One major proposal of 13 Bankers is setting caps on the size of financial institutions, with the goals of preventing the existence of banks that are too big to fail and of eliminating the competitive distortions created by megabanks. We propose that no institution should have assets greater than 4 percent of GDP, with lower limits for institutions with riskier positions, taking into account "derivatives, off-balance-sheet positions, and other factors that increase the damage a failing institution could cause to other financial institutions" (p. 215). Given the nature of the book, however, we did not attempt a detailed discussion of how to take these factors into account in setting size caps. So consider this a continuation of the book.
There are two, related dimensions to this problem. One is that assets, conventionally measured, do not accurately reflect a bank's total exposures (and, similarly, liabilities do not accurately reflect a bank's total obligations). The biggest reasons for these discrepancies are derivatives and off-balance-sheet entities (such as structured investment vehicles). The second dimension is that the systemic importance of a financial institution--the damage it could cause to the system by its failure--depends in part on whom its obligations are owed to; this is the simple meaning of the famous term "interconnectedness."
At a conference at Fordham Law School last month, I proposed a way of thinking about this problem, which I labeled the "blast radius" method. (Think about asteroids hurtling through space trying to figure out how big a hole in the Earth they would cause on impact.) The goal is to measure, for each institution, the potential damage that it could cause to the rest of the financial system.
Traditionally, bank regulation has focused on the asset side of the balance sheet, assessing the relative riskiness of different assets and converting them into "risk-weighted assets," which are then used as the basis for capital requirements. My proposal was similar, but applied to the liability side of the balance sheet: for each class of liability, determine how much investors or counterparties might stand to lose from the institution's failure. Of course, in order to do this assessment, it is necessary to look at both derivatives and off-balance sheet exposures, since they represent money that other institutions might be counting on--and might not get in a crisis.
In other words, traditional capital requirements attempt to minimize the chance that a bank will fail. The blast radius calculation -- and size caps -- attempt to minimize the chance that a failing bank will cause a mass extinction in the financial system.
Daniel Gros of the Centre for European Policy Studies has a helpful article on this topic at Vox. As it turns out, United States and European banks use different accounting standards when reporting their financial position; U.S. banks use GAAP while European banks use IFRS. These can produce wildly different results. For example, Deutsche Bank's assets are 1.0 trillion euros under GAAP but 2.2 trillion euros under IFRS.
This difference is predominantly caused by different ways of accounting for derivatives; to simplify things, GAAP allows a bank to net its positions against each other (if I have a $100 trade with one counterparty and an opposite $100 trade with another counterparty, that counts as zero assets and zero liabilities), while IFRS doesn't. Gros projects that Goldman Sachs, whose end-2008 balance sheet showed $900 billion in assets, might have had up to $4.6 trillion in assets under IFRS. Instead of a leverage ratio of 15, it would have had a leverage ratio of 72.
What does this mean? For starters, it means that we need to have a more accurate grip on just how big our largest financial institutions already are. Gros says that we need accounting standards that do not allow netting of derivatives, in order to show how much damage could be caused to the financial system in a crisis. We also need to have consistent standards across countries if we are to have any hope of harmonizing financial regulation internationally.
More generally, if we are going to restrict the size of our largest banks, we need to have a measure of size that captures all of their obligations to other institutions, and does so under crisis conditions, when counterparty risk is high. We don't claim to have the final word on how to value every financial position. But this is a tractable problem that smart people can get together to solve. The real question is whether we have the political will to rein in the megabanks.