What’s Inside America’s Banks?

Some four years after the 2008 financial crisis, public trust in banks is as low as ever. Sophisticated investors describe big banks as “black boxes” that may still be concealing enormous risks—the sort that could again take down the economy. A close investigation of a supposedly conservative bank’s financial records uncovers the reason for these fears—and points the way toward urgent reforms.

At the heart of the problem is a worry about the accuracy of banks’ financial statements. Some of the questions are basic: How do banks account for loans? Can investors accurately assess the value of those loans? Others are far more complicated: What risks are posed by complex financial instruments, such as the ones that caused JPMorgan’s massive loss? The answers are supposed to be found in the publicly available quarterly and annual reports that banks file with the Securities and Exchange Commission.

The Financial Accounting Standards Board, an independent private-sector organization, governs the accounting in these filings. Don Young, currently an investment manager, was a board member from 2005 to 2008. “After serving on the board,” he recently told us, “I no longer trust bank accounting.”

Accounting rules have proliferated as banks, and the assets and liabilities they contain, have become more complex. Yet the rules have not kept pace with changes in the financial system. Clever bankers, aided by their lawyers and accountants, can find ways around the intentions of the regulations while remaining within the letter of the law. What’s more, because these rules have grown ever more detailed and lawyerly—while still failing to cover every possible circumstance—they have had the perverse effect of allowing banks to avoid giving investors the information needed to gauge the value and risk of a bank’s portfolio. (That information is obscured by minutiae and legalese.) This is true for the complicated questions about financial innovation and trading, but it also is true for the basic questions, such as those involving loans.

At one point during Young’s tenure, some members of the Financial Accounting Standards Board wanted to make banks account for loans in the same way they do for securities, by recording them at current market values, a method known as “fair value.” Banks were instead recording the value of their loans at the initial loan amount, and setting aside a reserve based on their assumptions about how likely they were to get paid back. The rules also allowed banks to use different methods to measure the value of the same kind of loans, depending on whether the loans were categorized as ones they planned to keep for a long time or instead as ones they planned to sell. Many accounting experts believed that the reported numbers did not give investors an accurate or reliable picture of a bank’s health.

After bitter battles, turnover on the board, worries about acting in the middle of the financial crisis, and aggressive bank lobbying, the accounting mandarins preserved the existing approach instead of switching to fair-value accounting for loans. Young believes that the numbers are even less reliable now. “It’s gotten worse,” he says. When we asked another former board member, Ed Trott, whether he trusted bank accounting, he said, simply, “Absolutely not.”

The problem extends well beyond the opacity of banks’ loan portfolios—it involves almost every aspect of modern bank activity, much of which involves complex investment and trading, not merely lending. Kevin Warsh, an ex–Morgan Stanley banker and a former Federal Reserve Board member appointed by George W. Bush, says woeful disclosure is a major problem. Look at the financial statements a big bank files with the SEC, he says: “Investors can’t truly understand the nature and quality of the assets and liabilities. They can’t readily assess the reliability of the capital to offset real losses. They can’t assess the underlying sources of the firms’ profits. The disclosure obfuscates more than it informs, and the government is not just permitting it but seems to be encouraging it.”

Accounting rules are supposed to help investors understand the companies whose shares they buy. Yet current disclosure requirements don’t illuminate banks’ financial statements; instead, they let the banks turn out the lights. And in that darkness, all sorts of unsavory practices can breed.

We decided to go on an adventure through the financial statements of one bank, to explore exactly what they do and do not show, and to gauge whether it is possible to make informed judgments about the risks the bank may be carrying. We chose a bank that is thought to be a conservative financial institution, and an exemplar of what a large modern bank should be.

Wells Fargo was founded on trust. Its logo has long been a strongly sprung six-horse stagecoach, a fleet of which once thundered across the American West, loaded with gold. According to the firm’s official history, “In the boom and bust economy of the 1850s, Wells Fargo earned a reputation of trust by dealing rapidly and responsibly with people’s money.” People believed Wells Fargo would keep their money safe—the bank’s paper drafts were as good as the gold it shipped throughout the country.

For a century and a half, Wells Fargo stock was also like gold, which is what led Warren Buffett to buy a stake in the bank in 1990. Since then, Buffett and Wells Fargo have been inextricably linked. As of fall 2012, Buffett’s firm, Berkshire Hathaway, owned about 8 percent of Wells Fargo’s shares.

Today, Wells Fargo still prominently displays the stagecoach logo at branches, in advertising, on the 12,000-plus ATMs that dot the country, and even at the bank’s museum stores. There, visitors can buy wholesome, family-friendly items: a stagecoach night‑light; stagecoach salt and pepper shakers; a hand-painted ceramic stagecoach pillbox. These are more than tchotchkes. They are emblems of the bank’s honest and honorable mission.

Buffett’s impeccable reputation has rubbed off on the bank. Wells Fargo is widely regarded as the most conservative of the nation’s biggest banks. Many investors, regulators, and analysts still believe its financial reports reflect a full, fair, and accurate picture of its business. The market value of Wells Fargo’s shares is now the highest of any U.S. bank: $173 billion as of early December 2012. The enthusiasm for Wells Fargo reflects the bank’s good reputation, as well as one seemingly simple fact: the bank earned solid net income of nearly $16 billion in 2011, up 28 percent from 2010.

To find out what’s behind that fact, you have to read Wells Fargo’s annual report—and that is where we began our adventure. The annual report is a special document: it is the place where a bank sets forth the audited details of its business. Although banks also submit unaudited quarterly reports and other periodical documents to the SEC, and have conference calls with analysts and shareholders, the annual report gives investors the most complete and, supposedly, reliable picture.

(Today, big banks have to answer to a dizzying litany of regulators—not only the SEC, but also the Federal Reserve, the Office of the Comptroller of the Currency, the Federal Deposit Insurance Corporation, the Commodity Futures Trading Commission, the newly created Consumer Financial Protection Bureau, and so on. The disclosure regimes vary, adding to the confusion. Banks confidentially release additional information to these regulators, but investors do not have access to those details. That regulators have these extra, confidential disclosures isn’t much comfort: given the inability of regulators to police the banks in recent years, one of the only groups that investors trust less than bankers is bank regulators.)

Wells Fargo’s most recent annual report, covering 2011, is 236 pages long. It begins like a book an average person might enjoy: a breezy journey through a year in a bank’s life. On the cover, that stagecoach appears. The first page has a moving story about a customer. The next few pages are filled with images of guys in cowboy hats, a couple holding hands by the ocean, cupcakes, and solar panels. In bold 50‑point font, Wells Fargo reports that it contributed $213.5 million to nonprofits during the year, and it even does the math to make sure we appreciate its generosity: “$4.1 million every week or $585,000 every day or $24,000 every hour.” The introduction’s capstone is this: “We don’t take trust for granted. We know we have to earn it every day in our conversations and actions with our customers. Here’s how we try to do that.”

The sheer volume of “trading” at Wells Fargo suggests that the bank is not what it seems.

Fortunately for Wells Fargo, most people do not read past the introduction. In the pages that follow, the sunny faces of satisfied customers disappear. So do the stories. The narrative is replaced by details about the bank’s businesses that range from the incomprehensible to the disturbing. Wells Fargo told us it devotes “significant resources to fulfilling all reporting requirements of various regulators.” Nevertheless, these disclosures wouldn’t earn anyone’s trust. They are littered with language that says nothing, at length. The report is riddled with progressively more opaque footnotes—the financial equivalent of Dante’s descent into hell. Indeed, after the friendly introduction, the report ought to bear a warning to the inquisitive reader intent on truly understanding the bank’s financial positions: “Abandon all hope, ye who enter here.”

The first circle of Wells Fargo’s version of the Inferno, like Dante’s Limbo, merely hints at what is to come, yet it is nonetheless unsettling. One of the main purposes of an annual report is to tell investors how a company makes money. Along these lines, Wells Fargo splits its businesses into two apparently simple and distinct parts—“interest income” and “noninterest income.” At first blush, these two categories appear to parallel the two traditional sources of banking income: interest from loans and customer fees.

But here the descent begins. Suddenly, this folksy mortgage bank starts showing signs of a split personality. It turns out that trading activities, the type associated with Wall Street firms like Goldman Sachs and Morgan Stanley, contribute significantly to each of Wells Fargo’s two categories of income. Almost $1.5 billion of its “interest income” comes from “trading assets”; another $9.1 billion results from “securities available for sale.”

One billion dollars of the bank’s “noninterest income” are “net gains from trading activities.” Another $1.5 billion is income from “equity investments.” Up and down the ledger, abstruse, all-embracing categories appear: “other fees earned from related activities,” “other interest income,” and just plain “other.” The income statement’s “other” catchalls collectively amounted to $6.6 billion of Wells Fargo’s income in 2011. It will take the devoted reader 50 more pages to find out that the bank derives a big chunk of that “other” income from, yes, “trading activities.” The sheer volume of “trading” at Wells Fargo suggests that the bank is not what it seems.

Some bank analysts say these trading numbers are small relative to the bank’s overall revenue ($81 billion in 2011) and profit (again, $16 billion in 2011). Other observers don’t even bother to look at these details, because they assume Wells Fargo is protected from trading losses by its capital reserves of $148 billion. That number, assuming it is accurate, can make any particular loss appear minuscule. For example, buried at the bottom of page 164 of Wells Fargo’s annual report is the following statement: “In 2011, we incurred a $377 million loss on trading derivatives related to certain CDOs,” or collateralized debt obligations. Just a few years ago, a bank’s nine-figure loss on these sorts of complex financial instruments would have generated major headlines. Yet this one went unremarked‑upon in the media, even by top investors, analysts, and financial pundits. Perhaps they didn’t read all the way to page 164. Or perhaps they had become so numb from bigger bank losses that this one didn’t seem to matter. Whatever the reason, Wells Fargo’s massive CDO-derivatives loss was a multi-hundred-million-dollar tree falling silently in the financial forest. To paraphrase the late Senator Everett Dirksen, $377 million here and $377 million there, and pretty soon you’re talking about serious money.

Even conservatively run banks can be risky, as George Bailey learned in It’s a Wonderful Life. But the Bailey Building and Loan Association did not earn money from trading. Trading is an inherently opaque and volatile business. It is subject to the vagaries of the markets. And yet in the past two decades, as profits from traditional lending and brokering activities have been squeezed, banks have turned more and more to trading in order to make money.

Today, banks’ trading operations involve more leverage, or borrowed money, than in the past. Banks also obtain a form of leverage by promising to pay money in the future if some event doesn’t go their way (much like an insurance company must pay out a lot of money if a house it covers burns down). These promises come in the form of derivatives, financial instruments that can be used to hedge against various risks—like the possibility that interest rates will rise or the likelihood that a company will default on its debts—or simply to place bets on those same possibilities, hoping to profit. Because many of these bets are both large and complex, trading carries the potential for catastrophic losses.

The cryptic way Wells Fargo describes its trading raises many questions. The bank breaks what it calls “net gains from trading activities”—which doesn’t cover all of its trading income, but is an important part—into three subcategories, leaving the annual-report reader to play a kind of shell game.

Look first at “proprietary” trading—activity a firm undertakes to make money for its own account by buying or selling stocks, bonds, or more-exotic financial creations. Self-evidently, this activity might involve big risks. When this shell is lifted, the bank’s exposure seems reassuringly inconsequential: the reported loss is just $14 million. Still, there may be more under this shell than meets the eye: that $14 million might not be indicative of the bank’s true exposure. Was Wells Fargo just lucky to finish slightly down after a roller-coaster year of wild gambling with much bigger gains and losses? Without more information about the size of the bank’s bets, it is impossible to know.

A second subcategory is “economic hedging.” An activity labeled “hedging” might sound soothing. Wells Fargo says it lost an inconsequential $1 million from economic hedging in 2011. So maybe there is nothing to worry about under this shell, either. In its pure form, hedging is supposed to reduce risk. A person buys a house and then hedges the risk of a fire by purchasing insurance. But hedging in the world of finance is more complex—so much so that it requires advanced mathematics and computer modeling, and still can be little better than guesswork. It is difficult to anticipate how a portfolio of complicated financial instruments will respond as variables like interest rates and stock prices go up and down. As a result, hedges don’t always work as intended. They may not fully eliminate large risks that banks think they’ve taken care of. And they may inadvertently create new, hidden risks—“unknown unknowns,” if you will. Because of all this complexity, some traders can disguise speculative positions as “hedges” and claim their purpose is to reduce risk, when in fact the traders are purposely taking on more risk to try to make a profit. That is what the traders within JPMorgan’s Chief Investment Office appear to have been doing. Was Wells Fargo’s “economic hedging” like buying straightforward insurance? Or was it more like speculation—what JPMorgan did? Do the reported numbers suggest low risk when in fact the opposite is true? The bank’s disclosures don’t answer these questions.

Finally we come to a third shell—and there’s unquestionably something to see under this one. It carries an innocuous label: “customer accommodation.” Wells Fargo made more than $1 billion from customer-accommodation trading in 2011. How did it make so much money merely by helping customers? This should be a plain-vanilla business: a broker sits between a buyer and a seller and takes a little cut of the transaction. But what we learned from the 2008 financial crisis, and what we keep learning from incidents such as the JPMorgan scandal, is that seemingly innocuous activities that appear highly profitable can be dangerous to a bank’s health—and to our economy.

Don’t look to the annual report for clarity. Here is the bank’s definition: “Customer accommodation trading consists of security or derivative transactions conducted in an effort to help customers manage their market price risks and are done on their behalf or driven by their investment needs.”

That might seem safe, but the report notably fails to explain why this activity would be so profitable. In fact, at many large banks, customer accommodation can be a euphemism for “massive derivatives bets.” For Wells Fargo, the subcategory of “customer accommodation, trading and other free-standing derivatives” included derivatives trades of about $2.8 trillion in “notional amount” as of the end of 2011, meaning that the underlying positions referenced in the bank’s derivatives were that large then. By way of explanation: if we were to make a bet with you about how much the price of a $70 share of Walmart would change this year—we pay you any increase, you pay us any decrease—we’d say the “notional amount” of the bet is $70.

Wells Fargo doesn’t expect to gain or lose $2.8 trillion on its derivatives, any more than we would expect the payment on our Walmart bet to be $70. Bankers generally assume that the likely risk of gain or loss on derivatives is much smaller than their “notional amount,” and Wells Fargo says the concept “is not, when viewed in isolation, a meaningful measure of the risk profile of the instruments.” Moreover, Wells Fargo reports that many of its derivatives offset each other, as yours might if you placed several wagers that Walmart stock would go up, along with several other bets that it would go down.

Yet, as investors in bank stocks learned in 2008, it is possible to lose a large portion of the “notional amount” of a derivatives trade if a bet goes terribly wrong. In the future, if interest rates skyrocket or the euro unravels, Wells Fargo might sustain huge derivatives losses, just as you might lose the full $70 you bet on Walmart if the company went bust. Wells Fargo doesn’t tell investors how much of the $2.8 trillion it could lose in a worst-case scenario, nor is it required to. Even a savvy investor who reads the footnotes can only guess at what the bank’s potential risk exposure to derivatives might be.

One reason Wells Fargo is trusted more than other big banks is that its notional amount of derivatives is comparatively small. At the end of the third quarter of 2012, JPMorgan had $72 trillion in notional amount on its books—about five times the size of the U.S. economy. But even at Wells Fargo levels, the numbers are so large that they lose their meaning. And they put Wells Fargo’s seemingly immense capital reserves—$148 billion, you’ll recall—in a rather different light.

How much risk is the bank actually taking on these trades? For which customers does it place a requested bet, then negate its risk by taking an exactly offsetting position in the market, so that it is essentially acting as an agent simply taking a commission? And for all these trades, what risk is Wells Fargo taking on its customers? Many of these bets involve the customers’ promises to pay Wells Fargo depending on how certain financial numbers change in the future. But what happens if some of those customers go bankrupt? How much money would Wells Fargo lose if it “accommodates” customers who can’t pay what they owe?

We asked Wells Fargo officials if we could talk to someone at the bank about its disclosures, including those concerning its trading and derivatives. They declined. Instead, they suggested we submit questions in writing, which we did.

In response, Wells Fargo public-relations representatives wrote, “We believe our disclosures on the topics you raised are comprehensive and stand on their own.” In answering our written questions about the annual report, the representatives simply pointed us back to the annual report. For example, when we inquired about the bank’s trading activities, Wells Fargo responded: “We would ask you to refer to our discussion of ‘Market Risk-Trading Activities’ on pages 80–81 in the Management Discussion and Analysis section of the Wells Fargo 2011 Annual Report.”

Yet it was precisely those pages that generated our questions about the bank’s various categories of trading. When we specifically asked Wells Fargo to help us quantify the risks associated with customer-accommodation trading, its representatives pointed us to those same pages. But those pages don’t answer that question. Here is the most helpful of the bank’s disclosures related to customer-accommodation trading:

For the majority of our customer accommodation trading we serve as intermediary between buyer and seller. For example, we may enter into financial instruments with customers that use the instruments for risk management purposes and offset our exposure on such contracts by entering into separate instruments. Customer accommodation trading also includes net gains related to market-making activities in which we take positions to facilitate expected customer order flow.

Bankers, and their lawyers, are careful about the language they use in annual reports. So why did they use the word expected in discussing customer order flow in that last sentence? Is Wells Fargo speculating based on what one of its traders “expects” a customer to do, instead of responding to what a customer actually has done? The language the bank pointed to for answers to our questions only raises more questions.

Wells Fargo’s annual report is filled with similarly cryptic declarations, but not the crucial information that investors actually need. It doesn’t describe worst-case scenarios for customer-accommodation trades, or even include any examples of what such trades might involve. When we asked straightforward questions—such as “How much money would Wells Fargo lose from these trades under various scenarios?”—the bank’s representatives declined to answer.

Only a few people have publicly expressed concerns about customer-accommodation trades. Yet some banking experts are skeptical of these trades, and suspect that they hide huge risks. David Stockman, who was the federal budget director under President Reagan, an investment banker at Salomon Brothers, and a partner at the private-equity firm Blackstone Group, calls the big banks “massive trading operations.” Stockman has become so disillusioned by America’s financial system that he is now regarded, in some quarters, as a wild-eyed heretic, but his expertise is undeniable. He recently told reporters for “The Gold Report,” an online newsletter, “Whether they called it customer accommodation or proprietary is a distinction without a difference.”

Bankers and regulators today might dismiss warnings that customer-accommodation derivatives could bring down the financial system as implausible. But a few years ago, they said the same thing about credit-default swaps and collateralized debt obligations.

The penultimate stop on our expedition through Wells Fargo’s annual disclosures brings us to one of the most important concepts in bank reporting: fair value. It’s the topic that led Don Young to conclude that he could not trust banks’ accounting after fighting about it on the Financial Accounting Standards Board. Banks hold huge amounts of assets and liabilities, including derivatives, and are supposed to record them at their “fair value.” Fair enough? Not so fast.

Like other banks, Wells Fargo uses a three-level hierarchy to report the fair value of its securities. Level 1 includes securities traded in active, public markets; it isn’t too scary. At Level 1, fair value simply means the reported price of a security. If Wells Fargo owned a stock or bond traded on the New York Stock Exchange, fair value would be the closing price each day.

Level 2 is more worrisome. It includes some shadier characters, such as derivatives and mortgage-backed securities. There are no active, public markets for these investments—they are bought and sold privately, if at all, and are not listed on exchanges—so Wells Fargo uses other methods to figure out fair value, including what it calls “model-based valuation techniques, such as matrix pricing.” At Level 2, fair value is what accountants would charitably describe as an “estimate,” based on statistical computer models and what they call “observable” inputs, such as the prices of similar assets or other market data. At Level 2, fair value is more like an educated guess.

Many banks’ stocks are below “book value” today. This indicates that investors don’t believe the stated value of the assets on banks’ books, or don’t believe banks will be profitable in the future—or both.

Level 3 is hair-raising. The bank’s Level 3 estimates are “generated primarily from model-based techniques that use significant assumptions not observable in the market.” In other words, not only are there no data about the prices at which these types of assets have recently traded, but there are no observable data to inform the assumptions one might use to generate prices. Level 3 contains the most-esoteric financial instruments—including the credit-default swaps and synthetic collateralized debt obligations that became so popular and prevalent at the height of the housing boom, filling the balance sheets of Bear Stearns, Merrill Lynch, Citigroup, and many other banks.

At Level 3, fair value is a guess based on statistical models, but with inputs that are “not observable.” Instead of basing estimates on market data, banks use their own assumptions and internal information. At Level 3, fair value is an uneducated guess.

Surely, one would assume, Wells Fargo’s assets would mostly reside on Level 1, with perhaps a small amount on Level 2. It’s just a simple mortgage bank, right? And it seems inconceivable that Wells Fargo would be loaded with Level 3 investments long after regulators have supposedly purged the banks of toxic assets and nursed them back to health.

Yet only a small fraction of Wells Fargo’s assets are on Level 1. Most of what the bank holds is on Level 2. And a whopping $53 billion—equivalent to more than a third of the bank’s capital reserves—is on Level 3. All three categories include risky assets that might lose value in the future. But the additional concern with Level 2 and Level 3 assets is that banks might have errantly recorded them at values that were inflated to begin with. There is no way to check whether reported values are accurate; investors have to trust the bank’s managers and auditors. Scholarly research on Level 3 assets suggests that they can be misstated by as much as 15 percent at any given time, even if the market is stable. If Wells Fargo’s estimates are that far off, the bank could be sitting on billions of dollars of hidden losses.

Wells Fargo discloses in a quiet footnote in small print on page 133 of the annual report that its Level 3 assets include “collateralized loan obligations with both a cost basis and fair value of $8.1 billion, at December 31, 2011.” In English, that means that the bank is recording the value of some of its most complicated investments (composed of packages of loans to companies) at exactly the price it paid for them (the “cost basis”). Were these products bought a year ago? Two? Before the crash of 2008? Have they actually retained their value? Don Young finds it curious that the fair value and cost basis would be the same. “With interest rates much lower than most expected, why didn’t the CLOs rise in value?” he asks. But he’s the first to admit that he’s really in no position to say. Without more information about the composition of the loan packages and when they were purchased, an outsider cannot determine what these assets might be worth.

Accountants and regulators insist that categorizing an investment as Level 1, 2, or 3 is better than simply recording the investment’s original cost. But the current system permits bankers to use their own internally generated estimates. Who oversees those estimates? Auditors who are dependent on the bank for significant revenue, and regulators who are endemically behind the curve. Such a setup erodes trust. And when that trust disappears, so does any confidence in what the bank says its investments are worth.

The Level 3 issue isn’t simply theoretical. One major problem during the 2008 crisis was that banks and investors didn’t know what to trust about Level 3, so they panicked. We just suffered through a crisis in Level 3 assets. We can’t afford another.

Presented by

Jesse Eisinger is a senior reporter at ProPublica, an independent, nonprofit newsroom in New York City that produces investigative journalism. His email address is Jesse.eisinger@propublica.org.

Join the Discussion

After you comment, click Post. If you’re not already logged in you will be asked to log in or register with Disqus.

Please note that The Atlantic's account system is separate from our commenting system. To log in or register with The Atlantic, use the Sign In button at the top of every page.

blog comments powered by Disqus


The Horrors of Rat Hole Mining

"The river was our source of water. Now, the people won't touch it."


What's Your Favorite Slang Word?

From "swag" to "on fleek," tweens choose.


Cryotherapy's Dubious Appeal

James Hamblin tries a questionable medical treatment.


Confessions of Moms Around the World

In Europe, mothers get maternity leave, discounted daycare, and flexible working hours.


How Do Trees Know When It's Spring?

The science behind beautiful seasonal blooming

More in Business

More back issues, Sept 1995 to present.

Just In