by Adam Tooze
From its origins in the 1930s, the GSE model of subsidy separated the origination of a mortgage from its ultimate funding. The commercial banks that issued the original loans to American families were repaid when they sold the mortgages to Fannie Mae. That enabled them to make more loans. It was the debt issued by the GSEs to finance the mortgages they were holding on their balance sheets that ultimately funded the loan. This was the basic structure of what became known as “originate to distribute.” Mortgage lenders no longer needed to hold the mortgages on their balance sheets; they became brokers operating for a fee. The government-backed credit rating of the GSEs backstopped the entire system.
Starting in the 1970s, as they confronted the instability of interest rates and its damaging implications for America’s mortgage model, the GSEs took a further critical step. Working with the help of investment banks, they pioneered securitization.19 Rather than holding the locally originated mortgages on their own books and financing them by issuing bonds, they would sell the mortgages directly to investors. To do so they packaged the mortgages into pools, in which they sold shares—securities. The idiosyncratic risks of individual loans would be pooled. Investors looking to hold real estate could buy a portfolio of broadly based exposure without having to build a branch network necessary to make loans across the far-flung economy of the United States. They did so fully aware of the risks and returns generated by the fluctuation of interest rates. Rather than having small savings and loans gamble on what was a viable loan, securitization would let the collective process of market haggling determine funding costs.
In 1970 Ginnie Mae carried out the first securitization. It was a simple model—a so-called pass-through—under which the flows of revenue from a pool of mortgages were passed by way of the GSE to investors. Not satisfied that this should remain a public monopoly, Lewis Ranieri and his team at hard-driving investment bank Salomon Brothers put together the first private securitization of mortgages for Bank of America in 1977.20 But it took a brave investor to buy a package of fixed interest mortgages at that moment. It was in the aftermath of the interest rate shock of the early 1980s that securitization came to the fore. The mortgage lender stranded with portfolios of low-interest mortgages turned to the market to recover whatever value they would yield by securitizing them and selling them off. From the 1980s, the GSEs, working with the investment banks, created not just pass-through mortgage-backed securities (MBS) but so-called collateralized mortgage obligations (CMO) that allowed a pool of MBS to be tranched into separate risk tiers profiles. This was the origin of so-called structured finance. Those with the top-tier first claim on the revenue stream had low risk of both default and early repayment. Tranches lower down the pecking order could be sold off to investors looking for riskier investments. The top-tier senior tranches would pay out except in the highly unlikely event of massive, collective default. Those top tranches, even if they were based on a pool of high-yielding, high-risk debt, could be designated as low risk, and the ratings agencies obliged by classifying them as AAA (80 percent of most securitizations were designated as senior and given the AAA rating).
Not surprisingly, given the very high ratings they handed out for MBS, the role of the ratings agencies would later become highly controversial. By the 1990s, Moody’s Investors Service and Standard & Poor’s divided 80 percent of the global debt-rating business between them.21 Fitch took another 15 percent of the market. They did not attain that control of the global market by freely handing out top AAA ratings. In 2008 there were only six AAA-rated corporations and no more than a dozen countries enjoying that ranking. This was despite the fact that since the 1980s it was issuers of debt who paid the ratings agencies to make their classifications, not the subscribers to their information services. Payment by the issuer created a conflict of interest. But the agencies had much to lose if they appeared to be selling good ratings and relatively little to gain if they showed favor to a client that was involved in only occasional bond issues. The mortgage securitization business changed that calculus. The sheer volume of mortgage-backed security issuance, involving tens of thousands of tranches, combined with the fact that the flow was concentrated in the hands of a few issuers, gave the ratings agencies a significant incentive to be “helpful.”22 But even more important was the nature of MBS as such. What made rating MBS different was that the underlying assets were not bonds issued by a company facing the unpredictable force of global competition. The MBS bundled thousands of what were supposed to be predictable assets—regular domestic mortgages. The ratings agencies did not have to calculate the risks of default on the basis of more or less subjective evaluations of a company’s business prospects. Nor did they have to render a judgment on a country’s fiscal policy. Instead, they could apply standardized financial mathematics to a population of mortgages that was assumed to have well-known statistical properties. If you knew default rates and could make assumptions about the degree of correlation between them, once you assembled enough mortgages and tranched them, the likelihood of the top tranches not paying was infinitesimal. Tens of thousands of asset-backed securities thus qualified for ultrasafe AAA ratings. What happened to the tranches lower down the pile was another matter altogether. There the risk of failure was much higher than if one simply held the mortgage pool. But at the right yield they too would find a buyer.
The idea, in the wake of the savings-and-loans disaster, was to spread risk outward from those immediately involved in lending to mortgage borrowers and to attract investors by turning mortgages into securities that offered a wide range of yield-risk profiles. And it worked. In 1980, 67 percent of American mortgages had been held directly on the balance sheets of depository banks. By the end of the 1990s, the risks involved in America’s system of long-term, fixed interest, easy repayment mortgages were securitized and spread across a much wider segment of the financial system than had been the case in 1979 when Volcker made his shock announcement. The GSEs held them. Banks held them. But so too did pension and insurance funds.23
Compared with the model of the savings and loan, securitization thus did its job in spreading risk. But did it by the same token reduce the incentive to carefully monitor the underlying loans? By splitting origination from funding, had the new system eliminated the incentive to monitor loans carefully? Whereas a local lender that held a mortgage for its entire thirty-year duration had every reason to monitor its customer very carefully, by the 1990s American mortgages were passing through at least five different institutions—originators, wholesalers of packages of mortgages, underwriters who assessed risk, government-sponsored enterprises and servicers who managed the flow of interest income—before being sold to an investor. Along that chain, what confidence could an investor have that the job was being done correctly? At each step of the way the main concerns were volume and fees. Who had an interest in maintaining quality? Perhaps it was not the government subsidy but these perverse incentives that led to the huge boom in bad lending and the crisis of 2007–2008.24
It is a theory that would have a superficial plausibility if the 1990s model of GSE-centered mortgage finance had still been dominant in the early 2000s. But, in fact, in the early 2000s, when the subprime boom unfolded, the industry had changed again. Securitization was more dominant than ever. The GSEs were still responsible for buying and securitizing the top-tier conforming mortgages. But as a new range of actors entered the mortgage market with a more dynamic and expansive agenda, their principal business model was not to disaggregate and to spread the risk but to integrate every step of the process, including the holding of large quantities of securities on their own balance sheets.25 It was this growth model, based on integration, not disintegration, that would blow the system up.
II
The path that led to the supercharged private mortgage industry of the early 2000s was twisted, but it too goes back to the breakdown of Bretton Woods in the 1970s and the unfettering of currencies, prices, interest rates and capital movem
ents that followed. It was not just the savings and loans but the entire financial sector that was forced to rethink its business model, and this went for the investment banks of Wall Street as much as for the commercial banks.
It is barely too much to say that the new, deregulated world of national and international finance was made for the investment banks.26 Through their business of trading on their clients’ behalf and launching debt and other securities, they enjoyed an “edge” over all other participants in the market.27 In 1975 the abolition of fixed fees charged by Wall Street brokers for trading stocks led to fierce competition, wiped out smaller firms and forced the integration of trading, research and investment banking. In the 1980s, with interest rates coming down and bonds beginning their long bull run, trading in so-called fixed-income securities—as opposed to equities—became ever more important. Drexel Burnham Lambert pioneered the market in high-yield corporate bonds, also known as junk bonds. Meanwhile, Salomon Brothers helped the GSEs devise the securitization model and launch each new batch of mortgage-backed securities. For other clients, the investment bankers were hard at work figuring out how to hedge against fluctuations in currencies and interest rates. They developed swaps, for instance, that allowed clients to trade excessive exposures in currencies. They made instruments that allowed one client to take on the risk of fluctuating interest rates while another client opted for fixed rates. In the 1990s a team at J.P. Morgan devised the credit default swap (CDS), which offered protection against the risk of nonpayment and allowed lenders to fine-tune their lending risk.28 At the same time, the investment banks progressively increased their own trading activity. They discovered the profits to be made through volume and leverage. The returns were extraordinary. In the early 1980s America’s investment banking elite earned returns of more than 50 percent on equity.
But achieving scale raised the question of funding. Investment banks don’t have deposits. They borrow the money they lend on wholesale markets from other banks or institutional funds. In the aftermath of the inflation and interest rate shocks of the late 1970s and early 1980s, this put them in a sweet spot. If investment banks didn’t have depositors, that suited savers, who, in the wake of the inflation, no longer wanted to put their money in bank deposits either. They opted instead for money market mutual funds (MMF), that characteristic financial institution of the new age.29 These were highly attractive to affluent households looking for better rates than those on offer from bank deposits. The money market mutual funds offered instantly accessible accounts without official government guarantee, but pledged by their private sector operators to return at least a dollar on the dollar plus an attractive rate of interest on top. Bypassing the bankrupt savings and loans and the struggling commercial banks, cash deposits flowed into huge pools of professionally managed cash looking for good yields on Wall Street.
Nor were the MMFs the only ones. Corporations began to manage their cash pools more professionally. Ultrarich individuals who became more and more numerous from the 1970s onward had billions of dollars that were managed by funds and family offices. By the end of the 1990s perhaps as much as a trillion dollars had accumulated in these institutional cash pools, looking for highly liquid, interest-yielding investment opportunities that were absolutely, or close to absolutely, safe.30 Lending against security, or buying the commercial paper of well-known investment banks, was precisely the kind of safe short-term asset that the managers of the cash pools wanted. And acting as the banker to the funds, so-called prime brokerage, was ideal business for the investment banks.
These institutional cash pools and the liquidity they brought to wholesale funding markets were the rocket fuel for the rise of the modern investment bank. The more resources they mobilized, either by borrowing in the wholesale market or on deposit, the bigger the turnover, the larger the profits. Until the 1980s, investment banks were relatively small operations, partnerships, well known and respected on Wall Street and the City of London but not household names. The belief in the ability to manage risk inspired by the new derivative instruments, combined with access to the institutional cash pools, enabled them to scale up their size, creating the “New Wall Street.”31 Firms like Goldman Sachs, Morgan Stanley and Merrill Lynch went from obscurity to star status. Originally built as partnerships, the huge scaling up of trading activity and the derivatives business meant that they needed to issue shares and go public. Merrill Lynch had done so already in 1971. Bear Stearns followed in 1985, Morgan Stanley in 1986. Goldman Sachs was the last to launch its IPO in May 1999.32 With Robert Rubin a classic exponent of this new Wall Street, the investment banks even had their man in government. Goldman Sachs began to earn its nickname as “government Sachs.”
Since the 1980s the investment banks had built their businesses on navigating uncertainty. As asset market booms they leveraged up.33 But sometimes uncertainty bites back. Between 1994 in Mexico and 1998 in Russia the globalized American banks faced a series of major crises. In September 1998, but for concerted action by the major Wall Street firms, the implosion of Long-Term Capital Management triggered by the uncertainty spreading from Russia might have brought down the entire hedge fund industry.34 That was followed by the dot-com boom and bust in 1998–2001—a creation of the new Wall Street as much as of Silicon Valley. Finally there was the spectacular accounting scandal at Enron, which took down once legendary accountants and management consultants Arthur Andersen. By the early 2000s, after two decades of dramatic growth, Wall Street was facing a political and regulatory backlash and urgently needed the “next big thing.” Given the expertise of the investment banks in bond trading and their role in the securitization of mortgages on behalf of the GSEs, it was not hard to foresee where scrappy investment banks like Lehman and Bear Stearns would look next.
For the commercial banks, the post-Volcker age was tougher. They lost deposits. They lost mortgage business. Might they go the way of the savings and loans?35 To reestablish profitability in the 1990s America’s high street banks underwent spectacular consolidation. The top ten banks increased their share of total assets from 10 to 50 percent between 1990 and 2000. In addition they looked for a new business model.36 Rather than thinking of themselves as maintaining lifelong relationships with clients and their communities, they repurposed themselves as service providers for a fee. They had always originated mortgages but had generally sold them to the GSEs. Given the pressure that they were now under, the mortgage market, with its multiple layers of origination, securitization, selling and servicing, seemed like a natural bridge between their familiar high street banking business and their aspirations to high finance. But to engage in that full range of activities they needed regulatory relief. The New Deal–era regulations that separated retail from investment banking had to fall. The Clinton Treasury, first under Rubin and then under Larry Summers, gave them what they needed. In 1999 the last remnants of 1930s banking regulations were swept aside. Citigroup and Bank of America rushed through the opening to a new era of American universal banking. Stretching from the high street to Wall Street, it was a model more familiar from Continental Europe now applied to America.
The third group of actors in the mortgage boom of the early 2000s was already in the business in the 1990s. They were banks like Washington Mutual, a survivor of the savings-and-loan disaster, and specialized mortgage lenders like Countrywide.37 As feeders to the GSEs they were restricted to mortgage origination. But why limit their ambitions? Why not integrate the entire chain? By the late 1990s and early 2000s all three groups of banks—investment banks, commercial banks and mortgage lenders—were following this logic. Rather than organizing their mortgage business around the GSEs, they set out to build integrated mortgage securitization businesses. Countrywide expanded from origination to securitization. A giant bank like Citi could envision itself as a provider at every stage, originating, securitizing, selling, holding and dealing in MBS. Even more remarkable was the evolution of investment banks like Lehman and Bear Stearn
s that had previously defined themselves through their remoteness from ordinary retail customers. Already in the 1990s Bear added the mortgage originator and servicer EMC to its portfolio and Lehman added four small mortgage lenders to its investment bank.
By the early 2000s the corporate strategies centered on private mortgage securitization were fully in place. But Fannie Mae and Freddie Mac still enjoyed a dominant position in the market thanks to their funding advantage. What gave the private mortgage securitization machine its chance was another interest rate shock combined with a hiccup at the GSE.38
When the dot-com bubble was followed by the shock of 9/11, the Fed dropped interest rates to 1 percent. As Alan Greenspan clearly intended, this unleashed a scramble among borrowers to refinance as many long-term mortgages as possible at lower rates. This was painful from the point of view of the original lenders. But it triggered an immediate wave of consumer spending, and for the mortgage industry it generated a huge surge in fees. The industry churned as it had never churned before. As compared with $1 trillion in new mortgages issued in 2001, in 2003 mortgage origination soared to $3.8 trillion, of which $2.53 trillion were for refinancing. In this huge boom the GSEs were still the major players. They continued to monopolize the prime mortgage segment. Their share of the market reached its maximum point in 2003 at 57 percent. But at that point it stalled. In the huge surge of business in the early 2000s not everything at the GSEs was aboveboard. Accounting and regulatory irregularities piled up. Fearful of another Enron, regulators subjected first Freddie Mac and then Fannie Mae to capital surcharges. They either had to raise new capital or contract their balance sheets. To make sure it was the latter, caps were imposed on their total balance-sheet size.39 The door to the private issuers was opened.