Abstract
Authenticity is a perennial issue in art markets. This paper investigates the character of some art identifying it as meta-credence goods and utilizing a formal Bayesian model of how experts (or buyers) play a role in evaluating art works, as suggested in a recent paper by Ginsburgh et al. (J Econ Behav Organ 159:36–50, 2019). Experts or a consensus of experts determines credence status where present and future falsification (in the sense of Karl Popper) is impossible. Credence is always a matter of probabilistic degree, and we define a class of extreme credence goods called meta-credence. Consensus of expert opinion serves as verification in the art world where consensus substitutes for falsification. These opinions are relied on by buyers and the art-loving public. This paper outlines the process that the art expert undergoes to render a verdict on the authenticity of art.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Most often subjective expert opinion is the basis for deciding between authenticity and inauthenticity—so-called connoisseurship. There are many experts whose opinions are sometimes contradictory. A consensus of experts is often the basis for an individual patron’s decisions of whether to bid, and how much to bid. However, experts differ in their reputation for accuracy. An individual expert who has demonstrated more comprehensive knowledge of the artist’s style and portfolio may convince buyers of authenticity notwithstanding opposing opinions of most recognized experts. A buyer would likely employ a collection of experts and base his decision on their individual independent estimates. The alternative is self-assessment. In the present paper, we eschew the question of precisely how the buyer combines these independent evaluations into a decision to buy or not to buy or bid or not bid. If the buyer and the expert are one and the same, our approach and Ginsburgh et al.’s are closely related. Specifically, considering the same set of variables, Ginsburgh et al. look at their effect on price, while our model focusses on their effect on the probability that a painting is “right” (by the hand of the artist) or not. (Also see note 12).
The role of the expert in judging or providing standards of quality has been explored in the literature on cultural economics: see Ginsburgh and Ours (2003) relating to the role of prizes and their impact on future success of concert pianists entering a competition and Hilger et al. (2011) who studied expert opinion on wine purchases in Northern California finding in a controlled experiment that wines with expert opinion attached to them achieved a 25% increase in sales over a controlled sale. Also see Ginsburgh and Menger (1996).
Branding is a manner of providing information to consumers. A McDonnell’s “Big Mac” bought in Bangor, Maine, is known to be the same as the one purchased in Paris, France.
The market for a major characteristic of religion is an extreme example of the meta-credence good. The demand characteristic “assurances of eternal salvation” is clearly a meta-credence purchase (other characteristics are not) since no one has ever returned to verify the quality of that part of the good. Evidence must be provided or manufactured by suppliers at the point of sale.
Other examples are described in Ekelund and Thornton (2019).
As with any model, our characterization of the art expert is not a description of reality. It cannot be observed in practice, but we do indicate several alternatives at the end of this section, i.e., where the expert could be “biased” toward a Type I or a Type II error.
Note that “provenance” may be a portmanteau for things other than the trail of ownership such as location, reproduction, exhibition or publication. Some of these are listed separately above. According to Spencer (2004: 199), the basic inquiries recognized by courts in cases involving art are “provenance, connoisseurship and scientific testing.” Naturally, a sound or impressive provenance does not guarantee authenticity since all the elements of provenance can be manufactured and falsified.
We employ Bayesian updating to formalize this model. It is a reasonably well-known technique arising from Bayes rule, a ubiquitous result from introductory statistics. But, while the technique may be well known, the results, arising from its application to the expert’s decision calculus concerning alternative assumptions regarding error costs and likelihood function means, are original.
There are two caveats. First, the expert may provide his decision to the client and make a recommendation to buy or not to buy. Alternatively, the expert may be hired to provide an informed probability of authenticity. In the first instance, the expert must incorporate the preferences of the client, which on its face is likely to be difficult. The latter option, which we adopt here, is more realistic. Second, a decision to buy by the client is not the same as a purchase because, typically, fine art is purchased at auction.
In what follows, we assume that the primary concern of the expert is that the painting is right and, based on this null hypothesis, the expert defines Type I and Type II errors accordingly.
A formal graphical and narrative exposition of these cases is available from the authors on request.
Our model is a theoretical skeleton of the determinants of whether a painting is right or not right. One could argue, incorrectly, that it is not subject to empirical test. Our model of expert decision making might be operationalized as follows: Imagine that there is sample of paintings known with (practicable) certainty to be authored by Breughel the Younger (“autograph”) along with paintings by artists of the same school known with a comparable degree of certainty not to be by Breughel the Younger. For each painting in the sample, the corresponding values of relevant attributes, X, are known—e.g., size, medium and so on. Here, “relevance” may be suggested by an arbitrarily chosen expert, perhaps, based on his or her catalogue raisonné. Alternatively, as analysts, we may propose a set of relevant features. Further, based on assumptions about the actual or fictitious expert’s relative costs of Types I and II error and the specific form of the likelihood functions (normal or logistic probability density functions, for example), we would determine the maximum likelihood estimates of the conditional probability function, Prob(authentic/X). The predictive performance of the estimated model could be assessed based on its accuracy utilizing a ‘hold-out” subsample. Upon choosing the most accurate model, the probabilities of authenticity of paintings “attributed to” Breughel the Younger’s or supposed to be from his atelier could be computed and compared to the distribution of probabilities associated with the paintings presumed to be authentic. It is possible that the model would suggest that some of these non-autograph paintings are more likely to be authentic than some of the presumed autograph paintings. Common sense suggests paintings with similar attributes—e.g., large landscapes—that are more likely to be authentic would sell for higher prices. One test of our model would correlate the probabilities of authenticity predicted by our model with auction prices. Significant discrepancies may suggest some unobserved features highly relevant to buyers.
Recall that we are focusing on connoisseurship. Falsification based on the physical properties of the canvas, paint, etc., is certainly possible and has been used successfully to ferret out frauds.
In a US trial for murder, almost always the only possible basis for falsifying a unanimous opinion of 12 jurors is a confession. In the market for fine art in contrast by convention, the majority opinion of experts (perhaps weighted by reputation or notoriety) suffices for verification.
See Ekelund, Jackson and Tollison (Ekelund et al. 2017, Chapter 6).
One of the litigants against Ms. Freedman was Domenico De Sole, the chairman of Sotheby’s, and wife Eleanore, who purchased a fake Mark Rothko painting for $8.3 million in 2004. Ms. Freedman settled the case with a court proceeding, depriving the public of a judicial determination of possible fraud.
All this took place even though Motherwell and Barnett Newman Foundations and members of Diebenkorn’s family had expressed doubt about the “rightness” of the paintings. As Jasani (2016) concluded, “No amount of due diligence and superb contract drafting will ever fully protect buyers from being lied to.” One buyer was spared buying a fake Jackson Pollock by going to IFAR—the International Foundation for Art Research—an organization devoted to authentication, ownership, theft and other concerns.
To complicate matters ever more, a one-time assistant to O’Keeffe claimed that he had created some of the “Canyon Suite” while working for the artist.
Some writers (Jones 2018) argue that in the early striping, inpainting and restorations, the real work of Leonardo has been significantly “covered up,” suggesting that this is the “real reasons” the Da Vinci has not been displayed by the Louvre Abu Dhabi (as of April 2019). Further, examination of the initially stripped painting shows “pentimenti” in the hands of the picture, suggesting that this adds credence to the picture being by the hand of the master.
In August 2018, purportedly “exciting” new evidence was brought forward which added to the mix when it was discovered that the picture belonged to an English nobleman after Charles I of England. While interesting, the tracing of the picture does not prove that the picture was painted by da Vinci. Additionally, the much-touted etching by the English exile Wenceslaus Hollar, a British exile living in Holland in 1650, carries a suggestion that the painting he used as a model was by Leonardo but, again, this assumes that the painting followed was by the master (Cole 2018).
Certainly, some paintings have a high probability of being simple credence good where falsification is probable within a reasonable period. These may be works executed closer in time, where artists or their agents kept records, or where they were part of a school (for comparisons). Meta-credence works, those with a low or practically zero probability of falsification, are likely from more distant periods or from less famous artists. Nonetheless, these works may sell for substantial sums based on verification through consensus among expert connoisseurs.
Such cases have already affected groups of experts and authenticators. The following group of authenticators have either been disbanded or were embroiled in lawsuits by 2014: Pollock, Warhol, Lichtenstein, Rauschenberg, Rembrandt, Basquiat, Haring, Noguchi, Picasso (disbanded in 1993; reformed in 2012), Modigliani, Twombly, Calder; Giacometti and Motherwell (Sorgatz 2014).
In France, art is subject to the droit moral, an intellectual property law where the artist, his descendants or assigns possess the rights of attribution and other aspects of the market for his or her work. See Rushton (1998).
In his important paper on art and law, Finchham (2017: 24) argues that “Attribution should be a simple and straightforward inquiry, it identifies the creator of a work of art, as well as its place and period of origin” in place of judicial determination of “attribution.” He believes that the combined expertise of art historians, experts and scientists will determine with “absolute certainty” the authenticity of a work of art. While that might be possible for some art, we have argued that some, perhaps much, art will remain as meta-credence goods for long periods or forever.
John Rewald, author of the catalogue raisonné of Paul Cezanne, was noted by Walter Frilchenfeldt in an essay in the catalogue that “When in doubt, Rewald decided rather to include than to reject. He was guided by the philosophy of the German art historian Max J. Friedländer, whose famous remark he translated as follows: “It is indeed an error to collect a forgery but is it a sin to stamp a genuine piece with the seal of falsehood.” (Quoted in Findlay (2004) in Spencer 62).
Ginsburgh et al. (2019) cite evidence that misattributions even sprinkled the collection of the great connoisseur Bernard Berenson, citing Gary Schwartz that only “one of the eight-seven relevant entries is an original Berenson attribution that is still accepted” (p. 25).
Perhaps the most honest admission that some art is of a meta-credence character is in the catalogue raisonné which has a “problems for Study” section containing works that can neither be authenticated or rejected (as well as another section of outright fakes): see (Fincham 2017).
References
Akerlof, G. A. (1970). The market for ‘lemons’: Quality uncertainty and the market mechanism. Quarterly Journal of Economics,84, 488–500.
Calfas, J. (2017). A Leonardo da Vinci Painting Just Sold for $450 Million. Here’s how experts figured out it was real, Time (November 17). http://time.com/5028341/leonardo-da-vinci-salvator-mundi-authentication/. Accessed 5 June 2018.
Chamberlin, E. H. (1934). The theory of monopolistic competition: A re-orientation of the theory of value. Cambridge: Harvard University Press.
Cohen, R. (2013). Bernard Berenson: A life in the picture trade. New Haven: Yale University Press.
Cole. (2018). Leonardo’s Salvator Mundi: Expert uncovers ‘exciting’ new evidence. The Art Newspaper (August 30).
Darby, M., & Karni, E. (1973). Free competition and the optimal amount of fraud. Journal of Law and Economics,16, 67–88.
Dulleck, U., & Kerschbamer, R. (2006). On doctors, mechanics, and computer specialists: The economics of credence goods. Journal of Economic Literature,44, 5–42.
Dulleck, U., Rudolf, K., & Sutter, M. (2011). The economics of credence goods: An experiment on the role of liability, verifiability, reputation, and competition. American Economic Review,101, 526–555.
Dupuit, J. (1962 [1849]). On tolls and transport charges. International Economic Papers (E. Henderson, Trans.), No. 11. London: Macmillan.
Ekelund, R. B., Jackson, J. D., & Tollison, R. D. (2017). The economics of American art: Issues, artists and market institutions. Oxford: Oxford University Press.
Ekelund, R., & Thornton, M. (2019). Extreme credence and ‘imaginary goods’. Manuscript (Auburn University).
Fincham, D. (2017). Authenticating art by valuing art experts. Mississippi Law Journal,86, 586–626.
Findlay, M. (2004). The Catalogue Raisonné. In D. Spencer (Ed.), The expert vs the object judging fakes and false attributions in the visual arts. Oxford: Oxford University Press.
Ginsburgh, V., & Menger, P.-M. (Eds.). (1996). Economics of the arts: Selected essays. Amsterdam: Elsevier.
Ginsburgh, V., Radermecker, A.-S., & Tommasi, D. (2019). The Effect of experts’ opinion on prices of art works: The case of Peter Brueghel the younger. Journal of Economic Behavior & Organization,159, 36–50.
Ginsburgh, V., & Van Ours, J. (2003). Expert opinion and compensation: Evidence from a musical competition. American Economic Review,93, 289–296.
Hilger, J., Rafert, G., & Boas, S. V. (2011). Expert opinion and the demand for experience goods: An experimental approach in the retail wine market. The Review of Economics and Statistics.,93, 1289–1296.
Hoving, T. (1996). False impressions: The hunt for big-time art fales. New York: Simon and Schuster.
Jasani, A. (2016). De Salle v. Knoedler gallery—A field of red flags, Art@Law (March 31).
Jones, J. (2018). The Da Vinci Mystery: Why is his $450 m masterpiece really being kept under wraps? The Guardian (October 14). https://www.theguardian.com/artanddesign/2018/oct/14/leonardo-da-vinci-mystery-why-is-his-450m-masterpiece-really-being-kept-under-wraps-salvator-mundi. Accessed 3 Jan 2019.
Karczewski, L. (2017) Authenticating the $450.3 Million Leonardo da Vinci Masterpiece with Science, Art law (December 1). https://artlaw.foxrothschild.com/2017/12/articles/art-authentication/authenticating-the-450-3-million-leonardo-da-vinci-masterpiece-with-science/. Accessed 17 May 2018.
Lacy, G. S. (2011). Standardizing Warhol: Antitrust liability for denying the authenticity of artwork, Washington Journal of Law, Technology & Arts 6: 185. (Winter). http://digital.law.washington.edu/dspacelaw/bitstream/handle/1773.1/560/6wjlta185.pdf?sequence=4. Accessed 23 Mar 2015.
Landes, W., Posner R. (1996). The economics of legal disputes over the ownership of works of art and other collectibles in Ginsburgh and Menger. In Essays in the economics of the arts, pp. 177–219.
Nelson, P. (1970). Information and Consumer Behavior. Journal of Political Economy,78, 311–329.
Nelson, P. (1974). Advertising as Information. Journal of Political Economy,82, 729–754.
Neuendorf, H. (2018). Who Really Painted ‘Salvator Mundi’? An Oxford Art Historian Says It Was Leonardo’s Assistant, Art World (August 7). https://news.artnet.com/art-world/did-leonardos-assistant-paint-salvator-mundi-1329638. Accessed 10 Aug 2018.
Rushton, M. (1998). The Moral Rights of Artists: Droit Moral ou Droit Pécuniaire? Journal of Cultural Economics,22, 15–32.
Sauej, N. (2017). Artistic license? Experts doubt Leonardo da Vinci painted $450 m Salvator Mundi, The guardian (November 20). https://www.theguardian.com/artanddesign/2017/nov/20/artistic-license-experts-doubt-eonardo-da-vinci-painted-450m-salvator-mundi. Accessed 20 Sept 2018.
Sorgatz, R. (2014). The end of authentication, The message (July 17) https://medium.com/@rexsorgatz. Accessed 5 Jan 2015.
Spencer, R. D. (Ed.). (2004). The expert vs the object. Judging fakes and false attributions in the visual arts. Oxford: Oxford University Press.
Stigler, G. (1961). The economics of information. Journal of Political Economy,69, 213–235.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Ekelund, R.B., Higgins, R. & Jackson, J.D. ART as meta-credence: authentication and the role of experts. J Cult Econ 44, 155–171 (2020). https://doi.org/10.1007/s10824-019-09354-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10824-019-09354-3