How Useful Are Art Indices?

When you think about art indices, you need to consider carefully what they choose to include as well as what they are forced to leave out.

While making a laudable attempt to bring transparency to the opaque art market, they have always been hobbled by the lack of sales data available. All of them rely on data from just half the art market – the auction market – when 53% of the global art market is actually made up of private gallery and dealer sales, according to TEFAF’s latest report.
Typically, they also track the most successful art sales at auction, in some cases, artworks that have successfully sold at auction more than once, while omitting artworks that fail to sell at auction (around one quarter of lots). As Georgina Adam points out in her new book ‘Big Bucks: The Explosion of the Art Market in the 21st Century‘, “if 10 Warhols are offered for sale, nine are bought in [fail to find a buyer or meet the unknown reserve price] but one triples its estimate, then performance indices would record a fine result”.
Barnett Newman's Black Fire I that sold for $84.1 million at Christie's last November. Art indices focus on the best performing part of the auction market.
Barnett Newman’s Black Fire I that sold for $84.1 million at Christie’s last November. Art indices focus on the best performing part of the auction market.
Not all indices include atypical results. Art Market Research, for example, allows subscribers to create their own indices that cut out the top and bottom outlier results at auction, but Mike Moses, co-founder of the Mei Moses Art Indices, agrees that buy-ins are a particular challenge. “Every index provider faces the same problem of how to deal with buy-ins,” he says. “There’s no way to assign any value to these lots, other than making up a price.” I should add here the obvious fact that no art indices take into account all the art that is never deemed valuable enough to be resold in the first place.
These are thorny problems, but if you use art market indices to assess whether art is a good investment, you’re making investment assumptions about a market where half of the data from private sales is unknowable and where the tranche of data that indices do include from the other half, the auction market, has a selection bias.
The fact that many of the sales prices that result from online auctions are not published suggests that the data art indices rely on are becoming less representative of the broader market than before. Selling art online is “leading to less, not greater transparency in the market, running counter to the freer access to price data that the internet was supposed to foster,” writes Adam. “Couple this with the growing number of private sales by auction houses, then comparing prices is likely to become less easy, not more so.”
As you think about any art market index, it’s important to consider the very different methodologies they use to tackle a market where all artworks are unique, but also the usefulness of their underlying data, so this seems like a good time to outline what some of the main art indices do and don’t include.
I’m also including the latest year-to-date results from each providers’ post-war and contemporary art indices below, where that is available, as a point of comparison.
1. Mei Moses Indices
Indices available: Mei Moses publishes one World All Art Index and seven indices representing different collecting categories. All of them are updated annually, although Mei Moses will soon publish a semi-annual update for the World All Art Index. Mei Moses also issues quarterly tracking estimates for these indices, based on new results so far during the year.
Source of underlying data: To try to make useful comparisons in a market in which all artworks are unique, the Mei Moses database only include artworks that have sold more than once, using data collected from two companies, Sotheby’s and Christie’s, but not their online sales. The Mei Moses database includes over 45,000 repeat sale pairs for over 20,000 individual works of art sold around the world. Around 3,000 to 4,000 new repeat sale pairs are added each year, and while the data about repeat sales is just gleaned from Sotheby’s and Christie’s, the auction venue where the artwork sold in the first place can be anywhere.
Index methodology: Using this repeat sales data, the Mei Moses Indices track the price difference between each subsequent sale. The indices include all the repeat sales data available and are not weighted. Bought-in lots are not included.
Performance of the Mei Moses World Post-War and Contemporary Index: down 0.47% from January to July 2014.
2. Artnet Indices
Indices available: Indices that monitor the performance of various market categories, such as contemporary art, Impressionist art and modern art. Subscribers can also access artist-specific indices or indices devoted to a subset of that artist’s work.

Source of underlying data:
Artnet’s database includes sales from 1,600 auction houses worldwide, including the results of online auctions from the companies that make this information available. These include lots sold via Artnet’s website, but also through external sites such as Auctionata and Heritage Auctions.
These external sites do not include Sotheby’s and Christie’s, although Sotheby’s online-only auctions through its new partnership with EBay have yet to start. Jordan Quitko, director of product operations at Artnet, reports that Christie’s used to publish the prices from these sales, but has since stopped. Quitko says that Artnet is actively talking with Sotheby’s and Christie’s about obtaining these results.
Index methodology: Artnet organizes auction sales into comparable groups, which includes repeat sales at auction as well as art works that have only sold once, as long they share enough characteristics with the other artworks in that group. Index values for individual artists within Artnet’s market sector indices are calculated based on the median price per lot for each artist within these groups, excluding prints, multiplied by total lots sold for that artist.
The Artnet Contemporary C50 index, for example, ranks the top performing 50 contemporary artists based on their index values each year, subject to a decay formula that looks back five years, and gives each of those 50 artists equal weighting in the index. This approach is said to capture 75% of all sold lots in the Artnet database, but does not include bought-in lots.
Performance of the Artnet Contemporary C50: Up 120% from January to July 2014.
Performance of the Artnet Modern 25: Down 54% from January to July 2014.
3. Artprice Indices
Indices available: Artprice publishes the Artprice Global Index, from which it also extracts results for different art periods, such as contemporary and modern art. Subscribers to its website can also generate artist-specific indices.
Source of the underlying data: Sales from 4,500 auction houses, but generally not the results of online auctions, according to Jean Minguet, Artprice’s head of econometrics, either because those results aren’t published (as in the case of Christie’s) or because they are not sold through a live auction format, which includes sales on Artprice.com. However, its database does include results from companies like Heffel that do hold live online auctions and publish the results.
Index methodology: The Artprice Global Index also uses the repeat sale methodology to track the performance of individual artworks that have sold more than once at auction and bought-in lots are not included.
Performance of the Index: Artprice would not release the latest year-to-date results for its index, but its price index for contemporary art increased 102% in 2013, while post-war art rose 76.7%.
It’s worth ending by saying that although art indices have a selection bias, they’re not the only ones. Under performing stocks are periodically switched out in favor of better performing names in the S&P500, for example. A key difference here, though, is that art indices, which are an amalgam of historical art sales, are not investable indices.
You can buy an index fund that replicates the performance of the S&P500. You can’t buy a fund that gives you exposure to the Artnet C50, for example; you’d need billions of dollars and amazing powers of persuasion to prize all the artworks included in the index away from their current owners to achieve that. And as I’ve written before, and as Adam suggests in the Warhol example above, even if you could buy different artworks by all the artists included in the index, they would not be a reliable proxy for its performance.
All these indices are a useful resource as long as people understand the different ways in which they slice and dice the market: you might be following the results of 25 or 50 or 1000 artists. Ultimately, though, they all focus on the best performing slice of the auction market. If we had more data available from online auctions, from private sales at auction houses and from all the traditional dealer and gallery sales conducted worldwide, would the returns posted by art indices actually look better? Perhaps, but given that all this information is missing, it’s impossible to tell.

When we’re looking for useful information about prices in the art market, a lot of the time we’re still searching in the dark.

Credits

Contributor
Post a Comment

Popular Posts