Authenticating Art by the Numbers:

The Fine Art Authenticity Score (FAAS)

Numbers

Those of us who are engaged in the qualitative subjective fields of art, history or literature rarely come face to face with numbers. Yes, we have to know how much our art materials cost, the price we will charge for our completed work, our mortgage rates, our utility bills, and so on. Our backgrounds and education force our thinking and behavior along the lines of non-numerical approaches in understanding and decision making. Very rarely do historians of art, aestheticians, or artists deal with raw data, its uses, processes, or where the data leads. As it turns out, many of our day to day actions rely heavily on numbers which are present subconsciously, and if we begin to allow what is taking place beneath the veil of the subconscious to emerge, we would be amazed at the power of numbers to develop sound analyses and strong conclusions about the level of authenticity in a work of art. Numbers are slowly challenging the way we perceive the subject of authenticity, and ultimately will provide a greater level of trust in its ascertainment.

Daily Life Decisions and Probability

We don't seem to be aware that many or most of our daily decisions are decisions based on probability. If we want to find out the probability of our 7 A.M. train being on time daily over the course of one year we would divide the days the train runs by the days the train runs on time. If out of 250 days it reached the station at 7.A.M. only 25 times, the probability of on time performance would only be 0.10 or one out of ten. Further, one could also assert that if it runs on time only 10 percent, it follows that 90 percent of the time it would be late. The probability of lateness is a significant piece of information in that it is a variable in your decisions regarding your future behavior and choices. If it is going to be late most of the time you will have to get up earlier to perhaps drive to work or take an earlier train. Or you might decide to move to an area where the trains have a higher probability of on time performance.

The significance of probability information is that it can result in a different decision about future actions. Fine art authenticity works in a similar fashion.

Fine Art Authenticity

Fine art authenticity moves on a continuum from copies, to fakes, to works created by the artist apprentices from the schools operated by the master, to works attributed to the master, and finally to genuine works solely from the hand of the master artist. The categorization of status on this continuum emanates from scholarly examination consisting of scientific analysis, provenance, and connoisseurship studies. Inevitably the system and structures of examination are grounded in a qualitative rather than quantitative framework simply because art history circumscribes and permeates fine art scholarship. This is not to say that all areas are grounded in the qualitative. The examination of the material composition of fine art takes place in the realm of both digital and statistical analysis of various data arrangements found.

Thomas Hoving and the Metropolitan Museum of Art (the Met) Authenticity Score

Thomas Hoving was the Director of the Met for 17 years. During the course of his tenure he was involved in the authentication analysis of 50,000 works of art held by the museum. This analysis focused on sculptures, paintings, prints, and works of various sizes, shapes and colors. It also included restored and non-restored works. In this analysis, curators, historians, scientists, and other experts using up-to-date technology and an interdisciplinary approach drew conclusions based on the data and information found. On a weekly basis, under the tutelage of Hoving, staff participated in breakfast meetings, tweaking their expertise by examining an eye, ear, or hand from a famous work chosen by Hoving who asked them to identify the author of the work. An expert who failed to correctly name the author of a given work solely from a drawn or painted ear, eye, or hand, suffered by being the poorer when he walked out of the breakfast, since breakfast was on the loser. After many years, Hoving estimated that only about 60% of the works in the Met were authentic or from the hand of the artist to whom they were attributed. Those not contained under the identifier of authentic were relegated to the categories of questionable authenticity due to unresolved issues surrounding the work or bogus works.

The Rembrandt Project Authenticity Score

In 1965 in the Netherlands, the Rembrandt Project was born with a life expectancy of about 10 years. The project was dedicated to authenticating the works of Rembrandt. Using scientific analysis, provenance research, and connoisseurship, the project team examined over 600 works allegedly created by the hand of Rembrandt. The result of the many years of research was that only about half were deemed to be authentic.

Note that such determinations of authenticity are not etched in stone. As new technologies and documents surface, such conclusions may change. From the work of Hoving and the Rembrandt Project one could reasonably infer that between 40 and 50% of the paintings examined via strict empirical analysis are of questionable authenticity.

A Quantitative Approach: Probability of Authenticity

In both the Hoving and Rembrandt project the authenticity/inauthenticity score consisted of a numerical value or simply a number/s. Hoving's numbers were 60/40 in favor of inauthenticity; the Rembrandt numbers were slightly different in that they were 50/50. What both numbers provide for us is a probability ratio of authenticity to inauthenticity. The probability of authenticity/inauthenticity is no different from the probability of flipping a coin, or throwing a dice.

Assume that you want to purchase a work of art and the particular artist is known to have very few inauthentic works attributed to him in circulation. Let us say over a period of five years there are 100 works attributed to the artist with one of questionable authenticity. Following our simple formula of percentage of probability we would have 1 over 100. 0.01, or one percent, are likely to be questionable and 0.99 authentic. The probability of authenticity would be 99%. In this case the decision to purchase is straightforward, since the chances of purchasing a fake is one out of one hundred.

If we are collectors of museum quality fine art the choice is more complex. Research has found that in the case of museum quality artists, some have high probabilities of authenticity of works attributed to them and others meager. The Rembrandt Project noted that a collector about to purchase a Rembrandt work has a 50% percent chance of buying a dud and 50% chance of buying a genuine piece. Miro, Chagall, Dali, Picasso, Matisse, Corot, and so on, have all been categorized with lesser or greater probability of authenticity for their attributed works. Police raided a show in Sweden and found a few thousand fakes of the above artists' works. Over the last year police found thousands of fakes being sold from a gallery in Chicago. The significance of the probability score is that it is possible to generate a numerical ratio from existing information of numbers of authentic works versus fakes. If we can secure the number of works of a given artist from a catalogue and have a number secured from reliable sources about the number of fakes attributed to that artist, we can establish an authentication score.

The FAAS: Step by Step

Assume for a moment that you have just found out that your favorite gallery has come into possession of a work by the famous French Impressionist painter, Corot. Although the gallery insists that there are no problems with the works by Corot, you decide to carry out your own investigation of the Fine Art Probability Score of Corot's works.

Venice, The Piazza by Corot

1. You start your investigation by typing "Corot Forgeries" into your search engine. Here I used AOL.
2. Your first hit is Wikipedia and it says that the Jousseaume collection has 2414 forgeries in its collection. It does not provide an exact number of authentic works.
3. The third website, Artcult notes that Robaut's catalogue states that Corot painted 2,500 genuine works and that there are 10,000 fakes in existence. At this point you have enough data to develop your FAAS.
4. Taking the total number of works available divided by the number of authentic or 12,500/2.500, we get a FAAS of 0.20 or twenty percent authentic. This is a low score. It means that your chances of buying an authentic Corot are 1 out of 5 or that if you bought 100 Corots, 80 would probably be fakes and 20 would be genuine.

Some points to consider here:

1. Note here that since no information exists that the 2414 fakes existed in addition to the 10,000 fakes indicated by Robaut, we will make the assumption that they were part of the 10,000.
2. The data grounding the Corot example was easily secured in a few steps. This is the exception rather than the rule. Be prepared to spend some time ascertaining authenticity and inauthenticity numbers. Much information is supplied via web searches.
3. The FAAS is not an indicator of a given work but of an artist's oeuvre. FAAS tells us where a given artist stands in terms of the chances of purchasing a bogus work.

If a collector is interested in purchasing a work and has a choice, let us say, of buying a Picasso, a Dali, and a Rembrandt, an authenticity probability score could be derived for each one enabling the collector to base his or her purchase decision on another point of information rather than the salesperson's push to buy now.

One may argue here that it is logically fallacious to move from a given high probability of inauthenticity in one medium attributed to a given artist to conclude that all mediums of the artist have a high level of inauthenticity. The probability function is not really moving from a parts to whole fallacy; it is only noting the mathematical fact that comparing artists attributions quantitatively for all their works, some have higher probabilities of authenticity than others. This is no different than what Hoving's group of researchers did or what the Rembrandt project is all about. It is recognized that it would be great to be able to separate oils from lithos, and sculptures from collages to come up with a more refined probability score. This concept has a problem in that there is a difficulty in securing information about how many works a given artist authored let alone how many were in each medium category. More importantly, the focus of this article is the fact that authenticity determinations can be developed quantitatively but much work is needed to develop the criteria for authenticity and its application to different mediums of art.

Applying FAAS Analysis

Assume for a minute you are interested in buying one of three works by Rembrandt, Corot, and a contemporary artist you frequently collect. You have the money and are ready to buy. You may decide amongst the three in numerous ways:

1. You just like one of the three, that's it. Here, feelings, emotions, and likes serve as the backdrop for your decision. It is like buying a sweater or tie. You just like it! For art as an enjoyment to collect and an investment, I am not sure this would be the way to purchase. If you find out later that you bought a fake, you will probably deeply regret the purchase.
2. You can hire a connoisseur, and he may be able to ascertain from "feelings in his stomach." I might be wrong here but I am really not interested in art as a food dish so this might not work.
3. You are a critical thinker and believe that sound reasoning should serve as the determinant for making decisions about how and where your money goes. The first issue regarding the works has to do with authenticity. If I am going to lay down a few thousand on a work, I want to be sure that it is worth it. If it is not authentic, my purchase price is at risk since what I pay is not what the piece is worth. I ask myself the question: How do I determine the work's authenticity? Asking the gallery manager or owner to provide a Certificate of Authenticity collides with a conflict of interest if they provide one from their own hand, since managers and gallery owners are there to make a profit. Or, I could seek out an appraiser to provide a COA. The problem is that appraisers are not authentication oriented and may not be expert in the particular art or artist. One quick way would be to ascertain from research the quantity of works produced by the three artists versus the quantity of fakes or questionable attributions, (Wikipedia and Museum Security Network provide various citations and references about fakes, forgers, and forgery numbers). Rembrandt's probability of authenticity is about 50%, Corot is about 20%, and your favorite artist is almost 100% since no fakes exist to lessen the probability of authenticity. The choice seems obvious. The numbers say go with your favorite artist.

The above represents a simplified approach for instruction. Decisions about purchasing fine art may be complex but there is no reason not to utilize a quantitative approach in reaching a decision or at least in helping to reach one.

The Rembrandt Project and Thomas Hoving of the Met provided the foundation of authenticity by the numbers, and I see no reason why their concept cannot be developed further to focus on determining the likelihood of fine art authenticity by a simple probability number.

The FAAS Approach

The FAAS approach is one factor, by no means the only one, in an assemblage of factors aiding the fine art consumer in making a decision about purchasing a work of art. Taken together with provenance, scientific analysis, and connoisseurship, it represents somewhat of a challenge to the existing approach of qualitative subjectivism or relying on, "a funny feelings in the stomach" as the criterion of authenticity. The guiding principles for following the FAAS approach in purchasing a work of art are as follows:

1. Adopt a skeptical approach to fine art for sale.
2. Before you purchase the art critically secure any information regarding the artist being considered. Remember that data exists indicating to what extent a given artist has been forged. Remember the Met, Rembrandt Project, and Corot.
3. If available, take the total number of works and divide it by the total number of authentic works. The Fine Art Authenticity Score (FAAS) will be a percentage of the total number of works divided by the number of authentic works.
4. If the number hovers around 0.50 then your potential purchase is more of a coin toss. Would you buy a car this way?/li>
5. If the number is less than a coin toss then reason should bite you and make you move away from the purchase.
6. If the number is found somewhere in the 0.80+ then your consideration of the work may, if necessary, move to the next level of documentation, connoisseurship, and scientific analysis. At this point you have evidence that the train runs 0.80 percent on time, rather than waiting for a train that may or may not arrive. In point, your research and your FAAS number has indicated that your potential fine art purchase is not like a house purchased on a hazardous dumpsite but on land designated as environmentally clean.
7. If for some reason the number is low but you are smitten by the artwork, ask the seller to provide data proving that the work is authentic.

Summary

Using a probability score to ascertain the authenticity for the oeuvre of a given artist represents an additional tool in the armaments of examining a work of art for collecting purposes. It does not establish that a given work is indeed authentic or inauthentic, but provides its authenticity probability prior to purchase. Like purchasing a car in terms of reliability, the FAAS tells the buyer the chances of purchasing a used car in terms of reliability: some are most reliable, moderately reliable, and poorly reliable. Wouldn't you rather purchase the most authentic artist works rather than the least authentic works?

— by John Daab Ph.D.  |  April 24, 2009  |  Print Version - PDF

Print   |  

back to top