Inside Job (13 page)

But finally, there is the problem of reality—both in general, and specifically in relation to this bubble. The fact is that people do commit fraud, even when they know that it is
personally risky for them. To take an extreme example, we
know
that people still perpetrate literal Ponzi schemes (Allen Stanford, $8 billion; Bernard Madoff, $20 billion), even though every
Ponzi scheme is mathematically guaranteed to collapse at some point, with inevitable imprisonment to follow. And now, thanks primarily to private lawsuits and secondarily to a few government
investigations, we also know that during the housing bubble many people were, in fact, consciously committing fraud and selling defective products.

Moreover, in the case of the housing bubble—and unlike isolated noninstitutional Ponzi schemes—everyone got away with it. The people responsible for the bubble are still wealthy, out
of jail, socially accepted, and either retired or not, as they prefer. Even those who really did lose something—CEOs, eminences on boards of directors, business unit heads, people such as
Richard Fuld, Robert Rubin, Angelo Mozilo, Stan O’Neal, or Joseph Cassano—are still enormously wealthy as a result of having taken lots of cash out during the bubble, and/or from their
severance payments.

In fact on a net basis, many of these people made far more money by creating and participating in the bubble than they would have made by staying out of it, even though they destroyed their
firms. It is true that they lost the value of whatever shares they owned when their firms
collapsed in 2008. But over the previous seven years their incomes had been hugely
inflated by the bubble, and they cashed a lot of it outbehaviour, incidentally, not suggestive of deep faith in (or concern for) their firms’ futures. Some of them also received enormous
severance payments when they were fired, even
after
the nature and effects of their conduct became known. To a remarkable extent they have avoided social disgrace, and still occupy positions
of prestige, even power. So the defenders of the banks are in effect saying: who are you going to believe, me or your own lying eyes?

During the bubble, the Bush administration and the Federal Reserve were essentially AWOL; if anything, they made matters worse. In 2004 the SEC voted unanimously to allow the five largest
investment banks to calculate their own leverage limits, based on their internal risk models. This meant that by the time the party ended, several of them were leveraged at more than thirty to one,
meaning that if the value of their assets declined by just 3 percent, they would be bankrupt. As a result, when the crisis occurred, three of the banks, Bear Stearns, Lehman Brothers, and Merrill
Lynch, were insolvent by 2008. Only Goldman Sachs and Morgan Stanley survived, and that by grace of government rescue operations.
¹
During this time, the
SEC and other government regulators reduced their risk analysis and enforcement staffs, and basically left the investment banks unsupervised. The same was true of the industry’s several
“self-regulatory” organizations, such as the Financial Industry Regulatory Authority (FINRA), the Securities Investor Protection Corporation (SIPC), and others. FINRA, for example,
describes its mission thus on its website:

Well, FINRA didn’t do too well. (The head of FINRA during the bubble? Mary Shapiro, who became chair of the SEC in 2009, appointed by President Obama.) As a result of
the failure of both government and private regulation, during the entire bubble the inmates were in charge of the asylum. Given their incentives, a massive fraud was entirely rational for most of
them, even if they had known in advance of all the damage it would cause.

Was it all really that naked? Yes, it was. Some of what follows might be a little dry. But I ask readers to bear with me because later in this book, I will be saying some rather strong
things—things like, these people should be in jail, they should have their wealth taken from them and given to people whose lives they destroyed, they should live in disgrace for the rest of
their lives, it is shocking that they have not been prosecuted, and it is obscene and dangerous that they still occupy prominent positions in government, universities, companies, and civic
institutions. So in the next couple of chapters, there will be some rather detailed stuff, because when we come to the punch line, I want it to be convincing.

Let us start our survey of the investment banking industry’s conduct with a lawsuit filed by the discount brokerage and asset management firm Charles Schwab, Inc. The case provides an
unusually broad statistical picture of the securities that Wall Street produced, and also helps provide context for some of the truly stunning conduct we shall encounter later.

SCHWAB SUED THE
broker-dealer divisions of twelve major financial institutions. The suit is based on the representations made in the offering materials
for thirty-six securitizations—mortgage-backed securitiesthat were purchased from these banks by various units of Schwab between 2005 and 2007. The defendants are BNP Paribas, Countrywide,
Bank of America, Citigroup, Credit Suisse, Deutsche Bank,
Goldman Sachs, Greenwich Capital, HSBC, Wells Fargo, Morgan Stanley, and UBS.

Investors in residential mortgage-backed securities do not have access to individual loan files when they make the purchase; they are shown only summary data for each mortgage on the “loan
tapes” that accompany a sales prospectus. But when Schwab sued, it analysed the summary data for 75,144 loans included in the securitizations it had purchased. While not a random sample, it
is a broad one.
³
If anything, the sample is probably substantially above average in quality, because Schwab invested conservatively, and the defendants in
this lawsuit do not include Bear Stearns and Merrill Lynch, two firms that produced much of the worst junk. As we’ll see shortly, some of it was
much
worse.

Still, Schwab’s analysis estimated that 45 percent of the loans violated the representations made in selling the securities. Schwab used four separate tests; most of the suspect loans
failed more than one of them.

The results of Schwab’s tests suggest that all securitizers significantly understated LTV ratios, and that all of the securities included substantial numbers of loans with LTV ratios of
more than 100 percent—in other words, loans for more than the home was worth. In one security comprising 1,597 loans, the model estimated that 626 loans were overvalued by more than 5
percent, with only 69 undervalued by as much. The pool’s weighted-average LTV rose from a represented 73.8 percent to 90.5 percent, and the model estimated that 196 loans had LTVs of more
than 100 percent, although the documents represented that none did. Roughly similar outcomes applied to all the other pools. A second test was applied to properties that were subsequently sold. The
sales
prices were consistently below the values implied by the claimed LTVs, in a pattern consistent with the model. Schwab notes that given the average leverage of the loans,
a 10 percent LTV overstatement implies an 80 percent reduction in stated equity.
²

2. Stated Liens
.
All of the securities’ offering materials stipulated that all property liens were fully disclosed. No mortgage loan, supposedly, is
ever funded without a title search to spot any lien that might be senior to the bank’s lien. Schwab did new title searches on the properties. In one pool of 2,274 loans, 669 had undisclosed
liens that, on average, reduced stated equity by 91.5 percent. In some pools, as many as half the loans had undisclosed liens.

3. Occupancy Status
.
 Owner-occupied homes have lower default and foreclosure risk than second homes or investment properties. For this reason, all of
the securities’ offering materials contained representations as to the percentage of loans in the security that were for owner-occupied (primary) residences. Schwab researched the
properties listed as primary residences for indications that they were really not, including:

Using these tests, in one pool of 1,498 loans, in which only 99 were alleged to be nonprimary residences, the correct total was estimated to
be 598, with most of those
failing more than one test. Roughly similar results prevailed throughout all the securities.

4. Internal Guideline Compliance
.
Schwab collected data on early payment defaults (EPDs) on all securitized loans issued by the originators of the mortgages
in these securities from 2001 through 2007. In well-underwritten mortgages, the rate of EPDs, defined as default during the first six months of the loan, should be vanishingly small. Schwab
investigated whether and when EPD rates changed during the bubble, and if so, whether stated credit guidelines changed at the same time.

None of the lenders changed its represented credit guidelines during this period. Despite this,
every
lender experienced a sharp upward break in its EPD rate at some point, a trend that
persisted thereafter in every case. But the date at which the break occurred was different for
each originator
. For Countrywide, for example, EPDs suddenly quadrupled starting in the first
quarter of 2005 and stayed at consistently high rates from then on. All the other lenders showed comparable upward EPD breaks; the starting dates ranged from mid-2003 to mid-2007.

Since the breaks occurred at different times for different originators, they were not likely to have been caused by general economic conditions. Rather, they probably reflected an internal
policy change; but those policy changes were never reflected in the lending guidelines officially represented to investors. (Credit guideline representations were usually detailed and specific.)
The obvious interpretation is that, one by one, originators chose to go aggressively down-market to make more money, while officially denying that they had done so.

And finally, in order to justify the representations made to investors, all of the securitizing banks hired outside firms to review the quality of the loans they bought. The largest and most
comprehensive reviewer was Clayton Holdings, which examined 911,000 mortgages for twenty-three different securitizing banks, including all of the largest securitizers, between January 2006 and July
2007. In testimony and documents provided to the Financial Crisis Inquiry Commission (FCIC), senior Clayton executives revealed that 28 percent of the loans they examined did not meet even the
securitizers’ own internal guidelines. Despite this, 39 percent of all loans that failed the securitizers’ guidelines were purchased and securitized anyway, a fact never disclosed to
the investors purchasing the final securities.
⁴

But were the banks just sloppy passive conduits, or active co-conspirators? Did they understand the full implications of their actions? To understand that, we need to look inside them. This has
not been done nearly enough, because the US government hasn’t really tried. There have been no criminal prosecutions using large-scale subpoena power and sworn testimony. All we have are
civil suits, in which the banks ferociously (and often successfully) resist subpoenas and often succeed in keeping records sealed. Even the FCIC established by the Obama administration had
extremely limited subpoena power, and a tiny budget. If there ever were to be a truly serious investigation, I have no doubt that we would learn much, much more than we currently know.

But what we currently know is still quite impressive.

BEAR STEARNS WAS
one of the most experienced mortgage players on Wall Street. In the three years from 2004 through 2006, Bear securitized nearly a
million mortgage loans with a total value of $192 billion—serious money, even for Wall Street.

Bear Stearns had been in the mortgage-backing business for nearly two decades, mostly securitizing high-quality loans. But in 2001 it created a new mortgage conduit, EMC
Mortgage, to securitize Alt-A loans. In 2003 EMC started to securitize subprime, stated income, and “no doc” loans, then in 2005, second-lien loans, of the type that even Angelo Mozilo
called the “most toxic” he had seen in his entire career.