CRISIS: THE NEXT GENERATION ? By Paul KRUGMAN (2001)

CRISIS: THE NEXT GENERATION ?

By Paul KRUGMAN (2001)

A model for each contextGeneration Crisis model

1° GCM Crisis of Bretton Woods (1971) Krugman (‘79); Flood and Garber (‘84)2°GCM European Monetary System crisis (1992) Obstfeld (‘94)3°GCM Asian crisis (1997-98) Krugman (‘99)4°GCM Internet Bubble (2001), Depression 2007 Krugman (’01)

1. First generation

Effort of CB to peg an exchange rate using reserves

Irresponsible economic policy => Long-run upward trend in shadow price

Shadow price > peg =>speculation attack=> investors advance crisis

Irresponsible economic policy => Predectible and deterministic crisis

No => recession

2. Second generation• No run out of reserve but policy choice• Peg in uncomfortable level => strict monetary policy

=> reduction of employment in the short term• If peg ceases to be credible => i increases => worsen

employment + increase financial distress => selfulfilling crisis of confidence => more speculative attacks

• Crisis are not the result of irresponsible policy• Crisis can be sudden and no predictable• Crisis is still harmless

3. Third generationSimultaneous currency crisis and bank crisis => «twin crisis », three main variant models:Moral hazard driven investement: Mc Kinnon and Pill (199?)Open-economy version of Diamond-Dybvig bank-run model: Chang and Velasco(1998a,b)Balances sheet implications of currency depreciation

Third generationHypothesis: Small open economy One good Two périods a Mundell-Flemming framework with

(1) y = D(y, i) + NX(eP*/P, y)

(2) M/P = L(y, i)

(3) i = i*

A more realistic version of 3°GCM

To permit this simplest model to generate crisis, Krugman adds a strong balance sheet effect from currency depreciation

(1') y = D(y, i, eP*/P) + NX(eP*/P, y)

(2') M(e)/P = L(y, i)

Sudden large depreciation creates havoc on balance sheet => crisis

« Original Sin »: key problem of 3rd Generation Crisis models

• Why did local firms and banks borrowed in foreign currency, hence facing an enormous exchange rate risk (irrational).

• Foreign investors protect themselves of currency crisis by lending in dollars in emerging economies.

• Local firms and banks: Benefit: more foreign capital. Cost: much larger probability and costs of financial crisis.

Traditional answer to crisis

• IMF financial support: The “sterilized intervenor of last resort”

• Rollovers and standstills: Alteration of the composition of capital flight but not volume

• Monetary policy: To rule out self-fullfilling crisis

• Fiscal policy: an expansive policy to jump out of the bad equilibrium

• Structural reform: Increase confidence

Fourth generation (this paper)

4th generation model is a generalization of the 3°GCM: not only the price of foreign currency but all asset prices

•OPEN-ECONOMY

•CLOSED-ECONOMY

Fourth generation

The balance-sheet matter:

In an imperfect capital market the ability of firms to exploit even profitable investment opportunities may depend on their ability to provide sufficient collateral to let them borrow the needed funds. If the price of foreign exchange rises, and firms have foreign currency debt => their net worth falls => reduction of their collaterals Likewise a decline in confidence => declining asset prices => a fall in investment => validates both the decline in asset prices and the fall in confidence.

HYPOTHESIS

• a small open economy, producing a single tradeable good• N investors, each must borrow B of the good to get themeselves

into the business• Each investors is also endowed with an equal share of a productive

resource ( the total quantity of that resource equal to K)• 2 periods: in 0 investors can or cannot borrow “seed money”

necessary to get start; in 1 who has mad the initial investiment of B can choose to produce according to a production function F(k)

• K = amount of resources used• investors may use either more or less than he owns, selling any

surplus for a price q determinated ina competitive market (buyers=borrower and seller=lender)

• n<N = potential investors actually went ahead.• r = real interset rate (exogenous)

The case of perfect market…

(4) q = F’(K/n) price is increasing in K(5) EP = S(q)/(1+r) – B profitS(q) is the “surplus” earned in period 1 over and above the cost.S(q) will be decreasing in q;

so from (4) and (5) we see that the profitability of investing is decreasing in the number of actual investors. If capital markets is perfect => there is a unique equilibrium value n* between 0 and N

Discussion: equation (4)Valuation of assets.

• (4) q = F’(K/n) Price is decreasing in K if decreasing returns to scale.

• In intertemporal models: future marginal productivity net of marginal costs, discounted, for marginal q(t).

• Discount rate: (1) economics: time preference + risk premium.

• Discount rate: (2) finance: risk free return + risk premium of an alternative asset, which belong to the same class of risk of the asset to be valued (opportunity cost and no opportunities of arbitrage).

T

tti

tFlowCashEtice

1 )1()0(Pr

T

tt

T

tt

i

tKCtKFEtq

i

tFlowCashEtice

1

1

)1(

))(('))((')0(

)1()0(Pr

Asset Price EquationValuation of Assets (cf. Efficient Market

Hypothesis) C(K): adjustment costs.

Profit Equation

(5) EP = S(q)/(1+r) – B profit

Profit is discounted by (1+r) by Krugman.

Profits = pY-wL –q(t)(K(t)-(K(t-1)) – (1+r)(B(t).

Investment (new capital) is bought at the cost q(t).

S(q) will be decreasing in q.See Kiyotaki and Moore flow of funds equation.

…and the case of imperfect market

• Problem of monitoring • The lender’s only recourse in the case of non-payment is

the ability to seize the borrower’s marketable resource in period 1. So the lender will not lend more than the borrower’s collateral:

• (7) B < (qK/N)/(1+r)• Suppose that (7) is always binding => that investing is

always profitable, if the seed money can be borrowed. • But q is increasing in n => each investor will invest only if

enough other investors are also expected to invest => his collateral is worth enough to persuade lenders to give him the necessary seed money.

Collateral constraint

(7) B < (qK/N)/(1+r)

For dynamical models: The debt repayment is guaranted in expectation by the expected future value of collateral.

(1+r(t)).B(t) < m.E(q(t+1)).K(t)/N

What matters is the expected price next period for a collateral constraint, not the current one.

A transaction cost m matters. Not all durable assets K(t) are “good” collateral.

Two Equilibria

One equilibrium is with all N potential investors investing => a high q => each investor can offer sufficient collateral to raise the necessary seed money.

The other equilibrium is with no investment: In this hard-edged model q=0 => nobody has collateral => nobody can invest. Self-fulfilling pessimism can cause investment to collapse, not because of the exchange rate and transfer problems stressed in the third-generation crisis models, but because of the effects of confidence on domestic asset prices: FIRE SALES and ILLIQUID ASSET MARKETS.

4°GCM IN A CLOSED ECONOMY

Predominance of domestic assets market rather than currency

Some nominal stickiness, in which: ( q is tobin’s q) => Investments => output

(A1) y = y(q)

(A2) q = q(y,i) higher y => higher profits => higher

(A3) i = MIN(i(y),0) monetary reaction function, positive nominal rate constraint (liquidity trap).

Equation 1: Output and asset prices

Y=C(q1,q2)+I(q1)+H(q2)+G(q)+(X-I)(q5).H=housing investment, price q2.I=Corporate investment, price q1.H and I: Tobin’s q positive effect on

investment.C: Wealth effect on consumption.G: Fiscal policy endogenous reactions.(X-I): Exchange rate q5 policies.

Non linear assumption Y(q)

May be:

• Low output: increases of asset prices have little increasing effects on output.

• High output: increases of asset prices have little increasing effects on output.

Discussion: equation (4)Valuation of assets.

• Asset Price is increasing in expected cash flow, which increases with future output Ey(t+1) (possibly auto-correlated with current output).

• Asset Price is decreasing with the discount rate, which includes the monetary policy (risk-free) rate i. But Monetary Policy Reaction function (Taylor Rule) imply that the interest rate increases with expected output. So Ambiguous effect of OUTPUT on ASSET PRICES.

• With liquidity trap, the second effect (increasing output decreases asset price) disappears.

mRiskpremiutEyiMIN

tEyG

i

tFlowCashEtq

T

tt

]0)),1(([1

))1((

)1()0(

1

AA: Price decreases with output: discount rate (monetary policy) effect dominates

AA: Price increase with output: cash flow effect dominates, compatible with liquidity trap.

AA: The cash flow effect first offset the discount rate effect, then, for high output, the monetary policy reaction dominates

Low equilibria: because increases of asset prices have low effect on output: very large change of

output prices to shift to good equilibrium

3 or 2 equilibria

In the former diagram, a linear AA curve leads to 2 equilibria in Krugman’s static model.

Extensions to a dynamical system: 3 equilibria are often such that 1 and 3 are stable and 2 unstable.

If low equilibrium 1 is stable: large external shock to shift above equilibrium 2, in order to converge to equilibrium 3.

Conclusion

Fourth GCM consider all assets prices, considering that exchange rate is one asset price among others (housing prices, equity prices).

Fourth GCM stress the case of monetary policy irrelevance in case of liquidity trap as for Japan in the 1990’s.

The major role of housing asset price

• The liquidity trap risk

Source : www.federalreserve.gov

Evolution des taux directeurs de la Fed

0

1

2

3

4

5

6

07/2

007

08/2

007

09/2

007

10/2

007

11/2

007

12/2

007

01/2

008

02/2

008

03/2

008

04/2

008

05/2

008

06/2

008

07/2

008

08/2

008

09/2

008

10/2

008

11/2

008

12/2

008

01/2

009

02/2

009

03/2

009

04/2

009

05/2

009

06/2

009

07/2

009

date

tau

x d

'inté

rêt(

%)

Taux à 3 mois

Taux à 6 mois

Taux à 1 mois

Taux à 1 an