The Confidence Game

by Maria Konnikova

  As luck would have it, David Anfam, the world’s foremost Rothko specialist, was in town from England. He came over to take a look. He declared it a beautiful example of Rothko’s style and assured Freedman it was “right,” confirming her initial impression. Of course, one opinion does not a provenance make, but the same assurances soon came from others: Stephen Polcari, an art historian and AbEx specialist who had previously served as director of the Archives of American Art at the Smithsonian. E. A. Carmean, a former curator of twentieth-century American art at the National Gallery and director of the Modern Art Museum of Fort Worth. A slew of experts passing through the Knoedler. Rothko’s son, Christopher, loved it. The experts loved it. And though the collector was anonymous, that wasn’t all that uncommon. The art world is a murky place, with many opting to remain unknown, and bills of sale often elusive. Glafira did provide a signed statement that she was legally authorized to sell the art, and Freedman’s team of researchers had not uncovered any indication to the contrary. Backed by the weight of expert opinion, Freedman decided to proceed with a sale.

  More paintings from the collection began to appear. One by one, Rosales said, the accidental owner wanted to get rid of them all. Freedman was thrilled—each one passed close inspection—but, she said, she needed a better sense of the history. “Anonymous” and “Mr. X” simply wasn’t good enough. A name was suggested, Freedman can’t rightly recall by whom. It might well have come originally from her own researchers: Alfonso Ossorio, himself an Abstract Expressionist who had advised many a client on potential purchases. Ossorio fit the timeline and the artists, and the story seemed to hold. Rosales said she would check with the owner.

  Indeed, she came back to Freedman, Ossorio had been involved. It was soon after, Ossorio’s name now in the provenance, that a buyer, Jack Levy, asked for his prospective purchase, a $2 million Jackson Pollock, to be checked by the International Foundation for Art Research, IFAR. The sale would be conditional on the painting’s authentication. Freedman readily agreed. The art, she knew, spoke for itself.

  But IFAR wasn’t as confident. The provenance, the report said, just didn’t add up. Ossorio could not be the dealer. Because of the gaps in provenance, IFAR’s area of expertise, the foundation could neither confirm nor deny the painting’s authenticity. The sale fell through.

  Freedman showed the report to the experts. Bollocks, they said. It made no sense. They would disregard it. It cast no material light on the painting itself; everything hinged on the likelihood of Ossorio’s involvement. Read the actual report: no one had any material doubts as to the Pollock’s authenticity or quality. Some of the reviewers said they couldn’t render a firm opinion, true, but the majority of the doubts hinged on the dubious provenance. Convinced, Freedman bought the painting back herself, splitting the cost with David Mirvish, a prominent Canadian collector. Her belief was strong. She would take the financial risk. Mirvish, too, read the IFAR report. He went along with Freedman; her take seemed the correct one.

  What about Ossorio? Freedman asked Rosales. He had been involved, Rosales said, but wasn’t the actual dealer. There had been a miscommunication.

  Soon, another name surfaced, this time most likely from Knoedler itself. David Herbert. A prominent dealer who knew many of the painters personally, gay like Mr. X, secretive: he was the perfect link. Rosales confirmed the find. Yes, David Herbert had been an adviser.

  The story made sense. E. A. Carmean dug into the research. The threads came together. Herbert was in all the right places at all the right times. They had found, they thought, the missing link.

  The paintings kept selling. Knoedler continued to profit nicely from the sales—as did Julian Weissman, a former Knoedler employee who had left in 1997 to start his own gallery and to whom, unbeknownst to Freedman and Knoedler, Rosales was also feeding works from the collection. Knoedler received over $63.7 million, and Weissman over $17 million. Rosales’s share grew, as well. Between 2006 and 2008 alone, Rosales netted some $14 million from the sales of her prize collection. Between 1994 and 2008, she’d sold sixty-three paintings, forty through Knoedler, twenty-three through Weissman’s gallery. Rosales claimed that she’d held on to only a portion of the commission; the rest had gone to the client.

  Push as she might for more information, Freedman kept coming up against a dead end. Could she go to Mexico to meet Mr. X? she asked. She bought a ticket for her assistant and put him on a plane. Rosales was mortified. How could Freedman betray her trust? No, a meeting was not possible.

  In 2009, just after Rosales became a U.S. citizen, the Dedalus Foundation, dedicated to Robert Motherwell’s work, issued a report on the Rosales Motherwells. Originally, its experts had declared the Motherwells authentic. But as more paintings came to the surface, they began to voice doubts. The report stated, flat out, that Dedalus could no longer stand behind the work. In their opinion, it was not by Robert Motherwell.

  The same year, the FBI began its own probe into some art that did not seem altogether right. It took time, but when, in 2012, Eric Jonke, a special agent with the IRS Criminal Investigation division, was charged with investigating, he dug quickly and efficiently to the root of the problem.

  In the end, Rosales was charged simply with tax evasion, joining the long list of criminals like Al Capone who’d gone down on the same charge. In the initial complaint, Rosales was charged with failing to disclose all of her sales income from Glafira Rosales Fine Arts LLC on her tax returns and understating her overall income. She had also failed to disclose a foreign bank account, at Caja Madrid—a legal requirement in the United States if the account in question held over $10,000, as this one did. Most of the proceeds of the sales, in fact, had gone directly overseas, without so much as a disclosure. For the 2006–2008 period, during which she’d earned $14.74 million, she had failed to report at least $12.5 million to the U.S. government. Rosales was promptly arrested.

  Lawsuits appeared. Pierre Lagrange, who had bought a Pollock from Knoedler back in 2007, for $17 million, was now demanding a refund. A forensic specialist, he said, had declared it a forgery. Freedman was dismissed. Knoedler Gallery shut its doors. But Ann Freedman stood firm: the art was good. “I had every confidence that, one day, I’d be vindicated. They’d be the foolish ones,” she told me. “I believed in those paintings with all my heart.”

  And then came the bombshell. Glafira Rosales confessed. She’d done it: passed off fake paintings as real, with the help of Bergantiños, his brother, and an elderly Chinese immigrant in Queens, Pei-Shen Qian—the painter. He had created every last one of the AbEx masterpieces, and they were all blatant forgeries.

  As she would later admit, Rosales had known all along she was committing fraud. She “then and there knew the Rosales works were fake and were not by the hand of the artists that Rosales claimed they were.”

  * * *

  How is it possible that some of the most respected names in the art world were taken in for so long by a fraud of such magnitude? How did no one spot the stream of impostor paintings that was slowly making its way into reputable collections?

  The send is that part of the con where the victim is recommitted, that is, asked to invest increasingly greater time and resources into the con artist’s scheme—and in the touch, the con finally comes to its fruition and the mark is completely, irrevocably fleeced. It’s Glafira Rosales bringing more and more paintings to Ann Freedman without any further clarity on the murky provenance that Freedman finds problematic (the send), and getting her to sell all of them, for more and more money, while Rosales herself makes plans to quietly disappear from the stage, leaving Freedman and Knoedler to deal with any fallout (the touch; here, of course, it didn’t quite go according to plan. Rosales hung on too long and was unable to leave as cleanly as she might have done a few years earlier). And while it would seem a difficult task to get people to give more where they’ve already given without return, commit more where they’ve already committed without evident reciprocity—Freedman continuin
g to sell the paintings even though the information she has repeatedly requested is not forthcoming, and ignoring any red flags in the process—it ends up being far simpler than it looks. Once the send is in motion, with the mark recommitted to raising the stakes, the touch—the con’s end—is inevitable. Once we are in, well and good, we are all in.

  On the evening of June 3, 1976, personnel from the Teton Dam project in eastern Idaho were going through a routine inspection when they noticed two small leaks, one 1,300 feet and one 1,500 feet downstream of the toe of the dam. Clear water was rushing out from the two seeps, at sixty gallons and forty gallons per minute. Concerned, the inspectors reported back. The leaks, the managers concluded, were not serious. Besides, as of nine in the evening on the following day, there hadn’t been any additional reports.

  Early the next morning, about seven, some contractors from Gibbons and Reed who’d arrived to work on the dam noticed water flowing from one of the abutments. At seven forty-five, a group of surveyors from the Bureau of Reclamation arrived at the site. There, at the toe of the dam, was another leak. And a bit higher up, another one. Immediately, they informed their supervisor. By quarter past eight, the project construction engineer and field engineer, Robert “Robbie” Robison and Peter Aberle, had both been notified. By nine, they were at the dam in person. Another leak from the abutment, this time into the embankment’s rock fill. The men gave instructions for the leaks to be properly channelized.

  By half past ten, the dam’s downstream face had turned dark: a wet spot that spread slowly outward. A deafening sound, like a crash or explosion, a “loud roar,” Robison later said. What they heard next sounded like a waterfall. A massive one. The water was flowing rapidly, taking the embankment materials down with it. Quickly, two dozers were dispatched to the site, to push rocks into the fast-eroding holes. Robison looked inside: a tunnel some thirty, maybe forty feet long, six feet across, all the way into the embankment. “The water was flowing extremely muddy,” he recalled, “exiting from the hole in the embankment about fifteen to twenty feet from the abutment.” The dozers had only been working some twenty minutes before the soil became too slippery. Soon, one after the other, they tumbled over the edge, and downstream.

  A whirlpool began to form on the water’s surface. Like a slow-motion horror film, it slowly spread, widening and quickening as it went. They tried to stop it with rocks, but nothing seemed to help. Then the sinkholes started appearing. The embankment crest collapsed. At 11:57, three minutes before noon, the dam was breached. It had taken only five hours from that morning’s leak for the whole thing to come crashing down.

  The Teton Dam failure was one of the costliest in the nation’s history. The dam had cost $85.6 million to build. In a five-hour span, the damage topped half that, $40 million, at the site alone. But the destruction went much further. About three hundred square miles, eighty miles down the Teton and Snake rivers, all the way to the American Falls Reservoir, had been flooded. Eleven people died, and twenty-five thousand more became homeless. In the towns of Rexburg and Sugar City alone, somewhere between sixteen and twenty thousand livestock were lost to the rapidly flowing water—about the strength of the Mississippi River at peak flood—and over one hundred thousand acres of farmland were now unusable, covered by floodwater that showed no signs of receding. By March 16, 1977, the damage claims topped a quarter of a billion dollars. The total claims, however, were expected to rise to $400 million—not counting the damage to the dam and its structures. Eventually, about $300 million in claims were paid out, and the total cost of the damage, by some estimates, was a full $2 billion. In short, the cost of failure was over twenty-three times that of the project itself.

  What went wrong—and could it have been averted? In August 1976, Congressman Leo Ryan, appointed as the head of the House of Representatives committee set up to investigate the events at Teton Dam, initiated hearings into the failure. It was there that some alarming evidence came to light. There had been questions about the appropriateness of the site, about the structure of the dam itself, about how construction was progressing. Robert Curry, a geologist at the University of Montana, pointed out that the Bureau of Reclamation’s study, back in 1961, had barely mentioned permeability—though the high permeability of the geologic material at the site would have certainly contributed to the dam’s collapse. In his mind, the data for the project to move forward was “inadequate” at best. Harold Proska, a geologist for the U.S. Geological Survey, went even further: the site was in a “young and unstable” area. In fact, he pointed out, back in January 1973, over three years before the disaster, the U.S. Geological Survey team had sent a memo to the Bureau of Reclamation saying that the “safety of Teton Dam project is of immediate concern.” Construction, though, had already been started. The memo was summarily dismissed. The director of design and construction, Harold Arthur, did admit that there had been problems “with extensive jointing, the fractures or potential fractures in the rock.” But they weren’t deemed serious enough to halt the building.

  The committee chair had a theory as to what, precisely, was going through people’s minds as they pushed forward despite the obstacles. He called it “momentum theory” and wondered whether anything would have caused the halt of construction once it had already gotten under way—“that is, the inclination on the part of the Bureau of Reclamation to continue dam construction, once commenced, despite hazards which might emerge during the course of construction,” he explained. Arthur was adamant: that would never be the case, be it at Teton Dam or anywhere else. Safety was their first concern. When Ryan pushed, however, Arthur did admit to one thing: not once in the history of the Bureau of Reclamation had a project been stopped once ground on the construction had been broken.

  The Teton Dam seems far removed from the world of the confidence game, except for one key similarity: once we’ve invested heavily in something, we no longer see it clearly, no matter the costs. Things that are red flags in retrospect are dismissed as irrelevant once we’ve already sunk sufficient resources—money, time, reputation—into an endeavor. It doesn’t matter if we’re dealing with artwork or something as serious as losing human lives, taking away people’s livelihoods, creating billions in damage, and setting back the environment of an entire region by decades. One would think that nothing is as important to get right as a massive infrastructure project that has an equally massive potential to inflict harm. How could they not have known? How could they not have seen? How could they not have taken note of the warnings? The fact that they don’t see or know, of course, is the reason that the send and touch are able to unfold as they are: we are so invested that we become blind, and so we raise our commitment right until the moment when the whole thing comes crashing down, be it literally (the dam) or metaphorically (an art gallery).

  To victims of the Teton Dam disaster just as to victims of the Rosales art fraud, it seems obvious that people should have known. Look at the IFAR report, the Dedalus report, the shady Mr. X. You’d have to be daft not to realize you were dealing with fakes. Look at the geological reports, the warning letters. Construction should have been ended. Ann Freedman should have realized, if not immediately, then at least once the paintings just kept coming, that something wasn’t right. Because the flags. They are there. They are waving. And people aren’t blind, right? Except they absolutely are. And the more they have invested, the blinder they become.

  In the early 1980s, Paul Slovic and Richard Thaler were discussing the crazy things people do as they make decisions to buy and sell, invest and divest, in ways that seem completely illogical to an outside observer. Why, for instance, would a family drive sixty miles through a snowstorm to get to a basketball game they don’t even particularly want to see? Well, Teton Dam, Slovic offered. Teton Dam was the perfect case in point. The family had spent money on those tickets. If they’d been free, maybe they would have stayed home and avoided hours of frustration. But the weight of the cost made it seem too important to let the tickets go. T
hey had to make the three-hour drive. They’d already committed. And the builders at Teton Dam: maybe had they not green-lighted the project, they would have stopped on those same warnings. But the weight of the cost loomed too heavy after ground was first broken. As one senator put it in discussions of another project, the Tennessee-Tombigbee Waterway, “To terminate a project in which $1.1 billion has been invested represents an unconscionable mishandling of taxpayers’ dollars.” Thaler termed the phenomenon the sunk-cost fallacy.

  The sunk-cost effect gives us a continued, strong motivation to believe in something even when the landscape has changed significantly since we first invested ourselves in it. In theory, we should only care about new, incremental costs. What we’ve already put into something shouldn’t matter: it’s lost anyway, whatever “it” happens to be—time, money, energy, whatever else. We should stick with it only if it still seems worthwhile in light of new evidence. We should abandon a dam if we see signs that the available data have changed since we authorized it. The money has already been spent, true. But if the signs are accurate, we are heading for disaster. Why throw good money after bad? We should abandon a partnership with a collector if we see signs that he’s not what we first thought him to be. The paintings have already been sold, true. But if the signs are accurate, we are headed for an even greater hit to our reputation. Why not admit you were wrong, and emerge ahead of the inevitable scandal?

  Alas, that is not at all how our minds work. The more and the longer we’ve invested in something, the more likely the sunk-cost argument is to get the better of both our reason and our perception. We don’t ignore red flags. To us, they simply aren’t there. They might be right in front of us, but we literally don’t even see any signs of danger. In a famous study on inattentional blindness, Daniel Simons and Christopher Chabris found that a majority of people fail to perceive a gorilla pounding her chest in the middle of a basketball game when they’ve been instructed to count the number of basketball passes between certain players. So busy are they with their task that they fail to see something much more glaring. The exact same thing happens by the time we get to the send and the touch: when we should be running away, we don’t even see the danger, committing ourselves further and further until we have nothing left. To a disinterested observer, with nothing invested and no preconceived notions, the gorilla is there, plain as day. To someone invested in a specific task or engrossed in the drama of the confidence game, it is essentially invisible.