The results of these and other experiments permit certain observations:
Capital markets are complex dynamic systems.
Complex dynamics exhibit memory or feedback, called path dependence.
Risk in capital markets is an exponential function of scale.
Small changes in initial system conditions produce divergent results.
System output can be orderly or chaotic.
These observations are the scientific basis for what is popularly known as a black swan event. The term “black swan” is used widely to describe any surprising headline even by those who lack a theoretical understanding of the underlying dynamics. Black swan discussion tends to trivialize science with a fatalistic tinge, as if to say “stuff happens.” Stuff doesn’t just happen. Crises emerge because regulators don’t comprehend the statistical properties of the systems they regulate.
LTCM was a textbook case in ignoring complexity theory. For example, traders at LTCM frequently constructed two-sided strategies using real government notes and synthetic notes in swap form. If the spread between the two trades converged as expected, a profit was realized and the trade was unwound. The conventional way to unwind a trade is to sell the note and terminate the swap by negotiation with the counterparty.
Counterparties charge a small fee for early swap termination. LTCM did not want to pay the fees. Instead, LTCM neutralized the first swap by entering into another swap with exactly opposite terms. The fixed/floating payments, margin payments, and other obligations of the two swaps were exactly offsetting, so cash flows and market risk netted to zero. Economically, this was the same as if the first swap were canceled, except LTCM did not incur a termination fee. Value at risk in the twin trades was zero according to regulators. LTCM called this technique the “wedding cake” because traders kept adding new layers to neutralize swap positions instead of terminating positions. The layers stacked up to make a $1 trillion wedding cake.
A complexity theorist looks at the same trade termination technique and sees gross notional value doubled, because there were now two swaps instead of one. This meant the risk more than doubled, because risk is an exponential function of scale. While partners were at the golf course or racetrack thinking risk was under control, LTCM was a ticking time bomb. In August 1998, the bomb exploded.
Aftermath
Complexity theory is not understood by regulators today, so perhaps LTCM partners can be forgiven for not understanding complexity in 1998. Yet once that collapse occurred, it might be expected that thought leaders in finance, like Alan Greenspan, Bob Rubin, and Larry Summers, would have learned lessons and tried to avoid a similar collapse in the future. They did just the opposite.
In August 1998, as the LTCM debacle unfolded, Alan Greenspan was Federal Reserve chairman, Bob Rubin was treasury secretary, and Larry Summers was Rubin’s deputy, soon to be treasury secretary himself. In February 1999, just months after the LTCM catastrophe, Time magazine featured Greenspan, Rubin, and Summers on its cover with the headline “THE COMMITTEE TO SAVE THE WORLD.”
Far from learning the lessons of 1998, the three did everything possible to make the system riskier and more unstable. “The Committee to Destroy the World” would have been a more apt headline. In 2008, their combined efforts came close to doing just that.
What were the lessons of 1998? The first was that off-balance-sheet derivatives are dangerous because they are opaque. In a crisis, counterparties cannot find each other to net transactions due to this nontransparency. The second lesson was that leverage converts minute market moves into massive losses that threaten solvency. The third lesson was that banks should be kept out of derivatives businesses. Hedge funds and speculators can roll the dice, but banking is a special business based on bondholder and depositor trust. That trust should not be squandered with swaps speculation. The fourth lesson was that derivatives risk is in the gross value, not the net. When LTCM teetered on the brink of failure, banks were not worried about net exposure on their books, they were worried about their gross position with LTCM that would have to be replaced when LTCM failed. The last, and most important, lesson was that crises emerge from nowhere without warning. A 1997 currency crisis in Thailand is not immediately or obviously linked to a Greenwich hedge fund collapse the next year. Still, that is exactly what happened.
Based on these lessons, policy choices were obvious. Derivatives should be confined to exchanges where they are transparent and margined. Leverage should be limited, and on balance sheet where it can be observed. Banks should be banned from derivatives except for bona fide hedging. Value at risk should be scrapped as a metric due to obsolescence and statistical defects. Finally, capital requirements should be increased as a cushion against emergent crises that should be expected, but never are.
Greenspan, Rubin, and Summers pursued the exact opposite of those five policies. It is convenient to claim ignorance on their part. Yet they were explicitly warned at the time by the one regulator who understood what happened at LTCM. Her name is Brooksley Born, then chair of the Commodity Futures Trading Commission (CFTC).
In 1999, Born was a member of the President’s Working Group, PWG, formed after the 1987 stock market crash. The PWG was a recommendation of the Brady Commission. Ironically, the principal author of the Brady Commission Report was David W. Mullins Jr., a Harvard professor at the time, later vice chairman of the Federal Reserve, and still later a partner at LTCM.
The PWG consisted of the president of the United States, Federal Reserve chairman, treasury secretary, SEC chairman, and the chairman of the Commodity Futures Trading Commission. PWG’s purpose was to put banking, securities, and commodities regulators in one place to deal with crises. The 1987 crash involved complex interactions between the stock market, regulated by the SEC, and Chicago futures markets, regulated by the CFTC. The crisis then spread to the bank payments system because billion-dollar margin calls were required between the New York and Chicago markets. Banks were hesitant to initiate wire transfers to Chicago futures brokers for fear that they would not receive incoming wires from New York stockbrokers. The system started to freeze up. Noncoordination among securities, futures, and bank regulators made crisis resolution difficult. The PWG was intended to prevent those problems in the future.
In 1998, the PWG included Bill Clinton, Bob Rubin, Alan Greenspan, Arthur Levitt, and Brooksley Born. Contagion caused by LTCM arose from swaps, which were regulated by Born; the SEC had almost no jurisdiction over swaps at the time. The policy response to LTCM fell to Rubin, Greenspan, and Born, with Larry Summers as Rubin’s right hand.
Greenspan did not understand derivatives risk then and shows little understanding today. He held the view that swaps reduce risk. The analysis begins with an observation, largely correct, that traditional securities and banking transactions represent bundles of disparate risks. A single bank loan can be thought of as a bundle of risks including interest rate risk, credit risk, foreign exchange risk, liquidity risk, operations risk, sovereign risk, and other distinct risks bound together in the loan. Greenspan thought that derivatives allowed these risks to be unbundled. Using swaps, a lender could separate credit risk from foreign exchange risk, and transfer each distinct risk type to a specialized counterparty in the best position to bear that risk. The risk migrated from the bundled product to strong hands best able to bear specific subrisks. This made the system stronger and more resilient. Greenspan was correct in this analysis as far as it went.
What Greenspan missed was that there was no limit on the notional value of derivatives that could be created from a single unit of original risk. If a $1 billion loan was broken into ten subrisks via swaps, and traded separately so total notional value added up to $1 billion, that would vindicate Greenspan’s view. In fact, dealers create $10 billion or more of swaps from $1 billion of underlying securities. Dealers create swaps that have no underlying instrument at all, just an index or formula for p
aying off bets with no connection to real-world debt. Swaps created from thin air add to the gross notional amount of risk and increase systemic scale and complexity. Greenspan’s quaint notion of putting a risk quantum in strong hands was dominated by the fact that the quantum was infinitely elastic.
Swaps are economically identical to exchange-traded futures. Both are bets in which each party owes more or less to the other depending on future market levels. The main difference is that futures are traded on exchanges, and swaps are traded in private over-the-counter deals. Born, the futures regulator, understood swap risk and wanted to move swaps to futures exchanges where they are margined properly and traded transparently.
Greenspan, Rubin, and Summers treated Born as if she were a throwback ignorant of financial technology. Archaeologists have discovered Sumerian clay tablets from 4500 BC that record transactions in commodity futures such as forward delivery of livestock. Aristotle discussed how options were used for market manipulation in the fourth century BC. Futures markets Born regulated had changed little since 1848, when grain started trading on the Chicago Board of Trade. Born’s beat seemed stuck in the past.
The modern swaps markets was innovated by Meriwether, Scholes, and others in the 1980s. Swaps grew in scope and sophistication as the twenty-first century drew near. Greenspan, Rubin, and Summers saw themselves as firmly on the right side of history in treating swaps differently from commodity futures.
Misogyny played a role. Greenspan, Rubin, and Summers were a powerful boys’ club out to squash the voice of the one woman regulator. In 2005, Summers, while Harvard president, showed the same bias in notorious remarks about women’s incapacity in quantitative science.
Distressingly, misogyny emerged again in 2008 when the most powerful woman regulator, Sheila Bair, chairwoman of the FDIC, was sidelined by a new boys’ club of Summers, Tim Geithner, and Ben Bernanke. Bair’s dead-on advice to close insolvent banks was scientific. Instead the boys’ club bailed out bank cronies at taxpayer expense. Born in 1998, and Bair in 2008, both correctly analyzed response functions to financial crises. Summers demolished them with bias and bad advice both times.
Born’s recommendations to PWG were limited to derivatives; unlike Bair, she was not a bank regulator. Her advice was to continue restrictions on new swap types and move most existing swaps to futures exchanges. Not only was Born’s advice marginalized, the boys’ club did the opposite.
In 1999, the sixty-six-year-old Glass-Steagall Act was repealed. Glass-Steagall, enacted in the Great Depression, separated the banking business from securities underwriting. One cause of the Great Depression was that banks in the 1920s originated weak loans and sold these as securities to unsuspecting retail investors. In 1933, Congress passed Glass-Steagall, which said that banks could either accept deposits and make loans, or underwrite and sell securities, but they could not do both. This was due to conflicts that led to bad loans dumped on customers in securitized form. Bankers promptly separated themselves into commercial banks that took deposits and made loans, and investment banks that underwrote and sold securities.
Separation worked well for sixty-six years and saved the United States from major banking crises. Individual banks such as Continental Illinois in 1984 might fail, and there were still conflicts and loan losses such as the 1980s savings and loan crisis. Still, after Glass-Steagall there was no general banking crisis of the kind seen from 1929 to 1933.
Glass-Steagall worked for exactly the reason complexity theory suggests. By breaking the banking system into two parts, Glass-Steagall made each part stronger by shrinking systemic scale, diminishing dense connections, and truncating channels through which failure of one institution jeopardizes all. It was exactly like building watertight compartments in a ship’s hold. One compartment could flood, yet the entire vessel does not sink.
Glass-Steagall was repealed by an unholy alliance of Republicans and Democrats led by Senator Phil Gramm and President Bill Clinton. Reasons for repeal were not directly related to the LTCM crisis—they had been developing for years. Ratification of the otherwise illegal merger of Travelers and Citicorp, promoted by Sandy Weill, was the driving force for elimination of Glass-Steagall. Ironically, it was Weill’s order to terminate spread trades at Salomon to facilitate the merger that catalyzed LTCM’s demise; another example of dense networks in deep operation.
Repeal worked to increase risk in the financial system. By allowing new combinations of financial institutions, Glass-Steagall repeal increased systemic scale past the point where one big bank failure crashes the entire system. Repeal also allowed commercial banks to mimic the proprietary securities trading activity of investment banks, and the derivatives that went along with it.
On December 21, 2000, just a few weeks before he left office, Bill Clinton signed another law less well known than Glass-Steagall repeal, yet more insidious in spreading systemic risk. The Commodity Futures Modernization Act of 2000 repealed prohibitions on certain swap types, and allowed these swaps to be hidden off-exchange, and off-balance-sheet. Born was pushed out of the CFTC on June 1, 1999, replaced by a new chairman, Bill Rainer, a friend of Bill Clinton’s from Arkansas, handpicked to pursue derivatives deregulation. Passage of the commodities act was a bipartisan gift to big banks pushed by Republican Phil Gramm, and Democrats Rubin and Summers.
Prior to 2000, swaps were limited to returns on stocks, bonds, interest rates, and currencies. Bets on commodities such as oil, metals, and grains were traded on regulated futures exchanges. Other commodities did not trade on futures markets, and were not permitted in swaps markets, and therefore did not trade in derivative form at all. With the 2000 repeal of swaps regulation, the door was open to completely unregulated trading in all derivatives. Companies like Enron quickly created markets in off-exchange electricity futures, and collapsed in multibillion-dollar frauds courtesy of Gramm, Rubin, and Summers.
The combined impact of Glass-Steagall repeal and swaps regulation repeal was a financial witches’ brew. Repeal meant that banks could trade like hedge funds in an unlimited range of instruments. Still, one more ingredient was needed to complete the mix—leverage.
On November 17, 2003, Rubin protégé Tim Geithner was installed as head of the Federal Reserve Bank of New York. Geithner assiduously turned a blind eye as banks piled on risks. Commercial banks regulated by Geithner were allowed to own investment banks. Still, the investment bank entities were individually subject to more stringent SEC capital requirements. Getting the SEC to roll over on leverage was the banker’s next mission.
The Republican Bush administration replaced the Democratic Clinton administration in 2001. Party affiliation does not matter when it comes to Washington’s desire to deliver on bankers’ wish lists. Bankers own Washington. Pressure to ease up on broker-dealer capital requirements was coming not just from banks, but from brokers such as Bear Stearns and Lehman not owned by banks. They wanted a level playing field so they could compete with banks in the securities business.
Simultaneously, banks wanted their own looser capital requirements. These were set by the Basel Committee of the Bank for International Settlements (BIS) in Switzerland. The first global bank capital rules were issued in 1988 under the name Basel I. Within a few years, these rules seemed too restrictive. Banks began pushing for new formulas to allow them to carry more risk on a smaller capital base. They justified this by promoting defective VaR models as a partial substitute for strict capital ratios. The resulting revisions to bank capital requirements implemented in stages between 2001 and 2004 were called Basel II. These rules allowed greater bank leverage based partly on VaR models that showed extreme leverage was safe.
It was against this background of easier bank capital requirements that the SEC moved to revise broker-dealer capital rules in 2003 and 2004. In 2003, the SEC expanded the definition of eligible collateral for leverage purposes to include certain mortgage-backed securities. In 2004, SEC expanded its oversight to
include broker-dealer holding companies. This new comprehensive holding company oversight borrowed concepts from Basel II, including risk weightings based on security type. In particular, the SEC permitted higher leverage for certain mortgage-backed securities versus traditional stocks and bonds. The result of these two changes was that mortgage-backed securities were eligible collateral for leverage and minimal capital was required to support that leverage.
In 1998, LTCM collapsed under a critical mass of derivatives leverage. In 1999, Glass-Steagall was repealed. In 2000, derivatives regulation was repealed. In 2001, bank capital requirements were eased. In 2003 and 2004, broker-dealer capital requirements were eased. Throughout this entire period, the Fed kept interest rates artificially low. It was as if global regulators reacted to LTCM by working together to repeat that debacle on a larger, more dangerous scale. That’s exactly what happened. In 2008, the entire risky, leveraged, interconnected house of cards collapsed.
CHAPTER 5
FORESHOCK: 2008
A financial market is riddled with feedback. . . . Such intrinsic feedback does not arise when gambling. . . . Likewise even if everyone had the perfect prediction model of the weather, the weather would still do what the weather does. All that would happen is that everyone would know exactly what to wear the next day. However, this is not true in the markets. If everyone were to be given the perfect prediction model, it would immediately stop being the perfect prediction model because of this strong feedback effect.
Neil F. Johnson, Ph.D., professor of physics, University of Miami
A New Crisis
From the perspective of complexity theory, the 2008 collapse was easily foreseen. A dynamically identical collapse happened in 1998. The scope of the 2008 panic was greater than in the 1998 panic. That increased scope was to be expected because systemic scale had increased in the intervening ten years. Our provisional law states: Derivatives risk increases exponentially as a function of scale measured by gross notional value. Excessive leverage, nontransparency, and a densely interconnected bank network were common factors in both crises. The difference in catalysts—sovereign swap spreads in 1998, subprime mortgages in 2008—is irrelevant. What matters is the deep structure of financial risk. Differentiating causation into swaps and subprime is like chasing snowflakes while ignoring avalanche danger. Snowflakes don’t kill you; avalanches do.
The Road to Ruin Page 17
