In the financial crisis of 2007-08, Computer Business Systems (CBS) performed on a much bigger stage than any we have encountered so far. The scope and impact of the systems extended beyond the corporate and the national to the global, and the damage inflicted was correspondingly great. In the financial crisis, CBSs and their constituent technologies came together with an unprecedented malignancy. The operations of Wall Street’s mortgage machine before and during the crash, and the role of CBSs within the machine, closely fits what Joseph Schumpeter called the “mechanization of progress,” whereby innovation becomes “depersonalized and automated” and “bureau and committee work” replaces individual actions and judgments.
The Wall Street machine relied on information technologies to create a virtual assembly line on which something as simple as a single subprime mortgage at the start of the line could become by the time it reached the end a molecule within a financial derivative so complex that it was beyond the powers of the IT systems themselves to manage or keep track of. Amid these highly complex IT systems, it was easy to forget that this vast, inverted pyramid of financial manipulation pivoted on the creditworthiness of countless middle- and lower-income families obtaining mortgages for the first time for homes they could ill afford.
As with the making of Ford’s Model T, the making of a financial derivative moved through multiple stages, with each stage responsible for adding an essential component to the product. On the Wall Street line, in contrast to Ford’s, these way stations were also independent financial agencies, each exacting hefty fees and markups as the product passed through its segment of the line. The U.S. government—J. K. Galbraith’s “countervailing power”—which might have set limits on the machine’s operations, was in fact actively working on the machine’s behalf. In their book "Thirteen Bankers," Simon Johnson and James Kwak show in detail how the regulatory regime that allowed the machine to run amok was as much the work of Wall Street Democrats such as Robert Rubin and Larry Summers as it was of Reagan Republicans such as Donald Regan and Senator Phil Gramm of Texas.
As the machine worked flat-out in the run-up to the crisis, there were eight principal businesses at work along the line: the mortgage brokers who worked directly with the subprime clients; the mortgage bankers, who, benefiting from the recommendations of the brokers, underwrote the subprime mortgages, bundled them together, and passed them on to investment bankers as mortgage-backed securities (MBSs); the mortgage servicers responsible for collecting the monthly mortgage payments from the subprime clients, even as the ownership of the mortgages moved from one remote owner to another along the line; the investment bankers who bundled the MBSs with further bundles of debt—student loans, credit card debt—to form collateralized debt obligations (CDOs); the rating agencies—Moody’s, S&P, Fitch—who examined the CDOs and, with a stroke of financial alchemy worthy of Merlin the magician, transformed CDOs heavy with poorly rated MBSs into top-rated derivatives with a AAA rating; the insurers, such as AIG, who made it possible for anyone, whether they owned a CDO or not, to take out an insurance policy—a credit default swap (CDS)—against a CDO’s possible default; the CDO brokers who marketed the newly sanitized derivatives; and at the very end of the line the purchasers of the CDOs and MBSs beyond Wall Street—foundations, universities, pension funds, midwestern school districts, German regional banks, all very big losers once the MBSs and CDOs went bad.
The millions of mortgages, student loans, and other debt contracts that constituted the raw material of the machine had a dual existence both as documents in the safekeeping of legal custodians, themselves a minor component of the machine, and as electronic bits in digital space. Once the transformation from the physical to the digital had taken place, the transactions could be assembled, subassembled according to risk, and moved between way stations at high speeds.
As we have seen, speed is also achieved in mass production by automating as far as possible the cognitive functions that may have to be performed at points along the line. That is why call-center agents deal with their clients with the use of digital scripts and why HMOs are constantly pressing their physician clients to observe standard treatment protocols that, according to the HMO database, are faster and cheaper than the alternatives. Unless this happens, and if instead employees are allowed to exercise their own judgment free of rigorous time constraints, then the business process or subprocess will not achieve management’s targets for time and cost—its key performance indicators—and, from management’s perspectives, the system will have failed.
The Wall Street machine was able to achieve the speed that it did only by automating, at three critical points along the line, complex judgments about financial instruments that should have been subject to painstaking, time-consuming analysis. But once again, the rules of this automated decision making were always the outcome of executive decision making that, once embedded in the system, had to be followed by front-line employees. The Wall Street machine was therefore as much an example of digital industrialism as the call centers of the front office. The operatives of the Wall Street machine relied on software-born indexes of risk to pass favorable judgment on the derivatives as they moved along the line.
The first of the three indexes at the heart of the crisis was the FICO score, used to estimate the creditworthiness of mortgage borrowers, that could be gamed by system designers to show that a Mexican immigrant worker with an income of twenty thousand dollars could handle a subprime mortgage worth three hundred thousand dollars. The second of the indexes was the rating agencies that treated subprime borrowers as if they were small businesses and, looking at the historical record for small business failures, found that the probability of subprime default was low. Finally, the value-at-risk (VAR) indexes pioneered by Professor Philippe Jorion of the University of California were heavily relied on to assess the risk of CDOs. For its failure to allow for the unexpected and exceptional, Nassim Taleb of the Black Swan characterized this index, a decade before the debacle, as “charlatanism” and “potentially dangerous malpractice” for a “school for sitting ducks.”
* * *
We will now, in the language of CBSs, drill down and look at how integration within the mortgage machine shaped the conduct of one of Wall Street’s leading actors, Goldman Sachs. Goldman’s handling of the Wall Street crash during its critical, formative months between the middle of 2006 and the end of 2007 must be among the most heavily documented events in modern business history. Pride of place in this bibliography goes to the Senate Permanent Subcommittee on Investigations’ (the Levin Committee’s) 266-page report on Goldman, “Failing to Manage Conflicts of Interest: A Case Study of Goldman Sachs,” which is just one section within a 639-page report on the role of investment banks in the crisis. The report draws on the tens of thousands of e-mails subpoenaed from Wall Street firms by the committee and provides a day-to-day, and sometimes hour-by-hour, account of what went on at Goldman during those months. A companion volume to the report is the transcript and video footage of the hearings before the Levin Committee, which took place on October 27, 2010, when Goldman’s crisis team, from CEO Lloyd Blankfein down to the humblest traders, gave their side of the story.
In addition, there is the growing list of lawsuits against Goldman, all with their accompanying texts, brought by government agencies such as the Securities and Exchange Commission (SEC) and the Federal Housing Authority, and also by aggrieved clients of Goldman such as the now defunct Australian hedge fund Basis Capital and by ACAS Capital, which collaborated with Goldman in the creation of the Abacus CDO, notorious for the role of the hedge-fund manager John Paulson in selecting the securities to be included in the CDO. In July 2010, the SEC fined Goldman $550 million to settle charges that it “failed to disclose to investors vital information about the Abacus CDO . . . particularly the role that Paulson and Co. played in the portfolio selection process.” In other court actions against Goldman the actions are still pending and Goldman’s level of culpability has yet to be decided. But the Levin Committee’s report and hearings provide, I believe, strong evidence that the deception and manipulation of clients eventually became an integral part of Goldman’s trading strategy.
In trawling through the documents, it is essential never to lose sight of the role of the Goldman corporate hierarchy in the crisis and the highly disciplined way in which it managed the company. The answer to the old Watergate question “What did the president know, and when did he know it?” is, in the case of Goldman’s big three—Lloyd Blankfein, CEO; Gary Cohn and Jon Winkelried, co-presidents—that they knew everything that mattered (and in real time). The big three’s point man for the crisis, Goldman’s own H. R. Haldeman, was co-president Gary Cohn, whose name turns up frequently in the e-mail flow of his subordinates. Although the e-mail vocabulary between Blankfein, Cohn, and their subordinates has a certain locker-room familiarity about it, there is never any doubt about who is in charge.
Blankfein, Cohn, and their team were men of the machine and, as it turned out, among the most skillful manipulators of the machine on Wall Street. By the early 2000s, Goldman’s derivatives trading could no longer be called banking in any meaningful sense of the term, but had become an industrial activity, turning out virtual products whose fortunes depended on the efficient management and coordination of processes: the accumulation of mortgages and other forms of debt from bankers and brokers, their transformation into financial derivatives, and their selling on to clients.
In Goldman’s culture these processes were of supreme value because they were vehicles for the creation of Goldman’s earnings and profits, and so critical for the health of the Goldman stock price, the size of the salaries and bonuses paid out to Goldman executives, and the reach of Goldman’s power and fame on Wall Street and beyond. As long as house prices continued to rise and those along the process chain continued to make money, the model’s flaws could be ignored, notably the dismal quality of the subprime mortgages upon which the whole system depended.
But once the housing market turned, the system collapsed. The difference between Goldman and the other leading players on Wall Street was that Goldman saw it coming and was able to recalibrate its machine so that not only did it avoid the catastrophic losses that destroyed Lehman Brothers and crippled Citicorp, but it actually came out ahead. But to achieve this Goldman behaved, I will argue, with ruthless cynicism, above all in deceiving and exploiting its clients. Why did Goldman do this? The simple answer is that for Goldman, wealth creation on its own behalf took priority over everything else, and nothing was going to stand in its way. In histories that tell of how Goldman acted on its view of the deteriorating markets and came out ahead, the word warehouse often appears, a choice word that is revealing both as a pointer to the heavily industrial character of Goldman’s trading activities and as providing a mental diagram to locate the various financial instruments Goldman was dealing in. But warehouse does not fully capture the reality of what was going on at Goldman.
A real warehouse is a place where finished goods are stored before they are shipped off to customers or retailers, whereas Goldman’s “warehouse” was much more like a factory where industrial processing of financial instruments took place on a virtual assembly line. The Goldman “factory” was an electronic space where the “raw materials” of loans coming in from such mortgage brokers as New Century, Long Beach, and Countrywide were processed on the virtual line into financial instruments such as mortgage-backed securities, collateralized debt obligations, and credit default swaps, which then were marketed to clients. In the precrisis world, where it was assumed that house prices would go on rising indefinitely, the processes of “securitization,” that is, the processing of the raw loans coming in from the brokers, were relatively straightforward.
In the case of mortgage-backed securities, the incoming loans were bundled together by Goldman, their ownership vested in a trust, with the trust issuing securities to investors, which gave them the right to the cash flows generated by the loans as householders paid off their mortgages. With collateralized debt obligations, several MBSs would be bundled together, along with bundles of student, consumer, or corporate loans, again with ownership of the loans vested in a trust, with the trust issuing securities to investors. In addition, there was a whole superstructure of insurance, known as credit default swaps, attached to the MBSs and CDOs. The owners of the MBS and CDO securities, or indeed anybody, could take out an insurance policy against their loss of value and receive compensation if this happened. Equally, the owners of the securities, if they were confident that they would hold their value, could be the providers of insurance and receive regular premium payments from the policy holders. If the securities lost value, then, as with any insurance policy, the issuer of the insurance was obliged to compensate the policy holder for their loss.
Whether Goldman was pursuing a coherent trading strategy during these crisis months is a fraught issue, because it is claimed by those suing Goldman that it had a consistently pessimistic view of the market, and although this pessimism shaped its trading on its own behalf, Goldman hid this pessimism from many of its clients, fed them an upbeat view of the market that it did not believe in, and persuaded them to buy assets that it knew were flawed and would lose value, and indeed did. Goldman’s defense all along has been that it had no aggressive moneymaking strategy at all and was simply the prudent guardian of its clients and its own interests. In its own words, “The risk management of the firm’s exposures and the activities of our clients dictated the firm’s action, not any view of what might or might not happen to any security or market.”
This was also the line taken by Lloyd Blankfein himself during his appearance before the Levin Committee on October 27, 2010, where he spoke of Goldman as virtually a charitable organization, the passive counterparty in deals where strong-minded clients came in and told Goldman exactly what they wanted, and Goldman respectfully executed their instructions. So Blankfein: “The customers who are coming to us for risk in the housing market wanted to have a security that gave them exposure to the housing market, and that is what they get. . . . [T]he security itself delivered the specific exposure that the client wanted to have.” And again: “What clients are buying, or customers are buying, is they are buying an exposure. The thing that we are selling to them is supposed to give them the risk they want. They are not coming to us to represent what our views are. . . . [T]he institutional clients we have wouldn’t care what our views are. They shouldn’t care.”
At the Senate hearings the Goldman team from Blankfein on downward was much helped in its own defense by the extreme complexity and variety of the derivatives Goldman traded during the crisis period and the difficulty for laymen, including those on the Levin Committee, of distinguishing between them, and especially the differing legal obligations attached to each of them. It was here that the Levin Committee hearings fell short in ways that undermined the impact of the report and the hearings in the public policy debate. In his preamble to the hearings, Senator Carl Levin (D-Mich.) drew attention to Goldman’s differing obligations to its clients as market maker and underwriter, but Levin and his fellow senators lost sight of this distinction when questioning Blankfein and his colleagues, allowing them to slip away again and again behind a fog of obfuscation.
One way of cutting through this obfuscation is to imagine for a moment that Goldman was a real industrial company making and selling real products, rather than a virtual industrial company making and selling virtual products. Imagine Goldman for a moment as the Goldman Motor Manufacturing Company, or GMMC, a Detroit competitor of Ford in the early days of mass production in the 1920s. One day GMMC discovers to its horror that there is a serious flaw in its manufacturing processes so that a significant percentage of the engines installed in its best-selling model, the Model G, break down after just a few weeks on the road. The vice president for manufacturing tells the CEO that GMMC must immediately close its plant for retooling, recall the products it has already sold, and strip down the models it has in stock, selling off the uncontaminated engine parts to local component dealers.
But the vice president for finance quickly does his sums and persuades his colleagues that the cost of this plan A is too great, and they must decide instead on plan B. With plan B, GMMC instructs its salesmen to avoid its local Detroit dealers, whom the company suspects have gotten wind of problems at the plant, and tells the GMMC salesmen instead to ship the problem Model Gs to the South and sell them directly to unsuspecting farmers in rural Kentucky and Tennessee. Without telling these customers that there is anything wrong with the Model Gs, GMMC quietly takes out an insurance policy with a local Detroit company that will pay out to GMMC every time one of its cars goes wrong. When the owners of the stricken vehicles demand a refund, GMMC refuses. This is a simplified but essentially accurate account of what Goldman frequently did in its derivatives trading. Looked at day- to-day, Goldman’s trading strategies were complex, sometimes counterintuitive, and lacking in obvious direction. The GMMC fable can be a helpful guide as we try to make sense of what Goldman was doing.
* * *
Drawing on its vast e-mail trove, the Levin Committee report shows Blankfein’s image of Goldman as a charitable organization to be entirely fictitious and repeatedly quotes chapter and verse to prove it. The report shows that from late 2006 onward, Goldman’s senior executives had a consistently pessimistic view of the housing market and the financial instruments attached to it and thenceforth pursued an aggressive trading strategy to maximize its gains from the crisis, with the manipulation of clients becoming, I will argue, an integral part of Goldman’s chosen strategy. The evidence of how Goldman’s top executives really viewed the markets is therefore germane to this whole history, and some of it now follows.
In an internal “self-review” dated September 26, 2007, Michael J. Swenson, head of the Goldman Sachs Mortgage Department’s Structured Products Group, wrote that “during the early summer of 2006 it was clear that the market fundamentals of subprime and the highly leveraged nature of Collateralized Debt Obligations were going to have a very unhappy ending.” On December 7, 2006, Daniel Sparks, head of Goldman’s Mortgage Department and a key link between the trading floor and the Goldman big three, exchanged e-mails with Thomas Montag, a senior Goldman executive and co-head of Global Securities for the Americas, “about why Goldman was not doing more to reduce the firm’s risk associated with its net long positions” in housing-related assets.
On December 14, 2006, David Viniar, chief financial officer and therefore number four in the Goldman hierarchy after the big three, convened a meeting in the conference room next to his office on the thirtieth floor (the seat of power where the big three also had their offices), where they conducted an in-depth review of the Mortgage Department’s holdings because its “position in subprime mortgage related assets was too long, and its risk exposure was too great.”
The next day Viniar e-mailed Montag about the deteriorating markets and the opportunities it opened up: “My basic message was lets [sic] be aggressive distributing things because there will be very good opportunities as the markets [go] into what is likely to be even greater distress and we want to be in a position to take advantage of them.”
Then on February 11, 2007, and from the thirtieth floor itself, Blankfein urged the Mortgage Department to get on with the task of selling off its deteriorating assets: “Could/should we have cleaned up these books before and are we doing enough right now to sell off cats and dogs in other books throughout the divisions?” On February 14, 2007, Daniel Sparks reported further on the deteriorating markets and the trading opportunities it opened up: “Subprime environment—bad; and getting worse. Everyday [sic] is a major fight for some aspect of the business. Credit issues are worsening on deals and pain is broad. . . . [D]istressed opportunities will be real, but we aren’t close to that time yet.”
In 2006 and 2007 Goldman originated twenty-seven CDOs and ninety-three MBSs with a total value of about $100 billion. The problem for Goldman from the summer of 2006 onward was that its “factory” was clogged with components of MBSs and CDOs in varying stages of manufacture—raw loans just in from the brokers and not yet bundled, loans in the process of being bundled, bundles of MBSs not yet put together as CDOs, finished CDOs and MBSs not yet marketed, and securities of old CDOs and MBSs that Goldman had not yet sold and were still in the factory. Goldman had acquired and processed these assets on the assumption that the housing market was strong and that there would be an equally strong demand for the finished CDOs and MBSs.13 But now the market was about to collapse, and Goldman had billions’ worth of what it believed would become failing assets on its hands. So what to do with them?
Much of what Goldman then did was what any owner of a big stock portfolio would do if faced with a collapsing market. Goldman stopped taking in any more loans from the mortgage dealers; it abandoned some CDOs and MBSs that were still “under construction” and liquidated others that were fully formed, selling off their components in the markets, as it also did with some of the “raw” loans recently acquired from the brokers that had yet to enter the securitization process. If this is all that Goldman had done, there would be no Goldman story, Blankfein would be esteemed on Wall Street as the great survivor, and Goldman would not be the target of multiple lawsuits that could still cost it hundreds of millions, if not billions, of dollars.
But what Goldman also did, and this has been the source of its troubles, was to persist with the creation of new CDOs and MBSs and to continue with the marketing of existing ones, even bringing some of its own proprietary assets into the factory so that they too could become part of a new CDO. Goldman also began taking the insurance structure pivoting on the CDOs, the credit default swaps, much more seriously as a potential source of revenue. Although the trading strategies involved in these activities were sometimes complex, the motive underlying them was simple and straightforward. Goldman believed that it could make more money by disposing of its factory assets as components of CDOs, MBSs, and CDSs than by just selling them off unadorned in the market.
The problem that then arose for Goldman was that in marketing the three kinds of financial derivatives, the company was acting as underwriter and placement agent and not simply as market maker or trader on its own behalf. As underwriter and placement agent, Goldman was subject to rules on fair disclosure that as market maker it was not. In the Levin Committee hearings, the Goldman team, from Blankfein on down, went to very considerable lengths to blur the distinction between the different roles and to cast themselves, whenever possible, as humble market makers. There can be little doubt that in preparing for the hearings, a high-risk event for Blankfein et al., Goldman’s extremely high-priced lawyers got together with their clients and advised them that, in view of their record, obfuscation on the distinction between underwriter and market maker was advisable.
This evasion was much in evidence at the hearings and especially in the gladiatorial contest between Senator Levin and Blankfein on the final day, when Levin searched with increasing frustration for the smoking gun that would sink Blankfein but never quite managed to find it. In reading the committee transcript, and simultaneously watching the contest on a podcast available at the Levin Committee website, I was reminded not so much of Watergate and the search for the Nixonian smoking gun as a remarkable scene that appears in several versions on YouTube.
In it a gigantic Alaskan brown bear sits on a slab of rock overlooking a fast-moving river full of salmon. The bear is trying to scoop up one of the salmon with his paws, but most of the time the salmon are much too quick for him and he fails. But just occasionally, he succeeds and has a terrific meal. There was something bearlike about Senator Levin peering down at Blankfein from his senatorial perch as he flailed around, trying to land his quarry, but with the slippery Blankfein swimming clear every time. The blurring of the distinction between market maker and underwriter was central to Blankfein’s evasive strategy, as the following exchanges reveal:
Senator Levin: Is there not a conflict when you sell something to somebody and then are determined to bet against that same security and you didn’t disclose that to the person you are selling it to? Do you see a problem?
Lloyd Blankfein: In the context of market making, that is not a conflict. What clients are buying, or customers are buying, is they are buying an exposure. The thing that we are selling to them is supposed to give them the risk they want.
Senator Levin wasn’t satisfied:
Senator Levin: How about you are investing in these securities. This isn’t a market making deal. This is where you have a decision to bet against, to take the short side of a security that you are selling, and you don’t think there is any moral obligation here?
Lloyd Blankfein: Every transaction Senator, and this is—and I think it is important and again, I am not trying to be resistant but to make sure your terminology—when as a market maker, we are buying from sellers and selling to buyers . . .
Levin cuts him off but later returns to the attack:
Senator Levin (with increasing frustration): You are betting against that same security you are out selling. I have just got to keep repeating this. I am not talking about generally in the market. I am saying you have got a short bet against that security. You don’t think the client would care?
Lloyd Blankfein: I don’t, Senator. I can’t speak to what people would care. I would say that the obligations of a market maker are to make sure your clients are suitable and to make sure they understand it. But we are a part of a market process. We do hundreds of thousands, if not millions of transactions as a market maker.
Again the Uriah Heep side of Blankfein as he relegates Goldman to being just another humble market maker along with all the others on the trading floor at the New York Stock Exchange. This was a clever strategy, but also high risk. It was clever because anyone with a serious interest in the stock market, which presumably included most of the members of the Levin Committee, knew what market makers do. They hold a supply of a stock for those who wish to buy or sell it. They adjust the price with shifts in supply and demand, and they are obliged to buy and sell at prices broadly in line with the rest of the market. It is not their role to advise investors about the wisdom of buying a stock, and they are not at fault if a stock loses 20 percent of its value within an hour of its purchase.
With his answers to Senator Levin, Blankfein was trying hard to cloak Goldman and himself in the passive neutrality of the market maker. But this was also high risk because it would have taken only one senator with forensic lawyerly skills to rip through this defense and reach—at last—the smoking gun. With the marketing of CDOs and MBSs to its clients, Goldman was not a market maker but an underwriter or placement agent and was therefore subject to rules of disclosure about the suitability of its product for the investor that it had manifestly violated. To grasp the sheer chutzpah of Goldman’s marketing, one needs to look at one of these deals in detail. Among the most revealing was Goldman’s marketing of the Timberwolf CDO between September 2006 and June 2007, the very months when its own view of the housing market soured and when it began to dispose of its own flawed assets.
* * *
The role of the investment banks in the financial crisis of 2007– 2008, including the role of Goldman Sachs examined here, is the most disturbing recent example of the economic and the unethical coming together with catastrophic consequences for the former. The trading teams at Goldman and the executives who presided over them were inhabiting an extreme variant of a closed world, where their reality was populated by the electronic images and databases of their systems. Their obsessive focus on these images, driven by an equal obsession for the bottom line, excluded the human realities tethered to the symbols.
So they cast a blind eye to the interests of the householders being pumped for subprime mortgages; the unscrupulousness of the mortgage brokers lining up the subprime borrowers; the duplicity of the ratings agencies, handsomely rewarded, in transforming mortgage dross to AAA gold; and the interests of the bankers’ clients who, at the bankers’ persuasion and deceived by them, were the end purchasers of the flawed derivatives. The cumulative impact of this fusion of technology, greed, and moral blindness, duplicated from one end of Wall Street to the other, was global economic meltdown.
So far only a single Goldman employee, the hapless Frenchman Fabrice Tourre, has been held to account for what Goldman did during those years. In a lawsuit brought by the SEC, Tourre was convicted of civil fraud by a New York City jury in August 2013. But as a midlevel processor and salesman of derivatives, Tourre was a very small cog in the Goldman machine, a conveyor belt for orders coming down from above. When will the real culprits be held to account?
Excerpted with permission from “Mindless: Why Smarter Machines Are Making Dumber Humans” by Simon Head. Available from Basic Books, a member of the Perseus Books Group. Copyright © 2014. All rights reserved.