Wine market prices and investment under uncertainty: an econometric model for Bordeaux Crus Classés - PDF

Agricultural Economics 26 (2001) Wine market prices and investment under uncertainty: an econometric model for Bordeaux Crus Classés Gregory V. Jones a,, Karl-Heinz Storchmann b,c a Geography Department,

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.

Devices & Hardware

Publish on:

Views: 17 | Pages: 19

Extension: PDF | Download: 0

Agricultural Economics 26 (2001) Wine market prices and investment under uncertainty: an econometric model for Bordeaux Crus Classés Gregory V. Jones a,, Karl-Heinz Storchmann b,c a Geography Department, Southern Oregon University, 1250 Siskiyou Blvd., Ashland, OR, USA b Rhine-Westphalian Institute for Economic Research, Essen, Germany c Yale University, New Haven, CT, USA Received 17 June 1999; received in revised form 20 June 2000; accepted 13 July 2000 Abstract This paper describes an econometric assessment of wine market prices for 21 of the Crus Classés châteaux in the Bordeaux region of France. The model developed in the analysis attempt to define the relationship between factors that influence wine quality and those that influence wine prices. Characteristics of the models are: (1) climate influences on grape composition (acid and sugar levels), (2) grape composition influences on market prices, (3) subjective quality evaluations (Parker-points) on market prices, and (4) the effects of age of the wine on market prices. The results indicate that composition levels of Merlot-dominated wines are more climate sensitive than those from Cabernet Sauvignon-dominated wines. Overall, warm, dry summers result in high sugar and low acid levels at harvest which in turn lead to higher quality wines. Wine market price sensitivity to Parker-point ratings indicates that properties with high Cabernet Sauvignon-dominated wines are highly dependent on the external ratings while Merlot-dominated wines have a decreased rating sensitivity. Smaller properties tend to gain over proportionally from high ratings indicating great jumps in price from year to year. Additionally, châteaux that have experienced high ratings for past vintages exhibit great sensitivity to point steps in ratings for current vintages. Aging has a positive effect on Bordeaux wine pricing. This is due to the increasing maturity as well as the increasing absolute scarcity. Absolute scarcity of product is expressed by the size of the property, with small properties producing less per vintage and therefore having less in the market. Additionally, Merlot-dominated wines exhibit more maturing potential and profit more from aging than Cabernet Sauvignon-dominated wines. Average per château real annual profit ranges from 1 to 10%. High levels of grape ripeness, absolute scarcity, and smaller properties that are dominated by Merlot in their blend lead to the highest profits. Forecasts for a vintage not yet on the market indicates that 1995 is better than 1994 for both Cabernet Sauvignon and Merlot-dominated wines, but that 1996 and 1997 are not as good as 1995, especially for Merlot-dominated wines Elsevier Science B.V. All rights reserved. JEL classification: C31; G1; Q11 Keywords: Bordeaux; Wine pricing; Econometric model; Viticulture; Climate 1. Introduction Corresponding author. Tel.: ; fax: address: (G.V. Jones). Wine aficionados and collectors interests have been raised by the sensational prices which old /01/$ see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S (00)00102-X 116 G.V. Jones, K.-H. Storchmann / Agricultural Economics 26 (2001) wines sometimes achieve in auctions. Since many of the famous red wines like Châteaux Margaux, Mouton-Rothschild, Lafite-Rothschild or Petrus, which the average wine consumer has probably not experienced, but certainly heard of this occurs especially to Bordeaux wines. The interest of professional investors in Bordeaux wines has increased dramatically in the last few decades because the profit yielded per year can be tremendous. According to a ranking of the best investment-wines of 1996, yields of 500% and more are routinely achieved (Blättel and Stainless, 1997). Therefore, the economic literature has begun to examine this theme with the goal to recommend appropriate investment strategies and portfolios (i.e., Schubert, 1996). In general, these recommendations are based on experiences of the potential for the development of certain wines and on intuitive knowledge of the product. Even with this increasing interest in wine investment, empirical studies on wine pricing are still rare. This lack of economic assessment is in fact amazing, first, because of the wide spread variance of the prices of single châteaux, and secondly, in respect to the variance within single vintages. For example a bottle of Château Cheval Blanc from the 1972 vintage is available for less than 20US-$, while the 1982 vintage from the same château costs about 500US-$. Two of the rare empirical investigations on this topic are those of Combris et al. (1997) and Ashenfelter et al. (1995). After estimating a hedonic price function for many Bordeaux wines, Combris et al. (1997) come to the conclusion that wine prices are rarely caused by organoleptic (sensory) characteristics. Although it is true that experts, through many tastings, can build a complex quality profile for individual châteaux or vintages, the information is not overly useful for the average consumer since they would not want to repeat this procedure. Given that the average consumer only has general and sometimes imperfect information on quality, label notoriety seems to be much more important. Since Combris et al. (1997) investigated only young wines of all quality levels it is difficult to apply these results to top Bordeaux wines due to the enormous price variances of different vintages. This empirical shortage is remedied by the econometric approach of Ashenfelter et al. (1995). The authors developed a panel equation, termed the Bordeaux-equation, which defines the influence of aging and different weather characteristics on a generalized Bordeaux index. Beside the evaluation of particular determinants, this approach also provides a relative prediction of prices for wines which are not yet on the market but have been sold as futures. It is obvious that the generalized Bordeaux-equation, as an average index, cannot explain the price of a single châteaux wines with the necessary exactness needed in an investment. This problem could be solved in general by estimating not a region-wide average index but single equations for each château. On one hand this approach could easily be adopted to help explain and forecast prices of single château wines. On the other hand, the blend of each château cannot be neglected since the top wines of Bordeaux are cuvées made from different varieties that are expected to react in a different way to climatic factors. Therefore, it is probable that the quality and prices could also be different according to the particular blend. This should be taken into account allowing the objective quality of the single varieties to be estimated separately. However, according to Combris et al. (1997), the impact of the wine quality on the price is not compelling since the quality information is not available or obvious to the average consumer. This information shortage has been remedied by the development of numerous subjective quality point systems and rankings in the specialized literature (e.g., Broadbent, 1981; Penning-Roswell, 1989; Parker, 1985). Since the hedonic approach assumes that most wine consumers consider these rankings in their buying decisions, their inclusion in an economic model is warranted. Accordingly, the following paper introduces a recursive econometric multi-equation model on Bordeaux wine pricing. First, the model quantifies the dependence of Crus Classés prices on climatic influences, grape composition, Parker-points, and aging. Second, the model allows one to forecast wine prices in order to compare achievable market prices and actual future prices. In the first section the general structure of the model and the database used to construct it is explained. Next, the climatic influences on wine composition are described for the Bordeaux region. Then the recursive model blocks composition and price and the corresponding equation specifications from which the adaptability of a simultaneous solution is shown. The final section presents model simulations and sensitivity results. G.V. Jones, K.-H. Storchmann / Agricultural Economics 26 (2001) Table 1 Prices of selected Bordeaux wines in US-$ per bottle, 1996/1997 a Château Vintage Beychevelle Cheval Blanc Cos d Estournel Ducru Beaucaillou Grand Puy Lacoste Gruaud Larose Haut Brion Lafite Rothschild Latour Leoville Barton Leoville Las Cases Lynch Bages Margaux La Mission-Haut-Brion Montrose Mouton Rothschild Palmer Petrus Pichon Comtesse Talbot Troplong Mondot a Data from Blättel and Stainless (1997). A value of zero means the wine was not available at auction. 2. Data and methods One of the more typical features of great Bordeaux wines is their extraordinary longevity. It is through maturity, achieved during long storage, that most Bordeaux wines develop their typical character. Many vintages have life expectancies of several decades and have been known to remain viable for over a century. Together, the longevity and the high profit yield expectancy, make Bordeaux wines not only a consumption good in great demand but also a desirable international speculative good. Hence, the wines of many of the châteaux from Bordeaux (many other top wines from other regions are represented as well) are traded all over the world in established wine auctions. This system guarantees, similar to a stock market, a comparatively high price transparency. Therefore, it can be assumed that auction prices indicate the relative (economic) scarcity and therefore the international esteem for those wines. For this analysis, the price database refers to all relevant wine auctions held worldwide within 1996/1997 and is calculated as the weighted arithmetic average of all auction prices of one wine. 1 To obtain a complete set of data, only frequently traded wines are considered in this analysis. Therefore, the current investigation is restricted to the vintages and to 21 châteaux (Table 1 gives the price data set used in the analysis). The model is conceived as a pure panel model (cross-sectional model) which attempts to explain the wine prices of different châteaux and vintages. All equations are specified linearly, logarithmically or semi-logarithmically, and are estimated with the ordinary least square (OLS) technique. The model is recursive and consists of two sections: a composition and climate section and a price section. First it is assumed that climatic variables such as precipitation, insolation or temperature determine particular quality features of the grapes the composition of the raw material. Here, the model distinguishes between the two main varieties grown in Bordeaux: 1 All auction results are to be found in Blättel and Stainless (1997). 118 G.V. Jones, K.-H. Storchmann / Agricultural Economics 26 (2001) Fig. 1. Structure of the econometric model on wine pricing. Cabernet Sauvignon and Merlot (which account for over 80% of grapes grown in the region) and explains their specific sugar and acid content at harvest (four separate equations). These endogenously determined variables are then used as explanatory variables in the price section of the model, which ultimately consists of 21 equations, (i.e., one for every château in the analysis). According to the average blend of the particular château, the specification of the equation involves more Cabernet Sauvignon or more Merlot compositional features. In addition to these exogenous wine quality variables, variables for the best known international quality rankings for each château, the so-called Parker-points (Parker, 1985), and the relative ages of the particular wines are included in the equations. The complete structure of the model is given in Fig Climate, phenology, and composition effects on quality Grapevines are a geographically expressive crop, being grown in distinct climate regimes worldwide that provide the ideal situations to produce high quality grapes. This is nowhere more evident than in Bordeaux, a region that is synonymous with some of the best wines in the world. While the interactions between the local climate, soil, and site location (termed the terroir by the French) play a varied role in the growth and output of the grapevines, the general effect of the climate is well known. Mild to cool and wet winters followed by warm springs, then hot summers with little precipitation produce the best wines (see Jones (1997) for a review). Therefore, there is an optimum climate regime that contributes greatly to the overall quality of a given vintage. Occurring as a direct effect of climate, the grapevine s growth can be described by its phenological events. Phenology is the study of individual physiological events or growth stages of plants or animals that recur seasonally in response to climate. Understanding the phenology of a given plant system is important in determining the ability of a region to produce a crop within the confines of its climatic regime. From a husbandry viewpoint, knowledge of a plant s growth stages is advantageous as cultural and chemical practices can be applied at optimum times in a plant s annual growth cycle. Additionally, information regarding growth stages can be useful in estimating crop yields. Vitis vinifera grapevines ( wine-bearing vines ) are a phenologically distinct crop with the most important developmental stages being débourrement (bud break), floraison (flowering), véraison (color change and maturation nascent), and harvest (grape maturity). 2 The timing of these developmental stages is also related to the ability of the vine to yield fruit, with early and fully expressed (unhindered by extremes of heat or cold, storms, etc.) phenological events usually resulting in larger yields (Jones, 1997; Mullins et al., 1992). Additionally, the phenological timing has been related to vintage quality with early harvests generally resulting in higher quality vintages (Ribereau-Gayon and Guimberteau, 1996). Many studies looking at the relationship between climate and quality have employed monthly averages in temperature and precipitation as the independent variables (for a good review see Gladstones, 1992). Given that plants do not respond to a calendar division of climate data, and that phenological timing and quality are related (Jones, 1997), each vintage in the Bordeaux region is divided according to the 2 For the main red varieties grown in Bordeaux, bud break occurs in late March or early April, floraison occurs in early June, véraison occurs from mid to late August, and harvest usually commences at the end of September or in early October (Jones, 1997). G.V. Jones, K.-H. Storchmann / Agricultural Economics 26 (2001) major phenological events of bud break, floraison, véraison, and harvest, thereby creating four stages 3 based upon the grapevine s annual response to the prevailing climate (Jones and Davis, 2000a). This allows for a comprehensive analysis of the climatic influences on the quality using a physiological approach. Climate variables of precipitation, estimated potential evapotranspiration (PET), 4 and the number of days with temperatures more than 25 and 30 C are summed by day and phenological stage to produce up to four independent climate variables per stage that could ultimately play a role in quality levels. Nearing harvest time, key vintage characteristics are the chemical composition of the grapes. Two of the chief determinants of crop ripeness are the relative amounts of sugar and acid found in the berries leading up to harvest (Mullins et al., 1992). Sugar represents a measure of the potential alcohol content of the wine and total acidity is a measure of the fixed and volatile acids present in the berries and has a direct influence on wine color, the growth of yeast and bacteria, and its pleasing balance with sugars on flavor qualities. During maturation, the levels of these two measures generally proceed in opposite directions: sugar levels increase and acid levels decline (Amerine et al., 1980). Optimum ripeness levels of sugar and acid vary by variety and region but should range between g/l of sugar and 3 9 g/l of total acidity (Winkler et al., 1974). In general, relatively high sugar levels produce better quality while high acid levels produce lower quality (Ribereau-Gayon and Guimberteau, 1996; Jones and Davis, 2000b). Ideally, there exists a ratio of sugar to acid that determines proper ripeness and quality potential. Daily climate data of maximum temperature, minimum temperature, and precipitation for the Bordeaux station for are obtained from METEO-France (1998). The climate data are divided 3 For the remainder of this analysis, each of the four stages will be represented by D = the dormant stage (harvest of one year to bud break of the next), B = the bud break stage (bud break to floraison), F = the floraison stage (flowering to véraison), and V = the véraison stage (véraison to harvest). 4 The PET variable is equal to the sum of the average temperatures (T max T min )/2 minus precipitation. It is a derived temperature-related variable that has been used to study the climate/viticulture relationship. It is referred to as the Ribereau-Gayon and Peynaud Index. by the mean phenology of the grapevines 5 from reference vineyards observed by Ribereau-Gayon and Guimberteau (1996, pers. commun.). Composition values (acid and sugar levels) for Cabernet Sauvignon and Merlot grapes for are also supplied by Ribereau-Gayon and Guimberteau (1996). Acid and sugar levels are measured at the reference vineyards prior to harvest and are averaged to obtain a single value for each vintage and variety. Table 2 shows the values for the significant climate variables in the analysis and grape composition for the years used in the analysis. The estimated equations for the climate effects on Cabernet Sauvignon and Merlot grape composition are shown in Table 3. Cabernet Sauvignon sugar levels are influenced by precipitation throughout the growing season. Too much early season precipitation (BPREC) has a negative effect by delaying growth and late season precipitation (FPREC and VPREC), especially during the ripening period, has a negative impact by possibly diluting the berries and producing lower relative sugar levels. Also many days of warm temperatures during floraison would mean that the grapes ripen rapidly and to a higher degree. Increased precipitation over the growing season increases the Cabernet Sauvignon acid levels and results in unripe grapes. On the other hand, warm and dry conditions, especially from flowering to harvest, allow the acid levels to decline and the grapes to become ripe. A similar set of climate influences on Merlot grape composition is seen with warm and dry periods of growth providing higher sugar levels while wetter periods coincide with higher acid levels Wine prices The price of a wine is determined by many different factors. One of the more important factors is an assumption that the value of a wine is greater the older it is. For the majority of wines this generalized assumption is actually not correct. Only outstanding wines, to which it can be argued that all of the wines 5 The dates of floraison, véraison, and harvest are supplied by Ribereau-Gayon and Guimberteau (1996, pers. commun.) and the dates of bud break are estimated from climate data with bud growth considered to occur when the mean daily temperature is above 10 C for five consecutive days (Mullins et al., 1992; Jones, 1997). 120 G.V. Jones, K.-H. Storchmann / Agricultural Economics 26 (2001) G.V. Jones, K.-H. Storchmann / Agricultural Economics 26 (2001) Table 3 Climate and phenology influences on sugar, and acid for Cabernet Sauvignon and Merlot: equations and statistical tests a Independent variables Dependent variables (g/l) CABSUG CABACID MERSUG MERACID DPET log(x) (3.11) (1/x) (6.97) BPET (1/x) (3.78) BPREC (1/x) 0.81 log(x) (2.61) (6.84) FPET (1/x) (8.13) FPREC 0.06x (
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