Never Let a Serious Crisis Go to Waste

by Philip Mirowski

  The “losers” are not left high and dry. By determining the market clearing price, the auction increases liquidity . . . The auction has effectively aggregated information about the security’s value. This price information is the essential ingredient needed to restore the secondary market for mortgage backed securities.130

  What mattered, they insisted, was “information”: information would summon forth funds from private actors, thereby thawing frozen secondary markets. The basis for this claim was that the assets to be purchased were “common,” or objectively, valued. According to conventional theory, one should expect in such cases that purchasers of such assets should find themselves misjudging this objective value, resulting in a kind of undesirable behavior called the “winner’s curse.” Market designers believed they could mitigate such problems by designing markets that efficiently aggregate information, and thereby assist market participants to discover the true value of items being sold. Although one way of reading the market designers’ argument was that one should generally trust existing markets to do the best job of aggregating information about assets, there were specific flaws (resulting from the nature of the commodity exchanged) that necessitated a suitably trained economist to provide a helping visible hand. In circumstances like these, with the largest financial firms in the nation perched on the precipice of default, the stakes were dangerously high, making their participation all the more crucial.

  However, in practice, the auction design process would encounter serious difficulties. The Treasury had initially selected as a baseline auction a design that “although undoubtedly sub-optimal in the formal mechanism design sense, it was deemed simple, transparent, and robust enough to be implemented rapidly and effectively.”131 In a crisis, especially important was the speed of deployment, since from the perspective of the “Break the Glass” memo, the Treasury had already lost a week due to the House’s rejection of the first version of TARP. Unfortunately, the market designers responded to the Treasury’s call for assistance by submitting widely incompatible designs for the auctions, necessitating the Treasury to decide between the rival analyses. By itself, the presence of rival proposals was not an insuperable obstacle, but complicating matters was that from the perspective of the Treasury one could not tell on paper what the best auction form was.132 For example, one dispute broke out over whether to run an “open” or “sealed bid” auction. This had historically been one of the most basic issues that market designers grapple with, and, yet, one proposal called for an “open” auction, yet another for a “closed” auction. Which one was to be preferred was supposed to turn on which mechanism did the best job of aggregating information, but theory provided neither guidance about which form was better, nor guidance about whether either form would bring new bits of useful information into the market. There were an enormous number of distinct heterogeneous securities (more than 23,000 types), but apparently there was no reliable price information from either markets (which had ground to a halt) or standard simulation methods (which had proved unreliable). Therefore, there was no reason for any market participant to generalize from information released by getting the price “right” for one security to the thousands of others available. The market designers placed in charge of implementing the auction acknowledged, “the relevant issues could not be addressed directly with economic theory.”133 So much for the bracing clarification of microeconomics. The dispute over auction forms raised a second, more serious problem: there was no good reason to believe that the auctions would do what the market designers had said they would: namely, summon a chain of events that would eventually bring the economy out of crisis by, in the first place, aggregating dispersed information. After all, no work had been done previously by market designers on how to fix a collapsing economy.

  Since market designers could identify no single optimal auction, the Treasury decided to set up two teams and asked them to more fully develop their proposals. Whereas the “Break the Glass” memo called for announcement of auction terms within two weeks of TARP’s passage, followed by the commencement of auctions in another two weeks, it took until the end of October to even manage to narrow down the candidates to two alternatives. Projections at that point had the first auctions beginning no earlier than December. Some of the Treasury staff became increasingly nervous about performance, regarding the auction design process as a “science experiment” run amok: market designers had always insisted that the performance of the auctions was sensitive to even seemingly minor changes in rules, and yet they could not even agree about how rule changes would affect performance. They wanted to implement both of the alternative auction forms and use the first set of auctions as trial runs, a prospect that surely failed to inspire confidence. And this in the midst of a collapsing world economy.

  Meanwhile, markets themselves had turned against the TARP plan. Things initially had started out well for the Treasury. The first announcement of the toxic asset purchase plan led immediately (on September 18) to a gain on the Dow of 410 points, followed by another 369 points the very next day. Paulson observed that the Treasury’s plan had “acted like a tonic to the markets.” Unfortunately, matters went from bad to worse to catastrophic over the course of the next two weeks, at least if one was guided by the stars of markets. The Dow plummeted, and credit markets remained frozen. While it was tempting to attribute the declines to the initial failure to pass the TARP, its passage on Friday, October 3, made this a difficult position to maintain, since the declines continued unabated. When the declines resumed the following Monday and spread across the world, Paulson interpreted financial markets as having judged that “TARP would not provide a quick enough fix.” But by then the handwriting was on the wall: Bernanke and various Treasury staff had been for at least a week expressing doubts that the asset purchase program would work; Paulson himself intimated to President Bush that the Treasury would probably need to purchase equity in the banks on October 1, two days before TARP’s passage. On October 13, Paulson informed the CEOs of Citigroup, Wells Fargo, JP Morgan, Bank of America, Merrill Lynch, Goldman Sachs, Morgan Stanley, Bank of New York Mellon, and State Street Bank that the Treasury now intended to emphasize capital injections—and he instructed these nine banks to accept them. By the end of October, Paulson canceled the auctions and instructed his staff to concentrate on capital injections instead. When markets judged the prospective market-based program to be faulty, the Treasury heeded the markets, not the economists.134

  To the Treasury, it ultimately didn’t much matter whether it resorted to boutique auctions or capital injections: to be sure, Paulson and his staff might lose face through their reversal, but Paulson had been considering such action well before the TARP passed. What mattered was political efficacy, not austere academic notions of “efficiency.” The main policy objective was to stop the market free fall without succumbing to bank “nationalization,” and the capital injection program basically accomplished that. But to the economist market designers, it made all the difference in the world. The market designers still under contract responded to the Treasury’s volte-face in emphasis by insisting that there was no good reason the Treasury could not use auctions to purchase bank shares in addition to toxic assets, a position they maintained until the Treasury made it clear they had no intention to seek release of any additional TARP funds, therefore foreclosing any prospect for using auctions.135

  Once that happened, things turned ugly: the market designers for hire themselves became some of the most fierce critics of the TARP, denouncing Cash for Trash. In an interview for NPR, Ausubel complained, “Instead of conducting transparent auctions, the Treasury is going to instead distribute suitcases of cash”; for Cramton, “It really is moving down the path to crony capitalism, in my mind, where the government is picking winners and losers in a nontransparent way.”136 This turnabout, however turncoat, was easy to pull off because both the market designers and the anti-TARP petitioners now claimed to have shared very similar assumptions about the economi
c role of government. At times these shared views became apparent: during the period of the most heated disagreement, Charles Calomiris (a Cato-based member of the NTC) stated to NPR, “If Larry [Ausubel] can convince me that he’s got the right mechanism, that’s great”; Calomiris went on to point out that he and Ausubel actually agreed on many things.137 And so the neoliberals welcomed the market designers.

  In hopes of getting auctions back on the agenda for the incoming Obama administration, they publicly promoted what appeared to be a scientific demonstration that the Paulson Treasury had taken the wrong approach. Cramton claimed that his study had demonstrated “the auction was a success. The banks traded their toxic assets for solid capital, and the taxpayers got a fair deal.” The fact that these “banks,” “taxpayers” and “assets” were only sketchy constructs in a laboratory did not necessarily detract from the lesson, and might even have served to highlight the difference between the naked politics governing Paulson’s Treasury and the calm impartial science of the market designers. One can institute real markets in laboratories, removed from the noise of the real world, and these real markets deemed the market designers to have been correct. But what did getting a “fair deal” have to do with the market designers’ theory of the crisis? At most it seemed only to address the issue of whether one could consider the price generated to be a legitimate “market” price—a matter that really could not be settled by an experiment, anyhow. This was a far different claim than the one they had originally made to the Treasury: in their initial submission, Ausubel and Cramton argued not for the aggregation ability of markets in general, but instead for a very specific kind of “clock auction.” While their early public statements did take care to portray their auctions as marketlike, they tended to emphasize to the Treasury how their clock auction improved upon other designs:

  A frequent motivation for the use of dynamic auctions is reducing common-value uncertainty. In this setting there is a strong common-value element. A security’s value is closely related to its “hold to maturity value,” which is roughly the same for each bidder. Each bidder has an estimate of this value, but the true value is unknown. The dynamic auction, by revealing market supply as the price declines, lets the bidders condition their bids on the aggregate market information. As a result, common-value uncertainty is reduced and bidders will be comfortable bidding more aggressively without falling prey to the winner’s curse—the tendency in a procurement setting of naïve sellers to sell at prices below true value . . . A principal benefit of the clock auction is the inherent price-discovery feedback mechanism that is absent in any sealed-bid auction format. Specifically, as the auction progresses, participants learn how the aggregate demand changes with price, which allows bidders to update their own strategies and avoid the winner’s curse . . . Efficiency in the clock auction always exceeded 97%.138

  This passage corresponds to the point made above, that market designers viewed the toxic assets as “common” or objectively valued, and that such cases posed for the market designer the task of figuring out how to aggregate information. It also makes explicit the mandate of Bernanke’s warning to avoid selling assets at too-low prices (although market designers offered a different rationale for doing so—avoiding the “winner’s curse”). But what is most notable about this passage is that it advocates a very specific type of auction—a clock auction—and does so on the basis of its ability to avoid the “winner’s curse,” as evident by its demonstrably superior “efficiency.” The reason this claim is so notable is because it is incoherent on its own terms: it only makes sense to attribute “97 percent efficiency” in the case of private valued auctions.139 If toxic assets are common valued, then all distributions are efficient by definition, and therefore the efficiency criterion is useless, or at best, irrelevant. While the criterion does make sense in the case of private valued auctions, one can never suffer from the winner’s curse in such cases, again by definition, and therefore the argument to prefer the clock auction on the grounds of information aggregation is nonsense. Since the market designers’ claim that one could avert the crisis by increasing information about the value of assets implied that the assets must be common valued (or else the link between auction performance and crisis aversion is severed), the “efficiency” evidence was especially misleading.

  Now, experts in game theory during the past decade should immediately recognize that the claim advanced by Ausubel and Cramton to the Treasury in support of their “clock auction” was misleading. But perhaps the point of the exercise was never to get the particulars of the economics justification correct, but instead to get the Treasury to purchase their “clock auction.” Sifting through all the coverage of the TARP plan, one comes across an acute observation made by a Newsweek reporter:

  [Ausubel and Cramton] hope to convince officials that not only does a reverse auction work, but, in the event the Treasury conducts one, to run it off their patented software platform . . . Ausubel and Cramton own two auction-services companies, Power Auctions and Market Design, each of which handle the back end of auctions for companies and foreign governments. They’ve already helped the French government sell electricity off its grid and Dutch energy companies auction off natural gas.140

  In fact, Power Auctions and Market Design, Inc., held the patents for the stipulated clock auction. But the presentation delivered by Ausubel and Cramton for the Treasury listed several additional “Typical Auction Related Activities” (product design; definition of detailed auction rules; auction software specification, development and testing; bidder training; establishment of an auction “war room”; operation of auction; post-auction reports on success of auction and possible improvements for future auctions) for which Power Auctions and Market Design, Inc., could provide assistance.141 Of course one can’t know the exact provisions that would have been contracted between the parties, since the Treasury scrapped the plan, but given the previous record of market designers, it is entirely reasonable to believe that they shaped their claims with an eye on landing lucrative contracts.142 For years, market designers had made all sorts of fantastic claims for newfangled markets—They can reverse global warming! Improve access to health care! Redress racial and gender discrimination, without committing “reverse discrimination!” Even achieve “free lunch redistribution!”—so long as you hire their firms to build them to your exacting specifications (after all: “details matter”). They have almost always directed the pitch at cash-strapped governments, urging them in particular to sell off public assets to private oligopolistic concerns; in the case of toxic asset auctions one need only invert the logic.

  Unfortunately, no one could much be bothered to scrutinize the claims of market designers. After all, there was a crisis a-brewing. Only a relatively small coterie of market designers ever got invited to participate in market design exercises, and most were partners in a small set of firms with interlocking directorates. In the case of the toxic asset auctions, the job of judging the proposals was assigned to Jeremy Bulow and Paul Milgrom, both partners with Ausubel and Cramton in Market Design, Inc. So much for Chinese Walls and plausible deniability. It doesn’t verge on the wildly conspiratorial to suggest that such arrangements create some perverse incentives when it comes to reining in some of the more fantastical claims (gaining popular acceptance for them improves the firm’s prospects), a fact that has seemed only to encourage ever more extravagant claims:

  The crisis was caused by mispricing: investment bankers were able to sell poor securities for full value based on misleading ratings. This mispricing was supported by the absence of a transparent secondary market for these mortgage-related securities. If we had transparent prices, a lot of the bad things that happened would not have happened. In particular the housing bubble would have been much less, and the investment bankers would not have been able to make such clever use of the rating agencies and create tens-of-thousands of senseless securities obfuscating prices. Even a tiny bit of good market design would have averted the finan
cial crisis by preventing its root cause: the sale of subprime mortgages as near-riskless securities.143

  . . . Calls for sensible regulation and market design were met with condescension before the credit crisis, a condescension that is being reevaluated now.144

  Good auction design in complex environments . . . requires exploiting the substantial advances that we have seen in market design over the last fifteen years. The recent financial crisis is another example where the principles of market design, if effectively harnessed by regulators, could have prevented or at least mitigated the crisis.145

  Of course, there is no record of any market designers having actually successfully intervened to prevent the crisis, or helped anyone else to ameliorate it, but historical accuracy was never the name of the game. And at a 2012 meeting of the Southern Economic Association, Peter Cramton repeated his claim that market design could have solved the economic crisis; the officer in charge of NSF funding for economics was a co-panelist; her response to Cramton’s claim was to hold out market design as a perfect example of the kind of work the NSF likes to fund: referring to work done with the FCC, she gushed: “he made the government billions!”146

  To understand what elevates the public activities of market designers from the realm of mere puffery and self-promotion to the level of agnotology it is necessary to review how their public statements changed over the course of the crisis. Initially, market designers provided public defense of the Treasury plan in both its particulars (the decision to use reverse auctions) and its generalities (they signed a petition to congress in support of the proposition that government could act beneficially to correct for market failures). Especially important was the seeming independence of the academic support—as Swagel rightly noted, Wall Street economists were cheerleaders for the TARP as well, but no one would pay attention to them. Inevitably, market designers had to walk a tightrope in order to participate, at times stressing their ability to deliver competitive (low) market prices, at others, higher than fire-sale prices. But in the short term, the academics gave the Treasury the arguments it wanted for instrumental ends—to get the TARP passed. Market designers then persisted with their advocacy for auctions even after the Treasury insiders had themselves dismissed the plan as unrealistic; once the Treasury changed its plans, the market designers now devolved into free agents, turned on a dime, and attacked the Treasury. Their complaints, rendered loudly and often in the public sphere, resembled nothing so much as those being made by the opponents of the TARP. If the public was tracking the record of the economists (which it wasn’t—there was a crisis on), they would be justified in wondering: Did these market designers have any clear idea of what might have caused the crisis in the first place? Did they even have any expertise regarding the financial sector? What made this about-face so confusing is that the phantom public could not trace it to any major shift in the Treasury’s position: the market designers had registered no dissent about the Treasury’s plan to inject capital into the banks, at least so long as it was poised to use their auctions to do so. But once they were out of the running for any auction contracts, and kicked out into the cold, market designers merely flipped and adopted the rhetoric of the TARP opponents. So how much of their analytical stance can be traced to the icy slopes of logic, and how much to the fickle fiduciary considerations of their patrons?