Bank resilience: yet another missed opportunity

Along with enormous misery, the financial crisis brought an opportunity for long-needed reform. At the top of the list was the clear need for more bank capital. To ensure resilience of the financial system, and protect the public purse, banks’ owners had to have much more skin in the game. That is, potential losses to equity holders had to go way up.

Unfortunately, the 2010 Basel III agreement missed this rare opportunity to make the financial system safe. And now, with the publication of the standards for what has come to be known as total loss-absorbing capacity (TLAC), the disappointment continues to grow. To understand why, we need to step back and address the big question for bank capital: how much is enough?

Here is a simple calculation that can serve as a guide. Data compiled by Laeven and Valencia imply that if banks can absorb the losses arising from nonperforming loans equal to 20% of their assets (see Table Line [1]), 85% of post-1970 crises in OECD countries would have been avoided. (This would reduce the frequency of crisis in a country from an average of once every 20 years to once every 100 years.) Assuming that loan losses given default are 50% (Table Line [2]) and loss provisions are 1½% of total assets (Table Line [4]), this means that banks will need loss-absorbing capital equal to 8½% of assets (Table Line [5]). Adding a further margin-of-safety buffer of 1% (Table Line [6]) leads to a capital need of 9½% of total assets (Table Line [7]). This does not yet account for the riskiness of the assets themselves. In the United States, for example, total assets are roughly twice risk-weighted assets (Table [Line 8]).

How much bank capital is needed (as a share of risk-weighted assets) to reduce crisis risk?

Source: These computations are based on a presentation by Luc Laeven that used the IMF Systemic Banking Crises Database.

Source: These computations are based on a presentation by Luc Laeven that used the IMF Systemic Banking Crises Database.

The take-away from this simple exercise is that reducing the frequency of crises to a tolerable level means requiring banks to hold capital equal to nearly 20% of their risk-weighted assets. This is twice the standard in Basel III for global systemically important banks. And, according to the most recent Basel III quantitative impact study, at the end of 2014, the world’s 100 largest banks had capital equal to 11% of risk-weighted assets and an unweighted leverage ratio of only about 5%. (Keep in mind that Admati and Hellwig make a case for requiring an unweighted leverage ratio of at least 20% of total assets, or roughly double Line [7] in the table and four times the actual level.)

Now, there have been arguments that increasing capital is costly not only to banks but to society. We don’t doubt that there are private costs to banks, but as we discussed in an earlier post, there is virtually no evidence that the equity increases we have seen—bank capital has more than doubled since end-2009—have undermined the U.S. supply of bank credit. Instead, as U.S. banks have accumulated more capital, they have increased their lending. The same pattern of improved credit supply is evident among euro-area banks that experienced a windfall from the rebound of sovereign bond prices after 2012 (see, for example, here). That is, strong banks lend, weak banks don’t. Society benefits from well-capitalized banks because they are more able and willing to play their intermediation role come rain or shine.

So, it would have been a good start had the Basel III capital requirements been set at 20% of risk-weighted assets – double where they are. That would have corresponded to a simple leverage ratio requirement of something closer to 10% rather than the 3% minimum of Basel III. However, given that regulators failed to do this in 2010, when the financial crisis was fresh, it is unlikely to occur anytime soon. The reason is bank owners and managers love leverage (see our post here). And bankers are powerful, effective lobbyists.

To digress only slightly, the political economy of financial regulatory reform resembles the political economy of free trade. It is difficult (if not impossible) to implement economic policies that improve social welfare on average when the benefits are diffuse across the entire population, while the costs are borne by a relatively small well-coordinated group capable of influencing the political process. In the case of trade, opening the economy to imports means that uncompetitive domestic producers will lose. While society as a whole benefits from lower priced goods and services—with most people enjoying small gains—the owners of less productive firms and their employees face concentrated losses. And, despite the overall net gains to society, it is very difficult to implement an acceptable scheme to compensate the losers. As a result, incumbents (both owners and workers) often lobby for, and often get, trade protection that diminishes national income.

How does this work in finance? Here, higher equity capital levels mean that banks lose their implicit government guarantee. They also lose the tax deductibility of debt service. And, because a substantial portion of the benefits of equity issuance by highly leveraged entities (like banks!) flows to debtholders (who gain a bigger cushion against losses they would face in a bankruptcy), the owners of leveraged firms view equity issuance as a costly dilution. (In corporate finance, this agency problem is so well studied that it has a name: debt overhang.)

So, if banks can get away with low levels of equity—and instead, rely principally on debt to finance their activities—the owners and managers get the upside of the risks they take, while taxpayers bear the big down side. Collectively, banks form a small, well-heeled, organized group. Their typical claim is that higher capital levels will harm the economy and shift financial activity abroad. (A great example of this was the debate over the impact of Basel III summarized here.)

Now, financial regulators and supervisors are not naïve. Many of them understand that the current level of bank equity is insufficient to protect the system from frequent crisis. So, having lost the Basel III battle to make equity requirements suitably robust, they have sought to convert some portion of bank debt into a loss absorber in a crisis. The logic is that banks will be less resistant because they view debt as cheaper to issue than equity.

This brings us to TLAC—Total Loss-Absorbing Capital. Over the past month, the Financial Stability Board issued its TLAC standards and the Federal Reserve Board put out domestic U.S. regulations for comment. As we described in some detail a year ago, the idea behind TLAC is that, in addition to equity capital, large banks will be required to issue long-term debt liabilities that can be converted into equity to preserve financial stability without government support if the bank experiences losses that threaten its solvency. When initially proposed, the minimum TLAC level was supposed to be 16% to 20% of risk-weighted assets, excluding additional equity buffers of up to 5% for the largest global systemic intermediaries. The combination of equity and TLAC debt thus appeared to increase the losses that a bank can absorb without recourse to public assistance by as much as 2½ times compared to the 2010 Basel III standard.

Ignoring the daunting problems of actually using the TLAC debt—whether it will be possible to convert debt to equity without a bankruptcy/resolution proceeding—these numbers didn’t look all that bad. Or, as an optimist might think, it could have been worse.

Well, now it is. The regulators have watered down the new requirement in two ways. First, the minimum TLAC requirement has been set at 16% as of 2019 and 18% as of 2022.  Second, and more disturbingly, banks will be allowed to hold TLAC-eligible debt issued by other banks in an amount up to 10% of their own equity. The effect will be to reduce the effective loss-absorbing equity capital of the most systemic banks by up to 10%. Put differently, a bank with a reported leverage ratio of 5% may have an actual loss-absorbing ratio in a systemic crisis of only 4½%.

This stealth weakening of the system is particularly worrisome because it opens the door to further opaque forms of dilution that diminish the credibility of the system’s reported capital. Suppose, for example, that banks assert in coming years that the new TLAC debt is still too costly to issue because there are too few buyers. Will regulators raise the amount of TLAC debt that banks can purchase from other banks to 20% of their capital? 50%? 100%? Where do they draw the line? What about purchases by leveraged nonbank intermediaries? Will be they be unconstrained?

In our view, the stability of a financial system requires that banks be barred from holding one another’s loss-absorbing liabilities in the same way that U.S. banks today cannot purchase bank equity (see, for example, here). When cross-holdings of this type are allowed, stress will propagate quickly from one institution to the next through links that are unobservable by outsiders. Allowing banks to hold TLAC debt undermines the credibility of their capital buffers and creates a source of contagion.

Indeed, holdings of TLAC by any leveraged intermediary can make the financial system fragile and merit heightened regulatory scrutiny. Imagine, for example, that banks provide a large volume of credit to non-bank leveraged entities such as hedge funds, who then purchase the TLAC of another bank. Because such indirect exposures are difficult to detect, the potential for them to arise undermines confidence in the resilience of the system.

Keep in mind also that, under the Dodd-Frank Act, the FDIC plans to require U.S. bank holding companies to issue a sufficient quantity of long-term debt that can be converted into equity in the event of resolution of a systemic bank. Once again, if the FDIC utilizes the watered-down TLAC definition for the long-term debt eligible under its “single-point-of-entry” resolution scheme, then the remedy for one systemic institution in distress will weaken others (see here).

So, we are disappointed and troubled. A much larger share of bank assets should be funded by equity than Basel III provides. The 2014 TLAC proposal already was a compromise. Now it has been further watered down in a way that conceals leverage and creates scope for further mischief. From a political economy perspective, this may be unsurprising. For those who understand the cost of financial crises, it is deeply disturbing.