(Bloomberg) The world’s largest banks are incorrectly accounting for their swaps trades, locking up money that could otherwise be paid out as dividends to their shareholders, according to a bold new academic paper.

The transactions are funded with money that’s borrowed from the bank’s treasury, and currently that lending cost is deducted from the value of the derivatives. That’s a mistake, according to Darrell Duffie, one of the report’s authors who argues banks should instead charge their trading partners more up-front for the deals, freeing up cash.

By calling out Wall Street accounting conventions, Duffie is in familiar territory. The Stanford University finance professor has already helped spur changes in how banks value credit risk and debt. This time, he’s targeted a business worth many trillions of dollars, where even tiny accounting shortcomings can add up to serious money. The financial health of banks has been a primary concern of regulators since the 2008 credit crisis forced governments around the world to bail out the world’s largest dealers.

Wall Street first revealed a switch to this accounting method for swaps—called funding valuation adjustment, or FVA—in 2011 when Goldman Sachs Group Inc. mentioned it in its annual report. In 2013, JPMorgan Chase & Co., the second-largest U.S. derivatives dealer, took a $1.5 billion loss on its fourth-quarter net revenue after it adopted FVA derivatives accounting.

‘Gone Wrong’
“The accounting has gone wrong,” said Duffie, a research associate at the National Bureau of Economic Research and a former member of the Federal Reserve Bank of New York’s financial advisory round table.

It took the report’s authors three years to understand and devise the new method, he added during an interview. “I know this sounds strange, but it’s taken people a long time to get this,” he said. “The model is irrefutable. If you follow the math you can’t reach any other conclusion.”

Duffie’s formulas boil down to this: Banks are masking the funding cost of their swaps desks by applying the cost—in the form of a discount—to the value of those positions that they report in their financial results every quarter. He’s not accusing the banks of any wrongdoing; he just thinks there’s a better way to book swaps activity.

To illustrate his point, Duffie points to a separate market: U.S. Treasuries. A bank holding $100 million of Treasuries wouldn’t value those securities at $99.5 million just because it had to borrow $500,000 to fund the purchase, Duffie said. Yet that’s exactly what they do in swaps.

‘No Sense’
That “clearly makes no sense,” because the Treasuries haven’t lost value, the authors said in the report. “Instead, this trade should cause an addition of $100 million to the disclosed value of the bank’s assets and a reduction of nearly $0.5 million to the value of the bank’s equity.”

Nitpicking over a $500,000 cost on a $100 million trade might seem inconsequential, but the market for interest-rate swaps, the largest in the world, is $319 trillion in size, according to the Bank for International Settlements. Even small errors matter.

To avoid this problem, banks should be pickier about what swaps they enter into and charge more for trades that require a lot of collateral or cash from the firm’s treasury, Duffie said. That would have the positive effect of lowering borrowing costs. “The banks will be worth more, so they’ll be able to issue debt at lower credit spreads,” Duffie said.

The bank switch to FVA came about for several reasons. In 2008, funding costs for bank swap deals saw “severe deviations” from the historic risk-free borrowing rate, according to the paper, which will be released publicly Friday. Also, large accounting firms “have signaled a willingness to accept FVA disclosures in dealers’ financial statement,” the authors said.

Duffie was one of several academics who near the turn of the century helped change how Wall Street banks account for credit-valuation adjustments and debt-value adjustments. The former are used to measure capital requirements on derivatives trades while the latter reflect the gains or losses a bank would take if it bought back its debt.

Duffie wrote the paper with Leif Andersen and Yang Song. Andersen is co-head of the global quantitative strategy group at Bank of America Corp. while Song is a graduate student in the Ph.D. program at Stanford’s business school.

In the new paper, the authors created a model they argue can be used by Wall Street banks to determine if a swap trade is beneficial or harmful from an equity shareholder point of view. While the issue of the FVA treatment by banks being wrong has been brought up by other researchers, this is the first time a model has been created to address it, they said.

Duffie said he hasn’t shown the paper to Wall Street executives yet.

“I imagine there’s going to be some howling by the banks because they think their accounting practices are right,” he said.

Register or login for access to this item and much more

All Accounting Today content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access