Macroeconomics Class 7 Financial Crisis Financial markets the

Macroeconomics Class 7: Financial Crisis

Financial markets – the common view: an instrument for hedging and diversifying risks that are generated exogenously. The question: is it possible that in some cases the market itself becomes an endogenous source of extreme risk and instability? (And if so, what can be done about it? ) Exogenous risk Endogenous risk (a bank run) 2

Example: “contagion” during the Asian Crisis • “A contagious disease continues to spread through emerging markets. After the Thai baht was devalued in July 1997, stock-markets and currencies tumbled across Asia. Several Latin American markets then caught a nasty chill. And now turmoil in Russia is sending shivers through almost every emerging market. ” The Economist, August 20, 1998 3

Two theories: (in both, liquidity plays a key role) A theory of bank-runs: Diamond and Dybvig • the source of the problem: liquidity-creation by the banking system – via demand deposits – or, more generally through asset transformation • results, sometimes in a bank run (rational panic) The “Fisher effect” • financial crisis is a result of excessive leverage – resulting in fire-sales – price spirals, contagion. . . 4

Asset transformation: banks’ ability to create liquidity by issuing liquid liabilities against non-liquid assets Deposits tend to be much more liquid • relative to the investments Bank X Balance Sheet Particularly demand deposits • that can be drawn on demand – any time – no questions asked – at par value Reserves Investments Demand deposits Fixed-term deposits Subordinated debt Equity What happens if there is a panic • and depositors try to withdraw simultaneously? 5

Detour: liquidity transformation and the money supply. Inside money versus outside money Outside money – banknotes issued by the central bank • includes coins + paper money • also called the money base • or M 0 The money base is held by • the banking system as reserves • and by households and companies as currency in circulation The money supply (M 1) is all means of payments • currency in circulation + demand deposits – demand deposits are as liquid as money 6

The accounting of money creation The Central Bank Gold Money Base M 0 The Private Banking System Reserves R credit demand deposits D M 1 currency in circulation C Notice: demand deposits are not fully covered by reserves! 7

Conclusion: the financial system creates money M 1 > M 0 The Central Bank Gold Money Base M 0 The Private Banking System Reserves R credit demand deposits D M 1 currency in circulation C with some algebra • M 1= C+D = (C+R) + credit = M 0 + credit Namely, the Money supply is made of • outside money: M 0, • inside money: credit, extended by the private banks 8

A theory of liquidity transformation by banks: Diamond and Dybvig. What purpose does it serve? Banks intermediates between: Investors who need immediacy, cash at a short notice • due to various emergencies – due to change in personal circumstances, health, career etc. And long-term projects • offering a high yield • but investment is locked in for many years 9

A simple example: time, t = 0, 1, 2. The demand for immediacy Ex post: after uncertainty is realized Ex ante: before uncertainty is realized t=0 t=1 three consumers each has 50 units of wealth at this point, identical one of the three faces an emergency, call him Type E, and needs resources now! Type is random and non verifiable t=2 the others (Type N) can defer consumption 10

Technology: long-term (illiquid) investment is more productive than short-term investment (storage) t=0 t=2 t=1 0% rate of return Liquid investment 0% rate of return +100% return to maturity Long-term investment Denoted by K liquidation cost: 1/3 of the investment is lost 11

A demand deposit scheme can even offer a positive return to the E -type, though the N-type must be offered a higher rate, otherwise he will withdraw his money at t=1; suppose rates are 20% and 80%, respectively. The bank must keep liquid reserves! t=0 Reserves 60 Credit 90 t=1 Deposits 150 Invest. 90 t=2 Deposits 90 liquidity 180 Deposits 180 90 (1+100%)=180 Type-E withdraws 60 50 3 Type-E consumption: 60. Type-N consumption: 90 Type-N gets 90 each 12

Claim: demand deposits also create (endogenously) the risk of a bank-run or a panic! Technically: two Nash equilibria. Non-verifiability → clients served on a first-come first-go basis • Type E always attempt to withdraw • Hence, the game is played by the Type-N players: P 1 and P 2 – each may choose between ‘running’ and ‘staying’ (R and S respectively) Expected in a bank-run equilibrium liquid resources: 60 liquidated capital: (2/3) 90=60 probability of service: 2 out of 3 total consumption (three consumers): 120 The good equilibrium total consumption: 2 90+60=240 13

Conclusions (so far. . . ) Financial crisis is no madness • unlike Mackay (1841) and so many others. . . Running on the bank is rational • a best response to others’ run A bank-run may hit a fundamentally sound bank • solvent at maturity • a bank-run is a coordination failure by rational players 14

Claim: the government should resolve this market failure via public deposit insurance, which costs nothing If a bank-run occurs • a tax, T, is levied on all the players who withdraw • to fully compensate those who did not manage (or tried) to withdraw (2/3) (60 -T)+(1/3) 90=40 2 T=90 60 -T 2 T=90 Once a deposit-insurance scheme is put in place • the bank-run is no-longer an equilibrium outcome The scheme costs nothing, because the insurance eliminates the run! 15

Claim: the bank-run equilibrium can also be eliminated through a private policy of suspension of convertibility: the bank should stop serving clients once it is out of liquidity (1/3) 60+(2/3) 90=80 (1/2) 60+(1/2) 90=75 Notice: once suspension of convertibility is incorporated into the contract • it is never implemented in equilibrium! Implication: a bank-run is not really a market failure • just a result of a defective contract that markets can fix on their own 16

Example: suspension of convertibility by Alfa Bank (Russia) • In the summer of 2004, the authorities withdrew licenses from several Russian banks in distress. • On July 6 2004, a newspaper, Kommersant, has alleged – wrongly – that Alfa Bank was linked with Guta Bank, which has just failed. • Within 3 days, depositors withdrew $200 M, 14% of total deposits. • Alfa bank has responded by imposing a 10% penalty of withdrawing deposits. • At the same time, extra liquidity was arranged from HSBC and the Russian Central Bank. • Eventually, Alfa managed to regain clients confidence and carry-on operations. 17

Milton Friedman’s explanation of the Great Depression: State versus markets in dealing with crisis The FED was established in 1917 • before, the US relied on a private system: – the New-York Clearing House Association, established 1853 – essentially, a club of city banks During the years of its existence, it • coordinated suspension of convertibility • acted as a “lenders of last resort” • expelled underperforming banks from the club According to Friedman, the system worked quite well • although panics did occur (every 7. 5 years, on average) 18

Milton Friedman: State versus markets The FED was created, because – in theory – the State can do it better • it has greater resources • can impose taxes, print money and supply liquidity But when the Great Depression came • the FED has failed to act • and the private system was no longer there The typical Friedman conclusion • no regulation is better than bad regulation 19

Irving Fisher’s theory of the Great Depression • Due to unanticipated deflation – and given that debt is denominated nominally – the real value of corporate debt has gone up – thinning out companies’ capital base • and hence, increasing leverage – increasing the incidence financial distress and bankruptcy: recession 20

Fisher’s evidence From 1929 to 1933 • money price level fell by 43% • if debt is denominated in money • such a development may seriously weaken companies balance sheet Fisher, Econometrica 1933 For example: After prices fell: P=57 (denominated in money) Before prices fell: P=100 Assets 100 Debt 50 Assets 57 Capital 50 Leverage increased from 50% to 88%! Debt 50 Capital 7 21

According to Fisher’s own calculations • The US corporate sector • Was wiped out of capital due to deflation Fisher, Econometrica 1933 22

Same sort of calculation for the current crisis The Economist, February 12, 2009 23

In the last 20 years, the old pattern has re-appeared in US data Source: Federal reserve Bank of Cleveland 24

International evidence: 27 countries since 1980 % (debt-service to income) quarters from crisis Source: Drehmann and Juselius, BIS (2012) 25

Another important link: high leverage → financial distress → repossession and fire sale → depressed collateral value → higher (market) leverage. Evidence: the US narrow-body aircraft market Pulvino, JF, (1998) • Sample: 704 second-hand-market transactions of narrow-body aircraft sales by US companies during the years 1978 -1991 • Fire sales (by high-leverage owners) go at a discount • Providing great bargains for low-leverage 26

Pulvino (cont. ) i. e. low leverage i. e. high leverage boom bust 27

A modern exposition of Fisher effect (based on work co-authored with Alexander Guembel). Step 1: project valuation in a non M&M world • Under M&M, the value of a project is independent of leverage • In reality, banks would not lend to an entrepreneur with capital less than A value M&M NPV expected loss due to liquidation • At B, the bank would lend, but high leverage would pose a high risk of repossession due to temporary cash shortage, decreasing the project’s value A B C owner’s capital→ ← leverage

Project valuation in a non M&M world (cont. ) • Only at point C, where the project is self financed, will it reach the M&M valuation (and stay flat beyond that point) value • Clearly, if a B-like project suffers a fall in collateral value, market leverage would increase, and valuation would fall (due to a higher risk of repossession) M&M NPV A B expected loss due to liquidation C owner’s capital→ ← leverage

Numerical example: consider a (non-listed) business Suppose • q: market price of investment goods • W: wealth (equity) of owner • D: debt Under M&M 30

Numerical example (cont. ): a non M&M world • Suppose that the bank demands 50% (on loan) security, S • 9% probability of financial distress – PV is still 100 – cash shortage might lead to default • Owner has to pledge a 1/3 of business as security, – which he will lose in case of distress 31

Numerical example (cont. ): suppose that the business is funded by a revolving credit line; suppose that, after funding, market price of investment goods fell 50% • Now, giving the bank security with a market price of 30 means pledging 2/3 of the business! 32

Step 2: the supply of fire sales is downwards sloping! lower q → higher leverage → more fire sales fire-sale price At the point of zero supply • the amount of collateral is zero • expected loss from liquidation is zero • and therefore, the valuation is according to M&M supply of fire-sales fire sales 33

Step 3: the demand for fire sales, also downwards sloping. Who buys fire sales? Speculators who store liquidity and wait for bargains. fire-sale price • At low levels of fire sales they bidup the price to its re-deployment value, δ, (60 in the numerical example) demand curve for fire-sale assets δ • Once liquidity is fully used, prices start dropping • down to the level of q q rock-bottom price fire sales

Implication I: financial crisis comes about when a tiny (external) increase in liquidation start a down-spiral fire-sale price • Lower collateral prices supply • Higher leverage δ • More liquidations • Ending in a black hole fire sales

Implication II: in crisis, there is a multiplier effect fire-sale price • Small changes in demand supply • Bring about big changes in equilibrium prices and quantities δ fire sales

Note: the multiplier effect is a direct implication of both supply and demand slop downwards, a pathology fire-sale price • Had supply been upwards sloping supply • The multiplier effect would not exist δ A B fire sales 37

Implication III: the competitive-equilibrium probability of crisis must be positive • Speculators lose money in normal times fire-sale price • but they make money in crisis (buying at a rock-bottom prices) • and holding the investment goods until fire-sales recover to the redeploument value supply δ fire sales

Implication IV: multiple equilibria Possibly, • there will be three equilibrium points, A, B and C • clearly, point A is better for the economy – higher prices – smaller collaterals – less fire sales fire-sale price supply δ A B C fire sales 39

Multiple equilibria (cont. ): what pushes the economy from A to C? pessimistic self-fulfilling expectations fire-sale price If markets believe q (at C) • lenders will demand more collateral • which will increase liquidations • which will result in lower prices supply δ A Like in the Diamond-Dybvig model • by guaranteeing that “all is fine” • a government can lift the economy – up from point C to pint A • at no fiscal cost B C fire sales 40

. . . but that will work only if the shock is not too big fire-sale price supply δ fire sales 41

Implication V: most significantly, a competitive equilibrium is no longer Pareto optimal Unlike in a “normal” market fire-sale price owners NPV drops as prices fall • supply and demand • do not reflect “fundamental” values q – since fire sales are driven by temporary cash shortages – fundamentals are still at M&M values • Hence, market equilibrium is not economically efficient fire sales 42

. . . and a bailout policy – injecting capital into distressed companies in order to limit financial distress – creates value! Policy steps fire-sale price • bailouts, that would decrease the amount of fire sales • liquidity injections, that would prevent trading at a rock-bottom price supply q fire sales 43

. . . but that might transform a corporate-debt crisis to a sovereign-debt crisis Reinhart and Rogoff, 2008 44

Corporate-sovereign swap: evidence from Ireland September 30, 2008: the Irish bailout Source: Acharya, Drechsler and Schnabl, 2010 45

Policy questions • Which theory of the crisis is better suits the facts? • How suitable were the policy steps taken by western governments? • What other policy steps could have been deployed? 46

The most highly-levered industry Source: Alessandri and Haldane, 2009 47