LIQUIDITY RIDGE SUMMER SCHOOL Montevideo December 2015 RIDGE

LIQUIDITY RIDGE SUMMER SCHOOL Montevideo, December 2015 RIDGE SUMMER SCHOOL DECEMBER 2015 1

Contents 1. 2. 3. 4. 5. Motivation Holmstrom-Tirole approach Cash in the Market and Asset Fire Sales Interbank Market and Banks’ Liquidity Lender of Last Resort RIDGE SUMMER SCHOOL DECEMBER 2015 2

Motivation: the 2007 liquidity crash • Collapse of market liquidity with the so-called « toxic assets » . • Freeze of the interbank market. • Increased liquidity hoarding by banks. RIDGE SUMMER SCHOOL DECEMBER 2015 3

Liquidity freeze RIDGE SUMMER SCHOOL DECEMBER 2015 4

The Freezing of the interbank market • No pure liquidity risk: counterparty risk increases too. 5

Haircuts RIDGE SUMMER SCHOOL DECEMBER 2015 6

Holmstrom-Tirole liquidity shocks • Firms invest at time t=0 • Face a cost overrun/liquidity shock at time t=1, if liquidity is not injected it implies a zero return • Projects mature at time t=2 • No aggregate shock: firms invest in cash and at time t=1 there is a market for the liquid asset. • Aggregate supply shocks implies “cash in the market” RIDGE SUMMER SCHOOL DECEMBER 2015 7

Holmstrom Tirole (II) • Firms’ managers may choose a project with negative net present value and private benefits if leverage is too high. This implies there is a limit to the firms’ leverage and size. • The market solution is inefficient because of the lack of commitment of private investors. RIDGE SUMMER SCHOOL DECEMBER 2015 8

Holmstrom Tirole (III) • Case 1: no aggregate liquidity shock – If firms can carry liquidity at no cost the second best is reached. If this is not the case, the second best cannot be reached by holding securities on other firms. – Credit lines allow to reach the second best, which justifies the existence of banks. • Case 2: aggregate liquidity shock – The Treasury is able to provide liquidity because it can tax future generations RIDGE SUMMER SCHOOL DECEMBER 2015 9

Cash in the Market and Asset Fire Sales A simplified approach assumes a segmented market with limited resources, M, from investors. If S is the supply of assets, in equilibrium, p. S=M, for p below some threshold. RIDGE SUMMER SCHOOL DECEMBER 2015 10

Liquidity and cash in the market • • • Acharya and Yorulmazer (2008) n banks A number k (k<n) of banks fail These banks sell their assets to the surviving ones Their cash determines the price of the banks’ assets RIDGE SUMMER SCHOOL DECEMBER 2015 11

Acharya and Yorulmazer results • Banks’ available liquidity is limited and depends upon the number of banks failing. This will determine the asset price as a function of k. • If there is not enough cash outsiders may buy the banks’ assets and put the assets to an inefficient use • Bailouts are justified. RIDGE SUMMER SCHOOL DECEMBER 2015 12

Banks’ optimal selling policy • Diamond and Rajan(2011) raise the question: why don’t banks prefer to liquidate their assets during the previous period. • They show that solvent banks prefer to do so; banks in distress are better off by taking the risk of becoming illiquid. RIDGE SUMMER SCHOOL DECEMBER 2015 13

Haircuts Subsequent 10 % haircut LIABILITIES ASSETS LIABILITIES Initial 5% haircut ASSETS ABS 100 REPO 95 EQUITY 5 100 ABS 50 50 RIDGE SUMMER SCHOOL DECEMBER 2015 REPO 45 EQUITY 5 50 14

Repo borrowing and price drops RIDGE SUMMER SCHOOL DECEMBER 2015 15

Brunnermeier and Pedersen • Cash in the market leads to asset price volatility. • Could be mitigated by funding liquidity. • Speculators (Investment banks) need access to funding to intervene in the markets. They get these funds through repos • Margins (Haircuts) are set so as to preserve a given value-at-risk (therefore they depend on volatility that depends upon shocks to fundamentals) they depend upon the volatility of the assets • Total value of margins on their positions cannot exceed their capital • Limited access to credit by speculators implies that prices diverge from fundamentals. RIDGE SUMMER SCHOOL DECEMBER 2015 16

Liquidity Spirals RIDGE SUMMER SCHOOL DECEMBER 2015 17

Interbank Market and Banks’ Liquidity 1. The basic Diamond-Dybvig-Bhattacharya. Gale model 2. Interest rate formation 3. Unsecured interbank market freeze 4. Secured interbank market freeze RIDGE SUMMER SCHOOL DECEMBER 2015 18

Bhattacharya and Gale(1987) • No aggregate risk, but bank idiosyncratic liquidity risk related to the random amount of withdrawals at time t=1. • If banks operate in isolation, the bank can only offer contracts contingent on its own liquidity shocks • The creation of an interbank market allows to restore efficiency. RIDGE SUMMER SCHOOL DECEMBER 2015 19

Generalizing Bhattacharya and Gale • • • Freixas, Martin & Skeie (2010) Consider a probability ρ of idiosyncratic liquidity shocks. For ρ=0 Diamond Dybvig obtains For ρ=1 Bhattacharya and Gale obtains In between a continuum of REE because of the inelasticity of demand supply in the interbank market. • The efficient equilibrium is characterized by a low level of interest rate in the liquidity crisis state. • Implications for monetary policy in a (2007) liquidity crisis 20

Allen and Gale • Build upon Bhattacharya and Gale (87) • Case 1: Zero aggregate risk – To prevent unnecessary liquidation banks make cross deposits or commit to lend – This allow banks to hold less liquid assets • Case 2: With zero probability there is aggregate risk – The aggregate amount of liquidity is insufficient: systemic risk • Distinguish different financial architectures 21

Freixas Parigi Rochet (2000) • Concern on systemic risk • Is a liquid interbank market a sufficient guarantee against a liquidity crisis affecting one institution? Or should the LOLR step in? • What are the determinants of contagion through the interbank market? • Two solutions: the efficient one and the “gridlock solution” 22

EXAMPLES OF « TRAVEL PATTERNS » (MATRIX T) 1 Ø CREDIT CHAINS 3 2 Ø DIVERSIFIED LENDING Ø MONEY CENTER BANK 1 2 3 3 1 2 23

Rochet and Vives (2004) • The liquidity coordination problem: banks have different opinions about other banks solvency. • Banks are willing to lend to solvent banks provided they don’t have a strategic risk stemming from other agents withdrawing • Use the Van Damme Morris-Shin global games approach that allows to solve for coordination within a game. 24

Rochet and Vives results • In equilibrium there exists a critical value of banks’ assets such that, whenever the value of the bank’s assets falls below this threshold, the banks will not have access to liquidity. • This threshold is above the solvency threshold. 25

Asymmetric Information and Interbank Market Structure • Freixas and Holthausen (2004) • Motivation – Understanding Europe Money Markets integration – Understanding external debt crises in emerging countries (Tequila crisis, East Asian crises) 26

Asymmetric Information • Countries/Regions Assumption: – better soft information about domestic borrowers • Questions: – Can an integrated market always emerge? – Compare the combined role of repo and unsecured markets 27

Country H Country L 1 - (s. G, s. G) (s. G, s. B) L-captive H-captive (s. B, s. G) (s. B, s. B) No credit 28

Equilibrium (II) • Result: – Segmentation ( = 0) is always an equilibrium – Multiplicity of equilibria possible – For some parameters only the segmented equilibrium exists 29

Liquidity Dry-ups • Malherbe (2014) • Banks cover their liquidity needs with T-Bills or with the interbank market • Adverse selection in the interbank market • Multiplicity of equilibria 30

Liquidity and information sensitive assets • Why the toxic assets illiquidity of 2007? • Is transparency required for markets to function? • Dang Gorton and Holmstrom argue the issue is adverse selection • Distinguish: information insensitive debt vs. Information sensitive debt. 31

Dang, Gorton and Holmstrom • Motivation: complexity of ABS prior to the crisis • Opacity need not prevent trading • Adverse selection leads to a collapse of the market because of the Akerlof result (market for lemons) • Implication: distinguish information insensitive securities from information sensitive securities RIDGE SUMMER SCHOOL DECEMBER 2015 32

Dang Gorton Holmstrom Density Information sensitive Information insensitive Debt Holders return Asset value RIDGE SUMMER SCHOOL DECEMBER 2015 33

Lender of Last Resort RIDGE SUMMER SCHOOL DECEMBER 2015 34

Liquidity Injection Diagram 1: Central bank total liabilities RIDGE SUMMER SCHOOL DECEMBER 2015 35 35

Interest Rates Policy Diagram 2 RIDGE SUMMER SCHOOL DECEMBER 2015 36 36

New Channels for Liquidity Injection • In the Eurozone – Broadening of the class of assets that can be used as guarantees in repo operations. – LTRO: ECB buys sovereign bonds. • In the US: three types of targets 1. Banks liquidity provision (TAF, TSLF, PDCF). 2. Final user liquidity provision (TALF, CPFF, MMIFF). 3. Long term Debt acquisition. RIDGE SUMMER SCHOOL DECEMBER 2015 37

Bagehot’s rules • Only illiquid solvent institutions should have access to credit • Loans should be made against good collateral at a penalty rate • This rules should be made public 38

Are these rules realistic? • Goodhart(1985, 1987) asserts that the distinction between illiquidity and insolvency is a myth 39

Goodfriend and King • LOLR only at the aggregate level through open market operations • Unsecured interbank market will then redistribute liquidity to illiquid solvent banks • This implies market discipline • Humphrey (1984) argues this would be Bagehot position today 40

Central banks: Liquidity provider or Crisis manager? • Bagehot, Allen and Gale: because of the risk of aggregate illiquidity, the Central Bank has a monetary policy role. • Rochet and Vives(2004), FPR(2004), role as lender of last resort to solve market inefficiency. 41

Central banks: : Liquidity provider or Crisis manager? (II) • FPR (2000), because of the risk of gridlock the Central Bank has a coordinating role as crisis manager and liquidity provider • FPR (2000) because bank closure may lead to contagion, role in the orderly closure of the insolvent bank. 42

Maturity mismatch and the Central Bank put • Farhi and Tirole (2010) on collective moral hazard. • What are the incentives of Central bank liquidity injection on banks portfolio (liquidity) decisions? • What are the banks’ decisions? – Banks strategies regarding their holding of a liquid asset are strategic complements. – Two equilibria emerge: one with banks holding liquidity and the other with banks expecting to be bailed out by a generous monetary policy. RIDGE SUMMER SCHOOL DECEMBER 2015 43