Higher Bank Capital Requirements Would Come at a Price

A dangerous misconception appears to be taking root in the public debate about bank safety. A belief is growing that banks could be made much safer, at essentially no economic cost, by requiring shareholders to supply far more of the funding for banks with correspondingly less coming from debtholders and depositors. In fact, there would be significant economic costs, so there needs to be a debate centered on an examination of the trade-offs. Personally, I agree with the majority of analysts and policymakers that the costs would outweigh the benefits, but my key point here is that we need a debate on the trade-offs, wherever we come out on them.

The arguments start with a sound theoretical base, but important caveats and practical problems are dropped from the discussion somewhere in the transmission chain from the more careful academic studies to the popular discourse. This matters, because many of the simplistic proposals being aired would reduce lending and make what remains substantially more expensive. The recent severe recession is a reminder of how much damage a credit crunch can do, so we ought not to inflict one on ourselves voluntarily.

The proposals call for much greater levels of bank capital, mostly in the form of “shareholder equity”, which comes from the sale of common shares to investors in combination with bank profits that accumulate over time. Currently, common shareholders supply roughly 5% of the funding for most banks, while the proposals often call for increasing this up to 30%. A key attraction is that proponents frequently argue that this increase in capital is costless or nearly so, when measured properly.

I will argue that this is untrue, unless one assumes some major changes to law and public policy that are very unlikely to occur. Even if they do, there would remain quite difficult transition issues and a more permanent problem that the change would likely cause a massive shift of lending to less regulated sectors, reducing the benefits of the change, potentially to the point of making the financial system less stable in the aggregate, not more.

Once one accepts that there will be significant economic costs to sharply higher capital requirements, then a useful debate can take place about the right level of capital, given the trade-offs, and how best to achieve it. In fact, this is the debate that much of the policymaking and academic community has been involved in for some years, and to which I have contributed. My central point is that it is important not to be sidetracked by arguments that there is no real cost to the added capital.

The remainder of this paper will discuss the issues at a fairly high level, both because of space limitations and to ensure the key points are understandable for a non-specialist. For those wishing more explanation, I have included a list of my more detailed papers on this topic under References in the back. This includes a primer on bank capital, for those new to the topic.

Before beginning the substantive explanation, let me explain my background. I was a financial institutions investment banker for almost two decades, (until 2008), primarily at J.P. Morgan, which might appear to some to potentially bias me in favor of the banks. However, I have been a strong supporter of the core of the Dodd-Frank reforms and of the Basel III global agreement on bank capital and liquidity requirements, as well as other reforms, which many in my former industry lobbied against quite strongly. I have done very extensive analyses of the economic costs and benefits of higher capital requirements, including as the co-author of a year-long study for the IMF on this topic and as sole author of an earlier series of papers for a task force put together by the Pew Charitable Trusts and additional papers since.

The core of agreement

As noted, there is a sound theoretical basis for the argument that, under certain conditions, high levels of capital at a firm do not raise financing costs. Modigliani and Miller demonstrated this more than 50 years ago and both went on to win Nobel Prizes in Economics, in part for this critical insight. They found that, under idealized conditions, moving to higher levels of funding in the form of common stock, and therefore lower levels of debt, would leave the total cost of funding unchanged. Common stock (also referred to as “common equity” or sometimes “equity”) should always be more expensive than debt, because debt has greater legal protections, particularly the right to be paid off in bankruptcy prior to shareholders receiving payments. So, it might seem at first blush that more equity and less debt should raise the total cost. However, Modigliani and Miller showed that the cost of each unit of equity and each unit of debt would drop by an amount that exactly offset the additional cost from having more units of equity and fewer of debt. The price per unit drops because both equity and debt become safer, and therefore more attractive, when a firm has more equity to protect it from financial shocks and thereby avoid bankruptcy.

No one of note seriously argues with this overall point anymore, under the idealized conditions assumed in the analysis. The issue becomes the extent to which these idealized conditions hold true in the real world and what the implications are of divergences from it.

The first area of disagreement: tax effects

In the U.S. and most of the advanced economies, interest payments on debt are tax-deductible while dividend payments on common stock are not. This is partially offset by lower tax rates for capital gains on the sale of stock by investors, but the net tax effect remains substantially more favorable to debt than to equity. Adding tax effects based on U.S. law to the Modigliani-Miller framework results in an altered finding – the total after-tax cost of financing a company is lower with higher levels of debt and lower levels of equity. This is a major reason that banks fund with much more debt and deposits than equity.

Proponents of much higher bank capital requirements generally argue that this differential tax treatment is a policy distortion that should be eliminated. My impression is that most economists agree with this position, although the issue seems to me more complicated than it is often presented, even from a theoretical point of view. (Does it really make sense for a bank to have no tax benefit related to its main expense, funding itself, while being taxed on the interest it receives from making loans and owning bonds?)

Regardless of the theoretical conclusions, it behooves advocates of sharply higher bank capital to make clear what their policy conclusions would be if the tax law were not changed, since this outcome is highly unlikely. This question is too often sidestepped or downplayed.

Advocates do make the sound argument that higher tax bills for banks would not represent money being burned, but would be available for other public uses and therefore represents a private cost and not a public one. However, this would still have the effect of pushing banks to raise credit pricing and/or reduce credit availability, unless the higher tax revenue is returned to the banks or used to subsidize borrowing. That is, the tax regime for banks could be altered to lower their tax rate or in some other manner offset the higher tax bill resulting from holding more equity and less debt. (Belgium gives a tax advantage to bank issuance of common stock in order to achieve this objective.) Alternatively, borrowers could be granted a government subsidy to offset the higher costs banks would charge.

Absent these changes, we should acknowledge that credit would become pricier and potentially less available. This represents an economic cost that then has to be weighed against the societal benefits of greater financial stability and the gains from whatever is done with the additional tax revenue. The trade-off might be worth it, but it is a trade-off and needs to be analyzed as such.

The second area of disagreement: government guarantees

Another factor not present in the Modigliani-Miller model is that bank debt and deposits often receive explicit or implicit government guarantees that are not fully offset by insurance premiums. The most obvious example is the FDIC’s guarantee of a large portion of bank deposits. The FDIC charges premium rates set by Congress, which partially offset the economic advantage. However, the aggregate premiums do not fully offset the benefits and, equally importantly, the degree of risk-sensitivity of the premiums is fairly low. Put simply, many bank depositors treat their deposits as if they were government-guaranteed and completely safe. Therefore, they do not charge more for deposits with riskier banks and less for safer ones, as Modigliani-Miller assumes for debt. The FDIC premiums do vary modestly with risk, but not enough to substitute for the market pricing that would occur without government guarantees.

Similarly, most observers believe that investors in bank debt assume that their risk is lowered by the potential for a government rescue if the financial markets start to fall apart. They do not necessarily believe that an idiosyncratic problem at a single bank, no matter how large, will cause a rescue. However, their biggest risk is that we have a repeat of the recent financial crisis, when wide swathes of the financial system were put at risk. In such circumstance, there remains a belief that government help would be available to at least some extent. This, too, reduces the responsiveness of interest rates on bank debt to differing levels of risk. The level of risk of bank equity is much less influenced by guarantees, since it is observable that governments are willing for shareholders to lose a high percentage of their investments in banks, sometimes all.

Taken together, these explicit and implicit guarantees make bank debt and deposits cheaper and less responsive to changes in risk, thereby incentivizing banks to fund less with equity and more with these other sources.

Advocates of higher capital correctly point out that these subsidies represent policy distortions and ought to be done away with, or their price passed through to banks to eliminate the economic distortion. Dodd-Frank does go some ways to accomplish this, but it clearly does not eliminate the issue. Therefore, forcing banks to move away from cheap debt towards expensive equity would raise their costs, with some or all of that passed through to borrowers. Higher capital levels would make banks intrinsically safer, which would itself reduce the benefit of any remaining guarantees, but the advantage would not be eliminated.

It might be worth forcing higher capital levels and either accepting higher credit costs and lower availability or providing subsidies to offset the effect. My point is simply that there are actual trade-offs at play here, a fact often ignored or denied by advocates of very high capital ratios.

Third area of disagreement: efficiency of capital-raising

Modigliani-Miller assumes a frictionless financial system, in which there is effectively no transaction cost for raising funds and in which equity and debt markets price securities perfectly. In reality, there are transaction costs, although these tend not to be major in the grand scheme of things for stocks that are already publicly traded. (Initial public offerings have quite significant transaction costs, but additional sales after that are considerably cheaper. Banks are generally publicly traded already and therefore IPO costs will seldom arise.)

The bigger issue is that stock offerings normally come at a discount. This is observable in the market and there are also theoretical reasons to expect it. The key theoretical explanation is probably the one related to what economists refer to “asymmetric information.” Put simply, company managements know their firm’s situation better than anyone on the outside. If they are willing for their company to sell shares, then it is unlikely that they view the shares as underpriced by the market and they may even think the stock price is currently higher than warranted. This is particularly concerning, since managements tend to have an excessively optimistic view of the prospects of the businesses they run. So, if they think the stock price is reasonable or even too high, then the shares are unlikely to be a bargain. Recognizing this problem, investors normally demand a discount to protect them from the real possibility that they would otherwise be overpaying for the shares. (The same issue theoretically applies with debt issuance, but the practical effect is far smaller, for a variety of reasons[1].)

In addition, markets are not always fully efficient, with money ready to shift at a moment’s notice to the investment with the best risk/return tradeoff available. For example, a key market for bank stocks consists of dedicated funds that have developed the expertise to invest in that specialized area. There is a limit to the funds they have available for investment at any given time. Therefore, stock offerings also come at a discount out of the need to lure sufficient money in the limited time in which the offering is operational.

Some of the factors that create a need for a discount are of less significance when small amounts are raised than when larger offerings are undertaken. The informational asymmetry problem is also lessened in circumstances where managements are not given a choice, such as when operating under a government mandate.

Advocates of sharply higher capital requirements generally argue that each of the above issues are of minor importance, especially when spread out over the many years in which the bank will use the equity raised. They also sometimes argue that the informational asymmetry problem can be effectively eliminated by simply requiring banks to raise the capital, so that investors will see that it does not reflect managements’ views on stock prices. However, unless the government is willing to require that certain absolute amounts are raised, the more likely approach is to set minimum capital levels in relation to the size of the bank. In that case, bank managements could choose to shrink, in order to lower or eliminate their need to sell stock or hold back on dividend payments or share repurchases. Thus, investors would still see the choice as essentially voluntary.

Forcing an absolute level of capital may be a viable choice for regulators in the short-term, but it would become micromanagement of the banks in the medium- to long-term, by foreclosing the ability to modify business plans in a way that would reduce capital requirements.

The fourth area of disagreement: market perceptions of the safety benefits of capital

Modigliani-Miller relies on markets to correctly perceive the change in relative safety that results from adding more equity to the funding mix. However, there is a chance that markets will be too skeptical in this regard, in which case equity and debt costs will not fall as they should and total funding costs will go up more than would be required by the other factors described above. Higher funding costs would then be passed on to borrowers in whole or part.

Why might markets be too skeptical? First, markets may assume that banks will be able to “game” the system. If managements would prefer to target a lower ratio of capital to risk, they may be able to find ways to take on additional risk that are not reflected in the formulas used to determine required capital levels. At the extreme, they might be able to hold the effective capital to risk ratio constant, producing no net gain in safety. Second, and related, markets may fear that managements will take stupid risks in an attempt to keep profits up in the face of the cost pressures produced by the factors described earlier.

The “black box” nature of banks is a related problem. Investors must rely on the quality of lending, securities, and derivatives transactions that are difficult to understand from the outside. There is likely to be a limit as to how safe investors are willing to assume banks will be, at least in the proposed range of capital requirements. This may change in the long-term, if banks end up proving themselves to be very safe.

It also must be recognized that much of the empirical work in this area shows a weaker relationship between capital ratios and overall risk levels than theory suggests. There are many reasons for the inability to prove the stronger case, including real difficulties in measuring the true level of risk being taken. Nonetheless, one can understand why markets may be somewhat skeptical of something on which academics assure them of the truth, but have not conclusively proven with empirical evidence.

Assuming, as I do, that the academics are fundamentally right on this, the markets should adjust appropriately in the long run. However, the transitional problems discussed next could be considerably exacerbated for some years by the market’s need to see proof of the increased safety. In addition, problems from gaming the risk levels would not go away over time, unless regulators find better methods to catch such actions, which may not be possible. On the positive side, to the extent that banks find intelligent ways to increase expected profits while taking higher risk the result may be equivalent to regulators imposing a lower than anticipated capital ratio, which would also mean lesser effects on credit.

The fifth area of disagreement: transitional effects

Proponents of sharply higher bank capital often downplay the difficulty of the transition from our current rules to the proposed new standards. However, there are real dangers that would need to be addressed. If the transition is too short, a substantial number of banks may have to sell a considerable amount of stock to maintain their current lending levels, much less to accommodate increasing credit demand. However, raising bank equity is unlikely to look very attractive for some years, because of a combination of: the continuing effects of the financial crisis, including major litigation and regulatory risks; the ever-increasing capital requirements as a result of adopting the proposed changes; and the problems that markets can have in absorbing large offerings in a sector in a limited time period. If there is any room for discretion, many banks are likely to cut back on credit provision to avoid having to raise some or all of the new capital. If there is not room for discretion, it will mean that the government has essentially imposed credit quotas on individual banks, which seems unlikely and probably economically damaging.

If banks do cut back on credit provision, then either the economy is likely to be slowed down, or less regulated entities will pick up the lending slack, bringing up other risks that will be covered in the next section.

Previous sections mentioned some other issues that would be harder in the near and medium-term than in the long run and there are likely to be others as well.

The sixth area of disagreement: the growth of “shadow banking”

There is a danger in focusing solely on the highly regulated financial sector. Extremely high capital requirements may drive banking activity into institutions or financial arrangements that are not regulated as strongly, often referred to somewhat pejoratively as “shadow banks.” There are many ways in which “shadow banking” can occur, and different authors have different definitions. A quite incomplete list of the mechanisms and institutions includes: Structured Investment Vehicles (SIVs), repurchase agreements (repos), money market funds, finance companies, and some forms of securitizations and derivatives.

There is near universal agreement after the financial crisis that the shadow banking sector is potentially capable of creating massive problems and triggering a future crisis. Therefore, there is much discussion of how to control those institutions and types of transactions. However, the truth is that we are far from completely figuring out how to make this happen and it is unlikely that there will be an approach clever enough to provide the same level of systemic protection in regard to shadow banks as there will be for highly regulated entities. There are some types of institutions that are so much like banks that it is conceivable that they will end up with capital requirements quite similar to banks, such as finance companies, but there will always be room for activity to move still further away from arrangements that look like traditional banks.

Let me give just one example of the type of difficulty that could arise in trying to regulate shadow banking on a basis similar to standard banking. If banks, and everything that looks bank-like, have very high capital requirements, then there will be a strong incentive for major industrial and retail firms to provide credit directly to their customers and suppliers. They could simply provide credit directly, to the extent this is allowed by the new regulations without triggering treatment as a bank. Beyond that, it is well known that there are many different ways to provide supplier and customer financing without making a formal loan. For example, one could pay a supplier up-front for a shipment of goods that will not be provided for some time in the future. If that is regulated away by treating it as a loan, then it will likely still be possible in many cases to buy a year’s worth of goods in advance, with a refund mechanism if the buyer ends up wanting to take less than the agreed level. This would economically be equivalent to an informal intent to purchase goods, combined with a loan to the supplier. Regulators would have to dive deep into the regulation of the business practices of non-banks in order to avoid all the potential permutations and it is impossible to imagine that happening in the U.S.

On the other side of the ledger, these companies would find themselves borrowing large sums in order to fund the supplier and customer loans. There will be a strong temptation to do this primarily in the short-term money markets, such as the commercial paper market, since this is almost always the cheapest source of funding on average over time. Policymakers and analysts generally are concerned about the funding side as the primary source of risk to the financial system from shadow banks. After all, if an industrial company wants to loan out funds that it has obtained from shareholders or long-term bond investors, why should regulators worry? On the other hand, a “bank run” could result if short-term money markets freeze, resulting in contagion effects across the financial system. There is a great deal of truth to this, although I would suggest that a future financial system with a much larger role for lending from huge businesses to small ones could produce its own form of crisis and resulting credit crunch, if large losses started to result from making big volumes of bad loans over some future period of extended prosperity.

Current market conditions would limit how much leverage could be taken on by big industrial firms and how much of that could be short-term in nature, since wholesale markets are skittish after the debacle of the financial crisis. However, feasible risk levels could rise very substantially as memories fade.

There are many disadvantages to allowing shadow banking to supplant traditional banking as the main source of lending to small and medium-sized enterprises and, perhaps even families. (Lending to big corporations in the U.S. has already largely moved out of the hands of the banks, except for contingent lending, such as letters of credit or revolving loans or lines of credit.) The lenders would be subject to much less supervision and regulation and their activities would be less well understood by the monetary authorities and by regulators. They might also undertake lending activity with less knowledge and experience of how to do so safely. This would be a particular problem in the near to medium term, as the expertise is being acquired.

How big might the trade-offs be?

The primary intent of this paper is to underline the fact that there are trade-offs between higher capital and goals such as economic growth. It would require a much longer paper to quantify the trade-offs, as I have done in part in the papers listed below.

However, it is easy to demonstrate that the level of costs is significant enough to require serious investigation. The first-order effect of increasing the ratio of common equity to total assets for banks from 5% to 30% would clearly be very high. Assume that the annual cost of bank equity is 5 percentage points higher than the after-tax cost of bank deposits and debt. (There are arguments for a higher figure or for a lower one. This is just an example in the middle of the range.)

If one quarter of the funding for their assets (30% minus 5%) shifts to the more expensive funding source, then, all else equal, banks would have to earn about 1.25 percentage points more, after-tax, on their total assets.  This would translate into a need to collect nearly two percentage points more on their loans and other assets, all else equal, since the interest collected would be taxable. A two percentage point increase in credit pricing would have huge economic effects.

The good news is that this first-order effect would be offset by increased tax revenues, greater financial safety, a squeeze on bank cost, shifts of business away from the banks, and other factors. The debate needs to be about this set of trade-offs, rather than the false debate about why a seemingly costless approach to bank safety is being stifled by the power of the banks and those who do their bidding.


Elliott, Douglas, “A Primer on Bank Capital,” The Brookings Institution, (Washington: The Brookings Institution), January 2010,

Elliott, Douglas, Suzanne Salloy, and Andre Oliviera Santos, “Assessing the Cost of Financial Regulation,” IMF Working Paper 233, September 2012, available at

Elliott, Douglas, “Quantifying the effects of lending increased capital requirements,” (Washington: The Brookings Institution), September 2009,

Elliott, Douglas, “A Further Exploration of Bank Capital Requirements: Effects of Competition from Other Financial Sectors and Effects of Size of Bank or Borrower and of Loan Type,” (Washington: The Brookings Institution), January 2010,

[1] The deposit portion of “debt” is often guaranteed and therefore insensitive to the future prospects of the bank. The rest of the debt is insensitive to all variations in future performance in the range of outcomes that avoid bankruptcy. Stockholders, on the other hand, care greatly about whether they earn a zero or negative return or a strongly positive one. Knowing a bank is “safe” may effectively be enough for a bondholder, but is not nearly enough information for a stock investor.