects are not eligible. A larger green credit quota does not help Zone 2, because Zone 2 projects are not classified as green. A larger green insurance product does not help Zone 2, because the product is designed for agriculture. The instruments scale in Zone 1, where the need is smaller, and stay absent in Zone 2, where the need is larger.

The rest of the book develops the chain. Chapter 4 takes up the green premium, the price signal the chain is supposed to produce. Chapters 5 and 6 take up identification, with a focus on rating models for technology firms. Chapters 7 through 10 take up risk buffering, with a focus on insurance and mutual risk-sharing. Chapters 11 and 12 take up catalysis, with a focus on venture capital and government funds. Chapters 13 and 14 take up integration, with a focus on digital infrastructure and practice. Chapter 15 takes up policy.

3.5 Objections and Limits

3.5.1 The definitional objection

The three-way distinction between green, transition, and brown finance is arbitrary, the objection runs. A steel mill cutting emissions by 40 percent does more for the climate than a solar farm producing the same electricity, because the mill is reducing a larger baseline. Why should the solar farm get cheaper capital?

Correct on the physics, wrong on the finance. The solar farm gets cheaper capital because its risk is lower. Its output is predictable, its technology mature, its policy support stable. The steel retrofit is uncertain, its technology new, its policy support contested. The cost of capital reflects risk, not climate benefit. If the policy goal is to maximize emissions reduction per dollar, the steel mill may be the better investment. If the goal is to minimize investment risk, the solar farm is. The financial system is organized around the second goal.

The policy response is to subsidize the steel mill's risk, not to relabel it as green. A transition finance framework offering a guarantee or a concessional loan to the mill is more honest than a green label that misrepresents the risk. The label matters less than the risk transfer.

3.5.2 The additionality objection

Transition finance is not additional, the objection runs. A steel company that was going to retrofit its furnace anyway does not need a transition label. The label is a windfall, not an incentive: cheaper capital for a project that would have been financed regardless.

The evidence is mixed. A 2022 study by the Bank for International Settlements examined green bond pricing and found the green premium small—2 to 8 basis points—and concentrated in the primary market, where issuers choose the timing of issuance. The study concluded the premium was more consistent with a demand effect than a cost-of-capital effect. Investors wanted green assets and paid a little more for them; the premium did not change project viability.

The additionality problem is real but not unique to transition finance. It is the standard problem of any subsidized financial instrument. The solution is design: tie the subsidy to a behavior that would not have happened otherwise. A transition loan requiring the borrower to meet a quantified emissions target, with a penalty if the target is missed, is more additional than one requiring only a label. The penalty structure makes the subsidy conditional.

Calibration is the hard part. A target too easy is not additional; a target too hard is not bankable. The lender needs the borrower's baseline and the sector's technology frontier. That knowledge is the identification function, and it is why the chain starts with rating.

3.5.3 The greenwashing objection

Transition finance is a greenwashing loophole, the objection runs. A company that cannot qualify as green qualifies as transition, keeps its brown operations, and claims credit for a plan that may never be executed.

The regulatory response has been to tighten standards. The EU's 2023 proposal on transparency and integrity of ESG rating activities, and the 2024 corporate sustainability due diligence directive, both raise disclosure requirements for companies claiming transition status. The Singapore framework requires a quantified baseline, a quantified target, and independent verification. Japan's framework, developed by the Ministry of Economy, Trade and Industry, requires third-party review and annual updating.

Standards are tightening; enforcement is weak. A transition plan is a document, and a document can be written to satisfy a verifier without changing behavior. Verifiers are paid by the companies they verify, and the fee is small relative to the value of the label—the same conflict that produced the credit rating failures of 2008. The lesson from that failure is that verification must be backed by a liability regime. A verifier liable for a false statement has an incentive to be careful; a verifier with no liability has an incentive to be accommodating.

No jurisdiction has imposed a liability regime on transition plan verifiers. The EU due diligence directive imposes liability on companies, not verifiers. The Singapore framework requires verification but does not specify the verifier's liability. That gap is the most important unresolved issue in transition finance, and Chapter 12 returns to it.

3.5.4 The limits of the map

The map simplifies. It treats emissions intensity as a single number, when a project has multiple environmental dimensions—carbon, water, land, biodiversity—and a project good on one may be bad on another. A hydro plant is low-carbon and high-impact on rivers. A biomass plant is renewable and may compete with food production. The EU taxonomy's "do no significant harm" principle tries to handle the multi-dimensional problem, but the trade-offs are not commensurable.

The map also treats the transition plan as binary—credible or not—when credibility is a matter of degree. A quantified baseline and target beat a qualitative target; independent verification beats self-verification. The four quadrants are a coarse grid, useful for exposition and insufficient for pricing. Pricing needs a finer grid, which is what the rating models in Chapters 5 and 6 provide.

The final limit is that the map is static. A project moves across it as it decarbonizes. A Zone 2 project that delivers its plan becomes Zone 1. A Zone 1 project whose technology becomes obsolete becomes Zone 4. The financial system must track the movement, which means periodic re-assessment. A green bond is assessed once, at issuance, and not again. A transition loan is assessed at origination, and the assessment is rarely updated. Static assessment is a design flaw, and it is why the chain's feedback loop is weak.

3.6 The Chinese and European Pathways

3.6.1 Two starting points

China started with a credit-based system and built the taxonomy around the banking system's needs. Europe started with a disclosure-based system and built the taxonomy around the capital markets' needs. The two converge, but they are not the same.

China's system launched in 2015, when the People's Bank of China announced a green bond catalogue and a green credit statistical system. It was designed to serve the banking sector, which dominates China's financial system. The green credit catalogue classified projects into categories, banks reported green credit balances to the regulator, and the regulator used the reports to evaluate the banks, with the evaluation affecting access to central bank refinancing. The system was top-down, quantitative, and tied to bank incentives.

It moves capital quickly. The green credit balance grew from 8 trillion yuan in 2016 to 30 trillion yuan in 2023, driven by the banks' response to the regulator's evaluation. It is a classification system, not a risk assessment system. A project isgreen if it is on the list, and the bank's credit assessment is separate from the green classification. A bank can classify a loan as green and still price it as a conventional loan. The classification does not change the risk weight, and it does not change the credit committee's assessment.

Europe's system launched in 2018 with the Commission's Action Plan on Financing Sustainable Growth; the taxonomy regulation followed in 2020. It was designed to serve the capital markets, a larger share of Europe's financial system than in China. The taxonomy defines what counts as sustainable, and the disclosure regulation requires companies and financial products to report alignment with it. The system relies on investors to use the disclosure to allocate capital.

It is rigorous. The taxonomy's technical screening criteria are detailed, and "do no significant harm" is a real constraint. It is also slow. Disclosure requirements phase in over several years, the data is incomplete, and a minority of companies report taxonomy alignment—often as estimates.

3.6.2 The convergence

Both systems are converging on transition finance, driven by the same problem: neither has a framework for Zone 2 projects.

China extended the green credit catalogue to include transition categories. The 2021 revision added "clean utilization of fossil energy," covering projects that reduce the emissions of coal-fired power plants. The category is controversial because it lets coal projects be classified as green. The regulator's justification is that the reductions are real relative to baseline. Critics reply that the label lets the plant keep operating and that the reduction is insufficient for net-zero.

Europe's response has been a separate transition finance framework, still in draft. The Commission's 2023 proposal would let a company label a bond transition if the company has a credible transition plan and the proceeds fund a transition project. It is more restrictive than China's because it requires a company-level plan, not just a project, and more demanding because it requires the plan to be verified.

The convergence is incomplete, and the differences matter for cross-border capital. A bond labeled transition in China may not be labeled transition in Europe, and the reverse. The China-EU Common Ground Taxonomy reconciles the green categories but does not cover transition. Transition categories are the next frontier, and reconciliation will be harder there, because that is where the political stakes are highest.

3.6.3 The lesson

The institutional starting point determines the framework's design, and the design determines its blind spots. China's credit-based system mobilizes capital well and assesses risk poorly. Europe's disclosure-based system assesses risk well and mobilizes capital poorly. The transition needs both.

The chain combines the credit system's mobilization capacity with the capital market's assessment capacity. The rating link provides assessment. The insurance link provides risk buffering. The catalysis link provides mobilization. The chain is not a copy of either system; it is an assembly of functions each system performs partially.

3.7 The Instruments, Revisited

3.7.1 A taxonomy of instruments

The instruments can be classified by function and by zone.

The asymmetry is the empirical claim. Zone 1 has instruments in every function. Zone 2 has instruments in every function, but less developed. Zone 3 has instruments only in identification, because policy is to exclude, not to finance. The gap is the missing Zone 2 instruments, and the missing instruments are what the chain supplies.

3.7.2 The instrument mix

The mix for a Zone 2 project is a package with four components.

Identification. A transition rating assessing the project's baseline, target, and credibility. The rating is the input to the other components because it determines the risk they must absorb.

Risk absorption. An insurance policy or guarantee covering technology and performance risk, priced off the rating, with the premium a function of assessed risk.

Catalysis. Equity or a concessional loan covering the residual risk insurance does not, functioning as first-loss capital that makes debt financeable.

Debt. A loan or bond providing the bulk of the capital, priced off the rating, the insurance, and the catalysis, and cheaper than it would be without them.

Rating identifies, insurance buffers, catalysis catalyzes, debt finances. The package is the unit the financial system needs to build.

3.7.3 The missing market

The missing market is the market for the package. The components exist but are sold separately. A borrower gets a rating from one provider, insurance from another, a guarantee from a third, a loan from a fourth. Transaction costs are high and coordination poor. Borrowers often do not know which components exist; providers do not know which borrowers need them.

This is a market-design problem, and it is the kind digital infrastructure can solve. A platform connecting rating providers, insurers, guarantee funds, and lenders would lower transaction costs and raise take-up. Chapter 13 takes up the platform. It is not a panacea: it does not solve the definitional, additionality, or greenwashing problems. It solves coordination, which the current ecosystem handles worst.

3.8 The Mill, the Spread, and What Comes Next

The Hebei mill is still waiting. Its 1.4 billion yuan retrofit would cut emissions 40 percent and pay back in nine years, and the bank's credit committee still has no box to tick. Run the numbers on the mispricing. If the mill borrows at a conventional corporate rate of roughly 5.5 percent instead of a transition rate near 4.2 percent—the gap the green premium implies for a large issuer, scaled to a transition asset—the interest cost over the nine-year payback rises by about 160 million yuan. That is more than a tenth of the project cost, and it is enough to push the internal rate of return below the company's hurdle rate in most of the scenarios the finance director has run. The project does not fail because capital is scarce. It fails because the spread is wrong.

Three things follow from the chapter's argument, and they are concrete.

The identification problem is a data problem before it is a standards problem. The 38 percent figure—transition-labeled bonds with a quantified baseline and target—measures how few issuers can document what a credit committee would need to see. Fixing the taxonomy without fixing asset-level emissions reporting leaves the credit officer where he started.

The package is not a metaphor. It is a set of contracts with prices attached: a rating fee, an insurance premium, a guarantee fee, a loan spread. Each price depends on the one before it. The rating determines the premium; the premium and the guarantee determine the spread. When the package is assembled project by project, each fee carries the fixed cost of a bespoke negotiation, and the sum is why small borrowers are shut out. Standardization is the only route to a lower sum.

The Chinese and European systems will keep diverging on transition categories until a common transition taxonomy exists. Until then, a Chinese transition bond sold to European investors carries a labeling risk that shows up in the spread, and the mill pays for it.

The next chapter takes up the green premium, the price signal the chain is supposed to produce. If the premium is zero, the chain is not working. If the premium is large, the chain is working, and the cost of capital for brown assets is rising. Measuring the premium is the test of the chain.

Note on sources. The chapter draws on the EU Taxonomy Regulation (2020); the People's Bank of China green bond catalogue (2015, revised 2021); the China-EU Common Ground Taxonomy (2021); the Monetary Authority of Singapore transition finance framework (2023); the International Energy Agency's net-zero investment estimates; the European Banking Authority's 2021 report on ESG risk management and its 2019 report on undue short-term pressure; the Institute of International Finance's 2023 survey of bank climate risk practices; the Bank for International Settlements' 2022 study of green bond pricing; the Science Based Targets initiative's 2023 progress report; the Climate Bonds Initiative's 2023 standard and its 2022 assessment of a Southeast Asian utility transition framework; and the Wanzhou District cooperative data from the source materials. The ING Terra approach and the China Baowu 2021 transition bond are cited in the text. Figures are approximate and drawn from the cited sources.