Capital at Banks

Notes from the Vault

Gerald P. Dwyer
April 2011

  • There have been numerous calls for banks to raise their capital and decrease the probability of failure.
  • Instead of merely increasing capital directly, bank owners could be required to post a bond—assets that generate income to the owners as long as the bank is open.
  • Such a bond has the potential to lower the risk that bank owners are willing to undertake and thereby reduce the number of failed banks.

How much equity capital should banks hold? Absent government involvement in banking, this seems like a question to be resolved by bank owners, debt holders, and banking customers. Actual government involvement in banking, though, ranges from deposit insurance to capital injections in the middle of a financial crisis. Even before these 20th- and 21st-century developments, banks in the United States were treated differently than other firms. In the United States, banks have always been incorporated under bank charter laws and not general incorporation laws. The affairs of failed banks have been wound up by bank regulators and not in bankruptcy. Bank in the United States usually if not always have had minimum capital requirements.

Because of the failures and near-death experiences of banks in the United States since 2008, there have been various proposals to raise bank capital. For example, Alan Greenspan (2010) has argued that equity capital at banks should be twice or more its current level.

Bank capital
Bank capital has been falling in importance for much of the history of banking in the United States. Chart 1 shows the history of bank capital relative to bank assets since the first data available in 1834. In the early 1800s, equity capital provided more funds than any other source. Now, equity capital is a relatively small source of bank funds. This decline is associated with many other changes in banking and financial markets in the United States. In the early 1800s, stocks were traded in regional markets in various states. Banks issued notes that passed from hand to hand as currency. But much has changed since then.

chart 1
The data are from Historical Statistics of the United States and the Federal Deposit Insurance Corporation (FDIC). The ratio shown is the ratio of the book value of capital to total assets. The Historical Statistics data are for all banks and the FDIC data are for insured banks. The Historical Statistics data end in 1980. The ratios from the two sources differ by only 0.01 percentage point in 1976, so data from the FDIC is used in 1977 and later years.

The decline in capital in the last century (chart 2) is of more immediate interest. Both the passage in 1913 of the Federal Reserve Act and the introduction of deposit insurance in 1933 had the potential to reduce banks' holdings of capital. Because the Federal Reserve can be a lender of last resort in the event of liquidity difficulties, banks may well want to hold less capital because they do not need to hold extra liquidity or capital for a possible run on the bank. In addition, because the Federal Deposit Insurance Corporation (FDIC) insures—in effect, guarantees—deposits up to a maximum amount, depositors are less concerned about their banks' financial condition, including the amount of capital invested by owners. As a result, banks can hold less capital. While the Federal Reserve may or may not have reduced banks' demand for capital, a decline in capital since 1933 is evident in the figure, and empirical evidence supports the role of the FDIC in that decrease.1 And much modern banking regulation assumes that banks want to hold less capital than regulators would like them to hold.

chart 2
The data are from the Historical Statistics of the United States and the Federal Deposit Insurance Corporation. Chart 1 provides additional information.

Both the bank capital in charts 1 and 2 and the bank capital determined by banking regulators are book capital. This is quite different than the market value of bank capital. Chart 3 shows an index of the market value of bank capital from January 2, 2006, to April 21, 2011. This index shows a very wide range in the market value of bank capital, from a low of 87.2 to a high of 218.6. This wide range over a little more than four years no doubt is as extraordinary as is the range of the S&P 500 or Dow-Jones Industrial Average during these years. Still, the market price of bank equity always varies more than the book value regulated by banking regulators. Banks are more likely to fail when their stock prices and the market value of bank capital are low rather than high.

chart 3
The data are from Standard and Poor's. The price index is the S&P/TSX Equal Weight Diversified Banks Index, the equal-weighted version of the S&P/TSX Diversified Banks Index. This index includes commercial banks whose business is derived primarily from commercial lending operations with significant activity in retail banking and lending to small and medium corporations.

The market value of bank capital is a better estimate of the value of the owners' stake in a bank than is its book capital. The market value of bank capital is also probably a better estimate of any possible recovery of funds the FDIC might receive when it resolves a bank.

Using bank capital to reduce risk
Various proposals have been made concerning bank capital in the last few years. Greenspan (2010) suggested that the minimum bank capital be raised "substantially." In his view, this increase in capital would substantially reduce the probability of banks failing if banks held similar assets before and after the increases. Others also have suggested higher capital.

Instead of only raising capital requirements, others have recommended different forms of capital as well. Most prominently, Mark Flannery (2005) recommended convertible collateral bonds—bonds that convert to equity if bank capital falls sufficiently low (Dwyer 2010).

Viral Acharya, Hamid Mehran, and Anjan Thakor (2011) proposed that banks be required to have capital that is held but not used to fund typical banking business.2 If a bank fails idiosyncratically, then this capital is forfeited. Although a failure could be bad luck, it is more likely due to bad management, fraud, or careless monitoring of the bank by its owners. The loss of the capital raises the cost of such poor performance and makes failure less likely.

Losing capital in the event of a failure has an analytical similarity to the role of bank capital in a well-known paper by Keeley (1990).3 Keeley argues that banking competition erodes the value of banks' charters, and banks are less likely to take risks that might result in failure if the banks' charter has value that is lost on failure. Effectively, a bank's charter is a valuable asset that disappears on failure. The analysis in Acharya et al.'s paper suggests using Keeley's proposition in a different way than just to explain banks' failure rates.

A proposal
Regulators could require a new bank to post a bond when it receives a charter. This bond would be a bond in the sense of a posted dollar amount forfeited in the event of failure, not in the sense of U.S. government or other debt. The funds would be invested in marketable securities, and the stockholders would receive the income from the bond as long as the bank did not fail. If the bank closed, the funds from the bond would be available to regulators to pay off liabilities. Anything left over would be returned to the bank's stockholders.

Such a bond is similar to capital in some respects. This proposed bond would be an asset that generates income as long as the bank is in business. The assets would not be invested in banking though, so a decrease in the bank's business would not in and of itself affect the bank's income from the bond or the value of the bond. The bond would be lost if the bank failed and some holders of the bank's liabilities were not paid. Because the bank might lose the bond upon failure, failure would be less likely because banks would take fewer risks to avoid losing the income from the bond.

This proposal is a preliminary policy suggestion. Many details would need to be fleshed out. For example, in what assets should the bond be invested? How might different investments affect the behavior of banks? If all the funds are invested in Treasury securities, a bank might be inclined to hold a riskier portfolio in its banking business. On the other hand, investment in stocks might result in the value of bond fluctuating too much. Another issue is whether banks should be required to add to the bond when they become larger. A large bond for a small bank is a small bond for a large international bank. Having a bond that is a specific fraction of assets would be necessary for it to be equally effective and feasible for small and large banks.

This proposal is similar in its essentials to the proposal by Acharya et al., and different in some details. Acharya et al. seem to take it for granted that the funds should be invested in Treasury securities, which is not obvious. Such investment might induce undesirable asset substitution on a bank's balance sheets. The market portfolio of stocks might be more likely to generate a flow of income to banks' stockholders that would not induce banks to take on extra risk due to the income from Treasury securities. The authors also take it for granted that funds should be transferred from the bond account to the bank's capital if losses occur and bank capital falls. An alternative solution could be to permit no such transfers and require the bank to operate as if the bond account did not exist other than through the receipt of income by stockholders, as long as the bank is in business. In this way, the bond is a lump-sum receipt of income to the bank's owners independent of the bank's activities other than staying open.

Many regulators, economists, and other experts have called for additional capital in banks. The alternative suggested here is that banks hold additional capital in a particular way which could help make banks less likely to take risks that could result in their failure.

Gerald Dwyer is the director of the Center for Financial Innovation and Stability at the Atlanta Fed. Robert Eisenbeis provided helpful comments, which does not necessarily imply that he agrees with anything said here. The views expressed here are the author's and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System.


1 The most recent empirical study of which I am aware is Dwyer (1981).

2 Acharya presented this paper at the 2011 Financial Markets Conference sponsored by the Federal Reserve Bank of Atlanta.

3 This similarity also was noticed by Robert Eisenbeis.

Acharya, Viral, Hamid Mehran, and Anjan Thakor. 2011. Caught between Scylla and Charybdis? Regulating bank leverage when there is rent-seeking and risk-shifting. Paper presented at the 2011 Financial Market Conference. Available at

Dwyer Jr., Gerald P. 1981. The effects of the Banking Acts of 1933 and 1935 on capital investment in banking. Journal of Money, Credit and Banking 13 (2): 192–204.

Dwyer, Gerald P. 2010. The financial system after the crash: Policy fallout. Notes from the Vault. July.

Flannery, Mark J. 2005. "No pain, no gain: Effecting market discipline via reverse convertible debentures." In Capital adequacy beyond Basel: Banking securities and insurance, edited byHal. S. Scott, 171-92. Oxford: Oxford University Press.

Greenspan, Alan. 2010. The crisis. Brookings Papers on Economic Activity Spring: 201–46.

Keeley, Michael C. 1990. Deposit insurance, risk and market power in banking. American Economic Review 80 (5): 1183–200.