Global Imbalances and Currency Wars at the ZLB

Ricardo Caballero1 Emmanuel Farhi2 Pierre-Olivier Gourinchas3

1MIT & NBER

2Harvard & NBER

3UC Berkeley & NBER

Pacific Basin Research Conference, San Francisco, November 2016

1 / 37

Global Imbalances

-2.50

-2.00

-1.50

-1.00

-0.50

0.00

0.50

1.00

1.50

2.00

2.50

1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010 2013

%OFWORLDGDP

U.S. EuropeanUnion Japan OilProducers EmergingAsiaex-China China Restoftheworld

Financial CrisisAsian Crisis Eurozone Crisis

Figure: Current Account, % of World GDP

2 / 37

Global Interest Rates (Short and Long)

0

5

10

15

20

25

1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010 2013

percent

U.S. Eurozone U.K. Japan

Financial Crisis Eurozone Crisis

(a) policy rates

0

2

4

6

8

10

12

14

16

18

1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010 2013

percent

U.S. Germany U.K. Japan

Financial Crisis Eurozone Crisis

(b) 10-year nominal yields

3 / 37

Output Gap (Advanced Economies), percent

-8

-6

-4

-2

0

2

4

6

1990 1993 1996 1999 2002 2005 2008 2011 2014

UnitedStates Eurozone Japan UnitedKingdom

Financial Crisis Eurozone Crisis

4 / 37

Global Exchange Rates

-50

-40

-30

-20

-10

0

10

20

30

2007 2008 2009 2010 2011 2012 2013 2014 2015

percent

Euro-dollar Yen-dollar Yuan-dollar

Abenomics ECB QE

The figure reports ln(E/E2007m1) where E denotes the foreign currency value of the dollar.

5 / 37

U.S. Interest Rate and Equity Risk Premium

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

1980 1985 1990 1995 2000 2005 2010 2015

Percentannualized

One-yearTreasuryyield OneyearaheadERP

Financial Crisis

Source: one-year Treasury yield: Federal Reserve H.15; ERP: Duarte & Rosa (2015).

6 / 37

Goal

I Simple model to shed light on these developments:

I transparent, parsimonious

I closed-form solutions

I Capital flows, exchange rates, unemployment and risk premia

I Away from, or at Zero Lower Bound (ZLB)

I Policy

7 / 37

Main Ideas

I ZLB tipping point for Global Imbalances (benign to malign):

I no ZLB → propagation of low interest rates via CA surpluses

I ZLB → propagation of recessions via CA surpluses

I Regime of increased policy interdependence (± spillovers):

I FX (zero sum)

I inflation targets (positive sum)

I government spending (positive sum)

I public debt issuance (positive sum)

I helicopter drops of money (positive sum)

I some forms of QE (positive sum)

8 / 37

Literature Review

Four strands of related literature:

I Asset shortages and global imbalances (Bernanke (2005), Caballeroet al (2008), Mendoza et al (2009))

I Liquidity traps in NK models & open economy (Keynes (1936),Krugman (1998), Eggertsson & Woodford (2003), Eggertsson &Krugman (2012), Werning (2012), Jeanne (2009), Cook & Devereux(2013), Benigno & Romei (2014))

I Secular Stagnation (Summers 2014), Caballero & Farhi (2015),Eggertsson & Mehrotra (2014))

I Safety and public debt (Stein (2012), Gorton & Ordonez (2014),Caballero & Farhi (2015), Barro and Mollerus (2015))

Closest paper to ours: Eggertsson, Mehrotra, Sing & Summers (2015).

9 / 37

Basic Model: Two Countries, no Risk

I Home and Foreign

I Endowment X of H good grows at rate g

I Endowment X ∗ of F good grows at rate g

I Relative size (constant): x = XX+X∗ .

10 / 37

Home Assets

I Dividends δX capitalized by Lucas trees:

I rate of depreciation ρ

I rate of new trees creation ρ

I Public debt D = dX financed by taxes τ

11 / 37

Home Agents

I OLG “perpetual youth”with birth/death Poisson rate θ;

I Earn income at birth, save it, and consume at death;

I Consumption shares on (H,F): (x , 1− x);

I Income of newborns: (1− τ)(1− δ)X+ value of new trees

12 / 37

Financial Development/Securitization Capacity

I Interpret δ as financial development/securitization capacity, notcapital share

I Only small part of capital income pledgeable to outside investors as“dividend” on tradable assets

I Depends on financial development/securitization capacity

I Interpret ρ as technological churn and expropriation risk

I Vt/PVt depends on δ and ρ

PVt =

∫ ∞t

Xse−

∫ strududs

Vt = δ

∫ ∞t

Xte−

∫ st

(ru+ρ)duds

13 / 37

Nominal Rigidities and Monetary Policy

I Competitive CES final good sector in each country

I Reinterpret endowment as non-traded input

I transformed into variety of intermediate good sold monopolistically

I H prices rigid in H currency, F prices rigid in F currency (PCP)

I accommodate demand at posted price

I Capacity utilization ξ ∈ [0, 1]

I Truncated Taylor rule: i = max{rn − ψ(1− ξ), 0}

I Real interest rate r = i

14 / 37

Foreign

Same as H but different parameters:

I Financial development/securitization capacity: δ∗ 6= δ

I Public debt to GDP ratio d∗ 6= d and taxes τ∗ 6= τ

I Other differences (extensions):I demographics and credit constraints (savers/borrowers)I securitization capacity & demand for safe assetsI inflation targets

15 / 37

Equilibrium Equations (along BGP)I Asset pricing (V : value of H trees in H currency)

rwV = −ρV + δξX

rwV ∗ = −ρV ∗ + δ∗ξ∗X ∗

I Wealth accumulation (W : H financial wealth in H currency):

W = gW = −θW + (1− δ)(1− τ)ξX + rwW + (ρ+ g)V

W ∗ = gW ∗ = −θW ∗ + (1− δ∗)(1− τ∗)ξ∗X ∗ + rwW ∗ + (ρ+ g)V ∗

I Government budget constraints:

(rw − g)D = τ(1− δ)ξX

(rw − g)D∗ = τ∗(1− δ∗)ξ∗X ∗

I Goods market clearing: (E : nominal exchange rate)

xθ(W + EW ∗) = ξX

(1− x)θ(W + EW ∗) = Eξ∗X ∗

16 / 37

ZLB “Complementary Slackness”

I No liquidity trap

rw > 0 and ξ = ξ∗ = 1

I Global liquidity trap

rw = 0 and ξ, ξ∗ ≤ 1

I All or none world

17 / 37

No Liquidity Trap

I World interest rate as “average” of autarky interest rates

rw = rw ,n = −ρ+δθ

1− θd

with

r a,n = −ρ+δθ

1− θdand r a,n∗ = −ρ+

δ∗θ

1− θd∗

I Net Foreign Assets and Current Account

NFA

X=

(1− θd)(rw − r a,n)

(g + θ − rw )(ρ+ rw )and

CA

X= g

NFA

X

I Exchange rateE = 1

18 / 37

Standard Metzler Diagram - Global

The global equilibrium interest rate rw is such that world financial markets are

in equilibrium: NFAX

= x NFAX

+ (1− x)NFA∗

X∗ = 0.19 / 37

Global Liquidity Trap

I World interest raterw = 0

I Fixed-point equations for ξ and ξ∗

ξ =θ

g + θ[xξ(1 +

gδ

ρ) + (1− x)Eξ∗(1 +

gδ∗

ρ) + xgd + (1− x)gd∗]

ξ∗ =1

E

θ

g + θ[xξ(1 +

gδ

ρ) + (1−x)Eξ∗(1 +

gδ∗

ρ) +xgd + (1−x)gd∗]

I Multiple equilibria indexed by E ...(Kareken-Wallace)

E =ξ

ξ∗

20 / 37

Global Liquidity Trap

I Output gaps as “FX-weighted averages” of autarky output gaps

ξ = x1− δθ

ρ

1− δθρ

ξa,l + (1− x)1− δ∗θ

ρ

1− δθρ

Eξa,l∗

ξ∗ = x1− δθ

ρ

1− δθρ

1

Eξa,l + (1− x)

1− δ∗θρ

1− δθρ

ξa,l∗

with

ξa,l = 1 +1− θd1− δθ

ρ

r a,n

ρand ξa,l∗ = 1 +

1− θd∗

1− δ∗θρ

r a,n∗

ρ

I Net Foreign Assets and Current Account

NFA

X=

(1− δθρ )(ξ − ξa,l)g + θ

andCA

X= g

NFA

X

21 / 37

Output Determination in the Global ZLB

figure reports Home (ξ) and Foreign (ξ∗) output at the global ZLB, for differentvalues of the exchange rate E ∈ [E, E ].

22 / 37

Metzler Diagram in Quantities - Global

Given E , ξ is such that world financial markets are in equilibrium:NFAX

(E) = x NFAX

+ (1− x)E NFA∗

X∗ = 0.

23 / 37

Currency Wars and Reserve Currency Paradox

I E determined by market coordination or FX intervention (peg)

I Beggar-thy-neighbor devaluations (zero-sum)

E ↑ =⇒ ξ ↑ ξ∗ ↓ CA

X↑

I Reserve currency paradox

24 / 37

Inflation

I ‘Old’ Keynesian Phillips curves (downward sticky prices )

[πH,t + κ0 + κ1(1− ξt)](1− ξt) = 0

[π∗F ,t + κ∗0 + κ∗1(1− ξ∗t )](1− ξ∗t ) = 0

I Taylor rules with inflation targets π > 0 and π∗ > 0

it = max{0, rnt + π + φ(πH,t − π)}

i∗t = max{0, rn∗t + π∗ + φ∗(π∗F ,t − π∗)}

25 / 37

Inflation

I With rw ,n < 0, multiple equilibria with different TOT: S =EP∗

F

PH

I No liquidity traps equilibrium (i > 0, i∗ > 0) if inflation targetshigh enough: rw ,n + min{π, π∗} > 0

I Global liquidity trap equilibrium (i = i∗ = 0) with deflationaryspiral

I at world level, more wage flexibility → deeper recessionI at country level, more wage flexibility → shallower recession

I Asymmetric liquidity trap equilibrium (i = 0, i∗ > 0)I no recession in one countryI worse recession in the other

I Inflation targets (positive sum) vs. FX interventions (zero sum)

26 / 37

Public Debt and Helicopter Drops of Money

I Public debt expansion (positive sum)...

d ↑ =⇒ ξ ↑ ξ∗ ↑ CA

X↓

I ...but not if used to finance asset purchases(different in model with safe and risky assets)

I Larger multiplier if higher private asset supply δ

I Equivalent to helicopter drops of money

27 / 37

Government Spending

I Government spending (positive sum)

G ↑ =⇒ ξ ↑ ξ∗ ↑ CA

X↓

I Domestic multiplier > 1 in SR(net asset supply boost + inflation boost through stimulus)

I More foreign leakage in LR(TOT appreciation)

28 / 37

More in Paper

I Home bias

I Non-unitary trade elasticities

I Borrowers and savers

I aging

I deleveraging

I Safe assets and global safe asset shortages (zoom in)

29 / 37

U.S. MPK

30 / 37

U.S. Interest Rate and Equity Risk Premium

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

1980 1985 1990 1995 2000 2005 2010 2015

Percentannualized

One-yearTreasuryyield OneyearaheadERP

Financial Crisis

Source: one-year Treasury yield: Federal Reserve H.15; ERP: Duarte & Rosa (2015).

31 / 37

Safe Asset Imbalances

-20

-15

-10

-5

0

5

10

15

20

1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010

%OFWORLDGDP

U.S. EuroArea Japan

OilProducers EmergingAsiaexChina China

Switzerland U.K. RestoftheWorld

Financial CrisisAsian Crisis Eurozone Crisis

Note: Net Safe positions defined as the sum of Official Reserves (minus Gold), Portfolio Debt and

Other Assets, minus Portfolio Debt and Other Liabilities. Source: Lane & Milesi-Ferretti (2007).

Regions defined as in Figure 1.

32 / 37

Safe Assets and Global Safe Asset Shortages

I Endogenous risk premia, increases at the ZLB

I Links reserve currency paradox and exorbitant privilege

I Can have ZLB in one country but not other (6= real interest rates)

I Policy:

I QE issue debt/purchase risky (not safe!) assets (positive sum)

I support private securitization capacity (positive sum)

I forward guidance (reduced effectiveness)

33 / 37

Safe Assets: Shocks and Preferences

I Disaster shock /w Poisson rate λ→ 0: output drops µ < 1

I Set d = d∗ = 0 and δ = δ∗

I Fraction α ‘Knightians’ (infinitely risk averse), 1− α Risk Neutral.

I Knightians have full home bias.

I Neutrals have ‘some’ home bias

34 / 37

Safe Assets: Securitization & Tranching

I Fraction φ < 1 of H dividend tranched and recombined.:I Poisson puts (pay nothing until Poisson shock)I Poisson calls (pay only until the Poisson shock)

I Knightians invest in safe assets combining puts and calls

I Neutrals invest in the rest

I Constrained regime: safe assets are scarce & Knightians price safeassets at the margin (safety premium).

35 / 37

Modified UIP and Risk Premia

I Fix exchange rate immediately after the shock E+

I No-arbitrage requires:

rw − rK

rw − rK∗=

E

E+

I modified UIP equation: the country with a high safety premium(rK < rK∗) has a currency that will appreciate when the shockoccurs (E > E+).

I Reserve Currency Paradox: if Home’s currency is expected toappreciate in bad times (E > E+), then rK < rK∗ and Home ismore likely to experience a liquidity trap

I if φ > φ∗ then NFA/X < 0: exorbitant privilege.

I Metzler diagram in safe assets

36 / 37

Conclusion

This paper:

I Model of global and local, permanent or persistent liquidity traps(secular stagnation)

I Traps in one country propagate to other countries

I Powerful beggar-thy-neighbor effects vis FX

I Model accounts for decline in risk-free rate and increase in riskpremia

I Paradox of the reserve currency: reserve countries suffer adisproportionate share of the trap

37 / 37