The overarching narrative of last week’s marquee art auctions in New York was that all the key players—auction houses, consignors, and bidders—were being cautious. You could hear the prognostications of market watchers looking for the ominous clouds of a recession, gauging whether world events had finally precipitated the major market retraction that seems to be foretold every auction season.
In framing this narrative, it has become a common staple of arts journalism—from the New York Times
and the Wall Street Journal
to the specialized arts press
—to measure the strength or weakness of the auction season by comparing aggregate hammer prices (sometimes with buyer’s premium included) against the aggregate pre-sale auction estimates. If the sales overall come out at the high end or exceed those estimates, we have a robust market. But if they fall toward the low end or—god forbid—below those estimates, then the market is in a downward spiral.
By the same token, if one auction house has met or exceeded its pre-sale estimates and had a lower buy-in rate than its competitors, then this is often construed as a reliable economic indicator of its ascendancy in the agonism for market dominance among the three largest firms (Christie’s, Sotheby’s, and Phillips). In the absence of any other quantitative benchmark, then, meeting or exceeding pre-sale estimates is widely taken as a barometer of a rising or falling market.
However, this entire gambit of measuring pre-sale auction estimates against hammer prices is an unreliable indicator of any meaningful economic fact other than the relative success of individual auction house specialists to negotiate reasonable terms with their consignors. Sometimes they are successful at keeping pre-sale estimates low, and sometimes they are not. But that complex psychological negotiation is hardly a foundation for deriving an objective measure of the strength or weakness of the broader marketplace. We all crave quantitative certainty in measuring market trends, but building economic assumptions upon the success or failure in meeting pre-sale estimates is like forecasting in quicksand.
The uncanny prevalence of this economic measure in arts journalism may partly stem from a fundamental misunderstanding of what an auction estimate is—and the hidden alchemy used to create it. While it may appear that an auction estimate is an objective, unadulterated prediction of the prospective market range of the object by the auction house specialist, it is actually a marketing tool that is often loosely tethered to this intuition of fair market value, and is often buffeted by extraneous psychological factors well beyond that intuition.
What is an estimate?
Estimates are set by auction house specialists in negotiation with consignors. The only contractual obligation they have is to sell the work at or above the reserve price, which is not disclosed, but in the majority of cases is 50–100 percent of the low estimate. The auction estimate is thus a range set slightly above that minimum price the client will accept, but is in no way tethered to any objective autonomous economic factors to restrict its range. By comparison, the “bid” and “ask” associated with stocks before the market opens are actual offers from real market participants and can be interpreted as a true measure of open interest in a given commodity. But the auction house specialist is under no obligation to restrict the range of their estimates to any objective measure of fair market value, especially on the upside. It really all comes down to what the consignor will accept, and there is no autonomous regulation or economic condition constraining the proposed range in and of itself.
If, for example, I have an item that I construe from past market history, condition, provenance, and other factors to likely yield a hammer price of $100,000, I can express that in a number of ways. If I have a relatively compliant consignor who doesn’t push back, and no auction house competition, my preference would be to publish an estimate of $70,000–$90,000 with a $60,000 reserve, and hope that robust competition pushes it past the high estimate to my target price (or above). If my consignor is a little skittish and worried about the low reserve, I may alternatively adjust the estimate to $80,000–$120,000 with the target price right in the middle, and a reserve raised to the low estimate. If I have a particularly recalcitrant or overeager consignor and competition from other auction houses on the deal, I may ultimately concede to an estimate of $120,000–$180,000 and make my target price the reserve, in the hope the market will at least vouchsafe my intuition and allow me to get it sold at the reserve—but in such an approach, the object is inherently at risk of failing to reach its reserve and being bought in.
This scenario shows that estimates of $70,000–$90,000, $80,000–$120,000, and $120,000–$180,000 may in fact all be expressing the same underlying economic prediction. Even where the specialist in charge believes the hammer price should fall around $100,000, the final range may be adjusted significantly by a number of extraneous psychological pressures, or lack thereof. As such, meeting, exceeding, or falling short of estimates is hardly an indication of the strength of the underlying market unless this malleability is taken into account.
The four-person paradox, or how estimates are formed
“I am accustoming myself to the idea of regarding every sexual act as a process in which four persons are involved. We shall have a lot to discuss about that.” Sigmund Freud proposed this formulation in a letter to a friend. Lawrence Durrell employed this quotation for the epigraph of Justine, the first novel of his Alexandria Quartet, one of the great works of literature in the 20th century. I highlight Freud’s formulation here, as it is a useful heuristic device for understanding how auction estimates are formed. It is a process that involves four persons, so to speak: the consignor, the specialist, the (imputed) bidders, and (the prospect of) other auction house competitors.
The first step in unpacking this process is the core dynamic between consignor and specialist, where the actual negotiation takes place. As it happens, many consignors manifest a pathology that can best be described as a paradoxical interest: They often feel the need to demand the highest estimate possible, on the superficially reasonable supposition that the higher the target range, the higher the sale price. This is a paradoxical interest in the sense that an artificially inflated estimate will often asphyxiate demand, and yield a hammer price that is in fact well below their expectations in the final analysis—or worse, the lot ends up being bought in. As such, it is often the specialist’s primary responsibility to cajole the consignor into accepting the lowest possible estimate, as truly in her self-interest.
That is premised on the imputed behavior of potential bidders, our third agent at this party, whose motivations the auction house specialist will know well. The imputed bidder is often compromised by a paradoxical interest of her own: The lower the estimate, the greater the bargain may seem, but they are often not alone in having the same intuition, and that is precisely how auction specialists generate demand. If you open an auction catalog and the estimate of $10,000–$15,000 seems so incredibly cheap for the
woodcut you always coveted that you immediately call up to register to bid, it is very likely that 30 others responded similarly. Inculcating that demand is an essential part of the pressures to set estimates and, absent any other factors, auction house specialists would set them as low as reasonably possible to stimulate such demand.
The fourth party at the table, however, is the presence (or absence) of auction house competitors. A competitive situation to win a consignment almost inevitably yields inflated estimates, as every specialist in the hunt is aware of the peculiar paradoxical interest that entices consignors, and they each hope a sweetened pot will seal the deal. One major auction house used to be notorious for radically inflating the estimate for one item far beyond any reasonable interpretation of market value in a competitive bid process, in the hope of winning the consignment. Even if that one item would inevitably be bought in, the house got the rest of the collection over its competitors, and that is ultimately what matters.
The predilections and countervailing self-interests of these four agents all play a role in buffeting auction house estimates, as specialists negotiate the final range. These critical factors also help to illuminate why major private estates are so highly valued and pursued by auction houses: They represent fresh-to-market material, often in large quantities, and the consignors are typically more deferential to an auction house specialist’s proposed estimates as the “experts” in the room, and are thus much more likely to sell above estimate.
There’s always an upside
Specialists could not be happier when auction items blow way past their pre-sale estimates, even if they are perfectly aware of this possibility in advance. They may nonetheless feign surprise
that the results were so strong and a new benchmark has emerged.
A real-world example can illuminate this effect (though I have changed the circumstances to maintain confidentiality). During the height of the Chinese market craze some 5 to 10 years ago, I had a consignor with an important
vase who was not particularly involved in the art market, was unconcerned about the estimates, and had not called the competition. The specialists in charge of the sale proposed a pre-sale estimate of $500,000–$700,000, even though they were confident it would likely exceed that by a wide margin, possibly 10 times the estimated range. The vase ultimately sold for $9.8 million (with premium), which was not that uncommon for prized consignments during those frothy times. The key insight here is that auction estimates are, in some special circumstances, set well below what the specialist actually predicts lots will sell for, as they have no incentive to risk a buy-in or scare off potential buyers once that “bargain fever” sets in among bidders.
Auction estimates are primarily marketing tools tethered, to some extent, to prior market history, but ideally designed to entice maximal bidding. Together with the glossy images and the occasionally pseudo-scholarly exegesis of catalog entries, they are designed to buff the patina of the wares on offer for as many potential bidders as possible. As such, the premise
that exceeding or falling short of these targets is actually an independent barometer of market strength or weakness is mostly a canard. It indicates the relative success or failure of an auction house’s deal-making skills.
I take comfort in the fact that, season after season, the market marks itself, and is better at sniffing out anomalies or excesses than any prophetic bromides from forecasters. If the market interprets that an estimate is artificially overinflated because a specialist desperately needed to win a consignment or is trying to make his budget targets, the lot will remain unsold. All it means is that someone needs to be a better negotiator, and/or clients need to dampen their often-inflated expectations. It should not be taken for a harbinger of impending doom.