Art as an Asset Class

Ms. Luisa Ann Rubino 2001 MBA

Gustav Klimt, "The Kiss"

Primary Address (Home)

Via G. Meda 28

MILANO 20141

ITALY

Phone: 39-335-498003

Introduction

The goal of this study is to analyze art as an asset class. Specifically, I intend to investigate the risk/reward profile of art and the size of its risk-adjusted appreciation ("alpha"). I will also determine art's correlation with known benchmarks, its performance in "down markets", and therefore its ability to add diversification in a portfolio. Finally, I will examine the relation between art index, consumer sentiment, and future real activity.

Summary of Findings

Art as an asset class displays a low level of correlation with equity. Consequently, art has the ability to reduce the risk of a portfolio when combined with other asset classes.

What is Art?

Art can be viewed as the expression of culture, and art investing as an "investment in meaning". But what is its monetary value? Historical data shows that prices of paintings by well-known and less well-known artists have increased over centuries: this proves that true artistic relevance prevails in the long run and ultimately also finds an appropriate market value. Therefore the great challenge of the art investor appears to lie in identifying artistic relevance.

The first issue I would like to address is whether art is a financial asset in the first place. As Stein notes, works of art are extraordinary economic goods. They are at once durable consumer goods and financial assets. Their aggregate supply is non-augmentable for the works of deceased artists. They are extremely heterogeneous, with values that range from a few dollars to a few million dollars. When hung in museums, they are public goods in the sense that their viewing services are simultaneously available to everyone. They are speculative goods to the extent that demand determines future price appreciation and expected future price appreciation determined demand. In sum, they are the archetype of what might be called "collector goods".

So, is art a financial asset? There is a wide debate on this point. Those who agree that art is a financial asset seek to replicate the periodic cash flows generated using the rental market as a proxy. Alongside these cash flows, though, art undoubtedly offers implicit rents to ownership, that accrue from the aesthetic benefits of possession (the so-called "aesthetic dividends" and from the status derived from ownership and possession, for example from hanging up a painting in your house and displaying it to guests). In this sense, there has been a discussion (see Frey and Eichenberger) as to what extent, and under what circumstances, art is more likely to be a consumption good (with traditional collectors prevailing) or an investment good (with financial speculations prevailing). The fact that art may be a speculative asset appears to be confirmed by a 1953 survey, in which 63% of art buyers admit that speculation is their most important criterion. Those who disagree, point to the lack of liquidity in works of art. On the other hand there are other forms of investment, such as venture capital and real estate, that are considered financial assets without any doubt even though they probably have a lower level of liquidity.

So, given that art is an asset, and that investors are interested in the financial returns of this investment, the second issue I would like to address is the risk that this investment entails and what is the source of the risk. Analysis shows that the standard deviation of returns of investment in art is high. Goetzmann believes that the single largest source of risk for the art investor is taste - the possibility that a work of art will fall from fashion and become worthless. His analysis shows that purchasing a work by a well-known artist is just as likely to provide high returns as purchasing the work of a lesser-known artist. Also Pesando's work on prints disproves the popular notion that masterpieces outperform the market. As Proust wrote in 1922 in "Cities of the Plain" on the re-appreciation of art movements, "As on the Stock Exchange, when a rise occurs a whole group of securities profit by it, so a certain number of despised artists benefited from the reaction, either because they did not deserve such a scorn, or simply because they had incurred it."

Investing in Art

The difficulty of investing in the art market derives from the art market's inefficiency, from the know-how and expertise that the art market requires, and from the uncertainty surrounding tax issues.

The art market is definitely not an efficient market. The drivers of this are:

  • The traded products are differentiated: the art markets are seriously incomplete and often very thin, with very few works of art of a specific author traded each year.
  • Market transparency is low.
  • There are potentially large differences in expertise between buyers and sellers.
  • There is low liquidity: buyers and sellers are often distant in terms of space and time, and there is not always an available buyer or seller.
  • Transaction costs (for example, auction fees, insurance and handling costs) are by far larger than in other markets.
  • There are psychic benefits of owning art, which are largely absent in the case of owning financial assets.

Additionally, as Baumol notes, the art markets have a much weaker equilibration process (the process by which market prices tend to the equilibrium price) than other securities. This is because:

  1. Elasticity of supply is equal to zero for works of deceased artists.
  2. Each individual work of art is unique, while the inventory of a particular stock is made up of a large number of homogeneous securities, all perfect substitutes for one another.
  3. The owner of a work of art has a monopoly on that specific object, while a given stock is held by many individuals who are potentially independent traders on a near perfectly competitive stock market.
  4. The purchase and sale of a work of art is an infrequent occurrence and may happen even only once in a century.
  5. The acquisition price of a work of art is not generally public information and is often only known to the parties immediately involved.
  6. The equilibrium price is unknown, so an objective evaluation (such as a present value of future cash flows) is often impossible.

Successful investment in art requires not only extensive know-how about the artistic quality and authenticity of the object to be acquired but also about peculiarities of the art market. Additionally, it requires the investor to establish a scenario of future economic and social developments, also including international factors such as exchange rate movements, special cultural factors and market preferences.

Tax uncertainties add to the difficulty of investing in art and no study seriously takes into account the tax effect of investing in art. In many countries investment in art is one of the major possibilities of escaping or at least reducing the tax burden. And this is even more important as it is often unclear which country's tax code applies.

Past Studies of Returns in the Art Market

A. Studies

Many studies focus on analyzing and predicting price trends, both for individual artists and for schools of artists. I will briefly review some of the most significant studies.

Baumol concludes that art prices behave randomly. Further, he shows that large gains and losses occur with shorter holding periods, while the returns on longer holding periods are very close to zero (this is indicative of a random process with a mean of zero).

Stein treats works of art as a combination of consumer durables, yielding a flow of nonpecuniary viewing services, and capital assets, yielding a return from financial appreciation. He therefore divides the rate of return into two components, a return from nominal capital appreciation and a residual return from durable services (the return from viewing services, less insurance and maintenance costs, plus an annualized premium to account for any tax advantages, less an annualized premium to account for the illiquidity of this form of investment). The results show an estimated annual net return from durable services of 1.6% (i.e. to the investor primarily interested in financial gain and valuing the return from viewing pleasure at only about 1.6% per year, paintings are no more or less attractive than other assets, and yield the going rate for their systematic risk). In comparison, the 10.5% average annual appreciation of paintings in the US between 1946 and 1968 lends credibility to the assumption that collectors regard paintings as capital assets.

Goetzman finds a very high correlation between his art index and an index of London Stock Exchange shares; he therefore concludes that, while returns to art investment have exceeded inflation for long periods, and the returns in the second half of the 20th century have rivaled the stock market, they are no higher than would be justified by the extraordinary risks they represent. Additionally, he finds that the high correlation to both bonds and stock makes investment in art a poor vehicle for the purposes of diversification.

In a 1995 paper, Goetzmann uses a simulated portfolio method. This yields very different results with median annual returns of 8.24% in nominal terms and 2.42% in real terms, and with standard deviations respectively of 8% and 7.47%. In his sample of less than 100 paintings there are a few dramatic outliers but the dispersion of annualized returns is surprisingly small, suggesting that, at least for holding periods greater than a decade, the risk is modest. Additionally, this risk not only reflects the risk of the art market as a whole, but it includes the idiosyncratic risk of individual works, to the extent that it is not diversified away completely through inclusion in a portfolio. These returns clearly beat inflation, bond investment, and the capital appreciation of common stocks.

Weiland, Donaldson and Quintero examine the impact of US and Japanese equity markets on art prices, and find a significant relationship. In particular, their analysis shows that art price returns exhibited first-order autocorrelation and heteroskedasticity. Further, they find that art and stock prices shared a single, common, long-term trend. They believe, though, that investors can not make above-normal profits buying and selling art, because the art market is less liquid than stock markets.

Chanel's work shows that financial markets influence the art market (in particular they conclude that art never "causes" the stock markets, whereas English, Japanese and American stocks significantly "cause" art), even with a lag of about one year. He further uses a VAR (Vector Auto Regressive) model and concludes that lagged financial variables help predict art prices, even if the lag does not allow for systematic profits.

Candela and Sorcu study the Italian art market over the 1983-1994 period and concluded that art had lower returns than financial assets, explaining this with the aesthetic dividend and ownership effect. They also conclude that in the long run art prices are unrelated to financial asset prices, but find a positive correlation with real estate prices. Additionally, they show that the correlations are not contemporaneous, due to the different level of liquidity of the assets - equities being the most liquid and real estate being the least liquid.

Pesando finds a negative correlation of prints with T-bills and identifies a minimum variance portfolio - 94% Treasury bills and 6% modern prints - with a real return of 2.19% and a standard deviation of 3.19%. Wilke shows that for Contemporary Modern Art the hypothesis that prices of an artist's work only rise significantly after his or her death does not hold true.

Some of the results are summarized in the following table:

Article / Period / Mean / St. Dev. / Beta / Correlation
Baumol (1986) / 1652 - 1961 / 0.55 / 1.250000 (t=2.4)
Frey and Pommerehne (1989) / 1635 - 198700 1950 - 1987 / 1.8000 6.70 / 5.000000000 4.7
Goetzmann (1993) / 1716 - 198600 1850 - 198600 1900 - 1986 / 3.2000 6.2000 17.50 / 56.500 65.000 52.80 / 0.670000 0.790000000 0.78
Goetzmann (1995) / 1907 - 1987 / 5.75 - 11.13 / 8.00
Mastumo, Andoh, and Hoban (1994) / 1975 - 1989 / 16.00 / 17.00
Pesando (1993) / 1977 - 1992 / 1.51 / 19.90 / 0.310000 (t=1.8) / 0.30
Stein (1977) / 10.47 / 0.820000 (t=2.4)

B. Techniques

Studies that have focused on measuring the returns of art investment have generally encountered difficulties in building an adequate index. Another difficulty is the presence of the bias (both on the low and on the high end) that derives from using auction prices. The two techniques that have been used to build the indeces are:

  1. Repeated sales regressions that look at paintings appearing several times in the market
  2. Art price indices, some of which have been estimated with hedonic regression analysis. Some authors search for the underlying forces behind art movements (e.g. income, inflation, stock prices movement), while others look at the interdependencies between the markets for various types of paintings and of various locations.

Examples of indices that have been used in the past include:

  • Baumol builds an index based on Gerald Reitlinger's 1961 compendium of sales of art works by "...the best known painters of the world" over more than five centuries. He takes all the data relative to works of art that have been sold at least two times over a 300-year period and comes up with 640 transactions extending from 1652 to 1961. Once deflated, Baumol calculates the returns for each painting for the period between adjacent transactions.
  • Pesando builds an index based on the auction market for modern prints (which are multiples and therefore allow for a larger number of repeat sales) over a 16-year period from 1977 to 1992.
  • Stein builds his indices on US and UK auction prices.
  • Goetzmann constructs an index that includes paintings brought to market at least twice over the 1715 - 1986 period. He again uses the Gerald Reitlinger data and extends it with information from Enrique Mayer for the years 1971 - 1987. In another paper, he recognizes the survivorship bias of repeat sales methods, and therefore calculates appraisal-based returns to a simulated portfolio.
  • Weiland, Donaldson and Quintero use a sample of impressionist and modern paintings sold at auctions held by Sotheby's and Christie's in New York over the period May 1977 to May 1995. They calculate estimated prices based on a hedonistic model and use them to construct a price index.
  • Candela and Sorcu propose a price index based on estimated and auction prices. They use hedonic regression to construct a price index of a time-invariant "representative painting" (a refinement of the average painting method). Specifically, from the auctions in the period they consider, they calculate a grand distribution of the distributions of prices estimated by the auction house in all the auction sessions; this is meant to approximate the theoretical and unobservable price structure of the market. The grand distribution is used to identify the "representative painting", and the auction prices of the paintings included are then used to generate the "representative painting's" price.
  • Chanel uses the Mayer compendia (1963 to 1993) and bases his index on prices generated using hedonic regression.

Most of the data used in these analyses are based on auctions because the data are easily available and reliable, but disregard other sales, which may be quantitatively more important and show different price movements. Additionally, auction prices are really wholesale prices and do not form a good basis to calculate the returns to private collectors.

C. Summary

The long-term trend in inflation-adjusted art prices follows the general economic trend, with art prices rising above average compared to the prices of other goods. However, most segments of the art market react quickly to a worsening of the economic environment, and this is especially true for objects in the lower price category, with broad markets.

Results such as Goetzmann's have led researchers to conclude that art is an attractive investment only for nearly risk-neutral investors, and then only if the expected returns to art exceed the expected returns to stocks. The high correlation displayed in Goetzmann's studies between the art and the stock and bond markets clearly makes art a poor vehicle for the purposes of diversification.

According to an alternative interpretation of these results (see Frey and Eichenberger), art returns come from two different sources: financial return (change in monetary value) and phsychic return (consumption benefit of owning art). Therefore, if there is at least some consumption benefit, the financial rate of return of art objects should in equilibrium be lower than that in other markets with similar risks.

My approach

I have used Art Market Research data, run regressions and found that art has a low correlation with equity. It is therefore a good vehicle for diversification. I have attempted an optimization exercise, which gives a high prominence to art.

As far as opportunities for future research, if as many studies point out the market is not efficient, then there is the possibility to predict it and thereby generate excess returns, even more so than for more efficient financial markets.