One of the lessons that many people have taken from the financial crisis is that there was too much debt and leverage and too little equity capital.  Capital serves as a buffer against loss.  If this buffer is inadequate, the probability of financial failure is high.  During an economic boom, borrowers are eager to take on more debt so as to “lever” their returns higher, and lenders are willing and eager to accommodate them.   

Bank regulators are demanding greater capital levels.  The primary regulatory capital measure is Tier 1 capital to risk weighted assets.  The minimum standard for this measure formerly was 5% for well-capitalized banks, and is now in the process of moving higher to 7%.  In addition, the largest banks, those deemed systemically risky, are required to hold additional capital.

Meanwhile, banks are arguing that moderate increases in capital levels are acceptable, but large increases are not feasible and in fact are counterproductive.  The effect will be, they argue, to increase the cost of capital for banks and will reduce lending.  As a matter of arithmetic, higher leverage means higher return on equity (ROE), at least it does so long as the investment yield is greater than the cost of funds.  Banks generally operate with a fairly small spread between the yield on earning assets and the cost of deposits and borrowings, so bank ROE would be quite modest absent leverage. 

Former Federal Reserve Chairman Alan Greenspan agrees with the banks that there is an upward bound on capital requirements.  Here is his argument:  There is a minimum return that is required to induce investors to invest in banks.  This can be measured by the minimum historical ROE of 5%.  Furthermore, bank profitability can be measured by the historical average ROA of 75 basis points.  Putting these two factors together, the former Chairman concludes that maximum sustainable bank capital levels are 15% (.0075/.05) of assets.

I think the former Chairman is missing the fundamental point that the ROA will vary with leverage, in particular, as the ratio of equity to assets increases, ROA will increase as well.  Also, the minimum ROE will vary with the riskiness of the bank.  As the ratio of equity to assets increases, the minimum acceptable ROE will be lower.  So, his conclusion that 15% is the maximum feasible capital ratio is dubious.

In their new book “The Bankers’ New Clothes” economists Admati and Hellwig (AH)1 directly address and dismantle banker arguments about the negative consequences of higher bank equity levels.  In particular, AH address what they refer to as the main banker myths:  first, higher bank capital level will mean less lending and less economic growth, and second, equity is expensive.  AH argue that capital (equity) is not a cash reserve and has nothing to do with bank lending.  Capital requirements only affect the funding mix between debt and equity, and should have no bearing on how the assets are deployed.  In order to increase capital levels, banks can issue new stock or can reduce dividends and stock buybacks.  The effect will be to increase asset levels (for a given level of debt) thus enabling greater lending, not less.

Regarding the cost of equity, or the required return on equity, AH point out that the cost of equity is an increasing function of leverage, and that the weighted average cost of equity and debt capital is determined by the riskiness of the bank’s assets, not by the mix between equity and debt.  In short, reduced leverage will reduce the required return on equity.

AH argue that the appeal of debt (including deposits and borrowings) to bankers primarily stems from subsidies including deposit insurance and the “too big to fail” (TBTF) policy.  These policies artificially lower the cost of debt to large banks by eliminating depositor or bank debt investor concerns about being paid back.  The consequences of these policies include increased fragility in the financial system and susceptibility to financial crises. 

AH favor large increases in minimum bank capital levels (to 20% or 30% of assets).  This will improve financial stability and long-term economic growth.  The mechanism they propose for achieving this is to require banks to stop paying dividends or buying back stock until bank capital levels have increased to the new higher minimum levels.

One counter-argument

Douglas Elliott2 of the Brookings Institute (and former a bank analyst for JP Morgan) believes we need to be more careful in pushing for much higher bank capital.  He believes the effect of higher capital requirements will be to raise loan rates and slow economic growth.  He points out that while the Modigiliani-Miller (MM) theorem about the irrelevance of capital structure is correct under their rarified assumptions, it does not hold in the real world of taxes and deposit insurance.   The effect of tax deductibility of interest payments means that debt is subsidized and is lower cost than in the MM world.  This effect is doubled in the case of banks due to deposit insurance (not to mention “too big to fail” policies).  Thus, as leverage increases the cost of debt does not increase thanks to the guarantees.

It is possible that AH would accept this point, but still argue for much higher capital requirements because the consequence of high bank leverage is greater financial fragility.

How would this work out?  Let’s take an example

Let’s explore further the Elliott argument that higher bank capital means higher loan rates.  Suppose we start a bank with capital of $1 billion and we decide to maintain the equity to asset ratio at 50%, as was common in the 19th century (before deposit insurance).  So we set out to raise deposits of $1 billion.  In today’s market, we only have to pay about 1% interest rate to accomplish this.  So we have interest expense of $10 million per year and annual operating expenses of about 1% of deposits or $10 million.  We now have $2 billion of funds to deploy ($1 billion capital and $1 billion deposits).  Suppose we buy or originate $2.0 billion jumbo mortgage loans and hold them for investment (for this exercise we ignore the bank’s need to hold cash reserves).  The going rate today would be about 4% giving us $80 million of interest income.  Suppose lending and deposit fees and provisions for loan losses exactly offset each other.

Our bank appears to be quite profitable:  net interest income of $70 million less operating expenses of $10 million leaves a pre-tax income of $60 million and after-tax income of approximately $40 million.  This is a return on assets (ROA) of 2% or 200 basis points.  But it is a return on equity (ROE) of just 4%.  This seems pretty meager, and unlikely to be adequate to draw investors.  However, the professors say that the required return on equity will be low because the riskiness of this bank is low.

How fast can our bank grow?  This depends on our dividend payout strategy.  Suppose we pay out 50% of income, or $20 million per year.  This leaves retained earnings of $20 million and represents annual growth in capital of 2%.  So, how are our investors doing?  They have an asset with a 2% dividend yield and earnings that are growing at 2% per year.

What is the value of our bank?  Well, it depends on the cost of equity capital. In view of the relatively low risk profile of the bank, the cost of equity will be pretty low.  How low can it be?  Surely, it must be greater than the mortgage interest rate of 4%.  Assuming 4% cost of equity and applying the Gordon Dividend Discount model, the fair value of equity is next year’s dividend ($20.4 million) divided by .02 (the difference between 4% cost of equity and 2% growth rate).  This formula yields a value of $1.02 billion, equal to the book value of equity.  At any higher cost of equity capital the value of the bank’s equity would be below book value.  So, surely new investors will not be drawn to this investment, and if this bank did exist, the best course of action would be to wind down operations and return funds to investors.

This is an extreme case, with an extremely high ratio of equity to assets.   But our profitability assumptions have been pretty aggressive as well, with no non-earning assets and only $10 million of operating expenses.  It does appear that mortgage rates would have to be higher to make our bank attractive to investors.  Or, our bank would have to drop the “portfolio lending” strategy and focus on originating loans for sale (the “originate to distribute” model), or find some other source of fee income. 

1Admati and Hellwig, The Bankers’ New Clothes, 2013.

2Douglas Elliott, “Excessive Bank Equity Rules Would Slow the Economy,” Brookings Institution, 2013.