Environmental attribute certificates are becoming a popular emission reduction tool for many companies. However, can this trading model genuinely drive substantial emission reductions across industries? Addressing this core question, the Green Finance 60 Forum (GF60), in collaboration with SINA Corp, engaged in a 1.5°C Talk dialogue with Professor Lisa Sachs, Director of the Columbia Center on Sustainable Investment. Professor Sachs argues that environmental attribute certificates reflect a widespread, deep-seated misconception in the global push to transition hard-to-abate sectors: we are substituting the pursuit of real-world, physical changes with sophisticated methodological designs.

Professor Lisa Sachs is the Director of the Columbia Center on Sustainable Investment at the Columbia Climate School and the Director of the Columbia University Master's Program in Climate Finance. Since joining the center in 2008, she has established and led its interdisciplinary research and advisory work, focusing on aligning investment laws, policies, and practices with sustainable development goals. She is globally recognized for her expertise in how laws, policies, and business practices influence international investment flows and promote sustainable development. She works closely with governments, regional and international development agencies, financial institutions, corporations, civil society organizations, and academic institutions to deeply understand the relationship between investment flows and sustainable development, aiming to improve investment policies and practices to achieve the SDGs and the goals of the Paris Agreement on climate change.

The following presents Professor Sachs's views in full:

Earlier this year, Microsoft announced a deal to purchase the environmental attributes linked to green steel produced by Swedish manufacturer Stegra. This move was widely praised as "innovative" and "market-shaping." However, the reality is that this transaction transferred "claims" without altering the production process itself; it highlights a more general issue plaguing efforts to scale the green steel market.

Microsoft did not procure green steel for actual use within the market where it was produced. Instead, it purchased a type of environmental attribute, allowing it to label steel procured or used elsewhere as "green steel," particularly in markets like China—where a significant portion of global steel is produced—where green steel production is not yet viable. The physical systems within Asian steel markets remained unchanged. No new green steel capacity was added, no electricity price risks were mitigated, and no financing constraints were eased. While this procurement might improve a corporate narrative on paper, it does not alter the underlying industrial reality it represents.

This scenario is not unfamiliar. Structurally, it follows the same logic as carbon offsets: a methodological arrangement enabling companies to claim progress without changing the emission trajectory of the systems they operate within. Similar to carbon offsets, the praise for such practices reveals a deeper issue: we have become accustomed to rewarding progress that is methodologically sound and quantifiable, rather than changes that genuinely transform the physical world.

In recent years, discussions around green steel have increasingly centered on standards, disclosure, book-and-claim systems, and buyer commitments. These tools are popular because they are easily identifiable, neatly fit into corporate net-zero targets and sustainability narratives, can be accounted for at the individual firm level, and allow companies to publicly announce "progress." The problem, however, is that they sidestep the more difficult, yet crucial, question: is steel production itself becoming cleaner, cheaper, and easier to finance in the locations where it actually occurs?

In short, these tools do not address the real constraints.

If the primary barrier to scaling green steel was merely a lack of channels for buyers to signal demand or unclear market definitions, the market would have already developed. This is not the case. What we observe, especially in Asia, is that green steel remains largely in the pilot stage, almost entirely reliant on corporate balance sheet financing, with project finance largely absent.

This distinction is critical. Industrial markets scale primarily through financing. When a project cannot secure financing on a standalone basis, it is not due to a lack of ambition but because the project risks and cash flows fail to meet the underwriting or credit thresholds of financial institutions.

The book-and-claim system does nothing substantial to change this. It decouples environmental attributes from physical supply, but it does not lower the cost of producing green steel, stabilize electricity prices, or reduce investment risks associated with hydrogen. It merely reallocates attributes after the fact, rather than making the next project easier to finance or cheaper to build.

Steel production does not follow the rapidly declining cost curve, or S-curve, seen in solar photovoltaics, to which it is often compared. Solar energy scaled because each additional panel installed directly expanded manufacturing scale, triggered learning effects, lowered unit costs, and enhanced competitiveness within the same markets where deployment happened. Purchasing solar panels altered the economics of future panel production.

Green steel does not operate this way. Its costs are dominated by electricity markets (price and volatility), capital intensity, and process configuration. Purchasing environmental attributes from a project in Sweden does not reduce electricity price volatility risks for hydrogen-based ironmaking in China, nor does it change grid access rules, dispatch constraints, or financing conditions. There is no "learning-by-claiming" effect that propagates across different markets.

Buyer commitments suffer from a similar misjudgment. They are often described as the "missing demand signal," but in practice, their role is far more limited. Buyer commitments can help stabilize project utilization once supply-side risks are mostly resolved; however, they cannot make projects with unviable cost structures feasible, nor can they persuade lenders to finance assets exposed to electricity price volatility and operational uncertainties.

This is not merely theoretical criticism. Across green steel and other hard-to-abate sectors, short-term, conditional, or commissioning-timeline-disconnected commitments have repeatedly proven ineffective at unlocking financing. Tools like buyer commitments truly make a difference only when project economics are already near market-clearing levels, not before.

So why do we persistently place hope in these tools?

Part of the reason is that we are deeply entrenched in an "individual entity perspective" framework. Standards, targets, disclosure, transition plans, and commitments all ask, "What can a single company achieve on its own?" They align naturally with corporate accountability systems but treat the political and institutional complexities of power markets, infrastructure, and industrial policy as external dependencies. Consequently, "progress" can be claimed even in the absence of systemic change.

But green steel is not a challenge at the individual company level; it is a systemic challenge. Systemic problems are not solved by tools designed for individual actors.

Prompted by the Microsoft deal, my colleagues and I were determined to identify the genuine constraints to scaling green steel in China. After梳理 numerous interrelated risks, constraints, and incentives, one issue emerged as paramount: hydrogen-based ironmaking requires massive, continuous electricity inputs over an asset lifespan of 20 to 30 years. Even where average renewable energy prices are low, industrial users in China still face price volatility, curtailment risks, congestion, and grid access uncertainties. The decisive risk is not whether electricity prices are "high," but whether they are "stable."

Of course, while volatility risk is the most critical, it is not the only one. It combines with technological risks, capital intensity, policy risks, and others, most of which currently must be absorbed by the steel companies themselves. This over-concentration of risk ultimately confines green iron production to cautious pilot stages.

Although these real project risks can be disaggregated, we seldom do so in practice. We often vaguely label unfinanceable projects as "high risk" without specifying which risks block financing, which could be bounded or pooled, and which render a project infeasible even with support. The result is a proliferation of blunt interventions—such as generic blended finance, broad subsidies, and symbolic demand aggregation—that improve perceptions without altering financing realities or production outcomes.

These issues were the central focus of a recent workshop we co-convened in Shanghai with Hang Lung Properties and GF60, explicitly aimed at advancing the green steel market in Asia. The discussion did not center on standards or pledges; it centered on economics.

We discussed a series of concrete proposals focused on how to bound and pool risks for green ironmaking to reduce the production costs of green steel. When costs decrease, economics improve, narrowing the cost gap with conventional steel; at that point, credible off-take commitments from buyers can effectively address remaining utilization risks, thereby creating genuine market demand.

The same challenge appears in other hard-to-abate sectors and industries that have yet to begin their transition. To date, industry and its partners have invested significant effort in designing methodological tools that "allow progress to be claimed" rather than focusing on addressing the systemic constraints that change physical reality.

These transitions, even the difficult ones, are achievable—provided we stop mistaking "claimability" for "actual change." The real starting point must be economics, risk, and finance. We are not short of tools to address these issues, and there remains considerable room for innovation. Until these core problems are resolved, we will continue to reward transactions that appear innovative while the underlying system remains largely unchanged.