Many people enjoy fine wine, but for most this enjoyment comes from drinking the luxury product. For Burak Kazaz, The Steven R. Becker Professor of Supply Chain Management at Syracuse University’s Martin J. Whitman School of Management, this enjoyment comes from research. On Nov. 13, 2018, Professor Kazaz delivered a talk at the Whitman School to give faculty and Ph.D. students a deep look into his research on wine futures pricing.
This research process took him from Central New York all the way to the Bordeaux region of France, where he collected data including weather, historic wine prices and professional (barrel and bottle) tasting scores to build a mathematical model that would help predict wine futures prices. Kazaz’s passion for the subject was infectious throughout the presentation, as he took the complicated supply chain and market system surrounding fine wine and explained it in a way that made it digestible and enjoyable for his audience.
The talk began with a definition of what was considered fine wine for the purposes of Kazaz’s research. The French wine market, specifically the Bordeaux region, serves as the global leader in terms of setting pricing and industry standards, so his presentation featured primarily French wines. Additionally, the types of wine included in the study were of the utmost quality, with certain brands costing as much as 2,590 euros per single bottle.
Kazaz then provided an overview of the wine supply chain, which begins with the winemakers. Physically growing and fermenting the grapes, winemakers serve as the supply chain suppliers. Eventually, the wine is barreled, bottled and sold to distributors. The intermediary between these winemakers and the final distributors are the wine markets. Kazaz focused on (and collaborated with) the leading global wine market, the London International Vintner’s Exchange, or Live-ex, for his research.
It is important to note these wine markets do not sell only physical bottles of wine. They also sell wine futures. “En primeur,” or wine futures, is a purchase method that allows distributors to purchase wine while it is still in the barrel before it is bottled. Kazaz’s research originated with a 2015 research paper, written in conjunction with colleagues Tim Noparumpa Ph.D. ’12 and Scott Webster, that supported the finding that the use of wine futures benefits winemakers.
This report garnered a flurry of media attention. But it also gained the attention of Steven R Becker, the treasurer and CFO of Southern Glazer’s Wine and Spirits, a large premier distributor of wines and spirits. Becker, a Whitman alum, wondered if wine distributors stood to take a gain or loss from Kazaz’s findings. This question led Kazaz to head a 2017 research paper dedicated to studying the impact of wine futures on the distributor side of the market, which was created with the help of doctoral student M.H. Hekimoğlu Ph.D. ‘16 and Webster.
These two papers resulted in a important finding for the wine industry — they revealed that the use of wine futures benefits winemakers and distributors alike. While winemakers benefit from the added insights about sales tactics, distributors benefit because wine futures have higher liquidity than physical bottles of wine, meaning they can make investment decisions with increased precision as they gain more data over time. In other words, it was discovered that this market is not a zero-sum game.
This research also caught the attention of the third branch involved in the fine wine market: Live-ex, the market itself. This led Kazaz to the realization that price prediction methods for mature wines were sophisticated, but no accurate price prediction method existed for young wines. He wanted to create a price estimation model for wine futures.
After intense research with a passionate team, a final model was deemed a success by the team. This model utilized explanatory variables of temperature, rainfall, Live-ex 100 index scores, barrel tasting scores, as well as an interaction variable between the positive changes in temperature and the Liv-ex 100 index. The model predicted wine prices for wine futures of the 2017 vintage. Kazaz, in a bold act, placed these predictions in a sealed envelope and gave them to Live-ex, to be opened at the end of summer 2018 after the revelation of the 2017 vintage wine future prices.
The result was stunning. This new model predicted the wine futures with an average error rate of just about 10 percent, reducing error to almost a third of the previous benchmark error rate, 34 percent. Neil Taylor, vice president at Live-ex wrote to the team, “Your predictive model is certainly the most accurate I have seen of all of the work we have either done ourselves or participated with.”
Overall, the astonishing accuracy of this new model has wide-ranging implications. Now, thanks to the tireless research of Kazaz and his many team members, winemakers, markets and distributors can efficiently plan and release pricing information and make more informed buying and selling decisions. Consumers can use this information to understand whether a fine wine on the shelf is priced below or above its true value.
Moving forward, proposals have been made to apply similar modeling tactics to the olive oil and cattle industries. Regardless of any future work to be done in other industries, one future project is on a more near-term horizon: bringing a wine futures market to the United States.