The art auction houses proudly proclaim on their websites the remarkable achievement when a painting goes beyond 'expectations', but to me this shows that either they do not know the market or have deliberatively underestimated.
An example of this was at a Sunday Art Auction at Leonard Joel in Melbourne last year. The first lot of the day was an artwork by highly sought after Western Australian artist Robert Juniper titled 'Blue River, 1978'. At 43cm by 95.5cm it was a decent size for a Juniper, and with popular hues of blue and purple in a semi-abstract landscape it was obvious that it would achieve a good price. Obviously to many, but not actually obvious to the auction house providing the estimates, which provided an estimate of between $ 2,000 and $ 3,000. However, upon calling the auction house I was told that there had been a lot of interest in the painting, and that they now expected it to go for more that the high estimate. So why not raise the estimate at this point? I said I would be happy to pay double the upper estimate, and although the offer may not have been taken seriously, it was still an indication that the estimates were too low. The auction results for the previous year also indicated a much higher value. The average price for a Robert Juniper oil in 2008 was $ 16,260 with an average price per square cm of $ 2.15. Multiplying this figure with the actual size of this painting (4106 sq cms) would give a base valuation of $ 8,827. Making some adjustments for the general quality of this work and the smaller than average size (smaller paintings generally achieves a higher price per sq cm), I would have thought that an estimate of $ 10,000 – $ 15,000 would have been a more accurate estimate. The painting sold for $ 27,600 including buyers premium, which was still well above even my estimates. Even with a more realistic estimate the result would still have set the auction off to the desired dramatic start that they were probably looking for.
Valuing paintings is not always as simple as doing simple calculations based on past prices. I remember being at a Deutscher Menzies auction and watching a Celia Perceval painting 'Picking Daffodils' with a realistic high estimate of $ 5,000 getting knocked down for $ 38,400 after a long bidding war between three or four phone bidders, leaving the auctioneer to whisper to the perplexed audience 'do they think its a John Perceval'? The next lot was another Celia Perceval with a higher estimate that sold for only $ 5,520. These situations are intriguing and add to the drama of an auction, but we do not need the auction houses to artificially create them. As a buyer, I would prefer to have a more accurate indication of the expected price so that I do not waste my time with something that will extremely be beyond my budget.
Even the most distinguished auction houses can get in wrong. I remember at one of Christie's last auctions in Australia looking at the estimate of $ 5,000 to $ 7,000 on David Boyd's large 1960's work 'The Offering' from his important 'The Trial' series, and asking the staff it was a typo and was actually supposedly to be $ 50,000 to $ 70,000. I was told that the estimate was correct, and the painting sold for $ 50,190 including buyers premium.
For an up-coming example keep an eye on William Boissevain's 'Unititled, c.1970s' at Bay East's auction next week. Similar paintings usually sell for between $ 6,000 and $ 8,000, although this painting is offered with a high estimate of only $ 2,500.
I will continue to enjoy the theater of those genuine auction surprises, but with so many research tools available, feedback from the views and years of collective experience of the auction house staff, apart from the rare exception there should be no excuse for getting it so wrong.