The largest U.S. banks keep rolling out new arguments to counter the growing impression -- among lawmakers especially -- that they present an unacceptable threat to the economy.
The latest iteration comes from six analysts at the Goldman Sachs Global Markets Institute in the form of a paper entitled "Measuring the TBTF effect on bond pricing." For the uninitiated, TBTF stands for too-big-to-fail, the idea that some banks are so systemically important that the government has no choice but to rescue them and their creditors if they get into trouble.
Various studies have shown -- and officials including Federal Reserve Chairman Ben S. Bernanke have agreed -- that too-big-to-fail status allows banks to borrow money at artificially low interest rates. Bloomberg View and others have argued that this taxpayer-backed subsidy, worth tens of billions of dollars a year, could be setting the stage for a financial disaster by encouraging banks to become as large and as systemically threatening as possible.
The Goldman paper seeks to challenge the too-big-to-fail logic by demonstrating that big banks don't enjoy a funding advantage compared to small banks. It also asserts that even if such an advantage exists, there are explanations other than a taxpayer subsidy.
Before getting into the details, it's important to note that the Goldman analysts are posing the question in an imprecise -- and perhaps convenient -- way. They're asking whether big banks borrow at lower rates than small banks. The salient question is whether big banks, thanks to government support, are borrowing at rates lower than they otherwise would.
That said, let's look at their points one by one.
1) The biggest banks don't have a funding advantage over small banks.
The Goldman analysts compare the yields on bonds issued by two groups of banks -- the six largest U.S. institutions, and a few dozen smaller ones. They find that the bigger institutions' cost of borrowing was, on average, 0.31 percentage point lower from 1999 through early 2013, but has lately been about 0.10 percentage point higher.
The analysis ignores a crucial distinction: The biggest banks are riskier because they use a lot more borrowed money, or leverage. The best way to see this is in banks' tangible common equity ratios, the percentage of their assets funded by shareholders, as opposed to creditors. It is a more reliable measure than the risk-weighted capital ratios favored by banks and their regulators.
The typical small bank has a tangible common equity ratio of about 9 percent, meaning it borrows about $10 for every $1 in equity its shareholders provide. The big banks' median ratio is less than 7 percent. If the largest banks' derivatives positions are included, as international accounting rules require, the median ratio would fall to 3.6 percent. A bank with such a low ratio could be rendered insolvent by a mere 3.6 percent drop in the value of its assets.
The biggest banks are riskier in other ways as well. Their operations and accounting, for example, have become so complex and opaque that their own managers, let alone sophisticated investors, are often at a loss to understand them.
What, then, does the Goldman analysis really say? At best, it tells us that a group of big, risky banks are able to borrow at the same rate as smaller, safer institutions. This suggests the market isn't fully charging the big banks for their riskiness. In other words, the analysis corroborates the too-big-to-fail advantage that it claims to disprove.
One simple way to get a sense of the big banks' funding advantage would be to find some similarly levered small banks and compare borrowing rates. This isn't easy: Markets generally don't allow small banks to use as much leverage as large banks do (still more evidence that the latter enjoy special status).
The Bloomberg terminal offered up seven that came close and had comparable bonds outstanding: M&T Bank Corp., Zions Bancorp, First Niagara Financial Group, City National Corp., First Horizon National Corp., Associated Banc-Corp and Susquehanna Bancshares Inc. On average, as of Monday, the small banks' bonds were paying 0.53 percentage point more than those of the six largest banks -- suggesting that big banks borrow at significantly lower rates than similarly levered small banks.
Even this might be an underestimate. The big banks are about twice as levered as our group of small banks when derivatives are included. As JPMorgan's $6.2 billion loss on its "London Whale" bets amply demonstrated, derivative positions matter in judging a bank's risk.
One note: It's possible that the Goldman analysis doesn't tell us much at all. The "small" bank group includes some not-so-small and not-so-representative companies, such as U.S. Bancorp ($355 billion in assets), student-loan specialist Sallie Mae and investment management firm BlackRock Inc., according to a spreadsheet provided by Goldman. Excluding them may or may not materially change the results. In any case, it wouldn't fix the flaws in the analysis.
2) To the extent that big banks have a funding advantage, it can be explained largely by the huge volume of bonds they issue.
It's true that sheer volume can be an advantage. When a company's bonds are plentiful, they are easy to buy and sell and hence more attractive to investors. The Goldman analysts estimate the value of this liquidity premium at 0.20 percentage point.
What, though, does the liquidity premium tell us about the advantage conferred by too-big-to-fail status? Nothing. Large borrowers would get it whether or not their creditors expected government support. So ultimately it's irrelevant.
3) Big companies have an even larger funding advantage over small companies in other industries. So there's no special reason to worry about big banks' funding advantage.
The Goldman analysts look at bond yields in a range of industries, from chemicals to travel. They find that companies in the top 10 percent by revenue commonly enjoy lower borrowing costs than those in the bottom 25 percent. This, they say, suggests that a funding advantage to size is totally normal, and not necessarily the result of government subsidies.
Aside from the fact that their analysis mixes methodologies -- top six by assets versus all the rest for banks, top 10 percent versus bottom 25 percent by revenue for other sectors -- it again ignores some important differences in risk.
In most industries, it stands to reason that the companies with the highest revenues would also tend to be the least risky, and hence would enjoy lower borrowing costs. They're probably more successful than their low-revenue counterparts, and they often achieve economies of scale that increase their profit margins, making it easier for them to afford their debt payments.
In banking, by contrast, the largest institutions are the riskiest. Their leverage -- about $30 for each $1 in equity -- is more than 50 times greater than that of the average U.S. company. They're typically not the industry's best performers, and their size goes far beyond what can be justified by economies of scale.
The latest research from the Bank of England suggests it's very hard to find any scale advantages for banks beyond $100 billion in assets. Under international accounting standards, the top six U.S. banks average $2.5 trillion in assets.
Hence, it's hard to understand the rates at which the largest banks borrow money -- unless creditors are assuming that taxpayers are bearing part of the risk.
4) Big banks deserve to borrow at lower rates, because their creditors almost never suffer a loss.
The Goldman analysts note that when big banks get into trouble, they almost never have to dip into Federal Deposit Insurance Corp. funds to compensate their depositors. This, they conclude, means that big banks are safer and hence worthy of lower borrowing rates.
Wait a minute. Isn't this actually an argument supporting the too-big-to-fail logic? The fact that big banks haven't incurred heavy losses for the FDIC serves only to demonstrate that the government can't allow such an extreme event to occur. During the 2008 crisis, government officials were afraid to impose losses even on big banks' unsecured bond holders, who are far below insured depositors in the pecking order of creditors. Small banks, by contrast, have made more use of FDIC deposit insurance precisely because the government can allow them to fail.
5) The government actually made money on its bailouts of big banks during the last financial crisis.
Looking at the investments the government made in the largest U.S. banks under the Troubled Asset Relief Program, the Goldman analysts find that the Treasury made a profit of $24 billion. This, they say, demonstrates that big banks aren't a threat to taxpayers.
OK, we've heard this one before. Measuring the return on bailouts is an absurdly narrow way to look at the cost of financial disaster. Distress at big banks tends to trigger broader crises with powerful economic repercussions. Government deficits and debt rise sharply as tax revenues decline and expenditures increase to ease the troubles of millions of unemployed.
The Bank of England has estimated the economic impact of systemic banking crises at 10 percent of gross domestic product, about a quarter of which is permanent. For the U.S. today, that would amount to more than $15 trillion.
The Goldman analysis simply sidesteps this issue, noting that it's not designed to address "the externalities of bank failures."
Momentum has been building in Congress behind efforts to eliminate the too-big-to-fail subsidy. Most notably, U.S. Senator Sherrod Brown, an Ohio Democrat, and U.S. Senator David Vitter, a Louisiana Republican, are proposing to sharply increase equity requirements for the largest banks. This would make them less likely to fail and encourage them to get smaller.
Bank executives seem to think the best way to prevent Congress from taking away their subsidies is to pretend they don't have any. Let's hope lawmakers aren't fooled.
This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.
To contact the author on this story:
Mark Whitehouse at firstname.lastname@example.org