Friday, September 19, 2014

Capital Language

Sometimes the press deserves a little applause. Peter Coy at Business Week and Pat Regnier at Time both wrote articles very nicely explaining bank capital and many fallacies around it.

Both articles also have nice graphics, but I give Coy and BusinessWeek the A+, because it also explains that banks can build capital without cutting lending.

A few select quotes. From Coy at Business Week, two fallacies skewered:
So what exactly is capital? Sometimes it’s described as a rainy-day fund, which is wrong. More often it’s characterized as something banks “hold,” which can make it sound like a pile of money that has to be set aside so it can’t be lent out for a profit. That’s not right either.
The American Bankers Association says that higher capital requirements for big banks “reduce economic and job growth.” But banks can meet capital requirements without cutting back lending. They just have to sell more shares (cutting down on buybacks also works) or reduce cash-draining dividends (refraining from raising them also helps).  
Regnier at Time too, and passes on the useful housing analogy.
...As Admati frequently points out, banks have benefited from the misconception that higher capital requirements means banks would have to keep 20% or 30% of their money locked up in a vault, instead of lending it out to businesses or homeowners.
In fact, making banks “hold more capital” actually means they have to borrow less. In their book, Admati and Hellwig show that this is almost exactly like a homeowner making sure to build up equity in her house. 
To raise more capital, banks wouldn’t hold back lending. Rather, they’d tap their shareholders, either by issuing new stock or just by cutting the dividends they pay out of earnings, letting profits build up on the balance sheet. 
It's refreshing when professional writers explain things a lot more clearly and succinctly than us academics seem to do, and get the economics spot on.  Yes, words, stories, and ideas do matter, and the change in attitude about bank capital is a great example.

HT to Anat Admati who sent me the links.


  1. "The American Bankers Association says that higher capital requirements for big banks “reduce economic and job growth.”"
    That's not truth but that's increased the cost of capital therefore there will people (low income people) with more dificults to borrow money.

  2. But the risk-weighting of bank assets causes a problem: when bank capital requirements are binding, the marginal dollar of equity goes toward loans that have a low risk weight - such as home loans, and sovereign debt. This means that small increases in capital requirements don't lead to more business loans. Julien Noizet has several posts (tagged with RWA) on his blog at explaining housing bubbles as an unintended consequence of the system of risk-weights.

    This is why I consider bank capital regulation a dead end. You've said that you want banks to internalize the costs of their distress. I think that the necessary step is to provide them with suitable hedging instruments. If these instruments eliminate the need for bailouts, then banks will choose capital levels closer to the social optimum, and there will be no unfortunate byproducts of poorly-conceived constraints.

    1. What “hedging instruments”? Far as I can see you are saying that banks should insure against the risks involved in business loans. OK, but that just raises the costs of business loans to much the same extent as where banks are required to hold more capital against business loans.

  3. How did Peter Coy get a picture of my bed?

  4. John,

    Good description by Peter and Pat.

    The hard part is maintaining that capital cushion in real time. Suppose a bank does option #2 - pays back debt, sells more equity. What stops that sale from driving down the market value of existing shares thru dilution and driving up the market value of existing debt through buy backs?

    For instance, bank retains the assets of $100 without cutting lending. Market value of banks outstanding debt is $90. Market value of banks outstanding equity is $10. Bank tries to buy back $10 worth of debt and sell $10 worth of equity.

    Market responds by bidding up the value of a bank's remaining debt to $90 and bids down the bank's total equity back to $10.

    Or would you prefer some non-market method of valuing a bank's debt and equity when determining when a bank is over / under capitalized?

    I think Anwer is right, but I don't agree on all the reasons - bank capital regulation is a dead end. He is right that proper hedging instruments are required - but those instruments should be structured to benefit both borrower and lender.

    If we are only concerned about the fortunes of lenders, then that answer is easy - government does all borrowing for an economy and pays financing costs with taxation.

  5. There’s actually a flaw in Peter McCoy’s argument and chart, but not a serious one. The flaw is thus.

    If banks are made to hold more capital, the TOTAL AMOUNT that individuals, households etc need to invest in capital in ALL INDUSTRIES for the country as a whole must rise (all else equal). That means the RETURN that households will demand for holding or investing in capital will rise. Thus banks will have to charge more for loans.

    However, contrary to the claims of bankster criminals, that will not reduce GDP: in fact it will INCREASE it. Reason is that when banks have a low capital ratio they are more likely to fail, thus they are implicitly relying on taxpayer funded subsidies or guarantees. And subsidies distort markets and reduce GDP. Thus a rise in capital requirements, while it will reduce loans and debts, will actually increase GDP.

    Bank capital ratios in the 1800s when taxpayer funded bail outs for banks were unheard of were often around 50%: way above current levels. That 50% is some indication of what the genuine free market ratio would be.

  6. Something nags at me about this.

    Since banks would not be obligated to pay dividends on common stock, you would think banks of their own volition would issue more common stock, and borrow less.

    This would make banks more sturdy, and give management a lot of flexibility in adjusting costs. A bad year, you cut the dividend. it the tax code? Borrowing is a deductible cost of business, while paying dividends comes out of profits?

    Is the solution a simple elimination of the interest deduction provisions of the tax code, and perhaps tax-free dividends? At least for banks?

    1. Ben,

      "Since banks would not be obligated to pay dividends on common stock, you would think banks of their own volition would issue more common stock, and borrow less."

      Unless of course the owner(s) of existing common stock may be the same guys making the financing decisions for the bank. Greed and short sightedness play no small part here. I think it was John Mack (formerly of Morgan Stanley) that suggested that bank CEO / officer compensation should be divided between both bank debt and equity options / or outright securities.

      "Is the solution a simple elimination of the interest deduction provisions of the tax code, and perhaps tax-free dividends? At least for banks?"

      If only banks are forbidden from deducting interest in filing a tax return, I think you would run into a problem trying to define what a bank is. Are the major industrial players (General Electric, Ford, etc.) that do vendor financing banks?

      If you totally eliminate the interest deduction provisions for all businesses, I think you would find that the economy would gravitate towards a higher reliance on government debt. One nice thing that the interest deduction does is that it helps to overcome the default risk premium built into the credit markets.

  7. In Europe the banks are apparently issuing CoCo bonds (Contingent Convertible). When everything is good, the bond is a bond and pays interest. On specified events happening, the bonds get swapped for shares. This would be a pre-agreed automatic "bail in" for those bonds. The regulators are, appropriately, treating those bonds as "equity".

    The fundamental problem with Anwar's call for suitable "hedging" instruments is: who would have deep enough capital to support such instruments? AIG went broke from selling default swaps.

    1. I agree that there is no institution big enough to insure the banks. It must be done by bank depositors. Currently we have banks trying to offer risk-free deposits, and that is a very strong constraint on risky lending activity, especially in economic downturns. Cochrane and others have proposed 100% equity banks, and suggested that this will not yield a shortage of money, but I think that they have not understood why debt is the universal form of money in our times. Charles Calomiris, a major contributor to the work on CoCos, does understand the monetary problems caused by increasing capital requirements. Anat Admati is very good at explaining the problems with CoCos.

    2. Anwer - the point of CoCos is to preserve the tax advantages of debt while achieving the regulatory advantages of equity. It would come down to the drafting of the contractual terms of the CoCo. CoCos are apparently being snapped up by the market at yields of 4% - that makes them much cheaper for banks than common or preferred equity.

      Small depositers making demand deposits should not be asked to insure the banks. Much better to have the bank issue CoCo bonds to willing and knowledgeable investors and have those investors insure the banks through that mechanism.

    3. Indexed debt would have the tax advantages of debt, some of the loss-absorbing ability of equity, and would be cheaper funding than CoCos. People accept very low yields on cash deposits because they are so liquid, and indexed debt would also be very liquid because it has similar informational properties. It would be even more liquid if the banking system could agree on a common benchmark for it, such as GDP - but that is a collective action problem.

      Perhaps you are right that small depositors lack the financial sophistication to make deposits that are slightly risky. They should use John Cochrane's proposed arrangement for holdings of risk-free government debt. But I am thinking about people already used to using shadow banks, mutual funds, etc. They understand why the balance fluctuates in their brokerage account, and would not panic if they saw small fluctuations in the account provided by their bank. In other words, a "national income" account through the bank would be good for people like you.

    4. "good for people like you" ????

      Anwar - I am a lawyer with 30 years experience in commercial law. Before that I studied physics and math.

      The problem of identifying, calculating and tracking the index swamps any benefit you might otherwise get from using indexing for liquid debt investments. Debt that is going to be counted as capital has to be sub-ordinate to other debts of the bank. One way to achieve that sub-ordination is to force conversion to shares in the event of financial difficulty.

    5. Yes, I'm saying that if you comment on a blog like this, you might be interested in a high-yielding account, with a complete understanding of the risks involved. The low returns on guaranteed deposits might be considered a tax on financial illiteracy. And it would be great to bring high-yielding accounts into the regulated system.

      I'm not saying that indexed debt would be counted as capital. Indeed we want it to be a senior claim because that makes it information-insensitive, and thus more liquid than residual claims. And yet the liability shrinks at the times that are most stressful to banks under the current banking model. Many people don't like CoCos because they are a nonlinear instrument, and there are studies that consider the bad effects of the trigger in times of crisis.


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.