Price Dynamics by Auction House

By Joseph L. Breeden

How important is where you buy and sell to the price of the wine? Rarely can we compare the same wine sold at two houses in a similar time frame. Rather, to assess the importance of auction house, we included auction house as a predictive factor in our price model for fine wine. This has the affect of first normalizing the price of the wine for all the other factors in our model: age, vintage, rating, etc. Thereby we can make a kind of adjusted basket of fruit to basket of fruit comparison, if not quite apples to apples.

With nine auction houses in the study, the distribution of prices was measured separately for Lafite, Bordeaux excluding Lafite, and Burgundy. In all cases, the scaling coefficients by auction house were statistically significant, meaning that auction house is a useful predictive factor for price. For Lafite and Bordeaux excl. Lafite, the price spread was one third of an order of magnitude – a large number in dollar terms. For Burgundy, the spread was even higher at 1.4 orders of magnitude in price.


To make sure that the distribution of price deltas by auction house is not just random noise, the next figure shows the price deltas of Burgundy versus Bordeaux where each point represents one auction house. Although Burgundy has a greater price spread than Bordeaux, the results are highly correlated. Auction houses obtaining the highest prices for Bordeaux also get the highest prices for Burgundy, and vice versa.


As before, the effect is statistically significant and normalized for all prior effects (wine age, specific wine pricing, market conditions, and unit sizes). However, the causality is open to interpretation. Wine history and condition is not included in the analysis, so it could be that certain auction houses only carry wines in better condition. Conversely, it could be that the wines are the same, but that auction house brand and participation account for the higher prices.


Joseph L Breeden, PhD has been building forecasting models for over 20 years for such areas as currency futures, sporting events, agricultural commodities, and loans. The methods he pioneered in his book, Reinventing Retail Lending Analytics (2014) are considered are considered best practice in the industry and performed well through the US mortgage crisis and many other international economic crises over the last 20 years. His love of wine and data analysis led to his participation in