Response: The Limits of ESG in Assessing Nonprofit Control

We appreciate Schröder and Thomsen’s thoughtful response and valuable empirical study exploring ESG performance among foundation-owned firms. This topic is timely, as nonprofit control is under stress both in the U.S., with the ongoing governance debate surrounding OpenAI, and in Europe, with recent turmoil at Novo Nordisk—one of the world’s leading pharmaceutical companies and arguably the crown jewel of the enterprise foundation model (Bansal, 2025). Following recent underperformance by the for-profit pharmaceutical company, the Novo Foundation has intervened reportedly to accelerate the CEO’s succession primarily out of concern that the company has failed to respond to competition and generate sufficient profits.

This unusual governance intervention by the nonprofit—typically passive in its oversight—is not driven by social purpose, but by the risk that the for-profit will fail to produce the kind of cash flows that have sustained the foundation’s substantial philanthropic giving. The governance shift at Novo Nordisk reinforces our definition of the income-generating model of nonprofit control. This view of enterprise foundations as primarily cash-generating entities is even more evident in companies that sell global consumer brands, such as IKEA and Carlsberg. The Carlsberg Foundation’s charter includes numerous charitable goals in support of science and the arts, while its only arguable commitment to responsible business is to developing the art of making beer and keeping the brewing of beer on a high and honorable level (See 2025 Charter).

Ultimately, then, the question turns on whether ESG scores reflect a deeper social purpose embedded in the business itself. While Schröder and Thomsen’s careful study provides valuable evidence, we would like to highlight several important caveats and urge caution in interpreting their findings.

First, their main sample comprises 69 firms, many of which are concentrated in relatively wealthy European countries. The control group, which is based on industry and assets size, is susceptible to concerns about unobserved quality that correlates with ESG. Foundation-owned firms tend to be high-performing companies (hence their income-generating function), and their superior ESG scores may reflect this underlying strength. Although the authors apply controls and evaluate ESG performance during crisis, these methodological constraints limit the ability to draw strong causal inferences.

Second, ESG scores are, at best, a crude proxy for true social or environmental impact. Many studies have highlighted the inconsistencies, biases, and methodological weaknesses of ESG ratings (Berg et al., 2022; Christensen et al., 2022). For example, employee satisfaction at Carlsberg, based on Glassdoor reviews, ranks lower than at competitor Heineken. Notably, the average ESG score reported for foundation-owned firms is about 57—above the control group’s average of around 47, but still well below the 75+ scores typical of global ESG leaders. This suggests mid-range, not leading, ESG performance. We should be cautious in making broad claims about lofty purposes based solely on ESG scores.

Third, even if the study captures ESG differences accurately, the results are fully consistent with our theoretical distinction between income-generating and socially-oriented nonprofit control. Many foundation-owned firms, such as Carlsberg or Novo Nordisk, primarily aim to generate long-term profits to support their nonprofit parents. Whether selling beer aligns with broader social health goals is debatable. ESG outperformance in these cases likely reflects responsible business practices driven by financial prudence, not prioritization of social impact over profit.

Indeed, we acknowledge in our own analysis that income-generating nonprofit-controlled firms may tend to act responsibly—or at least not irresponsibly. This aligns with Schröder and Thomsen’s observations. But responsibility in this context is typically a byproduct of long-term business strategy, not a signal of embedded social purpose (see e.g., Gadinis and Amiazad, 2020).

The real test for nonprofit control arises when firms pursue broad social goals while relying on outside capital. In such cases, governance may become more fragile as these objectives begin to collide. The OpenAI case—discussed in our original piece and elsewhere (see also Weiss, 2025)—illustrates how outside investor pressure can compromise mission integrity. As more tech and blockchain firms experiment with foundation-based models, the risk of greenwashing rises.

To be clear, we think the law should facilitate the income-generating model of nonprofit control. These firms run effective businesses and contribute to charitable goals. We have even advocated relaxing restrictive U.S. legal rules to encourage broader adoption. However, when nonprofit-controlled firms invoke expansive missions without channeling cash flows back to their nonprofit sponsors, greater scrutiny is needed.

In sum, Schröder and Thomsen offer useful data, but their conclusions should be interpreted with care. Foundation-owned firms may perform reasonably well on ESG metrics, but this does not necessarily reflect a deeper commitment to public purpose. To fulfill their potential, nonprofit control structures must be carefully designed and governed to guard against mission drift and reputational misuse.