Abstract
Art, along with other “treasure assets,” has become a central object for investment opportunities. Investment return studies using hedonic and resale price regressions on different artistic periods and styles produce estimates of varying rates of return, predominately low rates with high standard deviations. The present study employs a new sample of American art sold at auction between 1987 and 2011—art created before 1950 by 33 artists born prior to 1900. Our study, unlike those that preceded it, considers works that are no-sales (those “bought-in” for failing to sell at auction at or above a predetermined and negotiated minimum price), in addition to full transaction costs—buyers and sellers premia on hammer prices. We conclude that investment return calculations are biased upward and may be negative when these factors are considered and that the “consumption utility” of art may be higher than previously thought. However, using a variant of the capital asset pricing model, we find that investment in early American art may still be desirable in a diversified portfolio of assets for when the price of stock assets falls, the price of art does not fall in the same proportion.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Art galleries have made use of these data in advertising and sales. See Questroyal (2013). The Fine Art Fund, a private equity firm in London, advises investors in the global art market, and a number of other organizations have arisen for that purpose and for other “treasure” assets. Investors are often advised of the vagaries and capriciousness of the art market as a monetary investment.
The most complete introduction to the economics of art market is found in Ashenfelter and Graddy (2003).
A seller’s reserve price is the (secret) price under which she would not be willing to sell. One exception to studies that exclude transaction costs relates to collectible British stamps. Dimson and Spaenjers (2011) calculate the impact of a 25 % transaction cost on return concluding that the asset must be held for 4 years to compare with an equity holding of just 3 months. Frey and Cueni (2013) note that a similar cost may apply to art investments but do not formally assess that possibility.
Until recently, few pre-1950 American works found their way into European museums, but there were exceptions. Whistler’s “mother” (“Arrangement in Grey and Black,” 1871) and an important work by Winslow Homer (“Summer Night,” 1890) were included in French collections (now in the D’Orsay) in Paris and London’s National Gallery, in a bid to include American works as part of their “western European art” collection, purchased a masterpiece by George Bellows (“Men at the Docks” 1912) from Randolph College for $25.5 million dollars in 2014.
An active market developed in American art that has grown to this day, spreading from the center in New York City throughout the nation. The number of galleries dealing in early American Art (prior to 1950) and the number of auction houses selling it grew steadily with an estimated 200 galleries selling American art between 1913 and 1940 (Hall 2001), with hundreds more to follow.
This distinction is not different than identifying an “American” stock or bond market although, obviously, Americans purchase stocks and bonds of other countries as their citizens do those created in the USA.
Only 19 of Agnello's (2002: Table 2) sample of 91 American artists were born in the twentieth century and none after 1935, replicating a larger but similarly composed sample as the one here.
Diversification naturally reduces risk in art as well as all investments. Hence, the publically traded art “funds” mentioned in the introduction.
Picasso’s work (along with Andy Warhol’s) is generally considered the gold standard in modern art markets. Picasso’s investment returns have been studied for both prints (Pesando 1993; Pesando and Shum 1996, 1999) and paintings (Czujack 1997). Picasso prints do better than prints in general (real return of 2.10–1.51 % between 1977 and 1992) with paintings at 8.30 % between 1966 and 1994.
The elimination of transaction costs is probably more serious than Mei and Moses appear to believe. They note that “The return could be further reduced by transaction costs [auction fees to both buyers and sellers at auction]. We like to note, however, that return estimates for financial assets, to some extent, also could suffer from the same biases, such as lack of market liquidity, transaction costs, and survival” (2002: 1659). Transaction costs alone relating to art sales or purchases at auction are significantly higher than those for sale of financial assets, however. At the present time, the standard buyer’s premium for art is 25 % (for purchases up to $100,000) and for sellers, at least 10 %. If art sales are through galleries, the gallery “take” may be 20–40 or more percent of the sale price all attributed to the seller.
Their sample of artists was from the Annual Art Sales Index (Hislop 1971–1992) and excludes bought-ins—those works that did not meet the reservation price.
However, Agnello notes that “If the purchaser is knowledgeable, lucky, and can afford to buy the best quality (high-end works), painting values can do much better than merely holding their real value. Paintings at the very high end of the price spectrum yield a nominal return of 9.9 % per year which exceeds all the benchmarks except the S&P 500” (2002: 460). Mei and Moses calculate fairly high returns for pre-1950 American art (Questroyal Fine Art 2013), but their data are strictly proprietary.
In principle, it is possible to incorporate the buyer’s premium into a hedonic model; it is simply not done. Perhaps the reason for not dealing with it is that it is not at all clear how to do so appropriately. One way might be to treat price as “premium price” and the buyer’s premium as a condition of sale variable, but this would build spurious correlation into the model. In any event, since data on the seller’s premium are not published, the seller’s premium will always be a source of omission.
For a full discussion of the details of RSR (Graddy et al. 2012).
Without doubt, there are many paintings that were offered for sale once at auction by these artists and sold. Including these data in our hedonic analysis would surely produce more (statistically) efficient return estimates. We did not include these data in order to retain comparability of return estimates across methods; i.e., so that both hedonic and resale estimates would be based on the same basic data set.
For paintings that were sold three times, the first and second sales constituted one resale, and the second and third sales were taken as a second resale. Paintings offered at auction three or four times during the sample period, but sold only twice were counted as one resale; the no-sales in that set of offers were ignored. This set of 105 resales constitutes our repeat sales sample.
Bowley notes that the buyer’s premium—which falls to a marginal 12 % for work receiving more than $2 million hammer price—may be split with the seller. Thus, to attract high-priced material, auction houses may offer an “enhanced premium” paid to the consignor or to a guarantor of the painting. Naturally, such practices reduce the overall return and profitability to the house.
For any painting, the premium price (PP) equals the hammer price (HP) plus the buyer’s premium (BP). The 2011 buyer’s premium rates were 25 % on the first $50,000, and for paintings hammered down for between $50,000 and $1,000,000, 20 % on the excess of hammer price over $50,000. Thus, $362,000 = HP + [$12,500 + 0.2 (HP-$50,000)]. Solving for HP gives $299,583.33.
The results are not quite as “identical” as Table 2 would suggest. There are differences in the arithmetic and geometric means, usually starting around the fourth significant digit. Most of the similarities in the Table arise due to rounding for tabular presentation.
As prefatory analysis, we regressed ln(premium price) on a full set of 30 artist dummies (Inness was the omitted artist) and a constant term. Twelve of the dummy coefficients did not differ significantly from zero, indicating that the pricing of their works did not differ significantly from those of Inness. The eighteen artist dummies that were significant are the ones included in our hedonic model.
This is an approximation, and due to the size of the coefficient, it may contain considerable error. An exact calculation reveals that the average price of an oil painting is about twice the average price of an otherwise identical painting in another medium: i.e., [(P oil − P other)/P other] = (e0.699 − 1) = 1.012. Thus, the percentage difference is 100(1.012) = 101.2 %.
While neither study expressly states they use hammer prices, both use the Hislop (1971–1996) data, which records only hammer prices.
The average (and marginal) buyer’s premium prior to 1993 was 10 % of hammer price. From 1993 on, the average buyer’s premium as a percent of hammer price was 18 %. Hence, the 80 % increase in average buyer’s premium.
Because of the higher standard errors, we note that the trend coefficient in Model V is statistically significant at the α = 0.055 level and that the trend coefficient in Model VII is statistically significant at the 0.065 level; both based on a one-tailed test that they are not negative.
Methods of marketing and transaction costs are important to the highly competitive auction houses Sotheby’s and Christie’s as well. A chief strategy of Sotheby’s, from 2006 until recently, was to focus on selling higher-priced material in art and in other collecting interests. Earlier (between 1999 and 2003) Sotheby’s partnered with Amazon (Amazon.com) to sell lower priced items at lower transaction costs, but incurred losses of $150 million (Vogel and Issac 2014). Now, Sotheby’s (and Christie’s) will again begin marketing art through the internet. Sotheby’s now partners with eBay (paying eBay a percentage) to create an internet business selling art, mainly lower priced material at first. Christie’s will develop its own internet site and offer art in addition to other items. It is hoped that the brand credence of the two houses will cause the value of materials sold on the internet to rise.
There are other reasons for holding a portfolio of American art (as in our sample). We conducted a portfolio investment procedure (a modified Capital Asset Pricing Model) with respect to our sample of early American artists. Accepting the most likely estimates of premia for these artists produces a beta value which is low relative to many traditional asset classes, but it is positive and significant. This means that there is a potential role for pre-1950 art in a portfolio of diversified assets. The risk-adjusted value of holding that art follows the business cycle but does not go down as much as the market in a downturn. (Results are available on request from the authors).
Art markets do not carry instant or rapid and continuous feedback as in the financial markets, but it may be that auction frequency in fact is sufficient to engage in long-term speculation. Use of the internet (see footnote 27) may in fact reduce those costs relative to holding other assets. Additionally, the quality of art may reduce competition, as in some studies showing “masterpieces” to give better returns over time (Mei and Moses 2002, 2005), as does the length of holding most assets (real estate and other assets).
References
Agnello, R. J. (2002). Investment returns and risk for art: Evidence from auctions of American paintings. Eastern Economic Journal, 28(4), 443–463.
Agnello, R. J., & Pierce, R. K. (1996). Financial returns, price determinants, and genre effects in american art investment. Journal of Cultural Economics, 20, 359–383.
Ashenfelter, O., & Graddy, K. (2003). Auctions and the price of art. Journal of Economic Literature, 41, 763–787.
Beggs, A., & Graddy, K. (2008). Failure to meet the reserve price: The impact on returns to art. Journal of Cultural Economics, 32, 301–320.
Bowley, G. (2014). The (auction) house doesn’t always win: Christie’s and Sotheby’s woo big sellers with a cut. The New York Times (January 15). http://www.nytimes.com/2014/01/16/arts/design. Accessed 24 February 2014.
Currier, W. (1991). Currier’s price guide to American artists (1645–1945) at auction. Brockton, MA: Currier Publications.
Czujack, D. (1997). Picasso Paintings at Auction, 1963-1994. Journal of Cultural Economics, 21, 229–249.
Dimson, E., & Spaenjers, C. (2011). Ex post: The investment performance of collectible stamps. Journal of Financial Economics, 100, 443–458.
Frey, B. S., & Cueni, R. (2013). Why invest in art. The Economists’ Voice, 10(1), 1–6.
Galenson, D., & Weinberg, B. A. (2000). Age and the quality of work: The case of modern American painters. Journal of Political Economy, 108, 661–677.
Goetzmann, W. (1993). Accounting for taste: Art and the financial markets over three centuries. American Economic Review, 83, 1370–1376.
Graddy, K., Hamilton, J., & Campbell, R. (2012). Repeat sales indexes: Estimation without assuming that errors in asset returns are independently distributed. Real Estate Economics, 40, 131–166.
Guilbaud, S. (1985). How New York stole the idea of modern art. Chicago: University of Chicago Press.
Guilbaud, S. (1992). Reconstructing modernism: Art in New York, Paris and Montreal 1945–1964. Cambridge, MA: MIT Press.
Hall, D. M. (2001). Purchasing power: The New York market for modern American painting 1913–1940. Ph.D. dissertation. University of Northumbria at Newcastle.
Heckman, J. (1979). Sample selection bias as a specification error. Econometrica, 47, 153–161.
Hislop, R. (Ed.) (1971–1992). Various years. The Annual Art Sales Index. Surrey: Art Sales Index, Ltd.
Mandel, B. R. (2009). Art as an investment and conspicuous consumption good. American Economic Review, 99, 1653–1663.
Mei, J., & Moses, M. (2002). Art as an investment and the underperformance of masterpieces. American Economic Review, 92, 1656–1668.
Mei, J., & Moses, M. (2005). Vested interest and biased price estimates: Evidence from an auction market. The Journal of Finance, 80(5), 2409–2435.
Pesando, J. E. (1993). Art as an investment: The market for modern prints. American Economic Review, 83, 1075–1089.
Pesando, J., & Shum, P. M. (1996). Price anomalies at auction: evidence from the market for modern prints. In V. Ginsburgh (Ed.), Economics of the arts (North Holland) (pp. 113–134).
Pesando, J. E., & Shum, P. M. (1999). The return to Picasso’s prints and to traditional financial assets, 1977–1996. Journal of Cultural Economics, 23, 181–199.
Questroyal (2013) Art vs. stocks and bonds: A comprehensive analysis of the investment potential of American art, 2nd edn. www.questroyalfineart.com.
Reitlinger, G. 1982 [1961]. The economics of taste: The rise and fall of picture prices 1760–1960, 3 volumes. New York: Hacker Art Books.
Renneboog, L., & Spaenjers, C. (2013). Buying beauty: On prices and returns in the art market. Management Science, 59, 36–53.
Stein, J. P. (1977). The monetary appreciation of paintings. Journal of Political Economy, 85(5), 1021–1035.
Thompson, D. (2008). The $12 million stuffed shark: The curious economics of contemporary art. New York: Palgrave’s Macmillan.
Vogel, C., & Issac, M. (2014). Sotheby’s, eBay team up for live auctions. The New York Times News Service (The Business Times), July 15. http://economictimes.indiatimes.com/news/international/business/sothebys-ebay-team-up-for-live-auctions/articleshow/38412473.cms?from=mdr.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Anderson, S.C., Ekelund, R.B., Jackson, J.D. et al. Investment in early American art: the impact of transaction costs and no-sales on returns. J Cult Econ 40, 335–357 (2016). https://doi.org/10.1007/s10824-015-9252-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10824-015-9252-7
Keywords
- American art investment
- Financial diversification
- Investment return with full transaction costs
- No-sales in art investment returns