Auction Guessing 1 (Part 2): Fallibility of Christie's and Sotheby's?

As you might imagine, my estimates for the works sold at Christie's were off by a substantial amount. The ratio of the actual price to the estimate provided by the auction house at Christie's was higher than at Sotheby's for two artists but lower for the third. If we assume that the market for Latin-American Art in New York City is fairly integrated, an assumption I believe is reasonable, what does it mean that their estimates were off in different ways?

I had assumed that the comparison between the actual prices and the estimates of the painting for both auction houses would simply reveal the difference between how the current market values the work and the "essential" or inherent value of the work.  But unless the demand for two of the artists increased drastically overnight while demand for the third dropped (possible?), it simply reminds us that no matter how knowledgeable one is, it is impossible to accurately determine an "essential" price.  Does this suggest there is no inherent value, only the value placed by the market?


Rufino Tamayo's Mujer Caminando outsold my estimate of $81,000 with its price of $98,500. Similarly Diego Rivera's Cargador outperformed my revised estimate of $45,800, realizing a price of $56,250.  Surprisingly it seems Matta's Thistles didn't sell. This is perhaps too narrow of a look though, and instead we should look at the price/estimate ratio of all these artists' work sold at the Christie's auction in comparison to those of Sotheby's.

The price to estimate ratio (price realized/average estimate), hereafter referred to as P/E, of Tamayo's work at the Sotheby's auction was 1.0136, but at the Christie's auction it was 1.1281.  The P/E of Rivera paintings at the Sotheby's was 1.144, while the P/E at Christie's was 1.08.  Though the work by Matta which I featured didn't sell, the 11 works by Matta which did sell at Christies sold at a P/E ratio of 1.267, significantly higher than the Sotheby's P/E of 1.128.

Though it isn't very surprising that these "price to estimate" ratios vary between the two auction houses, it is somewhat surprising to me that they don't vary in a consistent way.  The P/E of Tamayo and Matta were significantly higher for Christie's than they were at Sotheby's, while the P/E of Matta was much lower for Christie's.  This negates the hypothesis that one is simply being more conservative in their estimates than the other.  Though a lot of explanations could be hypothesized based on this evidence, for now it suffices to say that they show that neither auction house is infallible in making their estimates. Furthermore, the differences between their estimates and the prices achieved cannot be explained by simply saying that the aggregate demand for Latin American art in late May 2011 was x% higher than expected by the art (market) experts.