10/31/2013

Selection Bias In Art as an Asset


A new Stanford Graduate Business School report looks at art as an asset, repeat sale indexes and selection bias. The report cautions investors to be weary of reported art returns as they may be inflated and the risk involved underestimated.

The report states
In short, investors are embracing art-as-an-asset-class as if it were a newly discovered van Gogh. But is it?

Research we completed recently and presented in August 2013 at the European Finance Association conference shows investors would be wise to be wary. The returns of fine art have been significantly overestimated, and the risk, underestimated. Our research, based on the most complete auction database, BASI (Blouin Art Sales Index) shows the true annual return of art as an asset class over 1972 to 2010 was closer to 6.5%, instead of the 10% that the index shows. Moreover, holding an art fund in your portfolio does not increase the chances that the portfolio will outperform.

The underlying cause of the overestimation of returns (and an accompanying underestimation of risk) is what is known as selection bias. People have suspected this bias in the indices used to report returns of some alternative asset classes for years, but our analysis is the first, we believe, to find a way to account for it.

The selection bias arises when returns are based on indices built on repeat sales of fairly illiquid assets that are not sold at random. Many of the returns based on those kinds of indices—including the S&P/Case-Shiller Home Price Indices—may be biased upwards.

Not only are the returns of art lower than investors think, but also the risk is higher. Our analysis, of 20,538 paintings repeatedly sold between 1972 and 2010, found the Sharpe Ratio for art is 0.04, rather than the 0.24 that has been previously found. The Sharpe Ratio on U.S. equities over the same time period is 0.30. (The Sharpe Ratio is the risk-free rate of return — such as that of the 10-year U.S. Treasury bond — subtracted from the average rate of return for a portfolio or asset class, divided by the standard deviation of the return on the portfolio or class. It is the best tool for comparing asset classes on a risk-adjusted basis, and is often used by researchers to develop ideas for the best portfolios.)

The selection bias in art occurs for several reasons. Among them: Paintings that happen to be in high demand tend to go to auction more frequently and sell at higher prices. People also tend to sell the paintings that have increased in value the most since the time of purchase. A similar selection bias is probably at work in real estate, when, for instance, people sell houses after they have appreciated a lot in value.
Source: Quartz News


No comments: