Art market indexes
July 10, 2013
The growth of the global art market, the creation of Art Funds, online price information, and new platforms for trading art, have all led to an increase in the desire to track movements in the art market more accurately.
Image: © Nationale Beeldbank
Together with methodologies developed originally to track house price movements, improved computing power and the desire to have a variety of art market indices are fuelling the move towards tradable art market indices. Interest in the art market is growing rapidly.
It is all too familiar that auction sales hitting the headlines with record-breaking prices catch our attention, spurring the desire for an investment in a beautiful asset that would have been financially very attractive as well as aesthetically pleasing. Those artworks failing to hit their reserve price at auction go unsold and those selling at low prices fail to catch any media limelight at all. But to what extent do average art prices increase over time? Are the record-breakers the exception, or is there an overall trend for art prices to appreciate over time?
Art prices will appreciate over time
Generally speaking, as long as art holds its cultural value in society, art prices will appreciate over time, reflecting a general increase in consumer prices in line with the rate of inflation. Artworks with a rich art history will tend to maintain their prices well, with less variability in price movements over time. Such paintings fall less in and out of fashion and the co-movement with overall prices is higher. In terms of finance, the idiosyncratic risk is less. While for artworks which carry much more idiosyncratic risk, as with, for example the rise and fall of impressionist paintings in the 1980s and 1990s and Contemporary art in the last decade, the price variability is often much greater. Artworks, with a proven track record, maintain their cultural and economic value, and the likelihood of falling prices is lower, hence, can be considered therefore as less risky. However, for paintings that fall from fashion and result in low sales prices, the picture can be very different. If considered as an investment, the financial risk can be much greater. The financial return over the holding period can in some cases be negative, even before inflationary effects are taken into account.
Gauge overall market performance
Those artworks which obtain large positive returns are the ones we tend to hear about, and those that don’t sell well, fall to the wayside. So to what extent do the stories we hear about in the press represent the overall art market? This is when an art price index is a useful means to measure the extent to which prices have changed over time; so that we can gauge overall market performance. Also submarkets, or different genres, artists, and movements can be indexed.
One way to estimate such average price changes for the art market is to track a basket of artworks over time; or rather, a portfolio of paintings, but the concept is the same. In recent years the financial press have quoted various art indices to gauge the overall level of art sales prices. Online internet websites which collect auction house sales data from the major auction houses also provide this service. Artnet, Artsales, Art Market Research and the Mei Moses art indexes, all provide a variety of indices available to subscribers, generally with an annual fee. Most providers provide information on their chosen index methodology and the choice of artists in the various indexes. Currently it is only possible to track auction sales prices, as sales prices from dealers are not widely available. The reliability and accuracy of the data used is very important. Naturally some errors will occur in the data, however, there can be issues with currencies, the recording of old English pounds and guineas, hammer prices and prices with premiums, which are at times incorrectly recorded. Furthermore unless the identical painting is identified when using repeat sales, then the use of repeat sales estimates to estimate an index will be incorrect. Simply matching titles of artworks by the same artist is not effective at finding repeat sales, as often the artwork is reattributed, or the title of the artwork changes over time.
Market boom in the 1960s
There are various methodologies to create such an index to track the art market. The earliest such methods adopted are average price indexes, and were used during the art market boom in the 1960s to track the relative increase of painting prices at a particularly buoyant time in the art market. Art market indexes themselves are therefore not new. The methodology at the time was to take the average prices of a large number of similar quality artworks. The idea not to include extremely high priced paintings, or low valued pieces was to not skew the average prices with a few outliers. Comparing similar artists, and similar artworks was the first means to construct indexes of groups of artworks over time.
Efforts in the 1960s to capture the large appreciation in art prices during that decade resulted in the establishment of a number of art price indexes. One of the first to do this was the art investor, Richard Rush. He provided graphs of indexes using groups of similar artists in an attempt to provide an index for particular genres of art movements since the 1920s, up until the 1960s. Interestingly, he tried to follow popular artworks sold in the 1920s and compare their price movements, rather than back-fill the popular artists of the sixties. This was one of the first approaches to measure overall average price movements (increases and decreases, of course) was to take prices of established artists and note their sales prices. Using enough artists, and avoiding exclusive, expensive masterpieces, led to the creation of an art market index as far back as the mid 60’s.
In 1967, Geraldine Keen, working for The Times newspaper, together with Peter Wilson, the Chairman of the board of Sotheby and Company created The Times Sotheby Index of Fine Art Prices, with the aim of producing a monthly index in the Times newspaper using art auction sales from Sotheby’s Auction House. Their aim was to measure the extent to which art prices had changed over the years, on average in various different fields of the art market. Although popular during the 1960s and early 1970s, The Times stopped reporting the index in the 1970s, and refer to the Art Sales Index as the most comprehensive available guide on art prices available instead. More recently the index was taken over by the Louise Blouin Media group.
Indeed, during each art market boom over the last sixty years there have been continued efforts to provide an art market index to reflect the sometimes rapid appreciation in art prices. When the market frenzy dies down, and the market collapses, or rather dries up, with few art sales taking place, interest wanes, and only with a resurgence in art sales prices, and record breaking prices hitting the headlines is interest reignited once more.
The methodological developments in providing indices of housing prices in the real estate sector have been adopted by academics in economics and finance to develop similar price indexes for art. There are two main methodologies used for the construction of these indexes. Firstly the hedonic models, which aim to explain the influence of time on prices, over and above the price paid for acquiring certain characteristics, capture the overall market movement. Once the premium for the artist, the size of the paintings, the location of sale, the medium and the genre of the painting have all been taken into consideration, and any other relevant characteristics, which do not change over time, then the premium which is paid for the market sentiment is also included as a further characteristic.
The residual or the part which is unexplained by the painting characteristics must be determined by the time aspect. Using a variable for each year, then the yearly effect can be estimated accordingly. In the 1980s many indexes were constructed using data from the Gerald Reitlinger and Enriqe Mayer data collections on art sales prices. Both hedonic indexes and repeat sales indexes were estimated.
The Mei-Moses repeat sales index was developed in 2002, using the same repeat sales index technique as the Case-Shiller house price indexes, now commonly used in the U.S. and quoted on Bloomberg. Indeed, even financial derivatives are traded on these housing indexes. With fresh data from New York auction houses, interest was spurred in quoting these art prices in the financial media, with art prices for the US auction houses outperforming the S&P500 equity index.
Art being considered a financial investment
Although not taking into account transaction costs, the apparent outperformance of art has led to continued interest in art being considered a financial investment and even as an asset class in its own right. Although using auction house data alone, and the bias induced at just focusing only on leading auction houses can be criticised, it does provide a good indication of price movements over a long period in time. Until dealers provide information on their prices, the auction market will be the only market that is transparent enough for art sales prices to be collected and used to construct art price indexes.
The repeat sales index is a special case of the hedonic methodology. If all paintings are included in the repeat sales model, then the estimation technique collapses to the hedonic methodology. The advantage however, of using the repeat sales methodology is that because two sales prices are used (purchase price and sales price), the characteristics of the painting do not need to be included, and the repeat sales index captures the pure change in price, over time, of an identical artwork.
The disadvantage is that not all works have two sales prices available. Take for example contemporary art appearing at auction for the first time, or artwork which have not appeared at auction before, either having only been sold in the primary market by dealers and not before at auction. The disadvantage of the hedonic approach is that the variables used to characterise the painting have to be able to capture the full variation in price across artworks. The most important of these variables is invariably the artist him or herself, followed by attributes such as oil on canvas, surface size, age of the artist etc, but other attributes and characteristics that cannot be quantified quite so easily get lost in this approach.
Consider two similar artworks in their attributes, but where the price tag is significantly different for the same artist – a white Fontana with three slashes, versus a red one with six slashes, sells at a multiple of the price of the white one. Only if such characteristics are additionally employed can the residual price movement be correctly attributed to the time dimension. As long as the hedonic methodology does not omit important variables and captures all of a painting’s characteristics it will capture well the time variation of artworks. If the importance of these various characteristics changes over time, then the time variation will be incorrectly estimated.
Dropped from the sample
Tracking identical paintings over time, so that repeat sales can be found for the same painting, means that many paintings are dropped from the sample. This
can be unsatisfactory. The appropriate method can be adopted depending on the requirements for which the index is to be used. Increasingly, the greater sources
of data available have proved that the differences in the indexes constructed from the various approaches are limited in their differences. In economic parlance,
they are not too dissimilar or insignificantly different from each other. Including hedonic characteristics into a repeat sales regression to capture characteristics,
which change between the point of purchase and sale, represent a hybrid approach to index construction; for example different sales locations or auctioneers in addition to the repeat sales prices.
What is new is the availability of data and the speed and ease at which indexes can be updated. The various index providers now have online tools that give subscribers access to almost immediate information on changes in indexes; as well as tailor-made sector or specific artist indices.
Negative art market price movements
The use of indexes to evaluate art fund managers trading performance, and for insurance companies to correctly assess the value of their clients’ artworks, may be the natural first place where indexes are adopted. Hedging the risk of negative art market price movements is of less importance to collectors who always maintain the aesthetic beauty of their artworks, even in times of a market downturn, but is certainly highly important to banks and institutions who may have issued loans against art as collateral. If the financial services industry has a need to take such short positions on the art market, then financial products based on art prices will undoubtedly become available to interested parties on the market. Picking up a ‘cheap’ Old Master painting, or a Gerard Richter at a fraction of its current price, if the market turns down may be an attractive option not only for Art Funds, but also for Museums and collectors.
Although it is not possible to invest in a fund that track these indices exactly, as is commonplace in the equity share market, real-estate market and for other financial assets, there are a number of specialised funds which will do a good job in investing in a subset of the index, and provides a similar financial return. With the average price of a painting being a large multiple of the average share price, providing a diversified fund of artworks is not (yet) feasible. Attempts to provide a securitised market for art are under discussion, and the advent of such a trading place in Europe may not be far behind recent developments in China and the Far East. A securitised market place would provide a whole new index of its own!