This page considers art sale indexes and other data on
the sale of collectibles.
It covers -
It is complemented by a note
highlighting sources of data regarding nominal prices,
purchasing power, wages and inflation in different eras
Art sale records - sometimes characterised (disingenuously
or otherwise) as 'art indexes' - have become a marketing
tool for dealers, investors
and journalists. They have also become a subject of fascination
for academic economists and other scholars who, as highlighted
earlier in this note, have crunched sales date to identify
the most 'valuable' artist per square centimetre of canvas,
whether creators optimise their value through a short
career ('die young and leave a beautiful corpus'?) and
other expressions of postgrad Trivial Pursuit.
This page considers the nature of those indexes and of
data about other collectibles, along with comments regarding
information use and misuse.
Individual enthusiasts throughout history have recorded
the prices they paid for sculpture, paintings and other
collectibles (or for commodities as mundane as bread,
salt, beer and butter) or the prices paid by contemporaries.
That information can sometimes be disinterred from diaries,
wills, import declarations,
letters and bankruptcy appraisals. It features in some
of the retrospective wage/price indexes highlighted here.
Some vendors have publicly advertised prices. Newspapers
and other journals have often noted prices paid at auctions
or other public sales.
Compilations of comprehensive lists of sale prices and
analysis of that information to create composite figures
for classes of works, genres or individual artists is
however a recent phenomenon.
It essentially dates from the past forty years, ushered
in by the Times-Sotheby's Index (1967-1970) that collated
information from the major auction houses during a period
of exuberance. That index has been emulated by other compilations,
which have sought to attain the status enjoyed by more
traditional bond and share market indexes published by
or other financial specialists such as Moody, Standard
& Poor and Bloomberg.
Some sales indexes have remained records of prices in
public sales. In principle such records allow year by
year tracking of what has been paid for work by specific
artists, for classes of works (eg Chinese ceramics or
European paintings), for genres (Impressionist prints
and Expressionist prints) and even - when there are repeat
sales - for individual works. A crude version of such
tracking is provided by the chronology later
in this note.
Some publishers and consumers of sales records have processed
the data, offering a statistical analysis and figures
that are claimed to be representative of activity in markets.
That reflects measures such as the FTSE and Dow Jones
indexes that aggregate price movements across a class
of shares to indicate whether a sharemarket is 'up' or
'down' (and by inference may be up or down tomorrow, next
week, next month or next year and what rate the movement
will take place).
Unsurprisingly, there is a growing market for information
about the sale of fine art and, by extension, other collectibles
ranging from antique furniture to rare stamps, coins,
cars and violins. A sufficient
number of consumers are willing to pay for -
(eg lists of prices at a particular event or from an
auction house over several years)
publication (eg print and electronic sets of data that
cover several auction houses and dealers or that feature
statistical abstracts or that facilitate searching by
genre, artist, location, price bracket or other attribute).
of those consumers are institutions. Others are dealers,
auction houses, financial advisers, art market consultants
Prices for access to the data reflect perceptions of what
consumers will bear, the effort required to assemble the
data (eg through scrutiny of a flood of media releases
and even to have an observer physically present at a sale),
investment in a database or collection of files created
over many years, and the extent of competition. Some publishers
for example are offering substantial 'teaser' access to
online databases, with the expectation that revenue will
come by selling whole-of-site access to corporate clients
and from individual 'premium search' access.
Emphasis by some publishers on abstracts and search facilities
reflects recognition that competitors are scraping data
from their print and electronic publications for entry
into commercial databases.
Where does the data come from?
Contemporary indexes centre on data about public sales
of prices through attendance at auctions
scrutiny of prices published in media releases, annual
reports and promotional works from auction houses
of prices identified in media releases and catalogues
is important to note that those indexes do not comprehensively
cover private resales (eg direct from one collector to
another), first sales (by an artist or dealer to a collector)
and much work resold by dealers.
As a result they are biased towards 'high end' works and
to items sold by major auction houses.
Individual dealers/auction houses and specialist publishers
produce a range of print-format records - generally on
an annual basis - of sales records.
Those publications are useful for analysis of trends but
should be used with caution regarding individual sales,
given recurrent indications (including successful prosecutions)
that some reported sales to figures such as Alan Bond
were in fact only on paper.
analysis and action
Much index data is informative in retrospect, rather than
as a predictor. It will tell you where the market has
been (or where that market is claimed to have been), rather
than where it is going to go.
Art price indexes are problematical. Indexes typically
cover major works, rather than minor items. They cover
public sales (principally at auctions) rather than private
sales. Some figures are false (eg leading auction houses
have admitted that some reported sales actually didn't
occur or that the sale was strongly subsidised by a dealer
Consumers may draw the wrong conclusions from figures,
eg that prices will rise steadily, that prices for a particular
artist or genre will remain strong, that second-rank works
by a master will appreciate faster than those of 'safer'
first-rank works by that master or works by an emerging
Data in the 1870s encouraged investment in painters such
as Frith, Romney, Bougereau, Makart and Alma-Tadema. In
real terms those prices have never recovered. Buying 20
Van Goghs in 1900 looks very astute if they were held
for 70 years, unwise if sold after 5 years. Buying a cache
of Warhols or Pollocks in the 1960s for sale in 2007 looks
clever … but the investor would have got better
returns from some blue chips such as IBM and Xerox, or
from prescient investment in Apple and Microsoft.
Online data sources include -
Offline works of particular relevance for Australia and
New Zealand include Edward Craig's
multivolume Australian Art Auction Records (1975-2005),
with some 185,000 sale records from 54 auction houses.
faith in figures
The Art Newspaper editorialised in December 2006
in this murky world of rumour, affirmation, denial and
reaffirmation, are auction prices the only reliable
ones? After all these prices are made publicly, published
by the auction houses, listed on art data sites such
as artnet.com or artprice.com.
But even with auctions there's often more than meets
the eye. For example, Cézanne's Rideau, Cruchon
et Compotier, 1894, was sold at auction for $60.5m
(Sotheby's New York 1999), but the deal was never completed,
and the still-life was later resold for "significantly
less" to Steve Wynn. Most lists of the most expensive
paintings sold at auction still cite, somewhere in their
nether regions, Van Gogh's Irises, 1889, which
sold for $53.9m in 1997. In fact the work didn't sell
- Sotheby's lent half of the price to the Australian
Bond to buy it. He never paid up and the painting
ended up in the Getty, for - you've guessed it - an
In addition, auction houses are increasingly guaranteeing
works, meaning that they will give the vendor an agreed
price whatever the result of the sale. So, if a paintings
sells for $2m, but is guaranteed for $2.5m, the difference
is paid to the consignor. What then is the true price
of the painting?
All this demonstrates the poor transparency of the art
market, despite its huge - and growing - value. All
attempts to measure it accurately, to predict its performance,
to analyse it, either through sectors, individual artists
or "baskets" of works of art, have floundered.
As Sotheby's itself admitted in its annual financial
filing on 16 March this year: "It is not possible
to measure with any particular accuracy the entire international
Problems are not an aberration of the 1990s. In 1985 for
example Christie's New York executive David Bathurst admitted
that in 1981 he had reported selling two paintings that
had not in fact found buyers at auction: a Van Gogh at
a supposed price of US$2.1 million and a Gauguin at US$1.3
million. Bathurst said he had lied - nothing like noblesse
oblige - to protect the art market from depression.
next page (theft
and other loss)