This week Tim Geithner, US Treasury secretary, declared that the capital requirements in last year's Basel III deal must be raised for large banks. Three days earlier Daniel Tarullo, Federal Reserve Governor, went further, suggesting as much as a doubling of the Basel ratios for the very biggest banks. Meanwhile Michel Barnier, European Union financial services commissioner, takes the opposite position: “We can't rush forward,” he says. The radical reformers are lining up against gradualists. On any serious reading of the evidence, the radicals are right.
This fight is vital, because many other post-crisis financial reforms simply won't deliver much. Take regulators' new powers to wind up failing institutions and impose losses on bondholders – powers that are supposedly an alternative to taxpayer bail-outs. Such “resolution authority” counts for less than supervisors' will to use it. In 1998 the Fed had no formal authority to wind up the hedge fund Long-Term Capital Management, and yet it did so anyway; in 2008, the US government did have authority to resolve the failing bank Washington Mutual, but the risk of bond market contagion caused it to flinch. Besides, most significant financial institutions are international. The brave new national resolution mechanisms will not be up to handling the challenges that count.
The key bit is in the begining, where he notes that companies can find capital elsewhere. That is the key. Banks are intermediaries, but they aren't the only intermediaries and if their pricing becomes uncompetitive then investors and borrowers will go direct to the market (and banks will take stupid risks and employ regulatory arbitrage to fight their increasing irrelevancy- we already saw this leading up to 2008).
Do the math on a bank's capital structure. If equity is 20% (and investors demand 12% ROE) and l.t. debt is 20% (say roughly 4% currently for 10 yr AA bank debt), you have 40% of your liabilities costing you an average of 8%. Let's say the other 60% is comprised of deposits, s.t. borrowings and repo at an average of 1%. So your full capital structure right now costs you 3.8%.
On the asset side you'll need to keep about 20% in treasuries for liquidity and collateral purposes, paying around 1% (assuming 2-5yr avg life) and another 10% in s.t. instruments paying 1%. The other 70% of your book will be loans. Solve for the loan rate that makes it all work. I get 5%. How many companies are borrowing at 5% on a floating rate basis if short-term deposit rates are 1%?
If there is a limit on a bank's ability to charge below market rates on deposits and above market rates on loans, then the solution is for bank shareholders to require only, say a 6-8% ROE. More likely, bank shares will trade at a fraction of book value (so that an ROE of 12% incl. dividends is possible), while regulators will use actual book value to calcualte equity capital.
Would-be egg heads and central planners can devise all the "sweet spot" numbers they want, but we're dealing with a real market here involving investors/depositors and borrowers. Banks are not the market, they are only intermediaries. And a broker/intermediary who tries to charge too much goes out of business. Banking is more dangerous than broking- a broker selling at below costs feels the pain of loss on every transaction. A banker can live in hope that a bad loan will get paid off, the loss isn't realized for years. By making banks less competitive regulators ensure both more stupidity by banks and more rescues of mutual funds, SIVs, schemes and scams (as investors by pass banks and go "direct" to the risk assets).
Regulators need to take a holistic view of the whole market: investors, depositors, borrowers, VC and PE, banks, hedge funds, mutual funds, closed-end funds, SIVs, insurance, etc. Focusing on just one of the intermediaries (banks) and ensuring that they are "safe" (ha) but uncompetitive doesn't even come close to addressing the issue.
Or take another underwhelming innovation: the creation of systemic regulators to contain bubbles. Even conceding that these new supercops will know a bubble when they see one, it's not clear how they can help. Right now, for example, there may be a tech bubble. But if the authorities choked off credit to Silicon Valley, tech companies could probably raise money elsewhere – from deep-pocketed angel investors who operate below the regulatory radar, from venture capitalists in Hong Kong or London, from foreign sovereign wealth funds and so on.
Because these financial reforms will yield little, it is crucial to get capital requirements right. Last year's Basel III guidelines require banks to boost the ratio of equity to risk-weighted assets from 2 per cent to 7 per cent, an increase already large enough to cause banks to fight back. But the right target is much higher – indeed, higher than even Mr Tarullo has suggested. The sweet spot is somewhere between 15 and 20 per cent.
Three factors drive this estimate. First, recall how much equity can be destroyed in a crisis. The International Monetary Fund calculates that credit losses at US banks between 2007 and 2010 amounted to 7 per cent of assets, so banks must be in a position to lose that much again and survive. Second, consider how much residual equity banks must have left after a large hit. Here the answer is about 8 per cent of assets – that is the amount that the top four US banks felt it necessary to hold in early 2010 in order to retain market confidence. Third, remember that capital is held against risk-weighted assets, and that the calculation of risk weights is notoriously treacherous, so banks should hold a further buffer against “model error”, aka geeks who screw up. Adding these factors together, a 20 per cent equity capital ratio seems reasonable, even if some of this may take the form of “coco” bonds that convert to equity in a crisis.
Wait, say the gradualists; a 20 per cent ratio would hike lending costs dramatically, harming growth. But as basic theory tells us, the cost of equity falls as banks raise more of it, because the act of raising equity cuts risk. In a paper to be published in the Journal of Economic Perspectives, Samuel Hanson, Anil Kashyap and Jeremy Stein calculate that a substantial 10 percentage point increase in banks' equity capital would boost lending costs by only 25-35 basis points. Another paper by Stanford's Anat Admati and three co-authors dismisses the equity-is-expensive claim as “fallacious, irrelevant, or very weak”.
The bolder regulators accept these arguments. David Miles, an external member of the Bank of England monetary policy committee, has endorsed an equity ratio of 15 to 20 per cent. The Swiss authorities have already imposed a 19 per cent equity-plus-coco ratio on UBS and Credit Suisse. However many officials still shrink from going significantly beyond the Basel III guidelines. After all, bankers remain powerful. Voters are distracted. And regulators fear that squeezing banks aggressively will drive risk into shadow banks.
This last concern argues for more regulatory courage, not less. To prevent risk migration, the shadow banks (hedge funds, special investment vehicles, or whatever alias they go by) must be subjected to the equivalent of capital requirements. When they buy a securitised loan, they should pay cash for a decent chunk of it rather than financing the purchase almost entirely with debt. These sorts of “margin” rules already exist for equities and must be extended to the credit markets. The battle for financial stability is not over yet.
This article appears in full on CFR.org by permission of its original publisher. It was originally available here (Subscription required).