"Cash, Capital, and Bank Risk-Taking: Back To the Future"

Transcription

1 "Cash, Capital, and Bank Risk-Taking: Back To the Future" Charles W. Calomiris (Based on coauthored work with Florian Heider, Marie Hoerova, Mark Carlson, Andrew Powell, Berry Wilson, Joseph Mason, David Wheelock) World Bank June 3, 2013

2 Where We Are (since the 1980s) Failed system of government protection of banks, which replaces private market discipline with government prudential regulation => Pandemic of costly banking crises! Regulation now focuses in most countries on book equity ratio requirements, despite evidence showing that these do not work. This is subject to manipulation by both bankers and government officials, who often have strong incentives to overstate capital, and undertake excessive risks.

3

4 Importance of Risk Management Emerging market twin crises show increases in risk taking after recession begins (often FX risks) at great public expense, and great GDP loss when collapse comes. Subprime crisis: Cross-section of risk taking was the same as in 1998, and reflected conscious differences in risk management (CRO Comp / CEO Comp).

5 Where We Were (and Where Some Remain) Substantial (often 20-30%) cash or government security reserve requirements (relative to debt). In the U.S., these date back to at least the 1830s, and were common both as government requirements and clearing house self-regulation requirements historically, and continued to be important under the National Bank Act and Federal Reserve Act. Other developed countries (e.g., the UK) also relied on reserve/liquidity requirements. Many emerging market countries still do!

6 Why the Change? In the U.S. and the U.K., financial disintermediation from banks reflected zerointerest payments on reserves combined with high inflation, and other unrelated factors. The switch from reserve ratios to capital ratios avoided continuing taxation of banks. But paying the fed funds rate on reserves avoids that problem, without having to rely entirely on capital ratio requirements.

7 Role of Cash? Now reserve requirements (outside of emerging markets) are low, proposed reforms focus narrowly on need to deal with exogenous liquidity risk. (A legitimate goal as Brazil s recent experience shows but we will return to discuss the flaws of Basel III approach to this problem later.) The point of this lecture: Cash reserve requirements are a unique and important prudential tool that needs to be employed alongside capital ratio requirements to manage insolvency risk. Substantial cash reserves should be required. Interest should be paid on them to avoid taxing banks.

8

9 Cash in a Frictionless World In a frictionless world (perfect information, no transaction cost) a liquidity requirement would add nothing to a capital requirement In that world, they are two equally good ways to skin the cat of targeting a given default risk on bank debt.

10 Loans and cash as assets Simple Bank Deposits and equity as funding sources Riskiness of assets is given by riskiness (s L ) of loans multiplied by loan-to-asset ratio (L/A) = 1 - (R/A) Equity ratio is (E/A) Riskiness of assets = s A = (L/A) s L Given s, (R/A) and (E/A) are two alternative tools for reducing deposit default risk from loan loss

11 Two Ways to Skin Cat of Default Risk

12 Theories with Frictions and Cash s Role Exogenous liquidity risk relating to risk of debt market collapse. Lack of substitutability of debt capacity for cash during times of need. Cash is better than debt capacity when risk of assets jumps. But short-term debt markets don t just collapse; they collapse when insolvency risk rises. Perhaps cash s role should be seen as preventative of crisis, not just insurance against costs of crisis.

13 Theory of Cash as a Prudential Tool (Calomiris, Heider and Hoerova) Under market discipline, cash holdings create depositor confidence that capital alone cannot because cash is observable and because it cannot be made risky. Cash is both a signal of asset quality and a commitment device to good risk management. Across banks, those whose risks and risk management are harder to observe will rely relatively more on cash. Across time, more cash in recessions. In recession states of the world, banks increase in cash to demonstrate continuing asset value and to credibly commit to more costly management of their loan risks. This convinces depositors to stay in the bank. => Win/Win of continuing intermediation

14 Why Is Cash a Commitment Device? Cash observably and credibly raises the lower bound of liquidation value of the bank (what debtholders get in insolvency states of the world). Higher cash means the banker gets less upside from increasing risk, and depositors are less subject to loss on downside. This can eliminate asset substitution risk in bad states => promotes good risk management.

15

16

17 Theory of Cash (Cont d) In the presence of a government safety net (deposit insurance, other debt bailouts) market discipline is absent, so it no longer leads banks to choose optimal levels of cash on their own. Regulatory cash ratio requirement is needed. Cash ratio requirements should be set higher than voluntary cash ratios under autarky because monitoring of state is unreliable.

18 Empirical Evidence (Market Discipline) Cross-sectional implications Cash and capital are substitutes for solving risk problem. Banks tradeoff one against the other and choose optimal combination based on costs of each (which differs across banks, so some banks choose more cash, others more capital). Observability of risk and risk management Cost of raising equity Quasi rents from lending

19 Empirical Evidence (Cont d) Time-series implications During recessions, all banks will tend to rely more on cash. Asset substitution risks rise Cost of raising capital rises Quasi rents from lending fall

20 Tradeoff: Calomiris-Powell (Argentina) Dep Var: Argentina s Quarterly Deposit Growth Regressor Coefficient Stand.Error Eq Ratio (-1) Loan Int. Rate Loans/Cash Sample period: 1993:3-1999:1 Number of Observations: 1,138 Adjusted R-Squared: 0.31

21 In 1935, a Nat. Bank s Cash is higher if (Calomiris, Mason, Wheelock) Equity is lower Loan risk is higher Mix of deposits has higher exogenous liquidity risk

22 Bank Corporate Governance and Risk Management in a Different World (Calomiris and Carlson) The National Banking Era ( ) National banks are unit banks (~similar size) Identical charter governing powers and practices Identical (national law) courts with jurisdiction No deposit insurance, and no central bank Ownership structure is chosen by the organizers of the banks, with no limits (other than natural ones related to transaction and information costs) How were ownership, governance, and risk taking related to one another in this system?

23 Governance of National Banks Examined once-twice a year (semi-random arrival, spatial sequencing). Five times a year submit call reports detailing their balance sheets. No prudential capital requirements, prudential cash reserve requirements not strictly enforced (as a fraction of deposits; frequent ~15% violations revealed in exams, unclear penalties). Stock holders face double liability. Lots of voluntary corporate governance rules.

24 Ownership Structure, Risk, and Cash Ownership structure is a key influence on corporate governance structure, risk, and relative reliance on cash. Two kinds of banks. Closely held banks => informal governance, hard to observe risk, but lower default risk; main costs are adverse selection and asset substitution risk in bad states, which are solved with greater reliance on cash as risk control. Widely held banks => formal governance structure, greater tolerance for default risk, less risk of asset substitution in bad states, less reliance on cash. One of these is not better than the other.

25 Sample All National banks in 37 cities 207 total banks 22 failed in the panic and 36 suspended Cities are either Western or Southern reserve cities Kansas City, MO; Louisville, KY; Minneapolis, MN; New Orleans, LA; Omaha, NE Larger non-reserve cities Denver, CO; El Paso, TX; Los Angeles, CA; Portland, OR; Spokane, WA; Stillwater, MN Mid-size banks Assets of $164 thousand to $8.3 million Largest banks at the time had ~$35 million in assets

26 14 12 Share of sample (percent) Top Managers Ownership share (percent)

27 Voluntary Governance Decisions Independent directors (the number > one) Board size (4-23, mean of 9) Frequency of board meetings Bonding of cashier (60%), bonding of president (35%) Formal loan approval committee (if included independent director) Equity-to-assets ratio Cash-to-assets ratio

28 Correlation Matrix Board meets at least monthly High % Outsiders on Board Active discount comm. Pres. bonded Cashier bonded Management stock share Board meets at least monthly High % Outsiders on Board Active discount committee Pres. bonded 0.50 Note that all correlations in first row are negative. All other correlations positive.

29 Low Management Ownership Share Medium Management Ownership High Management Ownership Assets ($ millions) Governance Score Pres. Salary /Assets President Bonded Officers Loan Share Outs. Dir. Loan Share Dividends per share

30 Measures of Banks Risk Choices Probability of failure or suspension Reliance on hot debt market Percent Troubled loans (examiners opinions) Estimated loan losses (objective criteria for bad loans, plus examiners opinions about other troubled ) Ex ante measures of loan risk (loan types)

31 Bank Closures National All banks

32 First Stage: Governance Score Intercept 8.38 ** Top 3 Managers Ownership% ** Outside Directors Ownership% 0.84 Log assets 0.11 Log age ** Gov Score, other local banks 0.50 ** Reserve city 0.09 Log bank assets in state Log distance to NYC ** Percent county income from Ag 0.07 F-statistic Adj R-Square 0.28

33 Relationship to Risk Taking High management ownership, and lessformal governance, is associated with: Lower probability of failure, lower reliance on borrowed funds, lower loan losses/assets Fewer real estate loans => Formal governance is associated with more risk

34 Capital/Cash Mix in Risk Management More cash/less capital makes sense if asset substitution risk in bad states is higher (or if adverse-selection costs of raising equity are higher). More formal governance should make risk and risk management more observable and thus reduce relative reliance on cash. Choosing to be an inside-dominated bank means more information and control problems => greater use of cash, less of capital.

35 Cash to Asset Ratio OLS OLS IV IV Intercept -0.66** -0.51* ** Top 3 Managers Ownership% 0.11** Score -0.01** Score - predicted -0.08** Score - residual -0.01** Outside Directors Ownership% Log assets 0.03** 0.02** 0.03** 0.02** Deposits/Total Debts Checking/Total Deposits 0.09** 0.11** 0.12** 0.10** Log age * Gov Score, other local banks -0.01** -0.01* ** Reserve city Log banking assets in state 0.02** 0.02* ** Log distance to NY ** Percent county income from Ag

36 Equity to Asset Ratio OLS OLS IV IV Intercept 1.49** 1.37** ** Top 3 Managers Ownership% -0.14** Score 0.01 Score - predicted 0.10** Score - residual 0.00 Outside Directors Ownership% -0.09* -0.11** -0.17** -0.11* Log assets -0.06** -0.06** -0.07** -0.05** Deposits/Total Debts -0.20** -0.20** -0.21** -0.20** Checking/Deposits 0.08** 0.07* * Log age -0.04** -0.04** ** Gov Score, other local banks 0.01* 0.01* -0.04* 0.02** Reserve city -0.04** -0.04* -0.05* -0.04* Log banking assets in state -0.02* Log distance to NY ** 0.01 Percent county income from Ag

37 Time Series Implications (Calomiris-Wilson) Cash becomes more important for all banks in bad states of the world. With onset of Great Depression, NYC banks suffered a series of shocks. Their low-default risk target was constant, but they increasingly used cash assets to achieve it.

38 Table 2: NYC Banks Loans/Cash, Risk, Equity, Dividends Loans/(R+T) Ass.Risk Equity/Ass. p Dividends $392m $162m Source: Calomiris and Wilson (2004).

39 How High a Cash Requirement Is Needed? If this were only reform, it would have to be set very high, but it should not be the only reform. Each reform has unique costs; a mix ensures no single cost becomes excessive (e.g., reliance on cash reqs disadvantage banks with high quasi rents from lending and low costs of accessing equity capital). Reform should combine remunerative cash requirement (20% of debts) with several other reforms (all incentive-robust simple, transparent, and not reliant on bankers or supervisors discretion including CoCos with market triggers, higher book equity ratio requirement, reforms of measurement of loan and debt risks on bank balance sheets).

40 Basel III Liquidity Ratio Proposal Basel s LCR and NSFR penalize illiquidity by equating more cash with less short-term debt, and also have a very broad (politically driven) definition of cash assets. Banks are supposed to produce liquidity with short-term debt! Idiosyncratic exogenous liquidity shocks be handled by interbank market. Exogenous systemic liquidity shocks (imported crisis from another country) can be handled by LOLR unless there is a default risk problem. Controlling endogenous liquidity risk is all about controlling insolvency risk.