The Dollar Hegemon? Evidence and Implications for Policymakers - - PowerPoint PPT Presentation

The Dollar Hegemon?

Evidence and Implications for Policymakers

Pierre-Olivier Gourinchas UC Berkeley, NBER and CEPR

pog@berkeley.edu

Annual Conference Chaire Banque de France - Paris School of Economics Paris, September 2019

1 / 29

A Dollar World

We live in a dollar world

• Global trade is invoiced in dollars
• Cross-border financial flows and security issuances are in dollars
• Monetary authorities anchor to the dollar
• International reserves are held in dollars

Even more so now than at any time in the past!

2 / 29

A Dollar World

The dollar plays an essential role in three areas:

• 1. The expansion of global trade and capital markets
• 2. The coordination of macroeconomic policies
• 3. The provision of international liquidity

This largely shapes the policy choices of the rest of the world

3 / 29

Outline of the Talk

• 1. The Emergence of the Dollar Hegemon
• 2. Implications for Policymakers
• 3. Directions for the Future

4 / 29

The Dollar Hegemon

Charles Kindleberger (1981)

‘I argue(d) that for the world economy to be stable, it needs a stabilizer, some country that would undertake to provide a market for distress goods, a steady if not counter- cyclical flow of capital, and a rediscount mechanism for providing liquidity when the monetary system is frozen in panic. [...] Britain, with frequent assistance from France, furnished coherence to the world economy along these lines during the nineteenth century and through the “belle ´ epoque.” The United States did so from 1945 (or perhaps 1936) to 1968 (or 1963 or 1971).’ Viewed the collapse of BW as the end of the era of uncontested US dominance...

5 / 29

The Emergence of the Dollar Hegemon

The opposite happened: “[After BW], the dollar assumed greater monetary importance than gold.” Mundell (1973). Increased dollar dominance:

• For trade invoicing and settlement
• For securities issuance and holdings
• As an anchor and intervention currency

6 / 29

Trade Invoicing

• 1. dollar’s invoicing share ≈ 3.5 share in trade. Gopinath (2015), Goldberg & Tille (2009).

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Colombia United States Argentina Brazil Pakistan Indonesia India South Korea Thailand Malta Ukraine Canada Israel Australia Peru Japan Turkey Lithuania Cyprus Bulgaria Greece Norway Iceland United Kingdom Slovenia Latvia Finland Spain Netherlands Portugal Estonia Sweden Denmark Ireland Italy Poland Romania Hungary Germany France Czech Republic Switzerland Belgium Slovakia Austria

Ratio to Exports and Imports Imports USD Exports USD

7 / 29

Financial Transactions

• 2. The use of the dollar in global markets. Eichengreen and Xia (2019)

44 16 11 6 2 63 22 3 59 21 4 62 20 5 4 2 40 34 4 7 2

10 20 30 40 50 60 70 USD EUR JPY GBPRMB USD EUR JPY RMB USD EUR JPY RMB USD EUR JPY GBPRMB USD EUR JPY GBPRMB Foreign exchange turnover International debts International loans Foreign exchange reserves Global payment currency

percentage

8 / 29

Financial Transactions: Currency Home Bias

• 3. Share of Corporate Bonds in Issuer’s Local Currency. Maggiori et al (2019).

The red bar reports for each issuing country, the share of bonds denominated in the issuer’s local currency out of all domestic investment in its corporate bonds. The hollow blue bars show, for each issuing country, the share of bonds denominated in the issuer’s local currency out of all foreign investment in its corporate bonds.

9 / 29

Financial Transactions: Dollar Funding

• 4. International Credit to Non-Residents by Currency. BIS global liquidity indicators.

1 2 3 4 5 6 7 8 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018

Percent of Global Output USD EUR JPY USD EUR JPY Bank Loans International Debt Securities

10 / 29

The Geography of Anchor Currencies, 1950

• 5. The spread of the dollar as an anchor currency. Ilzetski, Reinhart and Rogoff (2019).

11 / 29

The Geography of Anchor Currencies, 2015

• 5. The spread of the dollar as an anchor currency. Ilzetski, Reinhart and Rogoff (2019).

12 / 29

Important Complementarities

• [invoicing → anchor] Dollar invoicing generates asymmetric spillovers and makes dollar

anchoring more desirable (Egorov & Mukhin, 2019)

• [anchor → invoicing] Dollar anchoring makes dollar pricing more desirable compared to

producer or local pricing (Mukhin, 2019).

• [funding → anchor] Dollar funding increases spillovers from US monetary policy and

makes dollar anchoring more desirable (global financial cycle, Rey 2013).

• [anchor → funding] Lower dollar volatility makes dollar funding more desirable.
• [anchor → reserves] Dollar anchoring requires larger dollar FX reserve holdings
• [invoicing ↔ funding ↔ holdings] Dollar invoicing makes dollar funding and dollar

holdings more desirable and vice versa (Gopinath & Stein 2018) Lock-in effect, in favor of the dollar

13 / 29

The Collapse of Bretton Woods Revisited

Yet, the current era of dollar hegemony started with a run on the dollar. Why? Three separate dynamics at play:

• 1. The collapse of the dollar-gold parity in 1968. Original Triffin dilemma.
• 2. A short term adjustment: depreciation of the dollar due to excess dollar liquidity in

1971-73. dollar glut

• 3. Deeper low frequency dynamics: with increased capital mobility and volatile exchange

rates, demand for safe assets relocated from local assets (‘shadow’ dollar safe assets) to dollar assets. Potential modern Triffin dilemma. Key insight: center country is a net provider of the global safe asset (Caballero et al, 2019)

14 / 29

Asian Countries and the Dollar Anchor

Country Start of the dollar anchor Previous Anchor Philippines 1945 none Japan July 1947 none Korea February 1953 none Indonesia April 1969 none Pakistan September 1971 UK pound Singapore June 1972 UK pound Hong Kong July 1972 UK pound China (Mainland) January 1974 none Malaysia September 1975 UK pound India August 1979 UK pound Vietnam January 1992 none Cambodia January 1994 none

Source: Ilzetzki et al (2019). The table reports the date at which latest dollar anchoring begins.

15 / 29

Asian Countries and the Dollar Anchor

Country Start of the dollar anchor Previous Anchor Philippines 1945 none Japan July 1947 none Korea February 1953 none Indonesia April 1969 none Pakistan September 1971 UK pound Singapore June 1972 UK pound Hong Kong July 1972 UK pound China (Mainland) January 1974 none Malaysia September 1975 UK pound India August 1979 UK pound Vietnam January 1992 none Cambodia January 1994 none

Source: Ilzetzki et al (2019). The table reports the date at which latest dollar anchoring begins.

Was the UK pound anchor a shadow US dollar anchor?

15 / 29

Implications for Policymakers

Implications for Policymakers

Five key implications:

1 Exchange rate passthrough and the effectiveness of flexible exchange rates 2 Dollar factor and global trade 3 The Trilemma vs Dilemma debate 4 Global real rates, safety traps and currency wars 5 Exorbitant Privilege and Exorbitant Duty 16 / 29

1.a Dollar price passthrough and trade elasticity

import prices import quantities (1) (2) (3) (4) ∆pij,t ∆pij,t ∆yij,t ∆yij,t ∆eij,t 0.757*** 0.164***

• 0.119***
• 0.0310*

(0.0132) (0.0126) (0.0139) (0.0160) ∆e$j,t 0.781***

• 0.186***

(0.0143) (0.0250) R-squared 0.356 0.398 0.069 0.071 Observations 46,820 46,820 52,272 52,272 Dyads 2,647 2,647 2,807 2,807

Source: Gopinath et al (2019).

“Dominant Currency Paradigm”: Weak expenditure switching (via imports only), flat terms

• f trade, less effective depreciation

17 / 29

1.b Inflation-Output Trade-Off

5 10 15 20

• 0.05

0.05 0.1 0.15 0.2 0.25

DCP PCP LCP

(a) Inflation

5 10 15 20 0.2 0.4 0.6 0.8 1

DCP PCP LCP

(b) Output Impulse response to a 25bps domestic monetary policy easing in a Small Open Economy under different invoicing regimes: Producer Currency Pricing (PCP, dashed-black), Local Currency Pricing (LCP, dashed blue) and Dominant Currency Pricing (DCP, solid). Source: Gopinath et al (2019).

18 / 29

• 2. Global Trade and the Dollar

−1 −.8 −.6 −.4 −1 −.8 −.6 −.4 1 2

unweighted trade−weighted

• cumul. response to 1% shock, percent

years after shock

Source: Gopinath et al (2019). Error bars: 95% confidence intervals, clustering by dyad.

• Bilateral trade not driven by bilateral

exchange rate, but by the dollar exchange rate.

• 1% appreciation of the dollar associated

with 0.6-0.8% decline in the volume of trade between countries in the ROW within a year.

19 / 29

• 3. Trilemma/Dilemma Debate
• The importance of dollar funding makes balance sheets worldwide sensitive to the

dollar and indirectly to US monetary policy (Rey 2013)

• Question: Does this make flexible exchange rates less desirable (Dilemma)?
• Answer: it depends on the strength of balance sheet effects. (Gourinchas 2018)
• Weak balance sheet effects. Standard Mundell/Friedman world.

Flexible exchange rate is desirable.

• Intermediate balance sheet effects: US tightening is contractionary, but domestic easing is

expansionary. Flexible exchange rate is more desirable.

• Strong balance sheet effects: US tightening is contractionary, and so is domestic easing.

Flexible exchange rate may become less desirable.

The last case is theoretically possible, but requires perverse transmission of monetary policy

20 / 29

• 3. Trilemma/Dilemma Debate

5 10 15

• 0.01
• 0.005

0.005 0.01 0.015 0.02 Output Manuf Total

(a) Weak

5 10 15

• 0.02
• 0.015
• 0.01
• 0.005

0.005 0.01 Output Manuf Total

(b) Intermediate

5 10 15

• 0.4
• 0.3
• 0.2
• 0.1

0.1 0.2 Output Manuf Total

(c) High Spillovers Impulse Response of EME output to a US Monetary Policy Tightening.

5 10 15

• 0.06
• 0.04
• 0.02

0.02 Output Manuf Total

(d) Weak

5 10 15

• 0.04
• 0.03
• 0.02
• 0.01

0.01 0.02 Output Manuf Total

(e) Intermediate

5 10 15

• 0.2
• 0.15
• 0.1
• 0.05

0.05 0.1 Output Manuf Total

(f) High Spillovers Impulse Response of EME output to a local Monetary Policy Tightening.

21 / 29

• 3. Trilemma/Dilemma Debate

5 10 15

• 0.01
• 0.005

0.005 0.01 0.015 0.02 Output Manuf Total

(a) Weak

5 10 15

• 0.02
• 0.015
• 0.01
• 0.005

0.005 0.01 Output Manuf Total

(b) Intermediate

5 10 15

• 0.4
• 0.3
• 0.2
• 0.1

0.1 0.2 Output Manuf Total

(c) High Spillovers Impulse Response of EME output to a US Monetary Policy Tightening.

5 10 15

• 0.06
• 0.04
• 0.02

0.02 Output Manuf Total

(d) Weak

5 10 15

• 0.04
• 0.03
• 0.02
• 0.01

0.01 0.02 Output Manuf Total

(e) Intermediate

5 10 15

• 0.2
• 0.15
• 0.1
• 0.05

0.05 0.1 Output Manuf Total

(f) High Spillovers Impulse Response of EME output to a local Monetary Policy Tightening.

Perverse Transmission of Monetary Policy

21 / 29

• 4. Global Real Rates, Safety Traps, Currency Wars
• Global safe assets are scarce. How do we know this? Because global real rates are low

and declining, while various risk premia remain elevated.

• Low levels of the safe global real rate brings the global economy up against the

effective lower bound

• Three major implications:
• This maintains the global economy on the edge of a regime of permanently depressed

demand: secular stagnation, liquidity or safety traps....

• This creates strong incentives for countries to attempt to reflate their economy at the

expense of their neighbors: Trade and currency wars.

• Policies that are optimal locally can be destabilizing globally: accumulating reserves, fiscal

austerity, excessively strict liquidity requirements on the financial sector.

Caballero, Farhi & Gourinchas (2019), Eggertsson et al (2017), Fornaro & Romei (2019).

22 / 29

The New Triffin Dilemma

0% 5% 10% 15% 20% 25%

1980 1984 1988 1992 1996 2000 2004 2008 2012 2016 2020 2024

Share of World Output

United States Euro (12) Japan United Kingdom China India EM-Asia (ex China and India) Projection

World Output Shares. Source: World Economic Outlook, April 2019. Output shares based on PPP.

23 / 29

• 5. Exorbitant Duty, Exorbitant Privilege

The US has a leveraged external balance sheet, long risky assets, short safe assets.

• 1. It collects an expected excess return (‘exorbitant privilege’)
• 2. It suffers outsized losses during times of global crisis, i.e. when resources are globally

scarce (‘exorbitant duty’)

• 3. The USD appreciates in times of global crisis (negative β)

Facts 1 and 2 are consistent with simple risk sharing model: cross country differences in indirect utility risk appetite (Gourinchas, Rey & Govillot 2019, Maggiori 2017). Fact 3 is harder. Standard model would suggest real depreciation in bad times. Suggests demand for insurance rises in bad times: risk on/off. Gourinchas, Rey and Govillot (2019)

24 / 29

A Simple Model of Risk Sharing with Risk On/Off Behavior

Gourinchas, Rey and Govillot (2019)

• Two countries: Home (US) and Foreign (*)
• Endowment: y T (traded) and y N (non-traded). RER as relative price of T and NT
• 3 states j: safe (s), fragile (f ) and disaster (d).

state d: decline in global output by factor 1 − b. ps→d < pf →d

• Differences in risk appetite: CRRA σ(j) ≤ σ∗(j)
• Risk On/Off:

σ(s) ≤ σ(f , d) ; σ∗(s) ≤ σ∗(f , d)

25 / 29

Estimating the Model

Estimate parameters of the model, including σ(j), σ∗(j), pj→j′ (total 12 parameters) to match 13 macro moments:

• unconditional probabilities in different states (Barro)
• equity risk premium
• price equity ratio
• trade balance/output (s)
• net equity /output (s)
• gross external assets/output (s)
• net foreign asset position/output (s)
• s.d. log output (detrended) (s)
• s.d. trade balance/output (s)
• appreciation of the RER s → f

26 / 29

Exorbitant Privilege, Exorbitant Duty

Moment Data Estimated Moment Data Estimated ¯ p(s) 0.80 0.79 ¯ p(f ) 0.15 0.14 Risk Premium (s, f ) (%) 4.00 5.39 PE(s, f ) 20 20.13 TB/Y (s) (%)

• 4.66
• 1.73

Net Equ./Y (s) (%) 7.59 7.65

• Ext. Assets/Y (s) (%)

78.92 78.96 NFA/Y (s) (%)

• 18.13
• 18.10

s.d. ln Y (s) 1.0795 1.723 s.d. TB/Y (s) 0.408 0.582 ∆ ln RERs→f (%) 8.00 7.74 ERNA 2.00 1.71 ∆VAL/Y (s → f ) (%)

• 14.28
• 11.92

27 / 29

Exorbitant Privilege, Exorbitant Duty

Moment Data Estimated Moment Data Estimated ¯ p(s) 0.80 0.79 ¯ p(f ) 0.15 0.14 Risk Premium (s, f ) (%) 4.00 5.39 PE(s, f ) 20 20.13 TB/Y (s) (%)

• 4.66
• 1.73

Net Equ./Y (s) (%) 7.59 7.65

• Ext. Assets/Y (s) (%)

78.92 78.96 NFA/Y (s) (%)

• 18.13
• 18.10

s.d. ln Y (s) 1.0795 1.723 s.d. TB/Y (s) 0.408 0.582 ∆ ln RERs→f (%) 8.00 7.74 ERNA 2.00 1.71 ∆VAL/Y (s → f ) (%)

• 14.28
• 11.92

27 / 29

Exorbitant Privilege, Exorbitant Duty

Parameter Estimated Value Parameter Estimated Value σ(s) 2.68 σ∗(s) 2.68 σ(f , d) 2.68 σ∗(f , d) 5.69 ps→f 0.0229 ps→d 0.0086 pf →d 0.0280 σy(s) 0.0172

28 / 29

Exorbitant Privilege, Exorbitant Duty

Parameter Estimated Value Parameter Estimated Value σ(s) 2.68 σ∗(s) 2.68 σ(f , d) 2.68 σ∗(f , d) 5.69 ps→f 0.0229 ps→d 0.0086 pf →d 0.0280 σy(s) 0.0172

28 / 29

Exorbitant Privilege, Exorbitant Duty

Parameter Estimated Value Parameter Estimated Value σ(s) 2.68 σ∗(s) 2.68 σ(f , d) 2.68 σ∗(f , d) 5.69 ps→f 0.0229 ps→d 0.0086 pf →d 0.0280 σy(s) 0.0172

• Model finds strong asymmetric risk On/Off.
• Matches Appreciation of the USD in bad times.
• Recovers Exorbitant privilege & Exorbitant duty (unmatched moments)
• Can account for decline in r ∗ (share of ROW)

28 / 29

Implications for the Future

Implications for the Future

I see three main implications

• 1. In the short to medium run, the dollar standard is here to stay, however dysfunctional:

Safe asset scarcity, low real rates, depressed global demand, monetary policy spillovers and the potential for FX or trade disruption.

• 2. In the (very) long run, the dollar dominance is bound to disappear.

Modern Triffin dilemma: eventually, the US will become too small and unable to expand the supply of safe dollar assets without exhausting its fiscal capacity.

• 3. The transition, when it happens, could be quite sudden.

As long as the RMB-USD rate remains stable, a dollar anchor could be a shadow RMB

• anchor. US dollar anchors could go the way of the UK pound anchors...

29 / 29

Thank You!