Campbell Art Investment

7/31/2019 Campbell Art Investment

1/19

A rt as a F i n a n c i a l I n v e s t m e n tR.AJ. CAMPBELL

R.A.J. CAMPBELLis ail assis tant professorot finance at MaastrichtUnivers i ty and T i lburg Un i-versiry in Th e Ne ther lands .rcampl)ell@'nnancp.unimaas.nl

The comparatively poor perfomianceof traditional asset classes in recentyears has driven the search forgreater returns via alternative assetclasses. The idea of reaping higher risk-adjustedreturns from diversification into assets that offerlow and even negative correlation v^ith equi-ties and bonds is extremely attractive. Therehas been significant growth in the traditionalalternative investments s uch as real estate, com-modity futures, private equity, and hedge fundinvestments. Additionally, a number of fundsspecializing in art have recently emerged. Theseappear to offer a highly beneficial diversifica-tion strategy with extremely low correlationwith traditional asset classes. It is important forinvestors to understand the risk and returncharacteristics of this new alternative asset class.In this article, the author takes a closelook at art as an alternative asset and examineshow this new alternative asset is expected toperform. The author focuses on bear markets,when the benefits of diversification are mostneeded. Th e autlior looks at the risk and returncharacteristics of art using art market indicesand analyzes the prospects for portfolio diver-sification in the art market by using a varietyof data across art market s ectors, including theOld Master, European Impress ionist, Mo dern,and Contemporary art markets. Because ofthelow correlation of art with other asset classes,the author finds opportunities for portfoliodiversification across art markets and across

asset classes. The results hold, even allowinfor the high transaction costs that are en cou ntered when trading art w hen s pread over longer time horizon.The possibility of investing in art harecently generated much interest amoninvestors worldwide. Direct investment in ar

is, of course, not new. However, structuresoiutions offered by art funds and a numbeot boutique funds offer investors the possibilitof investing in a diversified art portfolio whicactively trades in art purely for financial gainThe most established is The Fine Art Fundlaunched in 2003 in London, and since theARTESTATE, Socit General Asset Management, and, more recently, the Art TradinFund have all raised sufficient capital to provide investors with indirect investments in thart market. There is also a move towards morspecialized funds that focus on one or twmarkets, such as Indian Art, Chinese art, oContemporary artists. The majority of thesfunds actively trade their artworks, ARTESTATE being the current exception (it aims thold a limited nuniber of artworks for the duration ofthe closed end fund). ARTESTATalso has a low entry level at 82500, whereamany other funds are focused more towarwealthy investors. These funds use a widvariety of trading strategies, similar to botprivate equity and hedge funds, trading on thinefficiencies currently present in the armarket, which is typically characterized by low

64 A R T A S A F [N A N C :IA L IN V E S T M E N T ; 200


2/19

liquidity. O f course, in some cases, hedge fund managersaccess the art market indirectly.The interest in investing in art has received ati etaor-mous boost from the availability of art price data. Data-

bases, indices, and market reports are now essentialanalytical tools wi th wh ich art investors can assess finan-cial performance. A number ot indices show averagereturns for artists and market sectors with data rangingfrom the 17th century to today. This article will focuson the use of available indices for various art markets toassess arts performance in a diversified portfolio. Theseindices provide a reliable estimate of the historical risk-return profile, and serve as benchmarks that art fundsaim to outperform. Obviously, the more speculativetniding strategies of some art fund managers wiU aim tooutperform the market to a considerable extent, and therecent performance of the more established funds hasgenerally shown this to be the case. This article willexamine return patterns that an art investment portfoliowould have provided, taking a more conservative.ipproLich to e xam niing the financial gains that artworkshave generated historically. This study also includes trans-action costs, which can be considerable in the art market,althou gh, in practice, art finids typically are able to neg o-tiate on these costs.In the following sections, the author reviews the

current literature and discusses current data on art indicesand the associated methodologies. The author explores therisk and return characteristics of ftne art tnarkets and ana-lyzes art as an alternative asset class in an internationaldiversified portfolio. The author accounts for the hightransaction costs encountered when auctioning fine art-works as well as the implications of smoothed returns,which occur tor assets that are appraisal based. Because ofthe m oderate returns found for art investments in tlie last30 years and the low correlation with other asset classestliat art appears to exhibit, the author finds a case forholding a small percentage of the investment portfolio inart. Currently, it is inconceivable to hold an index trackingfund; however, there are a number of alternative ways tohold a diversified art portfolio as part of an overall wealthmanagement strategy.ART INDICES

Only a limited number of studies have attemptedto construct art price indices. The first significant studywas by Stein |1977], w ho looked at average prices. More

prom inent studies using repeat-sales regression followed:Bauml [1986]; C.oetzmann |1993], and Pesando 119931.Stein [1977] pointed out the selection bias derived fromlooking only at repeat sales on auction house data; how-ever, average prices also suffer from serious biases in ahighly heterogeneous market. Repeat-sales regressionsrequire artworks to be offered for sale at auction morethan once to be included as a repeat sale. The annual artreturn index provided by Goetzniann is fairly extensivedating troni 1715 to 1986, with art market returns since1850 providing a higher return than stocks or bo nds, albeitwith a much higher standard d eviation. Pesaiido provideda semianiuKil index for the shorter 1977-1992 period,which includes the collapse in prices at the end of the]9K0s, not covered in the Goetzmann index, and givesreturns for art well below both the equit\' and bond mar-kets, while the variance ot these returns is similar to equitymarkets. The most recent index is the Mei atid Moses[2(H)2| index using U.S. data.

Anderson [ 1974|, and later Beulens and Ginsburgh[1992] used hedonic pricing models that also analyze thespecific characteristics of the artworks, such as size, artist,and art style. More recently, there has been a move toincorporate hedonic pricing within a repeat sales frame-work. Zanola [2007] provided indices that more accu-rately establish the price determinants in art valuation.Data and MethodologyTh e M ei Moses and Art M arket Research art indicesare the two most widely quoted indicators of art marketperformance. Both rely on data from sales at the mainauction houses. However, auction results alone provide anincomplete picture of the market performance becausethey represent only a subset of the whole market.The dealer market is largely ignored because of anabsence of obtainable data. The re is some disagreement asto what percentage of the market is composed of dealers.Figures troni two recent studies range from a 50-50 splitbetween auction houses and dealers to a 7030 split infavor of dealers. In any event, it cannot be denied thatdealers have a significant, albeit un quantifiable, impact uponthe art market. The absence of dealers' transactions fromthe art indices may have a bearing on the rate of returnreflected by the indices. This is because dealers may buyat lower prices but sell at prices with higher transactioncosts, thereby reducing the art investors' rate of return,'ft is likely that art funds, which act more like private dealers

C 200 Tun JOURNAL OF ALTFUNATIVE INVESTMENI'S 6 5


3/19

than auction houses, adopt a similar strategy by using the irinsider know ledge and expertise to exploit inefficiencies inthe market. This is likely to produce art tnarket returnsmuch larger than the benchmarks used here.Four primary methodologies are used to constructart price indices: geometric means, average prices, repeat-sales regressiotis (RSR), and hedoiiic regressions. Chanel,Gerard-Varet, and Ginsburgh's [ 1996] study indicated thatover long periods, the respective m ethodologies are closelycorrelated. Issues regarding the various index pricingmethodologies were documented by Ginsburgh, Mei,and Moses [2006], which specifically compared hedonicto repeat sales regression. Ashenfelter and Graddy [2003]provided a survey of average returns estimated from artprice data, currently in the academic literature. Exhibit 1

gives the estimated art market performance found by theseand a tew additional studies.For the pu rpose of this comparison, this article willfocus on the data from Art Market Research because itprovides a wider and more frequent source of informa-tion. The author also provides some comparisons withthe Mei Moses All Art index. These indices show thathistorically, average real returns for art are moderate.Returns are above inflation and tend to be greater thangovernment bonds but less than equities. There has beena general upward trend in art prices in the market-The survey of art pricing methodologies in Exhibit 1tends to indicate that the repeat sales methodology providesslightly higher estimates ot average returns than the othermethodologies for similar time periods. For example,Anderson [1974] provided RSR and hedonic price indicestbr the periods 1780-1970 and 1780-1960 and Ghanel,Gerard-Varet, and Ginsburgh [I996| for the 1855-1969period. It is of interest to observe the long-run trend in tliemarket and note that there have been periods In whicli artreturns have been substantially higher than average.To evaluate the various index methodologies, the

author uses data from both Art Market Research (AMR)and M ei Moses (MM) All Art Index. A M R data are avail-able monthly but only go back to 1976. The authorincludes all available data for each sector. It is importantto include the entire distribution in the indices becausethis takes into account the extreme price movements inthe m arket that are vital in correlation estimation and theanalysis of diversification benefits. A M R data uses averagereturns on a 12-month moving average.Tlie MM series for the All Art Index dates back to1875, measured on an annual basis, and to 1965, on a

semiannual basis. The MM All Art Index is computusing repeat sales initially sold at auction by Sotheby's aChristie's.Exhibit 2 provides summary statistics for the twprice index methodologies. To compare the two serisemiannual data from 1976 to 2002 are used.- Using semannual data rather than m onthly increases the series' annuvolatility, and the shorter time period results in a slighlower average annual return.For all indices, calculate the return ofthe marketby continuously compounding returns. This is more apppriate than measuring cumulative returns. The returnthe natural log return ot the price index at time, /, such tlA/J.j denotes the rate of change ofp. i

AP =]n\ X 1UP.Exhibi t 2 shows that the average retu rn on the Mdata series for the 27-year period was mu ch h igher thw h e n using the A M R data . Using repeat sales, the averare turn on an annual basis is over 10% , wh ereas the A Mgeneral art index is jus t over 5.25% (8% for U.S . artists a5% for artists in the U.K.).-^ Stein | I 9 7 7 | , G o e t z m a[1993], and Ginsburgh [2006] acknowledged the se le

t ion bias that occu rs from focusing on repeat sales. To included in the calculat ion, repeat sales regression requiartworks to be otfered tor sale at auctions more than onIt is thought that artworks that fal l drast ically in valt end not to he resold at auctions.""C o m p u t i n g the c orre lation statistics for the tw o dferent index methodologies reveals that the corre la t io f t h e A M R ind ex wi th the M M Index is only 0 .2 (sPanel B of Exhibit 2). This is because of s moo th ing in tA M R i ndex .Tak ing a two-p eri od mo vin g average for the re tu

series increases the c orre lat ion dram atically. Thi s is espcially t rue for the All Art In dex and the U.S. 100 indicw h i c h have a correlat ion coefficient of 0.86 . The largn u m b e r of observat ions used for the moving averagprovides higher correlat ion coefficients (see Panel C Exh ib i t 2).T h e s e resul ts indicate that the two me thodo logiresult in indices that are good pro xies of art market pr icof auct ion sales data.T h e col lec t ion of infor ma t ion from databases however, problemat ic for a n um ber of reasons. Ashetifel

66 AKT AS A FINANCIAL INVESTMENT ; 2


4/19

E X H I B I T 1Estimated Fine Art Market Performance, 17th to 21st century (as reported by variotts academic papers, by periodof study)

AatborBauml 119861['rey and Pommerehne119891

Buelens and G insburgh119921

Goetzmann | I 9 9 3 |

Anderson | I974 |Chane l , Gerard-Vare t ,and Ginsburgh | I996 |

Meiand Moses |2OO2|

Coetzmann |1996|Fase II9961Stein | t977|Barre. Oocclo, andGinsburgh | I996 |Czujack |I997|Deutschman 1199I|Angnello and Pierce[19961Campbell |2OOS1

Pesando |1Q93|E'csando and Sbum119961Frey and Serna [1990|Candela and Scorcu[1997]

SamplePaintings in GeneralPaintings in General

Paintings in GeneralPaintings in GeneralPaintings in General

Paintings in General

Paintings in GeneralAmerican. Impressionist,and Old Masters

Paintings in General19th CenturyPaintings in GeneralGreat ImptessionistOther ImpressionistPicasso PaintingsOld Masters19th CenUiry USPaintings in GeneralUS PaintingsModem PrintsPieasso PrintsOld MastersModem and ContemporaryPaintings

Period1652-19611635-19491653-19871950-19871700-19611780-19701716-19861S50-I9861900-19861780-19601780-19701855-19691855-19691875-19991900-19861900-19991950-19991977-19911907-19771946-19661972-19921946-19681962-199]1962-19911966-19941971-19911971-19921976-20041976-20041977-19921977-19921981-19881983-1994

MethodRSRRSRRSRRSRHedonicRSRRSRRSRRSRHedonicRSRHedonicRSRRSRRSRRSRRSRRSRRSR

GeometricMeanHedonicHedonicHedonic

Average pricesAverage pricesRSRRSRHedonic

Nominal Return

3.70%3.20%6.20%17.50%3.30%3.70%

11.00%10.60%10.47%12.0%8.00%12.30%9.30%5.73%7.94%

12.00%10.59%3.89%

Real Return0.60%1.40%1.50%1.70%0.91%3.00%2.00%3.80%13.3%2.60%3.00%4.90%5.00%4.90%5.20%5.20%8.20%7.80%5.00%7.50%1.10%

5.00%1.00%8.30%6.04%3.25%1.44%3,66%1.51%2.10%3.20%0.21%

Standarddeviation

5.00%

**5.65%

6.50%5.19%

*

4.28%3.72%3.55%2.13%2.11%

*

8.27%8.73%19.94%23.38%

teiitrns estimaied addiiionally hy A.f:heiifelit'r and Gniddy.

Ti IL JOURNAI or AT'KNATIVE INVESTMENTS 6 7


5/19

E X H I B I T 2Comparison of AMR Average Price data and MM Repeat Sales Indices, 1976-2002

Panel A.Semiannual log return data 1976/01-2002/12

Annual A verage ReturnAnnual Average St DeviationAverageStandard DeviationSkew nessKurtosisPanel B. Correlation Matrix

ART 1005.27%17.11%0.0260.121-0.8371,694

Average PricesUS 1008.26%15.86%0.0410.112-0.8171.029

Semiannual log return data 1976/01-2002/12

Art 100US 100UK 100AI! Art IndexPanel C. Correlation Matrix

ART 1001,0000.8220.6510.210

Average PricesUS 1001.0000.5650.221

UK 1005.12%11.10%0.0260.078-0.097-1.083

UK 100

1.0000.250

Semiannual log returns2 period moving averages 1976/01-2002/12

Art 100US 100UK 100All Art Index

ART 1001.0000.8710.7160.857

Average PricesUS 1001.0000,6440,714

UK 100

1,0000.342

Repeat SalesAll Art Index

10.07%21.88%0.0500.155-0.277-0.395

Repeat SalesAll Art Index

1.000

Repeat SalesAU Art Index

1,000and Graddy's [203] study contended that an empiricaldiscrepancy in one year can materially alter tbe overall rateof return by up to 5%. Evidence o this phenomenoncan be found when the Mei & Moses All Art Index iscompared with the General Art Index of Art MarketResearch for the 19762002 period. A difference in theirestimations of the return after the art market bubble burstin 1991 results in a significant difference between theaverage return figures thereafter. This can be observedin Exhibit 3, where both indices are plotted together.The repeat sales index does not capture as significant adownturn as the AMR data does.

This difference also indicates the importance ofliq-uidity during downturns in the art market. The numberof art sales is likely to be greatly reduced in downturns.

with the market becoming more illiquid. The art investfaces a greater degree of liquidity risk tban investors other hnancial assets. When artworks tail to reach thereserve prices and are sold, the price indices are affecteFewer transactions result in larger estimation errors. Apresent, little information is available on market liquidiover the empirical time series. This problem is especialsignificant for repeat-sales regression estimations, whiare constructed with fewer observations. It is likely ththe price estimation error that occurred after the art markcrash in the early 1990s was because of this issue wirepeat-sales estimation. Mei and Moses suggested that adoes significantly better during wartime, using the exampof four U.S. wars (Forbes |2()()11). During these periodart appears to outperform stocks. This finding may al

68 Aii,T .\s A FINANCIAL INVESTMENT SP IUNG 2(


6/19

E X H I B I T 3Comparison of Art Price Indices, 1976-2002Repeat sales All Art Index versus the average price indices from Art Market Research tor the general art market (Art HKl), a basket ofU.S. artists(US 100). and a basket of British ardsts (UK 100).

Repeat Sales vs. Average Price IndicesSemi - annual data180001600014000

-ART 100 - US 100- - - UK 100

All An Index

(N f lS S 8M N 4average price data for the V.S., U.K., All .4rt Index, and MM Repeat Sales hidiics, 1976-2002, using semiatniuat data.

be because ofthe lack of liquidity during these periodsand is a highly interestitig point that requires further inves-tigation.

Evidendy, fundamental problems exist with art data-bases and indices. However, both databases and indicesare becoming more sophisticated and accurate at pro-viding objective information. Comfort can be taken fromthe fact that tbe Standard ik Poor's index was recentlyoverhauled. If well-established, traditional investmentindices are still tweaking their assumptions, art indicesshould be allowed to refme their models over time.

Although the information provided by the databasesand indices is not complete, it is the best market infor-mation that is currently available. Tbe information pro-vides us witb a somewhat robust indication ofthe generaltrends in tbe market. Moreover market anotnalies andinef iciencies may lead to much higher realized returns.

Investment skill hes in interpreting the available infor-mation, assessing wbether the risk-return ratio is acceptableand deciding whether the investment is appropriate to anexisting portfolio. Taste adds an additional unquantifiable

element of risk to art investment even after market analysishas been undertaken. Art as a direct investment presentsa risky investment opportunity, although purchasingaccording to personal taste results in an aesthetic benefitthat can potentially outweigh any fmancial benefit or lossincurred.

When considering art as an indirect investn^ent,where the non-pecuniary benefits are not obtained, aninvestor would be advised to opt for an alternative invest-ment vehicle (AIV) or art mutual fund (AMF), in wbichrisk diversification through tbe securitization of art-works is more likely to result in greater fmancial returns.

Fine Art Market PerformanceTo analyze tbe performance of a variety of art mar-

kets, tliis article focuses on the indices produced by AMR.These indices also allow for a breakdown ofthe fine artmarket into various sectors. For the various schools,movements, and periods, the average prices of sales byindividual artists are combined to form an equally

20()S T H E J O U R N A L (IF ALTf-.RNAllVh INVFSTMENTS 6 9


7/19

weighted portfolio. This analysis uses the General artindex as well as the following four sectors of the artmarket: Old Masters, European Impressionists, Modern,and Contemporary.

The General art index contains a mixed basket ofover 100 well-known artists ranging from Basquiat toCanaletto. The index covers a variety of artists from dif-ferent sectors and countries, constituting a diversifiedindex of art. The index comprises art sales data from over109,000 auction sales.

The Old Masters index consists of European artistsuntil the 18th century There are over 25,000 sales includedin the index with artists from Brueghel [1568-1625] toConstable [1776-1837].

The index for European Impressionist art containsa smaller sample of 25 artists, tor example, Manet[1832-1883] and Matisse [1869-1954]. The periodincludes European Impressionist artists in the late 19thcentury and also some Post-Impressionists. The numberof sales included in the index is lower than for other sec-tors, with just over 22,000 prices included.

Modern art contains a higher number of artistsand sales prices, with over 63,000 transactions included.

These range from Kandinsky [18661944] to Bac11 9091992], There may be some disagreement amoart historians about the exact definitions ofthe classication ot Modern art.

The final sector is Contemporary art, for whithere are over 21,000 sales included in the data. The indis newer with d;ita starting in 1985. Artists covered incluFreud [1922-] and Hirst [1965-].

The choice of artists, which is shown in tappendix, is a highly subjective but representative choicThe indices, therefore, provide a general indication of asector price movements.

Exhibit 4 shows that prices over the past 20 yeafor the various sectors have at times diverged quite sustantially, particularly during the period ofthe 1988-199bubble, which affected Impressionist art more than othsectors.

Including contemporary art and rebasing the indicto January 1985. when all series are complete, it can seen that the contemporary art sector has outpertormall others over the past 20 years. Impressionists were tlowest performers, v^nth their greatest returns having beemade iti the late 1970s.

E X H I B I T 4Performance of Various Fine Art Markets, 1985-2006

ARTSECTORSJAIN 1985-FEB2006

10000GENERAL ART

OLD MASTERS

EUROPEANIMPRESSIONSTS

MODERN 100

CONTEMPORARY 10

Note: AMR awriige price data for the irt market iectors usin^ iiioiithly i/uM from JiWimry 1976 to Fclmiiiry 2006.

7 0 ART .^s A FINANCIM INVESTMENT Si 'RINC 2(1


8/19

RISK AN D RETURNUsing data from January 1980 until February 2006gives 25 years of mo nthly return data for a variety of sec-tors. Over this period, the general art index has made an

average annual return of 6.5%. More specifically. OldMasters have generated 5.5%; European Impressionists,6.3%; and Modern, 7.5%. Contemporary (data starts in1985, so using the slightly shorter 20-year period) haveoffered the highest returns at 9% on an annual basis. Usinga representative, liypothetica! fund that holds a com pos i-tion ot 3(t% Old Masters, 15% European Impressionists,15% Mo dern, and 40% Contemporary, the average returnusing data frotn the various sectors would have been7.05%>. Again, this is for the 20-year p erio d, because dataon Contemporary art starts in 1985.

Descriptive statistics are shown in Exhibit 5. TheEuropean Impressionists have been the most volatilemarket with an annual average standard deviation oft he

series of mo re than 15%. Old Masters have been theleast volatile with only a 7% average annual standarddeviation.A closer look at the risk and return characteristics ofthe sectors, focusing specifically on the 25-year period liom1980 to 2006 (this period is chosen because other asset classdata is also only available from 1980) shows that the fine artindices are themselves not highly correlated. This gives anindication ot the potential benetits trom holding;! diversi-fied art portfolio across artists and across various art sec-tors. The highest correlation over the period is betweenthe general art index and all other sectors, most hkelybecause each of the individual sectors feeds into the gen-eral art index. T he correlation coefficients range betweeti(1.27 and 0.53 for the four individual art sectors. Modernart and the general art index have a correlation of 0.76.An examination ofthe return-risk ratio ofthe var-ious sectors shows that Modern and Contemporary artoffer the highest return for a tuiit of risk, where risk is

E X H I B I T 5A. Descriptive StLilislics lor Fine An Mitrkcls. I'lSO 2

AVERAGEANNUAL RETURNAVERAGEANNUAL ST. DEVAVERAGEMONTHLY RETAVERAGEMONTHLY ST DEVS K E W N E S SKURTOSISRETURN/RISK

FUNDC O M P O S I T I O N *

7.05%

6.92%

0.006

0.020

0.544

2.0991.02

G E N E R A LART

6.56%

8.08%

0.005

0.023

-0.5183.583

0.81B. Correlation Statistics for Fine Art Markets 198O-2006

GENERAL ARTOLD MASTERSEUROPEAN IMPRESSIONISTSMODERN 100CONTEMPORARY 100

O LDMASTERS

5.52%

7.09%

0.005

0.020

0.1490.469

0.78

EUR CON-IMPRESSION MODERN TEMPORAKY *

6.30%

13.12%

0.005

0.038

-0.1123.5100.48

GENERA L OLD EUROPEANART MASTERS IMPRESSIONISTS1.0000.4300.7400.7610.534

1.0000.3380.2990.279

1.0000.6110.346

7.55%

3K%

0.006

0.021-0.1463.1131.02

M O D E R N100

1.0000.530

9.00%

9.90%

0.007

0.029

1.0483.562

0.91

C O N T E M P O R A R V100

1.000Note: AXiR average price dala for the an market sectors usinj monthly data from iniary 1980 to Ffhni.iry 2006. The diiiii for ilic (Joiiivinponiry ,ntiiHii thv Fund composiiioii iisinj; the Contemporary an market do not sitirt until fiuiiiary 1985.

; 2008 TH E JOURNAL OF AITFRNATIVE INVESTMENTS 7 1


9/19

measured by the standard deviation of returns. Per unitof risk, the fund composition also offers an attractivereturn of 1.02. Although the average return is slightly lessthan tor the Modern and Contemporary markets, the riskis alleviated through a well-diversihed portfolio in whichreturns per unit of risk are as high as for Modern art.

The risk-return trade-off can also be depictedgraphically, as shown in Exhibit 6. Generally, there is apositive trade-off between risk and return. The relation-ship of a higher expected return required for an investorto face greater risk is an underpinning of modern fniancetheory. The higher the return and the lower the risk, themore desirable the index from a ftnancial point of view.In this case, the most attractive position from a ftnancialpoint of view is the top left-hand corner ofthe graph.This is illustrated tor both the Modern and Fund com-position markets.

Asset Class FrameworkThe financial markets analyzed represent the major

asset classes. The author uses the Morgan Stanley Capital

Indices^ for U.S. equity (MSCI US), U.K. equity (MSUK), and world equity (MSCI World), Lehman BrotheAggregate Corporate Bond Index, and North AmericReal Estate Investment Trust Index (NAREIT). For thedge fund data series, the author uses the CredSuisse/Tremont Hedge Fund data series dating from 199S&P GSCI commodity futures data is from Goldman SacThe author uses the U.S. and U.K. ID-year GovernmeBond Indices, and U.K. Govemtneiit Treasury Bills, whihave been available on a monthly basis only from 198Data are collected frotii Datastream, Global Financial DaNAREIT, and Credit Suisse/Tremont. Descriptistatistics are given for a variety of time horizons for aasset classes in Exhibit 7.

Exhibit 8 plots the general risk and return trade-ofor the variety of asset classes. Because of the smoothnature ofthe art market return series, the exhibit alincludes information for the desmoothed art index, whicaccounts for the moving average in the series. The risksubstantially higher for the same level of return and henshould be more reflective of the true volatility in t

E X H I B I T 6Risk-Return Trade-Off for Fine Art Markets, 1980-2006

10.00%9.00%8.00%7.00%6.00%5.00%4.00%3.00%2.00%1.00%0.00%

0.00%

Jan 1980-Feb2006

FundComposition*

Old Masters

ModernGeneral Art

Contemporary*

European

2.00% 4.00% 6.00% 8.00% 10.00%Average Annual Standard Deviation

12.00% 14.00%

*diitiifrom 1985.Note: AMR average price data for the art market sectors usii{ monthly data from January i 980 to February 2006. The data for the Contemporary aatid the Fund amipoiitiou using the Cofitcmporar)' art ntarkei do not sMri until January 1985.

72 ART AS A FINANCIAL INVESTMFNT SPRING 2


10/19

E X H I B r T 7Descriptive Statistics for Fine Art and Financial Markets

25 Years

A. 1980-2006Annual Av ReturnAnnual St DevAverageStDevSkewKurt15 Yean

B.1990-2006Annual Av ReturnAnnual St DevAverageStDevSkewKurtS Years

C.2000-2006Annual Av ReturnAnnual St DevAverageS t D e vSkewKurt

WorldEquity10.88%13.93%0.0090.040-1.1753.866

WorldEquit>'6.78%14.00%0.0060.040-0.8271.217

WorldEquit) '0.06%14.23%0.0000.04!-0.5963.SS5

USEquity12.39%15.16%0.0100.044-0.8473.404

USEquity10.28%14.35%0.0090.041-0.5890.935

USEquity-1.28%15.29%-0.0010.044-0.2743,381

UKEquity13.27%16.63%O.Oll0.048-1.2705.755

UKEquity8.60%14.27%0.0070.041-0.4910.623

UKEquity0.93%14.14%0.0010.041-0.8965.480

USCorporate CommodBonds Futures14.91%22.72%0.0120.0663.17450.560

USCorporateBonds8.33%4.53%0.0070.013-0.3760.590

USCorporateBonds7.18%4.20%0.0060.012-0.96742.897

8.42%17.44%0.0070.0500.0711.252

CommodFutures7.36%19.38%0.0060.0560.2470.910

CommodFutures13.11%22.55%0.0110.065-0.188I.I90

USIOY UKIOYGovt Bonds Govt Bonds8.36%8.62%0.0070.0250.1381.274

USIOYGovtBonds6.52%6.99%0.0050.020-0.4440.829

USIOYGovtBonds6.19%7.64%0.0050.022-0.7731.274

11.00%8.57%0.0090.025-0.2111.413

UKIOYGovtBonds NAREH9.02% 12.10%6.75% 12.89%0.008 0.0 H)0.019 0.037-0.076 -0.6690.631 2.123

NAREIT11 .98%12.88%0.0100.037-0.7662.918

Hedge' Funds*10.31%7.79%0.0090.022-0.0352.367

UKIOYGovt HedgeBonds NAREIT Funds6.35% 19.10%4.62% 14.06%0-005 0.0160.013 0.041-0.338 -1.4881.413 2.823

7.53%5.09%0.0060.0150.2042.367

Art6.56%8.08%0.0050.023-0.5183.583

Art-1.26%7.54%-0.0010.022-1.4594.751

Art3.56%5.17%0.00300.01491.38334,9021

Moifs: Bqiiiiy ifiilkes ar c from Morjâii Shuihy Cap ital iiJiirs for U.S. equily (MSCI US), U .K. i'i]tiity (MSCI UK), a mi world equity (MSCI IVorld).U-liinmi BrotluTs Aiît\i>atc Corporate Bond tndcx (am ilabk only for the U.S.) nnd Nortii America n Real Esiaic liweslmcnl Tntsl Index (NARHil'). Forthe Imi^êfund data series, tin- Credit Suisse/Tretmm t Hedge Fund data series datingom 994 is used. Vie U.S. and U.K. lO-year Goivrntnent Bondindices, and U.K. Covernment Tr easury Bills have only heen auailable on a monthly basis from 1980. S&P CSCI amm odity iturc data is amilahle fromColdnhiii Sachs. Other data were collected frotn Datastream . Global Financial Data. NARBIT, and Credit Suisse/Tremont. All .Art Index from Art Mar ketR.csamh. All other data is monthly from fanuary 980 to February 2006.* Hedge Funds are the Tremont Hed' Fund Data.

market. This desmoothing process is common in thefinance literature for real estate and hedge funds. Thisdesmoothed data is also used in the analysis on optimalportfolio allocation.Exhibit 9 gives correlation statistics for the 25-yearhorizon. Art has a low correlation with other asset classes:the highest being commodity futures with a monthly 0.09correlation and the most negatively correlated beingNAREIT, whose returns arc correlated at -0.08. The cor-relation with domestic real estate and art tends to be higher.

Correlations with other asset classes remain low evenafter accounting for various time horizons, as shown inExhibit 10. During the recent bear market for equities,when commodity futures prices, government bondindices, and real estate markets all rose, the correlationbetween art and oth er asset classes has been positive, albeitquite low (see Exhibit 11).Deriving the return per unit of risk for the variousasset classes shows that over the past 20 years, hedge fundshave ofiered the most attractive returns; U.K. governm ent

2(X)8 TH EJO URN AI OI AITFRNATIVE INVESTMENTS 7 3


11/19

E X H I B I T 8Risk-Return Trade-Off for Finaticial Asset Classes, 1980-2006

:un

0)

An

R

rae

1

16141210

864

2

.00%

.00%

.00%

.00%

.00%

.00%

.00%

.00%0.00%

UK Govt BondsHedge Funds'

Jan 1980-Feb2006Corporate Bonds,

l ^^^l^ US Epuity-UK Equity

World EquityGovtBonds

'General ArtCommodity Futures

DesmoothedArt

0.00% 5.00% 10.00% 15,00% 20.00%AverageAnnual Standard Deviation

25.00%

Noie: See Exhibit 1 notes.

bonds also showed a good return per unit of risk.NAREIT, art, and equity have also offered attractiveinvestment opportunities. The level to which these assetscan reduce risk in an asset portfolio depends crucially othe extent to which the returns are correlated with eachother. The lower the correlation, the higher the diversi-fication benefits, and the greater the ability ofthe port-folio to maintain returns while reducing risk. This resultsin more moderate returns being generated with a lowerstandard deviation around the expected mean.

For the lowest 10% of returns on the U.K. equitymarket for the last 25-year period, the average return onother financial assets varied between 6% for world equityand 1.4% tor U.S. corporate bonds. U.K. government bondsalso provided good protection with returns ciose to theaverage 9% over the same period {see Panei A of Exhibit 7).Art proxides significantly greater monthlyrcturnsduringthese months than the other asset classes. This is, of course,affected by the smoothing process inherent in the data.

Portfolio DiversificationDetermining optimal portfolio allocations requires

an assumption about the expected return distribution ofasset classes. The best prediction ofthe future is helpedby examining the historical distribution of returns as an

estimate ot future expected returns. This, of coursedepends on the time horizon chosen in the past. Thiarticle provides a number of descriptive statistics as weas correlation coefficients for the time horizons of 25 year15 years, and 5 years. Data on U.K. government bondare only available on a monthly basis since 198(1. Becausgovernment bonds are a crucial element of any weldiversified portfolio, the portfolio is optimized by usindata from the past 25 years.

Importantly, investing in art has large transactiocosts, sometimes as much as 30% ofthe sale price. Thexpense can be minimized by using a long time horizonsuch as 25 years.

In Exhibit 12. the risk-return trade-ofF betweethe various asset classes is shown along with the optimaportfolio when art is included. Also shown is the capitamarket line where the risk-free rate (where the risk assumed to be zero) intercepts with the y-axis and thoptimal portfoho of assets. Tbe investor can obtain anposition along the capital market line by holding a proportion of his wealth in cash, with an expected returequal to the risk-free rate and the optimal portfolio.

The optimal portfolio is derived from the perspective of a U.K. investor who has the possibility of investinin the following indices; World, U.S., and U.K. equityU.S. corporate bonds, Cotnmodity Futures Inciex

74 ART AS A FINANCIAL INVESTMUNT S P R I N C ; 2()(


12/19

oo0 0

Sc.ScC

-- o o o

- d d d d 3 r>- M (p sS r- o i*) S^ 9 d d d 9o i Irt M [M P> r-i q s s s d 9 9 9 9 9

S 2 g i g d o J ^

CQ

o CN * 03 CN f CD Irtq q q r j ' ^ q S q P q^ d d d d d d d d 9 dR q9

d o d d d d d d d d d d d

o o

y^ aouiijuiZ < t Zlu i S a.O

NAREIT, lO-year U.K. government bonds. Art. andhedge funds.The portfolio is first optimized excluding an invest-ment in art by using the risk-^return profile from the past

25 years; the results are presented in Panel A of Exhibit 13.When including General Art in the portfolio, thelow correlation with the other asset classes results in ahigh allocation to art in the portfolio (more than 20%).This is derived using the General Art index rather thanthe fund composition, which would be an even higherpercentage allocation in the optimal portfolio. Therefore,the more conservative return from theGeneral art indexdoes not overemphasize the art allocation.An important feature ofthe data methodology behindthe indices is the moving average, which results in a posi-

tively autocorrelated series. It is important in the analysison risk and return and on portfolio diversification tbat thetrue market risk and return levels be calculated. In the nextsection, the desmoothed data results in a more volatile returnseries that is more in line with the true art market volatility.Henceforth, this examination will take transactioncosts into account, which has the effect of reducing tbereturns generated on the series, and will look at how thesetwo eTects of greater risk and lower return affect theoptimal portfolio allocation. Finally, the analysis willinclude hedge funds in tbe optimal portfolio allocation,Desm oothing ReturnsAt first glance, the art market does not appear tobave been very volatile. However tbe lower volatility inth e art market is likely the result of appraisal-inducedbiases, which occur d uring tbe indexation of tbe art data.Tbe smoothing of the returns is a result is this as well.Tbis has the effect ot generating volatilities that are sub-stantially lower tban the true volatility ofthe market.Because tbe data tor the art indices are generallyappraisal based, the analysis needs to account for this.Although they are a highly valuable source of informationregarding behavior o fth e art market, there is of course adifference hetween the appraisal-based returns and the truemarket returns. It is tbe true market returns tliat actuallyrepresent the economic opportunity cost to investors, andthe statistical prop erties of w hich are directly comparableto alternative asset classes. The illiquid nature ofthe artmarket, with mfrequent valuations, and averaged pricequotes, leads to a smoothing in the returns. It is thereforeimperative that the series he "desmootbed" to eliminate,

Si'RING 2(K1H THEjOURNAt OF A L T L R N A I lVF. INVESTMENTS 7 5


13/19

E X H I B I T 1 0Risk-Return and Correlation Statistics for Fine Art and Financial Markets: 5 Years, 15 Years, 25 Years

MSCIUKUSCORPBOND GSCI

US 10YEARGOVT.BONDUK 10YEARGOVT.BOND REIT HEDGEFUND GENARTMSCI MSCIGENERAL ART WRLD USA

1980-2006 4.7% -3.2% 3-2% -1.3% 9.1% -3.0% -6.3% -7.5% -5.3% 100.0%1990-20062000-2006RETURN/RISK

Note: SeeExhibit 7 notes.

2.0%- 6 .1%0.78

-4.2%-5.0%0.82

-0-7%-2.9%0.8

-11-20

.8%

.5 %

.66

7.1%10.1%0.48

-5.4%3.8%0.70

-6-7%11.5%1.28

-7.0%4.4%0.93

-5-81

.3 %

.9%

.32

100.0%100.0%

0.81

E X H I B I T 1 1Fine Art Market Performance during U.K. Equity Bear Markets, 1980-2006

Asset Class Returns during Equity Down MarketsBased onMonthly Data: Jan 1980-Feb 2006

as far as possible, any underlying autocorrelation, whichtends to be character is t ic of these smoothed ser ies ofappraised returns. The most widely used approaches arethose of Geltner [1993|, trom the real-estate finance liter-ature and now also comm on in the hedge tund literature(Brooks and Kat |2001] , and Kat and Lu [20021). Geltneradjusts the return series to eliminate the first-order auto-correlation. Assuming that the observed (smoothed) returnon the art index, r^ , is a weig hted average ot the true un der-

lying return at t ime i, r^, and the observed (smoothedreturn at t ime - 1, f " , _ p '

r ] = ( l - (2Simply rearranging enables the determinat ion of th

actual return, which, if assumed to be an AR(1) procesacts to eliminate the f i rs t -order autocorrelat ion.

76 A R T AS A FINANCIAL INVESTMENT 200


14/19

E X H I B I T 12Risk-Return Trade-Off: Optimal Portfolio Allocation, 1980-2006

Jan 1980-Feb2006

UK Govn Bonds*

*

Hedge Funds

Ganeral Art

NAREIT US Equity*

World Equity

^UK EquityCorporate Bonds

Commodity Futures

0% 5% 10% 15%Average Annual Standard Deviation

20% 25%

Note: SeeExliihit 7 notes.

r. -r = 1- a (3)If che first-order autocorrelation ot the sinoothfd

series is positive, then the standard deviation ofthe actualreturn series will be greater. However, if the first-order.uitocorrelation ofthe smoothed series is negative, thentlic standard deviation ofthe actual series will be lower.If the autocorrelation structure is more comphcated, thenthe more rigorous process developed by Okunev andWhite [2()(l3| can be adopted to remove higher levels ofautocorrelation in the smoothed series:

+ "2 - (4)

where the constant, a, to desmooth the series, is a flinc-tioii of higher orders of autocorrelation.'' This approachis directly apphcable for art indices, which also exhibitexceptionally high autocorrelations in reported returns.There is indeed evidence of smoothing in the returns,and for series that are positively autocorrelated, thesmoothing has the effect of diminishing the risk apparentin the asset class; hence, it is necessary to correct for the

smoothing, resulting in amore volatile desmoothedreturn series.

Using the more simplified approach of Geltner[1993] does not completely eliminate the first-orderautocorrelation in the time series for art. The moresophisticated approach from Okunev and White [2003],which takes into account higher orders of autocorrela-tion, does result in a desmoothed series that no longersuffers from hrst-order autocorrelation. The high posi-tive autocorrelative structure present inthe art seriesresults in the desmoothed series exhibiting significantlyhigher volatility.

By desnioothing the returns toaccount tor theautocorrelation inthe data, the risk increases substan-tially from 6.5% to 11.5%,^ Taking a universal 5%increase inthe monthly standard deviation for the artseries can show how this increase affects the optimalportfolio allocation. It reduces the allocation in art sub-stantially, by roughly half, from over 20% to just under10%, with the reduction roughly equally spread amongthe other asset classes in the portfolio. The low corre-lation still results in art providing a highly attractive port-folio investment. World equity still remains unattractivegiven the slightly lower return-risk ratio than the otherasset classes and the relatively high correlation with theU.S. equity market (in this case, 90%). This finding is seen

; 2008 THE JouKNAi. ii\-AiTFRNArivi: INVHSTMKNTS 77


15/19

E X H I B I T 1 3Optimal Portfolio Allocation: 1980-2006

Panel A. Excluding Hedge FundsPORTFOLIO

NO ARTWITHARTDESMOOTHEDAR TDESMOOTHED ART& TRANS COSTS

Panel B. IncludingPORTFOLIO

NO ARTWITH ARTDESMOOTHEDARTDESMOOTHED ART& TRANS COSTS

Note: See Exhibit 7 notes.

MSCIUSEQUITY7.95%6.84%7.60%7.79%

HedgeMSCIusEQUITY0.00%0.00%0.00%0.00%

MSCIUKEQUITY3.77%1.55%2.34%3.17%

FundsMSCIUKEQUITY0.00%0.00%0.00%0.00%

usCORPBONDS7.45%5.63%6.54%7.07%

usCORPBONDS2.96%2.47%2.67%2.80%

COMMODFUTURES8.62%5.83%7.06%7.97%

COMMODFUTURES3.04%1.92%2.34%2.74%

USREALESTATE19.65%17.32%19.12%19.42%

usREALESTATE10.20%9.89%10.84%10.74%

UK 10YEARGOVT52.55%41.39%47.41%50.39%

UK 10YEARGOVT42.55%35.60%39.17%41.04%

AMRALL ARTINDEX

21.43%9.93%4.19%

AMRALL ARTINDEX

15.91%7.04%2.82%

HEDGEFUNDS

1

1

1

1

HEDGEFUNDS41.25%34.21%37.94%39.87%

in Panel B of Exhibit 13. Art's high transaction costsspread over 25 years equal 1.5% a year. Despite thesecosts, art still remains an attractive, although small, port-folio allocation.

Including Hedge FundsHedge funds provide an attractive return per unit of

risk, meaning that hedge funds also provide substantialrisk-return benefits in a diversified portfolio. Includinghedge funds in the portfolio allocation analysis results ina much higher allocation to hedge funds and art's alloca-tion is reduced to only 3%. As shov n in Exhibit 9, thecorrelation ot hedge fund returns and mainstream assetclasses is higher than between other alternative asset classes.This result is because hedge funds, rather than being analternative asset class, offer investment strategies forinvesting in mainstream assets, primarily equities, andfixed income.

Optimizing the portfolio with the inclusion of hedgefunds in the four scenarios1) without art, 2) with art,3) with desmoothed art, and 4) with desmoothed art and

transaction costs-^produces the portfolio allocationshown in Panel B of Exhibit 13.For each ot the four scenarios, the allocation intart is increasingly lower and there is a large percentagallocation into hedge flinds. Hedge funds over the perioanalyzed have been the preferred portfolio diversifier.

SUMMARY AND CONCLUSIONSFaced with un der performing portfolios, investor

are continually seeking alternative assets and sophisticatesolutions to reap high returns while minimizing risk. Thiarticle has taken a close look at the financial implicationof including art as an alternative asset class. This previousltiontransparent market is becoming more accessible vithe increasing availability of indices and data on the armarket. Additionally, art funds offer iivestors the opportunity to invest indirectly into the art market.

Indirect investment into the art market results ilosing the aesthetic pleasure from holding the art; however, financial gains can be made through pooling resourcewith the help of experts, while benefiting from diversfication. The art fund market is still in its infancy. Ther

78 AS .A F I N A N C I A L I N \ T . S T M E N T ; 2111


16/19

are few altfrnatives, and these are only available to investorswilling to invest at a substantial level. Entry levels are, atpresent, still high. In time, these funds may become moreaccessible to the mainstream investor through poolingjoint interests.The results in this article show that art 's low corre-lation with other asset classes offers diversification bene-fits from holding art in an investment portfolio. Optimalportfolio allocations using empirical returns over the past25 years provide support for investors to consider art as anattractive, albeit small, addition to their investment strategy.A P P E N D I XINDEX CONSTITUENTS

General all ArtPierre ALECHINSKY. Hek-ii ALLINGHAM, SirLawrence ALMA-TADEMA, Michael ANCHER, KarelAPPEL, Ccorg UASELITZ, Jean Muhcl BASQUIAT, AlbertIJIERSTADT, Pierre ONNARD, Fernando BOTERO,Francois B ( ^ U C : H E R , Eugene BOUDIN, Arttuir MerricBloomficld BOYD. Georges BRAQUE, Bernard BUFFET,Sir Edward Cotey BURNE-JONES. CANALETTO, MarcCHAGALL, Sandro CHA, Giorgio de CHIRICO, PieterCLAESZ, Jean Baptiste Camille COROT, GustaveCOURBET, Salvador DALI, Montague DAWSON, Otto

DIX. Jean HUBUFFET, Max ERNST, Henri FAN TIN-LATOUR, Lyoncl FEININGER, Lucio FONTANA, MylesBirket FOSTER, Jean Honore FRAGONARD, SamFI.ANC;iS, Thomas ClAINSBOROUGH, John Will iam(ODWARD, Jan van GOYEN. Jean-Bapt i s te GiUZE,Atkinson GRIMSHAW, Francesco GUARDI. KeithHARING, Henri HARPIGNIES, Childe HASSAM. Paul-i:esar HELLEUJohn Frederick (sur) HERRING, FerdinandHO l )LER , A ntonio JACO BSEN , Johan-Laurents JENSE N,Johan Barthold J O N C ; K I N D , AsgcrJRN. Jan van KESSEL,Ernst Ludwig KIRCHNER. Moise KISLING, Paul KLEE.Gustav KLIMT. Willem KOEKK OEK, O skar KO KOS CHK A,Willem df KOONING, Nicolas de LARGILLIERE, CarlLARSSON. Marie LAURENCIN, Fernand LEGER, LordFrederic LEKHTON, Sir Peter LELY, Bruno LILJEFORS,Nicolaes MAES, Kern- MAGI.ITTE, Michle MARIESCHI,Ben MA1.SHALL, Henri MATISSE, Sir John Everett MIL-LAIS. Joan MIRO. Claude MONET, Giorgio MORANDI,Sir Alfred MU NN IN GS , Enl NO LD E, A R PEN CK, PabloPICASSO, Serge POLIAKOFF, Pierre Auguste RENOIR,Sir Joshua RE YN OL DS , Jean-Paul RIO PEL LE, DiegoRIVERA, Hubert ROBERT, Dante Gabriel ROSSETTI,Salomon van RUYSDAEL, Gino SEVERINI, Dorothea

SHARP, Leon SPILLIAERT. Carl SPITZWEG, AlfredSTEVENS, Marcus STONE. Abraham STORCK, AntonioTAPIES. David (younger) TENIERS. Fri tz THAULOW.Archibald T HO 1U 3U RN , Giovanni Battista TIE PO LO , JamesJacquesJosepli TIS SOT . M aurice UT RIL LO . Louis VALTAT,Edouard VUILLARD, Andy WARHOL, Tom WESSEL-MANN. Jack Butler YEATS, Anders ZORN.O ld M a s t e r sOsias 1 BEERT. Nicolaes BERCHEM, Louis LeopoldBOILLY, Francois BOUCHER, Jan (elder) BRUECiHEL,Jan (younger) BRUEGHEL. CANALETTO, Annibate CAR-RACCI, John CONSTABLE, Aelbert CUYP, ArthurDEVIS. Carlo DO LC I, Sir Anthony van DY CK. Jean H onorcFRAGONARD. Fran s I FRANCKEN, Th o mas GAINS-

BOR OU GH , Theod o re C ; E R I C : A U L T . Luca GIORDANO,Jan van GOYEN, Jean-Bapti .ste GREUZE, FrancescoGUARDI, Giacomo GUARDI, Giovanni Francesco GUER-C I N O . Jan Davidsz de HEEM. Egbert van HEEMSKERK.Meindert HOBBEMA. Will iam HOGARTH, Mckhior deH ON DE CO ET ER ,Jea n Baptiste HUET, Jacob van HUL S-DONCK. Jan van HUYSUM. Julius Caesar IBBETSON.Antonio JOL I,Jacob JOR DA EN S. Jan van I KESSEL, NicolasLANCRET, Nicolas dc LARGILLIERE, Si r ThomasLAWRENCE. Sir Peter LELY, Carle van LOO, NicolaesMAES, Alessandro MAGNASCO, Michle MARIESC^HLBen MARSHALL, Adam Frans van der MEULEN, JanMiense MOLENAER. Klaos MOLENAER, Joos deMOMPER, Peter MONAMY,Jean Baptiste MONNtWER.Geo rg e MORLAND, Alex an d e r NASMYTH, Ch a r l e s -Joseph NATOIRE, Jean Marc NATTIER, Aert van derNEER. Adriaen van OSTADE, Isaac van OSTADE. JeanBapti.ste OUDRY, Giovanni Paolo PANINI, Jean BaptistePATER. Ciianibattista PIAZZETTA, Giovan BattistaPIRA NESI, Guido R EN I. Sir Joshua REYN OL DS , MarcoRICCI, Sebast iano RICCI, Hubert ROBERT, GeorgeROMNEY. Salvator ROSA. Thomas ROWLANDSON. SirPeter Paul KUBENS. Jacob van RUYSDAEL, Salomon vanRUYSI.:)AEL, Paul SANDBY, Francis (elder) SARTORIUS,John Not t SARTORIUS, Jan STEEN, George STUBBS.Giovanni Battista TIEPOLO, Ciovanni DomenicoTIEPOLO, Jacopo TINTORETTO, Joseph Mallord WilliamTURNER, Lucas van UDEN, Willem van de (elder)VELDE, Simon VERELST, Nicolas van VERENDAEL,Joseph VERNET, Paolo VERONESE, David VINCKE-B O O N S . Simon dc VLIEGER, Sebastian VRANCX. JeanAntoine WATTEAU, Jan WEENIX, Adam WILLAERTS,J o h n W O O T T O N , P h i l i p s W O U W E R M A N , J o s e p hW R I G H T O F D E R B Y . Jan W Y N A N T S . J oh an n Z O F -FAN Y, Francesco ZU CC AR EL LL

2008 TH E JciukNAi oi- Ai.Ti-:RNAllVf-: INVESTMINTS 7 9


17/19

European ImpressionistsLaureano BARRAU, Jean BEIJ^UD, Eugene BO UD IN,Gustave CAILLEBOTTE, Paul CEZANNE, Edgar DEGAS,Jean Louis FORAIN, Paul GAUGUIN, Armand GUIL-LAUMIN, Albert LEBOURG, Stanislas LEPINE, MaxLIEBERMANN, Edouard MANET, Henri -French MA1.TIN,Henri MATISSE, Maxime MAUFRA, Claude MONET. BertheMO RISO T, Rode nc O 'C O N O R , Camle FISSAlUi-O, PierreAuguste RENOIR, Theodore ROUSSEAU. Alfred SISLEY,Max SLEVOGT, Joaquin SOROLLA Y BASTIDA.Modern ArtistsPierre ALECHINSKY, Karel APPEL, Fernandez ARM AN ,Edouard AR ROY O, Fnnik AUERBACH , Francis BACO N, WilliBAUMEISTER, WiUiam BAZIOTES, Max BECKMANN,Joseph BEUYS, Max BILL, Jules BISSIER, Fernando BOTERO,Louise BOURGEOIS, Alberto BURRI, Reg BUTLER,Alexander CALDER, Giuseppe CAPOGROSSI, AnthonyCARO, Baldacdni CESAR , Lynn CHAD WICK, John CH AM -BERLAIN, Eduardo CHILLIDA, CHRISTO, CORNEILLE,Joseph CORNELL, Richard DIEBENKORN, Jim DIN E, PieroDORAZIO. Jean DUBUFFET, Jean FAUTRIER, LucioFONTANA, Sjni FRANCIS, Helen FRANKENTHALER,Alberto GIA COM ETTI, Arshe GORKY, Adolph GO TTLIEB,Hans HRTUNG, Dame Barbara HEPWORTH. PatrickHERON, Eva HESSE, David HOCKNEY, Hans HOFMANN,Fnedricb HUNDERTWASSER, Robert INDIANA. Asger

JRN, Wassily KANDINSKY, Paul KLEE, Yves KLEIN. FranzKLINE, Willem de KOONING, Wilfredo LAM, PeterLANYON, Roy LICHTENSTEIN, Richard LINDNER,Richard LONG, Morris LOUIS, Piero MANZONI, GiacomoMANZU, Marino MARINI, Agnes MARTIN, GeorgesMATHIEU, MATTA,Joan MITCHELL, Henry O M MOORE,Robert MOTHERWELL, Ernst Wilhelm NAY, LouiseNEVELSON, Ben NICHOLSON, Isaniu NOGUCHI. JulesOLITSKI. Victor PASMORE, Serge POLIAKOFF. JacksonPOLLOCK, Arnaldo POMODORO, Arnulf RAINER,ARNULF RAINER, Martial RAYSSE, Ad REINHARDT,Germaine R ICH IER , Bridget RILEY, Jean-Paul RIO PELLE,Diego RIVERA , James ROSENQUIST, Mark ROTHKO, DavidSIQUEIROS, Pierre SOULAGES, Daniel SPOERRI, Nicolasde STAEL, Rufino TAMAYO, Antonio TAPIES, WayneTHIEBAUD, Mark TOBEY, Gnther UECKER, EmilioVEDOVA, Bram van VELDE, Maria Elena VIEIRA DA SILVA,Andy WARHOL, Tom WESSELMANN.

Contemporary ArtistsCarl ANDRE, Richard ARTSCHWAGER, MiguelBARCELO, Matthew BARNEY, Georg BASELITZ, Jean

Michel BASQIAT, Vanesa BEECROFT, RosBLECKNER, Chris t ian BOLTANSKI, Maurizo CATTELAN, Sandro CHA, Francesco CLEMENTE, TonCRAGG, En zo CUCCHI, Ol iv i e r DEBRE, WimDELVOYE, THOMAS DEMAND, RINEKE DIJKSTRAPeter DOIG, STAN DOUGLAS, Marlene DUMAS, TraceEMIN, Luis FHITO, Rainer FETTING, Eric FISCHL, P.&WEISS FISCHLI, Dan FLAVIN, Gnther FORC;, LuciaFREUD, GILBERT and GEORGE, Robert GOBER, NANGOLLIIN, Fe l i x GONZALEZ-TORRES, Do u g laG O R D O N , DAN GRAHAM, Andreas GURSKY, KeitHARING, Dainien HIRST, Jenny HOLZER. Gary HUMEJrg IMMENDORF, Jasper JOHNS, Donald JUDD, AleKA TZ, M ike KELLEY, Ellsworth KELLY, Anselm KIE FERMart in KIPPENBERGER, JefT KOONS, JanniKOUNELLIS, Sol LEWITT, Robert LONGO, SaraLUCAS, Robert MANGOLD, Brice MARUEN, MariME RZ, Juan MU NO Z, Bruce NAUMA N, Shirin NESHATChris OFILI, Claes OLDENBURG, Gabrie l OROZCONam June PAIK, Mimmo PALADINO, P ANA MERE NKOA R PENCK, Michelangelo PISTOLETTO, Sigmar POLKERichard PRINCE, Robert 1J\USCHENBERG, CharlesAmerican RAY. Gcrbard RICHTER. Pipotti RIST, MimmROTELLA, Susan ROT HE NB ER G, Thomas RUFF, EdwarRU SCH A, N iki de SAINT-PHAL LE, David SALLE, AntoniSAU RA, Jenny SAVILLE. Julian SC HN AB EL, ThomaSCHUTTE, Sean SCULLY, George SEGAL, Richard -American SERRA, Andres SERI^J^NO, Joel SHAPIRO, CindSHERMAN, Jose Maria SICILIA, Frank STELLA, ThomaSTRUTH, Donald SULTAN, Rosemarie TROCKEL, LuT U Y M A N S , Cy TWOMBLY. Jeff WALL, Franz WESTChristopher WOOL,ENDNOTES

'Frey and Eichenberger [1995].^Data are only available until December 2(102 on tbe AArt Index.A description of tbe artists included in the various indicis in the appendix.^Similar survivorship bias is also apparent in other financial indice s."The M SCI indices are extremely highly correlated w ittbe national stock market indices, for example, tbe S&P 500 anthe FTSE lOU. Data on the MSCI indices arc available on monthly basis for the whole sample.''See Okunev and White [2003] for greater detail on thconditions that the autocorrelation function must fulfill and othe apphcation to remove higher orders of autocorrelation."See Campbell |2OO5] for a detailed analysis odesmoothing art series data.

80 AR T AS A FINANCIAL INVESTMENT ; 200


18/19

REFERENCESAnderson, R .C . "Painrings as an Investment ." Economic Inquiry,12(1974) . pp , 13-26 .Ashenfelter, O. "How Auctions Work for Wine ;uid An"Jouriia!of Ucoiioriiic Perspectives, 3 (1989) , pp. 23-36.Ashcnfckcr , O. , and K. Graddy. "Auct ions and the Price ofArt.'\lViriiil of Economic Literature, 41 (2003) , pp. 763-788.Baum l , WJ. "U nnatu ra l Value: O r Art Investment as Float ingCvAp Gnnie ." American Economic Review, 76 (1986) , pp. U)-14.l ind ens , N , , and V. Ginsb urgh. "R evisi t ing Baum ol ' s 'Art asFloat ing Crap Game' ." European Economic Revietv, 37 (1993) ,pp . 1 3 5 1 - 1 3 7 1 .Campbel l , R. "Art as an Al ternat ive Asset Class," Workingpaper, Maastricht University, 2005.Chan el , O. , L.A. Gerard-Varet , and V. Ginsburgh . "Th e Re l-evance of Hedonic Price Indices."_/[)nniii/ of Cultural Economics,20(1 996 ) , pp . 1 -24 .Frey. B.R .. and R. Eichenbcrgcr. "O n the R etu rn of Art Invest-m e n t R.eturn Ana.\yse%y Journal of Cultural Economics, 19 (1995),pp . 2 0 7 - 2 2 0 .CifitiKT. IX "Smoothing in Appraisal Based Returns."_/i>""iij/of Real Estate finance and Economics, 4 (1993) . pp . 327-3 45 .Ginsburgh , V . J . M e i , and M. Moses , "O n the C ompu ta t ionof Price Indices." In V. Ginsb urgh and D, Throsby, eds,. Hand-book ofEcotiotnics Art and Culture, 1, Am sterdam: Elsevier N ort li-Hol land, 2006.

Granipp, W, Pricinii the PricelessArt, Artists and E conomics. N e wYork: Basic Books, 1989.Kochugovindan, S. Barclays Capital Equity Gilt Study. L o n d o n :Barclays Bank, 2005,M ei . J . . and M . Moses. "Art as an Investment and the U nd er-p e r i o r m a n c e o M a s t e r p i e c e s , " American Economic Review,9 2 ( 2 0 0 2 ) , p p . 1 6 5 6 - 1 6 6 8 .

. '"Vested Interest and Biased Price Estimates: Evidence froman Auction MarkeC fournal of Finance, 60 (2005), pp. 2409-2435,Okunev,J. , and ]) . White. "Hedge Fund Risk Factors and Valueat Risk of Credi t Trading Stra tegies." Work ing paper . U niver-sity of New South Wales, 2 0 0 3 .Pesando , J .E . "Ar t as an Inves tment : Th e M od ern Marke tfo r Modern Pr in t s . " American Economic Revieir, 83 (1993) .pp . 1 0 7 5 - 1 0 8 9 .Ste in, J .P "T he M oneta ry Apprecia t ion of Paintings." Journalof Political Economy, 85 (1977) , pp. 1021-1035.Zanola , R. "The Dynamics of Art Pr ices: Select ion CorrectedRepeat-Sales Index." Working paper . Universi ty of EasternPiedmont , 2007 .

To order reprints of this article, please contact Dewey Palmieri [email protected] or 2i2-224'675

Sl'lUN(; 20(18 Ol Al.Ti-RNATIVF INVESTMENTS 8 1


19/19

Documents

Campbell Art Investment