Which artist brings the best return?

After five years of above-average growth, the art market has retreated.

Prices sank at Sotheby's and Christie's in November--in the case of contemporary art, to 2006 levels or below. A few works brought record highs, but these were far outnumbered by others that sold below their low estimate or failed to sell at all. Christie's offered "Study for Self-Portrait" by Francis Bacon for a low estimate of $40 million.

It was a Bacon no one brought home.

In Depth: Art That Beats The S&P

Over the long term, however, art as an investment has performed quite well. We know this from data we collect for our several Mei Moses Art Indexes, which track the market's performance back to 1957. Our complete analysis is available to subscribers at www.artasanasset.com. Henceforth, portions of it will also be available quarterly on Forbes.com.

Unlike stocks and bonds, works of art are unique. They trade infrequently. Thus, an index based on average prices over time may say more about the mix of objects that come to market than changes in values.

Behind the Numbers

To address this problem, the Mei Moses Indexes are based on repeat auction sales of the same works. Our database contains approximately 13,200 repeat sale pairs. (The Case-Shiller index for residential real estate works in a similar fashion.) Five subindexes cover the following categories of artwork: American before 1950; Impressionist and Modern; Old Master and 19th century; Postwar and Contemporary; and Latin American.

Comparing our All Art Index against the S&P 500 (with dividends reinvested) from 1958 through Oct. 31, 2008, shows that stocks have beaten art for the past quarter-century but have done worse than art for the past half-century. Also worth noting, if you think of art as pure investment rather than enjoyment, is that insurance, capital gains tax and transaction costs are much higher on art.

Two other facts of interest to investors: A diversified art portfolio would be less volatile than a stock index fund if you look at year-to-year returns but riskier if your holding period is decades. Also, the correlation between annual returns on art and those on stocks (or bonds) is close to zero. So art may play a positive role in wealth diversification.

What categories have fared best? During the first 10 months of 2008, the two top categories were Old Master and 19th century (with a gain of a little over 10%) and Impressionist and Modern (a little under 10%). Latin American also was one of the stronger performers.

The weakest category was American, with an initial loss of over 10% at mid-year and a current substantial loss of over 30%. Post War and Contemporary--hottest for the last 25 years--produced a lightly negative return for the first half of 2008 that has since become a loss of over 10%.

How much were unrealistic expectations to blame for November's poor results? Our data allow us to extrapolate the current value of a piece that hasn't sold in years. Take Mark Rothko's "No. 43 Mauve," which Christie's put up for sale in November with an estimate of $20 million to $30 million. The same work sold in 1988 for $1.5 million. If marked to market by our Postwar Index, its value would have been $7.6 million. No surprise it didn't sell.

We were curious to discover which artist has outperformed all others. To find out, we did the same thing Forbes does in evaluating new-issues underwriters: looking at all the price pairs we have for an artist, calculating an annualized performance and seeing how that stacks up against the All Art Index over the same time period.

Of the 25 most liquid artists--a group that includes Renoir, Picasso, Monet, Sargent and Hassam--Warhol wins, palette down. He averages 10.3 percentage points a year better than the Index. Of course, at any auction a lack of buyers or an inflated expectation by the owner may leave a work unsold. A Warhol portrait of Mao, supposed to fetch $4.5 million-plus failed to sell in a November auction.

Whether the art market's future will be like the last 10 years or the last 50 is, of course, impossible to predict. For persons wanting exposure to the broad market without having to buy art, a new instrument was created in November: futures contracts based on the 2008 closing value of our All Art Index.