Macroprudential analysis of financial institutions: section 3

Transcription

1 Macroprudential analysis of financial institutions: section 3 Master 2 course J.S. Mésonnier Banque de France, Financial Economics Research This version: october 2013 Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

2 Outline of part 3: regulating bank capital in a macroprudential perspective Why do banks hold (so little) capital? Why regulate bank capital? The macro importance of bank capital Bank capital regulation from Basel I to Basel III A critical view of the "Basel tower" Beyond Basel III Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

3 A puzzle: the high leverage of credit institutions Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

4 Leverage and profitability across industries Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

5 The capital structure decision of firms: Modigliani & Miller (1958) What is the cost of capital? Developing a theory of corporate investment when asset returns are uncertain. Restrictive assumptions: perfect competition, no assymetries of information, no opportunity of arbitrage, no taxation Firms grouped into homogenous classes of risk (correlated returns), so that the expected rate of return on assets in each class is a constant ρ A Then, M&M show (by arbitrage) that the market value of a firm is independent of its capital structure: ρ A.V A = ρ A.(E + D) As a consequence, the expected return on shares can be expressed as (WARCC equation): ρ E = ρ A.V A r D.D E = ρ A + (ρ A r). D E Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

6 The capital structure decision of firms: Modigliani & Miller (1958) M&M assume that the asset side of the balance sheet (projects, physical assets, loans, securities) is given: the financing-mix decision is separable from the investment decision An ex-ante view on the capital structure decision, concerned with having, not raising equity (see below). Different from pure accounting equation (book values and ex post): ROE = ROA + (ROA r).d/e M&M stringent assumptions generally do not hold, but M&M should be used to discipline the analyisis: one must be precise about what deviation from idealized conditions is at work. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

7 Deviations from M&M and capital structure of banks Debt is subsidized (tax deductibility of interest payments): true, but not specific to banks Failure is costly: limit to leverage (applies to all firms, but banks benefit from implicit/explicit state guarantees) Disciplining role of (short-term) debt and high leverage (Calomiris and Kahn, 1991, Diamond and Rajan, 2001): Debt returns are not contingent: no reason to monitor managers. However, with higher leverage, creditors are more exposed to losses: incentive to better monitoring and higher required returns Banks are "opaque": agency problems of creditors with management. Short-term debt (like deposits) is disciplining managers because of the constant threat of a run Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

8 Deviations from M&M and capital structure of banks (foll d) Note that this can go too far: high leverage induces risk-shifting moral hazard. Managers/Shareholders prefer riskier gambles that maximize the price of their equity (call) option on bank assets (Jensen and Meckling,1976). Intuition: suppose a project-firm yielding E (R), funded by E and D. Debtholders are senior. Compare expected utilities for shareholders and debtholders if project returns either value R = D σ or R = D + σ with equal probabilities. Tension between run-based disciplining role of leverage and risk-shifting incentives: basis for a "market-based capital requirement" Debate: is the latter enough? At all times? (risk assessment through the cycle...) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

9 Why regulate bank capital? General motivations Cost of negative shocks to bank capital: empirical findings Bank capital and the amplification of macro shocks: a DSGE approach Bank capital and systemic risk-taking: a macro model Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

10 Why regulate bank capital? Banks are fragile: prone to runs (Diamond and Dybvig, 1983). High social cost of bank failures and bank crisis (Laeven and Valencia, 2008, 2010 see Table below). This cost is not internalized by bank managers/shareholders: thus need for regulation! Deposit insurance creates adverse incentives: insured depositors do not need monitor bank managers anymore. Hence more risky decisions. Regulatory capital requirements must compensate. [What happens when managers are not shareholders and complete contracts are not feasible? Bank capital regulation needed to enforce second best solutions on behalf of depositors. See Dewatripont and Tirole (1994), also summary in FR, chap 9.] Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

11 Banking crises entail high social costs Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

12 Bank leverage in a historical perspective (US case) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

13 Why regulate bank capital? Note however: possible welfare costs of capital regulation (e.g. Van den Heuvel, 2010). Notably, social benefits of the issuance of "bank money". See also Gorton and Metrick (2010) for an analysis of private bank benefits from issuing collateralized short term debt (like repos) that command money-like convenience premiums (safety + transaction services). Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

14 Costs of bank capital shocks: empirical results How may adverse shocks to bank capital impact lending and economic activity? Challenge #1: disentangling loan demand and supply effects Bad economic conditions reflect in poorer lending opportunities, lower loan demand for investment, and deteriorated bank capital (write-offs). How to deal with reverse causality? Natural experiment: identify some truly exogenous source of capital depletion. Best exemple: Peek and Rosengreen (1997) Exploit differences in bank balance-sheets, assuming common lending opportunities: OK if differences across banks hinge only on their ability to rebuild capital. Ex.: Bernanke and Lown (1991), Hancock and Wicox (1993), Berrospide and Edge (2010) Statistical controls, e.g. lags of loans as IV or regressors in a VAR. Recent microeconometric literature using loan-level data in a diff-in-diff spirit (e.g. Jimenez et al., 2011, cf. Peydro, IJCB, 2010 for a survey). Converging results that capital shocks matter for loan supply Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

15 Costs of bank capital shocks: empirical results Challenge #2: assessing the impact on economic activity Many studies find that changes in capital fo some banks matter for their lending to some borrowers Evidence of aggregate impact? What if borrowers have other sources of funding? How to control for this? Natural experiment again: Peek and Rosengreen (2000) Some studies use VAR frameworks, thus building on the statistical approach above. Ex.: Berrospide and Edge (2010) Less concensus on overall economic impact of changes in bank capital Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

16 Bank capital shocks and lending: Peek and Rosengreen (1997) Focus on US branches of Japanese banks around the burst of the Japanese bubble (end 1980s-early 1990s) From 1988 on, Basel I allowed Japanese banks to count their unrealized gains on common equity holdings as Tier 2 capital, thus fueling the boom During the 80s, aggressive international expansion of Japanese banks, notably in the US (18% of US C&I loans in 1990) Burst if the bubble (Nikkei lost 50% over ) caused large Japanese banks to fall below the regulatory minimum of 8% P&R use this and regress loans (DL) by Japanese branches in the US on their parent company s capital ratio (PRBC) + controls: 1 pp decline in PRBC results in 6% in loans at the branch level Note: effect is much lower for Japanese banks subsidiaries Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

17 Bank capital shocks and lending: Bernanke and Lown (1991) Focus on the "credit crunch" associated with the recession in the US. Claims by observers that depletion in (book) bank capital following the burst of the real estate bubble in some states, notably in New England, interacting with the phasing-in of Basel I, caused a decline in lending that aggravated the recession (hence "capital crunch", R. Syron). Cross-sectional regression using state level data of loan growth during the recession on CAR in same regression at bank level for New Jersey banks suggest strong effect of lower capital ratios on lending, at least for small banks Consistent with capital crunch hypothesis, assuming that large banks can raise equity more easily, but not suffi cient (e.g. if loan markets are different for both types) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

18 Bank capital shocks and lending: Bernanke and Lown (1991) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

19 Bank capital shocks and lending: Berrospide and Edge (2010) Follow on methodology of Hancock and Wilcox (1993) and assume that banks adjust their leverage to changes in a predetermined target which hinges on bank characteristics: K i,t = λ(ki,t K i,t ) + ε i,t Ki,t = α i + θx i,t 1 Loan growth is then regressed at bank level on the estimated capital surplus (Ki,t K i,t)/ki,t and macro controls Use Call report data for 140 large US BHC over Find that $1 capital surplus results in $1.86 increase in loan volumes (for average capital asset ratio of 10% and average loans to assets of 60%) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

20 Bank capital shocks and activity Peek and Rosengreen (2000) show that loan supply identified in PR 1997 impacted real activity in US states where Japanese branches accounted for a significant market, notably market for commercial real estate (CRE). State level panel regressions CRE activity on lending by Japanese banks and controls, instrumenting the lending variable by the health of the relevant Japanese parent banks Find significant impact of lending by Japanese bank branches on real estate activity Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

21 Bank capital and amplification of macro shocks: theoretical results from a DSGE model Meh & Moran (JEDC, 2010): New Keynesian DSGE model with banking sector where bank capital solves asymmetric info problem bankers and depositors (à la Holmström Tirole, QJE 1997). Bank capital thus determines ability of banks to attract funds and fund investment: influence on business cycle through a bank capital channel of transmission. 3 types of agents: households (depositors), entrepreneurs, bankers and 3 types of goods: final and intermediate goods as in standard NK model + capital goods produced by entrepreneurs. Production of capital goods is plagued by two types of moral hazard problems: Entrepreneurs may choose type of project at the costs of banks Banks may choose intensity of monitoring at the costs of depositors Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

22 Capital good production: shirking entrepreneurs Entrepreneurs invest i t for a gross observable return of Ri t (success) of 0 (failure). They privately choose among different projects with different prob of success and levels of private benefits: Good guys: choose projects with proba α g and private benefits 0 Bad guys: choose α b < α g and private benefits bi t Ugly guys: choose α b and private benefits Bi t, B > b Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

23 Capital good production: bank monitoring Banks can monitor entrepreneurs to deter them from choosing projects with high-level shirking intensity. Monitoring is costly (cost µi t ) and monitoring effort non-observable to depositors: moral hazard problem again Solution: banks must invest their own capital in the projects to align their private incentives to the interests of depositors Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

24 Financing entrepreneurs: financial contract (1) Entrepreneur: net worth n, project of size i, needs external financing i n (subscript t here) Bank: combines deposits d with own equity (net worth) a (with required market rates of return: r d, r a ) Price of capital goods in terms of final goods: q One-period financial contract set in real terms, so that we select only equilibria where entrepreneurs must choose good projects (i.e. α g ) determines investment size i, contributions a and d, and respective (positive) shares of total return: R e, R b, R h. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

25 Financing entrepreneurs: financial contract (2) Contract aims at maximizing expected return to the entrepreneur s.t. incentive, participations and feasability constraints: max {i,a,d,r e,r b,r h } (qαg R e i) s.t. : qα g R e i qα b R e i + bi qα g R b i µi qα b R b i qα g R b i (1 + r a )a qα g R h i (1 + r d )d a + d µi i n R e + R b + R h = R Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

26 Financing entrepreneurs: financial contract (3) Solving for the return shares in equilibrium, they come out as linked to severity of moral hazard problems: R e = b/ α R b = µ/ α R h = R b/ α µ/ α Solving for i, one gets: i = (a + n)/g, where G = 1 + µ + qαg 1+r d (R b/ α µ/(q α)) 1/G is leverage of project over net worth of bank and entrepreneur. NB: constant across all contracts. Also note that G / q < 0 and G / r d > 0 Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

27 Financing entrepreneurs: financial contract (4) Important to see also that financial contracts determines a market-based capital adequacy ratio κ that banks have to meet and depends only on macro variables (r, q): κ = = a a + d µ µ + q α + r a (R b/ α µ/(q α)) 1+r d Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

28 Reality check: cyclical properties of modelled CAR After derivation of FOC, aggregation, computation of competitive equilibrium and calibration, how do the cyclical properties of model-based CAR compare with those observed for the US economy? Good! Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

29 What do we learn from this model? Simulated IRFs to shocks An active bank capital channel amplifies the size and persistence of the effects of adverse technology shocks and restrictive monetary policy shocks (compared to economy where µ = 0) These amplifying effects are dampened whenever banks have more capital ex ante (higher CAR of e.g. 20%) "Credit crunches" matter: negative bank capital shocks (=accelerated depreciation of bank capital have strong recessionary effects. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

30 Technology shocks and the bank capital channel Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

31 Technology shocks and the bank capital channel: economy with more bank capital Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

32 Credit crunch Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

33 Bank capital ratios in a macro model of systemic risk-taking Martinez-Miera and Suarez (mimeo, CEMFi, 2012) Objective: incorporate banks & systemic risk in macro analysis DSGE-style models explicit about banks do not yet have/share clear notion of endogenous systemic risk: Pre-crisis: Van den Heuvel, 2008; Meh-Moran, 2010 Post-crisis: Brunnermeier-Sannikov, 2011; Gertler-Karadi, 2011; Gertler-Kiyotaki, 2010;... In this paper, systemic risk results from banks voluntary exposure to an infrequent & large common shock... which is attractive to them due to standard risk-shifting incentives of levered firms [ Clear link to microeconomic literature on bank risk-taking, cf. Jensen & Meckling, 1976] Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

34 Main features of the model 1 Dynamic equilibrium model in which perfectly competitive firms need bank loans for their funding in advance of wages and working capital 2 Firms & banks make joint unobservable systemic risk-taking decision Systemic firms fail with proba 1 if systemic shock occurs (otherwise failure is iid) Systemic firms fail less conditionally of no shock, but more unconditionally 3 Bank capital dynamics: capital invested in systemic loans is lost if shock realizes: Last bank standing effect like in Perotti-Suarez (2002) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

35 Other features Many simplifications: risk-neutral agents, no physical capital accumulation, no money, no price rigidities, no labor market frictions... but structural description is rich enough for calibration Role for bank capital is different from other macro papers... Meh-Moran 10: Monitoring incentives a la Holmström-Tirole 97 Gertler-Kiyotaki 10: Preventing fund diversion a la Hart-Moore here it reduces bankers systemic gambling incentives Aggregate uncertainty is key, requiring full non-linear solution (rather than log-linearization around non-stochastic SS) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

36 Qualitative results 1 Strengthening capital requirements (CRs): Reduces systemic risk taking and, hence, the losses due to systemic shocks Reduces credit and output in normal periods (but also the size of their collapse during crises) 2 Welfare trade-offs produce a unique socially optimal CR [Concerns on costs of crises vs. Constrained supply of credit] Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

37 List of endogenous variables Marginal value of bank capital v t = v(e t ) 1 Fraction of systemic banks x t = x(e t ) [0, 1) Physical capital used by firms k t = k(e t ) 0 Wage rate w t = w(e t ) 0 ROE at non-systemic bank R 0t+1 = R 0 (e t ) 1 + r ROE at systemic bank if no shock R1t+1 1 ε = R1 ε 1 (e t ) 0 Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

38 Social welfare W as a function of capital requirement γ Social welfare % 8% 10% 12% 14% 16% 18% Capital requirement Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

39 v(e) and x(e) under low and optimal γ Value of one unit of bank capital (v) optimal capital requirement (14%) low capital requirement (7%) Systemic risk taking (x) optimal capital requirement (14%) low capital requirement (7%) Aggregate amount of bank capital (e) Figure 2a: v(e) Aggregate amount of bank capital (e) Figure 2b: x(e) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

40 Comments on Fig. 2: Greater scarcity of e due to higher γ implies larger v(e) Systemic risk-taking (increasing in e) is lower when γ is higher Why? γ affects position and slope of v(e) Gives greater incentives to preserve e after systemic shock Further intuition (esp. on welfare trade-offs) requires looking at (endogenous) dynamics of e Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

41 Quantitative results Optimal capital requirements: positive and large (14%) Comparison CR=7% CR=14% (unconditional means) Fraction of systemic loans: 71% 24% Loan rates increase: 4.1% 5.6% Macro aggregates fall: bank credit flows by 21%), GDP growth by 7% Yet, difference in social welfare +0.9% permanent consumption: static gains (less risk-shifting) dynamic gains (lower loss of economic activity after shock) Fall in year-after-shock aggregates: CR=7% loan rate (+11.6pp), bank credit (65%), GDP (32%) CR=14% loan rate (+2.5pp), bank credit (24%), GDP (10%) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

42 Take away of this model Very significant implications of capital requirements Setting them optimally requires economic risk-management view: static element: ineffi cient systemic risk-taking if leverage too high = larger unconditional failure rate dynamic element: macroeconomic effects of systemic shock propagated and amplified by loss of bank capital Standard macroeconomic aggregates evaluated in the PSS (credit, GDP) give bad indication of convenience of high γ: capital requirements have large costs in those terms. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

43 Bank capital regulation in a historical perspective: overview Focus on bank capital regulation is recent: late 1980s 1930s-1970s: stronger focus on market structure (e.g. US Glass Steagall Act), asset allocation rules, caps on interest rates (e.g. US Regulation Q) s-1980s: regulations made largely ineffective due to financial innovation, liberalization of international capital flows, IT. Basel Accord of 1988 (Basel I) introduced new harmonized minimum capital requirements for international banks. Risk weights by class (100% for C&I loans, 50% for mortgages, 0% for loans to sovereigns) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

44 Bank capital regulation in a historical perspective: overview 1993: proposed extension of Basel I to market risk (changes in value of trading books). Criticized by industry as imposing too bold ratios compared to predicaments of banks quantitative models Amendment to incorporate market risk: banks allowed to determine regulatory capital on the basis of their own risk models. 2004: Basel II Agreements (implementation started in 2006) extend this logic to credit risk also (IRB approach) 2010: New Basel III proposal (detailed below). Implementation phase runs up until Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

45 How to regulate bank capital? Risk adjusted capital requirements Portfolio approach of bank assets [Kareken and Wallace (1978), Crouhy and Galai (1986), Kim and Santomero (1988)]: if banks behave as portfolio managers when they choose the composition of their portfolio of assets, then regulator should use risk-related weights for the computation of the capital-to-asset ratio. Using a mean-variance model, Kim and Santomero compare the bank s portfolio choice before and after a solvency regulation is imposed. They show that the solvency regulation will entail a recomposition of the risky part of the bank s portfolio in such a way that its risk is increased while the volume is decreased (ineffi cient portfolio), unless risk weights are proportional to the market β s of assets. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

46 How to regulate bank capital? Risk adjusted capital requirements Rochet (1992) criticizes the Kim-Santomero approach and adds the limited liability option in the bank s objective function. If this option is correctly taken into account, the corrected utility function U LL (µ, σ 2 ) = 1/ 2π. µ/σ u(µ + tσ) exp( t2 /2)dt. is not always decreasing in σ 2. Then the effi ciency of solvency regulations is jeopardized even more. Even when market-based risk weights are used, it may be necessary to require an additional regulation in the form of an additional minimum capital requirement for banks (in absolute terms), independent of their size. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

47 Basel I: reminder Two powerful advantages: Level playing field for competition among international banks (case of Anglo-saxon banks, rise of Japanese banks in 80s) Simple rules: simple risk classes, RWA covered by minimum ratio of 4% Tier 1 capital (equity less goodwill) and 8% Tier 1 + Tier 2 capital. Note: demise of reserve requirements on deposits to insure liquidity and safety of banks. Focus on liabilities side (capital) only (+LLR function of CB and ample provision of sovereign bonds as liquidity buffers) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

48 Basel II: adjusting risk weights to fight regulatory arbitrage Three pillars: regulatory capital base, risk-adjustment using internal banks models (I), supervisory oversight (II), market discipline (III) Pillar I: coping with credit risk (CR, market risk (MR) and operational risk (OR): RWA = 12.5 (OR + MR) w i A i i Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

49 Basel II capital requirements for credit risk (Pillar I): how does it work? A menu of approaches, depending on sophistication of a banks activities and of its internal risk management capabilities. Standardized approach: based on external credit ratings when available, on Basel I charge (8%)otherwise. Internal ratings-based approach: banks assign exposure to different asset classes ; within each class, banks assign different internal rating grades / to the creditworthiness of (homogenous groups of) borrowers This implies estimating for each borrower: one-year ahead PD, LGD, EAD, Maturity (M) Foundation IRB approach: banks estimate only the PD and take other parameters from the supervisor Advanced IRB approach: banks estimate all parameters Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

50 Basel II: IRB approach Capital requirement based on VaR at 99.9% with one year horizon: Capital required= Maximal possible loss - expected loss (already incorporated in loan pricing) VaR computed using a one-factor Gaussian copula model of time-to-default: Suppose large portfolio of loans, same PD, copula correlation ρ then worst case default rate at 99.9% is: WCDR = N ( N 1 (PD) + ) ρ.n 1 (0.999) 1 ρ Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

51 Basel II: IRB approach (2) Total portfolio loss (worts case) will then be less (at 99.9%) than: ( i EAD i.lgd i ).WCDR Whereas expected loss (priced in or provisioned) is: i EAD i.lgd i.pd Hence capital requirement is: ( i EAD i.lgd i ).(WCDR PD) cf. for details Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

52 Basel II: IRB aproach for corporate, sovereign or bank exposures Assumed correlation between copula parameter ρ and PD (empirical research): ( 1 e 50PD ρ = e (1 ) e 50PD ) (1 1 ) e 50PD 1 e 50 Adjustment for Maturity ] (M) if longer than one year: MA = where b = [ ln(pd)] 2 [ 1+(M 2.5).b b idea: control for changes in creditworthiness during life of loan Bottom line: total risk-weighted assets: RWA = 12.5.[EAD.LGD.(WCDR PD).MA] Hence: required capital = 8% of RWA (half of which must be Tier 1 capital). Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

53 Basel II IRB approach: NB: Correlation decreases with PD (riskier loans supposed to have higher idiosyncratic risk) Source: Repullo, Suarez, Trucharte (2010) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

54 Basel II IRB: Example of capital charge computation Bank with assets $100 mns of loans to A-rated corporations, with PD=0.1% and LGD=60% and average maturity 2.5 years. Implies (see above): b = and MA = 1.59 Tabulated Gaussian WCDR for portfolio gives: WCDR = 3.4% Capital charge under Basel II is ($ mns): ( ).1.59 = 39.3 NB: under Basel I would have been $100 mns, and under standardized Basel II: $50 mns Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

55 What risk-weighting means: difference between capital adequacy and leverage ratios Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

56 Motivations for the Basel 3 proposals The depth and severity of the crisis were amplified by weaknesses in the banking sector: excessive leverage, inadequate and low-quality capital, and insuffi cient liquidity buffers. Amplification by a procyclical deleveraging process and the interconnectedness of systemically important financial institutions. Reforms aim "to improve the banking sector s ability to absorb shocks arising from financial and economic stress, whatever the source, thus reducing the risk of spillover from the financial sector to the real economy." Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

57 Motivation of Basel 3 (2) To strengthen bank-level, or micro prudential, regulation, so as to raise the resilience of individual banking institutions in periods of stress. Add a macro prudential focus, addressing system wide risks (cross-sectional correlations + procyclical amplification). Recognition that micro and macro prudential approaches to supervision are interrelated, as greater resilience at the individual bank level reduces the risk of system wide shocks. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

58 Reform design process October 2009: G20 guidelines December 2009: proposals made public Review of comments received (close to 300 comments from bankers, academics, governments, other standard setters and prudential supervisors, and various other market participants and interested parties) and redrafting of proposals. Validation by the Committee, then the Governors and Heads of Supervision between (July September 2010) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

59 Reform highlights stronger bank capital regulation. new bank liquidity regulation improvements in supervision, risk management and governance, as well as greater transparency and disclosure. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

60 The new enlarged framework of Basel III Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

61 Stronger capital regulation: corrected microprudential approach (1) Raising the quality of capital to ensure banks are better able to absorb losses on both a going concern and a gone concern basis: greater focus on common equity. (see Focus below). Raising the quantity of capital. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

62 Focus: what is "higher-quality capital"? What is capital? Common equity, preferred stock, unrealized gains on asset trades? Deferred tax assets, goodwill, subordinated debt? Tier 1? Tier 2? (see forms below) Most used metric so far: Tier 1 capital = common equity + preferred stock +... Is it the same quality (loss absorbtion capacity)? Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

63 What is capital? US case (FRY 9C form, 2010) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

64 What is capital? US case Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

65 Higher quality capital: Preferred stock vs common equity Easier to see in a dynamic perspective Bank A: Asset $100, capital $6, of which common equity 100% Bank B: Asset $100, capital $6, of which common equity $2 and preferred stock $4 Both banks lose $3 during the crisis. They want to recapitalize. Which bank finds it more diffi cult to issue the missing $3 of equity (to avoid deleveraging)? Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

66 Stronger capital regulation: corrected microprudential approach (2) Common Equity Tier 1 (CET1) = common shares (+ associated share premium + retained earnings + some minority interests - goodwill - deferred tax assets) Tier 1 (loss absorption on a going-concern basis) = CET1 + some perpetual preferred stocks + some CoCos +... Tier 2 (loss absorption on a gone-concern basis) = T1 + subordinated debt + other preferred stocks + other CoCos + general provisions +... Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

67 Stronger capital regulation: new ratios Common Equity Tier 1 (CET1) = 4.5% + 2.5% conservation buffer (+ countercyclical buffer up to 2.5% + G-SIB add-on 2%) Tier 1 = 6% Tier 1 + Tier 2 = 8% Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

68 Basel III: implementation calendar Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

69 Stronger capital regulation: new macroprudential approach New countercyclical buffer to protect the banking sector from periods of excess credit growth. Promoting the build up of capital buffers in good times that can be drawn down in periods of stress Numerous challenging implementation issues Introducing an internationally harmonised leverage ratio: backstop to the risk-based capital measure + aims to contain the build-up of excessive leverage in the system: proposed ratio of 3% from 2013 on. Denominator will encompass on- and off-balance sheet exposures and derivatives. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

70 The new countercyclical capital buffer A single objective: ensure that banking sector has enough capital to maintain the flow of credit when the system experiences stress consecutive to a period of excess credit growth dampening effect on credit growth viewed as "positive side benefit" National decisions and jurisdictional reciprocity Each jurisdiction will be able to use "judgement" to implement buffer add-on (see guidelines below) Upper limit common to all countries. 12-months preannouncement of add-on, immediate effect of reduction during stress Host authorities will set buffer add-on for exposure to counterparties in their jurisdiction. Home authorities will monitor that banks in their jurisdiction correctly calculate buffers according to the geographic location of their exposures. Consequences for international banks. Enforcement: 12 months to increase capital before restrictions on earning distributions. A common reference guide Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

71 The new countercyclical capital buffer (2) Simple guiding variable: credit/gdp gap (extracted using a one-sided HP filter). Rationale: on average good predictor of banking crisis in OECD countries over the past three decades (cf. Borio, Drehman, Tstatsaronis, IJCB, 2010). However, acknowledgement that "does not work well in all jurisdictions at all times". "Judgement coupled with proper communication is thus an integral part of the proposal". Step by step guide to calculating the proposed buffer add-on (cf. BCBS, consultative document, July 2010, Annex 2) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

72 The new countercyclical capital buffer (3) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

73 The new countercyclical capital buffer (4) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

74 Countercyclical buffer (5): simulation for two NL banks (A: large international bank, B: mid-sized local bank) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

75 Basel III: implementation calendar Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

76 Basel III: where do we stand? Basel III Monitoring Report (March 2014) Data as of June 2013, 27 participating countries (voluntary basis) Study covers 227 banks 102 Group 1 banks (international bank with Tier 1 Capital > 3 bn): very good coverage in most countries 125 Group 2 banks (others): not so good coverage Estimates of current capital shortfall induced by full implementation of Basel III requirements Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

77 Basel III: 2014 monitoring report (1) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

78 Basel III: 2014 monitoring report (2) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

79 Basel III: 2014 monitoring report (3) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

80 Basel III: 2014 monitoring report (4) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

81 Basel III: 2014 monitoring report (5) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

82 A critical view of the Basel Tower Some unfair critics, to be first debunked using M&M Pending issues: Complexity and opacity Regulatory arbitrage II and III Conceptual issues unsolved Procyclicality: analysis using Repullo and Suarez (RFS, 2013) Competition issues: shadow banking Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

83 Debunking some fallacies about higher capital ratios Cf. Admati, De Marzo, Hellwig, Pfleiderer (2011): "Fallacies, irrelevant facts and myths in the discussion of capital regulation" Fallacy 1: equity is "idle" (confusion between the two sides of the balance sheet Every dollar of capital is one less dollar working in the economy (Steve Bartlett, Financial Services Roundtable, reported by Floyd Norris, A Baby Step Toward Rules on Bank Risk, New York Times, Sep. 17, 2010). The British Bankers Association... calculated that demands by international banking regulators in Basel that they bolster their capital will require the UK s banking industry to hold an extra 600bn of capital that might otherwise have been deployed as loans to businesses or households. The Observer (July 11, 2010) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

84 "Fallacies, irrelevant facts and myths in the discussion of capital regulation" Fallacy 2: Increased capital requirements force banks to operate at a suboptimal scale and to restrict valuable lending and/or deposit taking More equity might increase the stability of banks. At the same time however, it would restrict their ability to provide loans to the rest of the economy. This reduces growth and has negative effects for all. Josef Ackermann, CEO of Deutsche Bank (November 20, 2009, interview). Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

85 "Fallacies, irrelevant facts and myths in the discussion of capital regulation" Fallacy 3: Increased equity requirements will hurt bank shareholders since it would lower the banks return on equity (ROE). Demands for Tier-1 capital ratio of 20%... could depress ROE to levels that make investment into the banking sector unattractive relative to other business sectors. Ackermann (2010, p. 5.) The problem with [equity] capital is that it is expensive. If capital were cheap, banks would be extremely safe because they would hold high levels of capital, providing full protection against even extreme events. Unfortunately, the suppliers of capital ask for high returns because their role, by definition, is to bear the bulk of the risk from a bank s loan book, investments and operations Elliott (2009, p. 12). Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

86 "Fallacies, irrelevant facts and myths in the discussion of capital regulation" Remember that ROE = ROA + (ROA r).d/e For a given capital structure, ROE does reflect the realized profitability of the bank s assets. But when comparing banks with different capital structures, ROE cannot be used to compare their underlying profitability. Higher equity capital requirements will tend to lower the bank s ROE only in good times when ROE is high. They will raise the ROE in bad times when ROE is low. From an ex ante perspective [as in M&M], the high ROE in good times that is induced by high leverage comes at the cost of having a very low ROE in bad times. Note: in other words, although equity has a higher required return, this does not imply that increased equity capital requirements would raise the banks overall funding costs. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

87 "Fallacies, irrelevant facts and myths in the discussion of capital regulation" Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

88 Costs of higher steady-state capital ratios: some empirics Little empirical evidence available Kashyap, Stein and Hanson (2010) test that the M&M intuition holds ("conservation of risk"): if leverage is lower, the market beta of the bank and the volatility of its stock should be lower too Regression of estimated β and σ of a panel of US bank stocks on annual measures of book equity to book assets (see table) Gives support to M&M principle for calibration purpose. Using the M&M logic, KSH then calibrate the effects of an increase of the required CAR on loan interest rates (see table) along three scenarios. Conclude that the increase in long-run level of interest rate on loans would be very small (below 45 bp for an increase in the capital requirement by 10 pp.) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

89 Kashyap, Stein and Hanson (2010) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

90 Kashyap, Stein and Hanson (2010) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

91 What are the effects of heightened capital requirements: short run costs Stock/Balance sheet costs of holding more equity vs flow costs of increasing capital Theory for balance sheet costs: see previous section on M&M Theory for flow costs: mainly lemon problem mentioned by Myers and Majluf (1984) due eg to opacity of relationship lending Note: association between stocks issues and price declines well documented empirically Theory implies that firms are not adverse to high capital ratio if can accumulate it over time by retained earnings Empirical evidence confirm that flow costs are contained (but potentially higher in crisis times!) Bottom line: no strong basis against higher capital requirements if phased in gradually by retained earnings Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

92 The tower of Basel: increasing complexity "With hindsight, a regulatory rubicon had been crossed [in 1996]. This was not so much the use of risk models as the blurring of the distinction between commercial and regulatory risk judgements. The acceptance of banks own models meant the baton had been passed. The regulatory backstop had been lifted, replaced by a complex, commercial judgement. The Basel regime became, if not self-regulating, then self-calibrating. A revised Basel Accord, Basel II, was agreed in It followed closely in the footsteps of the trading book amendment. Internal risk models were allowed as a means of calibrating credit risk. Indeed, not so much permitted as actively encouraged, with internal models designed to deliver lower capital charges. By design, Basel II served as an incentive device for banks to upgrade their risk management technology. (...)That meant greater detail and complexity." (Andrew Haldane, Jackson Hole Speech, 2012) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

93 The tower of Basel: increasing complexity Basel I: 30 pages Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

94 The tower of Basel: increasing complexity Basel I: 30 pages Basel II: 347 pages Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

95 The tower of Basel: increasing complexity Basel I: 30 pages Basel II: 347 pages Basel III: 616 pages... Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

96 The tower of Basel: increasing complexity Basel I: 30 pages Basel II: 347 pages Basel III: 616 pages... This but understates complexity and opacity: parameter space of a large bank s banking + trading books probably dimension of several millions (Haldane, 2012) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

97 The tower of Basel: increasing complexity Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

98 The tower of Basel: increasing complexity (2) Do not forget the legislative blanket that translates non-statutory Basel rules into national legislations: US: Dodd-Frank Act of 2010 = 848 pages (20 times more than Glass Steagall Act) + rule-making book set up by regulatory agencies (400 pieces) = 8,843 pages at mid 2012 (up to probably 25,000 if same proportions maintained) EU: a dozen of directives and associated rule-books, making up around 2,000 pages in mid 2012 (up to 60,000 when completed?) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

99 The tower of Basel: increasing costs Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

100 The tower of Basel: reporting burden Reporting burden for a large bank: UK: 1974: some 150 entries, 2012: some 7,500 entries US: 1930: some 80 entries, 2012: 2,271 Excel columns Maintenance costs: UK: compliance with Basel III reporting would imply 200 full-time jobs for a European mid-sized bank (McKinsey, 2010) US: compliance with DFA would imply some 1000 full-time job for large bank (Fin. Serv. Committee, 2010) How much is this investment worth in terms of averting future financial crises? Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

101 Horse race: complex RW vs simple capital rules Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

102 Horse race: complex RW vs simple capital rules Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

103 Basel II did not stop regulatory arbitrage The declining trend in RWA Micro evidence on arbitrage associated with the implementation of IRB models (Mariathasan and Merrouche, 2013) Analysis of risk-weighted assets for market risk An example about capital arbitrage and promise shifting using CDS (Blundell-Wignall and Atkeson, OECD Jl, 2010) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

104 Basel II did not stop regulatory arbitrage (2) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

105 Basel II did not stop regulatory arbitrage (3) Micro evidence on arbitrage associated with the implementation of IRB models (Mariathasan and Merrouche, 2013) using: 115 banks that have been approved for IRB adoption 21 OECD countries annual balance sheet data, (Bankscope) Idea: contrast ex post banks that eventually failed and banks that did not Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

106 Basel II did not stop regulatory arbitrage (3) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

107 Basel II did not stop regulatory arbitrage (3) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

108 Basel II did not stop regulatory arbitrage (3) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

109 Variation across banks of computed RWA for market risk (mrwa) Cf. Regulatory consistency assessment program, BCBS, January Objective: assessing potential for heterogeneity of mrwa across banks + highlight role of aspects of Basel rules Sample of 16 global banks. Observation period includes recent Basel 2.5 changes. Two joint exercises: analysis of published reports on mrwa + hypothetical test portfolio exercise coordinated by BCBS Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

110 Variation of mrwa: main findings/public reports Large variation in average published mrwas for trading books, only partially correlated with size/composition of trading positions. Average rwa ranging from 10% to 80% (mostly in 15%-45% bucket) Sizeable share of heterogeneity due to supervisory factors (country or bank levels), eg netting of derivative positions => debate "accuracy" of rules vs common level playing-field! Impact of modelling choices by banks (notably degree on reliance on internal models ranging from 10% to 80%) Overall, insuffi cient quality of disclosures (opacity for investors) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

111 Variation of mrwa: main findings/public reports Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

112 Variation of mrwa: main findings/public reports Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

113 Variation of mrwa: reliance on internal models Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

114 Variation of mrwa: reliance on internal models Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

115 Variation of mrwa: main findings/test portfolio exercise BCBS designd 26 dummy portfolios (covering all major risk factors, made of vanilla products mainly) for banks to compute associated IRB metrics (VaR, SVaR, IRC) over 20 trading days Variation of results largely due to banks model choices Higher for new, more complex IRC (incremental risk charges) models than for VaR and Stressed VaR Ex. of choices: length of data period used for calibration, aggregation approaches, scaling factors (from 1-day to n-day risk estimates), calibration of transition martices etc. Variations caused by differences in supervisory multipliers applied to raw output of models (from 3 to 5.5), explaining about 25% of total variation in estimated mrwas. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

116 Variation of mrwa: test exercise results for most diversified PTF Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

117 Variation of mrwa: test exercise results for most diversified PTF Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

118 Variation of mrwa: test exercise results (VaR approach) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

119 Capital arbitrage and promise shifting using CDS Still massive incentives to use "complete market techniques" (shorting credit by buying CDS) to avoid capital charges and reduce tax burdens for clients, thereby maximizing returns. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

120 Capital arbitrage and promise shifting using CDS: fictive example Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

121 Conceptual deficits of Basel capital regulation See Martin Hellwig (2010, "Business as usual?"): A lack of systematic analysis by the BCBS of why previous regulation failed and how the proposed regulation will succeed in preventing another crisis. However, Basel regulations have implied a tremendous amount of work and sophistications. Lots of even unasked questions: Why could major banks manage their risks and equity in a way that contributed to the crisis? Why was bank capital so low that doubts about sovency arouse so rapidly after the onset of the crisis (and interbank market froze)? What assurance do we have that, would Basel 3 fixes have been implemented, the system would have fare better? Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

122 Conceptual deficits of capital regulation "Regulatory capture by sophistication" (Hellwig): the successive versions of Basel all strived to improve risk calibration of capital requirements ("fascination" for model based approaches) Root of the problem may not be deficiencies in risk modelling but rather incentives of bank managers to "economize on equity" to increase "shareholder value" in the short-run. High leverage is the quickest way to achieve it. Basel has too much focused on risk calibration and neglected the issue of governance: regulation is needed for bankers to internalize large externalities of bank risk taking discrepancy between private interests of bank managers and public interest in financial stability Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

123 Conceptual deficits of risk models As acknowldged by BCBS, implemented risk models failed to account for: risks arising from correlations of credit risks in mortgages, MBS and CDO: role of systematic/common macro risk factor risks arising from correlations between counterparty credit risks and underlying risks in derivatives, when the counterparty concludes many similar contracts: case of AIG and monoline insurers asset fire sales by large institutions, asset price falls and liquidity spirals (reason for bailing out LTCM in 1998). As not acknowledged by BCBS however, more fundamental problems remain unaddressed: fragile empirical basis: short time-series, non-stationarities, rare events many risks are not exogenous but endogenous. Ex.: counterparty credit risk in derivatives and total exposure of the counterparty to similar contracts. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

124 Precise objective of bank capital regulation needs being clarified Three different views on how capital regulation may increase bank safety, generally named in regulatory sources: 1 capital is a buffer against losses 2 capital reduces incentives for taking excessive risks 3 capital buys time: more room for intervention by supervisor in case of stress Not mutually consistent: (1) calls for calibration / overall risk, (2) for calibration / incremental risks from marginal assets, (3) for calibration / ease with which assets can be disposed of during intervention Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

125 Dynamic effects of bank capital regulation require more attention Limits of an understanding of bank capital regulation based on three period models...new dynamic models are required. Regulatory buffers have no absorbing function if they have to be met at all times. In real world with ongoing financing and investment decisions, dynamics of implementation of capital requirements over time needs more attention. Cf. dynamic provisioning / deprovisioning schemes and Basel 3 countercyclical capital ratios. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

126 Where should we head? (Hellwig, 2010) Take distance with the principle that regulatory capital must be finely attuned to risks taken by banks risk calibration of capital requirements has largely contributed to decline in effectice capital ratios and systemic interconnectedness through incentive to (optically) put risks off balance-sheet (securitization) equity should aim at protecting against unknown risks (while provisions are set against recognized risks) cf. Basel 3 leverage ratio We should aim at substantial increase of capital requirements, based on systemic concerns with 20% equity ratio, doubts about solvency are unlikely to arise (key ingredient in market freeze of 2007) effects of deleveraging in times of stress mechanically lower Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

127 Procyclical effects of bank capital regulation Cf. notes on Repullo and Suarez, RFS, Procyclical effects of bank capital regulations feature high in regulatory reform agenda since 2007 Mechanism: in recession, losses erode banks capital while risk weights increase. Constrained issuance of new equity implies deleveraging and credit crunch. Model capturing key trade-offs in the debate. Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

128 Leverage and competition issues Puzzle: How to reconcile finding of small effects of lower bank leverage on rates charged by banks with observed high level of leverage at which banks operate (which implies some risks)? KSH argue that solution has to do with nature of competition in financial services. Competitive advantage relies much on cheap funding. Thus small differences may matter. Smaller banks less leveraged: role of relationship lending? Increased competition due to deregulation of intra and interstate branching and Higher capital requirements may reduce systemic risk by stopping one dangerous form of competition But they may also increase the competitive edge of unregulated entities (shadow banks) Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

129 Kashyap, Stein and Hanson (2010): impact of increased competition on distribution of leverage within US states Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135

130 Kashyap, Stein and Hanson (2010): impact of increased competition on distribution of leverage within US states Mésonnier-Renne (Banque de France, Financial Economics Macroprudential Research) analysis This version: october / 135