A Comparative Study of Wine Auction Prices: Mouton Rothschild Premier Cru Classé - PDF

May 2003 A Comparative Study of Wine Auction Prices: Mouton Rothschild Premier Cru Classé Jan Bentzen Valdemar Smith Department of Economics Aarhus School of Business Prismet, Silkeborgvej

Please download to get full document.

View again

of 19
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 25 | Pages: 19

Extension: PDF | Download: 0

May 2003 A Comparative Study of Wine Auction Prices: Mouton Rothschild Premier Cru Classé Jan Bentzen Valdemar Smith Department of Economics Aarhus School of Business Prismet, Silkeborgvej Aarhus C Denmark Abstract: Auctions of selected wines have regularly taken place internationally and from natural reasons they have mostly involved the finest wines as e.g. the top wines from Bordeaux. In order to analyse specific auction wine prices, the Mouton Rothschild (Medoc Premier Cru Classé) has been selected for investigation, where auction data have been collected from the USA (The Chicago Wine Company), Denmark (Bruun Rasmussen, Selected Wines Auctions) and from other sources as well. The price development of this specific icon wine is expected to be influenced by a lot of factors, although theoretically investment decisions concerning e.g. icon wines ought not to be highly sensitive to short-run business conditions. The empirical findings exhibit that the auction prices of the Mouton Rothschild differ relatively much between the auction houses, and the time series analysis reveals only weak evidence of co-movements between wine prices and selected business cycle indicators. JEL Classifications : E32, D44. Keywords: Wine auction prices, Mouton Rothschild, business cycles. Acknowledgements. The authors want to thank Bruun Rasmussen Auctioneers of Fine Art, Copenhagen, and the London branches of Sotheby's and Christie s for providing useful information concerning the wine auctions. Paper prepared for the OENOMETRICS IX Conference, Budapest, May 1. Introduction During the 1990s, the prices of Cru Classé wines from Bordeaux increased dramatically. Sustained economic growth in the OECD area and significant growth rates in East-Asia combined with a widening of the wine market were the most important factors behind the massive price jumps that took place. The growth in prices of Cru Classé wines took place irrespectable of the huge advances in quality and subsequent increases in supply and variety of premium wines coming from many overseas wine producing countries, i.e. California, Australia, Chile, South-Africa etc. in this period. Furthermore, wine producers in countries within EU, e.g. Italy and Spain, made a large effort in order to strengthen their position on the market by offering wines in nearly all price segments at reasonable prices. Moreover, a standard result from many wine magazines reporting from blind tastings including Cru Classé wines from Bordeaux and top wines from other wine producing areas is often that the Cru Classé wines were valued lower than their market price, whereas the opposite often is seen with wines from California, Australia and Italy. Still, the prices of Cru Classé wines did seem to live their own lives at least until the Millennium, with primeur prices of each new vintage never starting below the prices of the previous year. For investors of earlier vintages of the top wines, the 1990s certainly was a profitable period, where for example a Chateau Cheval Blanch 1990 that could be bought en primeur for less than $50 now sells for a little less than $500 at wine auctions. Indeed many buyers of Cru Classé wines most likely looked and still look at these wines as investment objects, believing that the prices of good vintages will keep rising over time. Consequently, believing that Cru Classé wines can be considered as investment objects like similar assets, e.g. paintings, jewels, real estate etc., economic factors will influence on prices, but with a long time horizon concerning investments in Cru Classé wines - some of these can be stored for decades and still be drinkable - especially short-run business conditions are not expected to be of major influence. The aim of this paper is to analyse the price development of a specific Cru Classé wine and the Mouton Rothschild 1982 vintage has been selected. This wine is regularly offered at international wine auctions, and hence a continuous time series data set more easily can be obtained compared to some of the other icon wines. Also, the wine has been top-rated by R. Parker with 100 points and 1 the wine should still be drinkable up till the year 2075 according to RP's Wine Advisor (2002). Auction wine prices are highly volatile where the same wine even may be sold at differing prices at the same auction as noted by Ashenfelter (1989), and therefore only one specific wine is selected for analysis in order to avoid too much uncertainty and volatility when constructing a consistent time series price data set. Data have been collected from the USA (The Chicago Wine Company, Wine Auctions) and Denmark (Bruun Rasmussen, Selected Wines Auctions) as well as other sources, and price series for the Mouton Rothschild 1982 vintage have been calculated representing the time period since In the following section, the international wine auctions and data sources are shortly described. Section three presents international price indices of the Mouton Rothschild and section four analyses the relationship between the auction wine prices and selected business cycle indicators, which is further investigated by causality tests in section five. Finally, section six concludes. 2. Wine auction data for French icon wines - Mouton Rothschild Auctions of selected wines have regularly taken place internationally and from natural reasons mostly involved the finest wines as e.g. the top wines from Bordeaux. Therefore, by now, a considerable amount of data exists for auction wine prices, i.e. both the assessed wine prices of the auctioneer and the eventual hammer prices at which the wines are sold. In order to analyse specific auction wine prices, the Mouton Rothschild (Medoc Premier Cru Classé) vintage 1982 has been selected for investigation, where auction data have been obtained from The Chicago Wine Company representing US prices, and for Denmark data from The Bruun Rasmussen Selected Wine Auctions (Copenhagen) are used. Bruun Rasmussen is the largest and most important auction house in Denmark, founded in 1948, and today it is the 8th largest auction house in the world. In the case of Chicago, at least one wine auction is taking place each month allowing a monthly data set to be constructed covering the time period 1996M3 to 2003M2. The Danish auctions are held less frequently and in this case a quarterly data price set for the Mouton Rothschild is obtained covering the time span 1995Q1 to 2003Q1. The Danish data go further back in time - actually, the Mouton 2 prices are available back to but these data are so sparse that no reliable price series can be constructed, and therefore they are not included in the analysis. For both the Chicago and the Copenhagen price data a few missing observations prevail and in these cases a simple linear interpolation has been applied to the data. Additionally, data have been made available from the London branches of Sotheby's and Christie s, and also the wine price database published by the Wine Market Journal has been accessed. As the latter sources only represent a limited number of time series observations, this information is appearing in connection with the Mouton Rothschild price series presented in part three. The price series from the Chicago and Copenhagen wine auctions are measured in terms of the final buyer s price with a 25% buyer s premium added to the Copenhagen hammer prices. The Chicago hammer prices are identical to the buyer s price because only the vendors that consign their wines to the Chicago auction are charged a fee, namely 28% for most of the sample period appearing here. For the London auction prices, a 10% fee has been added concerning both Sotheby's and Christie s in order to convert these prices to buyer s prices. Finally, all final buyer's prices have been converted to USD by applying the exchange rates related to the time of the auctions and thereby facilitating comparisons of the respective wine prices exhibited in part three. When testing for comovements between wine prices and business cycles - only including the US and Danish data, as mentioned before - the empirical tests are performed in national currencies as these from obvious reasons will be most related to the domestic buyers. For the Danish wine auction both data for the 'price estimate' made by the auctioneer - and appearing in the pre-auction information material - and the prices at which the wines are actually sold (hammer prices) are available, and therefore a further analysis of the interaction between these prices is included. For wines offered at any kind of auction, it is essential that they have been stored correctly and that their condition concerning label, level of wine in the bottle etc. are acceptable. For both the Chicago and Copenhagen auctions, some information is available about the latter condition - but no information about storage - and only data representing bottles of normal or close to normal conditions with respect to level of wine and labels have been included in the analysis. Also, only lots with Mouton Rothschild vintage 1982 exclusively have been selected for the data 3 sample as many auctions regularly offer bundles of both differing vintages and brands of icon wines at sale. 3. Auction price indices of the Mouton Rothschild 1982 vintage Information concerning the lot size, i.e. the number of bottles included in each auction number, is usually made available by the auction house. For the Chicago wine auctions the monthly prices for lot sizes ranging between one and twelve bottles are presented in Figure 1. Figure 1. Prices of the Chateaux Mouton Rothschild vintage 1982 from the Chicago Wine Auctions, 1996M3-2003M02 (monthly data, USD per bottle) HP_smoothed (average) price path Note: The dots represent separate lots of sizes 1, 2, 3, 4, 6, 8 and 12 bottles, respectively, sold at the auctions and the prices in nominal USD/bt. correspond to buyer s prices as no fee is charged. The solid line is the Hodrick-Prescott filter applied to the average prices. Source: Data from the Chicago Wine Trading Company (www.tcwc.com) Figure 1 reveals that the prices of identical wines may vary strongly even at the same auction which, of course, appears against any rational behaviour. There may be some unobserved characteristics related to the lots offered at the auction, e.g. whether the bottles are sold in original wood cases or not, but the high volatility of wine prices seems hard to explain and is certainly in contradiction with the law of one price. From the average prices of each auction, the Hodrick- Prescott filter has been applied, and the result is reported as the solid line in Figure 1. Usually, the 4 HP filter has been used in connection with analysing business cycle topics where the time series is decomposed into a trend component and a stationary component - the trend component being the solid line in Figure 1 - and a major advantage of the HP filter in connection with especially time series data for fine wines is that no particular functional form for the trend is assumed a priori. As mentioned before, the prices of icon wines have increased heavily from the late 1980s to the mid-1990s as also partly exhibited in Figure 1. The Mouton Rothschild obtained its maximum (average) price in the later part of 1997 with an absolute price well above USD 500 per bottle when evaluated solely from the smoothed HP trend. Later on, a weakening of prices is observed and presently the Mouton Rothschild prices may be facing a downward slide. The data from Figure 1 can be decomposed according to the lot sizes, and this is reported in Table 1. Table 1. Prices of the Chateaux Mouton Rothschild vintage 1982 from the Chicago Wine Auctions according to lot size, 1996M3-2002M02 (monthly data, USD per bottle). Average price (USD/bt) Standard deviation Lot size: 1 bottle bottles bottles bottles bottles bottles bottles Total Source: Data from the Chicago Wine Trading Company (www.tcwc.com). Table 1 indicates a tendency towards increasing price with respect to the lot size. This can be explained by the fact that for single bottles information about storing conditions is limited whereas larger lots must be assumed to have been stored more professionally in the preceding period as well as larger lots may be offered in original wood cases. The latter arguments also seem to be partly confirmed by the calculated standard deviations where the larger lots (eight and twelve bottles) are associated with smaller variance, but still the relatively high standard deviations reveal the highly volatile auction prices of even icon wines. 5 Figure 2. Buyer s price for Chateau Mouton Rothschild vintage 1982 at various auction houses, 1995Q1-2003Q1 (quarterly data, USD per bottle) Chicago London Copenhagen International Notes: Prices measured as average hammer price plus buyer's premium (auction fee) and converted to USD per bottle. For London, semi-annual data supplied by Christie s is used, and the 'International' price line is the average of all international auctions (primarily from the USA and Europe) as collected and published by the Wine Market Journal. Sources: The Chicago Wine Company, Christie's London, The Wine Market Journal and data (auction catalogues) from Bruun Rasmussen, Copenhagen. The Chicago auction prices presented in Figure 2 are calculated as auction-specific average prices from the data appearing in Figure 1, and actually the absolutely highest price achieved for the Mouton Rothschild is close to USD 600 in For the first part of the sample period the price data exhibited in Figure 2 are relatively identical - taking into consideration the highly volatile wine auction prices - but during recent years the prices obtained at the Copenhagen auctions seem to deviate systematically below especially the US prices. Especially from 1996 to 1998 the prices seem to explode which is in reasonable accordance with international tendencies of the Bordeaux Premier Cru wines. During this period, the primeur prices of Cru Classé wines kept rising from year to year irrespective of quality, e.g. the primeur price of the 1997 vintage was approximately 20% above the primeur price of the 1996 vintage despite the fact that 1996 was a significantly better vintage in Medoc as well as in Pomerol/St.Emilion. Consequently, higher primeur prices are expected to affect the auction prices on similar wines from earlier vintages in the same direction. 6 The price differences in Figure 1 probably reflect that the demand for fine and rare wines at the Danish market is relatively weak compared to e.g. the US and also the UK markets. 1 The average difference in buyer s prices is USD 100, and the question is whether this price gap is large enough to cover all costs in connection with arbitrage trade, i.e. bringing the bottles from Denmark to e.g. Chicago or London and offering them for sale there. Noting that the number of bottles supplied at the auctions normally is twelve or less and that the seller must include a fee concerning the price difference if he decides to offer the wine at e.g. Christie s in London or at the Chicago Wine Company, the margin is probably not large enough for arbitrage as seen from the sellers point of view. 2 On the other hand, the potential for making favourable bargains for foreign buyers seems to exist, because they can bid by telephone or other communication forms during the auctions in Copenhagen. Still, with permanent price difference existing between different geographical markets there seems to be barriers of trade, i.e. lack of information, transportation costs, storage costs of the wines, etc. Table 2. Average prices of the Mouton Rothschild vintage 1982 covering eight auction sites, (USD per bottle). Average price Standard deviation Low High Note: Data covering the following auction houses used for the calculation of average, annual prices: The Chicago Wine Company, Sothesby's (London), Sothesby's (USA), Christie s (USA), Butterfield s (USA), Acker Merrall & Condit (USA), Bruun Rasmussen (Copenhagen), Internet auction prices (average of international internet auctions). Sources: The Chicago Wine Company, the Wine Market Journal and Bruun Rasmussen (Copenhagen). The prices from Table 2 are calculated as simple averages from the eight wine auction houses listed in the note and presented as buyer s prices converted to USD per bottle. In contradiction with the earlier figures, prices seem to be slightly increasing during recent years, although part of the increase may be explained by the dollar exchange rates fluctuations as all prices have been converted to USD. Like before, a high standard deviation is observed in these cross-section data for 1 See Ashenfelter (1989) who compares auction prices in London, Chicago, Geneva and Amsterdam. 2 Part of the price differences is due to imperfect information as many suppliers actually do not have information about auctions outside Denmark, how to arrange for transport of the wines etc. 7 international prices of the Mouton Rothschild, and the gab between the lowest and highest prices observed in the respective periods is as high as USD 200. Hence, these icon wine prices are still at a historically high level - as well as volatile - but it is hard to believe in further price increases as this vintage is more than twenty years of age by now, and a long-run equilibrium price level ought to have been attained. 4. The icon wine prices and the business cycle In order to address the question of whether the auction wine prices are influenced by business cycle factors or not, different indicators either leading or coincident to the cycle have been selected for the analysis involving the monthly Chicago wine auction prices and the Copenhagen auction data with a quarterly frequency. Before investigating for eventual shorter run influences from the actual economic conditions, the time series properties of the wine price data are analysed, i.e. whether these can best be described as unit root, non-stationary processes or not. In Table 3, the results of the Dickey-Fuller test concerning unit roots are reported. Table 3. Unit root tests of the Mouton Rothschild prices. --- DF/ADF --- N No trend Trend Chicago Monthly data 1996M3-2003M2-3.06* {1} -3.58* {0} 82 (-2.90) (-3.46) Copenhagen Quarterly data 1996Q1-2003Q1-3.28* {1} {1} (-2.97) (-3.57) 29 Notes: Log values of the variables used in the tests and the critical values at the 5% level of significance reported in parentheses, according to MacKinnon (1991). The number of lags in the DF/ADF-test in {}-parentheses. The Dickey-Fuller tests for a unit root in the wine price series data have been performed also including a deterministic time trend in order to give strength of an alternative of trend-stationarity. The null hypothesis in these tests is non-stationarity of the variables, and as revealed in Table 3 the empirical evidence is in favour of rejecting a unit root in the wine prices - at least when no trend is included in the test procedure. Although the number of observations is reasonable, especially for the 8 monthly data, the absolute length of the time span is only seven years and therefore, also when considering the DF-tests statistics just exceeding the 5% critical values, the price variables might still show up to be non-stationary when testing for a longer time span. In Bentzen et al. (2002), a quarterly price index for is constructed for icon wines appearing at the Copenhagen wine auctions, and from this relatively long-run time series non-stationarity of wine prices is the most likely conclusion. Therefore, also first differences of the respective variables will be considered in the analysis involving causality between the business cycle and wine prices. Among the range of relevant economic indicators, the stock market indices and sales of new cars seem to be just as volatile as the Mouton Rothschild price indices in Figure 2. Hence, for bot
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks