But that started to change in the late 1990s. Stock trading was becoming commoditized, which meant that victory increasingly went to the lowest-cost player. The underwriting fees paid by corporations looked as though they’d be squeezed. The universal banks had more money to work with, which meant that they could use their balance sheets to offer commercial loans, enticing companies to the banks’ investment advisers when it came time to do deals. (That such blatant “tying” was against the law seemed beside the point.) Life as a stand-alone investment bank was presumed to be an unhappy one, the thinking went.
Then a surprising thing happened. The big financial supermarkets found that combining hotshot investment bankers with milquetoast commercial bankers flopped. Citigroup and J.P. Morgan languished. Deutsche Bank
and Credit Suisse remained Wall Street also-rans. Bank of America
retrenched to focus on retail banking.
But as the stock market emerged from the dotcom crash, most of the independents flowered. Their earnings and return on equity soared, as did their stocks. From the end of 2002 to the middle of this year, Goldman’s shares rose about 240 percent. From the end of that year to the beginning of this one, Lehman’s soared about 200 percent, and Bear’s increased a comparatively paltry 175 percent before giving back significant amounts amid the recent problems. All easily outperformed the markets and their peers.
Investors had noticed that the independent investment banks rarely fail and almost always thrive. Bear Stearns, formed in 1923, boasts it has never had an unprofitable year. Lehman had carved out a niche in fixed income and mortgages. Bear financed mortgages as well and served hedge funds. Goldman Sachs, brightest of them all, became the bank that runs the world, not just Wall Street.
How did they do it? Largely by taking more risks with their own money than they had before. Sure, they still do top-notch advisory work and act as dealers. But the growth in revenue and income has come mostly from betting with their own capital. They set up in-house private equity firms and spread money around on their own proprietary trading desks. In hindsight, it seems obvious that this couldn’t last.
It’s become a cliché in recent years to label Goldman Sachs a hedge fund. But that doesn’t mean the cliché is wrong. The percentage of Goldman’s revenue that came from trading for its own accounts and investing its own money rose to 63 percent in 2006, according to Portales. From 1999 to 2003, that percentage never broke 50 percent. The portion of Lehman’s revenue that came from its own investments was 55 percent last year, up from an average of 44 percent between 1999 and 2004.
In the most recent quarter, investors learned how vulnerable these revenues can be. Bear, Lehman, and Morgan Stanley all suffered big drops from such trading. Goldman somehow managed to thrive. We’ll see if that lasts.
As a group, the five independent investment banks got 45 percent of their revenue from investing their own capital last year. In 1999—at the height of the tech boom, when they were generating fees from more-conventional sources—that number was about 30 percent.
The assets on the independents’ balance sheets have ballooned since then, and they have become riskier institutions because of it. A recent paper by Tobias Adrian of the New York Federal Reserve and Hyun Song Shin of Princeton University makes it clear that the banks aren’t much different from those low-credit borrowers now haunting the housing market. As home prices went up, homeowners borrowed against their houses; as prices have fallen, some have found themselves underwater, having to default.
You might think the investment banks would have been smarter. But they weren’t. When times are good, more investors are willing to lend and take risks. This gives investment banks excess capital to invest. Prices on assets rise. A global search for yield begins. The banks load up on assets and increase their leverage. At its core, what the banks are doing is buying high—hardly the stuff of a great business model.
As Adrian and Shin point out, this activity can quickly spill over into the broader market: “There is the potential for a feedback effect in which weaker balance sheets lead to greater sales.”
Of course, banks protect themselves by hedging, which homeowners can rarely do. But even when this factor is taken into account, banks are still exposed. For instance, after trades that are netted out or hedged by other trades are subtracted, Bear had an “adjusted” leverage of 15.5 times its capital at the end of the second quarter. That’s way up from the 11-times level it had been at before the credit boom began. Its gross leverage declined slightly to 29.9 times capital in the third quarter. (By comparison, at the end of the second quarter, Goldman’s gross leverage ratio was 24.5 times, up from 18.7 in late 2003. Its adjusted leverage was up slightly from then.)
In Lehman’s third quarter, its leverage shot up by both measures. Its total leverage was more than 30 times capital, up from 25.8 times a year ago. Net leverage rose almost 20 percent, to 16 times, from a year earlier. (As a benchmark, it’s notable that Long-Term Capital Management, the hedge fund that notoriously blew up in 1998, typically had leverage of 25 to 1.)
When an investor has leverage of 30 to 1, all it takes is a reduction of about 3 percent in the assets to wipe out the equity. In the third quarter, the banks uniformly said that the worst was behind them and that none had lost money. But what happens if the market blindsides them again as it did in August?
Just as in the housing market, it’s hard to predict which mechanisms will cause an investment-banking meltdown. It’s almost impossible to tell how risky the assets on the banks’ balance sheets are, for instance, or how they have chosen to hedge or what the risks are that the parties on the other side of their trades will default.
All over Wall Street, investors scrambled this summer to get an education in all sorts of wondrous financial arcana. There are the collateralized debt obligations and their dodgy progeny, known as C.D.O.-squareds. There are the structured investment vehicles and their dubious cousins S.I.V.-lites, the off-balance-sheet vehicles that banks and financial companies use for short-term funding.
And then there’s the investment bankers’ dependence on so-called level-three assets. Thanks to a recent ruling by the Financial Accounting Standards Board, financial institutions now break out the fair value of various types of difficult-to-value assets into levels. There is level one, which is, to put it in the language of buffalo wings, “hot.” A level-one asset is relatively easy to value, with quoted prices in an active market for identical assets or liabilities. Then there is “extra hot,” officially known on Wall Street as level two. On this level, quoted prices for similar assets are out there, but accountants are needed to adjust their prices based on market criteria. And then there is “atomic”—level three. Nothing comparable to level-three assets trades on any market, and there isn’t even an objective standard by which to value them. Level-three assets include bespoke derivative contracts that never trade, complex distressed debt, and mortgages that have gone bad.
To value these assets, the banks have sophisticated internal models, which they point out are carefully pored over by their auditors. But it’s sort of like finding the asset’s value by sticking your finger in the wind.
It may not have been an accident that unraveling all this is so tough. Indeed, an official at one of the banks told me that the balance sheets are opaque on purpose: “We don’t make it easy to understand what instruments we use to make that leverage.”Like frat boys at a sports bar, Wall Street has binged on level-three assets. In the first half of this year, Bear Stearns ended up with $18 billion of level-three assets on its balance sheet.
That’s 135 percent of its entire book value, or net worth. For Lehman, level-three assets represented 104 percent of its book value. And for Goldman, 141 percent.
When things are going well, level-three assets are lucrative, since the investment banks have the expertise to dominate the market for them. But things might be changing. The canary in this particular mine shaft is Bear, which actually took losses on these risky assets in the first half of this year. In the third quarter, Lehman’s level-three assets rose both in dollar terms and as a percentage of assets, compared with the second quarter. Morgan Stanley’s and Goldman’s level-three assets rose as well. (Third-quarter level-three profitability disclosures aren’t available yet.)
It’s hard to find any corner of the U.S. investment-banking world that’s doing well. Structured finance issuance is down. Mergers-and-acquisitions activity is slowing. The banks are already making small numbers of employees redundant. Bonus season this year will make Montgomery Burns look generous. Of the five banks, only Goldman is looking solid.
It’s unlikely that Bear will be able to stay independent. Lehman has been trying to imitate Goldman, which logged blockbuster earnings in the third quarter even as its competitors were staggering. But whether Lehman will have any success depends on whether this crisis has truly passed, and that bet has long odds. In the meantime, expect the perennial takeover speculation surrounding Morgan Stanley and Merrill Lynch to reheat.
There is an end-of-era feel to the whole thing. After all those years of investment bankers being mistakenly lambasted as rogues, it will be ironic if the moment Wall Street finally embraced its reputation became its undoing.