By Fabian Bocart and Christian Hafner


Understanding the behavior of art prices is important to precisely anticipate the financial benefits of owning art. So far, a very common approach in the scientific literature is to assume that all artworks follow a common, deterministic path. What happens if we reverse the perspective assuming that art prices behave randomly?

Artworks are largely considered as dual consumption goods, that is, they can behave both as investment and consumption. For instance, paintings can be considered an investment because the value of art can raise as time goes by and they can be considered as consumption because collectors can enjoy a nice artwork on their walls or enjoy owning a piece of art history. Understanding the behavior of art prices is important to precisely anticipate the financial benefits of owning art.

A common problem with artworks is that transactions are rarely observed. Most pieces of art are totally unique and sell only very sporadically. Nevertheless, it is possible to find a common behavior of prices in the art market: by observing many sales, and correcting for individual characteristics of each artwork, one can identify the underlying common evolution of prices in the market. This is the principle behind most research on the economics of art. A very common approach in the scientific literature is to assume that all artworks follow a common, deterministic path. Simple statistical modelling is then used to estimate this path.

Unfortunately, we find that assuming that art prices are deterministic is not very realistic. Like many assets, such as stocks, bonds, or other types of commodities, it is not very convincing to assume that art prices are determined by some kind of fixed natural laws. If such was the case, then we could perfectly anticipate future prices. In practice, anticipating the future of prices is not easily feasible and largely contradicts fundamental laws of finance and economics.

Instead, we suggest a new model that assumes that art prices behave randomly, in a fashion similar to the stock market. This means that we acknowledge the fact that paths of art price may be somehow unpredictable, but we also make sure that, in case they are not and that trends can be identified, we fully capture that information. In our study we provide such a new methodology to measure the variation of prices of artworks.

By doing so, we offer two elegant measures for the art world. First, we develop a methodology that allows users to track the price of art at a very granular level: monthly steps. Other known methodologies offered only quarterly steps at best and are prone to imprecise results at higher frequencies. Second, we derive a precise formulation to compute volatility of art prices. Volatility is the deviation of financial returns with respect to the mean return. It is an important metric to compute for instance the proportion of art a financial portfolio should contain. Another application of a precise volatility metric is the valuation of insurances.

In our study, we exploit a database of auction sales of very famous artists. We show that volatility of the art market increased during the financial crisis of 2008, and during the debt crisis of 2011. Nevertheless, prices behaved differently: during the financial crisis, prices dropped along with many other asset classes, but during the European crisis of 2011, prices increased. We show that other known methods to compute returns of art fail to properly measure the pulse of the market and can lead to wrong conclusions. Finally, another property of the model for future applications is that it allows integrating other asset classes, such as stocks or bonds to predict the future of the art market in case of correlation. Because we have observed strong relationships in 2008 between the art market and the stock market, we expect to find relevant information in the stock market to predict the art market.

Nevertheless, we are aware this new methodology can be even further improved: for future research, we suggest waiving certain assumptions on the type of randomness that impacts art prices. In a recent research paper by Cagnone et al (2016), the authors offer a certain type of generalization of our methodology. They apply successfully this new method to tribal art prices. In general, this new generation of methodologies based on non-deterministic modelling can not only be applied to the art market, but can be used to estimate the evolution of prices of any asset class that exhibits heterogeneity in their characteristics: real estate, musical instruments, computers, cars, base-ball cards, yachts and planes, to name but a few.

This text is based on:
Bocart, F. Y. R. P., & Hafner, C. M. (2015). Volatility of price indices for heterogeneous goods with applications to the fine art market. Journal of Applied Econometrics, 30(2), 291-312.

Authors information:
Fabian Bocart is a specialist of collectibles as an asset class.
Christian Hafner is professor of econometrics at Université catholique de Louvain (UCL) and President of the Louvain School of Statistics, Biostatistics, and Actuarial Sciences (LSBA), Belgium.

Image source:
Francis Bacon, 1953, 153 x 118cm, Des Moines Art Center, courtesy Wikimedia Common.

May 13, 2024
The Economic Impact Of Unesco World Heritage Designations: Evidence From Italy
April 29, 2024
Rethinking Music Streaming Royalties – How Do We Better Align Consumer Value And Artist Payments?
April 15, 2024
Cultural Capital Externalities
April 1, 2024
Examining The Potential Disruption From Ad-supported Streaming Services

Become a member

Members of the ACEI will be part of a network of scholars, researchers and practitioners interested in advancing cultural economics.

Join Today