Tài liệu The Market for Sculptures: an Adjacent Year Regression Index pptx

The Market for Sculptures:an Adjacent Year Regression Index Department of Public Policy and Public Choice - Polis Corso Borsalino, 50 – 15100 Alessandria - Italy E-mail: zanola@sp.unipmn

The Market for Sculptures:

an Adjacent Year Regression Index

Department of Public Policy and Public Choice - Polis

Corso Borsalino, 50 – 15100 Alessandria - Italy

E-mail: zanola@sp.unipmn.it

ABSTRACT: This paper analyses the performance of an investment in sculptures during the period 1987-1995 by applying

the hedonic price technique with time dummy variables to a sample of over 27,000 sales occured at auctions The main finding is that the trend of the rate of return on an investment in sculptures is more stable than those associated to alternative forms of investment during the analysed period.

JEL Classification Numbers: C5, Z1.

Key Words: Auction; Hedonic Price; Sculpture; Return.

T he marketforsculptures has characteristics thatmake it di¤ erent fromothermarkets forcollectibles A lthough similarto the marketforpaintings,the marketforsculptures is smallerthan it, with about3-5 per-centoftotalcollectiblessoldatauctions A numberofreasons explainthis, suchas technicaldi¢culties and thehigh costs ofproduction as wellas thefactthatsculpturesneedappropriatespacestobeplacedorbecauseofthelowernumberofsculpturecollectors than paintingcollectors [L ocatelli-B ieyand Z anola, 19 9 9 a].

T hepurposeofthis paperis toexaminetheinvestmentvalueofsculptures

A lthough there areanumberofstudies thathave investigated the investmentvalueofcollectibles, suchaspaintings[A nderson, 19 7 4;Stein, 19 7 7 ;B uelensand

G insburgh, 19 9 3; de la B arre etal., 19 9 4;Chanel, 19 9 5; Chaneletal., 19 9 6;

M ossetto and L azzaro, 19 9 6; A gnello and P ierce, 19 9 6; Candela and Scorcu,

19 9 7 ; R enneboogand V an H outte, 19 9 8 ; A gnelloand P ierce, 19 9 8 ;L

ocatelli-B iey and Z anola, 19 9 9 b], prints [P esando, 19 9 3], violins [R oss and Z ondervan,

19 8 9 ], wine [N erlove, 19 9 5; Combis et al., 19 9 7 ], antique furniture [G raeser,

1 T herawmaterialdepends ontheconstructiontechnique H andmodellinguses terracotta, orrawmaterialthatallows toproduceglazed ceramics Castingis an ancienttechniquewhich uses bronze Carving is performed by usinga numberofdi¤ erentmaterials, such as marble, stone, wood, ivory, bone, and, more recently, plasterand resin.

19 9 3],photos [P ompe, 19 9 6],toourknowledgethis itis the…rsttimethattheperformanceofsculptures as …nancialassets is studied.

T his studyapplies thehedonicpricetechnique2 withtimedummyvariableswhich allows priceindexes forsculptures and shadowvalues forcharacteristics

to be simultaneously estimated D ata are drawn from auctions hold duringthe period 19 8 7 -19 9 5, as compiled by M ayerInternationalA uction R ecord on

CD -R om

T heremainderofthispaperisorganisedasfollows Section2 brie‡youtlinesthe methodology of the hedonic technique Section 3 de…nes the empiricalspeci…cation of the model Section 4 describes the data set used, which iscompiled from M ayerInternationalA uction R ecords on CD -R om Section 5summarises thebasic…ndings Conclusions aredrawn in Section 6

2 See G ordon (19 9 0);G riliches (19 9 0);and T riplett(19 9 0).

2 M ethodology

T heapproachusedtoestimateapriceindexforsculptures is thehedonicpricetechnique as developed in similar studies analysing the investment value ofcollectibles T hehedonicapproach involves theestimateofthe implicitpricestoeachcharacteristicincludedintheequationbyallowingtocontrolforpossiblenon-temporaldeterminants ofpricevariations

A setofj-characteristics, xjkt,withj = 1 ;:::n;areidenti…edforaregressionofthepriceofthesculpturek, withk= 1 ;:::;m, soldinyeart, witht= 1 ;:::;T ;

on its j-characteristics, such that:

lnpkt= ¯0 +

TX¡1 t=1

T wo observations are necessary First, an implicit assumption in (1) isthatthesocialvaluationofsculpturecharacteristics does notchangeovertimegiven the shortness ofthe analysed period Secondly, the modelis speci…ed

in semi-logtransformations in ordertohandle the severe skewness in auctionprices T ransformingthe dependentvariable, enables us touse ordinary leastsquares estimators withouthavingtoworryaboutthesensitivityoftheresultstoskewness

B asedonhedonicregressionequation(1), anumberofpriceindexes canbecomputed3 In this studyweadopttheadjacentyearregressions index, whichimplicitlyweights each observation equally[B erndtetal., 19 9 5]

3 In this respect, we are only considering sculptures as …nancialassets, withoutany

refer-B y subtractingfrom the logarithm ofthe auction price the implicitpricesgiven tospeci…c characteristics, the characteristic-free price ofsculpture kinyeartis equalto:

where» is equaltoe(b¯t+ 1¡b ¯t)¡1 :

ence to theirpsychic returns [B aumol, 19 8 6; Frey and Eichnberger, 19 9 5; Santagata, 19 9 8 ; Czujack, 19 9 7 ].

3 FunctionalForm

T hehedonicregressionframeworktakes intoaccountthee¤ ectofheterogeneity

on prices by controlling fora numberofdi¤ erences in characteristics amongsculptures T hedependentvariableis thelogarithm ofsaleprice, de‡ated andreduced by 15% to correctfortransaction fees charged by auction houses tosellers T heindependentvariables areclassi…ed as follows:

²P roduction A vailabledatadonotallowustodistinguishbetweendi¤ erentschools H owever, as suggested bysomeauctionexperts contacted byus,threedi¤ erentperiods ofproductionmaybeidenti…ed: old, forsculpturesproduced before 18 00; mod, forsculptures produced between 1801 and

19 30;and cont, forsculptures produced after19 31

²N ationality U sing the frequency with which the sculptors’ nationalitiesappearin the data set, the following dummy variables are introduced:France, natf;U sa, natus;G ermany, natd;Italy, natit;andG reatB ritain,natgb:

²Size T hereis nostandard measureofsizeduethevarietyofconstructiontechniques4 In what follows, foreach observation we use the highestdimension value(width, length, thickness), size, as aproxyforvolume

²M edia A setofdummy variables is used: bronze, bro, and marble, mar,are the traditional raw materials used, expecially for old and modernsculptures, whileresin and plaster, res, are used in contemporary sculp-tures Finally, giventhehighnumberofsculpturesofthis kind, alsoivory,ivo, and terra-cotta, ter, areconsidered

4 Forinstance, there are sculptures in the round, which can be viewed from any direction,

as wellas incised relief, in which the lines are cutintoa ‡atsurface.

²Salerooms and Cities of Sale Sotheby’s, soth, and Christie’s, chr; areknowntobetheleadingauctionhouses inthis kindoftransactions T heiroverallperformances di¤ eronly slightly T hemostimportantsculpturesauction’ markets are N ew Y ork, ny, L ondon, lon, and P aris, par Afurthercity is introduced, R ome and M ilan, rom, given the importanceoftheItalian marketforarts.

²M asterpieces Inordertotakeintoaccountthemostvaluedsculptors, lowingP esando(19 9 3), dummyvariables proxythemasterpieceportfolioofthe top 20 percentofsculptures by prices, top20 , as wellas the inex-pensiveportfolio, low20 ;theremainderis assignedtothemiddlemarket,mid20

fol-²P eriod: a setofdummies, d, is introduced foreach yearbetween 19 8 7and19 9 5

Formally, ourspeci…cation is given by:

lnp = ¯0 + ¯1old+ ¯2mod+ ¯3natf + ¯4natus+ ¯5natd+ ¯6natit+

¯7natgb+ ¯8size+ ¯9bro+ ¯1 0mar+ ¯1 1res+ ¯1 2pla+ ¯1 3ter+

¯1 4ivo+ ¯1 5soth+ ¯1 6chr+ ¯1 7ny + ¯1 8lon+ ¯1 9par+ ¯20rom +

4 D ata

T he data used in this paperare drawn from auctions hold duringthe periodoftime 19 8 7 -19 9 5 T he source ofthis data is the 19 9 5 edition ofthe M ayerInternationalA uction R ecords (M IA R ) onCD -R om, which contains records of

27 119 sculptures soldattheworld’s majorauctions5 Foreachsculptureinthedataset, anumberofinformations areprovided P rices aregross ofthebuyers’and sellers’ transaction fees paid to auction houses and are recorded in fourdi¤ erentcurrencies N oinformations is providedontheoriginofthesculpturesandexhibitions ofthesculptures Forthesakeofsimplicity, weassumethatallsales occurattheend ofeach period A llsculptures arepriced in U S dollars,de‡ated by using the U S consumerprice index (19 9 0 = 100)6toremove thegeneraltrend ofin‡ation P rices are 15% less to correctfortransaction feeschargedbyauctionhouses tosellers, butas inmoststudies, wedonottakeintoaccountstorageand insurancecosts

Summarystatistics ofthesamplearedisplayed in T able1

[T A B L E 1]

5 Even ifauction records alone donotre‡ecttheentire market- since auction houses have little incentive to sellout of fashion sculptures and, besides, ‘bought in’ works may in‡ate prices [G oetzmann, 19 9 3]- itis notclearwhetherprices are biased upwards ordownwards [A gnelloand P ierce, 19 9 8 ].

6 S ource: T he FedaralR eserve B ank.

5 R esults

FollowingB erndtetal (19 9 5), sinceheteroscedasticity may be present, darderrors andvariance-covariancematrices ofthecoe¢cients havebeencom-puted byusingtheW hite(19 80) heteroscedasticity-robustprocedure T able2displays themainresults obtained inthehedonicregression

stan-[T A B L E 2]

In almostallcases the coe¢cients aresigni…cantly di¤ erentfrom zeroat5percentoreven at1 percentprobability level T he coe¢cients associated toeach kind ofindependentvariablecan beused toranksculptures accordingtothepriceofa‘normalised’sculpture, thatis, theprice ofasculpturewhen alltheothervariables areassumed tobeatastandard level

Sculptures producedbefore18 00 arethemostpro…tableduetothescarcityofoldestsculptures, whilethoseproducedafter19 30 performsworsethansculp-tures producedbetween1801 and19 30 A s tonationality, theG erman, theEn-glish and the Italian sculptors secure the highestprices P rices are increasing

in size, as in thecaseofpaintings [B uelens and G insburg, 19 9 3;Chaneletal.,

19 9 6;A gnelloandP ierce, 19 9 6;CandelaandScorcu, 19 9 7 ;R enneboogandV an

H outte, 19 9 8 ;A gnelloand P ierce, 19 9 8 ]

T hemostexpensivemediais marbleduetothehigh costofproduction, aswellas technicaldi¢culties in producing sculptures by using this media B ycontrast, othermediadisplaynegativecoe¢cients, even ifbronzeand resin arenotstatisticallysigni…cant

P rices recorded atSotheby’s are higherthan those atChristie’s, and N ew

Y ork is the city where prices ofsculptures are the highest B y generalising

P esando’s analysis ofthe marketforprints, a possible reason forthis is due

to N ewY ork’s capacity toattracthigh quality sculptures because the bidder

audienceis trulyinternational Finally, masterpieces perform dlemarketartists;bycontrast, inexpensiveportfolioworks perform worsethanmiddlemarketportfolio.

betterthanmid-T he set of dummy variables introduced for each year between 19 8 8 and

19 9 5 is used in equation (4) to build the adjacentyearregressions index forsculptures, which is reproduced in T able3

[T A B L E 3]

A marketshockcharacterises theperiod19 8 7 -19 9 5 coveredbythedata-set,since after a boom period until 19 9 0, the market for collectibles registers anon-boom period from 19 9 1 to19 9 4 T he rate ofreturn ofan investmentonsculptures seems tofollowthe trend ofthe rate ofreturn on paintings duringthesameperiodoftime[CandelaandScorcu, 19 9 7 ;L ocatelli-B ieyandZ anola,

19 9 9 b] R atesofreturnarepositiveuntil19 9 0, followedbyanegativetrendfrom

19 9 1 to19 9 4, with apositivesign in 19 9 5, butstilllowerthan thepriceindex

19 8 7 H owever, the trend ofthe performance ofan investmentin sculptures

is lower than those associated to alternative forms of investment Figure 1compares the de‡ated index for sculptures with the de‡ated indexes of realreturns on U S stocks, U S 30 yeargovernmentbonds and gold

[ FIG U R E 1]

D uringtheboom periodinvestments insculptures earnedlowerrealreturnsthan U S stocks, U S 30 yeargovernmentbonds, and gold, forms ofinvestmentcharacterised byacomparabledegreeofrisk H owever, startingfrom 19 9 1, re-turnsonsculpturesdecrease, butatalowerratethanotherkindsofinvestment

T his resultmay be due tothe characteristics ofthe marketforsculptures, asillustrated in Section 1 In fact, the demand forthis kind ofcollectibles is ex-pectedtoberatherinelastic, sothatthereturns onsculptures arenotsomucha¤ ected byboom and non-boom periods

6 Conclusions

T his paperanalyses therateofreturnonsculptures, akindofcollectibleaboutwhich very little is known A lthough the market for sculptures only repre-sents about3-5 percentoftotalcollectibles sold atauctions, good investmentopportunities may exist In ordertoaddress this question, the adjacentyearregressions index forsculptures, sold atauctions duringtheperiod 19 8 7 -19 9 5,has been computed byusingan hedonicpricetechnique

O urresults canbesummarised as follows T herateofreturnonsculpturesseemstofollowthetrendofothercollectiblesduringtheanalysedperiod, suchaspaintings [Candelaand Scorcu, 19 9 7 ;L ocatelli-B ieyand Z anola, 19 9 9 b] H ow-ever, thetrendoftheperformanceofaninvestmentinsculpturesislowerduringtheanalysed period, duetotheinelasticity ofthedemand which characterisesthis market

A numberofissues remain forfurtherresearch Following similarstudieswhich applythehedonictechniquetoestimatereturns on paintings, itmaybeinterestingtoanalyse the production ofa restricted numberoffamous sculp-tures T his allows us bothtoisolatefactors thatin‡uenceprices and tocollectadditionaldata towhatis listed in published data sets [Czujack, 19 9 7 ] Fur-thermore, inthis studyweonlyfocuseonthereturns onsculptures as …nancialassets H owever, anumberofauthors [B aumol, 19 86;Frey and Eichenberger,

19 9 5;Santagata, 19 9 8;Czujack, 19 9 7 ]highlightthatvisualarts acterised by aesthetic returns, an analysis of which may represent a futuredevelopmentofthis paper

arealsochar-A cknowledgments

T heauthorsaregratefulfor…nancialsupportfrom CN R program oncultural

goods T his paperhas bene…ted from comments by G B rosio, M Ferreroand

W Santagata T heusualdisclaimers apply

7 R eferences

A gnelloR J and P.R K.P ierce(19 9 6), FinancialR eturns, P riceD eterminants,andG enreE¤ ects inA mericanA rtInvestment, JournalofCulturalEconomics,20(4), 359 -383

A gnelloR J andP.R K.P ierce(19 9 8), D isaggregatedreturns andriskin artpurchases T he case ofA merican P aintings sold atauction from 19 7 1-19 9 6,

U niversityofD elaware, mimeo

A nderson R C (19 7 4), P aintings as an Investment, Economic Inquiry, 12,13-26

B aumolW J (19 86), U nnaturalV alue, orA rtInvestmantas FloatingCrap

G ame, A merican EconomicR eview, 7 6,10-14

B erndtE.R , Z G riliches andN J.R appaport(19 9 5), Econometricestimatesofpriceindexes forpersonalcomputers inthe19 9 0’s, JournalofEconometrics,

e-Combris P., S.L ecocq and M U isser (19 9 7 ), Estimation of H edonic P riceEquation forB ordeaux: does Q uality M A tter?, Economic Journal, 107 , 39 0-402.

Czujack C (19 9 7 ), P icasso P aintings atA uction, (19 63-19 9 4), JournalofCulturalEconomics, 21, 3, 229 -247

delaB arreM , S.D occioandV G insburg(19 9 4), R eturns onImpressionist,

M odernandContemporaryEuropeanP aintings, 19 62-19 9 1, A nnalesd’EconomieetStatistique, 35, 143-181

FreyB S.andW W P ommerehene(19 8 9 ), M uses andM arkets;Explorations

in the Economics ofthe A rts, O xford, B asilB lackwell

FreyB S.andR Eichenberger(19 9 5), O ntherateofreturnintheartmarket:Surveyand evaluation, European EconomicR eview, 39 , 528 -537

G insburgh V and A F P enders, (19 9 7 ), L and A rtists and A rt M arkets,JournalofCulturalEconomics, 21, 3, 219 -228

G oetzmann W N (19 9 3), A ccounting ForT aste: A rt and the Financial

M arkets O verT hreeCenturies, A merican EconomicR eview, 83(5), 137 0-137 6

G ordonR (19 9 0), T hemeasurementofdurablegoodsprices, Chicago, Chicago

P ress

L ocatelli-B ieyM and R Z anola(19 9 9 a), Ilmercatodellesculturein Italia: