Sitting before Congress a few weeks ago, Goldman Sachs (NYSE: GS) CFO David Viniar said, "Based on the nature of markets ... we know that we will sometimes incur losses."

Sometimes. In the first quarter, Goldman, JPMorgan Chase (NYSE: JPM), Citigroup (NYSE: C), and Bank of America (NYSE: BAC) all scored perfect trading results, making money from their trading divisions every single day. And not just a little money, but piles of it. JPMorgan made an average of $118 million per day, while Goldman made over $100 million on 35 days.

Perfectly suspicious
This might not sound strange. Businesses are supposed to make money, after all. But trading is a horrendously competitive and fickle business. Unless you're acting on inside information or have Miss Cleo on your payroll, trading results should look something like a bell curve: A few days produce large losses, a few days produce large gains, and the bulk of days produce something in between. And that's if you're good.

That's exactly what trading results used to look like. Take this chart, made with data from Goldman's annual report, which shows how many days Goldman's trading division made specific amounts of money in 2003:












Source: Goldman Sachs annual report.

Again, it looks like a bell curve. Some big losses. Some big gains. Lots of in between. That's what you'd expect. Now compare that with 2009's breakdown of daily trading results:












Source: Goldman Sachs annual report.

No more bell curve. In 2009, Goldman's trading division made boatloads of money on most days, lots of money on many days, and ... that's about it. Raging success became the norm. In six years flat, the concept of "risk" was seemingly vaporized.

Why? What changed? It's hard to tell because Goldman is unwavering in its quest to keep trading information secret. Some of these results are from old-fashioned client-driven activity where Goldman serves as a market maker. This is inherently safe activity that could produce steady results. But the other part of trading is proprietary activity, where Goldman acts like a hedge fund and trades on its own behalf. That's the risky and controversial side. Details on each segment aren't disclosed.

Goldman has developed a cute habit of beating around the bush when investors and reporters ask about these details. Asked whether the recent trading success was from client-driven or proprietary trading, COO Gary Cohn replied, "Over the last 12 months we have only recorded 11 loss days. It is implausible that a proprietary-driven business model could be right 96 percent of the time."

Well, no it's not, and I'll tell you why. What has changed recently, and what has been a tailwind to banks' proprietary trading operations, are 1) the concept of "too big to fail," and 2) investment banks becoming bank holding companies (as Goldman did in the Fall of 2008), which provides access to the Federal Reserve's safety net.

Money for nothin'
Both of these fairly new developments result in artificially low borrowing costs. Basically, lenders are willing to lend banks money at ridiculously cheap rates because they know the government will bail them out when things go bad. Examples of this happening can be downright amusing. As Rolling Stone columnist Tim Dickinson recently pointed out, "In March ... the rating agency Moody's (NYSE: MCO) disclosed that it has upgraded Citi's debt solely because it believes the government will step in to prevent default."

Banks can then use that artificially cheap money to fund their trading divisions. Artificially cheap money leads to artificially high profits, which likely explains why trading has suddenly turned into a foolproof mint. When you can borrow at nearly 0%, losing money becomes a challenge.

And borrowing at nearly 0% isn't exaggerating. Last month, my colleague Alex Dumortier laid out the big banks' borrowing costs from Federal Reserve loans and repurchase agreements. JPMorgan's average borrowing cost on this debt is just 0.08%. Goldman's is 0.45%. Bank of America's is 0.52%. In 2009, Goldman's weighted average cost of unsecured long-term debt was 1.42%. Coca-Cola's (NYSE: KO), for comparison, was 5.3%. This is an enormous subsidy to banks, and it's no wonder they suddenly have money blowing out of their ears.

What you can do about it
The absurdity of banks using the Federal Reserve to borrow money on the cheap for trading purposes, rather than lending to the broader economy, is the foundation of the so-called Volcker Rule proposed earlier this year, named after former Fed chairman Paul Volcker. Under the Volcker Rule, banks tied to the Fed would be banned from proprietary trading. Period.

Congress votes sometime over the next few days on whether to include a version of the Volcker Rule in forthcoming financial regulation bills. If you want to let your Senator know how you feel on this issue, click here for their contact information.