Art Investment: An Empirical Inquiry

Abstract

Investment in and especially in paintings by the masters is often considered to be highly profitable. Particularly in the United States, but also in the richer European nations, there is an even larger number of investors who believe that buying works of is not only fun but also a good from a purely financial point of view. American banks have recently strengthened this trend toward art as an investment by employing art counselors, thus suggesting that it is a financially rewarding activity to engage in, and also suggesting that superior knowledge helps to improve financial success in this market. Buying and selling has in fact become an increasingly popular activity since the end of World War II. The auction houses report record turnovers and record prices are paid in increasingly rapid sequence. In April 1987 van Gogh's Sunflowers was sold at Christie's London for $39.9 million, only to be surpassed in November 1987 by van Gogh's Irises, sold by Sotheby's New York for $53.9 million (including the auction house's 10 percent commission). This last painting had been bought by the seller's mother in 1947 for $84,000, less than $0.5 million in today's money, which gives a real rate of return of about 12 percent per year. These enormous prices that are paid today for some masterpieces create a widespread belief that the rate of return on such investments is in general and on average very high. Books devoted to the subject [40, 385; 27, 24; 12, 215 et seq.] seem to confirm that the rate of return from in paintings, at least during the 1950s and 1960s, was far greater than the rate obtainable from comparable risky assets in financial markets. That the financial profitability in the 1970s and 1980s is believed to be, if anything, even larger than before is confirmed by many articles in financial and newspapers and journals. Up to now only very few scientific studies of the subject exist, but they also have serious limitations. The most recent one by Baumol [5] covers a period of over 300 years but ends in 1960. Two other studies by Anderson [1] and Stein [42] extend up to the end of the 1960s but do not go any further. All three studies are practically limited to transactions by Anglo-Saxon auction