Are Banks Special? International Risk Management Conference IRMC2015 Luxembourg, June 15 Michel Crouhy Natixis Wholesale Banking michel.crouhy@natixis.com and Dan Galai The Hebrew University and Sarnat School of Management dan@sigma-invest.co.il
Introduction Are banks so special that they can justify financing themselves mainly with debt? Prior to the crisis the risk-based capital ratio of banks fell as low as 1 to 3% The 2007-2009 GFC unveiled the vulnerability of banks. Excessive borrowing of banks was identified as a major factor in the crisis. Three effects seem to have been responsible for the global impact of the GFC: First, when one institution s asset sales depressed prices, other institutions were also affected because of their holdings of these assets. Second, because banks had little capital, solvency concerns arose with rising fear of contagion. Third, since much of the borrowing by banks was in the form of short-term debt, liquidy dried out and precipitated the collapse of banks like Bear Stearns and Lehman. As a result, regulators took steps with Basel 2.5, Basel III and Dodd-Frank to strengthen capital requirements of banks, especially systemic banks.
Introduction Are banks special firms that they can only achieve their goals with high leverage, above and beyond what is considered acceptable for an industrial corporation? If we accept Modigliani-Miller (M&M) proposition (1958) then the capital structure is irrelevant for both the cost of capital and the value of the bank. Hence, we ask the question: whether banks are special firms such that M&M Proposition I does not apply to bank? 3
Introduction Are banks special because of the services they provide to the economy? Should banks be subsidized in order for these services to be supplied in the right quantity? Is a subsidy in the form of government insurance efficient? Liquidity and payment services: banks channel people s savings into credit for economic investment. Depositors, samll businesses and other borrowers are not equipped to monitor their banks and ensure that they remain trustworthy. So there is an essential role for public regulation of banks to maintain stable trust in channels of credit that are vital for the economy. Banks are engaged into maturity transformation to reconcile the needs of the depositors and the corporates with long-term financing needs. Governments are, implicitly or explicitly, insuring banks creditors, then increasing leverage can increase the value of this government insurance at the expense of the tax-payers. 4
Introduction Information processing: people lend their savings to banks for investment because banks have better information about how to find good investments. Managing risks: banks have the expertise and tools to help investors and corporations to hedge their risks. A smooth functioning of the economy then requires a stable banking system: solvency of banks and trust in the financial health of the banks is critical. 5
Basel I Basel regulation of banks on a global basis started with the publication of Basel I in 1988 to be implemented from January 1992. Basel I dealt with minimum capital requirements against credit risk. It introduced the concept of Risk Weighted Assets (RWAs), against which banks were required to hold at least 8% capital. Basel I also introduced the concept of Tier 1, 2 and 3 of capital, where only Tier 1 fits the classical, strict definition of equity. In practice, in many banks before the GFC classical equity amounted to 1-3% only of RWAs. 6
Basel III New Basel III capital requirements call for a minimum of 7 % common equity capital to riskweighted assets. Basel III requires also a backstop 3 % leverage ratio, defined as common equity to total assets, or more precisely total exposure taking into account off-balance sheet derivatives. Risk-weighting is a complex system in which some assets count less against capital requirements than others. For example, a dollar of residential mortgage might count as 50 cents, but it might count as 10 cents if it is a complex mortgagebacked security, and zero if it is government debt. 7
Definition of Capital Tier 1 capital: tangible common equity and retained earnings and other instruments such as CoCos that have a loss absorbing capacity on a going concern basis. Tier 2 capital will typically consist of subordinated debt and contingent convertible capital, such as CoCo bonds. It will continue to provide loss absorption on a gone concern basis, i.e., following insolvency and upon liquidation. Tier 3 capital, or sub-supplementary capital, consisted of short-term subordinated debt with an original maturity of at least two years. It had to be unsecured and fully paid up. (Not anymore under Basel III) 8
II. Conflicting Views on the Appropriate Leverage Ratio in Banking Shortcomings of capital regulation: Basel definition of capital is based on accounting equity (calculations based on book values) Does it provide a good enough indication of the solvency of the bank? Before the GFC market values were twice the accounting values and after the crisis it was the other way around for many banks: which information to trust more in order to assess the risk of insolvency? Financial theory is based on PCM and market values. What is the rational to regulate banks capital (doesn t exist for other non-financial firms)? 9
II. Conflicting Views on the Appropriate Leverage Ratio in Banking Empirical Evidence on Leverage in Banking: Gropp and Heider (2010) The same economic factors affect the differences in leverage ratios of banks as well as non-bank corporations in the US and Europe. Deposit insurance and capital regulation considerations " were of second order importance in determining the capital structure of large U.S. and European banks during 1991 to 2004." the increase in leverage in the sample period stemmed from nondeposit liabilities. Berg and Gider (2014) though banks have much larger leverage ratios (86%) than nonbank corporations (49%), they have much lower volatility on their assets (5.0% compared to 25.1% ). 11
II. Conflicting Views on the Appropriate Leverage Ratio in Banking Shortcomings of capital regulation: Current capital regulatory framework has become too complex, too costly to operate and to supervise, and not very transparent (Haldane The dog and the frisbee; Admati and Hellwig The banker new clothes) : Investors and supervisors have lost confidence in the calculation of RWAs: becomes harder and harder to relate regulatory capital to balance sheet risks. An alternative to current regulation would be to raise capital to 20% or more. Recent empirical research supports this view: Haldane (2012) : the leverage ratio is a statistically significant predictor of bank failure, while Tier 1 risk-based ratios are not. Blundell Wignall and Roulet (2012): same conclusion. Kapan and Minoiu (2013): better capitalized banks decreased their supply of credit less than other banks. 12
II. Conflicting Views on the Appropriate Leverage Ratio in Banking Bankers arguments against more capital: Equity capital is expensive because shareholders require higher returns on capital than debt holders. More capital will force banks to charge higher interest rates on loans and restrict banks ability to finance the economy. But why non-financial firms borrowing represents less than 50% of assets? Why successful companies like Apple do not borrow at all? 13
II. Conflicting Views on the Appropriate Leverage Ratio in Banking Bankers arguments against more capital: Banks have to generate a minimum ROE that won t be achievable with higher equity capital: banks will miss opportunities that would be attractive if they could fund themselves with more debt. Myers-Majluf effect: managers will be more likely to issue new equity when they have adverse private information suggesting that the market has over-valued the stock of the company. The public may take a firm s decision to issue new equity as bad news that will have a negative impact on the stock price. This effect can indeed make equity financing more expensive. But this effect would not apply if the regulator requires transparent public information. 14
II. Conflicting Views on the Appropriate Leverage Ratio in Banking Why these arguments are false? Crouhy and Galai (1986, 1991) and Admati and Hellwig (2013) More equity reduces the ability of banks to lend money: Bank capital is not a cash reserve: both debt and equity can be used to make loans and other investments. Capital and reserves are on different sides of the balance sheet: reserve requirements restrict how banks use their funds. Crouhy and Galai (1991) analyze the trade-off between equity and cash reserves. 15
II. Conflicting Views on the Appropriate Leverage Ratio in Banking Why these arguments are false? Crouhy and Galai (1986, 1991) and Admati and Hellwig (2013) Equity is expensive because shareholders require higher returns than debt holders: Required rates of return on debt and equity are not fixed and depend on the risk of the assets. But the overall cost of capital is not expected to change with the capital ratio under the PCM assumption and no tax advantage. The cost of debt and equity cannot be considered in isolation with out referring to the debt-equity mix. According to M&M (1958) a change in the funding mix that does not affect the total risk of the investment cannot have any effect on the overall funding costs. Any impact on the funding costs is only due to deviations from the M&M framework: the fact that using a different mix may impact taxes, the subsidies banks receive or their investment decision. 16
II. Conflicting Views on the Appropriate Leverage Ratio in Banking Why these arguments are false? Crouhy and Galai (1986, 1991) and Admati and Hellwig (2013) Banks have to generate a minimum ROE that won t be achievable if they have to increase capital. ROE is not fixed to say 15%. It depends on the firm s leverage: the required ROE should be lower when there is more capital in the funding mix. M&M Proposition II shows that the expected cost of equity increases as a linear function of leverage to fully compensate shareholders for the additional financial risk. But the cost of equity will also increase as a function of the business risk for any given leverage ratio. Hence the ROE is not a good ex-ante measure of performance: it may lead to wrong incentives for managers with bonuses related to ROE. 17
III. Why Banks are Different? Banks are different because they benefit from subsidies and from underpriced guarantees that make M&M propositions irrelevant: Since the GFC, U.S. banks have been able to borrow at close to zero interest rates from the FED and invest in U.S. Government bonds that are perfectly safe and pay 2% Banks can count on being bailed out by the government when they cannot pay back their debt. Then creditors do not worry much about banks defaulting and are willing to lend them for lower interest rates. The costs of bank borrowing is then partly borne by taxpayers. Underpriced deposits insurance is also an incentive for banks to take on more risk through higher leverage and more risky investments.
IV. The Contingent Claim Approach (CCA) Crouhy and Galai (1986, 1991) proposed a Contingent Claim Approach (CCA) under the assumption of PCM: Loans originated by a bank should be priced based on the credit standing of the obligor and is independent of the capital structure of the bank. The causality is such that once the loan portfolio of the bank is determined and the bank has raised equity capital, then depositors in an unregulated environment will ask for a return on their deposits that will reflect the risk to which they are exposed. If not fairly compensated, depositors will leave the bank. Banks failed in the past for lack of liquidity rather than shortage of capital: some of these claims change in a regulated environment.
IV. The Contingent Claim Approach (CCA) The problem faced by the bank is the rolling, or refinancing, of deposits since deposits can be withdrawn any time on demand: In the 1991 paper we model this liquidity issue and show how in an unregulated environment depositors, to keep their deposits in the bank, require a return on their deposits that fully compensates them for the credit risk they face: it is a function of α, β and σ (volatility of the rate of return on the risky assets V t )
IV. The Contingent Claim Approach (CCA) The CCA is helpful to analyze the distortive effects of bank regulation and free subsidies and guarantees that benefit the banks and make banks special with regard to non-financial firms. 1. If deposits β are guaranteed and deposit insurance is underpriced, there is an incentive for the bank to shift away to the insurer more credit risk by lowering the amount of equity financing α. If deposits are not insured, the equilibrium rate of interest paid by the bank on deposits should compensate depositors for the risk of default:
IV. The Contingent Claim Approach (CCA) Equilibrium relationship between the equity ratio α and the required interest rate on deposits, r 0 This figure shows that when regulators try to control at the same time the capital Ratio of the bank and impose an interest rate ceiling on the interest paid on deposits, one of the constraints become ineffective.
IV. The Contingent Claim Approach (CCA) Optimal insurance premium required by the FDIC to insure deposits G is the cost of insuring 1 $ of deposits paid at time 0 by the shareholders: S 0 = α + (1 α) g This condition ensures that both equity and the insurance premium are fairly priced. A value of α below its equilibrium value α* means that there is a wealth transfer from the FDIC to the shareholders. There is therfore an incentive for the bank to increase leverage since the FDIC subsidizes equity holders.
IV. The Contingent Claim Approach (CCA) The CCA is helpful to analyze the distortive effects of bank regulation and free subsidies and guarantees that benefit the banks and make banks special with regard to non-financial firms. If depositors and lenders know the government will step in if the bank defaults, credit risk is then shifted to the tax payers. In this situation banks are incentivized to increase their leverage without depositors asking to be compensated for the increased risk of default. These free subsidies and guarantees have led depositors to be less concerned, or not concerned at all, by the risk of default of the bank.
The Contingent Claim Approach (CCA) Another facet of the situation that deposits are considered by depositors to be risk-free, is the uncertainty on the amount of funding provided by depositors that is inherent to the contract between depositors and the banks that allow depositors to withdraw any amount of their deposits on demand. This is a situation quite different than the one faced by non-financial firms. This uncertainty on the sources of funding means that the investment decisions of a bank cannot be assumed as given and known as required in M&M propositions.
Conclusion If M&M propositions are irrelevant for banks it is not because of the tax argument: Banks have more interest income on their loans than interest expenses. For banks interest expenses are like any production cost for an industrial firm. Hence, the tax effect is irrelevant for banks. It is due primarily to: the explicit and implicit guarantees that are underpriced or not priced at all: banks can borrow at low interest rates with little capital. Bank default risk becomes a public problem, then the requirement that banks should have adequate equity; The uncertainty on the size of the banks and investment portfolio.