By protecting banks from credit risk the bankruptcy amendments have the perverse effect of undermining the standard means by which banks protect themselves from risk. The two most important tools traditionally used by banks to protect themselves are (i) a strong capital base to protect the bank from bankruptcy in the event of unexpected losses and (ii) the careful management of credit risk. The safe harbor exemptions increase systemic risk by encouraging banks to view themselves as protected in the event of default – this encourages them to greatly increase their exposure to counterparty credit risk and to believe that unexpected losses are unlikely so that they do not need much capital and can operate with a high degree of leverage.
When the repo amendment was passed, it was precisely because dealers were highly leveraged that a bankruptcy could cause a chain of failures and disrupt the market. Thus it is possible to view development of the repo market in the early 80s critically: the development of a wholly unregulated financial market had led to excessive leverage and the risk of financial instability. The market was in need of being protected by regulation in the form of capital and liquidity requirements. Imposing such requirements on the repo market would, however, have restrained its growth and further increased capital requirements for banks that were already struggling. There is no evidence that this policy was ever considered.
Instead, the 1984 repo amendment exempted repurchase agreements on Treasuries, Agencies and certificates of deposit from standard bankruptcy procedures. By guaranteeing that these repos could not be tied up in bankruptcy the law reduced the risks faced by dealers who worked with small liquidity and capital margins and thus encouraged them to operate on a highly leveraged basis.
After the 2005 bankruptcy law was passed granting all repurchase agreements safe harbor, the leverage created by repurchase agreements soared. Charles Munger describes the situation in an interview:
Our regulators allowed the proprietary trading departments at investment banks to become hedge funds in disguise, using the “repo” system—one of the most extreme credit-granting systems ever devised. The amount of leverage was utterly awesome. The investment banks, to protect themselves, controlled, to some extent, the use of credit by customers that were hedge funds. But the internal hedge funds, owned by the investment banks, were subject to no effective credit control at all.[1]
Effectively after 2005, all securities became liquid assets. Gary Gorton reports that up through April 2007 most repurchase agreements were not subject to any haircut at all, that is, they were for the full market value of the underlying collateral.[2] A dealer who owned – or borrowed – an investment grade security could use it to raise 100% of its value. In short, every dealer bank had access to unlimited leverage, not only when investing in safe assets like Treasuries, but even when investing in bonds that were just one step away from junk. They had speculative margin accounts with no margin. This was only possible after the passage of the 2005 Bankruptcy Act.
In many ways what happened to the dealer banks in 2008 repeated precisely what had happened to stock market investors in 1929. Asset values fell at the same time that margin requirements rose. As borrowing levels became constrained, liquidity dried up and forced asset sales set off a vicious cycle.[3] The IMF explains that this dynamic played an important role in the failures of both Bear Stearns and Lehman Brothers.[4]
The safe harbor amendments not only encouraged financial institutions to become dangerously leveraged, they also encouraged them to become dangerously exposed to each other. After the 1984 repo amendment was passed, the primary dealers’ financing in the repo market grew from $133 billion in 1984 to $834 billion ten years later, clearly increasing the exposure of the largest dealers to the market.[5] These trends continued at least up through 2007.
It is astounding to note that in 1999 the President’s Working Group recognized the fact that expansions of safe harbor had a tendency to encourage the market to grow bigger, but chose to focus on the benefits of “liquidity” and to ignore the possibility that increasing the exposure of market participants to counterparties could adversely affect systemic risk. They write:
The ability to net may also contribute to market liquidity by permitting more activity between counterparties within prudent credit limits. This added liquidity can be important in minimizing market disruptions due to the failure of a market participant.[6]
In short, the President’s Working Group assumed that market participants would keep exposures to a prudent level once safe harbor was in place, even though they had not done so in the absence of safe harbor protection. After all, if prudent levels of exposure had been the norm before the bankruptcy exemptions were enacted, then the robust early growth of the repo and swaps markets demonstrated that these contracts did not need special treatment in bankruptcy to flourish. The fact that these markets needed protection after they had already grown large was evidence that banks were not keeping credit within prudent limits before the passage of the laws.
In retrospect it is easy to see that neither repurchase agreements, nor certain derivatives were kept within prudent credit limits. Bear Stearns was a beneficiary of the growth of repo markets. The company’s quarterly report ending February 28, 2008, just days before it failed, states that total assets were $399 billion – and that $303 billion had been pledged to Bear as collateral. Of this amount, Bear had “repledged or otherwise used” $211 billion. In short, half of Bear Stearns’ balance sheet was financed using repurchase agreements. That this lending was far from prudent was made clear in the second week of March when Bear was at the very edge of bankruptcy and was saved only by a shot-gun marriage with J.P. Morgan Chase and the assumption by the Federal Reserve of $29 billion worth of asset risk.
AIG is another example of a firm that failed because its counterparties were willing to take on excessive credit risk. In the case of AIG the problem was credit default swaps. In September 2008, AIG found that it could not afford to post the collateral required by its derivative contracts. This failure to honor its contracts would have driven it into bankruptcy, had the government not stepped into the breach. Given the Bear Stearns and AIG examples one can only conclude that, when the bankruptcy exemptions increased market liquidity, that increase was not tempered by the prudence of the counterparties.
The failures of Bear Stearns and AIG may have been exacerbated by the fact that the bankruptcy exemptions encourage counterparties to demand overcollateralization of their positions. As the President’s Working Group observed in reference to the Long Term Capital Management (LTCM) collapse:
If its collateral holdings did not reflect potential future exposure, then a firm selling collateral provided by LTCM in the event of a default would still have been exposed to the difference between the value of the collateral and the value of the closed-out financial contract at the time the collateral was sold.[7]
In fact one of the recommendations of the report on LTCM is that banks should consider requiring that “potential future exposure” be collateralized – or in other words that current positions be overcollateralized.[8]
To understand the problem of overcollateralization it is important to recognize how profoundly bankruptcy law is weakened by the safe harbor exemptions. Standard bankruptcy procedure guarantees that a secured lender gets the lesser of the amount due on the loan or the value of the collateral plus an unsecured claim for the remainder of the loan. If a loan is overcollateralized the bankruptcy trustee has the right to reclaim the excess amount for the benefit of the other creditors.[9] By contrast, safe harbor protections allow the lender to seize collateral – and it’s not obvious that the trustee is in a position to determine whether or not excess collateral was posted. While this already undercuts the very principles of bankruptcy, the safe harbor privileges go further: even if the transfer of collateral was fraudulent, as long as it was received in good faith, the other creditors in a bankruptcy action have no rights to it.[10]
In short the safe harbor laws are deliberately structured to encourage counterparties to demand that their positions be overcollateralized – especially when the danger is growing that a firm will in fact declare bankruptcy. The legal protection of overcollateralized derivative positions creates two problems: first, it interferes with the principles of fairness that bankruptcy laws are designed to protect and, second, it may have systemic implications.
Consider what happened in each case when Bear Stearns, Lehman Brothers and AIG were about to fail. Repo counterparties refused to roll over the contracts or derivative counterparties sought additional collateral. It is precisely because of the strong protections for collateral that counterparties are so aggressive in their demands. According to lawyers interviewed by the Financial Times “under the old rules, creditors of companies facing financial difficulties were wary of settling trades or seeking extra collateral because they knew such demands could precipitate a bankruptcy filing and potentially freeze their claims.”[11] In the modern financial regime, the demands of counterparties ensure that as soon as bankruptcy is suspected, failure becomes a certainty: no counterparty wants to be the one that didn’t demand collateral or withdraw its repurchase agreements fast enough. Without the extraordinary privileges granted by the safe harbor protections, counterparties would be much more reluctant to force a firm into bankruptcy.
[1] Stanford Lawyer, Spring 2009, p. 17 http://www.law.stanford.edu/publications/stanford_lawyer/issues/80/pdfs/sl80_munger.pdf
[2] Gary Gorton, Dec 31 2008, “Information Liquidity and the (Ongoing) Panic of 2007,” p. 10. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1324195
[3] Gorton, 2009, pp. 33 – 38 has a thorough description of this process in the 2008 repo market. For the 1929 crisis see J.K. Galbraith’s The Great Crash.
[4] IMF, “Assessing risks to global financial stability,” Global Financial Stability Report, Oct 2008, p. 25. http://www.imf.org/external/pubs/ft/gfsr/2008/02/pdf/chap1.pdf
[5] Data from Federal Reserve Bulletin. Available at http://fraser.stlouisfed.org/publications/frb/page/31488 The number of dealers doubled from the 1960s to the late 80s and declined from the late 80s onward. http://www.newyorkfed.org/aboutthefed/fedpoint/fed02.html
[6] PWG, p. 20.
[7] President’s Working Group on Fin. Mkts., Hedge Funds, Leverage, and the Lessons of Long-Term Capital Management (1999) http://www.treas.gov/press/releases/reports/hedgfund.pdf, p. 21.
[8] PWG, p. 34
[9] Kettering, p. 1599-1600.
[10] PWG p. E4.
[11] Francesco Guerrera, Nicole Bullock and Julie MacIntosh, Oct 30 2008, “Wall Street ‘made rod for own back’” Financial Times.
Synthetic Assets 