How big is the art market? It’s a question that seems relatively simple, yet is elusive. The difficulty of pinning down an estimate (mostly as a result of the fact that it’s in dealers’ best interests to keep their sales private) means that any figures that do get reported take on a life of their own as they are echoed across the industry. The size of the market is an important indicator of the health of the art world, how much it is growing (or shrinking), and its importance relative to other parts of the economies.
As one of the only professionals doing reliable and largely unbiased quantitative analysis, Clare McAndrew’s annual TEFAF-sponsored art market report is widely considered to include the definitive yearly figures in the industry. But not everyone agrees with her results.
In response to this year’s edition of the TEFAF paper, New York-based art lawyer and former associate counsel at Sotheby’s Michael McCollough wrote a blog post detailing “two major flaws from which it cannot recover completely.” First, he says that McAndrew has vastly underestimated the size of the private art market (galleries and dealers), and second, the volume of sales in each sector are not detailed. He thinks the private market (gallery sales) are actually twice the public auction houses' reported sales:
For instance, they say the public market (auctions) makes up 48% of the total market and the private market (galleries) comprises 52%. To people who actually work in the market, these number [sic] are off wildly; most people estimate the private market to be at least twice as large as the public market. If my estimates are correct — and we know the global public market was at $27 billion last year — then the global private market was around $54 billion, thus the total global market should have been somewhere around $81 billion.
ARTINFO reached out to McCullough to discuss his critique. “If you think about it intuitively, the secondary market is what most people call the churn in the market [buying and selling old works] — when people decide to deaccession works. The churn in the market could never be equal to the current output [the primary market],” he said. If that were true, he continued, then no one would have art on their walls because they would always be buying and selling.
Does this critique hold water? One issue with his argument is that it discounts the fact that world’s supply of art has been built up over a long period of time, making supply on the secondary market side much, much larger than on the primary market. Because people tend not to destroy art at the same rate they are buying it, the amount of art in the market is constantly expanding — therefore, the “churn” within this chunk of the market definitely can be larger than the annual turnover in the primary market.
For example, say in our alternative art world — ARTINFOLAND — artists produce 1,000 pieces of artwork per year, all of which get sold through the gallery system. Artists have been producing about 1,000 pieces of artwork per year for the last 150 years (let’s assume we’re only talking about Impressionist, modern, and contemporary to keep the numbers smaller). Some of those got destroyed or lost or were cast off for one reason or another, but there are still some 100,000 artworks sitting on the walls of ARTINFOLAND collectors. Every year, the gallery system sells 1,000 works of art. As long as at least one in every 100 of the existing works sitting on the walls of collectors get resold in the secondary market (and the prices of secondary market sales are roughly equal to art prices in the primary market, the auction system will be as large or larger to the gallery system.
HOW THE SAUSAGE GETS MADE
While McCullough’s critique of McAndrew has a few flaws, his post did trigger a really interesting discussion with the economist about her methods and the difficulty of answering the fairly straightforward question: How big is the art market?
McAndrew noted that she was surprised by McCullough’s critique because so many other people giving her feedback thought she had overestimated the size of the primary market. She said that she always takes the opinions of experts in the industry seriously, and their feedback can be very helpful, but she also pointed out that every person enmeshed in the market has an inherent bias because they spend a lot more time in their particular segment of the market than they do exploring the bigger picture.
McAndrew starts out with a list of dealers — which has doubled in size from 3,500 since she started doing this report to 7,000 this year — and she conducts a large poll that asks them a number of different questions about their business and their place in the industry. However, it isn’t perfect. “The problem with the survey is that even with the data I have, you are not covering every part of the market. It’s not a random sample, it’s a stratified sample,” she noted.
Therefore, there are a number of statistical workarounds she has to do in order to clean up the data and make it more accurately reflect the actual state of the market. For example, she has realized over time that the top 1-2 percent and the bottom 10 percent are usually underrepresented in the sample — either because they don’t belong to the dealers associations that she works with, or because they just don’t respond to her survey (I’m looking at you, megagalleries) — so she creates a normal distribution curve to estimate what she should be getting from those at the top and the bottom who were left out of her survey.
“To test it back again, I will ask dealers how their sales compare to the auction sector. Then I see how that compares to the figures I’ve generated from the opinion poll and the statistical model. When you cut out strategic answers [dealers clearly trying to game the system and inflate their responses], the two are very similar,” she said. The more the numbers line up, the closer she is to a realistic picture of the state of the market.
NARROWING IN ON A MORE ACCURATE PICTURE
Of course, there are markets where this is easier to do than others, which gets back to McCullough’s second point about the lack of detailed data by sector — in many cases, McAndrew’s sample sizes are just not large enough to publish a result with any statistical significance.
Though not without problems, it’s much easier to gather data from established art markets like in the U.S., U.K., and France than from smaller and emerging markets like China, Brazil, Spain, or Sweden. Every year she studies a few of these markets to try to make her results more accurate. “Each year I work on [a few of] the smaller countries that would get a gross aggregate applied to them [basically an estimate]. Every year I’m getting more and more specific for each country,” she explained.
“I got a few people with the bigger auction houses saying that the auction houses were much more dominant in terms of value,” she added.
As for why McAndrew has left her methodologies — which used to be included in an appendix — out of her report? TEFAF thought no one would be interested.
Watch our interview at TEFAF with Clare McAndrew: