Global megabanks and their friends are pushing back hard against the idea that additional reforms are needed – beyond what is supposed to be implemented as part of the Dodd-Frank 2010 financial legislation. The latest salvo comes from Goldman Sachs which, in a recent report, “Measuring the TBTF effect on bond pricing,” denied there is any such thing as downside protection provided by the official sector to creditors of “too big to fail” financial conglomerates.
The Goldman document appears hot on the heels of similar arguments in papers by such organizations as Davis Polk (a leading law firm for big banks), the Bipartisan Policy Center (where the writing is done by a committee comprised mostly of people who work closely with big banks), and JP Morgan Chase (a big bank). This is not any kind of conspiracy but rather parallel messages expressed by people with convergent interests, perhaps with the thought that a steady drumbeat will help sway the consensus back towards the banks’ point of view. But the Goldman Sachs team actually concedes, point blank, that too big to fail does exist — punching a big hole in the case painstakingly built by its allies.
Of course, the report puts a different spin on matters. But if you add in a little bit of recent Goldman history – not mentioned in the report – the dangers of relying on their read of the data become readily apparent.
The Goldman analysts make clear that, even under the most favorable interpretation (i.e., theirs), very large financial institutions were able to borrow more cheaply than other financial firms at the height of the crisis in 2008-09. The funding advantage they measure in the crisis looks (from their charts) to be in the range of 400-800 basis points.
In fact, there was further advantage to being one the very largest firms – remember that when hedge funds “ran” from Morgan Stanley, their destination was JP Morgan Chase (this point is not in the Goldman report, which conflates these two firms with other very large financial institutions).
The Goldman team shows there was an even larger advantage for huge non-bank financial companies than there was for banks, but they neglect to mention that their company was one of our large non-banks as the crisis intensified. Goldman was allowed to convert to become a bank holding company in September 2008, so that it could access the Fed’s discount window (i.e., increase its ability to borrow from the central bank.) This conversion was allowed – or perhaps even urged by officials – precisely because they feared the consequences of Goldman failing.
Goldman executives argued long and hard in September 2008 that they were too big – and complex and generally important – to be allowed to fail. Hank Paulson, then Secretary of the Treasury and former head of Goldman, felt strongly that the continued existence of his firm was essential to the well-functioning of the world economy.
The measured difference in spreads is obviously huge but even greater is the real funding advantage between not being able to borrow from the Fed (think CIT group, $80 billion total assets, which foundered and begged for assistance in fall 2009) and being able to borrow from the Fed (Goldman Sachs, $1.1 trillion total assets when it hit the rocks in mid-September 2008).
CIT, of course, was not a bank – hence the argument that it should not be allowed to borrow from the Fed. But Goldman was not a bank either at the relevant point in time.
The issue of TBTF is not about the rules or the cost of credit relative to small banks in a period of calm. It is about the availability of government support, broadly defined, when bad things happens – enabling the megabanks to borrow more cheaply than would otherwise be the case.
Goldman Sachs has downside protection available to it which a small or even medium-sized regional bank cannot hope to access. This helps increase liquidity in their bonds and generally makes it easier to borrow in good times as well as bad.
The Goldman team argues that big banks show lower losses than smaller banks both in the recent credit cycle and during the S&L crisis. But this uses F.D.I.C data and it necessarily masks all the other support provided by Fed support and various kinds of debt guarantees. And Goldman does not cover the emerging market debt crisis of the early 1980s, which was all about big bank losses (with Citi at the center, as it was in 2007-08).
Goldman also argues that the funding advantage for megabanks today is smaller than it was in the crisis, indicating that there is no longer a TBTF issue in the minds of creditors.
But this rather indicates two points. First, we are not in a crisis – so the immediate cash value of government support is lower, at least for short-term debt. But the availability of that support in the future or in various unspecified difficult scenarios is hugely valuable, and a compelling reason to buy megabank debt. Second, megabanks pay a complexity or opaqueness premium on their debt. Creditors charge more than they would otherwise because they cannot see inside the largest of these groups to understand the risks they are taking and the shocks to which they are vulnerable.
The probability that the government (including the Fed) will protect all creditors fully has never been one and will never be zero. Even at the height of the financial crisis, the credit default swap spread for some large financial institutions was high (e.g., for Morgan Stanley, so insuring its debt against default was expensive).
Creditors weigh the odds. The spread on megabank debt over benchmarks is the combination of downside guarantees (which tend to lower spreads) and the complexity premium (which tends to increase spreads).
The real issue for too big to fail is: by how much does the prospect of government support lower spreads compared to what they would otherwise be. The Goldman report acknowledges that too big to fail exists and distorts the market, but conveniently ignores the question of how big this distortion is really – and how it threatens to again bring down the economy.
This piece is cross-posted from Baseline Scenario with permission.