ESG Scores Miss What Matters: Can Companies Adapt?

• ESG scores measure what companies disclose—not necessarily how they respond to disruption.
• For investment analysts, resilience may reveal risks traditional ESG analysis misses.
• The implication for long-term investors: move from sustainability attributes to demonstrated capabilities.

Consider two regulated utilities with near-identical ESG ratings. Both publish a net-zero-by-2050 target. Both score similarly on major disclosure benchmarks. Both receive similarly strong governance ratings. On paper, they are interchangeable.

Look closer and they are nothing alike:

One of these companies is adapting. The other is reporting. Most ESG scores struggle to tell them apart. They standardize what companies disclose; they do not test what companies can do under pressure.

That gap matters now more than it once did. Geopolitical rivalry is redrawing supply chains and energy systems. Artificial intelligence is rewriting competitive advantage within a single investment cycle. Deglobalisation, ecological thresholds and shifting industrial policy interact and compound, in what analysts have started calling a polycrisis (World Economic Forum, 2023). In this environment, judging companies by their attributes alone is increasingly costly. Sustainable investment needs a narrative fit for it.

This is where attribute-based assessment runs out of road. It measures the policies, targets, certifications and disclosures that companies possess. These attributes are real and often useful. But possession is not capacity. A company can hold every certificate and still freeze when its regulatory, technological, geopolitical or ecological environment shifts discontinuously.

Survival under disruption depends on a different property: capability. Strategy scholars describe such higher-order capacities as dynamic capabilities (Teece et al., 1997; Teece, 2007). For analytical purposes, I organize them around three dimensions:

It helps to keep three related words apart, because people use them loosely. A robust company resists: it absorbs a shock without changing. A resilient company adapts: it responds and adjusts. A transforming company renews: it remakes itself when the environment demands it. The framing draws on resilience ecology, which began by distinguishing resilience from stability (Holling, 1973), but the three-part formulation here is an applied distinction for company analysis. All three responses have their place. None alone is sufficient, and an ESG score cannot tell you of which a company is capable.

My argument, developed in a new working paper, is simple to state: long-term investors should shift their primary unit of sustainability assessment from attributes to capabilities. From what companies have to what they can do under pressure.

Research on tourism firms after the Canterbury earthquakes found that adaptive resilience, the post-disaster capacity to lead, collaborate and improvise, influenced financial performance directly; planned resilience, the pre-disaster stock of preparations, helped only indirectly, by feeding that adaptive capacity (Prayag et al., 2018). Firms with stronger social and environmental practices weathered the 2008 financial crisis with greater resilience and recovered faster (DesJardine, Bansal and Yang, 2019). And a fifteen-year study of manufacturing firms found that sustainability practices predicted lower financial volatility and higher survival rates, with the strongest effects precisely when the environment turned turbulent (Ortiz-de-Mandojana and Bansal, 2016).

The attribute side of the ledger, meanwhile, is not standing still, and not obviously improving where it counts. A study of more than 11,000 Russell 3000 sustainability reports found that disclosures grew less specific and less quantitatively dense as reporting became mainstream practice, even as the reports themselves grew longer (Kim et al., 2026). More disclosure does not automatically mean more decision-useful information. The finding does not center on any single company; it describes what happens when a reporting regime scales faster than the discipline behind it. It is a second reason, alongside the capability evidence, to treat what a company discloses as a starting point for analysis rather than the analysis itself.

I read the resilience findings as capability stories. Sustainability practices help companies endure not because a certificate wards off disaster, but because developing those practices builds the muscles of sensing, seizing and reconfiguring. I should be honest about the limits: these studies make the capability lens plausible, but they do not yet validate the attribute-versus-capability distinction directly. That test now has a shape. Build the two assessments independently of each other and see whether they sort companies differently, then see whether capability evidence predicts outcomes under disruption that attributes miss. If independently built assessments keep ranking companies alike when material disruption hits, my distinction does not hold.

Nor do they prove that capability-rich companies will outperform in every sector or market. The better-supported claim is narrower: capability evidence may help investors identify companies with lower downside risk, faster recovery, lower volatility and higher survival probability when structural disruption occurs. This is a thesis about risk, recovery and survival, not a generic promise of alpha.

Capabilities cannot be read directly from a score, and they should not be collapsed into a new composite score. But they leave tracks over time. A minimum viable assessment can begin with three kinds of evidence already available to analysts.

For sensing, compare a company's risk discussion with its peers. Specific, scenario-informed and company-specific language suggests that management is genuinely updating its view as conditions change. Language that closely mirrors sector peers is a red flag.

For seizing, follow the money, not the target. Does capital expenditure over three to five years actually move towards the stated transition plan? An ambitious target accompanied by flat capital allocation is not a transition plan. It is a contradiction.

For reconfiguring, look for a track record of business-model change rather than product tweaks. Incremental optimization amid acknowledged disruption tells you the company can describe its problem but not act on it.

Engagement conversations sharpen all three. Ask how the company monitors emerging risks and how that monitoring changes strategy. Ask how capital allocation has actually shifted in response to the risks it has named. Ask what business-model changes it is making and which trade-offs it accepts. Ask for a case where an internal warning changed a strategic assumption or an investment decision, because sensing fails when employees detect change but do not feel safe enough to voice it. Disclosures show what a company reports; these questions probe how it decides. The gap between the two is where the investment insight lives.

Engagement is most defensible here as a way to test and reveal capability. It may also support capability development, but only where the investor has credible access, influence and governance rights.

None of this scales easily by hand. AI can help, not by forecasting ESG scores, which would reproduce the attribute logic, but by surfacing evidence for analyst judgment: gaps between stated targets and actual capital expenditure, risk language that echoes sector peers, and disposal patterns that do or do not match a transition plan. Used this way, AI extends analyst judgment rather than replacing it.

Even a portfolio built entirely from capability-rich companies is not automatically resilient. If every holding senses, seizes and reconfigures well, but all share the same structural exposure, a single technology, a single geography, a correlated regulatory risk, the portfolio can still fail as a whole when that exposure turns against it. Resilience, like capability, has to be assessed one level up.

Three design principles do that work (Bril and Schramade, 2023):

None of this comes free. Redundancy looks like inefficiency against a benchmark that rewards tight tracking, and a board that has not had an honest conversation about what its mandate actually values will read every buffer as a mistake. That conversation, more than any investment technique, is usually the binding constraint.

One more test disciplines the whole exercise, at both company and portfolio level. Suppose the first utility's coal divestment simply sold the plant to a less scrutinised operator that runs it harder, while its grid investment raised tariffs on households least able to pay. Adaptive, yes. Sustainable, no.

The test I use asks three plain questions. Who bears any risk that a company's adaptation displaces? Through what channel does it travel, a sale, a price increase, a relocated supply chain? And does the transfer reduce risk across the system, or merely move it somewhere less visible? A capability assessment without this second axis becomes a search for strategic survivors, and long-term investors, whose portfolios ultimately depend on the stability of the systems to which that risk is being transferred, should mark that difference down before it shows up in prices.

None of this discards existing practice. Leading ESG analysis is already moving towards forward-looking assessment of transition credibility; my framework systematises that movement rather than inventing it. ESG data remain relevant diagnostic evidence about disclosed commitments, governance structures and reported sustainability performance: some of it records intent, some records action, and none of it substitutes for adaptive capacity. The shift is in what investors treat as the primary object of assessment. Attributes record what a company has adopted, disclosed or achieved. Capabilities, though formed by history, describe its demonstrated capacity to act when the next disturbance arrives. For a long-term investor, whose returns depend on conditions that do not yet exist, that orientation toward the future is not a refinement of the ESG approach. It is a correction of its central limitation.

The full working paper extends this logic across three investment horizons and five connected activities, from structural analysis and company assessment through portfolio construction to active ownership and measurement. The next stage of sustainable investment will belong to investors who can distinguish companies that merely report sustainability from those that have built the capabilities to sense structural change, seize emerging opportunities and reconfigure themselves when the environment demands it.

I am publishing this, and the working paper behind it, because sustainable investment needs a sharper question at its center: has this company built the capacity to act, not only the paperwork to describe its intent

Bril, H., & Schramade, W. (2023). Strengthening investment portfolios through resilience – A primer [Working paper]. SSRN.

DesJardine, M., Bansal, P., & Yang, Y. (2019). Bouncing back: Building resilience through social and environmental practices in the context of the 2008 global financial crisis. Journal of Management, 45(4), 1434–1460.

Holling, C. S. (1973). Resilience and stability of ecological systems. Annual Review of Ecology and Systematics, 4(1), 1–23.

Kim, H., Li, N., Feldman, R., Feldman, Y., & Liu, Y. (2026). What sustainability disclosures disclose (Coase-Sandor Institute for Law and Economics Research Paper No. 26-7) [Working paper]. University of Chicago Law School.

Ortiz-de-Mandojana, N., & Bansal, P. (2016). The long-term benefits of organizational resilience through sustainable business practices. Strategic Management Journal, 37(8), 1615–1631.

Prayag, G., Chowdhury, M., Spector, S., & Orchiston, C. (2018). Organizational resilience and financial performance. Annals of Tourism Research, 73, 193–196.

Teece, D. J. (2007). Explicating dynamic capabilities: The nature and microfoundations of (sustainable) enterprise performance. Strategic Management Journal, 28(13), 1319–1350.

Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18(7), 509–533.

World Economic Forum. (2023). Global risks report 2023.