Fig. 1. Grace Glueck, "The Guggenheim Cancels Haacke's Show," New York Times, April 7, 1971

A Response to Barabási and Shekhtman


PDF: Levine, A Response

Editors’ Note: This response is part of the final “Toward a More Inclusive Digital Art History” project generously funded by the Terra Foundation for American Art. Click on the Toward a More Inclusive Digital Art History tag for all contributions to the series. 

Albert-László Barabási and Louis Shekhtman’s article on the funding sources of American art may not deliver major revelations, but it begins to probe institutional funding structures in novel ways. Neither institutional critique (cf. Hans Haacke; fig. 1) nor probing analyses of art-world finances are new,1 nor are analyses of board overlaps.2 Even so, the attempt to use computational science—namely, network analysis—to understand how museums are supported is a worthy development.

Newspaper clipping with headline "The Guggenheim Cancel's Haacke's Show"
Fig. 1. Grace Glueck, “The Guggenheim Cancels Haacke’s Show,” New York Times, April 7, 1971

Although the data are incomplete and the consideration of museums with an express focus on American art is narrow, that there are no great new insights is itself a meaningful finding. Indeed, it is an important result that the authors validate art-world intuition: that there are winner-take-all effects in the museum world. Some of this phenomenon likely has to do with local economies, as suggested in the paper, but much of it is also undoubtedly attributable to other reasons (e.g., greater brand visibility for larger museums, enhanced by greater exhibitions budgets; more dollars spent on public relations; staff and resources to travel and meet new donors, etc.).

Observing, however, that big organizations have differential access to resources (enabling them to get bigger) is a separate consideration from whether these larger organizations have healthy financial models or are more successful in fulfilling their missions. The amount of money raised and the number of donors at any given institution certainly speak to the presence (or not) of effective fundraising practices; whether those metrics indicate an efficient operation or the delivery of impact is a different question. It is possible, for example, that the Terra Foundation for American Art’s grants to the New Britain Museum of American Art were more impactful to that institution than the foundation’s gifts to the Whitney Museum of American Art or the Crystal Bridges Museum of American Art. The foregoing is a strictly hypothetical question—the Whitney and Crystal Brides are both extraordinary institutions—but future analyses should understand the distinction between the size of a gift and its use/impact.

It is also important to note that, given the nature of the data as reported in the 990, the analysis conflates gifts of art, gifts for art, gifts for capital, and gifts for endowment with gifts for the operating budget. The authors cannot be faulted for the vagueness of the data, but it would be helpful to segment giving by type. Consider a simple case: a museum normally draws from its endowment at 4–5 percent of a 12–20-quarter rolling average. That means a $1 million gift will yield $50,000 (at a 5 percent draw) when vested; however, in the first year, it will yield only $10,000 (assuming both that the money is given all at once and that the museum uses a 20-quarter average draw formula). Consider, then, that a $50,000 gift earmarked to be spent down immediately could drive more impact in mission fulfilment in the short term than a $1 million endowment donation.

The larger donation, of course, has greater long-range impact and enhances sustainability, which leads us to a crucial point: one great benefit of a private philanthropic model is the ability to build up an endowment, which ideally allows for sustainable funding with minimal restrictions.

There are myriad downsides to a private philanthropic model, including the chance that large donors will hold outsized sway, and there are legitimate equity issues around how endowments get created. However, such critiques apply to any model with a strong concentration of funders, including one in which a government is a primary supporter. Just as Kenneth Arrow empirically demonstrated that there is no perfect voting system,3 for similar reasons there is no perfect support structure for institutions serving the public good.

With this in mind, there is another possibility lurking in the data, one that at least typifies my experience as the director of the Toledo Museum of Art, and that is that we should allow for the possibility that at least sometimes, and perhaps much of the time, capital follows a vision rather than sets it.

I appreciate all the good reasons and real-life case studies that exist for people to “follow the money,” but Barabási and Shekhtman’s argument—like many other forms of institutional critique—elides the structure of a museum’s donations with the culture of its governance. An institution’s management has agency, and while those leaders are ultimately accountable to the board, healthy boards let directors and their staffs lead. Similarly, ongoing grant funding from a foundation could represent medium- to long-term alignment of institutional interests and missions rather than just friendship networks among trustees. It would be interesting to know, for example, how many grants the Terra Foundation gave to the Whitney before any current individual trustee sat on the board. Perhaps the relationship between the foundation funder and the museum is longer-lived than with any individual.

Barabási and Shekhtman are data scientists, and their helpful contribution gives visible and digestible structure to phenomena many of us have intuited in the art world. The article, though, from the outset identifies itself as taking a critical posture. Critique is healthy, and it is warranted, but there are other, deeper interrogations of the data that are required to give a better sense of if money flows influence strategic decision-making as much as is supposed. I suspect the answer will be that in some cases it does and in others it does not. There are questions we can ask of the data to tease some of that out, and we as a field can and should do a better job of reporting textured data to help researchers like Barabási and Shekhtman fill in the gaps.

Cite this article: Adam Levine, “A Response to Barabási and Shekhtman,” in “Reflecting on ‘Toward a More Inclusive Digital Art History,’” special section, ed. Diana S. Greenwald, Panorama: Journal of the Association of Historians of American Art 9, no. 2 (Fall 2023),


  1. Andrea Fraser, 2016 in Money, Museums, and Politics (Cambridge, MA: MIT Press, 2018).
  2. Joel H. Levine, “The Methodology of the Atlas of Corporate Interlocks,” Bulletin de Méthodologie Sociologique, no. 17 (1988): 20–58.
  3. Kenneth Arrow, “A Difficulty in the Concept of Social Welfare,” Journal of Political Economy 58, no. 4 (1950): 328–46.

About the Author(s): Adam Levine is the Edward Drummond and Florence Scott Libbey President, Director, and CEO of the Toledo Museum of Art